Article

Clinicians usually measure renal function by using surrogate markers because directly measuring glomerular filtration rate (GFR) is not routinely feasible in a clinical setting.^1,2 Creatinine (Cr) and cystatin C (CysC) are the 2 main surrogate molecules used to estimate GFR.³

Creatine is a molecule nonenzymatically converted into Cr, weighing only 113 Da in skeletal muscles.⁴ It is then filtered at the glomeruli and secreted at the proximal tubules of the kidneys. However, serum Cr (sCr) levels are affected by several factors, including age, biological sex, liver function, diet, and muscle mass.⁵ Historically, sCr levels also are affected by race.⁵ In an early study of factors affecting accurate GFR, researchers reported that self-identified African American patients had a 16% higher GFR than those who did not when using Cr.⁶ Despite this, the inclusion of Cr on a basic metabolic panel has allowed automatic reporting of an estimated GFR using sCr (eGFR_Cr) to be readily available.⁷

In comparison to Cr, CysC is an endogenous protein weighing 13 kDa produced by all nucleated cells.^8,9 CysC is filtered by the kidney at the glomeruli and completely reabsorbed and catabolized by epithelial cells at the proximal tubule.⁹ Since production is not dependent on skeletal muscle, there are fewer physiological impacts on serum concentration of CysC. Levels of CysC may be elevated by factors shown in the Table.

Estimating Glomerular Filtration Rates

Multiple equations were developed to mitigate the impact of extraneous factors on the accuracy of an eGFR_Cr. The first widely used equation that included a variable adjustment for race was the Modification of Diet in Renal Disease study, presented in 2006.¹⁰ The equation increased the accuracy of eGFR_Cr further by adjusting for sex and age. It was followed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation in 2009, which was more accurate at higher GFR levels.¹¹

CysC was simultaneously studied as an alternative to Cr with multiple equation iterations shown to be viable in various populations as early as 2003.^12-15 However, it was not until 2012 that an equation for the use of CysC was offered for widespread use as an alternative to Cr alongside further refinement of the CKD-EPI equation for Cr.¹⁶ A new formula was presented in 2021 to use both sCr and serum CysC levels to obtain a more accurate estimation of GFR.¹⁷ Research continues its effort to accurately estimate GFR for diagnosing kidney disease and assessing comorbidities relating to decreased kidney function.³

All historical equations attempted to mitigate the potential impact of race on sCr level when calculating eGFR_Crby assigning a separate variable for African American patients. As an unintended adverse effect, these equations may have led to discrimination by having a different equation for African American patients.¹⁸ Moreover, these Cr-based equations remain less accurate in patients with varied muscle mass, such as older patients, bodybuilders, athletes, and individuals with varied extremes of daily protein intake.^1,8,9,19Several medications can also directly affect Cr clearance, reducing its ability to act as a surrogate for kidney function.¹In this case report, we discuss an African American patient with high muscle mass and protein intake who was initially diagnosed with kidney disease based on an elevated Cr and found to be misdiagnosed based on the use of CysC for a more accurate GFR estimation.

A 35-year-old African American man serving in the military and recently diagnosed with HIV was referred to a nephrology clinic for further evaluation of an acute elevation in sCr. Before treatment for HIV, a brief record review showed a baseline Cr of about 1.3 mg/dL, with an eGFR_Cr of 75 mL/min/1.73 m².²⁰ In the same month, the patient was prescribed bictegravir/emtricitabine/tenofovir alafenamide, an HIV drug with nephrotoxic potential.²¹ The patient's total viral load remained low, and CD4 count remained > 500 after initiation of the HIV treatment. He was in his normal state of health and had no known contributory history before his HIV diagnosis. Cr readings peaked at 1.83 mg/dL after starting the HIV treatment and remained elevated to 1.73 mg/dL over the next few months, corresponding to CKD stage 3A. Because bictegravir/emtricitabine/tenofovir alafenamide is cleared by the kidneys and has a nephrotoxic profile, the clinical care team considered dosage adjustment or a medication switch given his observed elevated eGFR_Cr based on the CKD-EPI 2021 equation for Cr alone. It was also noted that the patient had a similar Cr spike to 1.83 mg/dL in 2018 without any identifiable renal insult or symptoms (Figure).

Diagnostic Evaluation

The primary care team ordered a renal ultrasound and referred the patient to the nephrology clinic. The nephrologist ordered the following laboratory studies: urine microalbumin to Cr ratio, basic metabolic panel (BMP), comprehensive metabolic panel (CMP), urinalysis, urine protein, urine Cr, parathyroid hormone level, hemoglobin A_1c, complement component 3/4 panels, antinuclear and antineutrophil cytoplasmic antibodies titers, glomerular basement membrane antibody titer, urine light chains, serum protein electrophoresis, κ/λ ratio, viral hepatitis panel, and rapid plasma reagin testing. Much of this laboratory evaluation served to rule out any secondary causes of kidney disease, including autoimmune disease, monoclonal or polyclonal gammopathies, diabetic nephropathy or glomerulosclerosis, and nephrotic or nephritic syndromes.

All laboratory studies returned within normal limits; no proteinuria was discovered on urinalysis, and no abnormalities were visualized on renal ultrasound. Bictegravir/emtricitabine/tenofovir alafenamide nephrotoxicity was highest among the differential diagnoses due to the timing of Cr elevation coinciding with the initiation of the medications. The patient's CysC level was 0.85 mg/dL with a calculated eGFR_Cys of 125 mL/min/1.73 m². The calculated sCR and serum cystatin C (eGFR_Cr-Cys) using the new 2021 equation and when adjusting for body surface area placed his eGFR at 92 mL/min/1.73 m².²⁰

The patient’s eGFR_Cysreassured the care team that the patient’s renal function was not acutely or chronically impacted by bictegravir/emtricitabine/tenofovir alafenamide, resulting in avoidance of unnecessary dosage adjustment or discontinuation of the HIV treatment. The patient reported a chronic habit of protein and creatine supplementation and bodybuilding, which likely further compounded the discrepancy between eGFR_Cr and eGFR_Cys and explained his previous elevation in Cr in 2018.

