Bridget E. Shields, MD

Vitamin D (VD) regulates keratinocyte proliferation and differentiation, modulates inflammatory pathways, and protects against cellular damage in the skin. ¹ In the setting of tissue injury and acute skin inflammation, active vitamin D—1,25(OH) ₂ D—suppresses signaling from pro-inflammatory chemokines and cytokines such as IFN- γ and IL-17. ^2,3 This suppression reduces proliferation of helper T cells (T _H 1, T _H 17) and B cells, decreasing tissue damage from reactive oxygen species release while enhancing secretion of the anti-inflammatory cytokine IL-10 by antigen-presenting cells. ^2-4

Suboptimal VD levels have been associated with numerous health consequences including malignancy, prompting interest in VD supplementation for improving cancer-related outcomes.⁵ Beyond disease prognosis, high-dose VD supplementation has been suggested as a potential therapy for adverse events (AEs) related to cancer treatments. In one study, mice that received oral vitamin D₃ supplementation of 11,500 IU/kg daily had fewer doxorubicin-induced cardiotoxic effects on ejection fraction (P<.0001) and stroke volume (P<.01) than mice that received VD supplementation of 1500 IU/kg daily.⁶

In this review, we examine the impact of chemoradiation on 25(OH)D levels—which more accurately reflects VD stores than 1,25(OH)₂D levels—and the impact of suboptimal VD on cutaneous toxicities related to chemoradiation. To define the suboptimal VD threshold, we used the Endocrine Society’s clinical practice guidelines, which characterize suboptimal 25(OH)D levels as insufficiency (21–29 ng/mL [52.5–72.5 nmol/L]) or deficiency (<20 ng/mL [50 nmol/L])⁷; deficiency can be further categorized as severe deficiency (<12 ng/mL [30 nmol/L]).⁸ This review also evaluates the evidence for vitamin D₃ supplementation to alleviate the cutaneous AEs of chemotherapy and radiation treatments.

Effects of Chemotherapy on Vitamin D Levels

A high prevalence of VD deficiency is seen in various cancers. In a retrospective review of 25(OH)D levels in 2098 adults with solid tumors of any stage (6% had metastatic disease [n=124]), suboptimal levels were found in 69% of patients with breast cancer (n=617), 75% with colorectal cancer (n=84), 72% with gynecologic cancer (n=65), 79% with kidney and bladder cancer (n=145), 83% with pancreatic and upper gastrointestinal tract cancer (n=178), 73% with lung cancer (n=73), 69% with prostate cancer (n=225), 61% with skin cancer (n=399), and 63% with thyroid cancer (n=172).⁵ Suboptimal VD also has been found in hematologic malignancies. In a prospective cohort study, mean serum 25(OH)D levels in 23 patients with recently diagnosed acute myeloid leukemia demonstrated VD deficiency (mean [SD], 18.6 [6.6] nmol/L).⁹ Given that many patients already exhibit a baseline VD deficiency at cancer diagnosis, it is important to understand the relationship between VD and cancer treatment modalities.⁵

In the United States, breast and colorectal cancers were estimated to be the first and fourth most common cancers, with 313,510 and 152,810 predicted new cases in 2024, respectively.¹⁰ This review will focus on breast and colorectal cancer when describing VD variation associated with chemotherapy exposure due to their high prevalence.

Effects of Chemotherapy on Vitamin D Levels in Breast Cancer—Breast cancer studies have shown suboptimal VD levels in 76% of females 75 years or younger with any T1, T2, or T3; N0 or N1; and M0 breast cancer, in which 38.5% (n=197) had insufficient and 37.5% (n=192) had deficient 25(OH)D levels.¹¹ In a study of female patients with primary breast cancer (stage I, II, or III and T1 with high Ki67 expression [≥30%], T2, or T3), VD deficiency was seen in 60% of patients not receiving VD supplementation.^12,13 A systematic review that included 7 studies of different types of breast cancer suggested that circulating 25(OH)D may be associated with improved prognosis.¹⁴ Thus, studies have investigated risk factors associated with poor or worsening VD status in individuals with breast cancer, including exposure to chemotherapy and/or radiation treatment.^12,15-18

A prospective cohort study assessed 25(OH)D levels in 95 patients with any breast cancer (stages I, II, IIIA, IIIB) before and after initiating chemotherapy with docetaxel, doxorubicin, epirubicin, 5-fluorouracil, or cyclophosphamide, compared with a group of 52 females without cancer.¹⁷ In the breast cancer group, approximately 80% (76/95) had suboptimal and 50% (47/95) had deficient VD levels before chemotherapy initiation (mean [SD], 54.1 [22.8] nmol/L). In the comparison group, 60% (31/52) had suboptimal and 30% (15/52) had deficient VD at baseline (mean [SD], 66.1 [23.5] nmol/L), which was higher than the breast cancer group (P=.03). A subgroup analysis excluded participants who started, stopped, or lacked data on dietary supplements containing VD (n=39); in the remaining 56 participants, a significant decrease in 25(OH)D levels was observed shortly after finishing chemotherapy compared with the prechemotherapy baseline value (mean, −7.9 nmol/L; P=.004). Notably, 6 months after chemotherapy completion, 25(OH)D levels increased (mean, +12.8 nmol/L; P<.001). Vitamin D levels remained stable in the comparison group (P=.987).¹⁷

Consistent with these findings, a cross-sectional study assessing VD status in 394 female patients with primary breast cancer (stage I, II, or III and T1 with high Ki67 expression [≥30%], T2, or T3), found that a history of chemotherapy was associated with increased odds of 25(OH)D levels less than 20 ng/mL compared with breast cancer patients with no prior chemotherapy (odds ratio, 1.86; 95% CI, 1.03-3.38).¹² Although the study data included chemotherapy history, no information was provided on specific chemotherapy agents or regimens used in this cohort, limiting the ability to detect the drugs most often implicated.

Both studies indicated a complex interplay between chemotherapy and VD levels in breast cancer patients. Although Kok et al¹⁷ suggested a transient decrease in VD levels during chemotherapy with a subsequent recovery after cessation, Fassio et al¹² highlighted the increased odds of VD deficiency associated with chemotherapy. Ultimately, larger randomized controlled trials are needed to better understand the relationship between chemotherapy and VD status in breast cancer patients.

Effects of Chemotherapy on Vitamin D Levels in Colorectal Cancer—Similar to patterns seen in breast cancer, a systematic review with 6 studies of different types of colorectal cancer suggested that circulating 25(OH)D levels may be associated with prognosis.¹⁴ Studies also have investigated the relationship between colorectal chemotherapy regimens and VD status.^15,16,18,19

A retrospective study assessed 25(OH)D levels in 315 patients with any colorectal cancer (stage I–IV).¹⁵ Patients were included in the analysis if they received less than 400 IU daily of VD supplementation at baseline. For the whole study sample, the mean (SD) VD level was 23.7 (13.71) ng/mL. Patients who had not received chemotherapy within 3 months of the VD level assessment were categorized as the no chemotherapy group, and the others were designated as the chemotherapy group; the latter group was exposed to various chemotherapy regimens, including combinations of irinotecan, oxaliplatin, 5-fluorouracil, leucovorin, bevacizumab, or cetuximab. Multivariate analysis showed that the chemotherapy group was 3.7 times more likely to have very low VD levels (≤15 ng/mL) compared with those in the no chemotherapy group (P<.0001).¹⁵

A separate cross-sectional study examined serum 25(OH)D concentrations in 1201 patients with any newly diagnosed colorectal carcinoma (stage I–III); 91% of cases were adenocarcinoma.¹⁸ In a multivariate analysis, chemotherapy plus surgery was associated with lower VD levels than surgery alone 6 months after diagnosis (mean, −8.74 nmol/L; 95% CI, −11.30 to −6.18 nmol/L), specifically decreasing by a mean of 6.7 nmol/L (95% CI, −9.8 to −3.8 nmol/L) after adjusting for demographic and lifestyle factors.¹⁸ However, a prospective cohort study demonstrated different findings.¹⁹ Comparing 58 patients with newly diagnosed colorectal adenocarcinoma (stages I–IV) who underwent chemotherapy and 36 patients who did not receive chemotherapy, there was no significant change in 25(OH)D levels from the time of diagnosis to 6 months later. Median VD levels decreased by 0.7 ng/mL in those who received chemotherapy, while a minimal (and not significant) increase of 1.6 ng/mL was observed in those without chemotherapy intervention (P=.26). Notably, supplementation was not restricted in this cohort, which may have resulted in higher VD levels in those taking supplements.¹⁹

Since time of year and geographic location can influence VD levels, one prospective cohort study controlled for differential sun exposure due to these factors in their analysis.¹⁶ Assessment of 25(OH)D levels was completed in 81 chemotherapy-naïve cancer patients immediately before beginning chemotherapy as well as 6 and 12 weeks into treatment. More than 8 primary cancer types were represented in this study, with breast (34% [29/81]) and colorectal (14% [12/81]) cancer being the most common, but the cancer stages of the participants were not detailed. Vitamin D levels decreased after commencing chemotherapy, with the largest drop occurring 6 weeks into treatment. From the 6- to 12-week end points, VD increased but remained below the original baseline level (baseline: mean [SD], 49.2 [22.3] nmol/L; 6 weeks: mean [SD], 40.9 [19.0] nmol/L; 12 weeks: mean [SD], 45.9 [19.7] nmol/L; P=.05).¹⁶

Although focused on breast and colorectal cancers, these studies suggest that various chemotherapy regimens may confer a higher risk for VD deficiency compared with VD status at diagnosis and/or prior to chemotherapy treatment. However, most of these studies only discussed stage-based differences, excluding analysis of the variety of cancer subtypes that comprise breast and colorectal malignancies, which may limit our ability to extrapolate from these data. Ultimately, larger randomized controlled trials are needed to better understand the relationship between chemotherapy and VD status across various primary cancer types.

Effects of Radiation Therapy on Vitamin D Levels

Unlike chemotherapy, studies on the association between radiation therapy and VD levels are minimal, with most reports in the literature discussing the use of VD to potentiate the effects of radiation therapy. In one cross-sectional analysis of 1201 patients with newly diagnosed stage I, II, or III colorectal cancer of any type (94% were adenocarcinoma), radiation plus surgery was associated with slightly lower 25(OH)D levels than surgery alone for tumor treatment 6 months after diagnosis (mean, −3.17; 95% CI, −6.07 to −0.28 nmol/L). However, after adjustment for demographic and lifestyle factors, this decrease in VD levels attributable to radiotherapy was not statistically significant compared with the surgery-only cohort (mean, −1.78; 95% CI, −5.07 to 1.52 nmol/L).¹⁸

Similarly, a cross-sectional study assessing VD status in 394 female patients with primary breast cancer (stage I, II, or III and T1 with high Ki67 expression [≥30%], T2, or T3), found that a history of radiotherapy was not associated with a difference in serum 25(OH)D levels compared with those with breast cancer without prior radiotherapy (odds ratio, 0.90; 95% CI, 0.52-1.54).¹² From the limited existing literature specifically addressing variations of VD levels with radiation, radiation therapy does not appear to significantly impact VD levels.

Vitamin D Levels and the Severity of Chemotherapy- or Radiation Therapy–Induced AEs

A prospective cohort of 241 patients did not find an increase in the incidence or severity of chemotherapy-induced cutaneous toxicities in those with suboptimal 1,25(OH)₂D₃ levels (≤75 nmol/L).²⁰ Eight different primary cancer types were represented, including breast and colorectal cancer; the tumor stages of the participants were not detailed. Forty-one patients had normal 1,25(OH)₂D₃ levels, while the remaining 200 had suboptimal levels. There was no significant association between serum VD levels and the following dermatologic toxicities: desquamation (P=.26), xerosis (P=.15), mucositis (P=.30), or painful rash (P=.87). Surprisingly, nail changes and hand-foot reactions occurred with greater frequency in patients with normal VD levels (P=.01 and P=.03, respectively).²⁰ Hand-foot reaction is part of the toxic erythema of chemotherapy (TEC) spectrum, which is comprised of a range of cytotoxic skin injuries that typically manifest within 2 to 3 weeks of exposure to the offending chemotherapeutic agents, often characterized by erythema, pain, swelling, and blistering, particularly in intertriginous and acral areas.^21-23 Recovery from TEC generally takes at least 2 to 4 weeks and may necessitate cessation of the offending chemotherapeutic agent.^21,24 Notably, this study measured 1,25(OH)₂D₃ levels instead of 25(OH)D levels, which may not reliably indicate body stores of VD.^7,20 These results underscore the complex nature between chemotherapy and VD; however, VD levels alone do not appear to be a sufficient biomarker for predicting chemotherapy-associated cutaneous AEs.

Interestingly, radiation therapy–induced AEs may be associated with VD levels. A prospective cohort study of 98 patients with prostate, bladder, or gynecologic cancers (tumor stages were not detailed) undergoing pelvic radiotherapy found that females and males with 25(OH)D levels below a threshold of 35 and 40 nmol/L, respectively, were more likely to experience higher Radiation Therapy Oncology Group (RTOG) grade acute proctitis compared with those with VD above these thresholds.²⁵ Specifically, VD below these thresholds was associated with increased odds of RTOG grade 2 or higher radiation-induced proctitis (OR, 3.07; 95% CI, 1.27-7.50 [P=.013]). Additionally, a weak correlation was noted between VD below these thresholds and the RTOG grade, with a Spearman correlation value of −0.189 (P=.031).²⁵

One prospective cohort study included 28 patients with any cancer of the oral cavity, oropharynx, hypopharynx, or larynx stages II, III, or IVA; 93% (26/28) were stage III or IVA.²⁶ The 20 (71%) patients with suboptimal 25(OH)D levels (≤75 nmol/L) experienced a higher prevalence of grade II radiation dermatitis compared with the 8 (29%) patients with optimal VD levels (χ²₂=5.973; P=.0505). This pattern persisted with the severity of mucositis; patients from the suboptimal VD group presented with higher rates of grades II and III mucositis compared with the VD optimal group (χ²₂=13.627; P=.0011).²⁶ Recognizing the small cohort evaluated in the study, we highlight the importance of further studies to clarify these associations.

Chemotherapy-Induced Cutaneous Events Treated with High-Dose Vitamin D

Chemotherapeutic agents are known to induce cellular damage, resulting in a range of cutaneous AEs that can invoke discontinuation of otherwise effective chemotherapeutic interventions.^27,28 Recent research has explored the potential of high-dose vitamin D₃ as a therapeutic agent to mitigate cutaneous reactions.^29,30

A randomized, double-blind, placebo-controlled trial investigated the use of a single high dose of oral ­25(OH)D to treat topical nitrogen mustard (NM)–induced rash.²⁹ To characterize baseline inflammatory responses from NM injury without intervention, clinical measures, serum studies, and tissue analyses from skin biopsies were performed on 28 healthy adults after exposure to topical NM—a chemotherapeutic agent classified as a DNA alkylator. Two weeks later, participants were exposed to topical NM a second time and were split into 2 groups: 14 patients received a single 200,000-IU dose of oral 25(OH)D while the other 14 participants were given a placebo. Using the inflammatory markers induced from baseline exposure to NM alone, posttreatment analysis revealed that the punch biopsies from the 25(OH)D group expressed fewer NM-induced inflammatory markers compared with the placebo group at both 72 hours and 6 weeks following NM injury (72 hours: 12 vs 17 inflammatory markers; 6 weeks: 4 vs 11 inflammatory markers). Notably, NM inflammatory markers were enriched for IL-17 signaling pathways in the placebo biopsies but not in the 25(OH)D intervention group. This study also identified mild and severe patterns of inflammatory responses to NM that were independent of the 25(OH)D intervention. Biomarkers specific to skin biopsies from participants with the severe response included CCL20, CCL2, and CXCL8 (adjusted P<.05). At 6 weeks posttreatment, the 25(OH)D group showed a 67% reduction in NM injury markers compared with a 35% reduction in the placebo group. Despite a reduction in tissue inflammatory markers, there were no clinically significant changes observed in skin redness, swelling, or histologic structure when comparing the 25(OH)D- supplemented group to the placebo group at any time during the study, necessitating further research into the mechanistic roles of high doses VD supplementation.²⁹

Although Ernst et al²⁹ did not observe any clinically significant improvements with VD treatment, a case series of 6 patients with either glioblastoma multiforme, acute myeloid leukemia, or aplastic anemia did demonstrate clinical improvement of TEC after receiving high-dose vitamin D₃.³⁰ The mean time to onset of TEC was noted at 8.5 days following administration of the inciting chemotherapeutic agent, which included combinations of anthracycline, antimetabolite, kinase inhibitor, B-cell lymphoma 2 inhibitor, purine analogue, and alkylating agents. A combination of clinical and histologic findings was used to diagnose TEC. Baseline 25(OH)D levels were not established prior to treatment. The treatment regimen for 1 patient included 2 doses of 50,000 IU of VD spaced 1 week apart, totaling 100,000 IU, while the remaining 5 patients received a total of 200,000 IU, also split into 2 doses given 1 week apart. All patients received their first dose of VD within a week of the cutaneous eruption. Following the initial VD dose, there was a notable improvement in pain, pruritus, or swelling by the next day. Reduction in erythema also was observed within 1 to 4 days.³⁰

No AEs associated with VD supplementation were reported, suggesting a potential beneficial role of high-dose VD in accelerating recovery from chemotherapy-induced rashes without evident safety concerns.

Radiation Therapy–Induced Cutaneous Events Treated with High-Dose Vitamin D

Radiation dermatitis is a common and often severe complication of radiation therapy that affects more than 90% of patients undergoing treatment, with half of these individuals experiencing grade 2 toxicity, according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events.^31,32 Radiation damage to basal keratinocytes and hair follicle stem cells disrupts the renewal of the skin’s outer layer, while a surge of free radicals causes irreversible DNA damage.³³ Symptoms of radiation dermatitis can vary from mild pink erythema to tissue ulceration and necrosis, typically within 1 to 4 weeks of radiation exposure.³⁴ The resulting dermatitis can take 2 to 4 weeks to heal, notably impacting patient quality of life and often necessitating modifications or interruptions in cancer therapy.³³

Prior studies have demonstrated the use of high-dose VD to improve the healing of UV-irradiated skin. A randomized controlled trial investigated high-dose vitamin D₃ to treat experimentally induced sunburn in 20 healthy adults. Compared with those who received a placebo, participants receiving the oral dose of 200,000 IU of vitamin D₃ demonstrated suppression of the pro-inflammatory mediators tumor necrosis factor α (P=.04) and inducible nitric oxide synthase (P=.02), while expression of tissue repair enhancer arginase 1 was increased (P<.005).³⁵ The mechanism of this enhanced tissue repair was investigated using a mouse model, in which intraperitoneal 25(OH)D was administered following severe UV-induced skin injury. On immunofluorescence microscopy, mice treated with VD showed enhanced autophagy within the macrophages infiltrating UV-irradiated skin.³⁶ The use of high-dose VD to treat UV-irradiated skin in these studies established a precedent for using VD to heal cutaneous injury caused by ionizing radiation therapy.

Some studies have focused on the role of VD for treating acute radiation dermatitis. A study of 23 patients with ductal carcinoma in situ or localized invasive ductal carcinoma breast cancer compared the effectiveness of topical calcipotriol to that of a standard hydrating ointment.³⁷ Participants were randomized to 1 of 2 treatments before starting adjuvant radiotherapy to evaluate their potential in preventing radiation dermatitis. In 87% (20/23) of these patients, no difference in skin reaction was observed between the 2 treatments, suggesting that topical VD application may not offer any advantage over the standard hydrating ointment for the prevention of radiation dermatitis.³⁷

Benefits of high-dose oral VD for treating radiation dermatitis also have been reported. Nguyen et al³⁸ documented 3 cases in which patients with neuroendocrine carcinoma of the pancreas, tonsillar carcinoma, and breast cancer received 200,000 IU of oral ergocalciferol distributed over 2 doses given 7 days apart for radiation dermatitis. These patients experienced substantial improvements in pain, swelling, and redness within a week of the initial dose. Additionally, a case of radiation recall dermatitis, which occurred a week after vinorelbine chemotherapy, was treated with 2 doses totaling 100,000 IU of oral ergocalciferol. This patient also had improvement in pain and swelling but continued to have tumor-related induration and ulceration.³⁹

Although topical VD did not show significant benefits over standard treatments for radiation dermatitis, high-dose oral VD appears promising in improving patient outcomes of pain and swelling more rapidly than current practices. Further research is needed to confirm these findings and establish standardized treatment protocols.

Final Thoughts

Suboptimal VD levels are prevalent in numerous cancer types. Chemotherapy often is associated with acute, potentially transient worsening of VD status in patients with breast and colorectal cancer. Although 25(OH)D levels have not corresponded with increased frequency of ­chemotherapy-related dermatologic AEs, suboptimal 25(OH)D levels appear to be associated with increased severity of radiation-induced mucositis and dermatitis.^20,25,26 The use of high-dose VD as a therapeutic agent shows promise in mitigating chemotherapy-induced and radiation therapy–induced rashes in multiple cancer types with reduction of inflammatory markers and a durable anti-inflammatory impact. Although the mechanisms of cellular injury vary among chemotherapeutic agents, the anti-inflammatory and tissue repair properties of VD may make it an effective treatment for chemotherapy-induced cutaneous damage regardless of injury mechanism.^2-4,35 However, reports of clinical improvement vary, and further objective studies to classify optimal dosing, administration, and outcome measures are needed. The absence of reported AEs associated with high-dose VD supplementation is encouraging, but selection of a safe and optimal dosing regimen can only occur with dedicated clinical trials.

Vitamin D (VD) regulates keratinocyte proliferation and differentiation, modulates inflammatory pathways, and protects against cellular damage in the skin. ¹ In the setting of tissue injury and acute skin inflammation, active vitamin D—1,25(OH) ₂ D—suppresses signaling from pro-inflammatory chemokines and cytokines such as IFN- γ and IL-17. ^2,3 This suppression reduces proliferation of helper T cells (T _H 1, T _H 17) and B cells, decreasing tissue damage from reactive oxygen species release while enhancing secretion of the anti-inflammatory cytokine IL-10 by antigen-presenting cells. ^2-4

Suboptimal VD levels have been associated with numerous health consequences including malignancy, prompting interest in VD supplementation for improving cancer-related outcomes.⁵ Beyond disease prognosis, high-dose VD supplementation has been suggested as a potential therapy for adverse events (AEs) related to cancer treatments. In one study, mice that received oral vitamin D₃ supplementation of 11,500 IU/kg daily had fewer doxorubicin-induced cardiotoxic effects on ejection fraction (P<.0001) and stroke volume (P<.01) than mice that received VD supplementation of 1500 IU/kg daily.⁶

In this review, we examine the impact of chemoradiation on 25(OH)D levels—which more accurately reflects VD stores than 1,25(OH)₂D levels—and the impact of suboptimal VD on cutaneous toxicities related to chemoradiation. To define the suboptimal VD threshold, we used the Endocrine Society’s clinical practice guidelines, which characterize suboptimal 25(OH)D levels as insufficiency (21–29 ng/mL [52.5–72.5 nmol/L]) or deficiency (<20 ng/mL [50 nmol/L])⁷; deficiency can be further categorized as severe deficiency (<12 ng/mL [30 nmol/L]).⁸ This review also evaluates the evidence for vitamin D₃ supplementation to alleviate the cutaneous AEs of chemotherapy and radiation treatments.

Effects of Chemotherapy on Vitamin D Levels

A high prevalence of VD deficiency is seen in various cancers. In a retrospective review of 25(OH)D levels in 2098 adults with solid tumors of any stage (6% had metastatic disease [n=124]), suboptimal levels were found in 69% of patients with breast cancer (n=617), 75% with colorectal cancer (n=84), 72% with gynecologic cancer (n=65), 79% with kidney and bladder cancer (n=145), 83% with pancreatic and upper gastrointestinal tract cancer (n=178), 73% with lung cancer (n=73), 69% with prostate cancer (n=225), 61% with skin cancer (n=399), and 63% with thyroid cancer (n=172).⁵ Suboptimal VD also has been found in hematologic malignancies. In a prospective cohort study, mean serum 25(OH)D levels in 23 patients with recently diagnosed acute myeloid leukemia demonstrated VD deficiency (mean [SD], 18.6 [6.6] nmol/L).⁹ Given that many patients already exhibit a baseline VD deficiency at cancer diagnosis, it is important to understand the relationship between VD and cancer treatment modalities.⁵

In the United States, breast and colorectal cancers were estimated to be the first and fourth most common cancers, with 313,510 and 152,810 predicted new cases in 2024, respectively.¹⁰ This review will focus on breast and colorectal cancer when describing VD variation associated with chemotherapy exposure due to their high prevalence.

Effects of Chemotherapy on Vitamin D Levels in Breast Cancer—Breast cancer studies have shown suboptimal VD levels in 76% of females 75 years or younger with any T1, T2, or T3; N0 or N1; and M0 breast cancer, in which 38.5% (n=197) had insufficient and 37.5% (n=192) had deficient 25(OH)D levels.¹¹ In a study of female patients with primary breast cancer (stage I, II, or III and T1 with high Ki67 expression [≥30%], T2, or T3), VD deficiency was seen in 60% of patients not receiving VD supplementation.^12,13 A systematic review that included 7 studies of different types of breast cancer suggested that circulating 25(OH)D may be associated with improved prognosis.¹⁴ Thus, studies have investigated risk factors associated with poor or worsening VD status in individuals with breast cancer, including exposure to chemotherapy and/or radiation treatment.^12,15-18

A prospective cohort study assessed 25(OH)D levels in 95 patients with any breast cancer (stages I, II, IIIA, IIIB) before and after initiating chemotherapy with docetaxel, doxorubicin, epirubicin, 5-fluorouracil, or cyclophosphamide, compared with a group of 52 females without cancer.¹⁷ In the breast cancer group, approximately 80% (76/95) had suboptimal and 50% (47/95) had deficient VD levels before chemotherapy initiation (mean [SD], 54.1 [22.8] nmol/L). In the comparison group, 60% (31/52) had suboptimal and 30% (15/52) had deficient VD at baseline (mean [SD], 66.1 [23.5] nmol/L), which was higher than the breast cancer group (P=.03). A subgroup analysis excluded participants who started, stopped, or lacked data on dietary supplements containing VD (n=39); in the remaining 56 participants, a significant decrease in 25(OH)D levels was observed shortly after finishing chemotherapy compared with the prechemotherapy baseline value (mean, −7.9 nmol/L; P=.004). Notably, 6 months after chemotherapy completion, 25(OH)D levels increased (mean, +12.8 nmol/L; P<.001). Vitamin D levels remained stable in the comparison group (P=.987).¹⁷

Consistent with these findings, a cross-sectional study assessing VD status in 394 female patients with primary breast cancer (stage I, II, or III and T1 with high Ki67 expression [≥30%], T2, or T3), found that a history of chemotherapy was associated with increased odds of 25(OH)D levels less than 20 ng/mL compared with breast cancer patients with no prior chemotherapy (odds ratio, 1.86; 95% CI, 1.03-3.38).¹² Although the study data included chemotherapy history, no information was provided on specific chemotherapy agents or regimens used in this cohort, limiting the ability to detect the drugs most often implicated.

Both studies indicated a complex interplay between chemotherapy and VD levels in breast cancer patients. Although Kok et al¹⁷ suggested a transient decrease in VD levels during chemotherapy with a subsequent recovery after cessation, Fassio et al¹² highlighted the increased odds of VD deficiency associated with chemotherapy. Ultimately, larger randomized controlled trials are needed to better understand the relationship between chemotherapy and VD status in breast cancer patients.

Effects of Chemotherapy on Vitamin D Levels in Colorectal Cancer—Similar to patterns seen in breast cancer, a systematic review with 6 studies of different types of colorectal cancer suggested that circulating 25(OH)D levels may be associated with prognosis.¹⁴ Studies also have investigated the relationship between colorectal chemotherapy regimens and VD status.^15,16,18,19

A retrospective study assessed 25(OH)D levels in 315 patients with any colorectal cancer (stage I–IV).¹⁵ Patients were included in the analysis if they received less than 400 IU daily of VD supplementation at baseline. For the whole study sample, the mean (SD) VD level was 23.7 (13.71) ng/mL. Patients who had not received chemotherapy within 3 months of the VD level assessment were categorized as the no chemotherapy group, and the others were designated as the chemotherapy group; the latter group was exposed to various chemotherapy regimens, including combinations of irinotecan, oxaliplatin, 5-fluorouracil, leucovorin, bevacizumab, or cetuximab. Multivariate analysis showed that the chemotherapy group was 3.7 times more likely to have very low VD levels (≤15 ng/mL) compared with those in the no chemotherapy group (P<.0001).¹⁵

A separate cross-sectional study examined serum 25(OH)D concentrations in 1201 patients with any newly diagnosed colorectal carcinoma (stage I–III); 91% of cases were adenocarcinoma.¹⁸ In a multivariate analysis, chemotherapy plus surgery was associated with lower VD levels than surgery alone 6 months after diagnosis (mean, −8.74 nmol/L; 95% CI, −11.30 to −6.18 nmol/L), specifically decreasing by a mean of 6.7 nmol/L (95% CI, −9.8 to −3.8 nmol/L) after adjusting for demographic and lifestyle factors.¹⁸ However, a prospective cohort study demonstrated different findings.¹⁹ Comparing 58 patients with newly diagnosed colorectal adenocarcinoma (stages I–IV) who underwent chemotherapy and 36 patients who did not receive chemotherapy, there was no significant change in 25(OH)D levels from the time of diagnosis to 6 months later. Median VD levels decreased by 0.7 ng/mL in those who received chemotherapy, while a minimal (and not significant) increase of 1.6 ng/mL was observed in those without chemotherapy intervention (P=.26). Notably, supplementation was not restricted in this cohort, which may have resulted in higher VD levels in those taking supplements.¹⁹

Since time of year and geographic location can influence VD levels, one prospective cohort study controlled for differential sun exposure due to these factors in their analysis.¹⁶ Assessment of 25(OH)D levels was completed in 81 chemotherapy-naïve cancer patients immediately before beginning chemotherapy as well as 6 and 12 weeks into treatment. More than 8 primary cancer types were represented in this study, with breast (34% [29/81]) and colorectal (14% [12/81]) cancer being the most common, but the cancer stages of the participants were not detailed. Vitamin D levels decreased after commencing chemotherapy, with the largest drop occurring 6 weeks into treatment. From the 6- to 12-week end points, VD increased but remained below the original baseline level (baseline: mean [SD], 49.2 [22.3] nmol/L; 6 weeks: mean [SD], 40.9 [19.0] nmol/L; 12 weeks: mean [SD], 45.9 [19.7] nmol/L; P=.05).¹⁶

Although focused on breast and colorectal cancers, these studies suggest that various chemotherapy regimens may confer a higher risk for VD deficiency compared with VD status at diagnosis and/or prior to chemotherapy treatment. However, most of these studies only discussed stage-based differences, excluding analysis of the variety of cancer subtypes that comprise breast and colorectal malignancies, which may limit our ability to extrapolate from these data. Ultimately, larger randomized controlled trials are needed to better understand the relationship between chemotherapy and VD status across various primary cancer types.

Effects of Radiation Therapy on Vitamin D Levels

Unlike chemotherapy, studies on the association between radiation therapy and VD levels are minimal, with most reports in the literature discussing the use of VD to potentiate the effects of radiation therapy. In one cross-sectional analysis of 1201 patients with newly diagnosed stage I, II, or III colorectal cancer of any type (94% were adenocarcinoma), radiation plus surgery was associated with slightly lower 25(OH)D levels than surgery alone for tumor treatment 6 months after diagnosis (mean, −3.17; 95% CI, −6.07 to −0.28 nmol/L). However, after adjustment for demographic and lifestyle factors, this decrease in VD levels attributable to radiotherapy was not statistically significant compared with the surgery-only cohort (mean, −1.78; 95% CI, −5.07 to 1.52 nmol/L).¹⁸

Similarly, a cross-sectional study assessing VD status in 394 female patients with primary breast cancer (stage I, II, or III and T1 with high Ki67 expression [≥30%], T2, or T3), found that a history of radiotherapy was not associated with a difference in serum 25(OH)D levels compared with those with breast cancer without prior radiotherapy (odds ratio, 0.90; 95% CI, 0.52-1.54).¹² From the limited existing literature specifically addressing variations of VD levels with radiation, radiation therapy does not appear to significantly impact VD levels.

Vitamin D Levels and the Severity of Chemotherapy- or Radiation Therapy–Induced AEs

A prospective cohort of 241 patients did not find an increase in the incidence or severity of chemotherapy-induced cutaneous toxicities in those with suboptimal 1,25(OH)₂D₃ levels (≤75 nmol/L).²⁰ Eight different primary cancer types were represented, including breast and colorectal cancer; the tumor stages of the participants were not detailed. Forty-one patients had normal 1,25(OH)₂D₃ levels, while the remaining 200 had suboptimal levels. There was no significant association between serum VD levels and the following dermatologic toxicities: desquamation (P=.26), xerosis (P=.15), mucositis (P=.30), or painful rash (P=.87). Surprisingly, nail changes and hand-foot reactions occurred with greater frequency in patients with normal VD levels (P=.01 and P=.03, respectively).²⁰ Hand-foot reaction is part of the toxic erythema of chemotherapy (TEC) spectrum, which is comprised of a range of cytotoxic skin injuries that typically manifest within 2 to 3 weeks of exposure to the offending chemotherapeutic agents, often characterized by erythema, pain, swelling, and blistering, particularly in intertriginous and acral areas.^21-23 Recovery from TEC generally takes at least 2 to 4 weeks and may necessitate cessation of the offending chemotherapeutic agent.^21,24 Notably, this study measured 1,25(OH)₂D₃ levels instead of 25(OH)D levels, which may not reliably indicate body stores of VD.^7,20 These results underscore the complex nature between chemotherapy and VD; however, VD levels alone do not appear to be a sufficient biomarker for predicting chemotherapy-associated cutaneous AEs.

Interestingly, radiation therapy–induced AEs may be associated with VD levels. A prospective cohort study of 98 patients with prostate, bladder, or gynecologic cancers (tumor stages were not detailed) undergoing pelvic radiotherapy found that females and males with 25(OH)D levels below a threshold of 35 and 40 nmol/L, respectively, were more likely to experience higher Radiation Therapy Oncology Group (RTOG) grade acute proctitis compared with those with VD above these thresholds.²⁵ Specifically, VD below these thresholds was associated with increased odds of RTOG grade 2 or higher radiation-induced proctitis (OR, 3.07; 95% CI, 1.27-7.50 [P=.013]). Additionally, a weak correlation was noted between VD below these thresholds and the RTOG grade, with a Spearman correlation value of −0.189 (P=.031).²⁵

One prospective cohort study included 28 patients with any cancer of the oral cavity, oropharynx, hypopharynx, or larynx stages II, III, or IVA; 93% (26/28) were stage III or IVA.²⁶ The 20 (71%) patients with suboptimal 25(OH)D levels (≤75 nmol/L) experienced a higher prevalence of grade II radiation dermatitis compared with the 8 (29%) patients with optimal VD levels (χ²₂=5.973; P=.0505). This pattern persisted with the severity of mucositis; patients from the suboptimal VD group presented with higher rates of grades II and III mucositis compared with the VD optimal group (χ²₂=13.627; P=.0011).²⁶ Recognizing the small cohort evaluated in the study, we highlight the importance of further studies to clarify these associations.

Chemotherapy-Induced Cutaneous Events Treated with High-Dose Vitamin D

Chemotherapeutic agents are known to induce cellular damage, resulting in a range of cutaneous AEs that can invoke discontinuation of otherwise effective chemotherapeutic interventions.^27,28 Recent research has explored the potential of high-dose vitamin D₃ as a therapeutic agent to mitigate cutaneous reactions.^29,30

A randomized, double-blind, placebo-controlled trial investigated the use of a single high dose of oral ­25(OH)D to treat topical nitrogen mustard (NM)–induced rash.²⁹ To characterize baseline inflammatory responses from NM injury without intervention, clinical measures, serum studies, and tissue analyses from skin biopsies were performed on 28 healthy adults after exposure to topical NM—a chemotherapeutic agent classified as a DNA alkylator. Two weeks later, participants were exposed to topical NM a second time and were split into 2 groups: 14 patients received a single 200,000-IU dose of oral 25(OH)D while the other 14 participants were given a placebo. Using the inflammatory markers induced from baseline exposure to NM alone, posttreatment analysis revealed that the punch biopsies from the 25(OH)D group expressed fewer NM-induced inflammatory markers compared with the placebo group at both 72 hours and 6 weeks following NM injury (72 hours: 12 vs 17 inflammatory markers; 6 weeks: 4 vs 11 inflammatory markers). Notably, NM inflammatory markers were enriched for IL-17 signaling pathways in the placebo biopsies but not in the 25(OH)D intervention group. This study also identified mild and severe patterns of inflammatory responses to NM that were independent of the 25(OH)D intervention. Biomarkers specific to skin biopsies from participants with the severe response included CCL20, CCL2, and CXCL8 (adjusted P<.05). At 6 weeks posttreatment, the 25(OH)D group showed a 67% reduction in NM injury markers compared with a 35% reduction in the placebo group. Despite a reduction in tissue inflammatory markers, there were no clinically significant changes observed in skin redness, swelling, or histologic structure when comparing the 25(OH)D- supplemented group to the placebo group at any time during the study, necessitating further research into the mechanistic roles of high doses VD supplementation.²⁹

Although Ernst et al²⁹ did not observe any clinically significant improvements with VD treatment, a case series of 6 patients with either glioblastoma multiforme, acute myeloid leukemia, or aplastic anemia did demonstrate clinical improvement of TEC after receiving high-dose vitamin D₃.³⁰ The mean time to onset of TEC was noted at 8.5 days following administration of the inciting chemotherapeutic agent, which included combinations of anthracycline, antimetabolite, kinase inhibitor, B-cell lymphoma 2 inhibitor, purine analogue, and alkylating agents. A combination of clinical and histologic findings was used to diagnose TEC. Baseline 25(OH)D levels were not established prior to treatment. The treatment regimen for 1 patient included 2 doses of 50,000 IU of VD spaced 1 week apart, totaling 100,000 IU, while the remaining 5 patients received a total of 200,000 IU, also split into 2 doses given 1 week apart. All patients received their first dose of VD within a week of the cutaneous eruption. Following the initial VD dose, there was a notable improvement in pain, pruritus, or swelling by the next day. Reduction in erythema also was observed within 1 to 4 days.³⁰

No AEs associated with VD supplementation were reported, suggesting a potential beneficial role of high-dose VD in accelerating recovery from chemotherapy-induced rashes without evident safety concerns.

Radiation Therapy–Induced Cutaneous Events Treated with High-Dose Vitamin D

Radiation dermatitis is a common and often severe complication of radiation therapy that affects more than 90% of patients undergoing treatment, with half of these individuals experiencing grade 2 toxicity, according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events.^31,32 Radiation damage to basal keratinocytes and hair follicle stem cells disrupts the renewal of the skin’s outer layer, while a surge of free radicals causes irreversible DNA damage.³³ Symptoms of radiation dermatitis can vary from mild pink erythema to tissue ulceration and necrosis, typically within 1 to 4 weeks of radiation exposure.³⁴ The resulting dermatitis can take 2 to 4 weeks to heal, notably impacting patient quality of life and often necessitating modifications or interruptions in cancer therapy.³³

Prior studies have demonstrated the use of high-dose VD to improve the healing of UV-irradiated skin. A randomized controlled trial investigated high-dose vitamin D₃ to treat experimentally induced sunburn in 20 healthy adults. Compared with those who received a placebo, participants receiving the oral dose of 200,000 IU of vitamin D₃ demonstrated suppression of the pro-inflammatory mediators tumor necrosis factor α (P=.04) and inducible nitric oxide synthase (P=.02), while expression of tissue repair enhancer arginase 1 was increased (P<.005).³⁵ The mechanism of this enhanced tissue repair was investigated using a mouse model, in which intraperitoneal 25(OH)D was administered following severe UV-induced skin injury. On immunofluorescence microscopy, mice treated with VD showed enhanced autophagy within the macrophages infiltrating UV-irradiated skin.³⁶ The use of high-dose VD to treat UV-irradiated skin in these studies established a precedent for using VD to heal cutaneous injury caused by ionizing radiation therapy.

Some studies have focused on the role of VD for treating acute radiation dermatitis. A study of 23 patients with ductal carcinoma in situ or localized invasive ductal carcinoma breast cancer compared the effectiveness of topical calcipotriol to that of a standard hydrating ointment.³⁷ Participants were randomized to 1 of 2 treatments before starting adjuvant radiotherapy to evaluate their potential in preventing radiation dermatitis. In 87% (20/23) of these patients, no difference in skin reaction was observed between the 2 treatments, suggesting that topical VD application may not offer any advantage over the standard hydrating ointment for the prevention of radiation dermatitis.³⁷

Benefits of high-dose oral VD for treating radiation dermatitis also have been reported. Nguyen et al³⁸ documented 3 cases in which patients with neuroendocrine carcinoma of the pancreas, tonsillar carcinoma, and breast cancer received 200,000 IU of oral ergocalciferol distributed over 2 doses given 7 days apart for radiation dermatitis. These patients experienced substantial improvements in pain, swelling, and redness within a week of the initial dose. Additionally, a case of radiation recall dermatitis, which occurred a week after vinorelbine chemotherapy, was treated with 2 doses totaling 100,000 IU of oral ergocalciferol. This patient also had improvement in pain and swelling but continued to have tumor-related induration and ulceration.³⁹

Although topical VD did not show significant benefits over standard treatments for radiation dermatitis, high-dose oral VD appears promising in improving patient outcomes of pain and swelling more rapidly than current practices. Further research is needed to confirm these findings and establish standardized treatment protocols.

Final Thoughts

Suboptimal VD levels are prevalent in numerous cancer types. Chemotherapy often is associated with acute, potentially transient worsening of VD status in patients with breast and colorectal cancer. Although 25(OH)D levels have not corresponded with increased frequency of ­chemotherapy-related dermatologic AEs, suboptimal 25(OH)D levels appear to be associated with increased severity of radiation-induced mucositis and dermatitis.^20,25,26 The use of high-dose VD as a therapeutic agent shows promise in mitigating chemotherapy-induced and radiation therapy–induced rashes in multiple cancer types with reduction of inflammatory markers and a durable anti-inflammatory impact. Although the mechanisms of cellular injury vary among chemotherapeutic agents, the anti-inflammatory and tissue repair properties of VD may make it an effective treatment for chemotherapy-induced cutaneous damage regardless of injury mechanism.^2-4,35 However, reports of clinical improvement vary, and further objective studies to classify optimal dosing, administration, and outcome measures are needed. The absence of reported AEs associated with high-dose VD supplementation is encouraging, but selection of a safe and optimal dosing regimen can only occur with dedicated clinical trials.

Vitamin D (VD) regulates keratinocyte proliferation and differentiation, modulates inflammatory pathways, and protects against cellular damage in the skin. ¹ In the setting of tissue injury and acute skin inflammation, active vitamin D—1,25(OH) ₂ D—suppresses signaling from pro-inflammatory chemokines and cytokines such as IFN- γ and IL-17. ^2,3 This suppression reduces proliferation of helper T cells (T _H 1, T _H 17) and B cells, decreasing tissue damage from reactive oxygen species release while enhancing secretion of the anti-inflammatory cytokine IL-10 by antigen-presenting cells. ^2-4

Suboptimal VD levels have been associated with numerous health consequences including malignancy, prompting interest in VD supplementation for improving cancer-related outcomes.⁵ Beyond disease prognosis, high-dose VD supplementation has been suggested as a potential therapy for adverse events (AEs) related to cancer treatments. In one study, mice that received oral vitamin D₃ supplementation of 11,500 IU/kg daily had fewer doxorubicin-induced cardiotoxic effects on ejection fraction (P<.0001) and stroke volume (P<.01) than mice that received VD supplementation of 1500 IU/kg daily.⁶

In this review, we examine the impact of chemoradiation on 25(OH)D levels—which more accurately reflects VD stores than 1,25(OH)₂D levels—and the impact of suboptimal VD on cutaneous toxicities related to chemoradiation. To define the suboptimal VD threshold, we used the Endocrine Society’s clinical practice guidelines, which characterize suboptimal 25(OH)D levels as insufficiency (21–29 ng/mL [52.5–72.5 nmol/L]) or deficiency (<20 ng/mL [50 nmol/L])⁷; deficiency can be further categorized as severe deficiency (<12 ng/mL [30 nmol/L]).⁸ This review also evaluates the evidence for vitamin D₃ supplementation to alleviate the cutaneous AEs of chemotherapy and radiation treatments.