Follow-up

The patient underwent serial monitoring that revealed a stable Cr and unremarkable eGFR, ruling out CKD. There has been no evidence of worsening kidney disease to date, and the patient remained on his initial HIV regimen.

This case shows the importance of using CysC as an alternative or confirmatory marker compared with sCr to estimate GFR in patients with high muscle mass and/or high creatine intake, such as many in the US Department of Defense (DoD) and US Department of Veterans Affairs (VA) patient populations. In the presented case, recorded Cr levels climbed from baseline Cr with the initiation of bictegravir/emtricitabine/tenofovir alafenamide. This raised the concern that HIV treatment was leading to the development of kidney damage.²²

Diagnosis of kidney disease as opposed to the normal decline of eGFR with age in individuals without intrinsic CKD requires GFR ≥ 60 mL/min/1.73 m² with kidney damage (proteinuria or radiological abnormalities, etc) or GFR < 135 to 140 mL/min/1.73 m²minus the patient’s age in years.²³ The patient’s Cr peak at 1.83 mg/dL in 2018 led to an inappropriate diagnosis of kidney disease stage 3a based on an eGFR_Cr (2021 equation) of 52 mL/min/1.73 m² when not corrected for body surface area.²⁰ However, using the new 2021 equation using both Cr and CysC, the patient’s eGFR_Cr-Cyswas 92 mL/min/1.73 m² after a correction for body surface area.

The 2009 CKD-EPI recommended the calculation of eGFR based on SCr concentration using age, sex, and race while the 2021 CKD-EPI recommended the exclusion of race.³ Both equations are less accurate in African American patients, individuals taking medications that interfere with Cr secretion and assay, and patients taking creatine supplements, high daily protein intake, or with high muscle mass.⁷ These settings result in a decreased eGFR_Cr without corresponding eGFR_Cys changes. Using SCr and CysC together, the eGFR_Cr-Cys yields improved concordance to measured GFR across race groups compared to GFR estimation based on Cr alone, which can avoid unnecessary expensive diagnostic workup, inappropriate kidney disease diagnosis, incorrect dosing of drugs, and accurately represent the military readiness of patients. Interestingly, in African American patients with recently diagnosed HIV, CKD-EPI using both Cr and CysC without race inclusion led to only a 2.9% overestimation of GFR and was the only equation with no statistically significant bias compared with measured GFR.²⁴

A March 2023 case involving an otherwise healthy 26-year-old male active-duty US Navy member with a history of excessive protein supplement intake and intense exercise < 24 hours before laboratory work was diagnosed with CKD after a measured Cr of 16 mg/dL and an eGFR_Cr of 4 mL/min/1.73 m² without any other evidence of kidney disease. His CysC remained within normal limits, resulting in a normal eGFR_Cys of 121 mL/min/1.73 m², indicating no CKD. His Cr and eGFR recovered 10 days after his clinic visit and cessation of his supplement intake. These findings may not be uncommon given that 65% of active-duty military use protein supplements and 38% use other performance-enhancing supplements, such as creatine, according to a study.²⁵

Unfortunately, the BMP/CMP traditionally used at VA centers use the eGFR_Cr equation, and it is unknown how many primary care practitioners recognize the limitations of these metabolic panels on accurate estimation of kidney function. However, in 2022 an expert panel including VA physicians recommended the immediate use of eGFR_Cr-Cys or eGFR_Cys for confirmatory testing and potentially screening of CKD.²⁶ A small number of VAs have since adopted this recommendation, which should lead to fewer misdiagnoses among US military members as clinicians should now have access to more accurate measurements of GFR.

The VA spends about $18 billion (excluding dialysis) for care for 1.1 to 2.5 million VA patients with CKD.²⁷ The majority of these diagnoses were undoubtedly made using the eGFR_Cr equation, raising the question of how many may be misdiagnosed. Assessment with CysC is currently relatively expensive, but it will likely become more affordable as the use of CysC as a confirmatory test increases.⁵ The cost of a sCr test is about $2.50, while CysC costs about $10.60, with variation from laboratory to laboratory.²⁸ By comparison, a renal ultrasound costs $99 to $140 for uninsured patients.²⁹ Furthermore, the cost of CysC testing is likely to trend downward as more facilities adopt the use of CysC measurements, which can be run on the same analytical equipment currently used for Cr measurements. Currently, most laboratories do not have established assays to use in-house and thus require CysC to be sent out to a laboratory, which increases result time and makes Cr a more attractive option. As more laboratories adopt assays for CysC, the cost of reagents will further decrease.

Given such considerations, confirmation testing of kidney function with CysC in specific patient populations with decreased eGFR_Cr without other features of CKD can offer great medical and financial benefits. A 2023 KDIGO report noted that many individuals may be mistakenly diagnosed with CKD when using eGFR_Cr.³ KDIGO noted that a 2013 meta-analysis of 90,000 individuals found that with a Cr-based eGFR of 45 to 59 mL/min/1.73 m² (42%) had a CysC-based eGFR of ≥ 60 mL/min/1.73 m². An eGFR_Cr of 45 to 59 represents 54% of all patients with CKD, amounting to millions of people (including current and former military personnel).^3,29-31 Correcting a misdiagnosis of CKD would bring significant relief to patients and save millions in health care spending.

Conclusions

In patients who meet CKD criteria using eGFR_Cr but without other features of CKD, we recommend using confirmatory CysC levels and the eGFR_Cr-Cys equation. This will align care with the KDIGO guidelines and could be a cost-effective step toward improving military patient care. Further work in this area should focus on determining the knowledge gaps in primary care practitioners’ understanding of the limits of eGFR_Cr, the potential mitigation of concomitant CysC testing in equivocal CKD cases, and the cost-effectiveness and increased utilization of CysC.