Effects of Chemotherapy on Vitamin D Levels

A high prevalence of VD deficiency is seen in various cancers. In a retrospective review of 25(OH)D levels in 2098 adults with solid tumors of any stage (6% had metastatic disease [n=124]), suboptimal levels were found in 69% of patients with breast cancer (n=617), 75% with colorectal cancer (n=84), 72% with gynecologic cancer (n=65), 79% with kidney and bladder cancer (n=145), 83% with pancreatic and upper gastrointestinal tract cancer (n=178), 73% with lung cancer (n=73), 69% with prostate cancer (n=225), 61% with skin cancer (n=399), and 63% with thyroid cancer (n=172).⁵ Suboptimal VD also has been found in hematologic malignancies. In a prospective cohort study, mean serum 25(OH)D levels in 23 patients with recently diagnosed acute myeloid leukemia demonstrated VD deficiency (mean [SD], 18.6 [6.6] nmol/L).⁹ Given that many patients already exhibit a baseline VD deficiency at cancer diagnosis, it is important to understand the relationship between VD and cancer treatment modalities.⁵

In the United States, breast and colorectal cancers were estimated to be the first and fourth most common cancers, with 313,510 and 152,810 predicted new cases in 2024, respectively.¹⁰ This review will focus on breast and colorectal cancer when describing VD variation associated with chemotherapy exposure due to their high prevalence.

Effects of Chemotherapy on Vitamin D Levels in Breast Cancer—Breast cancer studies have shown suboptimal VD levels in 76% of females 75 years or younger with any T1, T2, or T3; N0 or N1; and M0 breast cancer, in which 38.5% (n=197) had insufficient and 37.5% (n=192) had deficient 25(OH)D levels.¹¹ In a study of female patients with primary breast cancer (stage I, II, or III and T1 with high Ki67 expression [≥30%], T2, or T3), VD deficiency was seen in 60% of patients not receiving VD supplementation.^12,13 A systematic review that included 7 studies of different types of breast cancer suggested that circulating 25(OH)D may be associated with improved prognosis.¹⁴ Thus, studies have investigated risk factors associated with poor or worsening VD status in individuals with breast cancer, including exposure to chemotherapy and/or radiation treatment.^12,15-18

A prospective cohort study assessed 25(OH)D levels in 95 patients with any breast cancer (stages I, II, IIIA, IIIB) before and after initiating chemotherapy with docetaxel, doxorubicin, epirubicin, 5-fluorouracil, or cyclophosphamide, compared with a group of 52 females without cancer.¹⁷ In the breast cancer group, approximately 80% (76/95) had suboptimal and 50% (47/95) had deficient VD levels before chemotherapy initiation (mean [SD], 54.1 [22.8] nmol/L). In the comparison group, 60% (31/52) had suboptimal and 30% (15/52) had deficient VD at baseline (mean [SD], 66.1 [23.5] nmol/L), which was higher than the breast cancer group (P=.03). A subgroup analysis excluded participants who started, stopped, or lacked data on dietary supplements containing VD (n=39); in the remaining 56 participants, a significant decrease in 25(OH)D levels was observed shortly after finishing chemotherapy compared with the prechemotherapy baseline value (mean, −7.9 nmol/L; P=.004). Notably, 6 months after chemotherapy completion, 25(OH)D levels increased (mean, +12.8 nmol/L; P<.001). Vitamin D levels remained stable in the comparison group (P=.987).¹⁷

Consistent with these findings, a cross-sectional study assessing VD status in 394 female patients with primary breast cancer (stage I, II, or III and T1 with high Ki67 expression [≥30%], T2, or T3), found that a history of chemotherapy was associated with increased odds of 25(OH)D levels less than 20 ng/mL compared with breast cancer patients with no prior chemotherapy (odds ratio, 1.86; 95% CI, 1.03-3.38).¹² Although the study data included chemotherapy history, no information was provided on specific chemotherapy agents or regimens used in this cohort, limiting the ability to detect the drugs most often implicated.

Both studies indicated a complex interplay between chemotherapy and VD levels in breast cancer patients. Although Kok et al¹⁷ suggested a transient decrease in VD levels during chemotherapy with a subsequent recovery after cessation, Fassio et al¹² highlighted the increased odds of VD deficiency associated with chemotherapy. Ultimately, larger randomized controlled trials are needed to better understand the relationship between chemotherapy and VD status in breast cancer patients.

Effects of Chemotherapy on Vitamin D Levels in Colorectal Cancer—Similar to patterns seen in breast cancer, a systematic review with 6 studies of different types of colorectal cancer suggested that circulating 25(OH)D levels may be associated with prognosis.¹⁴ Studies also have investigated the relationship between colorectal chemotherapy regimens and VD status.^15,16,18,19

A retrospective study assessed 25(OH)D levels in 315 patients with any colorectal cancer (stage I–IV).¹⁵ Patients were included in the analysis if they received less than 400 IU daily of VD supplementation at baseline. For the whole study sample, the mean (SD) VD level was 23.7 (13.71) ng/mL. Patients who had not received chemotherapy within 3 months of the VD level assessment were categorized as the no chemotherapy group, and the others were designated as the chemotherapy group; the latter group was exposed to various chemotherapy regimens, including combinations of irinotecan, oxaliplatin, 5-fluorouracil, leucovorin, bevacizumab, or cetuximab. Multivariate analysis showed that the chemotherapy group was 3.7 times more likely to have very low VD levels (≤15 ng/mL) compared with those in the no chemotherapy group (P<.0001).¹⁵

A separate cross-sectional study examined serum 25(OH)D concentrations in 1201 patients with any newly diagnosed colorectal carcinoma (stage I–III); 91% of cases were adenocarcinoma.¹⁸ In a multivariate analysis, chemotherapy plus surgery was associated with lower VD levels than surgery alone 6 months after diagnosis (mean, −8.74 nmol/L; 95% CI, −11.30 to −6.18 nmol/L), specifically decreasing by a mean of 6.7 nmol/L (95% CI, −9.8 to −3.8 nmol/L) after adjusting for demographic and lifestyle factors.¹⁸ However, a prospective cohort study demonstrated different findings.¹⁹ Comparing 58 patients with newly diagnosed colorectal adenocarcinoma (stages I–IV) who underwent chemotherapy and 36 patients who did not receive chemotherapy, there was no significant change in 25(OH)D levels from the time of diagnosis to 6 months later. Median VD levels decreased by 0.7 ng/mL in those who received chemotherapy, while a minimal (and not significant) increase of 1.6 ng/mL was observed in those without chemotherapy intervention (P=.26). Notably, supplementation was not restricted in this cohort, which may have resulted in higher VD levels in those taking supplements.¹⁹

Since time of year and geographic location can influence VD levels, one prospective cohort study controlled for differential sun exposure due to these factors in their analysis.¹⁶ Assessment of 25(OH)D levels was completed in 81 chemotherapy-naïve cancer patients immediately before beginning chemotherapy as well as 6 and 12 weeks into treatment. More than 8 primary cancer types were represented in this study, with breast (34% [29/81]) and colorectal (14% [12/81]) cancer being the most common, but the cancer stages of the participants were not detailed. Vitamin D levels decreased after commencing chemotherapy, with the largest drop occurring 6 weeks into treatment. From the 6- to 12-week end points, VD increased but remained below the original baseline level (baseline: mean [SD], 49.2 [22.3] nmol/L; 6 weeks: mean [SD], 40.9 [19.0] nmol/L; 12 weeks: mean [SD], 45.9 [19.7] nmol/L; P=.05).¹⁶

Although focused on breast and colorectal cancers, these studies suggest that various chemotherapy regimens may confer a higher risk for VD deficiency compared with VD status at diagnosis and/or prior to chemotherapy treatment. However, most of these studies only discussed stage-based differences, excluding analysis of the variety of cancer subtypes that comprise breast and colorectal malignancies, which may limit our ability to extrapolate from these data. Ultimately, larger randomized controlled trials are needed to better understand the relationship between chemotherapy and VD status across various primary cancer types.

Effects of Radiation Therapy on Vitamin D Levels

Unlike chemotherapy, studies on the association between radiation therapy and VD levels are minimal, with most reports in the literature discussing the use of VD to potentiate the effects of radiation therapy. In one cross-sectional analysis of 1201 patients with newly diagnosed stage I, II, or III colorectal cancer of any type (94% were adenocarcinoma), radiation plus surgery was associated with slightly lower 25(OH)D levels than surgery alone for tumor treatment 6 months after diagnosis (mean, −3.17; 95% CI, −6.07 to −0.28 nmol/L). However, after adjustment for demographic and lifestyle factors, this decrease in VD levels attributable to radiotherapy was not statistically significant compared with the surgery-only cohort (mean, −1.78; 95% CI, −5.07 to 1.52 nmol/L).¹⁸

Similarly, a cross-sectional study assessing VD status in 394 female patients with primary breast cancer (stage I, II, or III and T1 with high Ki67 expression [≥30%], T2, or T3), found that a history of radiotherapy was not associated with a difference in serum 25(OH)D levels compared with those with breast cancer without prior radiotherapy (odds ratio, 0.90; 95% CI, 0.52-1.54).¹² From the limited existing literature specifically addressing variations of VD levels with radiation, radiation therapy does not appear to significantly impact VD levels.

Vitamin D Levels and the Severity of Chemotherapy- or Radiation Therapy–Induced AEs

A prospective cohort of 241 patients did not find an increase in the incidence or severity of chemotherapy-induced cutaneous toxicities in those with suboptimal 1,25(OH)₂D₃ levels (≤75 nmol/L).²⁰ Eight different primary cancer types were represented, including breast and colorectal cancer; the tumor stages of the participants were not detailed. Forty-one patients had normal 1,25(OH)₂D₃ levels, while the remaining 200 had suboptimal levels. There was no significant association between serum VD levels and the following dermatologic toxicities: desquamation (P=.26), xerosis (P=.15), mucositis (P=.30), or painful rash (P=.87). Surprisingly, nail changes and hand-foot reactions occurred with greater frequency in patients with normal VD levels (P=.01 and P=.03, respectively).²⁰ Hand-foot reaction is part of the toxic erythema of chemotherapy (TEC) spectrum, which is comprised of a range of cytotoxic skin injuries that typically manifest within 2 to 3 weeks of exposure to the offending chemotherapeutic agents, often characterized by erythema, pain, swelling, and blistering, particularly in intertriginous and acral areas.^21-23 Recovery from TEC generally takes at least 2 to 4 weeks and may necessitate cessation of the offending chemotherapeutic agent.^21,24 Notably, this study measured 1,25(OH)₂D₃ levels instead of 25(OH)D levels, which may not reliably indicate body stores of VD.^7,20 These results underscore the complex nature between chemotherapy and VD; however, VD levels alone do not appear to be a sufficient biomarker for predicting chemotherapy-associated cutaneous AEs.

Interestingly, radiation therapy–induced AEs may be associated with VD levels. A prospective cohort study of 98 patients with prostate, bladder, or gynecologic cancers (tumor stages were not detailed) undergoing pelvic radiotherapy found that females and males with 25(OH)D levels below a threshold of 35 and 40 nmol/L, respectively, were more likely to experience higher Radiation Therapy Oncology Group (RTOG) grade acute proctitis compared with those with VD above these thresholds.²⁵ Specifically, VD below these thresholds was associated with increased odds of RTOG grade 2 or higher radiation-induced proctitis (OR, 3.07; 95% CI, 1.27-7.50 [P=.013]). Additionally, a weak correlation was noted between VD below these thresholds and the RTOG grade, with a Spearman correlation value of −0.189 (P=.031).²⁵

One prospective cohort study included 28 patients with any cancer of the oral cavity, oropharynx, hypopharynx, or larynx stages II, III, or IVA; 93% (26/28) were stage III or IVA.²⁶ The 20 (71%) patients with suboptimal 25(OH)D levels (≤75 nmol/L) experienced a higher prevalence of grade II radiation dermatitis compared with the 8 (29%) patients with optimal VD levels (χ²₂=5.973; P=.0505). This pattern persisted with the severity of mucositis; patients from the suboptimal VD group presented with higher rates of grades II and III mucositis compared with the VD optimal group (χ²₂=13.627; P=.0011).²⁶ Recognizing the small cohort evaluated in the study, we highlight the importance of further studies to clarify these associations.

Chemotherapy-Induced Cutaneous Events Treated with High-Dose Vitamin D

Chemotherapeutic agents are known to induce cellular damage, resulting in a range of cutaneous AEs that can invoke discontinuation of otherwise effective chemotherapeutic interventions.^27,28 Recent research has explored the potential of high-dose vitamin D₃ as a therapeutic agent to mitigate cutaneous reactions.^29,30

A randomized, double-blind, placebo-controlled trial investigated the use of a single high dose of oral ­25(OH)D to treat topical nitrogen mustard (NM)–induced rash.²⁹ To characterize baseline inflammatory responses from NM injury without intervention, clinical measures, serum studies, and tissue analyses from skin biopsies were performed on 28 healthy adults after exposure to topical NM—a chemotherapeutic agent classified as a DNA alkylator. Two weeks later, participants were exposed to topical NM a second time and were split into 2 groups: 14 patients received a single 200,000-IU dose of oral 25(OH)D while the other 14 participants were given a placebo. Using the inflammatory markers induced from baseline exposure to NM alone, posttreatment analysis revealed that the punch biopsies from the 25(OH)D group expressed fewer NM-induced inflammatory markers compared with the placebo group at both 72 hours and 6 weeks following NM injury (72 hours: 12 vs 17 inflammatory markers; 6 weeks: 4 vs 11 inflammatory markers). Notably, NM inflammatory markers were enriched for IL-17 signaling pathways in the placebo biopsies but not in the 25(OH)D intervention group. This study also identified mild and severe patterns of inflammatory responses to NM that were independent of the 25(OH)D intervention. Biomarkers specific to skin biopsies from participants with the severe response included CCL20, CCL2, and CXCL8 (adjusted P<.05). At 6 weeks posttreatment, the 25(OH)D group showed a 67% reduction in NM injury markers compared with a 35% reduction in the placebo group. Despite a reduction in tissue inflammatory markers, there were no clinically significant changes observed in skin redness, swelling, or histologic structure when comparing the 25(OH)D- supplemented group to the placebo group at any time during the study, necessitating further research into the mechanistic roles of high doses VD supplementation.²⁹

Although Ernst et al²⁹ did not observe any clinically significant improvements with VD treatment, a case series of 6 patients with either glioblastoma multiforme, acute myeloid leukemia, or aplastic anemia did demonstrate clinical improvement of TEC after receiving high-dose vitamin D₃.³⁰ The mean time to onset of TEC was noted at 8.5 days following administration of the inciting chemotherapeutic agent, which included combinations of anthracycline, antimetabolite, kinase inhibitor, B-cell lymphoma 2 inhibitor, purine analogue, and alkylating agents. A combination of clinical and histologic findings was used to diagnose TEC. Baseline 25(OH)D levels were not established prior to treatment. The treatment regimen for 1 patient included 2 doses of 50,000 IU of VD spaced 1 week apart, totaling 100,000 IU, while the remaining 5 patients received a total of 200,000 IU, also split into 2 doses given 1 week apart. All patients received their first dose of VD within a week of the cutaneous eruption. Following the initial VD dose, there was a notable improvement in pain, pruritus, or swelling by the next day. Reduction in erythema also was observed within 1 to 4 days.³⁰

No AEs associated with VD supplementation were reported, suggesting a potential beneficial role of high-dose VD in accelerating recovery from chemotherapy-induced rashes without evident safety concerns.

Radiation Therapy–Induced Cutaneous Events Treated with High-Dose Vitamin D

Radiation dermatitis is a common and often severe complication of radiation therapy that affects more than 90% of patients undergoing treatment, with half of these individuals experiencing grade 2 toxicity, according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events.^31,32 Radiation damage to basal keratinocytes and hair follicle stem cells disrupts the renewal of the skin’s outer layer, while a surge of free radicals causes irreversible DNA damage.³³ Symptoms of radiation dermatitis can vary from mild pink erythema to tissue ulceration and necrosis, typically within 1 to 4 weeks of radiation exposure.³⁴ The resulting dermatitis can take 2 to 4 weeks to heal, notably impacting patient quality of life and often necessitating modifications or interruptions in cancer therapy.³³

Prior studies have demonstrated the use of high-dose VD to improve the healing of UV-irradiated skin. A randomized controlled trial investigated high-dose vitamin D₃ to treat experimentally induced sunburn in 20 healthy adults. Compared with those who received a placebo, participants receiving the oral dose of 200,000 IU of vitamin D₃ demonstrated suppression of the pro-inflammatory mediators tumor necrosis factor α (P=.04) and inducible nitric oxide synthase (P=.02), while expression of tissue repair enhancer arginase 1 was increased (P<.005).³⁵ The mechanism of this enhanced tissue repair was investigated using a mouse model, in which intraperitoneal 25(OH)D was administered following severe UV-induced skin injury. On immunofluorescence microscopy, mice treated with VD showed enhanced autophagy within the macrophages infiltrating UV-irradiated skin.³⁶ The use of high-dose VD to treat UV-irradiated skin in these studies established a precedent for using VD to heal cutaneous injury caused by ionizing radiation therapy.

Some studies have focused on the role of VD for treating acute radiation dermatitis. A study of 23 patients with ductal carcinoma in situ or localized invasive ductal carcinoma breast cancer compared the effectiveness of topical calcipotriol to that of a standard hydrating ointment.³⁷ Participants were randomized to 1 of 2 treatments before starting adjuvant radiotherapy to evaluate their potential in preventing radiation dermatitis. In 87% (20/23) of these patients, no difference in skin reaction was observed between the 2 treatments, suggesting that topical VD application may not offer any advantage over the standard hydrating ointment for the prevention of radiation dermatitis.³⁷

Benefits of high-dose oral VD for treating radiation dermatitis also have been reported. Nguyen et al³⁸ documented 3 cases in which patients with neuroendocrine carcinoma of the pancreas, tonsillar carcinoma, and breast cancer received 200,000 IU of oral ergocalciferol distributed over 2 doses given 7 days apart for radiation dermatitis. These patients experienced substantial improvements in pain, swelling, and redness within a week of the initial dose. Additionally, a case of radiation recall dermatitis, which occurred a week after vinorelbine chemotherapy, was treated with 2 doses totaling 100,000 IU of oral ergocalciferol. This patient also had improvement in pain and swelling but continued to have tumor-related induration and ulceration.³⁹

Although topical VD did not show significant benefits over standard treatments for radiation dermatitis, high-dose oral VD appears promising in improving patient outcomes of pain and swelling more rapidly than current practices. Further research is needed to confirm these findings and establish standardized treatment protocols.

Final Thoughts

Suboptimal VD levels are prevalent in numerous cancer types. Chemotherapy often is associated with acute, potentially transient worsening of VD status in patients with breast and colorectal cancer. Although 25(OH)D levels have not corresponded with increased frequency of ­chemotherapy-related dermatologic AEs, suboptimal 25(OH)D levels appear to be associated with increased severity of radiation-induced mucositis and dermatitis.^20,25,26 The use of high-dose VD as a therapeutic agent shows promise in mitigating chemotherapy-induced and radiation therapy–induced rashes in multiple cancer types with reduction of inflammatory markers and a durable anti-inflammatory impact. Although the mechanisms of cellular injury vary among chemotherapeutic agents, the anti-inflammatory and tissue repair properties of VD may make it an effective treatment for chemotherapy-induced cutaneous damage regardless of injury mechanism.^2-4,35 However, reports of clinical improvement vary, and further objective studies to classify optimal dosing, administration, and outcome measures are needed. The absence of reported AEs associated with high-dose VD supplementation is encouraging, but selection of a safe and optimal dosing regimen can only occur with dedicated clinical trials.

Hidradenitis suppurativa (HS) is a chronic inflammatory skin condition that affects approximately 0.1% of the US population.^1,2 Severe disease or HS flares can lead patients to seek care through the emergency department (ED), with some requiring inpatient admission. ³ Inpatient hospitalization of patients with HS has increased over the last 2 decades, and patients with HS utilize emergency and inpatient care more frequently than those with other dermatologic conditions.^4,5 Minority patients and those of lower socioeconomic status are more likely to present to the ED for HS management due to limited access to care and other existing comorbid conditions. ⁴ In a 2022 study of the Nationwide Readmissions Database, the authors looked at hospital readmission rates of patients with HS compared with those with heart failure—both patient populations with chronic debilitating conditions. Results indicated that the hospital readmission rates for patients with HS surpassed those of patients with heart failure for that year, highlighting the need for improved inpatient management of HS.⁶

Patients with HS present to the ED with severe pain, fever, wound care, or the need for surgical intervention. The ED and inpatient hospital setting are locations in which physicians may not be as familiar with the diagnosis or treatment of HS, specifically flares or severe disease. ⁷ The inpatient care setting provides access to certain resources that can be challenging to obtain in the outpatient clinical setting, such as social workers and pain specialists, but also can prove challenging in obtaining other resources for HS management, such as advanced medical therapies. Given the increase in hospital- based care for HS and lack of widespread inpatient access to dermatology and HS experts, consensus recommendations for management of HS in the acute hospital setting would be beneficial. In our study, we sought to generate a collection of expert consensus statements providers can refer to when managing patients with HS in the inpatient setting.

Methods

The study team at the Wake Forest University School of Medicine (Winston-Salem, North Carolina)(M.N., R.P., L.C.S.) developed an initial set of consensus statements based on current published HS treatment guidelines,^8,9 publications on management of inpatient HS,³ published supportive care guidelines for Stevens-Johnson syndrome, 10 and personal clinical experience in managing inpatient HS, which resulted in 50 statements organized into the following categories: overall care, wound care, genital care, pain management, infection control, medical management, surgical management, nutrition, and transitional care guidelines. This study was approved by the Wake Forest University institutional review board (IRB00084257).

Participant Recruitment—Dermatologists were identified for participation in the study based on membership in the Society of Dermatology Hospitalists and the Hidradenitis Suppurativa Foundation or authorship of publications relevant to HS or inpatient dermatology. Dermatologists from larger academic institutions with HS specialty clinics and inpatient dermatology services also were identified. Participants were invited via email and could suggest other experts for inclusion. A total of 31 dermatologists were invited to participate in the study, with 26 agreeing to participate. All participating dermatologists were practicing in the United States.

Delphi Study—In the first round of the Delphi study, the participants were sent an online survey via REDCap in which they were asked to rank the appropriateness of each of the proposed 50 guideline statements on a scale of 1 (very inappropriate) to 9 (very appropriate). Participants also were able to provide commentary and feedback on each of the statements. Survey results were analyzed using the RAND/ UCLA Appropriateness Method.¹¹ For each statement, the median rating for appropriateness, interpercentile range (IPR), IPR adjusted for symmetry, and disagreement index (DI) were calculated (DI=IPR/IPR adjusted for symmetry). The 30th and 70th percentiles were used in the DI calculation as the upper and lower limits, respectively. A median rating for appropriateness of 1.0 to 3.9 was considered “inappropriate,” 4.0 to 6.9 was considered “uncertain appropriateness,” and 7.0 to 9.0 was “appropriate.” A DI value greater than or equal to 1 indicated a lack of consensus regarding the appropriateness of the statement. Following each round, participants received a copy of their responses along with the group median rank of each statement. Statements that did not reach consensus in the first Delphi round were revised based on feedback received by the participants, and a second survey with 14 statements was sent via REDCap 2 weeks later. The RAND/UCLA Appropriateness Method also was applied to this second Delphi round. After the second survey, participants received a copy of anonymized comments regarding the consensus statements and were allowed to provide additional final commentary to be included in the discussion of these recommendations.

Results

Twenty-six dermatologists completed the first-round survey, and 24 participants completed the second-round survey. All participants self-identified as having expertise in either HS (n=22 [85%]) or inpatient dermatology (n=17 [65%]), and 13 (50%) participants self-identified as experts in both HS and inpatient dermatology. All participants, except 1, were affiliated with an academic health system with inpatient dermatology services. The average length of time in practice as a dermatologist was 10 years (median, 9 years [range, 3–27 years]).

Of the 50 initial proposed consensus statements, 26 (52%) achieved consensus after the first round; 21 statements revealed DI calculations that did not achieve consensus. Two statements achieved consensus but received median ratings for appropriateness, indicating uncertain appropriateness; because of this, 1 statement was removed and 1 was revised based on participant feedback, resulting in 13 revised statements (eTable 1). Controversial topics in the consensus process included obtaining wound cultures and meaningful culture data interpretation, use of specific biologic medications in the inpatient setting, and use of intravenous ertapenem. Participant responses to these topics are discussed in detail below. Of these secondround statements, all achieved consensus. The final set of consensus statements can be found in eTable 2.

Comment

Our Delphi consensus study combined the expertise of both dermatologists who care for patients with HS and those with inpatient dermatology experience to produce a set of recommendations for the management of HS in the hospital care setting. A strength of this study is inclusion of many national leaders in both HS and inpatient dermatology, with some participants having developed the previously published HS treatment guidelines and others having participated in inpatient dermatology Delphi studies.8-10 The expertise is further strengthened by the geographically diverse institutional representation within the United States.

The final consensus recommendations included 40 statements covering a range of patient care issues, including use of appropriate inpatient subspecialists (care team), supportive care measures (wound care, pain control, genital care), disease-oriented treatment (medical management, surgical management), inpatient complications (infection control, nutrition), and successful transition back to outpatient management (transitional care). These recommendations are meant to serve as a resource for providers to consider when taking care of inpatient HS flares, recognizing that the complexity and individual circumstances of each patient are unique.

Delphi Consensus Recommendations Compared to Prior Guidelines—Several recommendations in the current study align with the previously published North American clinical management guidelines for HS.^8,9 Our recommendations agree with prior guidelines on the importance of disease staging and pain assessment using validated assessment tools as well as screening for HS comorbidities. There also is agreement in the potential benefit of involving pain specialists in the development of a comprehensive pain management plan. The inpatient care setting provides a unique opportunity to engage multiple specialists and collaborate on patient care in a timely manner. Our recommendations regarding surgical care also align with established guidelines in recommending incision and drainage as an acute bedside procedure best utilized for symptom relief in inflamed abscesses and relegating most other surgical management to the outpatient setting. Wound care recommendations also are similar, with our expert participants agreeing on individualizing dressing choices based on wound characteristics. A benefit of inpatient wound care is access to skilled nursing for dressing changes and potentially improved access to more sophisticated dressing materials. Our recommendations differ from the prior guidelines in our focus on severe HS, HS flares, and HS complications, which constitute the majority of inpatient disease management. We provide additional guidance on management of secondary infections, perianal fistulous disease, and importantly transitional care to optimize discharge planning.

Differing Opinions in Our Analysis—Despite the success of our Delphi consensus process, there were some differing opinions regarding certain aspects of inpatient HS management, which is to be expected given the lack of strong evidence-based research to support some of the recommended practices. There were differing opinions on the utility of wound culture data, with some participants feeling culture data could help with antibiotic susceptibility and resistance patterns, while others felt wound cultures represent bacterial colonization or biofilm formation.

Initial consensus statements in the first Delphi round were created for individual biologic medications but did not achieve consensus, and feedback on the use of biologics in the inpatient environment was mixed, largely due to logistic and insurance issues. Many participants felt biologic medication cost, difficulty obtaining inpatient reimbursement, health care resource utilization, and availability of biologics in different hospital systems prevented recommending the use of specific biologics during hospitalization. The one exception was in the case of a hospitalized patient who was already receiving infliximab for HS: there was consensus on ensuring the patient dosing was maximized, if appropriate, to 10 mg/kg.12 Ertapenem use also was controversial, with some participants using it as a bridge therapy to either outpatient biologic use or surgery, while others felt it was onerous and difficult to establish reliable access to secure intravenous administration and regular dosing once the patient left the inpatient setting.¹³ Others said they have experienced objections from infectious disease colleagues on the use of intravenous antibiotics, citing antibiotic stewardship concerns.

Patient Care in the Inpatient Setting—Prior literature suggests patients admitted as inpatients for HS tend to be of lower socioeconomic status and are admitted to larger urban teaching hospitals.^14,15 Patients with lower socioeconomic status have increased difficulty accessing health care resources; therefore, inpatient admission serves as an opportunity to provide a holistic HS assessment and coordinate resources for chronic outpatient management.

Study Limitations—This Delphi consensus study has some limitations. The existing literature on inpatient management of HS is limited, challenging our ability to assess the extent to which these published recommendations are already being implemented. Additionally, the study included HS and inpatient dermatology experts from the United States, which means the recommendations may not be generalizable to other countries. Most participants practiced dermatology at large tertiary care academic medical centers, which may limit the ability to implement recommendations in all US inpatient care settings such as small community-based hospitals; however, many of the supportive care guidelines such as pain control, wound care, nutritional support, and social work should be achievable in most inpatient care settings.

Conclusion

Given the increase in inpatient and ED health care utilization for HS, there is an urgent need for expert consensus recommendations on inpatient management of this unique patient population, which requires complex multidisciplinary care. Our recommendations are a resource for providers to utilize and potentially improve the standard of care we provide these patients.

Acknowledgment—We thank the Wake Forest University Clinical and Translational Science Institute (Winston- Salem, North Carolina) for providing statistical help.

Hidradenitis suppurativa (HS) is a chronic inflammatory skin condition that affects approximately 0.1% of the US population.^1,2 Severe disease or HS flares can lead patients to seek care through the emergency department (ED), with some requiring inpatient admission. ³ Inpatient hospitalization of patients with HS has increased over the last 2 decades, and patients with HS utilize emergency and inpatient care more frequently than those with other dermatologic conditions.^4,5 Minority patients and those of lower socioeconomic status are more likely to present to the ED for HS management due to limited access to care and other existing comorbid conditions. ⁴ In a 2022 study of the Nationwide Readmissions Database, the authors looked at hospital readmission rates of patients with HS compared with those with heart failure—both patient populations with chronic debilitating conditions. Results indicated that the hospital readmission rates for patients with HS surpassed those of patients with heart failure for that year, highlighting the need for improved inpatient management of HS.⁶

Patients with HS present to the ED with severe pain, fever, wound care, or the need for surgical intervention. The ED and inpatient hospital setting are locations in which physicians may not be as familiar with the diagnosis or treatment of HS, specifically flares or severe disease. ⁷ The inpatient care setting provides access to certain resources that can be challenging to obtain in the outpatient clinical setting, such as social workers and pain specialists, but also can prove challenging in obtaining other resources for HS management, such as advanced medical therapies. Given the increase in hospital- based care for HS and lack of widespread inpatient access to dermatology and HS experts, consensus recommendations for management of HS in the acute hospital setting would be beneficial. In our study, we sought to generate a collection of expert consensus statements providers can refer to when managing patients with HS in the inpatient setting.

Methods

The study team at the Wake Forest University School of Medicine (Winston-Salem, North Carolina)(M.N., R.P., L.C.S.) developed an initial set of consensus statements based on current published HS treatment guidelines,^8,9 publications on management of inpatient HS,³ published supportive care guidelines for Stevens-Johnson syndrome, 10 and personal clinical experience in managing inpatient HS, which resulted in 50 statements organized into the following categories: overall care, wound care, genital care, pain management, infection control, medical management, surgical management, nutrition, and transitional care guidelines. This study was approved by the Wake Forest University institutional review board (IRB00084257).

Participant Recruitment—Dermatologists were identified for participation in the study based on membership in the Society of Dermatology Hospitalists and the Hidradenitis Suppurativa Foundation or authorship of publications relevant to HS or inpatient dermatology. Dermatologists from larger academic institutions with HS specialty clinics and inpatient dermatology services also were identified. Participants were invited via email and could suggest other experts for inclusion. A total of 31 dermatologists were invited to participate in the study, with 26 agreeing to participate. All participating dermatologists were practicing in the United States.

Delphi Study—In the first round of the Delphi study, the participants were sent an online survey via REDCap in which they were asked to rank the appropriateness of each of the proposed 50 guideline statements on a scale of 1 (very inappropriate) to 9 (very appropriate). Participants also were able to provide commentary and feedback on each of the statements. Survey results were analyzed using the RAND/ UCLA Appropriateness Method.¹¹ For each statement, the median rating for appropriateness, interpercentile range (IPR), IPR adjusted for symmetry, and disagreement index (DI) were calculated (DI=IPR/IPR adjusted for symmetry). The 30th and 70th percentiles were used in the DI calculation as the upper and lower limits, respectively. A median rating for appropriateness of 1.0 to 3.9 was considered “inappropriate,” 4.0 to 6.9 was considered “uncertain appropriateness,” and 7.0 to 9.0 was “appropriate.” A DI value greater than or equal to 1 indicated a lack of consensus regarding the appropriateness of the statement. Following each round, participants received a copy of their responses along with the group median rank of each statement. Statements that did not reach consensus in the first Delphi round were revised based on feedback received by the participants, and a second survey with 14 statements was sent via REDCap 2 weeks later. The RAND/UCLA Appropriateness Method also was applied to this second Delphi round. After the second survey, participants received a copy of anonymized comments regarding the consensus statements and were allowed to provide additional final commentary to be included in the discussion of these recommendations.

Results

Twenty-six dermatologists completed the first-round survey, and 24 participants completed the second-round survey. All participants self-identified as having expertise in either HS (n=22 [85%]) or inpatient dermatology (n=17 [65%]), and 13 (50%) participants self-identified as experts in both HS and inpatient dermatology. All participants, except 1, were affiliated with an academic health system with inpatient dermatology services. The average length of time in practice as a dermatologist was 10 years (median, 9 years [range, 3–27 years]).

Of the 50 initial proposed consensus statements, 26 (52%) achieved consensus after the first round; 21 statements revealed DI calculations that did not achieve consensus. Two statements achieved consensus but received median ratings for appropriateness, indicating uncertain appropriateness; because of this, 1 statement was removed and 1 was revised based on participant feedback, resulting in 13 revised statements (eTable 1). Controversial topics in the consensus process included obtaining wound cultures and meaningful culture data interpretation, use of specific biologic medications in the inpatient setting, and use of intravenous ertapenem. Participant responses to these topics are discussed in detail below. Of these secondround statements, all achieved consensus. The final set of consensus statements can be found in eTable 2.

Comment

Our Delphi consensus study combined the expertise of both dermatologists who care for patients with HS and those with inpatient dermatology experience to produce a set of recommendations for the management of HS in the hospital care setting. A strength of this study is inclusion of many national leaders in both HS and inpatient dermatology, with some participants having developed the previously published HS treatment guidelines and others having participated in inpatient dermatology Delphi studies.8-10 The expertise is further strengthened by the geographically diverse institutional representation within the United States.

The final consensus recommendations included 40 statements covering a range of patient care issues, including use of appropriate inpatient subspecialists (care team), supportive care measures (wound care, pain control, genital care), disease-oriented treatment (medical management, surgical management), inpatient complications (infection control, nutrition), and successful transition back to outpatient management (transitional care). These recommendations are meant to serve as a resource for providers to consider when taking care of inpatient HS flares, recognizing that the complexity and individual circumstances of each patient are unique.

Delphi Consensus Recommendations Compared to Prior Guidelines—Several recommendations in the current study align with the previously published North American clinical management guidelines for HS.^8,9 Our recommendations agree with prior guidelines on the importance of disease staging and pain assessment using validated assessment tools as well as screening for HS comorbidities. There also is agreement in the potential benefit of involving pain specialists in the development of a comprehensive pain management plan. The inpatient care setting provides a unique opportunity to engage multiple specialists and collaborate on patient care in a timely manner. Our recommendations regarding surgical care also align with established guidelines in recommending incision and drainage as an acute bedside procedure best utilized for symptom relief in inflamed abscesses and relegating most other surgical management to the outpatient setting. Wound care recommendations also are similar, with our expert participants agreeing on individualizing dressing choices based on wound characteristics. A benefit of inpatient wound care is access to skilled nursing for dressing changes and potentially improved access to more sophisticated dressing materials. Our recommendations differ from the prior guidelines in our focus on severe HS, HS flares, and HS complications, which constitute the majority of inpatient disease management. We provide additional guidance on management of secondary infections, perianal fistulous disease, and importantly transitional care to optimize discharge planning.

Differing Opinions in Our Analysis—Despite the success of our Delphi consensus process, there were some differing opinions regarding certain aspects of inpatient HS management, which is to be expected given the lack of strong evidence-based research to support some of the recommended practices. There were differing opinions on the utility of wound culture data, with some participants feeling culture data could help with antibiotic susceptibility and resistance patterns, while others felt wound cultures represent bacterial colonization or biofilm formation.

Initial consensus statements in the first Delphi round were created for individual biologic medications but did not achieve consensus, and feedback on the use of biologics in the inpatient environment was mixed, largely due to logistic and insurance issues. Many participants felt biologic medication cost, difficulty obtaining inpatient reimbursement, health care resource utilization, and availability of biologics in different hospital systems prevented recommending the use of specific biologics during hospitalization. The one exception was in the case of a hospitalized patient who was already receiving infliximab for HS: there was consensus on ensuring the patient dosing was maximized, if appropriate, to 10 mg/kg.12 Ertapenem use also was controversial, with some participants using it as a bridge therapy to either outpatient biologic use or surgery, while others felt it was onerous and difficult to establish reliable access to secure intravenous administration and regular dosing once the patient left the inpatient setting.¹³ Others said they have experienced objections from infectious disease colleagues on the use of intravenous antibiotics, citing antibiotic stewardship concerns.

Patient Care in the Inpatient Setting—Prior literature suggests patients admitted as inpatients for HS tend to be of lower socioeconomic status and are admitted to larger urban teaching hospitals.^14,15 Patients with lower socioeconomic status have increased difficulty accessing health care resources; therefore, inpatient admission serves as an opportunity to provide a holistic HS assessment and coordinate resources for chronic outpatient management.

Study Limitations—This Delphi consensus study has some limitations. The existing literature on inpatient management of HS is limited, challenging our ability to assess the extent to which these published recommendations are already being implemented. Additionally, the study included HS and inpatient dermatology experts from the United States, which means the recommendations may not be generalizable to other countries. Most participants practiced dermatology at large tertiary care academic medical centers, which may limit the ability to implement recommendations in all US inpatient care settings such as small community-based hospitals; however, many of the supportive care guidelines such as pain control, wound care, nutritional support, and social work should be achievable in most inpatient care settings.

Conclusion

Given the increase in inpatient and ED health care utilization for HS, there is an urgent need for expert consensus recommendations on inpatient management of this unique patient population, which requires complex multidisciplinary care. Our recommendations are a resource for providers to utilize and potentially improve the standard of care we provide these patients.

Acknowledgment—We thank the Wake Forest University Clinical and Translational Science Institute (Winston- Salem, North Carolina) for providing statistical help.

Hidradenitis suppurativa (HS) is a chronic inflammatory skin condition that affects approximately 0.1% of the US population.^1,2 Severe disease or HS flares can lead patients to seek care through the emergency department (ED), with some requiring inpatient admission. ³ Inpatient hospitalization of patients with HS has increased over the last 2 decades, and patients with HS utilize emergency and inpatient care more frequently than those with other dermatologic conditions.^4,5 Minority patients and those of lower socioeconomic status are more likely to present to the ED for HS management due to limited access to care and other existing comorbid conditions. ⁴ In a 2022 study of the Nationwide Readmissions Database, the authors looked at hospital readmission rates of patients with HS compared with those with heart failure—both patient populations with chronic debilitating conditions. Results indicated that the hospital readmission rates for patients with HS surpassed those of patients with heart failure for that year, highlighting the need for improved inpatient management of HS.⁶

Patients with HS present to the ED with severe pain, fever, wound care, or the need for surgical intervention. The ED and inpatient hospital setting are locations in which physicians may not be as familiar with the diagnosis or treatment of HS, specifically flares or severe disease. ⁷ The inpatient care setting provides access to certain resources that can be challenging to obtain in the outpatient clinical setting, such as social workers and pain specialists, but also can prove challenging in obtaining other resources for HS management, such as advanced medical therapies. Given the increase in hospital- based care for HS and lack of widespread inpatient access to dermatology and HS experts, consensus recommendations for management of HS in the acute hospital setting would be beneficial. In our study, we sought to generate a collection of expert consensus statements providers can refer to when managing patients with HS in the inpatient setting.

Methods

The study team at the Wake Forest University School of Medicine (Winston-Salem, North Carolina)(M.N., R.P., L.C.S.) developed an initial set of consensus statements based on current published HS treatment guidelines,^8,9 publications on management of inpatient HS,³ published supportive care guidelines for Stevens-Johnson syndrome, 10 and personal clinical experience in managing inpatient HS, which resulted in 50 statements organized into the following categories: overall care, wound care, genital care, pain management, infection control, medical management, surgical management, nutrition, and transitional care guidelines. This study was approved by the Wake Forest University institutional review board (IRB00084257).

Participant Recruitment—Dermatologists were identified for participation in the study based on membership in the Society of Dermatology Hospitalists and the Hidradenitis Suppurativa Foundation or authorship of publications relevant to HS or inpatient dermatology. Dermatologists from larger academic institutions with HS specialty clinics and inpatient dermatology services also were identified. Participants were invited via email and could suggest other experts for inclusion. A total of 31 dermatologists were invited to participate in the study, with 26 agreeing to participate. All participating dermatologists were practicing in the United States.

Delphi Study—In the first round of the Delphi study, the participants were sent an online survey via REDCap in which they were asked to rank the appropriateness of each of the proposed 50 guideline statements on a scale of 1 (very inappropriate) to 9 (very appropriate). Participants also were able to provide commentary and feedback on each of the statements. Survey results were analyzed using the RAND/ UCLA Appropriateness Method.¹¹ For each statement, the median rating for appropriateness, interpercentile range (IPR), IPR adjusted for symmetry, and disagreement index (DI) were calculated (DI=IPR/IPR adjusted for symmetry). The 30th and 70th percentiles were used in the DI calculation as the upper and lower limits, respectively. A median rating for appropriateness of 1.0 to 3.9 was considered “inappropriate,” 4.0 to 6.9 was considered “uncertain appropriateness,” and 7.0 to 9.0 was “appropriate.” A DI value greater than or equal to 1 indicated a lack of consensus regarding the appropriateness of the statement. Following each round, participants received a copy of their responses along with the group median rank of each statement. Statements that did not reach consensus in the first Delphi round were revised based on feedback received by the participants, and a second survey with 14 statements was sent via REDCap 2 weeks later. The RAND/UCLA Appropriateness Method also was applied to this second Delphi round. After the second survey, participants received a copy of anonymized comments regarding the consensus statements and were allowed to provide additional final commentary to be included in the discussion of these recommendations.

Results

Twenty-six dermatologists completed the first-round survey, and 24 participants completed the second-round survey. All participants self-identified as having expertise in either HS (n=22 [85%]) or inpatient dermatology (n=17 [65%]), and 13 (50%) participants self-identified as experts in both HS and inpatient dermatology. All participants, except 1, were affiliated with an academic health system with inpatient dermatology services. The average length of time in practice as a dermatologist was 10 years (median, 9 years [range, 3–27 years]).

Of the 50 initial proposed consensus statements, 26 (52%) achieved consensus after the first round; 21 statements revealed DI calculations that did not achieve consensus. Two statements achieved consensus but received median ratings for appropriateness, indicating uncertain appropriateness; because of this, 1 statement was removed and 1 was revised based on participant feedback, resulting in 13 revised statements (eTable 1). Controversial topics in the consensus process included obtaining wound cultures and meaningful culture data interpretation, use of specific biologic medications in the inpatient setting, and use of intravenous ertapenem. Participant responses to these topics are discussed in detail below. Of these secondround statements, all achieved consensus. The final set of consensus statements can be found in eTable 2.

Comment

Our Delphi consensus study combined the expertise of both dermatologists who care for patients with HS and those with inpatient dermatology experience to produce a set of recommendations for the management of HS in the hospital care setting. A strength of this study is inclusion of many national leaders in both HS and inpatient dermatology, with some participants having developed the previously published HS treatment guidelines and others having participated in inpatient dermatology Delphi studies.8-10 The expertise is further strengthened by the geographically diverse institutional representation within the United States.

The final consensus recommendations included 40 statements covering a range of patient care issues, including use of appropriate inpatient subspecialists (care team), supportive care measures (wound care, pain control, genital care), disease-oriented treatment (medical management, surgical management), inpatient complications (infection control, nutrition), and successful transition back to outpatient management (transitional care). These recommendations are meant to serve as a resource for providers to consider when taking care of inpatient HS flares, recognizing that the complexity and individual circumstances of each patient are unique.