Clinicians usually measure renal function by using surrogate markers because directly measuring glomerular filtration rate (GFR) is not routinely feasible in a clinical setting.^1,2 Creatinine (Cr) and cystatin C (CysC) are the 2 main surrogate molecules used to estimate GFR.³

Creatine is a molecule nonenzymatically converted into Cr, weighing only 113 Da in skeletal muscles.⁴ It is then filtered at the glomeruli and secreted at the proximal tubules of the kidneys. However, serum Cr (sCr) levels are affected by several factors, including age, biological sex, liver function, diet, and muscle mass.⁵ Historically, sCr levels also are affected by race.⁵ In an early study of factors affecting accurate GFR, researchers reported that self-identified African American patients had a 16% higher GFR than those who did not when using Cr.⁶ Despite this, the inclusion of Cr on a basic metabolic panel has allowed automatic reporting of an estimated GFR using sCr (eGFR_Cr) to be readily available.⁷

In comparison to Cr, CysC is an endogenous protein weighing 13 kDa produced by all nucleated cells.^8,9 CysC is filtered by the kidney at the glomeruli and completely reabsorbed and catabolized by epithelial cells at the proximal tubule.⁹ Since production is not dependent on skeletal muscle, there are fewer physiological impacts on serum concentration of CysC. Levels of CysC may be elevated by factors shown in the Table.

Estimating Glomerular Filtration Rates

Multiple equations were developed to mitigate the impact of extraneous factors on the accuracy of an eGFR_Cr. The first widely used equation that included a variable adjustment for race was the Modification of Diet in Renal Disease study, presented in 2006.¹⁰ The equation increased the accuracy of eGFR_Cr further by adjusting for sex and age. It was followed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation in 2009, which was more accurate at higher GFR levels.¹¹

CysC was simultaneously studied as an alternative to Cr with multiple equation iterations shown to be viable in various populations as early as 2003.^12-15 However, it was not until 2012 that an equation for the use of CysC was offered for widespread use as an alternative to Cr alongside further refinement of the CKD-EPI equation for Cr.¹⁶ A new formula was presented in 2021 to use both sCr and serum CysC levels to obtain a more accurate estimation of GFR.¹⁷ Research continues its effort to accurately estimate GFR for diagnosing kidney disease and assessing comorbidities relating to decreased kidney function.³

All historical equations attempted to mitigate the potential impact of race on sCr level when calculating eGFR_Crby assigning a separate variable for African American patients. As an unintended adverse effect, these equations may have led to discrimination by having a different equation for African American patients.¹⁸ Moreover, these Cr-based equations remain less accurate in patients with varied muscle mass, such as older patients, bodybuilders, athletes, and individuals with varied extremes of daily protein intake.^1,8,9,19Several medications can also directly affect Cr clearance, reducing its ability to act as a surrogate for kidney function.¹In this case report, we discuss an African American patient with high muscle mass and protein intake who was initially diagnosed with kidney disease based on an elevated Cr and found to be misdiagnosed based on the use of CysC for a more accurate GFR estimation.

A 35-year-old African American man serving in the military and recently diagnosed with HIV was referred to a nephrology clinic for further evaluation of an acute elevation in sCr. Before treatment for HIV, a brief record review showed a baseline Cr of about 1.3 mg/dL, with an eGFR_Cr of 75 mL/min/1.73 m².²⁰ In the same month, the patient was prescribed bictegravir/emtricitabine/tenofovir alafenamide, an HIV drug with nephrotoxic potential.²¹ The patient's total viral load remained low, and CD4 count remained > 500 after initiation of the HIV treatment. He was in his normal state of health and had no known contributory history before his HIV diagnosis. Cr readings peaked at 1.83 mg/dL after starting the HIV treatment and remained elevated to 1.73 mg/dL over the next few months, corresponding to CKD stage 3A. Because bictegravir/emtricitabine/tenofovir alafenamide is cleared by the kidneys and has a nephrotoxic profile, the clinical care team considered dosage adjustment or a medication switch given his observed elevated eGFR_Cr based on the CKD-EPI 2021 equation for Cr alone. It was also noted that the patient had a similar Cr spike to 1.83 mg/dL in 2018 without any identifiable renal insult or symptoms (Figure).

Diagnostic Evaluation

The primary care team ordered a renal ultrasound and referred the patient to the nephrology clinic. The nephrologist ordered the following laboratory studies: urine microalbumin to Cr ratio, basic metabolic panel (BMP), comprehensive metabolic panel (CMP), urinalysis, urine protein, urine Cr, parathyroid hormone level, hemoglobin A_1c, complement component 3/4 panels, antinuclear and antineutrophil cytoplasmic antibodies titers, glomerular basement membrane antibody titer, urine light chains, serum protein electrophoresis, κ/λ ratio, viral hepatitis panel, and rapid plasma reagin testing. Much of this laboratory evaluation served to rule out any secondary causes of kidney disease, including autoimmune disease, monoclonal or polyclonal gammopathies, diabetic nephropathy or glomerulosclerosis, and nephrotic or nephritic syndromes.

All laboratory studies returned within normal limits; no proteinuria was discovered on urinalysis, and no abnormalities were visualized on renal ultrasound. Bictegravir/emtricitabine/tenofovir alafenamide nephrotoxicity was highest among the differential diagnoses due to the timing of Cr elevation coinciding with the initiation of the medications. The patient's CysC level was 0.85 mg/dL with a calculated eGFR_Cys of 125 mL/min/1.73 m². The calculated sCR and serum cystatin C (eGFR_Cr-Cys) using the new 2021 equation and when adjusting for body surface area placed his eGFR at 92 mL/min/1.73 m².²⁰

The patient’s eGFR_Cysreassured the care team that the patient’s renal function was not acutely or chronically impacted by bictegravir/emtricitabine/tenofovir alafenamide, resulting in avoidance of unnecessary dosage adjustment or discontinuation of the HIV treatment. The patient reported a chronic habit of protein and creatine supplementation and bodybuilding, which likely further compounded the discrepancy between eGFR_Cr and eGFR_Cys and explained his previous elevation in Cr in 2018.

Follow-up

The patient underwent serial monitoring that revealed a stable Cr and unremarkable eGFR, ruling out CKD. There has been no evidence of worsening kidney disease to date, and the patient remained on his initial HIV regimen.

This case shows the importance of using CysC as an alternative or confirmatory marker compared with sCr to estimate GFR in patients with high muscle mass and/or high creatine intake, such as many in the US Department of Defense (DoD) and US Department of Veterans Affairs (VA) patient populations. In the presented case, recorded Cr levels climbed from baseline Cr with the initiation of bictegravir/emtricitabine/tenofovir alafenamide. This raised the concern that HIV treatment was leading to the development of kidney damage.²²

Diagnosis of kidney disease as opposed to the normal decline of eGFR with age in individuals without intrinsic CKD requires GFR ≥ 60 mL/min/1.73 m² with kidney damage (proteinuria or radiological abnormalities, etc) or GFR < 135 to 140 mL/min/1.73 m²minus the patient’s age in years.²³ The patient’s Cr peak at 1.83 mg/dL in 2018 led to an inappropriate diagnosis of kidney disease stage 3a based on an eGFR_Cr (2021 equation) of 52 mL/min/1.73 m² when not corrected for body surface area.²⁰ However, using the new 2021 equation using both Cr and CysC, the patient’s eGFR_Cr-Cyswas 92 mL/min/1.73 m² after a correction for body surface area.

The 2009 CKD-EPI recommended the calculation of eGFR based on SCr concentration using age, sex, and race while the 2021 CKD-EPI recommended the exclusion of race.³ Both equations are less accurate in African American patients, individuals taking medications that interfere with Cr secretion and assay, and patients taking creatine supplements, high daily protein intake, or with high muscle mass.⁷ These settings result in a decreased eGFR_Cr without corresponding eGFR_Cys changes. Using SCr and CysC together, the eGFR_Cr-Cys yields improved concordance to measured GFR across race groups compared to GFR estimation based on Cr alone, which can avoid unnecessary expensive diagnostic workup, inappropriate kidney disease diagnosis, incorrect dosing of drugs, and accurately represent the military readiness of patients. Interestingly, in African American patients with recently diagnosed HIV, CKD-EPI using both Cr and CysC without race inclusion led to only a 2.9% overestimation of GFR and was the only equation with no statistically significant bias compared with measured GFR.²⁴

A March 2023 case involving an otherwise healthy 26-year-old male active-duty US Navy member with a history of excessive protein supplement intake and intense exercise < 24 hours before laboratory work was diagnosed with CKD after a measured Cr of 16 mg/dL and an eGFR_Cr of 4 mL/min/1.73 m² without any other evidence of kidney disease. His CysC remained within normal limits, resulting in a normal eGFR_Cys of 121 mL/min/1.73 m², indicating no CKD. His Cr and eGFR recovered 10 days after his clinic visit and cessation of his supplement intake. These findings may not be uncommon given that 65% of active-duty military use protein supplements and 38% use other performance-enhancing supplements, such as creatine, according to a study.²⁵

Unfortunately, the BMP/CMP traditionally used at VA centers use the eGFR_Cr equation, and it is unknown how many primary care practitioners recognize the limitations of these metabolic panels on accurate estimation of kidney function. However, in 2022 an expert panel including VA physicians recommended the immediate use of eGFR_Cr-Cys or eGFR_Cys for confirmatory testing and potentially screening of CKD.²⁶ A small number of VAs have since adopted this recommendation, which should lead to fewer misdiagnoses among US military members as clinicians should now have access to more accurate measurements of GFR.

The VA spends about $18 billion (excluding dialysis) for care for 1.1 to 2.5 million VA patients with CKD.²⁷ The majority of these diagnoses were undoubtedly made using the eGFR_Cr equation, raising the question of how many may be misdiagnosed. Assessment with CysC is currently relatively expensive, but it will likely become more affordable as the use of CysC as a confirmatory test increases.⁵ The cost of a sCr test is about $2.50, while CysC costs about $10.60, with variation from laboratory to laboratory.²⁸ By comparison, a renal ultrasound costs $99 to $140 for uninsured patients.²⁹ Furthermore, the cost of CysC testing is likely to trend downward as more facilities adopt the use of CysC measurements, which can be run on the same analytical equipment currently used for Cr measurements. Currently, most laboratories do not have established assays to use in-house and thus require CysC to be sent out to a laboratory, which increases result time and makes Cr a more attractive option. As more laboratories adopt assays for CysC, the cost of reagents will further decrease.

Given such considerations, confirmation testing of kidney function with CysC in specific patient populations with decreased eGFR_Cr without other features of CKD can offer great medical and financial benefits. A 2023 KDIGO report noted that many individuals may be mistakenly diagnosed with CKD when using eGFR_Cr.³ KDIGO noted that a 2013 meta-analysis of 90,000 individuals found that with a Cr-based eGFR of 45 to 59 mL/min/1.73 m² (42%) had a CysC-based eGFR of ≥ 60 mL/min/1.73 m². An eGFR_Cr of 45 to 59 represents 54% of all patients with CKD, amounting to millions of people (including current and former military personnel).^3,29-31 Correcting a misdiagnosis of CKD would bring significant relief to patients and save millions in health care spending.

Conclusions

In patients who meet CKD criteria using eGFR_Cr but without other features of CKD, we recommend using confirmatory CysC levels and the eGFR_Cr-Cys equation. This will align care with the KDIGO guidelines and could be a cost-effective step toward improving military patient care. Further work in this area should focus on determining the knowledge gaps in primary care practitioners’ understanding of the limits of eGFR_Cr, the potential mitigation of concomitant CysC testing in equivocal CKD cases, and the cost-effectiveness and increased utilization of CysC.