Delphi Consensus Recommendations Compared to Prior Guidelines—Several recommendations in the current study align with the previously published North American clinical management guidelines for HS.^8,9 Our recommendations agree with prior guidelines on the importance of disease staging and pain assessment using validated assessment tools as well as screening for HS comorbidities. There also is agreement in the potential benefit of involving pain specialists in the development of a comprehensive pain management plan. The inpatient care setting provides a unique opportunity to engage multiple specialists and collaborate on patient care in a timely manner. Our recommendations regarding surgical care also align with established guidelines in recommending incision and drainage as an acute bedside procedure best utilized for symptom relief in inflamed abscesses and relegating most other surgical management to the outpatient setting. Wound care recommendations also are similar, with our expert participants agreeing on individualizing dressing choices based on wound characteristics. A benefit of inpatient wound care is access to skilled nursing for dressing changes and potentially improved access to more sophisticated dressing materials. Our recommendations differ from the prior guidelines in our focus on severe HS, HS flares, and HS complications, which constitute the majority of inpatient disease management. We provide additional guidance on management of secondary infections, perianal fistulous disease, and importantly transitional care to optimize discharge planning.

Differing Opinions in Our Analysis—Despite the success of our Delphi consensus process, there were some differing opinions regarding certain aspects of inpatient HS management, which is to be expected given the lack of strong evidence-based research to support some of the recommended practices. There were differing opinions on the utility of wound culture data, with some participants feeling culture data could help with antibiotic susceptibility and resistance patterns, while others felt wound cultures represent bacterial colonization or biofilm formation.

Initial consensus statements in the first Delphi round were created for individual biologic medications but did not achieve consensus, and feedback on the use of biologics in the inpatient environment was mixed, largely due to logistic and insurance issues. Many participants felt biologic medication cost, difficulty obtaining inpatient reimbursement, health care resource utilization, and availability of biologics in different hospital systems prevented recommending the use of specific biologics during hospitalization. The one exception was in the case of a hospitalized patient who was already receiving infliximab for HS: there was consensus on ensuring the patient dosing was maximized, if appropriate, to 10 mg/kg.12 Ertapenem use also was controversial, with some participants using it as a bridge therapy to either outpatient biologic use or surgery, while others felt it was onerous and difficult to establish reliable access to secure intravenous administration and regular dosing once the patient left the inpatient setting.¹³ Others said they have experienced objections from infectious disease colleagues on the use of intravenous antibiotics, citing antibiotic stewardship concerns.

Patient Care in the Inpatient Setting—Prior literature suggests patients admitted as inpatients for HS tend to be of lower socioeconomic status and are admitted to larger urban teaching hospitals.^14,15 Patients with lower socioeconomic status have increased difficulty accessing health care resources; therefore, inpatient admission serves as an opportunity to provide a holistic HS assessment and coordinate resources for chronic outpatient management.

Study Limitations—This Delphi consensus study has some limitations. The existing literature on inpatient management of HS is limited, challenging our ability to assess the extent to which these published recommendations are already being implemented. Additionally, the study included HS and inpatient dermatology experts from the United States, which means the recommendations may not be generalizable to other countries. Most participants practiced dermatology at large tertiary care academic medical centers, which may limit the ability to implement recommendations in all US inpatient care settings such as small community-based hospitals; however, many of the supportive care guidelines such as pain control, wound care, nutritional support, and social work should be achievable in most inpatient care settings.

Conclusion

Given the increase in inpatient and ED health care utilization for HS, there is an urgent need for expert consensus recommendations on inpatient management of this unique patient population, which requires complex multidisciplinary care. Our recommendations are a resource for providers to utilize and potentially improve the standard of care we provide these patients.

Acknowledgment—We thank the Wake Forest University Clinical and Translational Science Institute (Winston- Salem, North Carolina) for providing statistical help.

In 2023, ESPEN (the European Society for Clinical Nutrition and Metabolism) published consensus recommendations highlighting the importance of regular monitoring and treatment of nutrient deficiencies in patients with inflammatory bowel disease (IBD) for improved prognosis, mortality, and quality of life.¹ Suboptimal nutrition in patients with IBD predominantly results from inflammation of the gastrointestinal (GI) tract leading to malabsorption; however, medications commonly used to manage IBD also can contribute to malnutrition.^2,3 Additionally, patients may develop nausea and food avoidance due to medication or the disease itself, leading to nutritional withdrawal and eventual deficiency.⁴ Even with the development of diets focused on balancing nutritional needs and decreasing inflammation,⁵ offsetting this aversion to food can be difficult to overcome.²

Cutaneous manifestations of IBD are multifaceted and can be secondary to the disease, reactive to or associated with IBD, or effects from nutritional deficiencies. The most common vitamin and nutrient deficiencies in patients with IBD include iron; zinc; calcium; vitamin D; and vitamins B₆ (pyridoxine), B₉ (folic acid), and B₁₂.⁶ Malnutrition may manifest with cutaneous disease, and dermatologists can be the first to identify and assess for nutritional deficiencies. In this article, we review the mechanisms of these micronutrient depletions in the context of IBD, their subsequent dermatologic manifestations (Table), and treatment and monitoring guidelines for each deficiency.

Iron

A systematic review conducted from 2007 to 2012 in European patients with IBD (N=2192) found the overall prevalence of anemia in this population to be 24% (95% CI, 18%-31%), with 57% of patients with anemia experiencing iron deficiency.⁷ Anemia is observed more commonly in patients hospitalized with IBD and is common in patients with both Crohn disease and ulcerative colitis.⁸

Pathophysiology—Iron is critically important in oxygen transportation throughout the body as a major component of hemoglobin. Physiologically, the low pH of the duodenum and proximal jejunum allows divalent metal transporter 1 to transfer dietary Fe³⁺ into enterocytes, where it is reduced to the transportable Fe²⁺.^9,10 Distribution of Fe²⁺ ions from enterocytes relies on ferroportin, an iron-transporting protein, which is heavily regulated by the protein hepcidin.¹¹ Hepcidin, a known acute phase reactant, will increase in the setting of active IBD, causing a depletion of ferroportin and an inability of the body to utilize the stored iron in enterocytes.¹² This poor utilization of iron stores combined with blood loss caused by inflammation in the GI tract is the proposed primary mechanism of iron-deficiency anemia observed in patients with IBD.¹³

Cutaneous Manifestations—From a dermatologic perspective, iron-deficiency anemia can manifest with a wide range of symptoms including glossitis, koilonychia, xerosis and/or pruritus, and brittle hair or hair loss.^14,15 Although the underlying pathophysiology of these cutaneous manifestations is not fully understood, there are several theories assessing the mechanisms behind the skin findings of iron deficiency.

Atrophic glossitis has been observed in many patients with iron deficiency and is thought to manifest due to low iron concentrations in the blood, thereby decreasing oxygen delivery to the papillae of the dorsal tongue with resultant atrophy.^16,17 Similarly, decreased oxygen delivery to the nail bed capillaries may cause deformities in the nail called koilonychia (or “spoon nails”).¹⁸ Iron is a key co-factor in collagen lysyl hydroxylase that promotes collagen binding; iron deficiency may lead to disruptions in the epidermal barrier that can cause pruritus and xerosis.¹⁹ An observational study of 200 healthy patients with a primary concern of pruritus found a correlation between low serum ferritin and a higher degree of pruritus (r=−0.768; P<.00001).²⁰

Evidence for iron’s role in hair growth comes from a mouse model study with a mutation in the serine protease TMPRSS6—a protein that regulates hepcidin and iron absorption—which caused an increase in hepcidin production and subsequent systemic iron deficiency. Mice at 4 weeks of age were devoid of all body hair but had substantial regrowth after initiation of a 2-week iron-rich diet, which suggests a connection between iron repletion and hair growth in mice with iron deficiency.²¹ Additionally, a meta-analysis analyzing the comorbidities of patients with alopecia areata found them to have higher odds (odds ratio [OR]=2.78; 95% CI, 1.23-6.29) of iron-deficiency anemia but no association with IBD (OR=1.48; 95% CI, 0.32-6.82).²²

Diagnosis and Monitoring—The American Gastroenterological Association recommends a complete blood cell count (CBC), serum ferritin, transferrin saturation (TfS), and C-reactive protein (CRP) as standard evaluations for iron deficiency in patients with IBD. Patients with active IBD should be screened every 3 months,and patients with inactive disease should be screened every 6 to 12 months.²³

Although ferritin and TfS often are used as markers for iron status in healthy individuals, they are positive and negative acute phase reactants, respectively. Using them to assess iron status in patients with IBD may inaccurately represent iron status in the setting of inflammation from the disease.²⁴ The European Crohn’s and Colitis Organisation (ECCO) produced guidelines to define iron deficiency as a TfS less than 20% or a ferritin level less than 30 µg/L in patients without evidence of active IBD and a ferritin level less than 100 µg/L for patients with active inflammation.²⁵

A 2020 multicenter observational study of 202 patients with diagnosed IBD found that the ECCO guideline of ferritin less than 30 µg/L had an area under the receiver operating characteristic (AUROC) curve of 0.69, a sensitivity of 0.43, and a specificity of 0.95 in their population.²⁶ In a sensitivity analysis stratifying patients by CRP level (<10 or ≥10 mg/L), the authors found that for patients with ulcerative colitis and a CRP less than 10 mg/L, a cut-off value of ferritin less than 65 µg/L (AUROC=0.78) had a sensitivity of 0.78 and specificity of 0.76, and a TfS value of less than 16% (AUROC=0.88) had a sensitivity of 0.79 and a specificity of 0.9. In patients with a CRP of 10 mg/L or greater, a cut-off value of ferritin 80 µg/L (AUROC=0.76) had a sensitivity of 0.75 and a specificity of 0.82, and a TfS value of less than 11% (AUROC=0.69) had a sensitivity of 0.79 and a specificity of 0.88. There were no ferritin cut-off values associated with good diagnostic performance (defined as both sensitivity and specificity >0.70) for iron deficiency in patients with Crohn disease.²⁶

The authors recommended using an alternative iron measurement such as soluble transferrin receptor (sTfR)/log ferritin ratio (TfR-F) that is not influenced by active inflammation and has a good correlation with ferritin values (TfR-F: r=0.66; P<.001).²⁶ However, both sTfR and TfR-F have high costs and intermethod variability as well as differences in their reference ranges depending on which laboratory performs the analysis, limiting the accessibility and practicality of easily obtaining these tests.²⁷ Although there may be inaccuracies for standard ferritin or TfS under ECCO guidelines, proposed alternatives have their own limitations, which may make ferritin and TfS the most reasonable evaluations of iron status as long as disease activity status at the time of testing is taken into consideration.

Treatment—Treatment of underlying iron deficiency in patients with IBD requires reversing the cause of the deficiency and supplementing iron. In patients with IBD, the options to supplement iron may be limited by active disease, making oral intake less effective. Oral iron supplementation also is associated with notable GI adverse effects that may be exacerbated in patients with IBD. A systematic review of 43 randomized controlled trials (RCTs) evaluating GI adverse effects (eg, nausea, abdominal pain, diarrhea, constipation, and black or tarry stools) of oral ferrous sulfate compared with placebo or intravenous (IV) iron supplementation in healthy nonanemic individuals found a significant increase in GI adverse effects with oral supplementation (placebo: OR=2.32; P<.0001; IV: OR=3.05; P<.0001).²⁸

Therefore, IV iron repletion may be necessary in patients with IBD and may require numerous infusions depending on the formulation of iron. In an RCT conducted in 2011, patients with iron-deficiency anemia with quiescent or mild to moderate IBD were treated with either IV iron sulfate or ferric carboxymaltose.²⁹ With a primary end point of hemoglobin response greater than 2 g/dL, the authors found that 150 of 240 patients responded to ferric carboxymaltose vs 118 of 235 treated with iron sulfate (P=.004). The dosing for ferric carboxymaltose was 1 to 3 infusions of 500 to 1000 mg of iron and for iron sulfate up to 11 infusions of 200 mg of iron.²⁹

Zinc

A systematic review of zinc deficiency in patients with IBD identified 7 studies including 2413 patients and revealed those with Crohn disease had a higher prevalence of zinc deficiency compared with patients with ulcerative colitis (54% vs 41%).³⁰

Pathophysiology—Zinc serves as a catalytic cofactor for enzymatic activity within proteins and immune cells.³¹ The homeostasis of zinc is tightly regulated within the brush border of the small intestine by zinc transporters ZIP4 and ZIP1 from the lumen of enterocytes into the bloodstream.³² Inflammation in the small intestine due to Crohn disease can result in zinc malabsorption.

Ranaldi et al³³ exposed intestinal cells and zinc-depleted intestinal cells to tumor necrosis factor α media to simulate an inflammatory environment. They measured transepithelial electrical resistance as a surrogate for transmembrane permeability and found that zinc-depleted cells had a statistically significantly higher transepithelial electrical resistance percentage (60% reduction after 4 hours; P<1.10^–6) when exposed to tumor necrosis factor α signaling compared with normal intestinal cells. They concluded that zinc deficiency can increase intestinal permeability in the presence of inflammation, creating a cycle of further nutrient malabsorption and inflammation exacerbating IBD symptoms.³³

Cutaneous Manifestations—After absorption in the small intestine, approximately 5% of zinc resides in the skin, with the highest concentration in the stratum spinosum.³⁴ A cell study found that keratinocytes in zinc-deficient environments had higher rates of apoptosis compared with cells in normal media. The authors proposed that this higher rate of apoptosis and the resulting inflammation could be a mechanism for developing the desquamative or eczematous scaly plaques that are common cutaneous manifestations of zinc deficiency.³⁵

Other cutaneous findings may include angular cheilitis, stomatitis, glossitis, paronychia, onychodystrophy, generalized alopecia, and delayed wound healing.³⁶ The histopathology of these skin lesions is characterized by granular layer loss, epidermal pallor, confluent parakeratosis, spongiosis, dyskeratosis, and psoriasiform hyperplasia.³⁷

Diagnosis and Monitoring—Assessing serum zinc levels is challenging, as they may decrease during states of inflammation.³⁸ A mouse model study showed a 3.1-fold increase (P<.001) in ZIP14 expression in wild-type mice compared with an IL-6 -/- knock-down model after IL-6 exposure. The authors concluded that the upregulation of ZIP14 in the liver due to inflammatory cytokine upregulation decreases zinc availability in serum.³⁹ Additionally, serum zinc can overestimate the level of deficiency in IBD because approximately 75% of serum zinc is bound to albumin, which decreases in the setting of inflammation.^40-42

Alternatively, alkaline phosphatase (AP), a zinc-dependent metalloenzyme, may be a better evaluator of zinc status during periods of inflammation. A study in rats evaluated zinc through serum zinc levels and AP levels after a period of induced stress to mimic a short-term inflammatory state.⁴³ The researchers found that total body stores of zinc were unaffected throughout the experiment; only serum zinc declined throughout the experiment duration while AP did not. Because approximately 75% of serum zinc is bound to serum albumin,⁴² the researchers concluded the induced inflammatory state depleted serum albumin and redistributed zinc to the liver, causing the observed serum zinc changes, while total body zinc levels and AP were largely unaffected in comparison.⁴³ Comorbid conditions such as liver or bone disease can increase AP levels, which limits the utility of AP as a surrogate for zinc in patients with comorbidities.⁴⁴ However, even in the context of active IBD, serum zinc still is currently considered the best biomarker to evaluate zinc status.⁴⁵

Treatment—The recommended dose for zinc supplementation is 20 to 40 mg daily with higher doses (>50 mg/d) for patients with malabsorptive syndromes such as IBD.⁴⁶ It can be administered orally or parenterally. Although rare, zinc replacement therapy may be associated with diarrhea, nausea, vomiting, mild headaches, and fatigue.⁴⁶ Additional considerations should be taken when repleting other micronutrients with zinc, as calcium and folate can inhibit zinc reabsorption, while zinc itself can inhibit iron and copper reabsorption.⁴⁷

Vitamin D and Calcium

Low vitamin D levels (<50 nmol/L) and hypocalcemia (<8.8 mg/dL) are common in patients with IBD.^48,49

Pathophysiology—Vitamin D levels are maintained via 2 mechanisms. The first mechanism is through the skin, as keratinocytes produce 7-dehydrocholesterol after exposure to UV light, which is converted into previtamin D₃ and then thermally isomerizes into vitamin D₃. This vitamin D₃ is then transported to the liver on vitamin D–binding protein.⁵⁰ The second mechanism is through oral vitamin D₃ that is absorbed through vitamin D receptors in intestinal epithelium and transported to the liver, where it is hydroxylated into 25-hydroxyvitamin D (25[OH]D), then to the kidneys for hydroxylation to 1,25(OH)₂D for redistribution throughout the body.⁵⁰ This activated form of vitamin D regulates calcium absorption in the intestine, and optimal vitamin D levels are necessary to absorb calcium efficiently.⁵¹ Inflammation from IBD within the small intestine can downregulate vitamin D receptors, causing malabsorption and decreased serum vitamin D.⁵²

Vitamin D signaling also is vital to maintaining the tight junctions and adherens junctions of the intestinal epithelium. Weakening the permeability of the epithelium further exacerbates malabsorption and subsequent vitamin D deficiency.⁵² A meta-analysis of 27 studies including 8316 patients with IBD showed low vitamin D levels were associated with increased odds of disease activity (OR=1.53; 95% CI, 1.32-1.77), mucosal inflammation (OR=1.25; 95% CI, 1.06-1.47), and future clinical relapse (OR=1.23; 95% CI, 1.03-1.47) in patients with Crohn disease. The authors concluded that low levels of vitamin D could be used as a potential biomarker of inflammatory status in Crohn disease.⁵³

Vitamin D and calcium are further implicated in maintaining skeletal health,⁴⁷ while vitamin D specifically helps maintain intestinal homeostasis⁵⁴ and immune system modulation in the skin.⁵⁵

Cutaneous Manifestations—Vitamin D is thought to play crucial roles in skin differentiation and proliferation, cutaneous innate immunity, hair follicle cycling, photoprotection, and wound healing.⁵⁶ Vitamin D deficiency has been observed in a large range of cutaneous diseases including skin cancer, psoriasis, vitiligo, bullous pemphigoid, atopic dermatitis, and various types of alopecia.^56-59 It is unclear whether vitamin D deficiency facilitates these disease processes or is merely the consequence of a disrupted cutaneous surface with the inability to complete the first step in vitamin D processing. A 2014 meta-analysis of 290 prospective cohort studies and 172 randomized trials concluded that 25(OH)D deficiency was associated with ill health and did not find causal evidence for any specific disease, dermatologic or otherwise.⁶⁰ Calcium deficiency may cause epidermal changes including dry skin, coarse hair, and brittle nails.⁶¹

Diagnosis and Monitoring—The ECCO guidelines recommend obtaining serum 25(OH)D levels every 3 months in patients with IBD.⁶² Levels less than 75 nmol/L are considered deficient, and a value less than 30 nmol/L increases the risk for osteomalacia and nutritional rickets, constituting severe vitamin D deficiency.^63-65

An observational study of 325 patients with IBD showed a statistically significant negative correlation between serum vitamin D and fecal calprotectin (r=−0.19; P<.001), a stool-based marker for gut inflammation, supporting vitamin D as a potential biomarker in IBD.⁶⁶

Evaluation of calcium can be done through serum levels in patients with IBD.⁶⁷ Patients with IBD are at risk for hypoalbuminemia; therefore, consideration should be taken to ensure calcium levels are corrected, as approximately 50% of calcium is bound to albumin or other ions in the body,⁶⁸ which can be done by adjusting the calcium concentration by 0.02 mmol/L for every 1 g/L of albumin above or below 40 g/L. In the most critically ill patients, a direct ionized calcium blood level should be used instead because the previously mentioned correction calculations are inaccurate when albumin is critically low.⁶⁹

Treatment—The ECCO guidelines recommend calcium and vitamin D repletion of 500 to 1000 mg and 800 to 1000 U, respectively, in patients with IBD on systemic corticosteroids to prevent the negative effects of bone loss.⁶² Calcium repletion in patients with IBD who are not on systemic steroids are the same as for the general population.⁶⁵

Vitamin D repletion also may help decrease IBD activity. In a prospective study, 10,000 IU/d of vitamin D in 10 patients with IBD—adjusted over 12 weeks to a target of 100 to 125 nmol/L of serum 25(OH)D—showed a significant reduction in clinical Crohn activity (P=.019) over the study period.⁷⁰ In contrast, 2000 IU/d for 3 months in an RCT of 27 patients with Crohn disease found significantly lower CRP (P=.019) and significantly higher self-reported quality of life (P=.037) but nonsignificant decreases in Crohn activity (P=.082) in patients with 25(OH)D levels of 75 nmol/L or higher compared with those with 25(OH)D levels less than 75 nmol/L.⁷¹

These discrepancies illustrate the need for expanded clinical trials to elucidate the optimal vitamin D dosing for patients with IBD. Ultimately, assessing vitamin D and calcium status and considering repletion in patients with IBD, especially those with comorbid dermatologic diseases such as poor wound healing, psoriasis, or atopic dermatitis, is important.

Vitamin B₆ (Pyridoxine)

Pathophysiology—Pyridoxine is an important coenzyme for many functions including amino acid transamination, fatty acid metabolism, and conversion of tryptophan to niacin. It is absorbed in the jejunum and ileum and subsequently transported to the liver for rephosphorylation and release into its active form.³⁶ An observational study assessing the nutritional status of patients with IBD found that only 5.7% of 105 patients with food records had inadequate dietary intake of pyridoxine, but 29% of all patients with IBD had subnormal pyridoxine levels.⁷² Additionally, they found no significant difference in the prevalence of subnormal pyridoxine levels in patients with active IBD vs IBD in remission. The authors suggested that the subnormal pyridoxine levels in patients with IBD likely were multifactorial and resulted from malabsorption due to active disease, inflammation, and inadequate intake.⁷²

Cutaneous Manifestations—Cutaneous findings associated with pyridoxine deficiency include periorificial and perineal dermatitis,⁷³ angular stomatitis, and cheilitis with associated burning, redness, and tongue edema.³⁶ Additionally, pyridoxine is involved in the conversion of tryptophan to niacin, and its deficiency may manifest with pellagralike findings.⁷⁴

Because pyridoxine is critical to protein metabolism, its deficiency may disrupt key cellular structures that rely on protein concentrations to maintain structural integrity. One such structure in the skin that heavily relies on protein concentrations is the ground substance of the extracellular matrix—the amorphous gelatinous spaces that occupy the areas between the extracellular matrix, which consists of cross-linked glycosaminoglycans and proteins.⁷⁵ Without protein, ground substance increases in viscosity and can disrupt the epidermal barrier, leading to increased transepidermal water loss and ultimately inflammation.⁷⁶ Although this theory has yet to be validated fully, this is a potential mechanistic explanation for the inflammation in dermal papillae that leads to dermatitis observed in pyridoxine deficiency.

Diagnosis and Monitoring—Direct biomarkers of pyridoxine status are in serum, plasma, erythrocytes, and urine, with the most common measurement in plasma as pyridoxal 5′-phosphate (PLP).⁷⁷ Plasma PLP concentrations lower than 20 nmol/L are suggestive of deficiency.⁷⁸ Plasma PLP has shown inverse relationships with acute phase inflammatory markers CRP⁷⁹ and AP,⁷⁸ thereby raising concerns for its validity to assess pyridoxine status in patients with symptomatic IBD.⁸⁰

Alternative evaluations of pyridoxine include tryptophan and methionine loading tests,³⁶ which are measured via urinary excretion and require normal kidney function to be accurate. They should be considered in IBD if necessary, but routine testing, even in patients with symptomatic IBD, is not recommended in the ECCO guidelines. Additional considerations should be taken in patients with altered nutrient requirements such as those who have undergone bowel resection due to highly active disease or those who receive parenteral nutritional supplementation.⁸¹

Treatment—Recommendations for oral pyridoxine supplementation range from 25 to 600 mg daily,⁸² with symptoms typically improving on 100 mg daily.³⁶ Pyridoxine supplementation may have additional benefits for patients with IBD and potentially modulate disease severity. An IL-10 knockout mouse supplemented with pyridoxine had an approximately 60% reduction (P<.05) in inflammation compared to mice deficient in pyridoxine.⁸³ The authors suggest that PLP-dependent enzymes can inhibit further proinflammatory signaling and T-cell migration that can exacerbate IBD. Ultimately, more data is needed before determining the efficacy of pyridoxine supplementation for active IBD.

Vitamin B₁₂ and Vitamin B₉ (Folic Acid)

Pathophysiology—Vitamin B₁₂ is reabsorbed in the terminal ileum, the distal portion of the small intestine. The American Gastroenterological Association recommends that patients with a history of extensive ileal disease or prior ileal surgery, which is the case for many patients with Crohn disease, be monitored for vitamin B₁₂ deficiency.²³ Monitoring and rapid supplementation of vitamin B₁₂ can prevent pernicious anemia and irreversible neurologic damage that may result from deficiency.⁸⁴

Folic acid is primarily absorbed in the duodenum and jejunum of the small intestine. A meta-analysis performed in 2017 assessed studies observing folic acid and vitamin B₁₂ levels in 1086 patients with IBD compared with 1484 healthy controls and found an average difference in serum folate concentration of 0.46 nmol/L (P<.001).⁸⁴ Interestingly, this study did not find a significant difference in serum vitamin B₁₂ levels between patients with IBD and healthy controls, highlighting the mechanism of vitamin B₁₂ deficiency in IBD because only patients with terminal ileal involvement are at risk for malabsorption and subsequent deficiency.

Cutaneous Manifestations—Both vitamin B₁₂ and folic acid deficiency can manifest as cheilitis, glossitis, and/or generalized hyperpigmentation that is accentuated in the flexural areas, palms, soles, and oral cavity.^85,86 Systemic symptoms of patients with vitamin B₁₂ and folic acid deficiency include megaloblastic anemia, pallor, and fatigue. A potential mechanism for the hyperpigmentation observed from vitamin B₁₂ deficiency came from an electron microscope study that showed an increased concentration of melanosomes in a patient with deficiency.⁸⁷

Diagnosis and Monitoring—In patients with suspected vitamin B₁₂ and/or folic acid deficiency, initial evaluation should include a CBC with peripheral smear and serum vitamin B₁₂ and folate levels. In cases for which the diagnosis still is unclear after initial testing, methylmalonic acid and homocysteine levels can help differentiate between the 2 deficiencies. Methylmalonic acid classically is elevated (>260 nmol/L) in vitamin B₁₂ deficiency but not in folate deficiency.⁸⁸ Cut-off values for vitamin B₁₂ deficiency are less than 200 to 250 pg/mL forserum vitamin B₁₂ and/or an elevated level of methylmalonic acid (>0.271 µmol/L).⁸⁹ A serum folic acid value greater than 3 ng/mL and/or erythrocyte folate concentrations greater than 140 ng/mL are considered adequate, whereas an indicator of folic acid deficiency is a homocysteine level less than 10 µmol/L.⁹⁰ A CBC can screen for macrocytic megaloblastic anemias (mean corpuscular volume >100 fl), which are classic diagnostic signs of an underlying vitamin B₁₂ or folate deficiency.

Treatment—According to the Centers for Disease Control and Prevention, supplementation of vitamin B₁₂ can be done orally with 1000 µg daily in patients with deficiency. In patients with active IBD, oral reabsorption of vitamin B₁₂ can be less effective, making subcutaneous or intramuscular administration (1000 µg/wk for 8 weeks, then monthly for life) better options.⁸⁹

Patients with IBD managed with methotrexate should be screened carefully for folate deficiency. Methotrexate is a folate analog that sometimes is used for the treatment of IBD. Reversible competitive inhibition of dihydrofolate reductase can precipitate a systemic folic acid decrease.⁹¹ Typically, oral folic acid (1 to 5 mg/d) is sufficient to treat folate deficiency, with the ESPEN recommending 5 mg once weekly 24 to 72 hours after methotrexate treatment or 1 mg daily for 5 days per week in patients with IBD.¹ Alternative formulations—IV, subcutaneous, or intramuscular—are available for patients who cannot tolerate oral intake.⁹²

Final Thoughts

Dermatologists can be the first to observe the cutaneous manifestations of micronutrient deficiencies. Although the symptoms of each micronutrient deficiency discussed may overlap, attention to small clinical clues in patients with IBD can improve patient outcomes and quality of life. For example, koilonychia with glossitis and xerosis likely is due to iron deficiency, while zinc deficiency should be suspected in patients with scaly eczematous plaques in skin folds. A high level of suspicion for micronutrient deficiencies in patients with IBD should be followed by a complete patient history, review of systems, and thorough clinical examination. A thorough laboratory evaluation can pinpoint nutritional deficiencies in patients with IBD, keeping in mind that specific biomarkers such as ferritin and serum zinc also act as acute phase reactants and should be interpreted in this context. Co-management with gastroenterologists should be a priority in patients with IBD, as gaining control of inflammatory disease is crucial for the prevention of recurrent vitamin and micronutrient deficiencies in addition to long-term health in this population.

In 2023, ESPEN (the European Society for Clinical Nutrition and Metabolism) published consensus recommendations highlighting the importance of regular monitoring and treatment of nutrient deficiencies in patients with inflammatory bowel disease (IBD) for improved prognosis, mortality, and quality of life.¹ Suboptimal nutrition in patients with IBD predominantly results from inflammation of the gastrointestinal (GI) tract leading to malabsorption; however, medications commonly used to manage IBD also can contribute to malnutrition.^2,3 Additionally, patients may develop nausea and food avoidance due to medication or the disease itself, leading to nutritional withdrawal and eventual deficiency.⁴ Even with the development of diets focused on balancing nutritional needs and decreasing inflammation,⁵ offsetting this aversion to food can be difficult to overcome.²

Cutaneous manifestations of IBD are multifaceted and can be secondary to the disease, reactive to or associated with IBD, or effects from nutritional deficiencies. The most common vitamin and nutrient deficiencies in patients with IBD include iron; zinc; calcium; vitamin D; and vitamins B₆ (pyridoxine), B₉ (folic acid), and B₁₂.⁶ Malnutrition may manifest with cutaneous disease, and dermatologists can be the first to identify and assess for nutritional deficiencies. In this article, we review the mechanisms of these micronutrient depletions in the context of IBD, their subsequent dermatologic manifestations (Table), and treatment and monitoring guidelines for each deficiency.

Iron

A systematic review conducted from 2007 to 2012 in European patients with IBD (N=2192) found the overall prevalence of anemia in this population to be 24% (95% CI, 18%-31%), with 57% of patients with anemia experiencing iron deficiency.⁷ Anemia is observed more commonly in patients hospitalized with IBD and is common in patients with both Crohn disease and ulcerative colitis.⁸

Pathophysiology—Iron is critically important in oxygen transportation throughout the body as a major component of hemoglobin. Physiologically, the low pH of the duodenum and proximal jejunum allows divalent metal transporter 1 to transfer dietary Fe³⁺ into enterocytes, where it is reduced to the transportable Fe²⁺.^9,10 Distribution of Fe²⁺ ions from enterocytes relies on ferroportin, an iron-transporting protein, which is heavily regulated by the protein hepcidin.¹¹ Hepcidin, a known acute phase reactant, will increase in the setting of active IBD, causing a depletion of ferroportin and an inability of the body to utilize the stored iron in enterocytes.¹² This poor utilization of iron stores combined with blood loss caused by inflammation in the GI tract is the proposed primary mechanism of iron-deficiency anemia observed in patients with IBD.¹³

Cutaneous Manifestations—From a dermatologic perspective, iron-deficiency anemia can manifest with a wide range of symptoms including glossitis, koilonychia, xerosis and/or pruritus, and brittle hair or hair loss.^14,15 Although the underlying pathophysiology of these cutaneous manifestations is not fully understood, there are several theories assessing the mechanisms behind the skin findings of iron deficiency.

Atrophic glossitis has been observed in many patients with iron deficiency and is thought to manifest due to low iron concentrations in the blood, thereby decreasing oxygen delivery to the papillae of the dorsal tongue with resultant atrophy.^16,17 Similarly, decreased oxygen delivery to the nail bed capillaries may cause deformities in the nail called koilonychia (or “spoon nails”).¹⁸ Iron is a key co-factor in collagen lysyl hydroxylase that promotes collagen binding; iron deficiency may lead to disruptions in the epidermal barrier that can cause pruritus and xerosis.¹⁹ An observational study of 200 healthy patients with a primary concern of pruritus found a correlation between low serum ferritin and a higher degree of pruritus (r=−0.768; P<.00001).²⁰

Evidence for iron’s role in hair growth comes from a mouse model study with a mutation in the serine protease TMPRSS6—a protein that regulates hepcidin and iron absorption—which caused an increase in hepcidin production and subsequent systemic iron deficiency. Mice at 4 weeks of age were devoid of all body hair but had substantial regrowth after initiation of a 2-week iron-rich diet, which suggests a connection between iron repletion and hair growth in mice with iron deficiency.²¹ Additionally, a meta-analysis analyzing the comorbidities of patients with alopecia areata found them to have higher odds (odds ratio [OR]=2.78; 95% CI, 1.23-6.29) of iron-deficiency anemia but no association with IBD (OR=1.48; 95% CI, 0.32-6.82).²²

Diagnosis and Monitoring—The American Gastroenterological Association recommends a complete blood cell count (CBC), serum ferritin, transferrin saturation (TfS), and C-reactive protein (CRP) as standard evaluations for iron deficiency in patients with IBD. Patients with active IBD should be screened every 3 months,and patients with inactive disease should be screened every 6 to 12 months.²³

Although ferritin and TfS often are used as markers for iron status in healthy individuals, they are positive and negative acute phase reactants, respectively. Using them to assess iron status in patients with IBD may inaccurately represent iron status in the setting of inflammation from the disease.²⁴ The European Crohn’s and Colitis Organisation (ECCO) produced guidelines to define iron deficiency as a TfS less than 20% or a ferritin level less than 30 µg/L in patients without evidence of active IBD and a ferritin level less than 100 µg/L for patients with active inflammation.²⁵

A 2020 multicenter observational study of 202 patients with diagnosed IBD found that the ECCO guideline of ferritin less than 30 µg/L had an area under the receiver operating characteristic (AUROC) curve of 0.69, a sensitivity of 0.43, and a specificity of 0.95 in their population.²⁶ In a sensitivity analysis stratifying patients by CRP level (<10 or ≥10 mg/L), the authors found that for patients with ulcerative colitis and a CRP less than 10 mg/L, a cut-off value of ferritin less than 65 µg/L (AUROC=0.78) had a sensitivity of 0.78 and specificity of 0.76, and a TfS value of less than 16% (AUROC=0.88) had a sensitivity of 0.79 and a specificity of 0.9. In patients with a CRP of 10 mg/L or greater, a cut-off value of ferritin 80 µg/L (AUROC=0.76) had a sensitivity of 0.75 and a specificity of 0.82, and a TfS value of less than 11% (AUROC=0.69) had a sensitivity of 0.79 and a specificity of 0.88. There were no ferritin cut-off values associated with good diagnostic performance (defined as both sensitivity and specificity >0.70) for iron deficiency in patients with Crohn disease.²⁶

The authors recommended using an alternative iron measurement such as soluble transferrin receptor (sTfR)/log ferritin ratio (TfR-F) that is not influenced by active inflammation and has a good correlation with ferritin values (TfR-F: r=0.66; P<.001).²⁶ However, both sTfR and TfR-F have high costs and intermethod variability as well as differences in their reference ranges depending on which laboratory performs the analysis, limiting the accessibility and practicality of easily obtaining these tests.²⁷ Although there may be inaccuracies for standard ferritin or TfS under ECCO guidelines, proposed alternatives have their own limitations, which may make ferritin and TfS the most reasonable evaluations of iron status as long as disease activity status at the time of testing is taken into consideration.

Treatment—Treatment of underlying iron deficiency in patients with IBD requires reversing the cause of the deficiency and supplementing iron. In patients with IBD, the options to supplement iron may be limited by active disease, making oral intake less effective. Oral iron supplementation also is associated with notable GI adverse effects that may be exacerbated in patients with IBD. A systematic review of 43 randomized controlled trials (RCTs) evaluating GI adverse effects (eg, nausea, abdominal pain, diarrhea, constipation, and black or tarry stools) of oral ferrous sulfate compared with placebo or intravenous (IV) iron supplementation in healthy nonanemic individuals found a significant increase in GI adverse effects with oral supplementation (placebo: OR=2.32; P<.0001; IV: OR=3.05; P<.0001).²⁸

Therefore, IV iron repletion may be necessary in patients with IBD and may require numerous infusions depending on the formulation of iron. In an RCT conducted in 2011, patients with iron-deficiency anemia with quiescent or mild to moderate IBD were treated with either IV iron sulfate or ferric carboxymaltose.²⁹ With a primary end point of hemoglobin response greater than 2 g/dL, the authors found that 150 of 240 patients responded to ferric carboxymaltose vs 118 of 235 treated with iron sulfate (P=.004). The dosing for ferric carboxymaltose was 1 to 3 infusions of 500 to 1000 mg of iron and for iron sulfate up to 11 infusions of 200 mg of iron.²⁹

Zinc

A systematic review of zinc deficiency in patients with IBD identified 7 studies including 2413 patients and revealed those with Crohn disease had a higher prevalence of zinc deficiency compared with patients with ulcerative colitis (54% vs 41%).³⁰

Pathophysiology—Zinc serves as a catalytic cofactor for enzymatic activity within proteins and immune cells.³¹ The homeostasis of zinc is tightly regulated within the brush border of the small intestine by zinc transporters ZIP4 and ZIP1 from the lumen of enterocytes into the bloodstream.³² Inflammation in the small intestine due to Crohn disease can result in zinc malabsorption.

Ranaldi et al³³ exposed intestinal cells and zinc-depleted intestinal cells to tumor necrosis factor α media to simulate an inflammatory environment. They measured transepithelial electrical resistance as a surrogate for transmembrane permeability and found that zinc-depleted cells had a statistically significantly higher transepithelial electrical resistance percentage (60% reduction after 4 hours; P<1.10^–6) when exposed to tumor necrosis factor α signaling compared with normal intestinal cells. They concluded that zinc deficiency can increase intestinal permeability in the presence of inflammation, creating a cycle of further nutrient malabsorption and inflammation exacerbating IBD symptoms.³³

Cutaneous Manifestations—After absorption in the small intestine, approximately 5% of zinc resides in the skin, with the highest concentration in the stratum spinosum.³⁴ A cell study found that keratinocytes in zinc-deficient environments had higher rates of apoptosis compared with cells in normal media. The authors proposed that this higher rate of apoptosis and the resulting inflammation could be a mechanism for developing the desquamative or eczematous scaly plaques that are common cutaneous manifestations of zinc deficiency.³⁵

Other cutaneous findings may include angular cheilitis, stomatitis, glossitis, paronychia, onychodystrophy, generalized alopecia, and delayed wound healing.³⁶ The histopathology of these skin lesions is characterized by granular layer loss, epidermal pallor, confluent parakeratosis, spongiosis, dyskeratosis, and psoriasiform hyperplasia.³⁷

Diagnosis and Monitoring—Assessing serum zinc levels is challenging, as they may decrease during states of inflammation.³⁸ A mouse model study showed a 3.1-fold increase (P<.001) in ZIP14 expression in wild-type mice compared with an IL-6 -/- knock-down model after IL-6 exposure. The authors concluded that the upregulation of ZIP14 in the liver due to inflammatory cytokine upregulation decreases zinc availability in serum.³⁹ Additionally, serum zinc can overestimate the level of deficiency in IBD because approximately 75% of serum zinc is bound to albumin, which decreases in the setting of inflammation.^40-42

Alternatively, alkaline phosphatase (AP), a zinc-dependent metalloenzyme, may be a better evaluator of zinc status during periods of inflammation. A study in rats evaluated zinc through serum zinc levels and AP levels after a period of induced stress to mimic a short-term inflammatory state.⁴³ The researchers found that total body stores of zinc were unaffected throughout the experiment; only serum zinc declined throughout the experiment duration while AP did not. Because approximately 75% of serum zinc is bound to serum albumin,⁴² the researchers concluded the induced inflammatory state depleted serum albumin and redistributed zinc to the liver, causing the observed serum zinc changes, while total body zinc levels and AP were largely unaffected in comparison.⁴³ Comorbid conditions such as liver or bone disease can increase AP levels, which limits the utility of AP as a surrogate for zinc in patients with comorbidities.⁴⁴ However, even in the context of active IBD, serum zinc still is currently considered the best biomarker to evaluate zinc status.⁴⁵

Treatment—The recommended dose for zinc supplementation is 20 to 40 mg daily with higher doses (>50 mg/d) for patients with malabsorptive syndromes such as IBD.⁴⁶ It can be administered orally or parenterally. Although rare, zinc replacement therapy may be associated with diarrhea, nausea, vomiting, mild headaches, and fatigue.⁴⁶ Additional considerations should be taken when repleting other micronutrients with zinc, as calcium and folate can inhibit zinc reabsorption, while zinc itself can inhibit iron and copper reabsorption.⁴⁷

Vitamin D and Calcium

Low vitamin D levels (<50 nmol/L) and hypocalcemia (<8.8 mg/dL) are common in patients with IBD.^48,49

Pathophysiology—Vitamin D levels are maintained via 2 mechanisms. The first mechanism is through the skin, as keratinocytes produce 7-dehydrocholesterol after exposure to UV light, which is converted into previtamin D₃ and then thermally isomerizes into vitamin D₃. This vitamin D₃ is then transported to the liver on vitamin D–binding protein.⁵⁰ The second mechanism is through oral vitamin D₃ that is absorbed through vitamin D receptors in intestinal epithelium and transported to the liver, where it is hydroxylated into 25-hydroxyvitamin D (25[OH]D), then to the kidneys for hydroxylation to 1,25(OH)₂D for redistribution throughout the body.⁵⁰ This activated form of vitamin D regulates calcium absorption in the intestine, and optimal vitamin D levels are necessary to absorb calcium efficiently.⁵¹ Inflammation from IBD within the small intestine can downregulate vitamin D receptors, causing malabsorption and decreased serum vitamin D.⁵²

Vitamin D signaling also is vital to maintaining the tight junctions and adherens junctions of the intestinal epithelium. Weakening the permeability of the epithelium further exacerbates malabsorption and subsequent vitamin D deficiency.⁵² A meta-analysis of 27 studies including 8316 patients with IBD showed low vitamin D levels were associated with increased odds of disease activity (OR=1.53; 95% CI, 1.32-1.77), mucosal inflammation (OR=1.25; 95% CI, 1.06-1.47), and future clinical relapse (OR=1.23; 95% CI, 1.03-1.47) in patients with Crohn disease. The authors concluded that low levels of vitamin D could be used as a potential biomarker of inflammatory status in Crohn disease.⁵³

Vitamin D and calcium are further implicated in maintaining skeletal health,⁴⁷ while vitamin D specifically helps maintain intestinal homeostasis⁵⁴ and immune system modulation in the skin.⁵⁵

Cutaneous Manifestations—Vitamin D is thought to play crucial roles in skin differentiation and proliferation, cutaneous innate immunity, hair follicle cycling, photoprotection, and wound healing.⁵⁶ Vitamin D deficiency has been observed in a large range of cutaneous diseases including skin cancer, psoriasis, vitiligo, bullous pemphigoid, atopic dermatitis, and various types of alopecia.^56-59 It is unclear whether vitamin D deficiency facilitates these disease processes or is merely the consequence of a disrupted cutaneous surface with the inability to complete the first step in vitamin D processing. A 2014 meta-analysis of 290 prospective cohort studies and 172 randomized trials concluded that 25(OH)D deficiency was associated with ill health and did not find causal evidence for any specific disease, dermatologic or otherwise.⁶⁰ Calcium deficiency may cause epidermal changes including dry skin, coarse hair, and brittle nails.⁶¹

Diagnosis and Monitoring—The ECCO guidelines recommend obtaining serum 25(OH)D levels every 3 months in patients with IBD.⁶² Levels less than 75 nmol/L are considered deficient, and a value less than 30 nmol/L increases the risk for osteomalacia and nutritional rickets, constituting severe vitamin D deficiency.^63-65

An observational study of 325 patients with IBD showed a statistically significant negative correlation between serum vitamin D and fecal calprotectin (r=−0.19; P<.001), a stool-based marker for gut inflammation, supporting vitamin D as a potential biomarker in IBD.⁶⁶

Evaluation of calcium can be done through serum levels in patients with IBD.⁶⁷ Patients with IBD are at risk for hypoalbuminemia; therefore, consideration should be taken to ensure calcium levels are corrected, as approximately 50% of calcium is bound to albumin or other ions in the body,⁶⁸ which can be done by adjusting the calcium concentration by 0.02 mmol/L for every 1 g/L of albumin above or below 40 g/L. In the most critically ill patients, a direct ionized calcium blood level should be used instead because the previously mentioned correction calculations are inaccurate when albumin is critically low.⁶⁹

Treatment—The ECCO guidelines recommend calcium and vitamin D repletion of 500 to 1000 mg and 800 to 1000 U, respectively, in patients with IBD on systemic corticosteroids to prevent the negative effects of bone loss.⁶² Calcium repletion in patients with IBD who are not on systemic steroids are the same as for the general population.⁶⁵

Vitamin D repletion also may help decrease IBD activity. In a prospective study, 10,000 IU/d of vitamin D in 10 patients with IBD—adjusted over 12 weeks to a target of 100 to 125 nmol/L of serum 25(OH)D—showed a significant reduction in clinical Crohn activity (P=.019) over the study period.⁷⁰ In contrast, 2000 IU/d for 3 months in an RCT of 27 patients with Crohn disease found significantly lower CRP (P=.019) and significantly higher self-reported quality of life (P=.037) but nonsignificant decreases in Crohn activity (P=.082) in patients with 25(OH)D levels of 75 nmol/L or higher compared with those with 25(OH)D levels less than 75 nmol/L.⁷¹

These discrepancies illustrate the need for expanded clinical trials to elucidate the optimal vitamin D dosing for patients with IBD. Ultimately, assessing vitamin D and calcium status and considering repletion in patients with IBD, especially those with comorbid dermatologic diseases such as poor wound healing, psoriasis, or atopic dermatitis, is important.