Clinicians usually measure renal function by using surrogate markers because directly measuring glomerular filtration rate (GFR) is not routinely feasible in a clinical setting.^1,2 Creatinine (Cr) and cystatin C (CysC) are the 2 main surrogate molecules used to estimate GFR.³

Creatine is a molecule nonenzymatically converted into Cr, weighing only 113 Da in skeletal muscles.⁴ It is then filtered at the glomeruli and secreted at the proximal tubules of the kidneys. However, serum Cr (sCr) levels are affected by several factors, including age, biological sex, liver function, diet, and muscle mass.⁵ Historically, sCr levels also are affected by race.⁵ In an early study of factors affecting accurate GFR, researchers reported that self-identified African American patients had a 16% higher GFR than those who did not when using Cr.⁶ Despite this, the inclusion of Cr on a basic metabolic panel has allowed automatic reporting of an estimated GFR using sCr (eGFR_Cr) to be readily available.⁷

In comparison to Cr, CysC is an endogenous protein weighing 13 kDa produced by all nucleated cells.^8,9 CysC is filtered by the kidney at the glomeruli and completely reabsorbed and catabolized by epithelial cells at the proximal tubule.⁹ Since production is not dependent on skeletal muscle, there are fewer physiological impacts on serum concentration of CysC. Levels of CysC may be elevated by factors shown in the Table.

Estimating Glomerular Filtration Rates

Multiple equations were developed to mitigate the impact of extraneous factors on the accuracy of an eGFR_Cr. The first widely used equation that included a variable adjustment for race was the Modification of Diet in Renal Disease study, presented in 2006.¹⁰ The equation increased the accuracy of eGFR_Cr further by adjusting for sex and age. It was followed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation in 2009, which was more accurate at higher GFR levels.¹¹

CysC was simultaneously studied as an alternative to Cr with multiple equation iterations shown to be viable in various populations as early as 2003.^12-15 However, it was not until 2012 that an equation for the use of CysC was offered for widespread use as an alternative to Cr alongside further refinement of the CKD-EPI equation for Cr.¹⁶ A new formula was presented in 2021 to use both sCr and serum CysC levels to obtain a more accurate estimation of GFR.¹⁷ Research continues its effort to accurately estimate GFR for diagnosing kidney disease and assessing comorbidities relating to decreased kidney function.³

All historical equations attempted to mitigate the potential impact of race on sCr level when calculating eGFR_Crby assigning a separate variable for African American patients. As an unintended adverse effect, these equations may have led to discrimination by having a different equation for African American patients.¹⁸ Moreover, these Cr-based equations remain less accurate in patients with varied muscle mass, such as older patients, bodybuilders, athletes, and individuals with varied extremes of daily protein intake.^1,8,9,19Several medications can also directly affect Cr clearance, reducing its ability to act as a surrogate for kidney function.¹In this case report, we discuss an African American patient with high muscle mass and protein intake who was initially diagnosed with kidney disease based on an elevated Cr and found to be misdiagnosed based on the use of CysC for a more accurate GFR estimation.

A 35-year-old African American man serving in the military and recently diagnosed with HIV was referred to a nephrology clinic for further evaluation of an acute elevation in sCr. Before treatment for HIV, a brief record review showed a baseline Cr of about 1.3 mg/dL, with an eGFR_Cr of 75 mL/min/1.73 m².²⁰ In the same month, the patient was prescribed bictegravir/emtricitabine/tenofovir alafenamide, an HIV drug with nephrotoxic potential.²¹ The patient's total viral load remained low, and CD4 count remained > 500 after initiation of the HIV treatment. He was in his normal state of health and had no known contributory history before his HIV diagnosis. Cr readings peaked at 1.83 mg/dL after starting the HIV treatment and remained elevated to 1.73 mg/dL over the next few months, corresponding to CKD stage 3A. Because bictegravir/emtricitabine/tenofovir alafenamide is cleared by the kidneys and has a nephrotoxic profile, the clinical care team considered dosage adjustment or a medication switch given his observed elevated eGFR_Cr based on the CKD-EPI 2021 equation for Cr alone. It was also noted that the patient had a similar Cr spike to 1.83 mg/dL in 2018 without any identifiable renal insult or symptoms (Figure).

Diagnostic Evaluation

The primary care team ordered a renal ultrasound and referred the patient to the nephrology clinic. The nephrologist ordered the following laboratory studies: urine microalbumin to Cr ratio, basic metabolic panel (BMP), comprehensive metabolic panel (CMP), urinalysis, urine protein, urine Cr, parathyroid hormone level, hemoglobin A_1c, complement component 3/4 panels, antinuclear and antineutrophil cytoplasmic antibodies titers, glomerular basement membrane antibody titer, urine light chains, serum protein electrophoresis, κ/λ ratio, viral hepatitis panel, and rapid plasma reagin testing. Much of this laboratory evaluation served to rule out any secondary causes of kidney disease, including autoimmune disease, monoclonal or polyclonal gammopathies, diabetic nephropathy or glomerulosclerosis, and nephrotic or nephritic syndromes.

All laboratory studies returned within normal limits; no proteinuria was discovered on urinalysis, and no abnormalities were visualized on renal ultrasound. Bictegravir/emtricitabine/tenofovir alafenamide nephrotoxicity was highest among the differential diagnoses due to the timing of Cr elevation coinciding with the initiation of the medications. The patient's CysC level was 0.85 mg/dL with a calculated eGFR_Cys of 125 mL/min/1.73 m². The calculated sCR and serum cystatin C (eGFR_Cr-Cys) using the new 2021 equation and when adjusting for body surface area placed his eGFR at 92 mL/min/1.73 m².²⁰

The patient’s eGFR_Cysreassured the care team that the patient’s renal function was not acutely or chronically impacted by bictegravir/emtricitabine/tenofovir alafenamide, resulting in avoidance of unnecessary dosage adjustment or discontinuation of the HIV treatment. The patient reported a chronic habit of protein and creatine supplementation and bodybuilding, which likely further compounded the discrepancy between eGFR_Cr and eGFR_Cys and explained his previous elevation in Cr in 2018.