Vitamin B₆ (Pyridoxine)

Pathophysiology—Pyridoxine is an important coenzyme for many functions including amino acid transamination, fatty acid metabolism, and conversion of tryptophan to niacin. It is absorbed in the jejunum and ileum and subsequently transported to the liver for rephosphorylation and release into its active form.³⁶ An observational study assessing the nutritional status of patients with IBD found that only 5.7% of 105 patients with food records had inadequate dietary intake of pyridoxine, but 29% of all patients with IBD had subnormal pyridoxine levels.⁷² Additionally, they found no significant difference in the prevalence of subnormal pyridoxine levels in patients with active IBD vs IBD in remission. The authors suggested that the subnormal pyridoxine levels in patients with IBD likely were multifactorial and resulted from malabsorption due to active disease, inflammation, and inadequate intake.⁷²

Cutaneous Manifestations—Cutaneous findings associated with pyridoxine deficiency include periorificial and perineal dermatitis,⁷³ angular stomatitis, and cheilitis with associated burning, redness, and tongue edema.³⁶ Additionally, pyridoxine is involved in the conversion of tryptophan to niacin, and its deficiency may manifest with pellagralike findings.⁷⁴

Because pyridoxine is critical to protein metabolism, its deficiency may disrupt key cellular structures that rely on protein concentrations to maintain structural integrity. One such structure in the skin that heavily relies on protein concentrations is the ground substance of the extracellular matrix—the amorphous gelatinous spaces that occupy the areas between the extracellular matrix, which consists of cross-linked glycosaminoglycans and proteins.⁷⁵ Without protein, ground substance increases in viscosity and can disrupt the epidermal barrier, leading to increased transepidermal water loss and ultimately inflammation.⁷⁶ Although this theory has yet to be validated fully, this is a potential mechanistic explanation for the inflammation in dermal papillae that leads to dermatitis observed in pyridoxine deficiency.

Diagnosis and Monitoring—Direct biomarkers of pyridoxine status are in serum, plasma, erythrocytes, and urine, with the most common measurement in plasma as pyridoxal 5′-phosphate (PLP).⁷⁷ Plasma PLP concentrations lower than 20 nmol/L are suggestive of deficiency.⁷⁸ Plasma PLP has shown inverse relationships with acute phase inflammatory markers CRP⁷⁹ and AP,⁷⁸ thereby raising concerns for its validity to assess pyridoxine status in patients with symptomatic IBD.⁸⁰

Alternative evaluations of pyridoxine include tryptophan and methionine loading tests,³⁶ which are measured via urinary excretion and require normal kidney function to be accurate. They should be considered in IBD if necessary, but routine testing, even in patients with symptomatic IBD, is not recommended in the ECCO guidelines. Additional considerations should be taken in patients with altered nutrient requirements such as those who have undergone bowel resection due to highly active disease or those who receive parenteral nutritional supplementation.⁸¹

Treatment—Recommendations for oral pyridoxine supplementation range from 25 to 600 mg daily,⁸² with symptoms typically improving on 100 mg daily.³⁶ Pyridoxine supplementation may have additional benefits for patients with IBD and potentially modulate disease severity. An IL-10 knockout mouse supplemented with pyridoxine had an approximately 60% reduction (P<.05) in inflammation compared to mice deficient in pyridoxine.⁸³ The authors suggest that PLP-dependent enzymes can inhibit further proinflammatory signaling and T-cell migration that can exacerbate IBD. Ultimately, more data is needed before determining the efficacy of pyridoxine supplementation for active IBD.

Vitamin B₁₂ and Vitamin B₉ (Folic Acid)

Pathophysiology—Vitamin B₁₂ is reabsorbed in the terminal ileum, the distal portion of the small intestine. The American Gastroenterological Association recommends that patients with a history of extensive ileal disease or prior ileal surgery, which is the case for many patients with Crohn disease, be monitored for vitamin B₁₂ deficiency.²³ Monitoring and rapid supplementation of vitamin B₁₂ can prevent pernicious anemia and irreversible neurologic damage that may result from deficiency.⁸⁴

Folic acid is primarily absorbed in the duodenum and jejunum of the small intestine. A meta-analysis performed in 2017 assessed studies observing folic acid and vitamin B₁₂ levels in 1086 patients with IBD compared with 1484 healthy controls and found an average difference in serum folate concentration of 0.46 nmol/L (P<.001).⁸⁴ Interestingly, this study did not find a significant difference in serum vitamin B₁₂ levels between patients with IBD and healthy controls, highlighting the mechanism of vitamin B₁₂ deficiency in IBD because only patients with terminal ileal involvement are at risk for malabsorption and subsequent deficiency.

Cutaneous Manifestations—Both vitamin B₁₂ and folic acid deficiency can manifest as cheilitis, glossitis, and/or generalized hyperpigmentation that is accentuated in the flexural areas, palms, soles, and oral cavity.^85,86 Systemic symptoms of patients with vitamin B₁₂ and folic acid deficiency include megaloblastic anemia, pallor, and fatigue. A potential mechanism for the hyperpigmentation observed from vitamin B₁₂ deficiency came from an electron microscope study that showed an increased concentration of melanosomes in a patient with deficiency.⁸⁷

Diagnosis and Monitoring—In patients with suspected vitamin B₁₂ and/or folic acid deficiency, initial evaluation should include a CBC with peripheral smear and serum vitamin B₁₂ and folate levels. In cases for which the diagnosis still is unclear after initial testing, methylmalonic acid and homocysteine levels can help differentiate between the 2 deficiencies. Methylmalonic acid classically is elevated (>260 nmol/L) in vitamin B₁₂ deficiency but not in folate deficiency.⁸⁸ Cut-off values for vitamin B₁₂ deficiency are less than 200 to 250 pg/mL forserum vitamin B₁₂ and/or an elevated level of methylmalonic acid (>0.271 µmol/L).⁸⁹ A serum folic acid value greater than 3 ng/mL and/or erythrocyte folate concentrations greater than 140 ng/mL are considered adequate, whereas an indicator of folic acid deficiency is a homocysteine level less than 10 µmol/L.⁹⁰ A CBC can screen for macrocytic megaloblastic anemias (mean corpuscular volume >100 fl), which are classic diagnostic signs of an underlying vitamin B₁₂ or folate deficiency.

Treatment—According to the Centers for Disease Control and Prevention, supplementation of vitamin B₁₂ can be done orally with 1000 µg daily in patients with deficiency. In patients with active IBD, oral reabsorption of vitamin B₁₂ can be less effective, making subcutaneous or intramuscular administration (1000 µg/wk for 8 weeks, then monthly for life) better options.⁸⁹

Patients with IBD managed with methotrexate should be screened carefully for folate deficiency. Methotrexate is a folate analog that sometimes is used for the treatment of IBD. Reversible competitive inhibition of dihydrofolate reductase can precipitate a systemic folic acid decrease.⁹¹ Typically, oral folic acid (1 to 5 mg/d) is sufficient to treat folate deficiency, with the ESPEN recommending 5 mg once weekly 24 to 72 hours after methotrexate treatment or 1 mg daily for 5 days per week in patients with IBD.¹ Alternative formulations—IV, subcutaneous, or intramuscular—are available for patients who cannot tolerate oral intake.⁹²

Final Thoughts

Dermatologists can be the first to observe the cutaneous manifestations of micronutrient deficiencies. Although the symptoms of each micronutrient deficiency discussed may overlap, attention to small clinical clues in patients with IBD can improve patient outcomes and quality of life. For example, koilonychia with glossitis and xerosis likely is due to iron deficiency, while zinc deficiency should be suspected in patients with scaly eczematous plaques in skin folds. A high level of suspicion for micronutrient deficiencies in patients with IBD should be followed by a complete patient history, review of systems, and thorough clinical examination. A thorough laboratory evaluation can pinpoint nutritional deficiencies in patients with IBD, keeping in mind that specific biomarkers such as ferritin and serum zinc also act as acute phase reactants and should be interpreted in this context. Co-management with gastroenterologists should be a priority in patients with IBD, as gaining control of inflammatory disease is crucial for the prevention of recurrent vitamin and micronutrient deficiencies in addition to long-term health in this population.

In 2023, ESPEN (the European Society for Clinical Nutrition and Metabolism) published consensus recommendations highlighting the importance of regular monitoring and treatment of nutrient deficiencies in patients with inflammatory bowel disease (IBD) for improved prognosis, mortality, and quality of life.¹ Suboptimal nutrition in patients with IBD predominantly results from inflammation of the gastrointestinal (GI) tract leading to malabsorption; however, medications commonly used to manage IBD also can contribute to malnutrition.^2,3 Additionally, patients may develop nausea and food avoidance due to medication or the disease itself, leading to nutritional withdrawal and eventual deficiency.⁴ Even with the development of diets focused on balancing nutritional needs and decreasing inflammation,⁵ offsetting this aversion to food can be difficult to overcome.²

Cutaneous manifestations of IBD are multifaceted and can be secondary to the disease, reactive to or associated with IBD, or effects from nutritional deficiencies. The most common vitamin and nutrient deficiencies in patients with IBD include iron; zinc; calcium; vitamin D; and vitamins B₆ (pyridoxine), B₉ (folic acid), and B₁₂.⁶ Malnutrition may manifest with cutaneous disease, and dermatologists can be the first to identify and assess for nutritional deficiencies. In this article, we review the mechanisms of these micronutrient depletions in the context of IBD, their subsequent dermatologic manifestations (Table), and treatment and monitoring guidelines for each deficiency.

Iron

A systematic review conducted from 2007 to 2012 in European patients with IBD (N=2192) found the overall prevalence of anemia in this population to be 24% (95% CI, 18%-31%), with 57% of patients with anemia experiencing iron deficiency.⁷ Anemia is observed more commonly in patients hospitalized with IBD and is common in patients with both Crohn disease and ulcerative colitis.⁸

Pathophysiology—Iron is critically important in oxygen transportation throughout the body as a major component of hemoglobin. Physiologically, the low pH of the duodenum and proximal jejunum allows divalent metal transporter 1 to transfer dietary Fe³⁺ into enterocytes, where it is reduced to the transportable Fe²⁺.^9,10 Distribution of Fe²⁺ ions from enterocytes relies on ferroportin, an iron-transporting protein, which is heavily regulated by the protein hepcidin.¹¹ Hepcidin, a known acute phase reactant, will increase in the setting of active IBD, causing a depletion of ferroportin and an inability of the body to utilize the stored iron in enterocytes.¹² This poor utilization of iron stores combined with blood loss caused by inflammation in the GI tract is the proposed primary mechanism of iron-deficiency anemia observed in patients with IBD.¹³

Cutaneous Manifestations—From a dermatologic perspective, iron-deficiency anemia can manifest with a wide range of symptoms including glossitis, koilonychia, xerosis and/or pruritus, and brittle hair or hair loss.^14,15 Although the underlying pathophysiology of these cutaneous manifestations is not fully understood, there are several theories assessing the mechanisms behind the skin findings of iron deficiency.

Atrophic glossitis has been observed in many patients with iron deficiency and is thought to manifest due to low iron concentrations in the blood, thereby decreasing oxygen delivery to the papillae of the dorsal tongue with resultant atrophy.^16,17 Similarly, decreased oxygen delivery to the nail bed capillaries may cause deformities in the nail called koilonychia (or “spoon nails”).¹⁸ Iron is a key co-factor in collagen lysyl hydroxylase that promotes collagen binding; iron deficiency may lead to disruptions in the epidermal barrier that can cause pruritus and xerosis.¹⁹ An observational study of 200 healthy patients with a primary concern of pruritus found a correlation between low serum ferritin and a higher degree of pruritus (r=−0.768; P<.00001).²⁰

Evidence for iron’s role in hair growth comes from a mouse model study with a mutation in the serine protease TMPRSS6—a protein that regulates hepcidin and iron absorption—which caused an increase in hepcidin production and subsequent systemic iron deficiency. Mice at 4 weeks of age were devoid of all body hair but had substantial regrowth after initiation of a 2-week iron-rich diet, which suggests a connection between iron repletion and hair growth in mice with iron deficiency.²¹ Additionally, a meta-analysis analyzing the comorbidities of patients with alopecia areata found them to have higher odds (odds ratio [OR]=2.78; 95% CI, 1.23-6.29) of iron-deficiency anemia but no association with IBD (OR=1.48; 95% CI, 0.32-6.82).²²

Diagnosis and Monitoring—The American Gastroenterological Association recommends a complete blood cell count (CBC), serum ferritin, transferrin saturation (TfS), and C-reactive protein (CRP) as standard evaluations for iron deficiency in patients with IBD. Patients with active IBD should be screened every 3 months,and patients with inactive disease should be screened every 6 to 12 months.²³

Although ferritin and TfS often are used as markers for iron status in healthy individuals, they are positive and negative acute phase reactants, respectively. Using them to assess iron status in patients with IBD may inaccurately represent iron status in the setting of inflammation from the disease.²⁴ The European Crohn’s and Colitis Organisation (ECCO) produced guidelines to define iron deficiency as a TfS less than 20% or a ferritin level less than 30 µg/L in patients without evidence of active IBD and a ferritin level less than 100 µg/L for patients with active inflammation.²⁵

A 2020 multicenter observational study of 202 patients with diagnosed IBD found that the ECCO guideline of ferritin less than 30 µg/L had an area under the receiver operating characteristic (AUROC) curve of 0.69, a sensitivity of 0.43, and a specificity of 0.95 in their population.²⁶ In a sensitivity analysis stratifying patients by CRP level (<10 or ≥10 mg/L), the authors found that for patients with ulcerative colitis and a CRP less than 10 mg/L, a cut-off value of ferritin less than 65 µg/L (AUROC=0.78) had a sensitivity of 0.78 and specificity of 0.76, and a TfS value of less than 16% (AUROC=0.88) had a sensitivity of 0.79 and a specificity of 0.9. In patients with a CRP of 10 mg/L or greater, a cut-off value of ferritin 80 µg/L (AUROC=0.76) had a sensitivity of 0.75 and a specificity of 0.82, and a TfS value of less than 11% (AUROC=0.69) had a sensitivity of 0.79 and a specificity of 0.88. There were no ferritin cut-off values associated with good diagnostic performance (defined as both sensitivity and specificity >0.70) for iron deficiency in patients with Crohn disease.²⁶

The authors recommended using an alternative iron measurement such as soluble transferrin receptor (sTfR)/log ferritin ratio (TfR-F) that is not influenced by active inflammation and has a good correlation with ferritin values (TfR-F: r=0.66; P<.001).²⁶ However, both sTfR and TfR-F have high costs and intermethod variability as well as differences in their reference ranges depending on which laboratory performs the analysis, limiting the accessibility and practicality of easily obtaining these tests.²⁷ Although there may be inaccuracies for standard ferritin or TfS under ECCO guidelines, proposed alternatives have their own limitations, which may make ferritin and TfS the most reasonable evaluations of iron status as long as disease activity status at the time of testing is taken into consideration.

Treatment—Treatment of underlying iron deficiency in patients with IBD requires reversing the cause of the deficiency and supplementing iron. In patients with IBD, the options to supplement iron may be limited by active disease, making oral intake less effective. Oral iron supplementation also is associated with notable GI adverse effects that may be exacerbated in patients with IBD. A systematic review of 43 randomized controlled trials (RCTs) evaluating GI adverse effects (eg, nausea, abdominal pain, diarrhea, constipation, and black or tarry stools) of oral ferrous sulfate compared with placebo or intravenous (IV) iron supplementation in healthy nonanemic individuals found a significant increase in GI adverse effects with oral supplementation (placebo: OR=2.32; P<.0001; IV: OR=3.05; P<.0001).²⁸

Therefore, IV iron repletion may be necessary in patients with IBD and may require numerous infusions depending on the formulation of iron. In an RCT conducted in 2011, patients with iron-deficiency anemia with quiescent or mild to moderate IBD were treated with either IV iron sulfate or ferric carboxymaltose.²⁹ With a primary end point of hemoglobin response greater than 2 g/dL, the authors found that 150 of 240 patients responded to ferric carboxymaltose vs 118 of 235 treated with iron sulfate (P=.004). The dosing for ferric carboxymaltose was 1 to 3 infusions of 500 to 1000 mg of iron and for iron sulfate up to 11 infusions of 200 mg of iron.²⁹

Zinc

A systematic review of zinc deficiency in patients with IBD identified 7 studies including 2413 patients and revealed those with Crohn disease had a higher prevalence of zinc deficiency compared with patients with ulcerative colitis (54% vs 41%).³⁰

Pathophysiology—Zinc serves as a catalytic cofactor for enzymatic activity within proteins and immune cells.³¹ The homeostasis of zinc is tightly regulated within the brush border of the small intestine by zinc transporters ZIP4 and ZIP1 from the lumen of enterocytes into the bloodstream.³² Inflammation in the small intestine due to Crohn disease can result in zinc malabsorption.

Ranaldi et al³³ exposed intestinal cells and zinc-depleted intestinal cells to tumor necrosis factor α media to simulate an inflammatory environment. They measured transepithelial electrical resistance as a surrogate for transmembrane permeability and found that zinc-depleted cells had a statistically significantly higher transepithelial electrical resistance percentage (60% reduction after 4 hours; P<1.10^–6) when exposed to tumor necrosis factor α signaling compared with normal intestinal cells. They concluded that zinc deficiency can increase intestinal permeability in the presence of inflammation, creating a cycle of further nutrient malabsorption and inflammation exacerbating IBD symptoms.³³

Cutaneous Manifestations—After absorption in the small intestine, approximately 5% of zinc resides in the skin, with the highest concentration in the stratum spinosum.³⁴ A cell study found that keratinocytes in zinc-deficient environments had higher rates of apoptosis compared with cells in normal media. The authors proposed that this higher rate of apoptosis and the resulting inflammation could be a mechanism for developing the desquamative or eczematous scaly plaques that are common cutaneous manifestations of zinc deficiency.³⁵

Other cutaneous findings may include angular cheilitis, stomatitis, glossitis, paronychia, onychodystrophy, generalized alopecia, and delayed wound healing.³⁶ The histopathology of these skin lesions is characterized by granular layer loss, epidermal pallor, confluent parakeratosis, spongiosis, dyskeratosis, and psoriasiform hyperplasia.³⁷

Diagnosis and Monitoring—Assessing serum zinc levels is challenging, as they may decrease during states of inflammation.³⁸ A mouse model study showed a 3.1-fold increase (P<.001) in ZIP14 expression in wild-type mice compared with an IL-6 -/- knock-down model after IL-6 exposure. The authors concluded that the upregulation of ZIP14 in the liver due to inflammatory cytokine upregulation decreases zinc availability in serum.³⁹ Additionally, serum zinc can overestimate the level of deficiency in IBD because approximately 75% of serum zinc is bound to albumin, which decreases in the setting of inflammation.^40-42

Alternatively, alkaline phosphatase (AP), a zinc-dependent metalloenzyme, may be a better evaluator of zinc status during periods of inflammation. A study in rats evaluated zinc through serum zinc levels and AP levels after a period of induced stress to mimic a short-term inflammatory state.⁴³ The researchers found that total body stores of zinc were unaffected throughout the experiment; only serum zinc declined throughout the experiment duration while AP did not. Because approximately 75% of serum zinc is bound to serum albumin,⁴² the researchers concluded the induced inflammatory state depleted serum albumin and redistributed zinc to the liver, causing the observed serum zinc changes, while total body zinc levels and AP were largely unaffected in comparison.⁴³ Comorbid conditions such as liver or bone disease can increase AP levels, which limits the utility of AP as a surrogate for zinc in patients with comorbidities.⁴⁴ However, even in the context of active IBD, serum zinc still is currently considered the best biomarker to evaluate zinc status.⁴⁵

Treatment—The recommended dose for zinc supplementation is 20 to 40 mg daily with higher doses (>50 mg/d) for patients with malabsorptive syndromes such as IBD.⁴⁶ It can be administered orally or parenterally. Although rare, zinc replacement therapy may be associated with diarrhea, nausea, vomiting, mild headaches, and fatigue.⁴⁶ Additional considerations should be taken when repleting other micronutrients with zinc, as calcium and folate can inhibit zinc reabsorption, while zinc itself can inhibit iron and copper reabsorption.⁴⁷

Vitamin D and Calcium

Low vitamin D levels (<50 nmol/L) and hypocalcemia (<8.8 mg/dL) are common in patients with IBD.^48,49

Pathophysiology—Vitamin D levels are maintained via 2 mechanisms. The first mechanism is through the skin, as keratinocytes produce 7-dehydrocholesterol after exposure to UV light, which is converted into previtamin D₃ and then thermally isomerizes into vitamin D₃. This vitamin D₃ is then transported to the liver on vitamin D–binding protein.⁵⁰ The second mechanism is through oral vitamin D₃ that is absorbed through vitamin D receptors in intestinal epithelium and transported to the liver, where it is hydroxylated into 25-hydroxyvitamin D (25[OH]D), then to the kidneys for hydroxylation to 1,25(OH)₂D for redistribution throughout the body.⁵⁰ This activated form of vitamin D regulates calcium absorption in the intestine, and optimal vitamin D levels are necessary to absorb calcium efficiently.⁵¹ Inflammation from IBD within the small intestine can downregulate vitamin D receptors, causing malabsorption and decreased serum vitamin D.⁵²

Vitamin D signaling also is vital to maintaining the tight junctions and adherens junctions of the intestinal epithelium. Weakening the permeability of the epithelium further exacerbates malabsorption and subsequent vitamin D deficiency.⁵² A meta-analysis of 27 studies including 8316 patients with IBD showed low vitamin D levels were associated with increased odds of disease activity (OR=1.53; 95% CI, 1.32-1.77), mucosal inflammation (OR=1.25; 95% CI, 1.06-1.47), and future clinical relapse (OR=1.23; 95% CI, 1.03-1.47) in patients with Crohn disease. The authors concluded that low levels of vitamin D could be used as a potential biomarker of inflammatory status in Crohn disease.⁵³

Vitamin D and calcium are further implicated in maintaining skeletal health,⁴⁷ while vitamin D specifically helps maintain intestinal homeostasis⁵⁴ and immune system modulation in the skin.⁵⁵

Cutaneous Manifestations—Vitamin D is thought to play crucial roles in skin differentiation and proliferation, cutaneous innate immunity, hair follicle cycling, photoprotection, and wound healing.⁵⁶ Vitamin D deficiency has been observed in a large range of cutaneous diseases including skin cancer, psoriasis, vitiligo, bullous pemphigoid, atopic dermatitis, and various types of alopecia.^56-59 It is unclear whether vitamin D deficiency facilitates these disease processes or is merely the consequence of a disrupted cutaneous surface with the inability to complete the first step in vitamin D processing. A 2014 meta-analysis of 290 prospective cohort studies and 172 randomized trials concluded that 25(OH)D deficiency was associated with ill health and did not find causal evidence for any specific disease, dermatologic or otherwise.⁶⁰ Calcium deficiency may cause epidermal changes including dry skin, coarse hair, and brittle nails.⁶¹

Diagnosis and Monitoring—The ECCO guidelines recommend obtaining serum 25(OH)D levels every 3 months in patients with IBD.⁶² Levels less than 75 nmol/L are considered deficient, and a value less than 30 nmol/L increases the risk for osteomalacia and nutritional rickets, constituting severe vitamin D deficiency.^63-65

An observational study of 325 patients with IBD showed a statistically significant negative correlation between serum vitamin D and fecal calprotectin (r=−0.19; P<.001), a stool-based marker for gut inflammation, supporting vitamin D as a potential biomarker in IBD.⁶⁶

Evaluation of calcium can be done through serum levels in patients with IBD.⁶⁷ Patients with IBD are at risk for hypoalbuminemia; therefore, consideration should be taken to ensure calcium levels are corrected, as approximately 50% of calcium is bound to albumin or other ions in the body,⁶⁸ which can be done by adjusting the calcium concentration by 0.02 mmol/L for every 1 g/L of albumin above or below 40 g/L. In the most critically ill patients, a direct ionized calcium blood level should be used instead because the previously mentioned correction calculations are inaccurate when albumin is critically low.⁶⁹

Treatment—The ECCO guidelines recommend calcium and vitamin D repletion of 500 to 1000 mg and 800 to 1000 U, respectively, in patients with IBD on systemic corticosteroids to prevent the negative effects of bone loss.⁶² Calcium repletion in patients with IBD who are not on systemic steroids are the same as for the general population.⁶⁵

Vitamin D repletion also may help decrease IBD activity. In a prospective study, 10,000 IU/d of vitamin D in 10 patients with IBD—adjusted over 12 weeks to a target of 100 to 125 nmol/L of serum 25(OH)D—showed a significant reduction in clinical Crohn activity (P=.019) over the study period.⁷⁰ In contrast, 2000 IU/d for 3 months in an RCT of 27 patients with Crohn disease found significantly lower CRP (P=.019) and significantly higher self-reported quality of life (P=.037) but nonsignificant decreases in Crohn activity (P=.082) in patients with 25(OH)D levels of 75 nmol/L or higher compared with those with 25(OH)D levels less than 75 nmol/L.⁷¹

These discrepancies illustrate the need for expanded clinical trials to elucidate the optimal vitamin D dosing for patients with IBD. Ultimately, assessing vitamin D and calcium status and considering repletion in patients with IBD, especially those with comorbid dermatologic diseases such as poor wound healing, psoriasis, or atopic dermatitis, is important.

Vitamin B₆ (Pyridoxine)

Pathophysiology—Pyridoxine is an important coenzyme for many functions including amino acid transamination, fatty acid metabolism, and conversion of tryptophan to niacin. It is absorbed in the jejunum and ileum and subsequently transported to the liver for rephosphorylation and release into its active form.³⁶ An observational study assessing the nutritional status of patients with IBD found that only 5.7% of 105 patients with food records had inadequate dietary intake of pyridoxine, but 29% of all patients with IBD had subnormal pyridoxine levels.⁷² Additionally, they found no significant difference in the prevalence of subnormal pyridoxine levels in patients with active IBD vs IBD in remission. The authors suggested that the subnormal pyridoxine levels in patients with IBD likely were multifactorial and resulted from malabsorption due to active disease, inflammation, and inadequate intake.⁷²

Cutaneous Manifestations—Cutaneous findings associated with pyridoxine deficiency include periorificial and perineal dermatitis,⁷³ angular stomatitis, and cheilitis with associated burning, redness, and tongue edema.³⁶ Additionally, pyridoxine is involved in the conversion of tryptophan to niacin, and its deficiency may manifest with pellagralike findings.⁷⁴

Because pyridoxine is critical to protein metabolism, its deficiency may disrupt key cellular structures that rely on protein concentrations to maintain structural integrity. One such structure in the skin that heavily relies on protein concentrations is the ground substance of the extracellular matrix—the amorphous gelatinous spaces that occupy the areas between the extracellular matrix, which consists of cross-linked glycosaminoglycans and proteins.⁷⁵ Without protein, ground substance increases in viscosity and can disrupt the epidermal barrier, leading to increased transepidermal water loss and ultimately inflammation.⁷⁶ Although this theory has yet to be validated fully, this is a potential mechanistic explanation for the inflammation in dermal papillae that leads to dermatitis observed in pyridoxine deficiency.

Diagnosis and Monitoring—Direct biomarkers of pyridoxine status are in serum, plasma, erythrocytes, and urine, with the most common measurement in plasma as pyridoxal 5′-phosphate (PLP).⁷⁷ Plasma PLP concentrations lower than 20 nmol/L are suggestive of deficiency.⁷⁸ Plasma PLP has shown inverse relationships with acute phase inflammatory markers CRP⁷⁹ and AP,⁷⁸ thereby raising concerns for its validity to assess pyridoxine status in patients with symptomatic IBD.⁸⁰

Alternative evaluations of pyridoxine include tryptophan and methionine loading tests,³⁶ which are measured via urinary excretion and require normal kidney function to be accurate. They should be considered in IBD if necessary, but routine testing, even in patients with symptomatic IBD, is not recommended in the ECCO guidelines. Additional considerations should be taken in patients with altered nutrient requirements such as those who have undergone bowel resection due to highly active disease or those who receive parenteral nutritional supplementation.⁸¹

Treatment—Recommendations for oral pyridoxine supplementation range from 25 to 600 mg daily,⁸² with symptoms typically improving on 100 mg daily.³⁶ Pyridoxine supplementation may have additional benefits for patients with IBD and potentially modulate disease severity. An IL-10 knockout mouse supplemented with pyridoxine had an approximately 60% reduction (P<.05) in inflammation compared to mice deficient in pyridoxine.⁸³ The authors suggest that PLP-dependent enzymes can inhibit further proinflammatory signaling and T-cell migration that can exacerbate IBD. Ultimately, more data is needed before determining the efficacy of pyridoxine supplementation for active IBD.

Vitamin B₁₂ and Vitamin B₉ (Folic Acid)

Pathophysiology—Vitamin B₁₂ is reabsorbed in the terminal ileum, the distal portion of the small intestine. The American Gastroenterological Association recommends that patients with a history of extensive ileal disease or prior ileal surgery, which is the case for many patients with Crohn disease, be monitored for vitamin B₁₂ deficiency.²³ Monitoring and rapid supplementation of vitamin B₁₂ can prevent pernicious anemia and irreversible neurologic damage that may result from deficiency.⁸⁴

Folic acid is primarily absorbed in the duodenum and jejunum of the small intestine. A meta-analysis performed in 2017 assessed studies observing folic acid and vitamin B₁₂ levels in 1086 patients with IBD compared with 1484 healthy controls and found an average difference in serum folate concentration of 0.46 nmol/L (P<.001).⁸⁴ Interestingly, this study did not find a significant difference in serum vitamin B₁₂ levels between patients with IBD and healthy controls, highlighting the mechanism of vitamin B₁₂ deficiency in IBD because only patients with terminal ileal involvement are at risk for malabsorption and subsequent deficiency.

Cutaneous Manifestations—Both vitamin B₁₂ and folic acid deficiency can manifest as cheilitis, glossitis, and/or generalized hyperpigmentation that is accentuated in the flexural areas, palms, soles, and oral cavity.^85,86 Systemic symptoms of patients with vitamin B₁₂ and folic acid deficiency include megaloblastic anemia, pallor, and fatigue. A potential mechanism for the hyperpigmentation observed from vitamin B₁₂ deficiency came from an electron microscope study that showed an increased concentration of melanosomes in a patient with deficiency.⁸⁷

Diagnosis and Monitoring—In patients with suspected vitamin B₁₂ and/or folic acid deficiency, initial evaluation should include a CBC with peripheral smear and serum vitamin B₁₂ and folate levels. In cases for which the diagnosis still is unclear after initial testing, methylmalonic acid and homocysteine levels can help differentiate between the 2 deficiencies. Methylmalonic acid classically is elevated (>260 nmol/L) in vitamin B₁₂ deficiency but not in folate deficiency.⁸⁸ Cut-off values for vitamin B₁₂ deficiency are less than 200 to 250 pg/mL forserum vitamin B₁₂ and/or an elevated level of methylmalonic acid (>0.271 µmol/L).⁸⁹ A serum folic acid value greater than 3 ng/mL and/or erythrocyte folate concentrations greater than 140 ng/mL are considered adequate, whereas an indicator of folic acid deficiency is a homocysteine level less than 10 µmol/L.⁹⁰ A CBC can screen for macrocytic megaloblastic anemias (mean corpuscular volume >100 fl), which are classic diagnostic signs of an underlying vitamin B₁₂ or folate deficiency.

Treatment—According to the Centers for Disease Control and Prevention, supplementation of vitamin B₁₂ can be done orally with 1000 µg daily in patients with deficiency. In patients with active IBD, oral reabsorption of vitamin B₁₂ can be less effective, making subcutaneous or intramuscular administration (1000 µg/wk for 8 weeks, then monthly for life) better options.⁸⁹

Patients with IBD managed with methotrexate should be screened carefully for folate deficiency. Methotrexate is a folate analog that sometimes is used for the treatment of IBD. Reversible competitive inhibition of dihydrofolate reductase can precipitate a systemic folic acid decrease.⁹¹ Typically, oral folic acid (1 to 5 mg/d) is sufficient to treat folate deficiency, with the ESPEN recommending 5 mg once weekly 24 to 72 hours after methotrexate treatment or 1 mg daily for 5 days per week in patients with IBD.¹ Alternative formulations—IV, subcutaneous, or intramuscular—are available for patients who cannot tolerate oral intake.⁹²

Final Thoughts

Dermatologists can be the first to observe the cutaneous manifestations of micronutrient deficiencies. Although the symptoms of each micronutrient deficiency discussed may overlap, attention to small clinical clues in patients with IBD can improve patient outcomes and quality of life. For example, koilonychia with glossitis and xerosis likely is due to iron deficiency, while zinc deficiency should be suspected in patients with scaly eczematous plaques in skin folds. A high level of suspicion for micronutrient deficiencies in patients with IBD should be followed by a complete patient history, review of systems, and thorough clinical examination. A thorough laboratory evaluation can pinpoint nutritional deficiencies in patients with IBD, keeping in mind that specific biomarkers such as ferritin and serum zinc also act as acute phase reactants and should be interpreted in this context. Co-management with gastroenterologists should be a priority in patients with IBD, as gaining control of inflammatory disease is crucial for the prevention of recurrent vitamin and micronutrient deficiencies in addition to long-term health in this population.

Medication-induced telogen effluvium (TE) is a nonscarring alopecia that typically is reversible. Appropriate management requires identification of the underlying trigger and cessation of potential culprit medications. In part 2 of this series, we review anticoagulant and antihypertensive medications as potential contributors to TE.

Anticoagulants

Anticoagulants target various parts of the coagulation cascade to prevent clot formation in patients with conditions that increase their risk for thromboembolic events. Common indications for initiating anticoagulant therapy include atrial fibrillation,¹ venous thromboembolism,² acute myocardial infarction,³ malignancy,⁴ and hypercoagulable states.⁵ Traditional anticoagulants include heparin and warfarin. Heparin is a glycosaminoglycan that exerts its anticoagulant effects through binding with antithrombin, greatly increasing its inactivation of thrombin and factor Xa of the coagulation cascade.⁶ Warfarin is a coumarin derivative that inhibits activation of vitamin K, subsequently limiting the function of vitamin K–dependent factors II, VII, IX, and X.^7,8 Watras et al⁹ noted that heparin and warfarin were implicated in alopecia as their clinical use became widespread throughout the mid-20th century. Onset of alopecia following the use of heparin or warfarin was reported at 3 weeks to 3 months following medication initiation, with most cases clinically consistent with TE.⁹ Heparin and warfarin both have alopecia reported as a potential adverse effect in their structured product labeling documents.^10,11

Heparin is further classified into unfractionated heparin (UFH) and low-molecular-weight heparin (LMWH); the latter is a heterogeneous group of medications derived from chemical or enzymatic depolymerization of UFH.¹² In contrast to UFH, LMWH exerts its anticoagulant effects through inactivation of factor Xa without the ability to bind thrombin.¹² An animal study using anagen-induced mice demonstrated that intraperitoneal administration of heparin inhibited the development of anagen follicles, while in vitro studies showed that the addition of heparin inhibited mouse dermal papilla cell proliferation.¹³ Other animal and in vitro studies have examined the inhibitory effects of heparin on signaling pathways in tumor lymphangiogenesis, including the vascular endothelial growth factor C/vascular endothelial growth factor receptor 3 axis.^14,15 Clinically, it has been demonstrated that heparin, especially LMWHs, may be associated with a survival benefit among certain cancer patients,^16,17 with the impact of LMWHs attributed to antimitotic and antimetastatic effects of heparin on tumor growth.¹⁴ It is hypothesized that such antiangiogenic and antimitotic effects also are involved in the pathomechanisms of heparin-induced alopecia.¹⁸

More recently, the use of direct oral anticoagulants (DOACs) such as dabigatran, rivaroxaban, and apixaban has increased due to their more favorable adverse-effect profile and minimal monitoring requirements. Bonaldo et al¹⁹ conducted an analysis of reports submitted to the World Health Organization’s VigiBase database of alopecia associated with DOACs until May 2, 2018. They found 1316 nonduplicate DOAC-induced cases of alopecia, with rivaroxaban as the most reported drug associated with alopecia development (58.8% [774/1316]). Only 4 cases demonstrated alopecia with DOAC rechallenge, suggesting onset of alopecia may have been unrelated to DOAC use or caused by a different trigger. Among 243 cases with a documented time to onset of alopecia, the median was 28 days, with an interquartile range of 63 days. Because TE most commonly occurs 3 to 4 months after the inciting event or medication trigger, there is little evidence to suggest DOACs as the cause of TE, and the observed cases of alopecia may be attributable to another preceding medical event and/or medication exposure.¹⁹ More studies are needed to examine the impact of anticoagulant medications on the hair cycle.

Antihypertensives

Hypertension is a modifiable risk factor for several ­cardiovascular diseases.²⁰ According to the 2019 American College of Cardiology/American Heart Association Guideline on the Primary Prevention of Cardiovascular Disease,²¹ first-line medications include thiazide diuretics, calcium channel blockers, angiotensin-converting enzyme (ACE) inhibitors, and angiotensin receptor ­blockers (ARBs).

Angiotensin-converting enzyme inhibitors exert their antihypertensive effects by reducing conversion of angiotensin I to angiotensin II, thereby limiting the downstream effects of vasoconstriction as well as sodium and water retention. Given the proven mortality benefit of ACE inhibition in patients with congestive heart failure, ACE inhibitors are used as first-line therapy in these patients.^22,23 Alopecia associated with ACE inhibitors is rare and limited to case reports following their introduction and approval in 1981.^24-28 In one case, a woman in her 60s with congestive heart failure initiated captopril with development of an erythematous pruritic rash on the extremities and diffuse scalp hair loss 2 months later; spontaneous hair growth resumed 1 month following captopril discontinuation.²⁵ In this case, the hair loss may be secondary to the drug eruption rather than true medication-induced TE. Initiation of enalapril in a woman in her 30s with hypertension was associated with diffuse scalp alopecia 4 weeks later that resolved with cessation of the suspected culprit, enalapril; rechallenge with enalapril several months later reproduced the hair loss.²⁷ Given limited reports of ACE inhibitor–associated hair loss relative to their pervasive use, a direct causal role between ACE inhibition and TE is unlikely, or it has not been rigorously identified. The structured product labeling for captopril includes alopecia in its list of adverse effects reported in approximately 0.5% to 2% of patients but did not appear at increased frequency compared to placebo or other treatments used in controlled trials.²⁹ Alternative inciting causes of alopecia in patients prescribed ACE inhibitors may include use of other medications, hospitalization, or metabolic derangements related to their underlying cardiac disease.

Although not indicated as a primary treatment for hypertension, β-blockers have US Food and Drug Administration approval for the treatment of certain arrhythmias, hypertension, heart failure, myocardial infarction, hyperthyroidism, and other conditions.³⁰β-Blockers are competitive antagonists of β-adrenergic receptors that limit the production of intracellular cyclic adenosine monophosphate, but the mechanism of β-blockers as antihypertensives is unclear.³¹ Evidence supporting the role of β-adrenergic antagonists in TE is limited to case reports. Widespread alopecia across the scalp and arms was noted in a man in his 30s several months after starting propranolol.³² Biopsy of an affected area of the scalp demonstrated an increased number of telogen follicles with no other abnormalities. Near-complete resolution of alopecia was seen 4 months following cessation of propranolol, which recurred within 4 weeks of rechallenge.³² Although the histopathologic features are compatible with TE, the loss of body hair and rapid recurrence within 4 weeks of rechallenge are atypical for TE. As such, the use of propranolol and the reported alopecia may be coincidental or evidence of an atypical drug reaction distinct from medication-induced TE. Only a handful of other case reports have been published describing TE in patients treated with β-blockers, including metoprolol and propranolol.^33,34 Alopecia has been reported with the use of carvedilol in up to 0.1% of participants.³⁵ Although cases have been reported, TE appears to be an uncommon occurrence following β-blocker therapy.

Minoxidil—Oral minoxidil originally was approved for use in patients with resistant hypertension, defined as blood pressure elevated above goal despite concurrent use of the maximum dose of 3 classes of antihypertensives.³⁶ Unlike other antihypertensive medications, minoxidil appears to cause reversible hypertrichosis that affects nearly all patients using oral minoxidil for longer than 1 month.³⁷ This common adverse effect was a desired outcome in patients affected by hair loss, and a topical formulation of minoxidil was approved for androgenetic alopecia in men and women in 1988 and 1991, respectively.³⁸ Since its approval, topical minoxidil has been commonly prescribed in the treatment of several types of alopecia, though evidence of its efficacy in the treatment of TE is limited.^39,40 Low-dose oral minoxidil also has been reported to aid hair growth in androgenetic alopecia and TE.⁴¹ Taken orally, minoxidil is converted by sulfotransferases in the liver to minoxidil sulfate, which causes opening of plasma membrane adenosine ­triphosphate–sensitive potassium channels.^42-44 The subsequent membrane hyperpolarization reduces calcium ion influx, which also reduces cell excitability, and inhibits contraction in vascular smooth muscle cells, which results in the arteriolar vasodilatory and antihypertensive effects of minoxidil.^43,45 The potassium channel–opening effects of minoxidil may underly its hair growth stimulatory action. Unrelated potassium channel openers such as diazoxide and pinacidil also cause hypertrichosis.^46-48 An animal study showed that topical minoxidil, cromakalim (potassium channel opener), and P1075 (pinacidil analog) applied daily to the scalps of balding stump-tailed macaques led to significant increases in hair weight over a 20-week treatment period compared with the vehicle control group (P<.05 for minoxidil 100 mM and 250 mM, cromakalim 100 mM, and P1075 100 mM and 250 mM).⁵⁰ For minoxidil, this effect on hair growth appears to be dose dependent, as cumulative hair weights for the study period were significantly greater in the 250-mM concentration compared with 100-mM minoxidil (P<.05).⁴⁹ The potassium channel–opening activity of minoxidil may induce stimulation of microcirculation around hair follicles conducive to hair growth.⁵⁰ Other proposed mechanisms for hair growth with minoxidil include effects on keratinocyte and fibroblast cell proliferation,^51-53 collagen synthesis,^52,54 and prostaglandin activity.^44,55

Final Thoughts

Medication-induced TE is an undesired adverse effect of many commonly used medications, including retinoids, azole antifungals, mood stabilizers, anticoagulants, and antihypertensives. In part 1⁵⁶ of this 2-part series, we reviewed the existing literature on hair loss from retinoids, antifungals, and psychotropic medications. Herein, we focused on anticoagulant and antihypertensive medications as potential culprits of TE. Heparin and its derivatives have been associated with development of diffuse alopecia weeks to months after the start of treatment. Alopecia associated with ACE inhibitors and β-blockers has been described only in case reports, suggesting that they may be unlikely causes of TE. In contrast, minoxidil is an antihypertensive that can result in hypertrichosis and is used in the treatment of androgenetic alopecia. It should not be assumed that medications that share an indication or are part of the same medication class would similarly induce TE. The development of diffuse nonscarring alopecia should prompt suspicion for TE and thorough investigation of medications initiated 1 to 6 months prior to onset of clinically apparent alopecia. Suspected culprit medications should be carefully assessed for their likelihood of inducing TE.