Follow-up

The patient underwent serial monitoring that revealed a stable Cr and unremarkable eGFR, ruling out CKD. There has been no evidence of worsening kidney disease to date, and the patient remained on his initial HIV regimen.

This case shows the importance of using CysC as an alternative or confirmatory marker compared with sCr to estimate GFR in patients with high muscle mass and/or high creatine intake, such as many in the US Department of Defense (DoD) and US Department of Veterans Affairs (VA) patient populations. In the presented case, recorded Cr levels climbed from baseline Cr with the initiation of bictegravir/emtricitabine/tenofovir alafenamide. This raised the concern that HIV treatment was leading to the development of kidney damage.²²

Diagnosis of kidney disease as opposed to the normal decline of eGFR with age in individuals without intrinsic CKD requires GFR ≥ 60 mL/min/1.73 m² with kidney damage (proteinuria or radiological abnormalities, etc) or GFR < 135 to 140 mL/min/1.73 m²minus the patient’s age in years.²³ The patient’s Cr peak at 1.83 mg/dL in 2018 led to an inappropriate diagnosis of kidney disease stage 3a based on an eGFR_Cr (2021 equation) of 52 mL/min/1.73 m² when not corrected for body surface area.²⁰ However, using the new 2021 equation using both Cr and CysC, the patient’s eGFR_Cr-Cyswas 92 mL/min/1.73 m² after a correction for body surface area.

The 2009 CKD-EPI recommended the calculation of eGFR based on SCr concentration using age, sex, and race while the 2021 CKD-EPI recommended the exclusion of race.³ Both equations are less accurate in African American patients, individuals taking medications that interfere with Cr secretion and assay, and patients taking creatine supplements, high daily protein intake, or with high muscle mass.⁷ These settings result in a decreased eGFR_Cr without corresponding eGFR_Cys changes. Using SCr and CysC together, the eGFR_Cr-Cys yields improved concordance to measured GFR across race groups compared to GFR estimation based on Cr alone, which can avoid unnecessary expensive diagnostic workup, inappropriate kidney disease diagnosis, incorrect dosing of drugs, and accurately represent the military readiness of patients. Interestingly, in African American patients with recently diagnosed HIV, CKD-EPI using both Cr and CysC without race inclusion led to only a 2.9% overestimation of GFR and was the only equation with no statistically significant bias compared with measured GFR.²⁴

A March 2023 case involving an otherwise healthy 26-year-old male active-duty US Navy member with a history of excessive protein supplement intake and intense exercise < 24 hours before laboratory work was diagnosed with CKD after a measured Cr of 16 mg/dL and an eGFR_Cr of 4 mL/min/1.73 m² without any other evidence of kidney disease. His CysC remained within normal limits, resulting in a normal eGFR_Cys of 121 mL/min/1.73 m², indicating no CKD. His Cr and eGFR recovered 10 days after his clinic visit and cessation of his supplement intake. These findings may not be uncommon given that 65% of active-duty military use protein supplements and 38% use other performance-enhancing supplements, such as creatine, according to a study.²⁵

Unfortunately, the BMP/CMP traditionally used at VA centers use the eGFR_Cr equation, and it is unknown how many primary care practitioners recognize the limitations of these metabolic panels on accurate estimation of kidney function. However, in 2022 an expert panel including VA physicians recommended the immediate use of eGFR_Cr-Cys or eGFR_Cys for confirmatory testing and potentially screening of CKD.²⁶ A small number of VAs have since adopted this recommendation, which should lead to fewer misdiagnoses among US military members as clinicians should now have access to more accurate measurements of GFR.

The VA spends about $18 billion (excluding dialysis) for care for 1.1 to 2.5 million VA patients with CKD.²⁷ The majority of these diagnoses were undoubtedly made using the eGFR_Cr equation, raising the question of how many may be misdiagnosed. Assessment with CysC is currently relatively expensive, but it will likely become more affordable as the use of CysC as a confirmatory test increases.⁵ The cost of a sCr test is about $2.50, while CysC costs about $10.60, with variation from laboratory to laboratory.²⁸ By comparison, a renal ultrasound costs $99 to $140 for uninsured patients.²⁹ Furthermore, the cost of CysC testing is likely to trend downward as more facilities adopt the use of CysC measurements, which can be run on the same analytical equipment currently used for Cr measurements. Currently, most laboratories do not have established assays to use in-house and thus require CysC to be sent out to a laboratory, which increases result time and makes Cr a more attractive option. As more laboratories adopt assays for CysC, the cost of reagents will further decrease.

Given such considerations, confirmation testing of kidney function with CysC in specific patient populations with decreased eGFR_Cr without other features of CKD can offer great medical and financial benefits. A 2023 KDIGO report noted that many individuals may be mistakenly diagnosed with CKD when using eGFR_Cr.³ KDIGO noted that a 2013 meta-analysis of 90,000 individuals found that with a Cr-based eGFR of 45 to 59 mL/min/1.73 m² (42%) had a CysC-based eGFR of ≥ 60 mL/min/1.73 m². An eGFR_Cr of 45 to 59 represents 54% of all patients with CKD, amounting to millions of people (including current and former military personnel).^3,29-31 Correcting a misdiagnosis of CKD would bring significant relief to patients and save millions in health care spending.