Medication-induced telogen effluvium (TE) is a nonscarring alopecia that typically is reversible. Appropriate management requires identification of the underlying trigger and cessation of potential culprit medications. In part 2 of this series, we review anticoagulant and antihypertensive medications as potential contributors to TE.

Anticoagulants

Anticoagulants target various parts of the coagulation cascade to prevent clot formation in patients with conditions that increase their risk for thromboembolic events. Common indications for initiating anticoagulant therapy include atrial fibrillation,¹ venous thromboembolism,² acute myocardial infarction,³ malignancy,⁴ and hypercoagulable states.⁵ Traditional anticoagulants include heparin and warfarin. Heparin is a glycosaminoglycan that exerts its anticoagulant effects through binding with antithrombin, greatly increasing its inactivation of thrombin and factor Xa of the coagulation cascade.⁶ Warfarin is a coumarin derivative that inhibits activation of vitamin K, subsequently limiting the function of vitamin K–dependent factors II, VII, IX, and X.^7,8 Watras et al⁹ noted that heparin and warfarin were implicated in alopecia as their clinical use became widespread throughout the mid-20th century. Onset of alopecia following the use of heparin or warfarin was reported at 3 weeks to 3 months following medication initiation, with most cases clinically consistent with TE.⁹ Heparin and warfarin both have alopecia reported as a potential adverse effect in their structured product labeling documents.^10,11

Heparin is further classified into unfractionated heparin (UFH) and low-molecular-weight heparin (LMWH); the latter is a heterogeneous group of medications derived from chemical or enzymatic depolymerization of UFH.¹² In contrast to UFH, LMWH exerts its anticoagulant effects through inactivation of factor Xa without the ability to bind thrombin.¹² An animal study using anagen-induced mice demonstrated that intraperitoneal administration of heparin inhibited the development of anagen follicles, while in vitro studies showed that the addition of heparin inhibited mouse dermal papilla cell proliferation.¹³ Other animal and in vitro studies have examined the inhibitory effects of heparin on signaling pathways in tumor lymphangiogenesis, including the vascular endothelial growth factor C/vascular endothelial growth factor receptor 3 axis.^14,15 Clinically, it has been demonstrated that heparin, especially LMWHs, may be associated with a survival benefit among certain cancer patients,^16,17 with the impact of LMWHs attributed to antimitotic and antimetastatic effects of heparin on tumor growth.¹⁴ It is hypothesized that such antiangiogenic and antimitotic effects also are involved in the pathomechanisms of heparin-induced alopecia.¹⁸

More recently, the use of direct oral anticoagulants (DOACs) such as dabigatran, rivaroxaban, and apixaban has increased due to their more favorable adverse-effect profile and minimal monitoring requirements. Bonaldo et al¹⁹ conducted an analysis of reports submitted to the World Health Organization’s VigiBase database of alopecia associated with DOACs until May 2, 2018. They found 1316 nonduplicate DOAC-induced cases of alopecia, with rivaroxaban as the most reported drug associated with alopecia development (58.8% [774/1316]). Only 4 cases demonstrated alopecia with DOAC rechallenge, suggesting onset of alopecia may have been unrelated to DOAC use or caused by a different trigger. Among 243 cases with a documented time to onset of alopecia, the median was 28 days, with an interquartile range of 63 days. Because TE most commonly occurs 3 to 4 months after the inciting event or medication trigger, there is little evidence to suggest DOACs as the cause of TE, and the observed cases of alopecia may be attributable to another preceding medical event and/or medication exposure.¹⁹ More studies are needed to examine the impact of anticoagulant medications on the hair cycle.

Antihypertensives

Hypertension is a modifiable risk factor for several ­cardiovascular diseases.²⁰ According to the 2019 American College of Cardiology/American Heart Association Guideline on the Primary Prevention of Cardiovascular Disease,²¹ first-line medications include thiazide diuretics, calcium channel blockers, angiotensin-converting enzyme (ACE) inhibitors, and angiotensin receptor ­blockers (ARBs).

Angiotensin-converting enzyme inhibitors exert their antihypertensive effects by reducing conversion of angiotensin I to angiotensin II, thereby limiting the downstream effects of vasoconstriction as well as sodium and water retention. Given the proven mortality benefit of ACE inhibition in patients with congestive heart failure, ACE inhibitors are used as first-line therapy in these patients.^22,23 Alopecia associated with ACE inhibitors is rare and limited to case reports following their introduction and approval in 1981.^24-28 In one case, a woman in her 60s with congestive heart failure initiated captopril with development of an erythematous pruritic rash on the extremities and diffuse scalp hair loss 2 months later; spontaneous hair growth resumed 1 month following captopril discontinuation.²⁵ In this case, the hair loss may be secondary to the drug eruption rather than true medication-induced TE. Initiation of enalapril in a woman in her 30s with hypertension was associated with diffuse scalp alopecia 4 weeks later that resolved with cessation of the suspected culprit, enalapril; rechallenge with enalapril several months later reproduced the hair loss.²⁷ Given limited reports of ACE inhibitor–associated hair loss relative to their pervasive use, a direct causal role between ACE inhibition and TE is unlikely, or it has not been rigorously identified. The structured product labeling for captopril includes alopecia in its list of adverse effects reported in approximately 0.5% to 2% of patients but did not appear at increased frequency compared to placebo or other treatments used in controlled trials.²⁹ Alternative inciting causes of alopecia in patients prescribed ACE inhibitors may include use of other medications, hospitalization, or metabolic derangements related to their underlying cardiac disease.

Although not indicated as a primary treatment for hypertension, β-blockers have US Food and Drug Administration approval for the treatment of certain arrhythmias, hypertension, heart failure, myocardial infarction, hyperthyroidism, and other conditions.³⁰β-Blockers are competitive antagonists of β-adrenergic receptors that limit the production of intracellular cyclic adenosine monophosphate, but the mechanism of β-blockers as antihypertensives is unclear.³¹ Evidence supporting the role of β-adrenergic antagonists in TE is limited to case reports. Widespread alopecia across the scalp and arms was noted in a man in his 30s several months after starting propranolol.³² Biopsy of an affected area of the scalp demonstrated an increased number of telogen follicles with no other abnormalities. Near-complete resolution of alopecia was seen 4 months following cessation of propranolol, which recurred within 4 weeks of rechallenge.³² Although the histopathologic features are compatible with TE, the loss of body hair and rapid recurrence within 4 weeks of rechallenge are atypical for TE. As such, the use of propranolol and the reported alopecia may be coincidental or evidence of an atypical drug reaction distinct from medication-induced TE. Only a handful of other case reports have been published describing TE in patients treated with β-blockers, including metoprolol and propranolol.^33,34 Alopecia has been reported with the use of carvedilol in up to 0.1% of participants.³⁵ Although cases have been reported, TE appears to be an uncommon occurrence following β-blocker therapy.

Minoxidil—Oral minoxidil originally was approved for use in patients with resistant hypertension, defined as blood pressure elevated above goal despite concurrent use of the maximum dose of 3 classes of antihypertensives.³⁶ Unlike other antihypertensive medications, minoxidil appears to cause reversible hypertrichosis that affects nearly all patients using oral minoxidil for longer than 1 month.³⁷ This common adverse effect was a desired outcome in patients affected by hair loss, and a topical formulation of minoxidil was approved for androgenetic alopecia in men and women in 1988 and 1991, respectively.³⁸ Since its approval, topical minoxidil has been commonly prescribed in the treatment of several types of alopecia, though evidence of its efficacy in the treatment of TE is limited.^39,40 Low-dose oral minoxidil also has been reported to aid hair growth in androgenetic alopecia and TE.⁴¹ Taken orally, minoxidil is converted by sulfotransferases in the liver to minoxidil sulfate, which causes opening of plasma membrane adenosine ­triphosphate–sensitive potassium channels.^42-44 The subsequent membrane hyperpolarization reduces calcium ion influx, which also reduces cell excitability, and inhibits contraction in vascular smooth muscle cells, which results in the arteriolar vasodilatory and antihypertensive effects of minoxidil.^43,45 The potassium channel–opening effects of minoxidil may underly its hair growth stimulatory action. Unrelated potassium channel openers such as diazoxide and pinacidil also cause hypertrichosis.^46-48 An animal study showed that topical minoxidil, cromakalim (potassium channel opener), and P1075 (pinacidil analog) applied daily to the scalps of balding stump-tailed macaques led to significant increases in hair weight over a 20-week treatment period compared with the vehicle control group (P<.05 for minoxidil 100 mM and 250 mM, cromakalim 100 mM, and P1075 100 mM and 250 mM).⁵⁰ For minoxidil, this effect on hair growth appears to be dose dependent, as cumulative hair weights for the study period were significantly greater in the 250-mM concentration compared with 100-mM minoxidil (P<.05).⁴⁹ The potassium channel–opening activity of minoxidil may induce stimulation of microcirculation around hair follicles conducive to hair growth.⁵⁰ Other proposed mechanisms for hair growth with minoxidil include effects on keratinocyte and fibroblast cell proliferation,^51-53 collagen synthesis,^52,54 and prostaglandin activity.^44,55

Final Thoughts

Medication-induced TE is an undesired adverse effect of many commonly used medications, including retinoids, azole antifungals, mood stabilizers, anticoagulants, and antihypertensives. In part 1⁵⁶ of this 2-part series, we reviewed the existing literature on hair loss from retinoids, antifungals, and psychotropic medications. Herein, we focused on anticoagulant and antihypertensive medications as potential culprits of TE. Heparin and its derivatives have been associated with development of diffuse alopecia weeks to months after the start of treatment. Alopecia associated with ACE inhibitors and β-blockers has been described only in case reports, suggesting that they may be unlikely causes of TE. In contrast, minoxidil is an antihypertensive that can result in hypertrichosis and is used in the treatment of androgenetic alopecia. It should not be assumed that medications that share an indication or are part of the same medication class would similarly induce TE. The development of diffuse nonscarring alopecia should prompt suspicion for TE and thorough investigation of medications initiated 1 to 6 months prior to onset of clinically apparent alopecia. Suspected culprit medications should be carefully assessed for their likelihood of inducing TE.

Medication-induced telogen effluvium (TE) is a nonscarring alopecia that typically is reversible. Appropriate management requires identification of the underlying trigger and cessation of potential culprit medications. In part 2 of this series, we review anticoagulant and antihypertensive medications as potential contributors to TE.

Anticoagulants

Anticoagulants target various parts of the coagulation cascade to prevent clot formation in patients with conditions that increase their risk for thromboembolic events. Common indications for initiating anticoagulant therapy include atrial fibrillation,¹ venous thromboembolism,² acute myocardial infarction,³ malignancy,⁴ and hypercoagulable states.⁵ Traditional anticoagulants include heparin and warfarin. Heparin is a glycosaminoglycan that exerts its anticoagulant effects through binding with antithrombin, greatly increasing its inactivation of thrombin and factor Xa of the coagulation cascade.⁶ Warfarin is a coumarin derivative that inhibits activation of vitamin K, subsequently limiting the function of vitamin K–dependent factors II, VII, IX, and X.^7,8 Watras et al⁹ noted that heparin and warfarin were implicated in alopecia as their clinical use became widespread throughout the mid-20th century. Onset of alopecia following the use of heparin or warfarin was reported at 3 weeks to 3 months following medication initiation, with most cases clinically consistent with TE.⁹ Heparin and warfarin both have alopecia reported as a potential adverse effect in their structured product labeling documents.^10,11

Heparin is further classified into unfractionated heparin (UFH) and low-molecular-weight heparin (LMWH); the latter is a heterogeneous group of medications derived from chemical or enzymatic depolymerization of UFH.¹² In contrast to UFH, LMWH exerts its anticoagulant effects through inactivation of factor Xa without the ability to bind thrombin.¹² An animal study using anagen-induced mice demonstrated that intraperitoneal administration of heparin inhibited the development of anagen follicles, while in vitro studies showed that the addition of heparin inhibited mouse dermal papilla cell proliferation.¹³ Other animal and in vitro studies have examined the inhibitory effects of heparin on signaling pathways in tumor lymphangiogenesis, including the vascular endothelial growth factor C/vascular endothelial growth factor receptor 3 axis.^14,15 Clinically, it has been demonstrated that heparin, especially LMWHs, may be associated with a survival benefit among certain cancer patients,^16,17 with the impact of LMWHs attributed to antimitotic and antimetastatic effects of heparin on tumor growth.¹⁴ It is hypothesized that such antiangiogenic and antimitotic effects also are involved in the pathomechanisms of heparin-induced alopecia.¹⁸

More recently, the use of direct oral anticoagulants (DOACs) such as dabigatran, rivaroxaban, and apixaban has increased due to their more favorable adverse-effect profile and minimal monitoring requirements. Bonaldo et al¹⁹ conducted an analysis of reports submitted to the World Health Organization’s VigiBase database of alopecia associated with DOACs until May 2, 2018. They found 1316 nonduplicate DOAC-induced cases of alopecia, with rivaroxaban as the most reported drug associated with alopecia development (58.8% [774/1316]). Only 4 cases demonstrated alopecia with DOAC rechallenge, suggesting onset of alopecia may have been unrelated to DOAC use or caused by a different trigger. Among 243 cases with a documented time to onset of alopecia, the median was 28 days, with an interquartile range of 63 days. Because TE most commonly occurs 3 to 4 months after the inciting event or medication trigger, there is little evidence to suggest DOACs as the cause of TE, and the observed cases of alopecia may be attributable to another preceding medical event and/or medication exposure.¹⁹ More studies are needed to examine the impact of anticoagulant medications on the hair cycle.

Antihypertensives

Hypertension is a modifiable risk factor for several ­cardiovascular diseases.²⁰ According to the 2019 American College of Cardiology/American Heart Association Guideline on the Primary Prevention of Cardiovascular Disease,²¹ first-line medications include thiazide diuretics, calcium channel blockers, angiotensin-converting enzyme (ACE) inhibitors, and angiotensin receptor ­blockers (ARBs).

Angiotensin-converting enzyme inhibitors exert their antihypertensive effects by reducing conversion of angiotensin I to angiotensin II, thereby limiting the downstream effects of vasoconstriction as well as sodium and water retention. Given the proven mortality benefit of ACE inhibition in patients with congestive heart failure, ACE inhibitors are used as first-line therapy in these patients.^22,23 Alopecia associated with ACE inhibitors is rare and limited to case reports following their introduction and approval in 1981.^24-28 In one case, a woman in her 60s with congestive heart failure initiated captopril with development of an erythematous pruritic rash on the extremities and diffuse scalp hair loss 2 months later; spontaneous hair growth resumed 1 month following captopril discontinuation.²⁵ In this case, the hair loss may be secondary to the drug eruption rather than true medication-induced TE. Initiation of enalapril in a woman in her 30s with hypertension was associated with diffuse scalp alopecia 4 weeks later that resolved with cessation of the suspected culprit, enalapril; rechallenge with enalapril several months later reproduced the hair loss.²⁷ Given limited reports of ACE inhibitor–associated hair loss relative to their pervasive use, a direct causal role between ACE inhibition and TE is unlikely, or it has not been rigorously identified. The structured product labeling for captopril includes alopecia in its list of adverse effects reported in approximately 0.5% to 2% of patients but did not appear at increased frequency compared to placebo or other treatments used in controlled trials.²⁹ Alternative inciting causes of alopecia in patients prescribed ACE inhibitors may include use of other medications, hospitalization, or metabolic derangements related to their underlying cardiac disease.

Although not indicated as a primary treatment for hypertension, β-blockers have US Food and Drug Administration approval for the treatment of certain arrhythmias, hypertension, heart failure, myocardial infarction, hyperthyroidism, and other conditions.³⁰β-Blockers are competitive antagonists of β-adrenergic receptors that limit the production of intracellular cyclic adenosine monophosphate, but the mechanism of β-blockers as antihypertensives is unclear.³¹ Evidence supporting the role of β-adrenergic antagonists in TE is limited to case reports. Widespread alopecia across the scalp and arms was noted in a man in his 30s several months after starting propranolol.³² Biopsy of an affected area of the scalp demonstrated an increased number of telogen follicles with no other abnormalities. Near-complete resolution of alopecia was seen 4 months following cessation of propranolol, which recurred within 4 weeks of rechallenge.³² Although the histopathologic features are compatible with TE, the loss of body hair and rapid recurrence within 4 weeks of rechallenge are atypical for TE. As such, the use of propranolol and the reported alopecia may be coincidental or evidence of an atypical drug reaction distinct from medication-induced TE. Only a handful of other case reports have been published describing TE in patients treated with β-blockers, including metoprolol and propranolol.^33,34 Alopecia has been reported with the use of carvedilol in up to 0.1% of participants.³⁵ Although cases have been reported, TE appears to be an uncommon occurrence following β-blocker therapy.

Minoxidil—Oral minoxidil originally was approved for use in patients with resistant hypertension, defined as blood pressure elevated above goal despite concurrent use of the maximum dose of 3 classes of antihypertensives.³⁶ Unlike other antihypertensive medications, minoxidil appears to cause reversible hypertrichosis that affects nearly all patients using oral minoxidil for longer than 1 month.³⁷ This common adverse effect was a desired outcome in patients affected by hair loss, and a topical formulation of minoxidil was approved for androgenetic alopecia in men and women in 1988 and 1991, respectively.³⁸ Since its approval, topical minoxidil has been commonly prescribed in the treatment of several types of alopecia, though evidence of its efficacy in the treatment of TE is limited.^39,40 Low-dose oral minoxidil also has been reported to aid hair growth in androgenetic alopecia and TE.⁴¹ Taken orally, minoxidil is converted by sulfotransferases in the liver to minoxidil sulfate, which causes opening of plasma membrane adenosine ­triphosphate–sensitive potassium channels.^42-44 The subsequent membrane hyperpolarization reduces calcium ion influx, which also reduces cell excitability, and inhibits contraction in vascular smooth muscle cells, which results in the arteriolar vasodilatory and antihypertensive effects of minoxidil.^43,45 The potassium channel–opening effects of minoxidil may underly its hair growth stimulatory action. Unrelated potassium channel openers such as diazoxide and pinacidil also cause hypertrichosis.^46-48 An animal study showed that topical minoxidil, cromakalim (potassium channel opener), and P1075 (pinacidil analog) applied daily to the scalps of balding stump-tailed macaques led to significant increases in hair weight over a 20-week treatment period compared with the vehicle control group (P<.05 for minoxidil 100 mM and 250 mM, cromakalim 100 mM, and P1075 100 mM and 250 mM).⁵⁰ For minoxidil, this effect on hair growth appears to be dose dependent, as cumulative hair weights for the study period were significantly greater in the 250-mM concentration compared with 100-mM minoxidil (P<.05).⁴⁹ The potassium channel–opening activity of minoxidil may induce stimulation of microcirculation around hair follicles conducive to hair growth.⁵⁰ Other proposed mechanisms for hair growth with minoxidil include effects on keratinocyte and fibroblast cell proliferation,^51-53 collagen synthesis,^52,54 and prostaglandin activity.^44,55

Final Thoughts

Medication-induced TE is an undesired adverse effect of many commonly used medications, including retinoids, azole antifungals, mood stabilizers, anticoagulants, and antihypertensives. In part 1⁵⁶ of this 2-part series, we reviewed the existing literature on hair loss from retinoids, antifungals, and psychotropic medications. Herein, we focused on anticoagulant and antihypertensive medications as potential culprits of TE. Heparin and its derivatives have been associated with development of diffuse alopecia weeks to months after the start of treatment. Alopecia associated with ACE inhibitors and β-blockers has been described only in case reports, suggesting that they may be unlikely causes of TE. In contrast, minoxidil is an antihypertensive that can result in hypertrichosis and is used in the treatment of androgenetic alopecia. It should not be assumed that medications that share an indication or are part of the same medication class would similarly induce TE. The development of diffuse nonscarring alopecia should prompt suspicion for TE and thorough investigation of medications initiated 1 to 6 months prior to onset of clinically apparent alopecia. Suspected culprit medications should be carefully assessed for their likelihood of inducing TE.

Alopecia is a commonly reported side effect of various medications. Anagen effluvium and telogen effluvium (TE) are considered the most common mechanisms underlying medication-related hair loss. Anagen effluvium is associated with chemotherapeutic agents and radiation therapy, with anagen shedding typically occurring within 2 weeks of medication administration.^1,2 Medication-induced TE is a diffuse nonscarring alopecia that is a reversible reactive process.^3-5 Telogen effluvium is clinically apparent as a generalized shedding of scalp hair 1 to 6 months after an inciting cause.⁶ The underlying cause of TE may be multifactorial and difficult to identify given the delay between the trigger and the onset of clinically apparent hair loss. Other known triggers of TE include acute illness,^7,8 nutritional deficiencies,^4,9 and/or major surgery.¹⁰

Each hair follicle independently and sequentially progresses through anagen growth, catagen transition, and telogen resting phases. In the human scalp, the telogen phase typically lasts 3 months, at the end of which the telogen hair is extruded from the scalp. Anagen and telogen follicles typically account for an average of 90% and 10% of follicles on the human scalp, respectively.¹¹ Immediate anagen release is hypothesized to be the mechanism underlying medication-induced TE.¹² This theory suggests that an increased percentage of anagen follicles prematurely enter the telogen phase, with a notable increase in hair shedding at the conclusion of the telogen phase approximately 1 to 6 months later.¹² First-line management of medication-induced TE is identification and cessation of the causative agent, if possible. Notable regrowth of hair is expected several months after removal of the inciting medication. In part 1 of this 2-part series, we review the existing literature to identify common culprits of medication-induced TE, including retinoids, antifungals, and psychotropic medications.

Retinoids

Retinoids are vitamin A derivatives used in the treatment of a myriad of dermatologic and nondermatologic conditions.^13,14 Retinoids modulate sebum production,¹⁵ keratinocyte proliferation,¹⁶ and epithelial differentiation through signal transduction downstream of the ligand-activated nuclear retinoic acid receptors and retinoid X receptors.^13,14,17 The recommended daily dosage of retinol is 900 µg retinol activity equivalent (3000 IU) for men and 700 µg retinol activity equivalent (2333 IU) for women. Retinoids are used in the treatment of acne vulgaris,¹⁸ psoriasis,¹⁹ and ichthyosis.²⁰ The most commonly reported adverse effects of systemic retinoid therapy include cheilitis, alopecia, and xerosis.²¹ Retinoid-associated alopecia is dose and duration dependent.^19,21-24 A prospective study of acitretin therapy in plaque psoriasis reported that more than 63% (42/66) of patients on 50 mg or more of acitretin daily for 6 months or longer experienced alopecia that reversed with discontinuation.²³ A systematic review of isotretinoin use in acne showed alopecia was seen in 3.2% (18/565) of patients on less than 0.5 mg/kg/d of isotretinoin and in 5.7% (192/3375) of patients on 0.5 mg/kg/d or less of isotretinoin.²⁴ In a phase 2 clinical trial of orally administered 9-cis-retinoic acid (alitretinoin) in the treatment of Kaposi sarcoma related to AIDS, 42% (24/57) of adult male patients receiving 60, 100, or 140 mg/m² alitretinoin daily (median treatment duration, 15.1 weeks) reported alopecia as an adverse effect of treatment.²⁵ In one case report, a patient who ingested 500,000 IU of vitamin A daily for 4 months and then 100,000 IU monthly for 6 months experienced diffusely increased shedding of scalp hair along with muscle soreness, nail dystrophy, diffuse skin rash, and refractory ascites; he was found to have severe liver damage secondary to hypervitaminosis A that required liver transplantation.²⁶ Regarding the pathomechanism of retinoid-induced alopecia, animal and in vitro studies similarly have demonstrated that all-trans-retinoic acid appears to exert its inhibitory effects on hair follicle growth via the influence of the transforming growth factor β2 and SMAD2/3 pathway influence on dermal papillae cells.^14,27 Development of hair loss secondary to systemic retinoid therapy may be managed with dose reduction or cessation.

Antifungals

Azole medications have broad-spectrum fungistatic activity against a wide range of yeast and filamentous fungi. Azoles inhibit sterol 14α-demethylase activity, impairing ergosterol synthesis and thereby disrupting plasma membrane synthesis and activity of membrane-bound enzymes.²⁸ Fluconazole is a systemic oral agent in this class that was first approved by the US Food and Drug Administration (FDA) for use in the 1990s.²⁹ A retrospective study by the National Institute of Allergy and Infectious Disease Mycoses Study Group followed the clinical course of 33 patients who developed alopecia while receiving fluconazole therapy for various mycoses.³⁰ The majority (88% [29/33]) of patients received 400 mg or more of fluconazole daily. The median time to hair loss after starting fluconazole was 3 months, and the scalp was involved in all cases. In 97% (32/33) of patients, resolution of alopecia was noted following discontinuation of fluconazole or a dose reduction of 50% or more. In 85% (28/33) of patients, complete resolution of alopecia occurred within 6 months of fluconazole cessation or dose reduction.³⁰ Fluconazole-induced TE was reproducible in an animal model using Wistar rats³¹; however, further studies are required to clarify the molecular pathways of its effect on hair growth.

Voriconazole is an azole approved for the treatment of invasive aspergillosis, candidemia, and fungal infections caused by Scedosporium apiospermum and Fusarium species. A retrospective survey study of patients who received voriconazole for 1 month or longer found a considerable proportion of patients developed diffuse reversible hair loss.³² Scalp alopecia was noted in 79% (120/152) of patients who completed the survey, with a mean (SD) time to alopecia of 75 (54) days after initiation of voriconazole. Notable regrowth was reported in 69% (79/114) of patients who discontinued voriconazole for at least 3 months. A subgroup of 32 patients were changed to itraconazole or posaconazole, and hair loss stopped in 84% (27/32) with regrowth noted in 69% (22/32) of patients.³² Voriconazole and fluconazole share structural similarity not present with other triazoles.^33,34 Because voriconazole-associated alopecia was reversed in the majority of patients who switched to itraconazole or posaconazole, the authors hypothesized that structural similarity of fluconazole and voriconazole may underly the greater risk for TE that is not a class effect of azole medications.³¹

Psychotropic Medications

Various psychotropic medications have been associated with hair loss. Valproic acid (or sodium valproate) is an anticonvulsant and mood-stabilizing agent used for the treatment of seizures, bipolar disorder (BD), migraines, and neuropathic pain.^35,36 Divalproex sodium (or divalproex) is an enteric-coated formulation of sodium valproate and valproic acid with similar indications. Valproate is a notorious culprit of medication-induced hair loss, with alopecia listed among the most common adverse reactions (reported >5%) on its structure product labeling document.³⁷ A systemic review and meta-analysis by Wang et al³⁸ estimated the overall incidence of valproate-related alopecia to be 11% (95% CI, 0.08-0.13). Although this meta-analysis did not find an association between incidence of alopecia and dose or duration of valproate therapy,³⁸ a separate review suggested that valproate-induced alopecia is dose dependent and can be managed with dose reduction.³⁹ A 12-month, randomized, double-blind study of treatment of BD with divalproex (valproate derivative), lithium, or placebo (2:1:1 ratio) showed a significantly higher frequency of alopecia in the divalproex group compared with placebo (16% [30/187] vs 6% [6/94]; P=.03).⁴⁰ Valproate-related hair loss is characteristically diffuse and nonscarring, often noted 3 to 6 months following initiation of valproate.^41,42 The proposed mechanism of valproate-induced alopecia includes chelation of zinc and selenium,⁴³ and a reduction in serum biotinidase activity, thereby decreasing the availability of these essential micronutrients required for hair growth.⁴¹ Studies examining the effects of valproate administration and serum biotinidase activity in patients have yielded conflicting results.^44-46 In a study of children with seizures including 57 patients treated with valproic acid, 17 treated with carbamazepine, and 75 age- and sex-matched healthy controls, the authors found no significant differences in serum biotinidase enzyme activity across the 3 groups.⁴⁴ In contrast, a study of 75 children with seizures on valproic acid therapy stratified by dose (mean [SD])—group A: 28.7 [8.5] mg/kg/d; group B: 41.6 [4.9] mg/kg/d; group C: 64.5 [5.8] mg/kg/d—found that patients receiving higher doses (groups B and C) had significantly reduced serum biotinidase activity (1.22 [1.11] and 0.97 [0.07] mmol/min/L, respectively) compared with 50 healthy pediatric controls (5.20 [0.90] mmol/min/L; P<.001). The same study found biotin supplementation at 10 mg/d for 20 days led to resolution of alopecia in 22% (2/9) of patients with alopecia on valproic acid therapy.⁴⁵ Despite hypothesized effects of valproate on micronutrients, the role of mineral supplementation in treating valproate-associated hair loss remains unclear. There is evidence to suggest that valproic acid–associated alterations in serum biotinidase activity may be transient. In a study of 32 pediatric patients receiving valproic acid for the treatment of epilepsy, serum biotinidase activity was significantly lower after 3 months of valproic acid therapy compared with pretreatment levels (P<.05); at 6 months, the serum biotinidase activity was increased compared with 3 months (P<.05) and not significantly different from pretreatment levels (P>.05).⁴⁶ Hair regrowth has been observed following discontinuation or dose reduction of valproate therapy in some cases.^39,47

Lithium carbonate (lithium) is used in the treatment of BD. Despite its efficacy and low cost, its potential for adverse effects, narrow therapeutic index, and subsequent need for routine monitoring are factors that limit its use.⁴⁸ Some reported dermatologic adverse reactions on its structure product labeling include xerosis, thinning of hair, alopecia, xerosis cutis, psoriasis onset/exacerbation, and generalized pruritus.⁴⁹ A systematic review and meta-analysis of 385 studies identified 24 publications reporting adverse effects of lithium on hair with no significantly increased risk of alopecia overall.⁵⁰ The analysis included 2 randomized controlled trials comparing the effects of lithium and placebo on hair loss in patients with BD. Hair loss was reported in 7% (7/94) of patients taking lithium and 6% (6/94) of the placebo group in the 12-month study⁴⁰ and in 3% (1/32) of the lithium group and 0% (0/28) of the divalproex group in the 20-month study.⁵¹ Despite anecdotal reports of alopecia associated with lithium, there is a lack of high-quality evidence to support this claim. Of note, hypothyroidism is a known complication of lithium use, and serum testing of thyroid function at 6-month intervals is recommended for patients on lithium treatment.⁵² Because thyroid abnormalities can cause alopecia distinct from TE, new-onset alopecia during lithium use should prompt serum testing of thyroid function. The development of hypothyroidism secondary to lithium is not a direct contraindication to its use⁵³; rather, treatment should be focused on correction with thyroid replacement therapy (eg, supplementation with thyroxine).⁵⁴

Commonly prescribed antidepressant medications include selective serotonin reuptake inhibitors (SSRIs) and bupropion. Selective serotonin reuptake inhibitors affect the neuronal serotonin transporter, increasing the concentration of serotonin in the synaptic cleft available for stimulation of postsynaptic serotonin receptors^55,56; bupropion is an antidepressant medication that inhibits norepinephrine and dopamine reuptake at the synaptic cleft.⁵⁷ Alopecia is an infrequent (1 in 100 to 1 in 1000 patients) adverse effect for several SSRIs.^58-62 A recent systematic review identified a total of 71 cases of alopecia associated with SSRI use including citalopram (n=11), escitalopram (n=7), fluoxetine (n=27), fluoxvamine (n=5), paroxetine (n=4), and sertraline (n=20), with a median time to onset of hair shedding of 8.6 weeks (range, 3 days to 5 years). Discontinuation of the suspected culprit SSRI led to improvement and/or resolution in 63% (51/81) episodes of alopecia, with a median time to improvement and/or resolution of 4 weeks.⁶³ A comparative retrospective cohort study using a large US health claims database from 2006 to 2014 included more than 1 million new and mutually exclusive patients taking fluoxetine, fluvoxamine, sertraline, citalopram, escitalopram, paroxetine, duloxetine, venlafaxine, desvenlafaxine, and bupropion.⁶⁴ Overall, 1% (1569/150,404) of patients treated with bupropion received 1 or more physician visits for alopecia. Patients on SSRIs generally had a lower risk for hair loss compared with patients using bupropion (citalopram: hazard ratio [HR], 0.80 [95% CI, 0.74-0.86]; escitalopram: HR, 0.79 [95% CI, 0.74-0.86]; fluoxetine: HR, 0.68 [95% CI, 0.63-0.74]; paroxetine: HR, 0.68 [95% CI, 0.62-0.74]; sertraline: HR, 0.74 [95% CI, 0.69-0.79]), with the exception of fluvoxamine (HR, 0.93 [95% CI, 0.64-1.37]). However, the type of alopecia, time to onset, and time to resolution were not reported, making it difficult to assess whether the reported hair loss was consistent with medication-induced TE. Additionally, the authors acknowledged that bupropion may have been prescribed for smoking cessation, which may carry a different risk profile for the development of alopecia.⁶⁴ Several other case reports have described alopecia following treatment with SSRIs, including sertraline,⁶⁵ fluvoxamine,⁶⁶ paroxetine,⁶⁷ fluoxetine,⁶⁸ and escitalopram.⁶⁹

Overall, it appears that the use of SSRIs portends relatively low risk for alopecia and medication-induced TE. Little is known regarding the molecular effects of SSRIs on hair growth and the pathomechanism of SSRI-induced TE. The potential benefits of discontinuing a suspected culprit medication should be carefully weighed against the risks of medication cessation, and consideration should be given to alternative medications in the same class that also may be associated with TE. In patients requiring antidepressant therapy with suspected medication-induced TE, consider transitioning to a different class of medication with lower risk of medication-induced alopecia; for example, discontinuing bupropion in favor of an SSRI.

Final Thoughts

Medication-induced alopecia is an undesired side effect of many commonly used drugs and drug classes, including retinoids, azole antifungals, and mood stabilizers. Although the precise pathomechanisms of medication-induced TE remain unclear, the recommended management often requires identification of the likely causative agent and its discontinuation, if possible. Suspicion for medication-induced TE should prompt a thorough history of recent changes to medications, risk factors for nutritional deficiencies, underlying illnesses, and recent surgical procedures. Underlying nutritional, electrolyte, and/or metabolic disturbances should be corrected. In part 2 of this series, we will discuss medication-induced alopecia associated with anticoagulant and antihypertensive medications.

Alopecia is a commonly reported side effect of various medications. Anagen effluvium and telogen effluvium (TE) are considered the most common mechanisms underlying medication-related hair loss. Anagen effluvium is associated with chemotherapeutic agents and radiation therapy, with anagen shedding typically occurring within 2 weeks of medication administration.^1,2 Medication-induced TE is a diffuse nonscarring alopecia that is a reversible reactive process.^3-5 Telogen effluvium is clinically apparent as a generalized shedding of scalp hair 1 to 6 months after an inciting cause.⁶ The underlying cause of TE may be multifactorial and difficult to identify given the delay between the trigger and the onset of clinically apparent hair loss. Other known triggers of TE include acute illness,^7,8 nutritional deficiencies,^4,9 and/or major surgery.¹⁰

Each hair follicle independently and sequentially progresses through anagen growth, catagen transition, and telogen resting phases. In the human scalp, the telogen phase typically lasts 3 months, at the end of which the telogen hair is extruded from the scalp. Anagen and telogen follicles typically account for an average of 90% and 10% of follicles on the human scalp, respectively.¹¹ Immediate anagen release is hypothesized to be the mechanism underlying medication-induced TE.¹² This theory suggests that an increased percentage of anagen follicles prematurely enter the telogen phase, with a notable increase in hair shedding at the conclusion of the telogen phase approximately 1 to 6 months later.¹² First-line management of medication-induced TE is identification and cessation of the causative agent, if possible. Notable regrowth of hair is expected several months after removal of the inciting medication. In part 1 of this 2-part series, we review the existing literature to identify common culprits of medication-induced TE, including retinoids, antifungals, and psychotropic medications.

Retinoids

Retinoids are vitamin A derivatives used in the treatment of a myriad of dermatologic and nondermatologic conditions.^13,14 Retinoids modulate sebum production,¹⁵ keratinocyte proliferation,¹⁶ and epithelial differentiation through signal transduction downstream of the ligand-activated nuclear retinoic acid receptors and retinoid X receptors.^13,14,17 The recommended daily dosage of retinol is 900 µg retinol activity equivalent (3000 IU) for men and 700 µg retinol activity equivalent (2333 IU) for women. Retinoids are used in the treatment of acne vulgaris,¹⁸ psoriasis,¹⁹ and ichthyosis.²⁰ The most commonly reported adverse effects of systemic retinoid therapy include cheilitis, alopecia, and xerosis.²¹ Retinoid-associated alopecia is dose and duration dependent.^19,21-24 A prospective study of acitretin therapy in plaque psoriasis reported that more than 63% (42/66) of patients on 50 mg or more of acitretin daily for 6 months or longer experienced alopecia that reversed with discontinuation.²³ A systematic review of isotretinoin use in acne showed alopecia was seen in 3.2% (18/565) of patients on less than 0.5 mg/kg/d of isotretinoin and in 5.7% (192/3375) of patients on 0.5 mg/kg/d or less of isotretinoin.²⁴ In a phase 2 clinical trial of orally administered 9-cis-retinoic acid (alitretinoin) in the treatment of Kaposi sarcoma related to AIDS, 42% (24/57) of adult male patients receiving 60, 100, or 140 mg/m² alitretinoin daily (median treatment duration, 15.1 weeks) reported alopecia as an adverse effect of treatment.²⁵ In one case report, a patient who ingested 500,000 IU of vitamin A daily for 4 months and then 100,000 IU monthly for 6 months experienced diffusely increased shedding of scalp hair along with muscle soreness, nail dystrophy, diffuse skin rash, and refractory ascites; he was found to have severe liver damage secondary to hypervitaminosis A that required liver transplantation.²⁶ Regarding the pathomechanism of retinoid-induced alopecia, animal and in vitro studies similarly have demonstrated that all-trans-retinoic acid appears to exert its inhibitory effects on hair follicle growth via the influence of the transforming growth factor β2 and SMAD2/3 pathway influence on dermal papillae cells.^14,27 Development of hair loss secondary to systemic retinoid therapy may be managed with dose reduction or cessation.

Antifungals

Azole medications have broad-spectrum fungistatic activity against a wide range of yeast and filamentous fungi. Azoles inhibit sterol 14α-demethylase activity, impairing ergosterol synthesis and thereby disrupting plasma membrane synthesis and activity of membrane-bound enzymes.²⁸ Fluconazole is a systemic oral agent in this class that was first approved by the US Food and Drug Administration (FDA) for use in the 1990s.²⁹ A retrospective study by the National Institute of Allergy and Infectious Disease Mycoses Study Group followed the clinical course of 33 patients who developed alopecia while receiving fluconazole therapy for various mycoses.³⁰ The majority (88% [29/33]) of patients received 400 mg or more of fluconazole daily. The median time to hair loss after starting fluconazole was 3 months, and the scalp was involved in all cases. In 97% (32/33) of patients, resolution of alopecia was noted following discontinuation of fluconazole or a dose reduction of 50% or more. In 85% (28/33) of patients, complete resolution of alopecia occurred within 6 months of fluconazole cessation or dose reduction.³⁰ Fluconazole-induced TE was reproducible in an animal model using Wistar rats³¹; however, further studies are required to clarify the molecular pathways of its effect on hair growth.

Voriconazole is an azole approved for the treatment of invasive aspergillosis, candidemia, and fungal infections caused by Scedosporium apiospermum and Fusarium species. A retrospective survey study of patients who received voriconazole for 1 month or longer found a considerable proportion of patients developed diffuse reversible hair loss.³² Scalp alopecia was noted in 79% (120/152) of patients who completed the survey, with a mean (SD) time to alopecia of 75 (54) days after initiation of voriconazole. Notable regrowth was reported in 69% (79/114) of patients who discontinued voriconazole for at least 3 months. A subgroup of 32 patients were changed to itraconazole or posaconazole, and hair loss stopped in 84% (27/32) with regrowth noted in 69% (22/32) of patients.³² Voriconazole and fluconazole share structural similarity not present with other triazoles.^33,34 Because voriconazole-associated alopecia was reversed in the majority of patients who switched to itraconazole or posaconazole, the authors hypothesized that structural similarity of fluconazole and voriconazole may underly the greater risk for TE that is not a class effect of azole medications.³¹

Psychotropic Medications

Various psychotropic medications have been associated with hair loss. Valproic acid (or sodium valproate) is an anticonvulsant and mood-stabilizing agent used for the treatment of seizures, bipolar disorder (BD), migraines, and neuropathic pain.^35,36 Divalproex sodium (or divalproex) is an enteric-coated formulation of sodium valproate and valproic acid with similar indications. Valproate is a notorious culprit of medication-induced hair loss, with alopecia listed among the most common adverse reactions (reported >5%) on its structure product labeling document.³⁷ A systemic review and meta-analysis by Wang et al³⁸ estimated the overall incidence of valproate-related alopecia to be 11% (95% CI, 0.08-0.13). Although this meta-analysis did not find an association between incidence of alopecia and dose or duration of valproate therapy,³⁸ a separate review suggested that valproate-induced alopecia is dose dependent and can be managed with dose reduction.³⁹ A 12-month, randomized, double-blind study of treatment of BD with divalproex (valproate derivative), lithium, or placebo (2:1:1 ratio) showed a significantly higher frequency of alopecia in the divalproex group compared with placebo (16% [30/187] vs 6% [6/94]; P=.03).⁴⁰ Valproate-related hair loss is characteristically diffuse and nonscarring, often noted 3 to 6 months following initiation of valproate.^41,42 The proposed mechanism of valproate-induced alopecia includes chelation of zinc and selenium,⁴³ and a reduction in serum biotinidase activity, thereby decreasing the availability of these essential micronutrients required for hair growth.⁴¹ Studies examining the effects of valproate administration and serum biotinidase activity in patients have yielded conflicting results.^44-46 In a study of children with seizures including 57 patients treated with valproic acid, 17 treated with carbamazepine, and 75 age- and sex-matched healthy controls, the authors found no significant differences in serum biotinidase enzyme activity across the 3 groups.⁴⁴ In contrast, a study of 75 children with seizures on valproic acid therapy stratified by dose (mean [SD])—group A: 28.7 [8.5] mg/kg/d; group B: 41.6 [4.9] mg/kg/d; group C: 64.5 [5.8] mg/kg/d—found that patients receiving higher doses (groups B and C) had significantly reduced serum biotinidase activity (1.22 [1.11] and 0.97 [0.07] mmol/min/L, respectively) compared with 50 healthy pediatric controls (5.20 [0.90] mmol/min/L; P<.001). The same study found biotin supplementation at 10 mg/d for 20 days led to resolution of alopecia in 22% (2/9) of patients with alopecia on valproic acid therapy.⁴⁵ Despite hypothesized effects of valproate on micronutrients, the role of mineral supplementation in treating valproate-associated hair loss remains unclear. There is evidence to suggest that valproic acid–associated alterations in serum biotinidase activity may be transient. In a study of 32 pediatric patients receiving valproic acid for the treatment of epilepsy, serum biotinidase activity was significantly lower after 3 months of valproic acid therapy compared with pretreatment levels (P<.05); at 6 months, the serum biotinidase activity was increased compared with 3 months (P<.05) and not significantly different from pretreatment levels (P>.05).⁴⁶ Hair regrowth has been observed following discontinuation or dose reduction of valproate therapy in some cases.^39,47

Lithium carbonate (lithium) is used in the treatment of BD. Despite its efficacy and low cost, its potential for adverse effects, narrow therapeutic index, and subsequent need for routine monitoring are factors that limit its use.⁴⁸ Some reported dermatologic adverse reactions on its structure product labeling include xerosis, thinning of hair, alopecia, xerosis cutis, psoriasis onset/exacerbation, and generalized pruritus.⁴⁹ A systematic review and meta-analysis of 385 studies identified 24 publications reporting adverse effects of lithium on hair with no significantly increased risk of alopecia overall.⁵⁰ The analysis included 2 randomized controlled trials comparing the effects of lithium and placebo on hair loss in patients with BD. Hair loss was reported in 7% (7/94) of patients taking lithium and 6% (6/94) of the placebo group in the 12-month study⁴⁰ and in 3% (1/32) of the lithium group and 0% (0/28) of the divalproex group in the 20-month study.⁵¹ Despite anecdotal reports of alopecia associated with lithium, there is a lack of high-quality evidence to support this claim. Of note, hypothyroidism is a known complication of lithium use, and serum testing of thyroid function at 6-month intervals is recommended for patients on lithium treatment.⁵² Because thyroid abnormalities can cause alopecia distinct from TE, new-onset alopecia during lithium use should prompt serum testing of thyroid function. The development of hypothyroidism secondary to lithium is not a direct contraindication to its use⁵³; rather, treatment should be focused on correction with thyroid replacement therapy (eg, supplementation with thyroxine).⁵⁴

Commonly prescribed antidepressant medications include selective serotonin reuptake inhibitors (SSRIs) and bupropion. Selective serotonin reuptake inhibitors affect the neuronal serotonin transporter, increasing the concentration of serotonin in the synaptic cleft available for stimulation of postsynaptic serotonin receptors^55,56; bupropion is an antidepressant medication that inhibits norepinephrine and dopamine reuptake at the synaptic cleft.⁵⁷ Alopecia is an infrequent (1 in 100 to 1 in 1000 patients) adverse effect for several SSRIs.^58-62 A recent systematic review identified a total of 71 cases of alopecia associated with SSRI use including citalopram (n=11), escitalopram (n=7), fluoxetine (n=27), fluoxvamine (n=5), paroxetine (n=4), and sertraline (n=20), with a median time to onset of hair shedding of 8.6 weeks (range, 3 days to 5 years). Discontinuation of the suspected culprit SSRI led to improvement and/or resolution in 63% (51/81) episodes of alopecia, with a median time to improvement and/or resolution of 4 weeks.⁶³ A comparative retrospective cohort study using a large US health claims database from 2006 to 2014 included more than 1 million new and mutually exclusive patients taking fluoxetine, fluvoxamine, sertraline, citalopram, escitalopram, paroxetine, duloxetine, venlafaxine, desvenlafaxine, and bupropion.⁶⁴ Overall, 1% (1569/150,404) of patients treated with bupropion received 1 or more physician visits for alopecia. Patients on SSRIs generally had a lower risk for hair loss compared with patients using bupropion (citalopram: hazard ratio [HR], 0.80 [95% CI, 0.74-0.86]; escitalopram: HR, 0.79 [95% CI, 0.74-0.86]; fluoxetine: HR, 0.68 [95% CI, 0.63-0.74]; paroxetine: HR, 0.68 [95% CI, 0.62-0.74]; sertraline: HR, 0.74 [95% CI, 0.69-0.79]), with the exception of fluvoxamine (HR, 0.93 [95% CI, 0.64-1.37]). However, the type of alopecia, time to onset, and time to resolution were not reported, making it difficult to assess whether the reported hair loss was consistent with medication-induced TE. Additionally, the authors acknowledged that bupropion may have been prescribed for smoking cessation, which may carry a different risk profile for the development of alopecia.⁶⁴ Several other case reports have described alopecia following treatment with SSRIs, including sertraline,⁶⁵ fluvoxamine,⁶⁶ paroxetine,⁶⁷ fluoxetine,⁶⁸ and escitalopram.⁶⁹

Overall, it appears that the use of SSRIs portends relatively low risk for alopecia and medication-induced TE. Little is known regarding the molecular effects of SSRIs on hair growth and the pathomechanism of SSRI-induced TE. The potential benefits of discontinuing a suspected culprit medication should be carefully weighed against the risks of medication cessation, and consideration should be given to alternative medications in the same class that also may be associated with TE. In patients requiring antidepressant therapy with suspected medication-induced TE, consider transitioning to a different class of medication with lower risk of medication-induced alopecia; for example, discontinuing bupropion in favor of an SSRI.