Conclusions

In patients who meet CKD criteria using eGFR_Cr but without other features of CKD, we recommend using confirmatory CysC levels and the eGFR_Cr-Cys equation. This will align care with the KDIGO guidelines and could be a cost-effective step toward improving military patient care. Further work in this area should focus on determining the knowledge gaps in primary care practitioners’ understanding of the limits of eGFR_Cr, the potential mitigation of concomitant CysC testing in equivocal CKD cases, and the cost-effectiveness and increased utilization of CysC.

To the Editor:

Phototherapy regularly is utilized in the outpatient setting to address various skin pathologies, including atopic dermatitis, psoriasis, pruritus, vitiligo, and mycosis fungoides.^1,2 Phototherapy is broadly defined by the measured administration of nonionizing radiation within the UV range including wavelengths within the UVA (eg, psoralen sensitizer plus UVA-1) and UVB (eg, broadband UVB, narrowband UVB) spectrums.^1,3 Generally, the mechanism of action is derived from effects on inflammatory components of cutaneous disorders and the induction of apoptosis, both precipitating numerous downstream events.⁴

From 2015 to 2018, there were more than 1.3 million outpatient phototherapy visits in the United States, with the most common procedural indications being dermatitis not otherwise specified, atopic dermatitis, and pruritus.⁵ From 2000 to 2015, the quantity of phototherapy services billed to Medicare trended upwards by an average of 5% per year, increasing from 334,670 in the year 2000 to 692,093 in 2015.⁶ Therefore, an illustration of associated costs would be beneficial. Additionally, because total cost and physician reimbursement fluctuate from year to year, studies demonstrating overall trends can inform both US policymakers and physicians. There is a paucity of research on geographical trends for procedural reimbursements in dermatology for phototherapy. Understanding geographic trends of reimbursement could duly serve to optimize dermatologist practice patterns involving access to viable and quality care for patients seeking treatment as well as draw health policymakers’ attention to striking adjustments in physician fees. Therefore, in this study we aimed to illustrate the most recent regional payment trends in phototherapy procedures for Medicare B patients.

We queried the Centers for Medicare & Medicaid Services Medicare Physician Fee Schedule (MPFS) database (https://www.cms.gov/medicare/payment/fee-schedules/physician/lookup-tool) for the years 2010 to 2023 for Current Procedural Terminology (CPT) codes common to phototherapy procedures: actinotherapy (96900); photochemotherapy by Goeckerman treatment or using petrolatum and UVB (96910); photochemotherapy using psoralen plus UVA (96912); and photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision (96913). Nonfacility prices for these procedures were analyzed. For 2010, due to midyear alterations to Medicare reimbursement (owed to bills HR 3962 and HR 4872), the mean price data of MPFS files 2010A and 2010B were used. All dollar values were converted to January 2023 US dollars using corresponding consumer price index inflation data. The Medicare Administrative Contractors were used to group state pricing information by region in accordance with established US Census Bureau subdivisions (https://www.census.gov/programs-surveys/economic-census/guidance-geographies/levels.html). Weighted percentage change in reimbursement rate was calculated using physician (MD or DO) utilization (procedure volume) data available in the 2020 Physician and Other Practitioners Public Use File (https://data.cms.gov/provider-summary-by-type-of-service/medicare-physician-other-practitioners/medicare-physician-other-practitioners-by-provider-and-service). All descriptive statistics and visualization were generated using R software (v4.2.2)(R Development Core Team).

Table 1 provides physician utilization data and the corresponding number of Part B beneficiaries for phototherapy procedures in 2020. There were 65,045 services of actinotherapy provided to a total of 6855 unique Part B beneficiaries, 173,979 services of photochemotherapy by Goeckerman treatment or using petrolatum and UVB provided to 13,122 unique Part B beneficiaries, 2524 services of photochemotherapy using psoralen plus UVA provided to a total of 357 unique Part B beneficiaries, and 37 services of photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision provided to a total of 27 unique Part B beneficiaries.

On average (unweighted), phototherapy reimbursement rates in the North increased by 0.68% between 2010 and 2023 (Table 2). After weighting for 2020 physician utilization, the average change in reimbursement rate was +19.37%. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+31.45%)($98.12 to $128.98; compound annual growth rate [CAGR], +0.0213), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−12.76%)($126.09 to $109.97; CAGR, −0.0105). For CPT code 96900, the reported adjusted decrease in reimbursement was −11.68% ($30.21 to $26.68; CAGR, −0.0095), and for CPT code 96913, the reported adjusted decrease in reimbursement was −4.27% ($174.03 to $166.60; CAGR, −0.0034).

On average (unweighted), phototherapy reimbursement rates in the Midwest increased by 8.40% between 2010 and 2023 (Table 3). After weighting for 2020 physician utilization, the average change in reimbursement rate was +28.53%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+41.48%)($80.42 to $113.78; CAGR, +0.0270), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−6.14%)($103.28 to $97.03; CAGR, −0.0049). For CPT code 96900, the reported adjusted decrease in reimbursement was −4.73% ($24.69 to $23.52; CAGR, −0.0037), and for CPT code 96913, the reported adjusted increase in reimbursement was +2.99% ($142.72 to $146.99; CAGR, +0.0023).

On average (unweighted), phototherapy reimbursement rates in the South decreased by 2.62% between 2010 and 2023 (Table 4). After weighting for 2020 physician utilization, the average change in reimbursement rate was +15.41%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+27.26%)($90.40 to $115.04 USD; CAGR, +0.0187), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−15.50%)($116.08 to $98.09; CAGR, −0.0129). For CPT code 96900, the reported adjusted decrease in reimbursement was −15.06% ($28.02 to $23.80; CAGR, −0.0125), and for CPT code 96913, the reported adjusted decrease in reimbursement was −7.19% ($160.11 to $148.61; CAGR, −0.0057).