Final Thoughts

Medication-induced alopecia is an undesired side effect of many commonly used drugs and drug classes, including retinoids, azole antifungals, and mood stabilizers. Although the precise pathomechanisms of medication-induced TE remain unclear, the recommended management often requires identification of the likely causative agent and its discontinuation, if possible. Suspicion for medication-induced TE should prompt a thorough history of recent changes to medications, risk factors for nutritional deficiencies, underlying illnesses, and recent surgical procedures. Underlying nutritional, electrolyte, and/or metabolic disturbances should be corrected. In part 2 of this series, we will discuss medication-induced alopecia associated with anticoagulant and antihypertensive medications.

Alopecia is a commonly reported side effect of various medications. Anagen effluvium and telogen effluvium (TE) are considered the most common mechanisms underlying medication-related hair loss. Anagen effluvium is associated with chemotherapeutic agents and radiation therapy, with anagen shedding typically occurring within 2 weeks of medication administration.^1,2 Medication-induced TE is a diffuse nonscarring alopecia that is a reversible reactive process.^3-5 Telogen effluvium is clinically apparent as a generalized shedding of scalp hair 1 to 6 months after an inciting cause.⁶ The underlying cause of TE may be multifactorial and difficult to identify given the delay between the trigger and the onset of clinically apparent hair loss. Other known triggers of TE include acute illness,^7,8 nutritional deficiencies,^4,9 and/or major surgery.¹⁰

Each hair follicle independently and sequentially progresses through anagen growth, catagen transition, and telogen resting phases. In the human scalp, the telogen phase typically lasts 3 months, at the end of which the telogen hair is extruded from the scalp. Anagen and telogen follicles typically account for an average of 90% and 10% of follicles on the human scalp, respectively.¹¹ Immediate anagen release is hypothesized to be the mechanism underlying medication-induced TE.¹² This theory suggests that an increased percentage of anagen follicles prematurely enter the telogen phase, with a notable increase in hair shedding at the conclusion of the telogen phase approximately 1 to 6 months later.¹² First-line management of medication-induced TE is identification and cessation of the causative agent, if possible. Notable regrowth of hair is expected several months after removal of the inciting medication. In part 1 of this 2-part series, we review the existing literature to identify common culprits of medication-induced TE, including retinoids, antifungals, and psychotropic medications.

Retinoids

Retinoids are vitamin A derivatives used in the treatment of a myriad of dermatologic and nondermatologic conditions.^13,14 Retinoids modulate sebum production,¹⁵ keratinocyte proliferation,¹⁶ and epithelial differentiation through signal transduction downstream of the ligand-activated nuclear retinoic acid receptors and retinoid X receptors.^13,14,17 The recommended daily dosage of retinol is 900 µg retinol activity equivalent (3000 IU) for men and 700 µg retinol activity equivalent (2333 IU) for women. Retinoids are used in the treatment of acne vulgaris,¹⁸ psoriasis,¹⁹ and ichthyosis.²⁰ The most commonly reported adverse effects of systemic retinoid therapy include cheilitis, alopecia, and xerosis.²¹ Retinoid-associated alopecia is dose and duration dependent.^19,21-24 A prospective study of acitretin therapy in plaque psoriasis reported that more than 63% (42/66) of patients on 50 mg or more of acitretin daily for 6 months or longer experienced alopecia that reversed with discontinuation.²³ A systematic review of isotretinoin use in acne showed alopecia was seen in 3.2% (18/565) of patients on less than 0.5 mg/kg/d of isotretinoin and in 5.7% (192/3375) of patients on 0.5 mg/kg/d or less of isotretinoin.²⁴ In a phase 2 clinical trial of orally administered 9-cis-retinoic acid (alitretinoin) in the treatment of Kaposi sarcoma related to AIDS, 42% (24/57) of adult male patients receiving 60, 100, or 140 mg/m² alitretinoin daily (median treatment duration, 15.1 weeks) reported alopecia as an adverse effect of treatment.²⁵ In one case report, a patient who ingested 500,000 IU of vitamin A daily for 4 months and then 100,000 IU monthly for 6 months experienced diffusely increased shedding of scalp hair along with muscle soreness, nail dystrophy, diffuse skin rash, and refractory ascites; he was found to have severe liver damage secondary to hypervitaminosis A that required liver transplantation.²⁶ Regarding the pathomechanism of retinoid-induced alopecia, animal and in vitro studies similarly have demonstrated that all-trans-retinoic acid appears to exert its inhibitory effects on hair follicle growth via the influence of the transforming growth factor β2 and SMAD2/3 pathway influence on dermal papillae cells.^14,27 Development of hair loss secondary to systemic retinoid therapy may be managed with dose reduction or cessation.

Antifungals

Azole medications have broad-spectrum fungistatic activity against a wide range of yeast and filamentous fungi. Azoles inhibit sterol 14α-demethylase activity, impairing ergosterol synthesis and thereby disrupting plasma membrane synthesis and activity of membrane-bound enzymes.²⁸ Fluconazole is a systemic oral agent in this class that was first approved by the US Food and Drug Administration (FDA) for use in the 1990s.²⁹ A retrospective study by the National Institute of Allergy and Infectious Disease Mycoses Study Group followed the clinical course of 33 patients who developed alopecia while receiving fluconazole therapy for various mycoses.³⁰ The majority (88% [29/33]) of patients received 400 mg or more of fluconazole daily. The median time to hair loss after starting fluconazole was 3 months, and the scalp was involved in all cases. In 97% (32/33) of patients, resolution of alopecia was noted following discontinuation of fluconazole or a dose reduction of 50% or more. In 85% (28/33) of patients, complete resolution of alopecia occurred within 6 months of fluconazole cessation or dose reduction.³⁰ Fluconazole-induced TE was reproducible in an animal model using Wistar rats³¹; however, further studies are required to clarify the molecular pathways of its effect on hair growth.

Voriconazole is an azole approved for the treatment of invasive aspergillosis, candidemia, and fungal infections caused by Scedosporium apiospermum and Fusarium species. A retrospective survey study of patients who received voriconazole for 1 month or longer found a considerable proportion of patients developed diffuse reversible hair loss.³² Scalp alopecia was noted in 79% (120/152) of patients who completed the survey, with a mean (SD) time to alopecia of 75 (54) days after initiation of voriconazole. Notable regrowth was reported in 69% (79/114) of patients who discontinued voriconazole for at least 3 months. A subgroup of 32 patients were changed to itraconazole or posaconazole, and hair loss stopped in 84% (27/32) with regrowth noted in 69% (22/32) of patients.³² Voriconazole and fluconazole share structural similarity not present with other triazoles.^33,34 Because voriconazole-associated alopecia was reversed in the majority of patients who switched to itraconazole or posaconazole, the authors hypothesized that structural similarity of fluconazole and voriconazole may underly the greater risk for TE that is not a class effect of azole medications.³¹

Psychotropic Medications

Various psychotropic medications have been associated with hair loss. Valproic acid (or sodium valproate) is an anticonvulsant and mood-stabilizing agent used for the treatment of seizures, bipolar disorder (BD), migraines, and neuropathic pain.^35,36 Divalproex sodium (or divalproex) is an enteric-coated formulation of sodium valproate and valproic acid with similar indications. Valproate is a notorious culprit of medication-induced hair loss, with alopecia listed among the most common adverse reactions (reported >5%) on its structure product labeling document.³⁷ A systemic review and meta-analysis by Wang et al³⁸ estimated the overall incidence of valproate-related alopecia to be 11% (95% CI, 0.08-0.13). Although this meta-analysis did not find an association between incidence of alopecia and dose or duration of valproate therapy,³⁸ a separate review suggested that valproate-induced alopecia is dose dependent and can be managed with dose reduction.³⁹ A 12-month, randomized, double-blind study of treatment of BD with divalproex (valproate derivative), lithium, or placebo (2:1:1 ratio) showed a significantly higher frequency of alopecia in the divalproex group compared with placebo (16% [30/187] vs 6% [6/94]; P=.03).⁴⁰ Valproate-related hair loss is characteristically diffuse and nonscarring, often noted 3 to 6 months following initiation of valproate.^41,42 The proposed mechanism of valproate-induced alopecia includes chelation of zinc and selenium,⁴³ and a reduction in serum biotinidase activity, thereby decreasing the availability of these essential micronutrients required for hair growth.⁴¹ Studies examining the effects of valproate administration and serum biotinidase activity in patients have yielded conflicting results.^44-46 In a study of children with seizures including 57 patients treated with valproic acid, 17 treated with carbamazepine, and 75 age- and sex-matched healthy controls, the authors found no significant differences in serum biotinidase enzyme activity across the 3 groups.⁴⁴ In contrast, a study of 75 children with seizures on valproic acid therapy stratified by dose (mean [SD])—group A: 28.7 [8.5] mg/kg/d; group B: 41.6 [4.9] mg/kg/d; group C: 64.5 [5.8] mg/kg/d—found that patients receiving higher doses (groups B and C) had significantly reduced serum biotinidase activity (1.22 [1.11] and 0.97 [0.07] mmol/min/L, respectively) compared with 50 healthy pediatric controls (5.20 [0.90] mmol/min/L; P<.001). The same study found biotin supplementation at 10 mg/d for 20 days led to resolution of alopecia in 22% (2/9) of patients with alopecia on valproic acid therapy.⁴⁵ Despite hypothesized effects of valproate on micronutrients, the role of mineral supplementation in treating valproate-associated hair loss remains unclear. There is evidence to suggest that valproic acid–associated alterations in serum biotinidase activity may be transient. In a study of 32 pediatric patients receiving valproic acid for the treatment of epilepsy, serum biotinidase activity was significantly lower after 3 months of valproic acid therapy compared with pretreatment levels (P<.05); at 6 months, the serum biotinidase activity was increased compared with 3 months (P<.05) and not significantly different from pretreatment levels (P>.05).⁴⁶ Hair regrowth has been observed following discontinuation or dose reduction of valproate therapy in some cases.^39,47

Lithium carbonate (lithium) is used in the treatment of BD. Despite its efficacy and low cost, its potential for adverse effects, narrow therapeutic index, and subsequent need for routine monitoring are factors that limit its use.⁴⁸ Some reported dermatologic adverse reactions on its structure product labeling include xerosis, thinning of hair, alopecia, xerosis cutis, psoriasis onset/exacerbation, and generalized pruritus.⁴⁹ A systematic review and meta-analysis of 385 studies identified 24 publications reporting adverse effects of lithium on hair with no significantly increased risk of alopecia overall.⁵⁰ The analysis included 2 randomized controlled trials comparing the effects of lithium and placebo on hair loss in patients with BD. Hair loss was reported in 7% (7/94) of patients taking lithium and 6% (6/94) of the placebo group in the 12-month study⁴⁰ and in 3% (1/32) of the lithium group and 0% (0/28) of the divalproex group in the 20-month study.⁵¹ Despite anecdotal reports of alopecia associated with lithium, there is a lack of high-quality evidence to support this claim. Of note, hypothyroidism is a known complication of lithium use, and serum testing of thyroid function at 6-month intervals is recommended for patients on lithium treatment.⁵² Because thyroid abnormalities can cause alopecia distinct from TE, new-onset alopecia during lithium use should prompt serum testing of thyroid function. The development of hypothyroidism secondary to lithium is not a direct contraindication to its use⁵³; rather, treatment should be focused on correction with thyroid replacement therapy (eg, supplementation with thyroxine).⁵⁴

Commonly prescribed antidepressant medications include selective serotonin reuptake inhibitors (SSRIs) and bupropion. Selective serotonin reuptake inhibitors affect the neuronal serotonin transporter, increasing the concentration of serotonin in the synaptic cleft available for stimulation of postsynaptic serotonin receptors^55,56; bupropion is an antidepressant medication that inhibits norepinephrine and dopamine reuptake at the synaptic cleft.⁵⁷ Alopecia is an infrequent (1 in 100 to 1 in 1000 patients) adverse effect for several SSRIs.^58-62 A recent systematic review identified a total of 71 cases of alopecia associated with SSRI use including citalopram (n=11), escitalopram (n=7), fluoxetine (n=27), fluoxvamine (n=5), paroxetine (n=4), and sertraline (n=20), with a median time to onset of hair shedding of 8.6 weeks (range, 3 days to 5 years). Discontinuation of the suspected culprit SSRI led to improvement and/or resolution in 63% (51/81) episodes of alopecia, with a median time to improvement and/or resolution of 4 weeks.⁶³ A comparative retrospective cohort study using a large US health claims database from 2006 to 2014 included more than 1 million new and mutually exclusive patients taking fluoxetine, fluvoxamine, sertraline, citalopram, escitalopram, paroxetine, duloxetine, venlafaxine, desvenlafaxine, and bupropion.⁶⁴ Overall, 1% (1569/150,404) of patients treated with bupropion received 1 or more physician visits for alopecia. Patients on SSRIs generally had a lower risk for hair loss compared with patients using bupropion (citalopram: hazard ratio [HR], 0.80 [95% CI, 0.74-0.86]; escitalopram: HR, 0.79 [95% CI, 0.74-0.86]; fluoxetine: HR, 0.68 [95% CI, 0.63-0.74]; paroxetine: HR, 0.68 [95% CI, 0.62-0.74]; sertraline: HR, 0.74 [95% CI, 0.69-0.79]), with the exception of fluvoxamine (HR, 0.93 [95% CI, 0.64-1.37]). However, the type of alopecia, time to onset, and time to resolution were not reported, making it difficult to assess whether the reported hair loss was consistent with medication-induced TE. Additionally, the authors acknowledged that bupropion may have been prescribed for smoking cessation, which may carry a different risk profile for the development of alopecia.⁶⁴ Several other case reports have described alopecia following treatment with SSRIs, including sertraline,⁶⁵ fluvoxamine,⁶⁶ paroxetine,⁶⁷ fluoxetine,⁶⁸ and escitalopram.⁶⁹

Overall, it appears that the use of SSRIs portends relatively low risk for alopecia and medication-induced TE. Little is known regarding the molecular effects of SSRIs on hair growth and the pathomechanism of SSRI-induced TE. The potential benefits of discontinuing a suspected culprit medication should be carefully weighed against the risks of medication cessation, and consideration should be given to alternative medications in the same class that also may be associated with TE. In patients requiring antidepressant therapy with suspected medication-induced TE, consider transitioning to a different class of medication with lower risk of medication-induced alopecia; for example, discontinuing bupropion in favor of an SSRI.

Final Thoughts

Medication-induced alopecia is an undesired side effect of many commonly used drugs and drug classes, including retinoids, azole antifungals, and mood stabilizers. Although the precise pathomechanisms of medication-induced TE remain unclear, the recommended management often requires identification of the likely causative agent and its discontinuation, if possible. Suspicion for medication-induced TE should prompt a thorough history of recent changes to medications, risk factors for nutritional deficiencies, underlying illnesses, and recent surgical procedures. Underlying nutritional, electrolyte, and/or metabolic disturbances should be corrected. In part 2 of this series, we will discuss medication-induced alopecia associated with anticoagulant and antihypertensive medications.

Ferritin is an iron storage protein crucial to human iron homeostasis. Because serum ferritin levels are in dynamic equilibrium with the body’s iron stores, ferritin often is measured as a reflection of iron status; however, ferritin also is an acute-phase reactant whose levels may be nonspecifically elevated in a wide range of inflammatory conditions. The various processes that alter serum ferritin levels complicate the clinical interpretation of this laboratory value. In this article, we review the structure and function of ferritin and provide a guide for clinical use.

Overview of Iron

Iron is an essential element of key biologic functions including DNA synthesis and repair, oxygen transport, and oxidative phosphorylation. The body’s iron stores are mainly derived from internal iron recycling following red blood cell breakdown, while 5% to 10% is supplied by dietary intake.^1-3 Iron metabolism is of particular importance in cells of the reticuloendothelial system (eg, spleen, liver, bone marrow), where excess iron must be appropriately sequestered and from which iron can be mobilized.⁴ Sufficient iron stores are necessary for proper cellular function and survival, as iron is a necessary component of hemoglobin for oxygen delivery, iron-sulfur clusters in electron transport, and enzyme cofactors in other cellular processes.

Although labile pools of biologically active free iron exist in limited amounts within cells, excess free iron can generate free radicals that damage cellular proteins, lipids, and nucleic acids.^5-7 As such, most intracellular iron is captured within ferritin molecules. The excretion of iron is unregulated and occurs through loss in sweat, menstruation, hair shedding, skin desquamation, and enterocyte turnover.⁸ The lack of regulated excretion highlights the need for a tightly regulated system of uptake, transportation, storage, and sequestration to maintain iron homeostasis.

Overview of Ferritin Structure and Function

Ferritin is a key regulator of iron homeostasis that also serves as an important clinical indicator of body iron status. Ferritin mainly is found as an intracellular cytosolic iron storage and detoxification protein structured as a hollow 24-subunit polymer shell that can sequester up to 4500 atoms of iron within its core.^9,10 The 24-mer is composed of both ferritin L (FTL) and ferritin H (FTH) subunits, with dynamic regulation of the H:L ratio dependent on the context and tissue in which ferritin is found.⁶

Ferritin H possesses ferroxidase, which facilitates oxidation of ferrous (Fe²⁺) iron into ferric (Fe³⁺) iron, which can then be incorporated into the mineral core of the ferritin heteropolymer.¹¹ Ferritin L is more abundant in the spleen and liver, while FTH is found predominantly in the heart and kidneys where the increased ferroxidase activity may confer an increased ability to oxidize Fe²⁺ and limit oxidative stress.⁶

Regulation of Ferritin Synthesis and Secretion

Ferritin synthesis is regulated by intracellular nonheme iron levels, governed mainly by an iron response element (IRE) and iron response protein (IRP) translational repression system. Both FTH and FTL messenger RNA (mRNA) contain an IRE that is a regulatory stem-loop structure in the 5´ untranslated region. When the IRE is bound by IRP1 or IRP2, mRNA translation of ferritin subunits is suppressed.⁶ In low iron conditions, IRPs have greater affinity for IRE, and binding suppresses ferritin translation.¹² In high iron conditions, IRPs have a decreased affinity for IRE, and ferritin mRNA synthesis is increased.¹³ Additionally, inflammatory cytokines such as tumor necrosis factor α and IL-1α transcriptionally induce FTH synthesis, resulting in an increased population of H-rich ferritins.^11,14-16 A study using cultured human primary skin fibroblasts demonstrated UV radiation–induced increases in free intracellular iron content.^17,18 Pourzand et al¹⁸ suggested that UV-mediated damage of lysosomal membranes results in leakage of lysosomal proteases into the cytosol, contributing to degradation of intracellular ferritin and subsequent release of iron within skin fibroblasts. The increased intracellular iron downregulates IRPs and increases ferritin mRNA synthesis,¹⁸ consistent with prior findings of increased ferritin synthesis in skin that is induced by UV radiation.¹⁹

Molecular analysis of serum ferritin in iron-overloaded mice revealed that extracellular ferritin found in the serum is composed of a greater fraction of FTL and has lower iron content than intracellular ferritin. The low iron content of serum ferritin compared with intracellular ferritin and transferrin suggests that serum ferritin is not a major pathway of systemic iron transport.¹⁰ However, locally secreted ferritins may play a greater role in iron transport and release in selected tissues. Additionally, in vitro studies of cell cultures from humans and mice have demonstrated the ability of macrophages to secrete ferritin, suggesting that macrophages are an important cellular source of serum ferritin.^10,20 As such, serum ferritin generally may reflect body iron status but more specifically reflects macrophage iron status.¹⁰ Although the exact pathways of ferritin release are unknown, it is hypothesized that ferritin secretion occurs through cytosolic autophagy followed by secretion of proteins from the lysosomal compartment.^10,18,21

Low Ferritin and Iron Deficiency—Although bone marrow biopsy with iron staining remains the gold standard for diagnosis of iron deficiency, serum ferritin is a much more accessible and less invasive tool for evaluation of iron status. A serum ferritin level below 12 μg/L is highly specific for iron depletion,²² with a higher cutoff recommended in clinical practice to improve diagnostic sensitivity.^23,24 Conditions independent of iron deficiency that may reduce serum ferritin include hypothyroidism and ascorbate deficiency, though neither condition has been reported to interfere with appropriate diagnosis of iron deficiency.²⁵ Guyatt et al²⁶ conducted a systematic review of laboratory tests used in the diagnosis of iron deficiency anemia and identified 55 studies suitable for inclusion. Based on an area under the receiver operating characteristic curve (AUROC) of 0.95, serum ferritin values were superior to transferrin saturation (AUROC, 0.74), red cell protoporphyrin (AUROC, 0.77), red cell volume distribution width (AUROC, 0.62), and mean cell volume (AUROC, 0.76) for diagnosis of IDA, verified by histologic examination of aspirated bone marrow.²⁶ The likelihood ratio of iron deficiency begins to decrease for serum ferritin values of 40 μg/L or greater. For patients with inflammatory conditions—patients with concomitant chronic renal failure, inflammatory disease, infection, rheumatoid arthritis, liver disease, inflammatory bowel disease, and malignancy—the likelihood of iron deficiency begins to decrease at serum ferritin levels of 70 μg/L or greater.²⁶ Similarly, the World Health Organization recommends that in adults with infection or inflammation, serum ferritin levels lower than 70 μg/L may be used to indicate iron deficiency.²⁴ A serum ferritin level of 41 μg/L or lower was found to have a sensitivity and specificity of 98% for discriminating between iron-deficiency anemia and anemia of chronic disease (diagnosed based on bone marrow biopsy with iron staining), with an AUROC of 0.98.²⁷ As such, we recommend using a serum ferritin level of 40 μg/L or lower in patients who are otherwise healthy as an indicator of iron deficiency.

The threshold for iron supplementation may vary based on age, sex, and race. In women, ferritin levels increase during menopause and peak after menopause; ferritin levels are higher in men than in women.^28-30 A multisite longitudinal cohort study of 70 women in the United States found that the mean (SD) ferritin valuewas 69.5 (81.7) μg/L premenopause and 128.8 (125.7) μg/L postmenopause (P<.01).³¹ A separate longitudinal survey study of 8564 patients in China found that the mean (SE) ferritin value was 201.55 (3.60) μg/L for men and 80.46 (1.64) μg/L for women (P<.0001).³² Analysis of serum ferritin levels of 3554 male patients from the third National Health and Nutrition Examination Survey demonstrated that patients who self-reported as non-Hispanic Black (n=1616) had significantly higher serum ferritin levels than non-Hispanic White patients (n=1938)(serum ferritin difference of 37.1 μg/L)(P<.0001).³³ The British Society for Haematology guidelines recommend that the threshold of serum ferritin for diagnosing iron deficiency should take into account age-, sex-, and race-based differences.³⁴ Ferritin and Hair—Cutaneous manifestations of iron deficiency include koilonychia, glossitis, pruritus, angular cheilitis, and telogen effluvium (TE).¹ A case-control study of 30 females aged 15 to 45 years demonstrated that the mean (SD) ferritin level was significantly lower in patients with TE than those with no hair loss (16.3 [12.6] ng/mL vs 60.3 [50.1] ng/mL; P<.0001). Using a threshold of 30 μg/L or lower, the investigators found that the odds ratio for TE was 21.0 (95% CI, 4.2-105.0) in patients with low serum ferritin.³⁵

Another retrospective review of 54 patients with diffuse hair loss and 55 controls compared serum vitamin B₁₂, folate, thyroid-stimulating hormone, zinc, ferritin, and 25-hydroxy vitamin D levels between the 2 groups.³⁶ Exclusion criteria were clinical diagnoses of female pattern hair loss (androgenetic alopecia), pregnancy, menopause, metabolic and endocrine disorders, hormone replacement therapy, chemotherapy, immunosuppressive therapy, vitamin and mineral supplementation, scarring alopecia, eating disorders, and restrictive diets. Compared with controls, patients with diffuse nonscarring hair loss were found to have significantly lower ferritin (mean [SD], 14.72 [10.70] ng/mL vs 25.30 [14.41] ng/mL; P<.001) and 25-hydroxy vitamin D levels (mean [SD], 14.03 [8.09] ng/mL vs 17.01 [8.59] ng/mL; P=.01).³⁶

In contrast, a separate case-control study of 381 cases and 76 controls found no increase in the rate of iron deficiency—defined as ferritin ≤15 μg/L or ≤40 μg/L—among women with female pattern hair loss or chronic TE vs controls.³⁷ Taken together, these studies suggest that iron status may play a role in TE, a process that may result from nutritional deficiency, trauma, or physical or psychological stress³⁸; however, there is insufficient evidence to suggest that low iron status impacts androgenetic alopecia, in which its multifactorial pathogenesis implicates genetic and hormonal factors.³⁹ More research is needed to clarify the potential associations between iron deficiency and types of hair loss. Additionally, it is unclear whether iron supplementation improves hair growth parameters such as density and caliber.⁴⁰

Low serum ferritin (<40 μg/L) with concurrent symptoms of iron deficiency, including fatigue, pallor, dyspnea on exertion, or hair loss, should prompt treatment with supplemental iron.^41-43 Generally, ferrous (Fe2⁺) salts are preferred to ferric (Fe3⁺) salts, as the former is more readily absorbed through the duodenal mucosa⁴⁴ and is the more common formulation in commercially available supplements in the United States.⁴⁵ Oral supplementation with ferrous sulfate 325 mg (65 mg elemental iron) tablets is the first-line therapy for iron deficiency anemia.¹ Alternatively, ferrous gluconate 324 mg (38 mg elemental iron) over-the-counter and its liquid form has demonstrated superior absorption compared to ferrous sulfate tablets in a clinical study with peritoneal dialysis patients.^1,46 One study suggested that oral iron 40 to 80 mg should be taken every other day to increase absorption.⁴⁷ Due to improved bioavailability, intravenous iron may be utilized in patients with malabsorption, renal failure, or intolerance to oral iron (including those with gastric ulcers or active inflammatory bowel disease), with the formulation chosen based on underlying comorbidities and potential risks.^1,48 The theoretical risk for potentiating bacterial growth by increasing the amount of unbound iron in the blood raises concerns of iron supplementation in patients with infection or sepsis. Although far from definitive, existing data suggest that risk for infection is greater with intravenous iron supplementation and should be carefully considered prior to use.^49,50Elevated Ferritin—Elevated ferritin may be difficult to interpret given the multitude of conditions that can cause it.^23,51,52 Elevated serum ferritin can be broadly characterized by increased synthesis due to iron overload, increased synthesis due to inflammation, or increased ferritin release from cellular damage.³⁴ Further complicating interpretation is the potential diurnal fluctuations in serum iron levels dependent on dietary intake and timing of laboratory evaluation, choice of assay, differences in reference standards, and variations in calibration procedures that can lead to analytic variability in the measurement of ferritin.^23,53,54

Among healthy patients, serum ferritin is directly proportional to iron status.^9,51 A study utilizing weekly phlebotomy of 22 healthy participants to measure serum ferritin and calculate mobilizable storage iron found a strong positive correlation between the 2 variables (r=0.83, P<.001), with each 1-μg/L increase of serum ferritin corresponding to approximately an 8-mg increase of storage iron; the initial serum ferritin values ranged from 2 to 83 μg/L in females and 36 to 224 μg/L in males.⁵⁵ The correlation of ferritin with iron status also was supported by the significant correlation between the number of transfusions received in patients with transfusion-related iron overload and serum ferritin levels (r=0.89, P<.001), with an average increase of 60 μg/L per transfusion.⁵¹

Clinical guidelines on the interpretation of serum ferritin levels by Cullis et al³⁴ recommend a normal upper limit of 200 μg/L for healthy females and 300 μg/L for healthy males. Outside of clinical syndromes associated with iron overload, Lee and Means⁵⁶ found that serum ferritin of 1000 μg/L or higher was a nonspecific marker of disease, including infection and/or neoplastic disorders. We have adapted these guidelines to propose a workflow for evaluation of serum ferritin levels (Figure). In patients with inflammatory conditions or those affected by metabolic syndrome, elevated serum ferritin does not correlate with body iron status.^57,58 It is believed that inflammatory cytokines, including tumor necrosis factor α and IL-1α, can upregulate ferritin synthesis independent of cellular iron stores.^15,16 Several studies have examined the relationship between insulin resistance and/or metabolic syndrome with serum ferritin levels.^31,32 Han et al³² found that elevated serum ferritin was significantly associated with higher risk for metabolic syndrome in men (P<.0001) but not in women.

Although cutaneous manifestations of iron overload can be seen as skin hyperpigmentation due to increased iron deposits and increased melanin production,²² the effects of elevated ferritin on the skin and hair are not well known. Iron overload is a known trigger of porphyria cutanea tarda (PCT),⁵⁹ a condition in which reduced or absent enzymatic activity of uroporphyrinogen decarboxylase (UROD) leads to build up of toxic porphyrins in various organs.⁶⁰ In the skin, PCT manifests as a blistering photosensitive eruption that may resolve as dyspigmentation, scarring, and milia.⁶¹ Phlebotomy is first-line therapy in PCT to reduce serum iron and subsequent formation of UROD inhibitors, with guidelines suggesting discontinuation of phlebotomy when serum ferritin levels reach 20 ng/mL or lower.⁶⁰ Hyperferritinemia (serum ferritin >500 μg/L) is a common finding in several inflammatory disorders often accompanied by clinically apparent cutaneous symptoms such as adult-onset Still disease,⁶² hemophagocytic lymphohistiocytosis,^63,64 and anti-melanoma differentiation-associated gene 5 dermatomyositis.⁶⁵ Among these conditions, serum ferritin levels have been reported to correlate with disease activity, raising the question of whether ferritin is a bystander or a driver of the underlying pathology.^62,66,67 However, rapid decline of serum ferritin levels with treatment and control of inflammatory cytokines suggest that ferritin is unlikely to contribute to pathology.^62,67

Final Thoughts

Many clinical studies have examined the association between hair health and body iron status, the collective findings of which suggest that iron deficiency may be associated with TE. Among commonly measured serum iron parameters, low ferritin is a highly specific and sensitive marker for diagnosing iron deficiency. Serum ferritin may be a clinically useful tool for ruling out underlying iron deficiency in patients presenting with hair loss. Despite advances in our understanding of the molecular mechanisms of ferritin synthesis and regulation, whether ferritin itself contributes to cutaneous pathology is poorly understood.^35,36,68-74 For patients who are otherwise healthy with low suspicion for inflammatory disorders, chronic systemic illnesses, or malignancy, serum ferritin can be used as an indicator of body iron status. The workup for slightly elevated serum ferritin should be interpreted in the context of other laboratory findings and should be reassessed over time. Serum ferritin levels above 1000 μg/L warrant further investigation into causes such as iron overload conditions and underlying inflammatory conditions or malignancy.

Ferritin is an iron storage protein crucial to human iron homeostasis. Because serum ferritin levels are in dynamic equilibrium with the body’s iron stores, ferritin often is measured as a reflection of iron status; however, ferritin also is an acute-phase reactant whose levels may be nonspecifically elevated in a wide range of inflammatory conditions. The various processes that alter serum ferritin levels complicate the clinical interpretation of this laboratory value. In this article, we review the structure and function of ferritin and provide a guide for clinical use.

Overview of Iron

Iron is an essential element of key biologic functions including DNA synthesis and repair, oxygen transport, and oxidative phosphorylation. The body’s iron stores are mainly derived from internal iron recycling following red blood cell breakdown, while 5% to 10% is supplied by dietary intake.^1-3 Iron metabolism is of particular importance in cells of the reticuloendothelial system (eg, spleen, liver, bone marrow), where excess iron must be appropriately sequestered and from which iron can be mobilized.⁴ Sufficient iron stores are necessary for proper cellular function and survival, as iron is a necessary component of hemoglobin for oxygen delivery, iron-sulfur clusters in electron transport, and enzyme cofactors in other cellular processes.

Although labile pools of biologically active free iron exist in limited amounts within cells, excess free iron can generate free radicals that damage cellular proteins, lipids, and nucleic acids.^5-7 As such, most intracellular iron is captured within ferritin molecules. The excretion of iron is unregulated and occurs through loss in sweat, menstruation, hair shedding, skin desquamation, and enterocyte turnover.⁸ The lack of regulated excretion highlights the need for a tightly regulated system of uptake, transportation, storage, and sequestration to maintain iron homeostasis.

Overview of Ferritin Structure and Function

Ferritin is a key regulator of iron homeostasis that also serves as an important clinical indicator of body iron status. Ferritin mainly is found as an intracellular cytosolic iron storage and detoxification protein structured as a hollow 24-subunit polymer shell that can sequester up to 4500 atoms of iron within its core.^9,10 The 24-mer is composed of both ferritin L (FTL) and ferritin H (FTH) subunits, with dynamic regulation of the H:L ratio dependent on the context and tissue in which ferritin is found.⁶

Ferritin H possesses ferroxidase, which facilitates oxidation of ferrous (Fe²⁺) iron into ferric (Fe³⁺) iron, which can then be incorporated into the mineral core of the ferritin heteropolymer.¹¹ Ferritin L is more abundant in the spleen and liver, while FTH is found predominantly in the heart and kidneys where the increased ferroxidase activity may confer an increased ability to oxidize Fe²⁺ and limit oxidative stress.⁶

Regulation of Ferritin Synthesis and Secretion

Ferritin synthesis is regulated by intracellular nonheme iron levels, governed mainly by an iron response element (IRE) and iron response protein (IRP) translational repression system. Both FTH and FTL messenger RNA (mRNA) contain an IRE that is a regulatory stem-loop structure in the 5´ untranslated region. When the IRE is bound by IRP1 or IRP2, mRNA translation of ferritin subunits is suppressed.⁶ In low iron conditions, IRPs have greater affinity for IRE, and binding suppresses ferritin translation.¹² In high iron conditions, IRPs have a decreased affinity for IRE, and ferritin mRNA synthesis is increased.¹³ Additionally, inflammatory cytokines such as tumor necrosis factor α and IL-1α transcriptionally induce FTH synthesis, resulting in an increased population of H-rich ferritins.^11,14-16 A study using cultured human primary skin fibroblasts demonstrated UV radiation–induced increases in free intracellular iron content.^17,18 Pourzand et al¹⁸ suggested that UV-mediated damage of lysosomal membranes results in leakage of lysosomal proteases into the cytosol, contributing to degradation of intracellular ferritin and subsequent release of iron within skin fibroblasts. The increased intracellular iron downregulates IRPs and increases ferritin mRNA synthesis,¹⁸ consistent with prior findings of increased ferritin synthesis in skin that is induced by UV radiation.¹⁹

Molecular analysis of serum ferritin in iron-overloaded mice revealed that extracellular ferritin found in the serum is composed of a greater fraction of FTL and has lower iron content than intracellular ferritin. The low iron content of serum ferritin compared with intracellular ferritin and transferrin suggests that serum ferritin is not a major pathway of systemic iron transport.¹⁰ However, locally secreted ferritins may play a greater role in iron transport and release in selected tissues. Additionally, in vitro studies of cell cultures from humans and mice have demonstrated the ability of macrophages to secrete ferritin, suggesting that macrophages are an important cellular source of serum ferritin.^10,20 As such, serum ferritin generally may reflect body iron status but more specifically reflects macrophage iron status.¹⁰ Although the exact pathways of ferritin release are unknown, it is hypothesized that ferritin secretion occurs through cytosolic autophagy followed by secretion of proteins from the lysosomal compartment.^10,18,21

Low Ferritin and Iron Deficiency—Although bone marrow biopsy with iron staining remains the gold standard for diagnosis of iron deficiency, serum ferritin is a much more accessible and less invasive tool for evaluation of iron status. A serum ferritin level below 12 μg/L is highly specific for iron depletion,²² with a higher cutoff recommended in clinical practice to improve diagnostic sensitivity.^23,24 Conditions independent of iron deficiency that may reduce serum ferritin include hypothyroidism and ascorbate deficiency, though neither condition has been reported to interfere with appropriate diagnosis of iron deficiency.²⁵ Guyatt et al²⁶ conducted a systematic review of laboratory tests used in the diagnosis of iron deficiency anemia and identified 55 studies suitable for inclusion. Based on an area under the receiver operating characteristic curve (AUROC) of 0.95, serum ferritin values were superior to transferrin saturation (AUROC, 0.74), red cell protoporphyrin (AUROC, 0.77), red cell volume distribution width (AUROC, 0.62), and mean cell volume (AUROC, 0.76) for diagnosis of IDA, verified by histologic examination of aspirated bone marrow.²⁶ The likelihood ratio of iron deficiency begins to decrease for serum ferritin values of 40 μg/L or greater. For patients with inflammatory conditions—patients with concomitant chronic renal failure, inflammatory disease, infection, rheumatoid arthritis, liver disease, inflammatory bowel disease, and malignancy—the likelihood of iron deficiency begins to decrease at serum ferritin levels of 70 μg/L or greater.²⁶ Similarly, the World Health Organization recommends that in adults with infection or inflammation, serum ferritin levels lower than 70 μg/L may be used to indicate iron deficiency.²⁴ A serum ferritin level of 41 μg/L or lower was found to have a sensitivity and specificity of 98% for discriminating between iron-deficiency anemia and anemia of chronic disease (diagnosed based on bone marrow biopsy with iron staining), with an AUROC of 0.98.²⁷ As such, we recommend using a serum ferritin level of 40 μg/L or lower in patients who are otherwise healthy as an indicator of iron deficiency.

The threshold for iron supplementation may vary based on age, sex, and race. In women, ferritin levels increase during menopause and peak after menopause; ferritin levels are higher in men than in women.^28-30 A multisite longitudinal cohort study of 70 women in the United States found that the mean (SD) ferritin valuewas 69.5 (81.7) μg/L premenopause and 128.8 (125.7) μg/L postmenopause (P<.01).³¹ A separate longitudinal survey study of 8564 patients in China found that the mean (SE) ferritin value was 201.55 (3.60) μg/L for men and 80.46 (1.64) μg/L for women (P<.0001).³² Analysis of serum ferritin levels of 3554 male patients from the third National Health and Nutrition Examination Survey demonstrated that patients who self-reported as non-Hispanic Black (n=1616) had significantly higher serum ferritin levels than non-Hispanic White patients (n=1938)(serum ferritin difference of 37.1 μg/L)(P<.0001).³³ The British Society for Haematology guidelines recommend that the threshold of serum ferritin for diagnosing iron deficiency should take into account age-, sex-, and race-based differences.³⁴ Ferritin and Hair—Cutaneous manifestations of iron deficiency include koilonychia, glossitis, pruritus, angular cheilitis, and telogen effluvium (TE).¹ A case-control study of 30 females aged 15 to 45 years demonstrated that the mean (SD) ferritin level was significantly lower in patients with TE than those with no hair loss (16.3 [12.6] ng/mL vs 60.3 [50.1] ng/mL; P<.0001). Using a threshold of 30 μg/L or lower, the investigators found that the odds ratio for TE was 21.0 (95% CI, 4.2-105.0) in patients with low serum ferritin.³⁵

Another retrospective review of 54 patients with diffuse hair loss and 55 controls compared serum vitamin B₁₂, folate, thyroid-stimulating hormone, zinc, ferritin, and 25-hydroxy vitamin D levels between the 2 groups.³⁶ Exclusion criteria were clinical diagnoses of female pattern hair loss (androgenetic alopecia), pregnancy, menopause, metabolic and endocrine disorders, hormone replacement therapy, chemotherapy, immunosuppressive therapy, vitamin and mineral supplementation, scarring alopecia, eating disorders, and restrictive diets. Compared with controls, patients with diffuse nonscarring hair loss were found to have significantly lower ferritin (mean [SD], 14.72 [10.70] ng/mL vs 25.30 [14.41] ng/mL; P<.001) and 25-hydroxy vitamin D levels (mean [SD], 14.03 [8.09] ng/mL vs 17.01 [8.59] ng/mL; P=.01).³⁶

In contrast, a separate case-control study of 381 cases and 76 controls found no increase in the rate of iron deficiency—defined as ferritin ≤15 μg/L or ≤40 μg/L—among women with female pattern hair loss or chronic TE vs controls.³⁷ Taken together, these studies suggest that iron status may play a role in TE, a process that may result from nutritional deficiency, trauma, or physical or psychological stress³⁸; however, there is insufficient evidence to suggest that low iron status impacts androgenetic alopecia, in which its multifactorial pathogenesis implicates genetic and hormonal factors.³⁹ More research is needed to clarify the potential associations between iron deficiency and types of hair loss. Additionally, it is unclear whether iron supplementation improves hair growth parameters such as density and caliber.⁴⁰

Low serum ferritin (<40 μg/L) with concurrent symptoms of iron deficiency, including fatigue, pallor, dyspnea on exertion, or hair loss, should prompt treatment with supplemental iron.^41-43 Generally, ferrous (Fe2⁺) salts are preferred to ferric (Fe3⁺) salts, as the former is more readily absorbed through the duodenal mucosa⁴⁴ and is the more common formulation in commercially available supplements in the United States.⁴⁵ Oral supplementation with ferrous sulfate 325 mg (65 mg elemental iron) tablets is the first-line therapy for iron deficiency anemia.¹ Alternatively, ferrous gluconate 324 mg (38 mg elemental iron) over-the-counter and its liquid form has demonstrated superior absorption compared to ferrous sulfate tablets in a clinical study with peritoneal dialysis patients.^1,46 One study suggested that oral iron 40 to 80 mg should be taken every other day to increase absorption.⁴⁷ Due to improved bioavailability, intravenous iron may be utilized in patients with malabsorption, renal failure, or intolerance to oral iron (including those with gastric ulcers or active inflammatory bowel disease), with the formulation chosen based on underlying comorbidities and potential risks.^1,48 The theoretical risk for potentiating bacterial growth by increasing the amount of unbound iron in the blood raises concerns of iron supplementation in patients with infection or sepsis. Although far from definitive, existing data suggest that risk for infection is greater with intravenous iron supplementation and should be carefully considered prior to use.^49,50Elevated Ferritin—Elevated ferritin may be difficult to interpret given the multitude of conditions that can cause it.^23,51,52 Elevated serum ferritin can be broadly characterized by increased synthesis due to iron overload, increased synthesis due to inflammation, or increased ferritin release from cellular damage.³⁴ Further complicating interpretation is the potential diurnal fluctuations in serum iron levels dependent on dietary intake and timing of laboratory evaluation, choice of assay, differences in reference standards, and variations in calibration procedures that can lead to analytic variability in the measurement of ferritin.^23,53,54

Among healthy patients, serum ferritin is directly proportional to iron status.^9,51 A study utilizing weekly phlebotomy of 22 healthy participants to measure serum ferritin and calculate mobilizable storage iron found a strong positive correlation between the 2 variables (r=0.83, P<.001), with each 1-μg/L increase of serum ferritin corresponding to approximately an 8-mg increase of storage iron; the initial serum ferritin values ranged from 2 to 83 μg/L in females and 36 to 224 μg/L in males.⁵⁵ The correlation of ferritin with iron status also was supported by the significant correlation between the number of transfusions received in patients with transfusion-related iron overload and serum ferritin levels (r=0.89, P<.001), with an average increase of 60 μg/L per transfusion.⁵¹

Clinical guidelines on the interpretation of serum ferritin levels by Cullis et al³⁴ recommend a normal upper limit of 200 μg/L for healthy females and 300 μg/L for healthy males. Outside of clinical syndromes associated with iron overload, Lee and Means⁵⁶ found that serum ferritin of 1000 μg/L or higher was a nonspecific marker of disease, including infection and/or neoplastic disorders. We have adapted these guidelines to propose a workflow for evaluation of serum ferritin levels (Figure). In patients with inflammatory conditions or those affected by metabolic syndrome, elevated serum ferritin does not correlate with body iron status.^57,58 It is believed that inflammatory cytokines, including tumor necrosis factor α and IL-1α, can upregulate ferritin synthesis independent of cellular iron stores.^15,16 Several studies have examined the relationship between insulin resistance and/or metabolic syndrome with serum ferritin levels.^31,32 Han et al³² found that elevated serum ferritin was significantly associated with higher risk for metabolic syndrome in men (P<.0001) but not in women.

Although cutaneous manifestations of iron overload can be seen as skin hyperpigmentation due to increased iron deposits and increased melanin production,²² the effects of elevated ferritin on the skin and hair are not well known. Iron overload is a known trigger of porphyria cutanea tarda (PCT),⁵⁹ a condition in which reduced or absent enzymatic activity of uroporphyrinogen decarboxylase (UROD) leads to build up of toxic porphyrins in various organs.⁶⁰ In the skin, PCT manifests as a blistering photosensitive eruption that may resolve as dyspigmentation, scarring, and milia.⁶¹ Phlebotomy is first-line therapy in PCT to reduce serum iron and subsequent formation of UROD inhibitors, with guidelines suggesting discontinuation of phlebotomy when serum ferritin levels reach 20 ng/mL or lower.⁶⁰ Hyperferritinemia (serum ferritin >500 μg/L) is a common finding in several inflammatory disorders often accompanied by clinically apparent cutaneous symptoms such as adult-onset Still disease,⁶² hemophagocytic lymphohistiocytosis,^63,64 and anti-melanoma differentiation-associated gene 5 dermatomyositis.⁶⁵ Among these conditions, serum ferritin levels have been reported to correlate with disease activity, raising the question of whether ferritin is a bystander or a driver of the underlying pathology.^62,66,67 However, rapid decline of serum ferritin levels with treatment and control of inflammatory cytokines suggest that ferritin is unlikely to contribute to pathology.^62,67

Final Thoughts

Many clinical studies have examined the association between hair health and body iron status, the collective findings of which suggest that iron deficiency may be associated with TE. Among commonly measured serum iron parameters, low ferritin is a highly specific and sensitive marker for diagnosing iron deficiency. Serum ferritin may be a clinically useful tool for ruling out underlying iron deficiency in patients presenting with hair loss. Despite advances in our understanding of the molecular mechanisms of ferritin synthesis and regulation, whether ferritin itself contributes to cutaneous pathology is poorly understood.^35,36,68-74 For patients who are otherwise healthy with low suspicion for inflammatory disorders, chronic systemic illnesses, or malignancy, serum ferritin can be used as an indicator of body iron status. The workup for slightly elevated serum ferritin should be interpreted in the context of other laboratory findings and should be reassessed over time. Serum ferritin levels above 1000 μg/L warrant further investigation into causes such as iron overload conditions and underlying inflammatory conditions or malignancy.

Ferritin is an iron storage protein crucial to human iron homeostasis. Because serum ferritin levels are in dynamic equilibrium with the body’s iron stores, ferritin often is measured as a reflection of iron status; however, ferritin also is an acute-phase reactant whose levels may be nonspecifically elevated in a wide range of inflammatory conditions. The various processes that alter serum ferritin levels complicate the clinical interpretation of this laboratory value. In this article, we review the structure and function of ferritin and provide a guide for clinical use.

Overview of Iron

Iron is an essential element of key biologic functions including DNA synthesis and repair, oxygen transport, and oxidative phosphorylation. The body’s iron stores are mainly derived from internal iron recycling following red blood cell breakdown, while 5% to 10% is supplied by dietary intake.^1-3 Iron metabolism is of particular importance in cells of the reticuloendothelial system (eg, spleen, liver, bone marrow), where excess iron must be appropriately sequestered and from which iron can be mobilized.⁴ Sufficient iron stores are necessary for proper cellular function and survival, as iron is a necessary component of hemoglobin for oxygen delivery, iron-sulfur clusters in electron transport, and enzyme cofactors in other cellular processes.

Although labile pools of biologically active free iron exist in limited amounts within cells, excess free iron can generate free radicals that damage cellular proteins, lipids, and nucleic acids.^5-7 As such, most intracellular iron is captured within ferritin molecules. The excretion of iron is unregulated and occurs through loss in sweat, menstruation, hair shedding, skin desquamation, and enterocyte turnover.⁸ The lack of regulated excretion highlights the need for a tightly regulated system of uptake, transportation, storage, and sequestration to maintain iron homeostasis.

Overview of Ferritin Structure and Function

Ferritin is a key regulator of iron homeostasis that also serves as an important clinical indicator of body iron status. Ferritin mainly is found as an intracellular cytosolic iron storage and detoxification protein structured as a hollow 24-subunit polymer shell that can sequester up to 4500 atoms of iron within its core.^9,10 The 24-mer is composed of both ferritin L (FTL) and ferritin H (FTH) subunits, with dynamic regulation of the H:L ratio dependent on the context and tissue in which ferritin is found.⁶

Ferritin H possesses ferroxidase, which facilitates oxidation of ferrous (Fe²⁺) iron into ferric (Fe³⁺) iron, which can then be incorporated into the mineral core of the ferritin heteropolymer.¹¹ Ferritin L is more abundant in the spleen and liver, while FTH is found predominantly in the heart and kidneys where the increased ferroxidase activity may confer an increased ability to oxidize Fe²⁺ and limit oxidative stress.⁶

Regulation of Ferritin Synthesis and Secretion

Ferritin synthesis is regulated by intracellular nonheme iron levels, governed mainly by an iron response element (IRE) and iron response protein (IRP) translational repression system. Both FTH and FTL messenger RNA (mRNA) contain an IRE that is a regulatory stem-loop structure in the 5´ untranslated region. When the IRE is bound by IRP1 or IRP2, mRNA translation of ferritin subunits is suppressed.⁶ In low iron conditions, IRPs have greater affinity for IRE, and binding suppresses ferritin translation.¹² In high iron conditions, IRPs have a decreased affinity for IRE, and ferritin mRNA synthesis is increased.¹³ Additionally, inflammatory cytokines such as tumor necrosis factor α and IL-1α transcriptionally induce FTH synthesis, resulting in an increased population of H-rich ferritins.^11,14-16 A study using cultured human primary skin fibroblasts demonstrated UV radiation–induced increases in free intracellular iron content.^17,18 Pourzand et al¹⁸ suggested that UV-mediated damage of lysosomal membranes results in leakage of lysosomal proteases into the cytosol, contributing to degradation of intracellular ferritin and subsequent release of iron within skin fibroblasts. The increased intracellular iron downregulates IRPs and increases ferritin mRNA synthesis,¹⁸ consistent with prior findings of increased ferritin synthesis in skin that is induced by UV radiation.¹⁹

Molecular analysis of serum ferritin in iron-overloaded mice revealed that extracellular ferritin found in the serum is composed of a greater fraction of FTL and has lower iron content than intracellular ferritin. The low iron content of serum ferritin compared with intracellular ferritin and transferrin suggests that serum ferritin is not a major pathway of systemic iron transport.¹⁰ However, locally secreted ferritins may play a greater role in iron transport and release in selected tissues. Additionally, in vitro studies of cell cultures from humans and mice have demonstrated the ability of macrophages to secrete ferritin, suggesting that macrophages are an important cellular source of serum ferritin.^10,20 As such, serum ferritin generally may reflect body iron status but more specifically reflects macrophage iron status.¹⁰ Although the exact pathways of ferritin release are unknown, it is hypothesized that ferritin secretion occurs through cytosolic autophagy followed by secretion of proteins from the lysosomal compartment.^10,18,21

Low Ferritin and Iron Deficiency—Although bone marrow biopsy with iron staining remains the gold standard for diagnosis of iron deficiency, serum ferritin is a much more accessible and less invasive tool for evaluation of iron status. A serum ferritin level below 12 μg/L is highly specific for iron depletion,²² with a higher cutoff recommended in clinical practice to improve diagnostic sensitivity.^23,24 Conditions independent of iron deficiency that may reduce serum ferritin include hypothyroidism and ascorbate deficiency, though neither condition has been reported to interfere with appropriate diagnosis of iron deficiency.²⁵ Guyatt et al²⁶ conducted a systematic review of laboratory tests used in the diagnosis of iron deficiency anemia and identified 55 studies suitable for inclusion. Based on an area under the receiver operating characteristic curve (AUROC) of 0.95, serum ferritin values were superior to transferrin saturation (AUROC, 0.74), red cell protoporphyrin (AUROC, 0.77), red cell volume distribution width (AUROC, 0.62), and mean cell volume (AUROC, 0.76) for diagnosis of IDA, verified by histologic examination of aspirated bone marrow.²⁶ The likelihood ratio of iron deficiency begins to decrease for serum ferritin values of 40 μg/L or greater. For patients with inflammatory conditions—patients with concomitant chronic renal failure, inflammatory disease, infection, rheumatoid arthritis, liver disease, inflammatory bowel disease, and malignancy—the likelihood of iron deficiency begins to decrease at serum ferritin levels of 70 μg/L or greater.²⁶ Similarly, the World Health Organization recommends that in adults with infection or inflammation, serum ferritin levels lower than 70 μg/L may be used to indicate iron deficiency.²⁴ A serum ferritin level of 41 μg/L or lower was found to have a sensitivity and specificity of 98% for discriminating between iron-deficiency anemia and anemia of chronic disease (diagnosed based on bone marrow biopsy with iron staining), with an AUROC of 0.98.²⁷ As such, we recommend using a serum ferritin level of 40 μg/L or lower in patients who are otherwise healthy as an indicator of iron deficiency.

The threshold for iron supplementation may vary based on age, sex, and race. In women, ferritin levels increase during menopause and peak after menopause; ferritin levels are higher in men than in women.^28-30 A multisite longitudinal cohort study of 70 women in the United States found that the mean (SD) ferritin valuewas 69.5 (81.7) μg/L premenopause and 128.8 (125.7) μg/L postmenopause (P<.01).³¹ A separate longitudinal survey study of 8564 patients in China found that the mean (SE) ferritin value was 201.55 (3.60) μg/L for men and 80.46 (1.64) μg/L for women (P<.0001).³² Analysis of serum ferritin levels of 3554 male patients from the third National Health and Nutrition Examination Survey demonstrated that patients who self-reported as non-Hispanic Black (n=1616) had significantly higher serum ferritin levels than non-Hispanic White patients (n=1938)(serum ferritin difference of 37.1 μg/L)(P<.0001).³³ The British Society for Haematology guidelines recommend that the threshold of serum ferritin for diagnosing iron deficiency should take into account age-, sex-, and race-based differences.³⁴ Ferritin and Hair—Cutaneous manifestations of iron deficiency include koilonychia, glossitis, pruritus, angular cheilitis, and telogen effluvium (TE).¹ A case-control study of 30 females aged 15 to 45 years demonstrated that the mean (SD) ferritin level was significantly lower in patients with TE than those with no hair loss (16.3 [12.6] ng/mL vs 60.3 [50.1] ng/mL; P<.0001). Using a threshold of 30 μg/L or lower, the investigators found that the odds ratio for TE was 21.0 (95% CI, 4.2-105.0) in patients with low serum ferritin.³⁵

Another retrospective review of 54 patients with diffuse hair loss and 55 controls compared serum vitamin B₁₂, folate, thyroid-stimulating hormone, zinc, ferritin, and 25-hydroxy vitamin D levels between the 2 groups.³⁶ Exclusion criteria were clinical diagnoses of female pattern hair loss (androgenetic alopecia), pregnancy, menopause, metabolic and endocrine disorders, hormone replacement therapy, chemotherapy, immunosuppressive therapy, vitamin and mineral supplementation, scarring alopecia, eating disorders, and restrictive diets. Compared with controls, patients with diffuse nonscarring hair loss were found to have significantly lower ferritin (mean [SD], 14.72 [10.70] ng/mL vs 25.30 [14.41] ng/mL; P<.001) and 25-hydroxy vitamin D levels (mean [SD], 14.03 [8.09] ng/mL vs 17.01 [8.59] ng/mL; P=.01).³⁶

In contrast, a separate case-control study of 381 cases and 76 controls found no increase in the rate of iron deficiency—defined as ferritin ≤15 μg/L or ≤40 μg/L—among women with female pattern hair loss or chronic TE vs controls.³⁷ Taken together, these studies suggest that iron status may play a role in TE, a process that may result from nutritional deficiency, trauma, or physical or psychological stress³⁸; however, there is insufficient evidence to suggest that low iron status impacts androgenetic alopecia, in which its multifactorial pathogenesis implicates genetic and hormonal factors.³⁹ More research is needed to clarify the potential associations between iron deficiency and types of hair loss. Additionally, it is unclear whether iron supplementation improves hair growth parameters such as density and caliber.⁴⁰

Low serum ferritin (<40 μg/L) with concurrent symptoms of iron deficiency, including fatigue, pallor, dyspnea on exertion, or hair loss, should prompt treatment with supplemental iron.^41-43 Generally, ferrous (Fe2⁺) salts are preferred to ferric (Fe3⁺) salts, as the former is more readily absorbed through the duodenal mucosa⁴⁴ and is the more common formulation in commercially available supplements in the United States.⁴⁵ Oral supplementation with ferrous sulfate 325 mg (65 mg elemental iron) tablets is the first-line therapy for iron deficiency anemia.¹ Alternatively, ferrous gluconate 324 mg (38 mg elemental iron) over-the-counter and its liquid form has demonstrated superior absorption compared to ferrous sulfate tablets in a clinical study with peritoneal dialysis patients.^1,46 One study suggested that oral iron 40 to 80 mg should be taken every other day to increase absorption.⁴⁷ Due to improved bioavailability, intravenous iron may be utilized in patients with malabsorption, renal failure, or intolerance to oral iron (including those with gastric ulcers or active inflammatory bowel disease), with the formulation chosen based on underlying comorbidities and potential risks.^1,48 The theoretical risk for potentiating bacterial growth by increasing the amount of unbound iron in the blood raises concerns of iron supplementation in patients with infection or sepsis. Although far from definitive, existing data suggest that risk for infection is greater with intravenous iron supplementation and should be carefully considered prior to use.^49,50Elevated Ferritin—Elevated ferritin may be difficult to interpret given the multitude of conditions that can cause it.^23,51,52 Elevated serum ferritin can be broadly characterized by increased synthesis due to iron overload, increased synthesis due to inflammation, or increased ferritin release from cellular damage.³⁴ Further complicating interpretation is the potential diurnal fluctuations in serum iron levels dependent on dietary intake and timing of laboratory evaluation, choice of assay, differences in reference standards, and variations in calibration procedures that can lead to analytic variability in the measurement of ferritin.^23,53,54

Among healthy patients, serum ferritin is directly proportional to iron status.^9,51 A study utilizing weekly phlebotomy of 22 healthy participants to measure serum ferritin and calculate mobilizable storage iron found a strong positive correlation between the 2 variables (r=0.83, P<.001), with each 1-μg/L increase of serum ferritin corresponding to approximately an 8-mg increase of storage iron; the initial serum ferritin values ranged from 2 to 83 μg/L in females and 36 to 224 μg/L in males.⁵⁵ The correlation of ferritin with iron status also was supported by the significant correlation between the number of transfusions received in patients with transfusion-related iron overload and serum ferritin levels (r=0.89, P<.001), with an average increase of 60 μg/L per transfusion.⁵¹

Clinical guidelines on the interpretation of serum ferritin levels by Cullis et al³⁴ recommend a normal upper limit of 200 μg/L for healthy females and 300 μg/L for healthy males. Outside of clinical syndromes associated with iron overload, Lee and Means⁵⁶ found that serum ferritin of 1000 μg/L or higher was a nonspecific marker of disease, including infection and/or neoplastic disorders. We have adapted these guidelines to propose a workflow for evaluation of serum ferritin levels (Figure). In patients with inflammatory conditions or those affected by metabolic syndrome, elevated serum ferritin does not correlate with body iron status.^57,58 It is believed that inflammatory cytokines, including tumor necrosis factor α and IL-1α, can upregulate ferritin synthesis independent of cellular iron stores.^15,16 Several studies have examined the relationship between insulin resistance and/or metabolic syndrome with serum ferritin levels.^31,32 Han et al³² found that elevated serum ferritin was significantly associated with higher risk for metabolic syndrome in men (P<.0001) but not in women.

Although cutaneous manifestations of iron overload can be seen as skin hyperpigmentation due to increased iron deposits and increased melanin production,²² the effects of elevated ferritin on the skin and hair are not well known. Iron overload is a known trigger of porphyria cutanea tarda (PCT),⁵⁹ a condition in which reduced or absent enzymatic activity of uroporphyrinogen decarboxylase (UROD) leads to build up of toxic porphyrins in various organs.⁶⁰ In the skin, PCT manifests as a blistering photosensitive eruption that may resolve as dyspigmentation, scarring, and milia.⁶¹ Phlebotomy is first-line therapy in PCT to reduce serum iron and subsequent formation of UROD inhibitors, with guidelines suggesting discontinuation of phlebotomy when serum ferritin levels reach 20 ng/mL or lower.⁶⁰ Hyperferritinemia (serum ferritin >500 μg/L) is a common finding in several inflammatory disorders often accompanied by clinically apparent cutaneous symptoms such as adult-onset Still disease,⁶² hemophagocytic lymphohistiocytosis,^63,64 and anti-melanoma differentiation-associated gene 5 dermatomyositis.⁶⁵ Among these conditions, serum ferritin levels have been reported to correlate with disease activity, raising the question of whether ferritin is a bystander or a driver of the underlying pathology.^62,66,67 However, rapid decline of serum ferritin levels with treatment and control of inflammatory cytokines suggest that ferritin is unlikely to contribute to pathology.^62,67

Final Thoughts

Many clinical studies have examined the association between hair health and body iron status, the collective findings of which suggest that iron deficiency may be associated with TE. Among commonly measured serum iron parameters, low ferritin is a highly specific and sensitive marker for diagnosing iron deficiency. Serum ferritin may be a clinically useful tool for ruling out underlying iron deficiency in patients presenting with hair loss. Despite advances in our understanding of the molecular mechanisms of ferritin synthesis and regulation, whether ferritin itself contributes to cutaneous pathology is poorly understood.^35,36,68-74 For patients who are otherwise healthy with low suspicion for inflammatory disorders, chronic systemic illnesses, or malignancy, serum ferritin can be used as an indicator of body iron status. The workup for slightly elevated serum ferritin should be interpreted in the context of other laboratory findings and should be reassessed over time. Serum ferritin levels above 1000 μg/L warrant further investigation into causes such as iron overload conditions and underlying inflammatory conditions or malignancy.

Eating disorders (EDs) and feeding disorders refer to a wide spectrum of complex biopsychosocial illnesses. The spectrum of EDs encompasses anorexia nervosa (AN), bulimia nervosa (BN), binge eating disorder, and other specified feeding or eating disorders. Feeding disorders, distinguished from EDs based on the absence of body image disturbance, include pica, rumination syndrome, and avoidant/restrictive food intake disorder (ARFID).¹

This spectrum of illnesses predominantly affect young females aged 15 to 45 years, with recent increases in the rates of EDs among males, patients with skin of color, and adolescent females.^2-5 Patients with EDs are at an elevated lifetime risk of suicidal ideation, suicide attempts, and other psychiatric comorbidities compared to the general population.⁶ Specifically, AN and BN are associated with high psychiatric morbidity and mortality. A meta-analysis by Arcelus et al⁷ demonstrated the weighted annual mortality for AN was 5.10 deaths per 1000 person-years (95% CI, 3.57-7.59) among patients with EDs and 4.55 deaths for studies that selected inpatients (95% CI, 3.09-6.28); for BN, the weighted mortality was 1.74 deaths per 1000 person-years (95% CI, 1.09-2.44). Unfortunately, ED diagnoses often are delayed or missed in clinical settings. Patients may lack insight into the severity of their illness, experience embarrassment about their eating behaviors, or actively avoid treatment for their ED.⁸

Pica—compulsive eating of nonnutritive substances outside the cultural norm—and rumination syndrome—regurgitation of undigested food—are feeding disorders more commonly recognized in childhood.^9-11 Pregnancy, intellectual disability, iron deficiency, and lead poisoning are other conditions associated with pica.^6,9,10 Avoidant/restrictive food intake disorder, a new diagnosis added to the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5)¹ in 2013, is an eating or feeding disturbance resulting in persistent failure to meet nutritional or energy needs. Etiologies of ARFID may include sensory sensitivities and/or a traumatic event related to eating, leading to avoidance of associated foods.¹²

Patients with an ED or a feeding disorder frequently experience malnutrition, including deficiencies, excesses, or imbalances in nutritional intake, which may lead to nutritional dermatoses.¹³ As a result, the skin may present the first visible clues to an ED diagnosis.^8,14-19 Gupta et al¹⁸ organized the skin signs of EDs into 4 categories: (1) those secondary to starvation or malnutrition; (2) cutaneous injury related to self-induced vomiting; (3) dermatoses due to laxative, diuretic, or emetic use; and (4) other concomitant psychiatric illnesses (eg, hand dermatitis from compulsive handwashing, dermatodaxia, onychophagia, trichotillomania). This review will focus on the effects of malnutrition and starvation on the skin.

Skin findings in patients with EDs offer the treating dermatologist a special opportunity for early diagnosis and appropriate consultation with specialists trained in ED treatment. It is important for dermatologists to be vigilant in looking for skin findings of nutritional dermatoses, especially in populations at an increased risk for developing an ED, such as young female patients. The approach to therapy and treatment must occur through a collaborative multidisciplinary effort in a thoughtful and nonjudgmental environment.

Xerosis

Xerosis, or dry skin, is the most common dermatologic finding in both adult and pediatric patients with AN and BN.^14,19 It presents as skin roughness, tightness, flaking, and scaling, which may be complicated by fissuring, itching, and bleeding.²⁰ In healthy skin, moisture is maintained by the stratum corneum and its lipids such as ceramides, cholesterol, and free fatty acids.²¹ Natural moisturizing factor (NMF) within the skin is composed of amino acids, ammonia, urea, uric acid, inorganic salts, lactic acid derivatives, and pyrrolidine-3-carboxylic acid.^20-22 Disruptions to this system result in increased transepidermal water loss and impaired barrier function.²³

In patients with ED, xerosis arises through several mechanisms. Chronic illness or starvation can lead to euthyroid sick syndrome with decreased peripheral conversion of thyroxine (T₄) to triiodothyronine (T₃).^24,25 In the context of functional hypothyroidism, xerosis can arise from decreased eccrine gland secretion.²⁶ Secretions of water, lactate, urea, sodium, and potassium from eccrine glands help to maintain NMF for skin hydration.²⁷ Persistent laxative or diuretic abuse and fluid intake restriction, which are common behaviors across the spectrum of EDs, lead to dehydration and electrolyte imbalances that can manifest as skin dryness.²⁰ Disrupted keratinocyte differentiation due to insufficient stores of vitamins and minerals involved in keratinocyte differentiation, such as vitamins A and C, selenium, and zinc, also may contribute to xerosis.^25,28,29

Severely restrictive eating patterns may lead to development of protein energy malnutrition (PEM). Cutaneous findings in PEM occur due to dysmaturation of epidermal keratinocytes and epidermal atrophy.³⁰ Patients with severe persistent depletion of macronutrients—carbohydrates, fat, and protein—may experience marasmus, resulting in loss of subcutaneous fat that causes the appearance of dry loose skin.^29,31

Xerosis is exceedingly common in the general population and has no predictive value in ED diagnosis; however, this finding should be noted in the context of other signs suggestive of an ED. Treatment of xerosis in the setting of an ED should focus on correction of the underlying malnutrition. Symptomatic alleviation requires improving skin hydration and repairing barrier function. Mild xerosis may not need treatment or can be ameliorated with over-the-counter moisturizers and emollients. Scaling secondary to dry skin can be improved by ingredients such as glycerol, urea, lactic acid, and dexpanthenol.^20,32 Glycerol and urea are small hydrophilic molecules that penetrate the stratum corneum and help to bind moisture within the skin to reduce transepidermal water loss. Urea and lactic acid are keratolytics of NMF commonly found in moisturizers and emollients.^33,34 Dexpanthenol may be used for soothing fissures and pruritus; in vitro and in vivo studies have demonstrated its ability to upregulate dermal fibroblast proliferation and epidermal re-epithelization to promote faster wound healing.³⁵

Lanugo

Lanugo is clinically apparent as a layer of fine, minimally pigmented hair. It is physiologically present on the skin surface of fetuses and newborns. In utero, lanugo plays an essential role in fetal skin protection from amniotic fluid, as well as promotion of proper hydration, thermoregulation, and innate immune development.^36-38 Although it may be found on approximately 30% of newborns as normal variation, its presence beyond the neonatal period signals underlying systemic disease and severe undernutrition.^16,36,39 Rarely, hypertrichosis lanuginosa acquisita has been reported in association with malignancy.^40,41 The finding of lanugo beyond the neonatal period should prompt exclusion of other medical disorders, including neoplasms, chronic infections, hyperthyroidism, malabsorption syndromes, and inflammatory bowel disease.^41-47

There is a limited understanding of the pathomechanism behind lanugo development in the context of malnutrition. Intentional starvation leads to loss of subcutaneous fat and a state of functional hypothyroidism.⁴⁸ Studies hypothesize that lanugo develops as a response to hypothermia, regulated by dermal papillae cell–derived exosomes that may stimulate hair growth via paracrine signaling to outer root sheath cells.^36,49 Molecular studies have found that T₃ impacts skin and hair differentiation and proliferation by modulating thyroid hormone receptor regulation of keratin expression in epithelial cells.^50,51 Lanugo may be a clinical indicator of severe malnutrition among ED patients, especially children and adolescents. A study of 30 patients aged 8 to 17 years with AN and BN who underwent a standard dermatologic examination found significant positive correlation between the presence of lanugo hair growth and concomitant amenorrhea (P<.01) as well as between lanugo hair and body mass index lower than 16 kg/m² (P<.05).¹⁹ Discovery of lanugo in the dermatology clinical setting should prompt a thorough history, including screening questions about eating patterns; attitudes on eating, exercise, and appearance; personal and family history of EDs or other psychiatric disorders; and screening for depression and anxiety. Given its association with other signs of severe malnutrition, a clinical finding of lanugo should prompt close physical examination for other potential signs of an ED and laboratory evaluation for electrolyte levels and blood counts.⁵² Resolution of lanugo secondary to an ED is achieved with restoration of normal total body fat.¹⁸ Treatment should be focused on appropriate weight gain with the guidance of an ED specialist.

Pruritus

The prevalence and pathomechanism of pruritus secondary to EDs remains unclear.^16,53,54 There have been limited reports of pruritus secondary to ED, with Gupta et al⁵³ providing a case series of 6 patients with generalized pruritus in association with starvation and/or rapid weight loss. The study reported remission of pruritus with nutritional rehabilitation and/or weight gain of 5 to 10 pounds. Laboratory evaluation ruled out other causes of pruritus such as cholestasis and uremia.⁵³ Other case reports have associated pruritus with iron deficiency, with anecdotal evidence of pruritus resolution following iron supplementation.^55-59 Although we found no studies specifically relating iron deficiency, EDs, and pruritus, iron deficiency routinely is seen in ED patients and has a known association with pica.^9,10,60 As such, iron deficiency may be a contributing factor in pruritus in ED patients. A UK study of 19 women with AN and a body mass index lower than 16 kg/m² found that more than half of the patients (11/19 [57.9%]) described pruritus on the St. Thomas’ Itch Questionnaire, postulating that pruritus may be a clinical feature of AN.⁶¹ Limited studies with small samples make it difficult to conclude whether pruritus arises as a direct consequence of malnutrition.

Treatment of pruritus should address the underlying ED, as the pathophysiology of itch as it relates to malnutrition is poorly understood. Correction of existing nutritional imbalances by iron supplementation and appropriate weight gain may lead to symptom resolution. Because xerosis may be a contributing factor to pruritus, correction of the xerosis also may be therapeutic. More studies are needed on the connection between pruritus and the nutritional imbalances encountered in patients with EDs.

Acrocyanosis

Acrocyanosis is clinically seen as bluish-dusky discoloration most commonly affecting the hands and feet but also may affect the nose, ears, and nipples. Acrocyanosis typically is a sign of cold intolerance, hypothesized to occur in the context of AN due to shunting of blood centrally in response to hypothermia.^39,62 The diminished oxyhemoglobin delivery to extremity sites leads to the characteristic blue color.⁶³ In a study of 211 adolescent females (age range, 13–17 years) with AN, physical examination revealed peripheral hypothermia and peripheral cyanosis in 80% and 43% of patients, respectively.⁴⁸ Cold intolerance seen in EDs may be secondary to a functional hypothyroid state similar to euthyroid sick syndrome seen in conditions of severe caloric deficit.²⁵

It is possible that anemia and dehydration can worsen acrocyanosis due to impaired delivery of oxyhemoglobin to the body’s periphery.⁶³ In a study of 14 ED patients requiring inpatient care, 6 were found to have underlying anemia following intravenous fluid supplementation.⁶⁴ On admission, the mean (SD) hemoglobin and hematocrit across 14 patients was 12.74 (2.19) and 37.42 (5.99), respectively. Following intravenous fluid supplementation, the mean (SD) hemoglobin and hematocrit decreased to 9.88 (1.79)(P<.001) and 29.56 (4.91)(P=.008), respectively. Most cases reported intentional restriction of dietary sodium and fluid intake, with 2 patients reporting a history of diuretic misuse.⁶⁴ These findings demonstrate that hemoglobin and hematocrit may be falsely normal in patients with AN due to hemoconcentration, suggesting that anemia may be underdiagnosed in inpatients with AN.

Beyond treatment of the underlying ED, acrocyanosis therapy is focused on improvement of circulation and avoidance of exacerbating factors. Pharmacologic intervention rarely is needed. Patients should be reassured that acrocyanosis is a benign condition and often can be improved by dressing warmly and avoiding exposure to cold. Severe cases may warrant trial treatment with nicotinic acid derivatives, α-adrenergic blockade, and topical minoxidil, which have demonstrated limited benefit in treating primary idiopathic acrocyanosis.⁶³

Carotenoderma

Carotenoderma—the presence of a yellow discoloration to skin secondary to hypercarotenemia—has been described in patients with EDs since the 1960s.^65,66 Beyond its clinical appearance, carotenoderma is asymptomatic. Carotenoids are lipid-soluble compounds present in the diet that are metabolized by the intestinal mucosa and liver to the primary conversion product, retinaldehyde, which is further converted to retinol, retinyl esters, and other retinoid metabolites.^67,68 Retinol is bound by lipoproteins and transported in the plasma, then deposited in peripheral tissues,⁶⁹ including in intercellular lipids in the stratum corneum, resulting in an orange hue that is most apparent in sites of increased skin thickness and sweating (eg, palms, soles, nasolabial folds).⁷⁰ In an observational study of ED patients, Glorio et al¹⁴ found that carotenoderma was present in 23.77% (29/122) and 25% (4/16) of patients with BN and other specified feeding or eating disorder, respectively; it was not noted among patients with AN. Prior case reports have provided anecdotal evidence of carotenoderma in AN patients.^66,71 In the setting of an ED, increased serum carotenoids likely are due to increased ingestion of carotene-rich foods, leading to increased levels of carotenoid-bound lipoproteins in the serum.⁷⁰ Resolution of xanthoderma requires restriction of carotenoid intake and may take 2 to 3 months to be clinically apparent. The lipophilic nature of carotenoids allows storage in body fat, prolonging resolution.⁷¹

Hair Changes

Telogen effluvium (TE) and hair pigmentary changes are clinical findings that have been reported in association with EDs.^14,16,19,72 Telogen effluvium occurs when physiologic stress causes a large portion of hairs in the anagen phase of growth to prematurely shift into the catagen then telogen phase. Approximately 2 to 3 months following the initial insult, there is clinically apparent excessive hair shedding compared to baseline.⁷³ Studies have demonstrated that patients with EDs commonly have psychiatric comorbidities such as mood and anxiety disorders, obsessive compulsive disorder, posttraumatic stress disorder, and panic disorder compared to the general population.^6,74-76 As such, stress experienced by ED patients may contribute to TE. Despite TE being commonly reported in ED patients,^16-18 there is a lack of controlled studies of TE in human subjects with ED. An animal model for TE demonstrated that stressed mice exhibited further progression in the hair cycle compared with nonstressed mice (P<.01); the majority of hair follicles in stressed mice were in the catagen phase, while the majority of hair follicles in nonstressed mice were in the anagen phase.⁷⁷ Stressed mice demonstrated an increased number of major histocompatibility complex class II⁺ cell clusters, composed mostly of activated macrophages, per 12.5-mm epidermal length compared to nonstressed mice (mean [SEM], 7.0 [1.1] vs 2.0 [0.3][P<.05]). This study illustrated that stress can lead to inflammatory cell recruitment and activation in the hair follicle microenvironment with growth-inhibitory effects.⁷⁷

The flag sign, or alternating bands of lesser and greater pigmentation in the hair, has been reported in cases of severe PEM.³¹ In addition, PEM may lead to scalp alopecia, dry and brittle hair, and/or hypopigmentation with periods of inadequate nutrition.^29,78 Scalp hair hypopigmentation, brittleness, and alopecia have been reported in pediatric patients with highly selective eating and/or ARFID.^79,80 Maruo et al⁸⁰ described a 3-year-old boy with ASD who consumed only potato chips for more than a year. Physical examination revealed reduced skin turgor overall and sparse red-brown hair on the scalp; laboratory testing showed deficiencies of protein, vitamin A, vitamin D, copper, and zinc. The patient was admitted for nutritional rehabilitation via nasogastric tube feeding, leading to resolution of laboratory abnormalities and growth of thicker black scalp hair over the course of several months.⁸⁰

Neuroendocrine control of keratin expression by thyroid-stimulating hormone (TSH) and thyroid hormones likely plays a role in the regulation of hair follicle activities, including hair growth, structure, and stem cell differentiation.^81,82 Altered thyroid hormone activity, which commonly is seen in patients with EDs,^24,25 may contribute to impaired hair growth and pigmentation.^26,51,83-85 Using tissue cultures of human anagen hair follicles, van Beek et al⁸⁵ provided in vitro evidence that T₃ and T₄ modulate scalp hair follicle growth and pigmentation. Both T₃- and T₄-treated tissue exhibited increased numbers of anagen and decreased numbers of catagen hair follicles in organ cultures compared with control (P<.01); on quantitative Fontana-Masson histochemistry, T₃ and T₄ significantly stimulated hair follicle melanin synthesis compared with control (P<.001 and P<.01, respectively).⁸⁵ Molecular studies by Bodó et al⁸³ have shown that the human scalp epidermis expresses TSH at the messenger RNA and protein levels. Both studies showed that intraepidermal TSH expression is downregulated by thyroid hormones.^83,85 Further studies are needed to examine the impact of malnutrition on local thyroid hormone signaling and action at the level of the dermis, epidermis, and hair follicle.

Discovery of TE, hair loss, and/or hair hypopigmentation should prompt close investigation for other signs of thyroid dysfunction, specifically secondary to malnutrition. Imbalances in TSH, T₃, and T₄ should be corrected. Nutritional deficiencies and dietary habits should be addressed through careful nutritional rehabilitation and targeted ED treatment.

Oral and Mucosal Symptoms

Symptoms of the oral cavity that may arise secondary to EDs and feeding disorders include glossitis, stomatitis, cheilitis, and dental erosions. Mucosal symptoms have been observed in patients with vitamin B deficiencies, inflammatory bowel disease, and other malabsorptive disorders, including patients with EDs.^86-88 Patients following restrictive diets, specifically strict vegan diets, without additional supplementation are at risk for developing vitamin B₁₂ deficiency. Because vitamin B₁₂ is stored in the liver, symptoms of deficiency appear when hepatic stores are depleted over the course of several years.⁸⁹ Insufficient vitamin B₁₂ prevents the proper functioning of methionine synthase, which is required for the conversion of homocysteine to methionine and for the conversion of methyl-tetrahydrofolate to tetrahydrofolate.⁸⁹ Impairment of this process impedes the synthesis of pyrimidine bases of DNA, disrupting the production of rapidly proliferating cells such as myeloid cells or mucosal lining cells. In cases of glossitis and/or stomatitis due to vitamin B₁₂ deficiency, resolution of lesions was achieved within 4 weeks of daily oral supplementation with vitamin B₁₂ at 2 μg daily.^90,91 Iron deficiency, a common finding in EDs, also may contribute to glossitis and angular cheilitis.²⁹ If uncovered, iron deficiency should be corrected by supplementation based on total deficit, age, and sex. Oral supplementation may be done with oral ferrous sulfate (325 mg provides 65 mg elemental iron) or with other iron salts such as ferrous gluconate (325 mg provides 38 mg elemental iron).²⁹ Mucosal symptoms of cheilitis and labial erythema may arise from irritation due to self-induced vomiting.⁸⁸

Dental erosion refers to loss of tooth structure via a chemical process that does not involve bacteria; in contrast, dental caries refer to tooth damage secondary to bacterial acid production. Patients with EDs who repeatedly self-induce vomiting have persistent introduction of gastric acids into the oral cavity, resulting in dissolution of the tooth enamel, which occurs when teeth are persistently exposed to a pH less than 5.5.⁹² Feeding disorders also may predispose patients to dental pathology. In a study of 60 pediatric patients, those with rumination syndrome were significantly more likely to have dental erosions than age- and sex-matched healthy controls (23/30 [77%] vs 4/30 [13%][P<.001]). The same study found no difference in the frequency of dental caries between children with and without rumination syndrome.⁹² These findings suggest that rumination syndrome increases the risk for dental erosions but not dental caries. The distribution of teeth affected by dental erosions may differ between EDs and feeding disorders. Patients with BN are more likely to experience involvement of the palatal surfaces of maxillary teeth, while patients with rumination syndrome had equal involvement of maxillary and mandibular teeth.⁹²

There is limited literature on the role of dentists in the care of patients with EDs and feeding disorders, though existing studies suggest inclusion of a dental care professional in multidisciplinary treatment along with emphasis on education around a home dental care regimen and frequent dental follow-up.^76,93,94 Prevention of further damage requires correction of the underlying behaviors and ED.

Other Dermatologic Findings

Russell sign refers to the development of calluses on the dorsal metacarpophalangeal joints of the dominant hand due to self-induced vomiting. Due to its specificity in purging-type EDs, the discovery of Russell sign should greatly increase suspicion for an ED.¹⁷ Patients with EDs also are at an increased risk for self-harming and body-focused repetitive behaviors, including skin cutting, superficial burning, onychophagia, and trichotillomania.¹⁹ It is important to recognize these signs in patients for whom an ED is suspected. The role of the dermatologist should include careful examination of the skin and documentation of findings that may aid in the diagnosis of an underlying ED.

Final Thoughts

A major limitation of this review is the reliance on small case reports and case series reporting cutaneous manifestations of ED. Controlled studies with larger cohorts are challenging in this population but are needed to substantiate the dermatologic signs commonly associated with EDs. Translational studies may help elucidate the pathomechanisms underlying dermatologic diseases such as lanugo, pruritus, and alopecia in the context of EDs and malnutrition. The known association between thyroid dysfunction and skin disease has been substantiated by clinical and basic science investigation, suggesting a notable role of thyroid hormone and TSH signaling in the skin local environment. Further investigation into nutritional and neuroendocrine regulation of skin health will aid in the diagnosis and treatment of patients impacted by EDs.

The treatment of the underlying ED is key in correcting associated skin disease, which requires interdisciplinary collaboration that addresses the psychological, behavioral, and social components of the condition. Following a diagnosis of ED, assessment should be made of the nutritional rehabilitation required to restore weight and nutritional status. Inpatient treatment may be indicated for patients requiring close monitoring to avoid refeeding syndrome, or those who meet the criteria for extreme AN in the DSM-5 (ie, body mass index <15 kg/m²),¹ or demonstrate signs of medical instability or organ failure secondary to malnutrition.⁶² Long-term recovery for ED patients should focus on behavioral therapy with a multidisciplinary team consisting of a psychiatrist, therapist, dietitian, and primary care provider. Comparative studies in large-scale trials of cognitive behavioral therapy, focal psychodynamic psychotherapy, and specialist supportive clinical management have shown little to no difference in efficacy in treating EDs.^75,95,96

Dermatologists may be the first providers to observe sequelae of nutritional and behavioral derangement in patients with EDs. Existing literature on the dermatologic findings of EDs report great heterogeneity of skin signs, with a very limited number of controlled studies available. Each cutaneous symptom described in this review should not be interpreted as an isolated pathology but should be placed in the context of patient predisposing risk factors and the constellation of other skin findings that may be suggestive of disordered eating behavior or other psychiatric illness. The observation of multiple signs and symptoms at the same time, especially of symptoms uncommonly encountered or suggestive of a severe and prolonged imbalance (eg, xanthoderma with vitamin A excess, aphthous stomatitis with vitamin B deficiency), should heighten clinical suspicion for an underlying ED. A clinician’s highest priority should be to resolve life-threatening medical emergencies and address nutritional derangements with the assistance of experts who are well versed in EDs. The patient should undergo workup to rule out organic causes of their nutritional dermatoses. Given the high psychiatric morbidity and mortality of patients with an ED and the demonstrated benefit of early intervention, recognition of cutaneous manifestations of malnutrition and EDs may be paramount to improving outcomes.

Eating disorders (EDs) and feeding disorders refer to a wide spectrum of complex biopsychosocial illnesses. The spectrum of EDs encompasses anorexia nervosa (AN), bulimia nervosa (BN), binge eating disorder, and other specified feeding or eating disorders. Feeding disorders, distinguished from EDs based on the absence of body image disturbance, include pica, rumination syndrome, and avoidant/restrictive food intake disorder (ARFID).¹

This spectrum of illnesses predominantly affect young females aged 15 to 45 years, with recent increases in the rates of EDs among males, patients with skin of color, and adolescent females.^2-5 Patients with EDs are at an elevated lifetime risk of suicidal ideation, suicide attempts, and other psychiatric comorbidities compared to the general population.⁶ Specifically, AN and BN are associated with high psychiatric morbidity and mortality. A meta-analysis by Arcelus et al⁷ demonstrated the weighted annual mortality for AN was 5.10 deaths per 1000 person-years (95% CI, 3.57-7.59) among patients with EDs and 4.55 deaths for studies that selected inpatients (95% CI, 3.09-6.28); for BN, the weighted mortality was 1.74 deaths per 1000 person-years (95% CI, 1.09-2.44). Unfortunately, ED diagnoses often are delayed or missed in clinical settings. Patients may lack insight into the severity of their illness, experience embarrassment about their eating behaviors, or actively avoid treatment for their ED.⁸

Pica—compulsive eating of nonnutritive substances outside the cultural norm—and rumination syndrome—regurgitation of undigested food—are feeding disorders more commonly recognized in childhood.^9-11 Pregnancy, intellectual disability, iron deficiency, and lead poisoning are other conditions associated with pica.^6,9,10 Avoidant/restrictive food intake disorder, a new diagnosis added to the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5)¹ in 2013, is an eating or feeding disturbance resulting in persistent failure to meet nutritional or energy needs. Etiologies of ARFID may include sensory sensitivities and/or a traumatic event related to eating, leading to avoidance of associated foods.¹²

Patients with an ED or a feeding disorder frequently experience malnutrition, including deficiencies, excesses, or imbalances in nutritional intake, which may lead to nutritional dermatoses.¹³ As a result, the skin may present the first visible clues to an ED diagnosis.^8,14-19 Gupta et al¹⁸ organized the skin signs of EDs into 4 categories: (1) those secondary to starvation or malnutrition; (2) cutaneous injury related to self-induced vomiting; (3) dermatoses due to laxative, diuretic, or emetic use; and (4) other concomitant psychiatric illnesses (eg, hand dermatitis from compulsive handwashing, dermatodaxia, onychophagia, trichotillomania). This review will focus on the effects of malnutrition and starvation on the skin.

Skin findings in patients with EDs offer the treating dermatologist a special opportunity for early diagnosis and appropriate consultation with specialists trained in ED treatment. It is important for dermatologists to be vigilant in looking for skin findings of nutritional dermatoses, especially in populations at an increased risk for developing an ED, such as young female patients. The approach to therapy and treatment must occur through a collaborative multidisciplinary effort in a thoughtful and nonjudgmental environment.

Xerosis

Xerosis, or dry skin, is the most common dermatologic finding in both adult and pediatric patients with AN and BN.^14,19 It presents as skin roughness, tightness, flaking, and scaling, which may be complicated by fissuring, itching, and bleeding.²⁰ In healthy skin, moisture is maintained by the stratum corneum and its lipids such as ceramides, cholesterol, and free fatty acids.²¹ Natural moisturizing factor (NMF) within the skin is composed of amino acids, ammonia, urea, uric acid, inorganic salts, lactic acid derivatives, and pyrrolidine-3-carboxylic acid.^20-22 Disruptions to this system result in increased transepidermal water loss and impaired barrier function.²³

In patients with ED, xerosis arises through several mechanisms. Chronic illness or starvation can lead to euthyroid sick syndrome with decreased peripheral conversion of thyroxine (T₄) to triiodothyronine (T₃).^24,25 In the context of functional hypothyroidism, xerosis can arise from decreased eccrine gland secretion.²⁶ Secretions of water, lactate, urea, sodium, and potassium from eccrine glands help to maintain NMF for skin hydration.²⁷ Persistent laxative or diuretic abuse and fluid intake restriction, which are common behaviors across the spectrum of EDs, lead to dehydration and electrolyte imbalances that can manifest as skin dryness.²⁰ Disrupted keratinocyte differentiation due to insufficient stores of vitamins and minerals involved in keratinocyte differentiation, such as vitamins A and C, selenium, and zinc, also may contribute to xerosis.^25,28,29

Severely restrictive eating patterns may lead to development of protein energy malnutrition (PEM). Cutaneous findings in PEM occur due to dysmaturation of epidermal keratinocytes and epidermal atrophy.³⁰ Patients with severe persistent depletion of macronutrients—carbohydrates, fat, and protein—may experience marasmus, resulting in loss of subcutaneous fat that causes the appearance of dry loose skin.^29,31

Xerosis is exceedingly common in the general population and has no predictive value in ED diagnosis; however, this finding should be noted in the context of other signs suggestive of an ED. Treatment of xerosis in the setting of an ED should focus on correction of the underlying malnutrition. Symptomatic alleviation requires improving skin hydration and repairing barrier function. Mild xerosis may not need treatment or can be ameliorated with over-the-counter moisturizers and emollients. Scaling secondary to dry skin can be improved by ingredients such as glycerol, urea, lactic acid, and dexpanthenol.^20,32 Glycerol and urea are small hydrophilic molecules that penetrate the stratum corneum and help to bind moisture within the skin to reduce transepidermal water loss. Urea and lactic acid are keratolytics of NMF commonly found in moisturizers and emollients.^33,34 Dexpanthenol may be used for soothing fissures and pruritus; in vitro and in vivo studies have demonstrated its ability to upregulate dermal fibroblast proliferation and epidermal re-epithelization to promote faster wound healing.³⁵

Lanugo

Lanugo is clinically apparent as a layer of fine, minimally pigmented hair. It is physiologically present on the skin surface of fetuses and newborns. In utero, lanugo plays an essential role in fetal skin protection from amniotic fluid, as well as promotion of proper hydration, thermoregulation, and innate immune development.^36-38 Although it may be found on approximately 30% of newborns as normal variation, its presence beyond the neonatal period signals underlying systemic disease and severe undernutrition.^16,36,39 Rarely, hypertrichosis lanuginosa acquisita has been reported in association with malignancy.^40,41 The finding of lanugo beyond the neonatal period should prompt exclusion of other medical disorders, including neoplasms, chronic infections, hyperthyroidism, malabsorption syndromes, and inflammatory bowel disease.^41-47

There is a limited understanding of the pathomechanism behind lanugo development in the context of malnutrition. Intentional starvation leads to loss of subcutaneous fat and a state of functional hypothyroidism.⁴⁸ Studies hypothesize that lanugo develops as a response to hypothermia, regulated by dermal papillae cell–derived exosomes that may stimulate hair growth via paracrine signaling to outer root sheath cells.^36,49 Molecular studies have found that T₃ impacts skin and hair differentiation and proliferation by modulating thyroid hormone receptor regulation of keratin expression in epithelial cells.^50,51 Lanugo may be a clinical indicator of severe malnutrition among ED patients, especially children and adolescents. A study of 30 patients aged 8 to 17 years with AN and BN who underwent a standard dermatologic examination found significant positive correlation between the presence of lanugo hair growth and concomitant amenorrhea (P<.01) as well as between lanugo hair and body mass index lower than 16 kg/m² (P<.05).¹⁹ Discovery of lanugo in the dermatology clinical setting should prompt a thorough history, including screening questions about eating patterns; attitudes on eating, exercise, and appearance; personal and family history of EDs or other psychiatric disorders; and screening for depression and anxiety. Given its association with other signs of severe malnutrition, a clinical finding of lanugo should prompt close physical examination for other potential signs of an ED and laboratory evaluation for electrolyte levels and blood counts.⁵² Resolution of lanugo secondary to an ED is achieved with restoration of normal total body fat.¹⁸ Treatment should be focused on appropriate weight gain with the guidance of an ED specialist.

Pruritus

The prevalence and pathomechanism of pruritus secondary to EDs remains unclear.^16,53,54 There have been limited reports of pruritus secondary to ED, with Gupta et al⁵³ providing a case series of 6 patients with generalized pruritus in association with starvation and/or rapid weight loss. The study reported remission of pruritus with nutritional rehabilitation and/or weight gain of 5 to 10 pounds. Laboratory evaluation ruled out other causes of pruritus such as cholestasis and uremia.⁵³ Other case reports have associated pruritus with iron deficiency, with anecdotal evidence of pruritus resolution following iron supplementation.^55-59 Although we found no studies specifically relating iron deficiency, EDs, and pruritus, iron deficiency routinely is seen in ED patients and has a known association with pica.^9,10,60 As such, iron deficiency may be a contributing factor in pruritus in ED patients. A UK study of 19 women with AN and a body mass index lower than 16 kg/m² found that more than half of the patients (11/19 [57.9%]) described pruritus on the St. Thomas’ Itch Questionnaire, postulating that pruritus may be a clinical feature of AN.⁶¹ Limited studies with small samples make it difficult to conclude whether pruritus arises as a direct consequence of malnutrition.

Treatment of pruritus should address the underlying ED, as the pathophysiology of itch as it relates to malnutrition is poorly understood. Correction of existing nutritional imbalances by iron supplementation and appropriate weight gain may lead to symptom resolution. Because xerosis may be a contributing factor to pruritus, correction of the xerosis also may be therapeutic. More studies are needed on the connection between pruritus and the nutritional imbalances encountered in patients with EDs.

Acrocyanosis

Acrocyanosis is clinically seen as bluish-dusky discoloration most commonly affecting the hands and feet but also may affect the nose, ears, and nipples. Acrocyanosis typically is a sign of cold intolerance, hypothesized to occur in the context of AN due to shunting of blood centrally in response to hypothermia.^39,62 The diminished oxyhemoglobin delivery to extremity sites leads to the characteristic blue color.⁶³ In a study of 211 adolescent females (age range, 13–17 years) with AN, physical examination revealed peripheral hypothermia and peripheral cyanosis in 80% and 43% of patients, respectively.⁴⁸ Cold intolerance seen in EDs may be secondary to a functional hypothyroid state similar to euthyroid sick syndrome seen in conditions of severe caloric deficit.²⁵

It is possible that anemia and dehydration can worsen acrocyanosis due to impaired delivery of oxyhemoglobin to the body’s periphery.⁶³ In a study of 14 ED patients requiring inpatient care, 6 were found to have underlying anemia following intravenous fluid supplementation.⁶⁴ On admission, the mean (SD) hemoglobin and hematocrit across 14 patients was 12.74 (2.19) and 37.42 (5.99), respectively. Following intravenous fluid supplementation, the mean (SD) hemoglobin and hematocrit decreased to 9.88 (1.79)(P<.001) and 29.56 (4.91)(P=.008), respectively. Most cases reported intentional restriction of dietary sodium and fluid intake, with 2 patients reporting a history of diuretic misuse.⁶⁴ These findings demonstrate that hemoglobin and hematocrit may be falsely normal in patients with AN due to hemoconcentration, suggesting that anemia may be underdiagnosed in inpatients with AN.

Beyond treatment of the underlying ED, acrocyanosis therapy is focused on improvement of circulation and avoidance of exacerbating factors. Pharmacologic intervention rarely is needed. Patients should be reassured that acrocyanosis is a benign condition and often can be improved by dressing warmly and avoiding exposure to cold. Severe cases may warrant trial treatment with nicotinic acid derivatives, α-adrenergic blockade, and topical minoxidil, which have demonstrated limited benefit in treating primary idiopathic acrocyanosis.⁶³

Carotenoderma

Carotenoderma—the presence of a yellow discoloration to skin secondary to hypercarotenemia—has been described in patients with EDs since the 1960s.^65,66 Beyond its clinical appearance, carotenoderma is asymptomatic. Carotenoids are lipid-soluble compounds present in the diet that are metabolized by the intestinal mucosa and liver to the primary conversion product, retinaldehyde, which is further converted to retinol, retinyl esters, and other retinoid metabolites.^67,68 Retinol is bound by lipoproteins and transported in the plasma, then deposited in peripheral tissues,⁶⁹ including in intercellular lipids in the stratum corneum, resulting in an orange hue that is most apparent in sites of increased skin thickness and sweating (eg, palms, soles, nasolabial folds).⁷⁰ In an observational study of ED patients, Glorio et al¹⁴ found that carotenoderma was present in 23.77% (29/122) and 25% (4/16) of patients with BN and other specified feeding or eating disorder, respectively; it was not noted among patients with AN. Prior case reports have provided anecdotal evidence of carotenoderma in AN patients.^66,71 In the setting of an ED, increased serum carotenoids likely are due to increased ingestion of carotene-rich foods, leading to increased levels of carotenoid-bound lipoproteins in the serum.⁷⁰ Resolution of xanthoderma requires restriction of carotenoid intake and may take 2 to 3 months to be clinically apparent. The lipophilic nature of carotenoids allows storage in body fat, prolonging resolution.⁷¹

Hair Changes

Telogen effluvium (TE) and hair pigmentary changes are clinical findings that have been reported in association with EDs.^14,16,19,72 Telogen effluvium occurs when physiologic stress causes a large portion of hairs in the anagen phase of growth to prematurely shift into the catagen then telogen phase. Approximately 2 to 3 months following the initial insult, there is clinically apparent excessive hair shedding compared to baseline.⁷³ Studies have demonstrated that patients with EDs commonly have psychiatric comorbidities such as mood and anxiety disorders, obsessive compulsive disorder, posttraumatic stress disorder, and panic disorder compared to the general population.^6,74-76 As such, stress experienced by ED patients may contribute to TE. Despite TE being commonly reported in ED patients,^16-18 there is a lack of controlled studies of TE in human subjects with ED. An animal model for TE demonstrated that stressed mice exhibited further progression in the hair cycle compared with nonstressed mice (P<.01); the majority of hair follicles in stressed mice were in the catagen phase, while the majority of hair follicles in nonstressed mice were in the anagen phase.⁷⁷ Stressed mice demonstrated an increased number of major histocompatibility complex class II⁺ cell clusters, composed mostly of activated macrophages, per 12.5-mm epidermal length compared to nonstressed mice (mean [SEM], 7.0 [1.1] vs 2.0 [0.3][P<.05]). This study illustrated that stress can lead to inflammatory cell recruitment and activation in the hair follicle microenvironment with growth-inhibitory effects.⁷⁷

The flag sign, or alternating bands of lesser and greater pigmentation in the hair, has been reported in cases of severe PEM.³¹ In addition, PEM may lead to scalp alopecia, dry and brittle hair, and/or hypopigmentation with periods of inadequate nutrition.^29,78 Scalp hair hypopigmentation, brittleness, and alopecia have been reported in pediatric patients with highly selective eating and/or ARFID.^79,80 Maruo et al⁸⁰ described a 3-year-old boy with ASD who consumed only potato chips for more than a year. Physical examination revealed reduced skin turgor overall and sparse red-brown hair on the scalp; laboratory testing showed deficiencies of protein, vitamin A, vitamin D, copper, and zinc. The patient was admitted for nutritional rehabilitation via nasogastric tube feeding, leading to resolution of laboratory abnormalities and growth of thicker black scalp hair over the course of several months.⁸⁰

Neuroendocrine control of keratin expression by thyroid-stimulating hormone (TSH) and thyroid hormones likely plays a role in the regulation of hair follicle activities, including hair growth, structure, and stem cell differentiation.^81,82 Altered thyroid hormone activity, which commonly is seen in patients with EDs,^24,25 may contribute to impaired hair growth and pigmentation.^26,51,83-85 Using tissue cultures of human anagen hair follicles, van Beek et al⁸⁵ provided in vitro evidence that T₃ and T₄ modulate scalp hair follicle growth and pigmentation. Both T₃- and T₄-treated tissue exhibited increased numbers of anagen and decreased numbers of catagen hair follicles in organ cultures compared with control (P<.01); on quantitative Fontana-Masson histochemistry, T₃ and T₄ significantly stimulated hair follicle melanin synthesis compared with control (P<.001 and P<.01, respectively).⁸⁵ Molecular studies by Bodó et al⁸³ have shown that the human scalp epidermis expresses TSH at the messenger RNA and protein levels. Both studies showed that intraepidermal TSH expression is downregulated by thyroid hormones.^83,85 Further studies are needed to examine the impact of malnutrition on local thyroid hormone signaling and action at the level of the dermis, epidermis, and hair follicle.

Discovery of TE, hair loss, and/or hair hypopigmentation should prompt close investigation for other signs of thyroid dysfunction, specifically secondary to malnutrition. Imbalances in TSH, T₃, and T₄ should be corrected. Nutritional deficiencies and dietary habits should be addressed through careful nutritional rehabilitation and targeted ED treatment.

Oral and Mucosal Symptoms

Symptoms of the oral cavity that may arise secondary to EDs and feeding disorders include glossitis, stomatitis, cheilitis, and dental erosions. Mucosal symptoms have been observed in patients with vitamin B deficiencies, inflammatory bowel disease, and other malabsorptive disorders, including patients with EDs.^86-88 Patients following restrictive diets, specifically strict vegan diets, without additional supplementation are at risk for developing vitamin B₁₂ deficiency. Because vitamin B₁₂ is stored in the liver, symptoms of deficiency appear when hepatic stores are depleted over the course of several years.⁸⁹ Insufficient vitamin B₁₂ prevents the proper functioning of methionine synthase, which is required for the conversion of homocysteine to methionine and for the conversion of methyl-tetrahydrofolate to tetrahydrofolate.⁸⁹ Impairment of this process impedes the synthesis of pyrimidine bases of DNA, disrupting the production of rapidly proliferating cells such as myeloid cells or mucosal lining cells. In cases of glossitis and/or stomatitis due to vitamin B₁₂ deficiency, resolution of lesions was achieved within 4 weeks of daily oral supplementation with vitamin B₁₂ at 2 μg daily.^90,91 Iron deficiency, a common finding in EDs, also may contribute to glossitis and angular cheilitis.²⁹ If uncovered, iron deficiency should be corrected by supplementation based on total deficit, age, and sex. Oral supplementation may be done with oral ferrous sulfate (325 mg provides 65 mg elemental iron) or with other iron salts such as ferrous gluconate (325 mg provides 38 mg elemental iron).²⁹ Mucosal symptoms of cheilitis and labial erythema may arise from irritation due to self-induced vomiting.⁸⁸

Dental erosion refers to loss of tooth structure via a chemical process that does not involve bacteria; in contrast, dental caries refer to tooth damage secondary to bacterial acid production. Patients with EDs who repeatedly self-induce vomiting have persistent introduction of gastric acids into the oral cavity, resulting in dissolution of the tooth enamel, which occurs when teeth are persistently exposed to a pH less than 5.5.⁹² Feeding disorders also may predispose patients to dental pathology. In a study of 60 pediatric patients, those with rumination syndrome were significantly more likely to have dental erosions than age- and sex-matched healthy controls (23/30 [77%] vs 4/30 [13%][P<.001]). The same study found no difference in the frequency of dental caries between children with and without rumination syndrome.⁹² These findings suggest that rumination syndrome increases the risk for dental erosions but not dental caries. The distribution of teeth affected by dental erosions may differ between EDs and feeding disorders. Patients with BN are more likely to experience involvement of the palatal surfaces of maxillary teeth, while patients with rumination syndrome had equal involvement of maxillary and mandibular teeth.⁹²

There is limited literature on the role of dentists in the care of patients with EDs and feeding disorders, though existing studies suggest inclusion of a dental care professional in multidisciplinary treatment along with emphasis on education around a home dental care regimen and frequent dental follow-up.^76,93,94 Prevention of further damage requires correction of the underlying behaviors and ED.

Other Dermatologic Findings

Russell sign refers to the development of calluses on the dorsal metacarpophalangeal joints of the dominant hand due to self-induced vomiting. Due to its specificity in purging-type EDs, the discovery of Russell sign should greatly increase suspicion for an ED.¹⁷ Patients with EDs also are at an increased risk for self-harming and body-focused repetitive behaviors, including skin cutting, superficial burning, onychophagia, and trichotillomania.¹⁹ It is important to recognize these signs in patients for whom an ED is suspected. The role of the dermatologist should include careful examination of the skin and documentation of findings that may aid in the diagnosis of an underlying ED.

Final Thoughts

A major limitation of this review is the reliance on small case reports and case series reporting cutaneous manifestations of ED. Controlled studies with larger cohorts are challenging in this population but are needed to substantiate the dermatologic signs commonly associated with EDs. Translational studies may help elucidate the pathomechanisms underlying dermatologic diseases such as lanugo, pruritus, and alopecia in the context of EDs and malnutrition. The known association between thyroid dysfunction and skin disease has been substantiated by clinical and basic science investigation, suggesting a notable role of thyroid hormone and TSH signaling in the skin local environment. Further investigation into nutritional and neuroendocrine regulation of skin health will aid in the diagnosis and treatment of patients impacted by EDs.

The treatment of the underlying ED is key in correcting associated skin disease, which requires interdisciplinary collaboration that addresses the psychological, behavioral, and social components of the condition. Following a diagnosis of ED, assessment should be made of the nutritional rehabilitation required to restore weight and nutritional status. Inpatient treatment may be indicated for patients requiring close monitoring to avoid refeeding syndrome, or those who meet the criteria for extreme AN in the DSM-5 (ie, body mass index <15 kg/m²),¹ or demonstrate signs of medical instability or organ failure secondary to malnutrition.⁶² Long-term recovery for ED patients should focus on behavioral therapy with a multidisciplinary team consisting of a psychiatrist, therapist, dietitian, and primary care provider. Comparative studies in large-scale trials of cognitive behavioral therapy, focal psychodynamic psychotherapy, and specialist supportive clinical management have shown little to no difference in efficacy in treating EDs.^75,95,96

Dermatologists may be the first providers to observe sequelae of nutritional and behavioral derangement in patients with EDs. Existing literature on the dermatologic findings of EDs report great heterogeneity of skin signs, with a very limited number of controlled studies available. Each cutaneous symptom described in this review should not be interpreted as an isolated pathology but should be placed in the context of patient predisposing risk factors and the constellation of other skin findings that may be suggestive of disordered eating behavior or other psychiatric illness. The observation of multiple signs and symptoms at the same time, especially of symptoms uncommonly encountered or suggestive of a severe and prolonged imbalance (eg, xanthoderma with vitamin A excess, aphthous stomatitis with vitamin B deficiency), should heighten clinical suspicion for an underlying ED. A clinician’s highest priority should be to resolve life-threatening medical emergencies and address nutritional derangements with the assistance of experts who are well versed in EDs. The patient should undergo workup to rule out organic causes of their nutritional dermatoses. Given the high psychiatric morbidity and mortality of patients with an ED and the demonstrated benefit of early intervention, recognition of cutaneous manifestations of malnutrition and EDs may be paramount to improving outcomes.

Eating disorders (EDs) and feeding disorders refer to a wide spectrum of complex biopsychosocial illnesses. The spectrum of EDs encompasses anorexia nervosa (AN), bulimia nervosa (BN), binge eating disorder, and other specified feeding or eating disorders. Feeding disorders, distinguished from EDs based on the absence of body image disturbance, include pica, rumination syndrome, and avoidant/restrictive food intake disorder (ARFID).¹

This spectrum of illnesses predominantly affect young females aged 15 to 45 years, with recent increases in the rates of EDs among males, patients with skin of color, and adolescent females.^2-5 Patients with EDs are at an elevated lifetime risk of suicidal ideation, suicide attempts, and other psychiatric comorbidities compared to the general population.⁶ Specifically, AN and BN are associated with high psychiatric morbidity and mortality. A meta-analysis by Arcelus et al⁷ demonstrated the weighted annual mortality for AN was 5.10 deaths per 1000 person-years (95% CI, 3.57-7.59) among patients with EDs and 4.55 deaths for studies that selected inpatients (95% CI, 3.09-6.28); for BN, the weighted mortality was 1.74 deaths per 1000 person-years (95% CI, 1.09-2.44). Unfortunately, ED diagnoses often are delayed or missed in clinical settings. Patients may lack insight into the severity of their illness, experience embarrassment about their eating behaviors, or actively avoid treatment for their ED.⁸

Pica—compulsive eating of nonnutritive substances outside the cultural norm—and rumination syndrome—regurgitation of undigested food—are feeding disorders more commonly recognized in childhood.^9-11 Pregnancy, intellectual disability, iron deficiency, and lead poisoning are other conditions associated with pica.^6,9,10 Avoidant/restrictive food intake disorder, a new diagnosis added to the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5)¹ in 2013, is an eating or feeding disturbance resulting in persistent failure to meet nutritional or energy needs. Etiologies of ARFID may include sensory sensitivities and/or a traumatic event related to eating, leading to avoidance of associated foods.¹²

Patients with an ED or a feeding disorder frequently experience malnutrition, including deficiencies, excesses, or imbalances in nutritional intake, which may lead to nutritional dermatoses.¹³ As a result, the skin may present the first visible clues to an ED diagnosis.^8,14-19 Gupta et al¹⁸ organized the skin signs of EDs into 4 categories: (1) those secondary to starvation or malnutrition; (2) cutaneous injury related to self-induced vomiting; (3) dermatoses due to laxative, diuretic, or emetic use; and (4) other concomitant psychiatric illnesses (eg, hand dermatitis from compulsive handwashing, dermatodaxia, onychophagia, trichotillomania). This review will focus on the effects of malnutrition and starvation on the skin.

Skin findings in patients with EDs offer the treating dermatologist a special opportunity for early diagnosis and appropriate consultation with specialists trained in ED treatment. It is important for dermatologists to be vigilant in looking for skin findings of nutritional dermatoses, especially in populations at an increased risk for developing an ED, such as young female patients. The approach to therapy and treatment must occur through a collaborative multidisciplinary effort in a thoughtful and nonjudgmental environment.

Xerosis

Xerosis, or dry skin, is the most common dermatologic finding in both adult and pediatric patients with AN and BN.^14,19 It presents as skin roughness, tightness, flaking, and scaling, which may be complicated by fissuring, itching, and bleeding.²⁰ In healthy skin, moisture is maintained by the stratum corneum and its lipids such as ceramides, cholesterol, and free fatty acids.²¹ Natural moisturizing factor (NMF) within the skin is composed of amino acids, ammonia, urea, uric acid, inorganic salts, lactic acid derivatives, and pyrrolidine-3-carboxylic acid.^20-22 Disruptions to this system result in increased transepidermal water loss and impaired barrier function.²³

In patients with ED, xerosis arises through several mechanisms. Chronic illness or starvation can lead to euthyroid sick syndrome with decreased peripheral conversion of thyroxine (T₄) to triiodothyronine (T₃).^24,25 In the context of functional hypothyroidism, xerosis can arise from decreased eccrine gland secretion.²⁶ Secretions of water, lactate, urea, sodium, and potassium from eccrine glands help to maintain NMF for skin hydration.²⁷ Persistent laxative or diuretic abuse and fluid intake restriction, which are common behaviors across the spectrum of EDs, lead to dehydration and electrolyte imbalances that can manifest as skin dryness.²⁰ Disrupted keratinocyte differentiation due to insufficient stores of vitamins and minerals involved in keratinocyte differentiation, such as vitamins A and C, selenium, and zinc, also may contribute to xerosis.^25,28,29

Severely restrictive eating patterns may lead to development of protein energy malnutrition (PEM). Cutaneous findings in PEM occur due to dysmaturation of epidermal keratinocytes and epidermal atrophy.³⁰ Patients with severe persistent depletion of macronutrients—carbohydrates, fat, and protein—may experience marasmus, resulting in loss of subcutaneous fat that causes the appearance of dry loose skin.^29,31

Xerosis is exceedingly common in the general population and has no predictive value in ED diagnosis; however, this finding should be noted in the context of other signs suggestive of an ED. Treatment of xerosis in the setting of an ED should focus on correction of the underlying malnutrition. Symptomatic alleviation requires improving skin hydration and repairing barrier function. Mild xerosis may not need treatment or can be ameliorated with over-the-counter moisturizers and emollients. Scaling secondary to dry skin can be improved by ingredients such as glycerol, urea, lactic acid, and dexpanthenol.^20,32 Glycerol and urea are small hydrophilic molecules that penetrate the stratum corneum and help to bind moisture within the skin to reduce transepidermal water loss. Urea and lactic acid are keratolytics of NMF commonly found in moisturizers and emollients.^33,34 Dexpanthenol may be used for soothing fissures and pruritus; in vitro and in vivo studies have demonstrated its ability to upregulate dermal fibroblast proliferation and epidermal re-epithelization to promote faster wound healing.³⁵

Lanugo

Lanugo is clinically apparent as a layer of fine, minimally pigmented hair. It is physiologically present on the skin surface of fetuses and newborns. In utero, lanugo plays an essential role in fetal skin protection from amniotic fluid, as well as promotion of proper hydration, thermoregulation, and innate immune development.^36-38 Although it may be found on approximately 30% of newborns as normal variation, its presence beyond the neonatal period signals underlying systemic disease and severe undernutrition.^16,36,39 Rarely, hypertrichosis lanuginosa acquisita has been reported in association with malignancy.^40,41 The finding of lanugo beyond the neonatal period should prompt exclusion of other medical disorders, including neoplasms, chronic infections, hyperthyroidism, malabsorption syndromes, and inflammatory bowel disease.^41-47

There is a limited understanding of the pathomechanism behind lanugo development in the context of malnutrition. Intentional starvation leads to loss of subcutaneous fat and a state of functional hypothyroidism.⁴⁸ Studies hypothesize that lanugo develops as a response to hypothermia, regulated by dermal papillae cell–derived exosomes that may stimulate hair growth via paracrine signaling to outer root sheath cells.^36,49 Molecular studies have found that T₃ impacts skin and hair differentiation and proliferation by modulating thyroid hormone receptor regulation of keratin expression in epithelial cells.^50,51 Lanugo may be a clinical indicator of severe malnutrition among ED patients, especially children and adolescents. A study of 30 patients aged 8 to 17 years with AN and BN who underwent a standard dermatologic examination found significant positive correlation between the presence of lanugo hair growth and concomitant amenorrhea (P<.01) as well as between lanugo hair and body mass index lower than 16 kg/m² (P<.05).¹⁹ Discovery of lanugo in the dermatology clinical setting should prompt a thorough history, including screening questions about eating patterns; attitudes on eating, exercise, and appearance; personal and family history of EDs or other psychiatric disorders; and screening for depression and anxiety. Given its association with other signs of severe malnutrition, a clinical finding of lanugo should prompt close physical examination for other potential signs of an ED and laboratory evaluation for electrolyte levels and blood counts.⁵² Resolution of lanugo secondary to an ED is achieved with restoration of normal total body fat.¹⁸ Treatment should be focused on appropriate weight gain with the guidance of an ED specialist.

Pruritus

The prevalence and pathomechanism of pruritus secondary to EDs remains unclear.^16,53,54 There have been limited reports of pruritus secondary to ED, with Gupta et al⁵³ providing a case series of 6 patients with generalized pruritus in association with starvation and/or rapid weight loss. The study reported remission of pruritus with nutritional rehabilitation and/or weight gain of 5 to 10 pounds. Laboratory evaluation ruled out other causes of pruritus such as cholestasis and uremia.⁵³ Other case reports have associated pruritus with iron deficiency, with anecdotal evidence of pruritus resolution following iron supplementation.^55-59 Although we found no studies specifically relating iron deficiency, EDs, and pruritus, iron deficiency routinely is seen in ED patients and has a known association with pica.^9,10,60 As such, iron deficiency may be a contributing factor in pruritus in ED patients. A UK study of 19 women with AN and a body mass index lower than 16 kg/m² found that more than half of the patients (11/19 [57.9%]) described pruritus on the St. Thomas’ Itch Questionnaire, postulating that pruritus may be a clinical feature of AN.⁶¹ Limited studies with small samples make it difficult to conclude whether pruritus arises as a direct consequence of malnutrition.

Treatment of pruritus should address the underlying ED, as the pathophysiology of itch as it relates to malnutrition is poorly understood. Correction of existing nutritional imbalances by iron supplementation and appropriate weight gain may lead to symptom resolution. Because xerosis may be a contributing factor to pruritus, correction of the xerosis also may be therapeutic. More studies are needed on the connection between pruritus and the nutritional imbalances encountered in patients with EDs.

Acrocyanosis

Acrocyanosis is clinically seen as bluish-dusky discoloration most commonly affecting the hands and feet but also may affect the nose, ears, and nipples. Acrocyanosis typically is a sign of cold intolerance, hypothesized to occur in the context of AN due to shunting of blood centrally in response to hypothermia.^39,62 The diminished oxyhemoglobin delivery to extremity sites leads to the characteristic blue color.⁶³ In a study of 211 adolescent females (age range, 13–17 years) with AN, physical examination revealed peripheral hypothermia and peripheral cyanosis in 80% and 43% of patients, respectively.⁴⁸ Cold intolerance seen in EDs may be secondary to a functional hypothyroid state similar to euthyroid sick syndrome seen in conditions of severe caloric deficit.²⁵

It is possible that anemia and dehydration can worsen acrocyanosis due to impaired delivery of oxyhemoglobin to the body’s periphery.⁶³ In a study of 14 ED patients requiring inpatient care, 6 were found to have underlying anemia following intravenous fluid supplementation.⁶⁴ On admission, the mean (SD) hemoglobin and hematocrit across 14 patients was 12.74 (2.19) and 37.42 (5.99), respectively. Following intravenous fluid supplementation, the mean (SD) hemoglobin and hematocrit decreased to 9.88 (1.79)(P<.001) and 29.56 (4.91)(P=.008), respectively. Most cases reported intentional restriction of dietary sodium and fluid intake, with 2 patients reporting a history of diuretic misuse.⁶⁴ These findings demonstrate that hemoglobin and hematocrit may be falsely normal in patients with AN due to hemoconcentration, suggesting that anemia may be underdiagnosed in inpatients with AN.

Beyond treatment of the underlying ED, acrocyanosis therapy is focused on improvement of circulation and avoidance of exacerbating factors. Pharmacologic intervention rarely is needed. Patients should be reassured that acrocyanosis is a benign condition and often can be improved by dressing warmly and avoiding exposure to cold. Severe cases may warrant trial treatment with nicotinic acid derivatives, α-adrenergic blockade, and topical minoxidil, which have demonstrated limited benefit in treating primary idiopathic acrocyanosis.⁶³

Carotenoderma

Carotenoderma—the presence of a yellow discoloration to skin secondary to hypercarotenemia—has been described in patients with EDs since the 1960s.^65,66 Beyond its clinical appearance, carotenoderma is asymptomatic. Carotenoids are lipid-soluble compounds present in the diet that are metabolized by the intestinal mucosa and liver to the primary conversion product, retinaldehyde, which is further converted to retinol, retinyl esters, and other retinoid metabolites.^67,68 Retinol is bound by lipoproteins and transported in the plasma, then deposited in peripheral tissues,⁶⁹ including in intercellular lipids in the stratum corneum, resulting in an orange hue that is most apparent in sites of increased skin thickness and sweating (eg, palms, soles, nasolabial folds).⁷⁰ In an observational study of ED patients, Glorio et al¹⁴ found that carotenoderma was present in 23.77% (29/122) and 25% (4/16) of patients with BN and other specified feeding or eating disorder, respectively; it was not noted among patients with AN. Prior case reports have provided anecdotal evidence of carotenoderma in AN patients.^66,71 In the setting of an ED, increased serum carotenoids likely are due to increased ingestion of carotene-rich foods, leading to increased levels of carotenoid-bound lipoproteins in the serum.⁷⁰ Resolution of xanthoderma requires restriction of carotenoid intake and may take 2 to 3 months to be clinically apparent. The lipophilic nature of carotenoids allows storage in body fat, prolonging resolution.⁷¹

Hair Changes

Telogen effluvium (TE) and hair pigmentary changes are clinical findings that have been reported in association with EDs.^14,16,19,72 Telogen effluvium occurs when physiologic stress causes a large portion of hairs in the anagen phase of growth to prematurely shift into the catagen then telogen phase. Approximately 2 to 3 months following the initial insult, there is clinically apparent excessive hair shedding compared to baseline.⁷³ Studies have demonstrated that patients with EDs commonly have psychiatric comorbidities such as mood and anxiety disorders, obsessive compulsive disorder, posttraumatic stress disorder, and panic disorder compared to the general population.^6,74-76 As such, stress experienced by ED patients may contribute to TE. Despite TE being commonly reported in ED patients,^16-18 there is a lack of controlled studies of TE in human subjects with ED. An animal model for TE demonstrated that stressed mice exhibited further progression in the hair cycle compared with nonstressed mice (P<.01); the majority of hair follicles in stressed mice were in the catagen phase, while the majority of hair follicles in nonstressed mice were in the anagen phase.⁷⁷ Stressed mice demonstrated an increased number of major histocompatibility complex class II⁺ cell clusters, composed mostly of activated macrophages, per 12.5-mm epidermal length compared to nonstressed mice (mean [SEM], 7.0 [1.1] vs 2.0 [0.3][P<.05]). This study illustrated that stress can lead to inflammatory cell recruitment and activation in the hair follicle microenvironment with growth-inhibitory effects.⁷⁷

The flag sign, or alternating bands of lesser and greater pigmentation in the hair, has been reported in cases of severe PEM.³¹ In addition, PEM may lead to scalp alopecia, dry and brittle hair, and/or hypopigmentation with periods of inadequate nutrition.^29,78 Scalp hair hypopigmentation, brittleness, and alopecia have been reported in pediatric patients with highly selective eating and/or ARFID.^79,80 Maruo et al⁸⁰ described a 3-year-old boy with ASD who consumed only potato chips for more than a year. Physical examination revealed reduced skin turgor overall and sparse red-brown hair on the scalp; laboratory testing showed deficiencies of protein, vitamin A, vitamin D, copper, and zinc. The patient was admitted for nutritional rehabilitation via nasogastric tube feeding, leading to resolution of laboratory abnormalities and growth of thicker black scalp hair over the course of several months.⁸⁰

Neuroendocrine control of keratin expression by thyroid-stimulating hormone (TSH) and thyroid hormones likely plays a role in the regulation of hair follicle activities, including hair growth, structure, and stem cell differentiation.^81,82 Altered thyroid hormone activity, which commonly is seen in patients with EDs,^24,25 may contribute to impaired hair growth and pigmentation.^26,51,83-85 Using tissue cultures of human anagen hair follicles, van Beek et al⁸⁵ provided in vitro evidence that T₃ and T₄ modulate scalp hair follicle growth and pigmentation. Both T₃- and T₄-treated tissue exhibited increased numbers of anagen and decreased numbers of catagen hair follicles in organ cultures compared with control (P<.01); on quantitative Fontana-Masson histochemistry, T₃ and T₄ significantly stimulated hair follicle melanin synthesis compared with control (P<.001 and P<.01, respectively).⁸⁵ Molecular studies by Bodó et al⁸³ have shown that the human scalp epidermis expresses TSH at the messenger RNA and protein levels. Both studies showed that intraepidermal TSH expression is downregulated by thyroid hormones.^83,85 Further studies are needed to examine the impact of malnutrition on local thyroid hormone signaling and action at the level of the dermis, epidermis, and hair follicle.

Discovery of TE, hair loss, and/or hair hypopigmentation should prompt close investigation for other signs of thyroid dysfunction, specifically secondary to malnutrition. Imbalances in TSH, T₃, and T₄ should be corrected. Nutritional deficiencies and dietary habits should be addressed through careful nutritional rehabilitation and targeted ED treatment.

Oral and Mucosal Symptoms

Symptoms of the oral cavity that may arise secondary to EDs and feeding disorders include glossitis, stomatitis, cheilitis, and dental erosions. Mucosal symptoms have been observed in patients with vitamin B deficiencies, inflammatory bowel disease, and other malabsorptive disorders, including patients with EDs.^86-88 Patients following restrictive diets, specifically strict vegan diets, without additional supplementation are at risk for developing vitamin B₁₂ deficiency. Because vitamin B₁₂ is stored in the liver, symptoms of deficiency appear when hepatic stores are depleted over the course of several years.⁸⁹ Insufficient vitamin B₁₂ prevents the proper functioning of methionine synthase, which is required for the conversion of homocysteine to methionine and for the conversion of methyl-tetrahydrofolate to tetrahydrofolate.⁸⁹ Impairment of this process impedes the synthesis of pyrimidine bases of DNA, disrupting the production of rapidly proliferating cells such as myeloid cells or mucosal lining cells. In cases of glossitis and/or stomatitis due to vitamin B₁₂ deficiency, resolution of lesions was achieved within 4 weeks of daily oral supplementation with vitamin B₁₂ at 2 μg daily.^90,91 Iron deficiency, a common finding in EDs, also may contribute to glossitis and angular cheilitis.²⁹ If uncovered, iron deficiency should be corrected by supplementation based on total deficit, age, and sex. Oral supplementation may be done with oral ferrous sulfate (325 mg provides 65 mg elemental iron) or with other iron salts such as ferrous gluconate (325 mg provides 38 mg elemental iron).²⁹ Mucosal symptoms of cheilitis and labial erythema may arise from irritation due to self-induced vomiting.⁸⁸

Dental erosion refers to loss of tooth structure via a chemical process that does not involve bacteria; in contrast, dental caries refer to tooth damage secondary to bacterial acid production. Patients with EDs who repeatedly self-induce vomiting have persistent introduction of gastric acids into the oral cavity, resulting in dissolution of the tooth enamel, which occurs when teeth are persistently exposed to a pH less than 5.5.⁹² Feeding disorders also may predispose patients to dental pathology. In a study of 60 pediatric patients, those with rumination syndrome were significantly more likely to have dental erosions than age- and sex-matched healthy controls (23/30 [77%] vs 4/30 [13%][P<.001]). The same study found no difference in the frequency of dental caries between children with and without rumination syndrome.⁹² These findings suggest that rumination syndrome increases the risk for dental erosions but not dental caries. The distribution of teeth affected by dental erosions may differ between EDs and feeding disorders. Patients with BN are more likely to experience involvement of the palatal surfaces of maxillary teeth, while patients with rumination syndrome had equal involvement of maxillary and mandibular teeth.⁹²

There is limited literature on the role of dentists in the care of patients with EDs and feeding disorders, though existing studies suggest inclusion of a dental care professional in multidisciplinary treatment along with emphasis on education around a home dental care regimen and frequent dental follow-up.^76,93,94 Prevention of further damage requires correction of the underlying behaviors and ED.

Other Dermatologic Findings

Russell sign refers to the development of calluses on the dorsal metacarpophalangeal joints of the dominant hand due to self-induced vomiting. Due to its specificity in purging-type EDs, the discovery of Russell sign should greatly increase suspicion for an ED.¹⁷ Patients with EDs also are at an increased risk for self-harming and body-focused repetitive behaviors, including skin cutting, superficial burning, onychophagia, and trichotillomania.¹⁹ It is important to recognize these signs in patients for whom an ED is suspected. The role of the dermatologist should include careful examination of the skin and documentation of findings that may aid in the diagnosis of an underlying ED.

Final Thoughts

A major limitation of this review is the reliance on small case reports and case series reporting cutaneous manifestations of ED. Controlled studies with larger cohorts are challenging in this population but are needed to substantiate the dermatologic signs commonly associated with EDs. Translational studies may help elucidate the pathomechanisms underlying dermatologic diseases such as lanugo, pruritus, and alopecia in the context of EDs and malnutrition. The known association between thyroid dysfunction and skin disease has been substantiated by clinical and basic science investigation, suggesting a notable role of thyroid hormone and TSH signaling in the skin local environment. Further investigation into nutritional and neuroendocrine regulation of skin health will aid in the diagnosis and treatment of patients impacted by EDs.

The treatment of the underlying ED is key in correcting associated skin disease, which requires interdisciplinary collaboration that addresses the psychological, behavioral, and social components of the condition. Following a diagnosis of ED, assessment should be made of the nutritional rehabilitation required to restore weight and nutritional status. Inpatient treatment may be indicated for patients requiring close monitoring to avoid refeeding syndrome, or those who meet the criteria for extreme AN in the DSM-5 (ie, body mass index <15 kg/m²),¹ or demonstrate signs of medical instability or organ failure secondary to malnutrition.⁶² Long-term recovery for ED patients should focus on behavioral therapy with a multidisciplinary team consisting of a psychiatrist, therapist, dietitian, and primary care provider. Comparative studies in large-scale trials of cognitive behavioral therapy, focal psychodynamic psychotherapy, and specialist supportive clinical management have shown little to no difference in efficacy in treating EDs.^75,95,96

Dermatologists may be the first providers to observe sequelae of nutritional and behavioral derangement in patients with EDs. Existing literature on the dermatologic findings of EDs report great heterogeneity of skin signs, with a very limited number of controlled studies available. Each cutaneous symptom described in this review should not be interpreted as an isolated pathology but should be placed in the context of patient predisposing risk factors and the constellation of other skin findings that may be suggestive of disordered eating behavior or other psychiatric illness. The observation of multiple signs and symptoms at the same time, especially of symptoms uncommonly encountered or suggestive of a severe and prolonged imbalance (eg, xanthoderma with vitamin A excess, aphthous stomatitis with vitamin B deficiency), should heighten clinical suspicion for an underlying ED. A clinician’s highest priority should be to resolve life-threatening medical emergencies and address nutritional derangements with the assistance of experts who are well versed in EDs. The patient should undergo workup to rule out organic causes of their nutritional dermatoses. Given the high psychiatric morbidity and mortality of patients with an ED and the demonstrated benefit of early intervention, recognition of cutaneous manifestations of malnutrition and EDs may be paramount to improving outcomes.