On average (unweighted), phototherapy reimbursement rates in the West increased by 27.53% between 2010 and 2023 (Table 5). After weighting for 2020 physician utilization, the average change in reimbursement rate was +51.16%. Reimbursement for all analyzed procedures increased in the western United States. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+66.56%)($80.84 to $134.65; CAGR, +0.0400), and CPT code 96912 reported the lowest adjusted increase in reimbursement (+10.64%)($103.88 to $114.93; CAGR, +0.0078). For CPT code 96900, the reported adjusted increase in reimbursement was 11.54% ($24.88 to $27.75; CAGR, +0.0084), and for CPT code 96913, the reported adjusted increase in reimbursement was 21.38% ($143.39 to $174.04; CAGR, +0.0150).

In this study evaluating geographical payment trends for phototherapy from 2010 to 2023, we demonstrated regional inconsistency in mean inflation-adjusted Medicare reimbursement rates. We found that all phototherapy procedures had increased reimbursement in the western United States, whereas all other regions reported cuts in reimbursement rates for at least half of the analyzed procedures. After adjusting for procedure utilization by physicians, weighted mean reimbursement for phototherapy increased in all US regions.

In a cross-sectional study that explored trends in the geographic distribution of dermatologists from 2012 to 2017, dermatologists in the northeastern and western United States were more likely to be located in higher-income zip codes, whereas dermatologists in the southern United States were more likely to be located in lower-income zip codes,⁷ suggesting that payment rate changes are not concordant with cost of living. Additionally, Lauck and colleagues⁸ observed that 75% of the top 20 most common procedures performed by dermatologists had decreased reimbursement (mean change, −10.8%) from 2011 to 2021. Other studies on Medicare reimbursement trends over the last 2 decades have reported major decreases within other specialties, suggesting that declining Medicare reimbursements are not unique to dermatology.^9,10 It is critical to monitor these developments, as the Centers for Medicare & Medicaid Services emphasized health care policy changes aimed at increasing reimbursements for evaluation and management services with compensatory payment cuts in billing for procedural services.¹¹

Mazmudar et al¹² previously reported a mean reimbursement decrease of −6.6% for laser/phototherapy procedures between 2007 and 2021, but these data did not include the heavily utilized Goeckerman treatment. Changes in reimbursement pose major ramifications for dermatologists—for practice size, scope, and longevity—as rates influence changes in commercial insurance reimbursements.¹³ Medicare plays a major role in the US health care system as the second largest expenditure¹⁴; indeed, between 2000 and 2015, Part B billing volume for phototherapy procedures increased 5% annually. However, phototherapy remains inaccessible in many locations due to unequal regional distribution of phototherapy clinics.⁶ Moreover, home phototherapy units are not yet widely utilized because of safety and efficacy concerns, lack of physician oversight, and difficulty obtaining insurance coverage.¹⁵ Acknowledgment and consideration of these geographical trends may persuasively allow policymakers, hospitals, and physicians to facilitate cost-effective phototherapy reimbursements that ensure continued access to quality and sustainable dermatologic care in the United States that tailor to regional needs.

In sum, this analysis reveals regional trends in Part B physician reimbursement for phototherapy procedures, with all US regions reporting a mean increase in phototherapy reimbursement after adjusting for utilization, albeit to varying degrees. Mean reimbursement for photochemotherapy by Goeckerman treatment or using petrolatum and UVB increased most among phototherapy procedures. Mean reimbursement for both actinotherapy and photochemotherapy using psoralen plus UVA decreased in all regions except the western United States.

Limitations include the restriction to Part B MPFS and the reliance on single-year (2020) physician utilization data to compute weighted changes in average reimbursement across a multiyear range, effectively restricting sweeping conclusions. Still, this study puts forth actionable insights for dermatologists and policymakers alike to appreciate and consider.

To the Editor:

Phototherapy regularly is utilized in the outpatient setting to address various skin pathologies, including atopic dermatitis, psoriasis, pruritus, vitiligo, and mycosis fungoides.^1,2 Phototherapy is broadly defined by the measured administration of nonionizing radiation within the UV range including wavelengths within the UVA (eg, psoralen sensitizer plus UVA-1) and UVB (eg, broadband UVB, narrowband UVB) spectrums.^1,3 Generally, the mechanism of action is derived from effects on inflammatory components of cutaneous disorders and the induction of apoptosis, both precipitating numerous downstream events.⁴

From 2015 to 2018, there were more than 1.3 million outpatient phototherapy visits in the United States, with the most common procedural indications being dermatitis not otherwise specified, atopic dermatitis, and pruritus.⁵ From 2000 to 2015, the quantity of phototherapy services billed to Medicare trended upwards by an average of 5% per year, increasing from 334,670 in the year 2000 to 692,093 in 2015.⁶ Therefore, an illustration of associated costs would be beneficial. Additionally, because total cost and physician reimbursement fluctuate from year to year, studies demonstrating overall trends can inform both US policymakers and physicians. There is a paucity of research on geographical trends for procedural reimbursements in dermatology for phototherapy. Understanding geographic trends of reimbursement could duly serve to optimize dermatologist practice patterns involving access to viable and quality care for patients seeking treatment as well as draw health policymakers’ attention to striking adjustments in physician fees. Therefore, in this study we aimed to illustrate the most recent regional payment trends in phototherapy procedures for Medicare B patients.

We queried the Centers for Medicare & Medicaid Services Medicare Physician Fee Schedule (MPFS) database (https://www.cms.gov/medicare/payment/fee-schedules/physician/lookup-tool) for the years 2010 to 2023 for Current Procedural Terminology (CPT) codes common to phototherapy procedures: actinotherapy (96900); photochemotherapy by Goeckerman treatment or using petrolatum and UVB (96910); photochemotherapy using psoralen plus UVA (96912); and photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision (96913). Nonfacility prices for these procedures were analyzed. For 2010, due to midyear alterations to Medicare reimbursement (owed to bills HR 3962 and HR 4872), the mean price data of MPFS files 2010A and 2010B were used. All dollar values were converted to January 2023 US dollars using corresponding consumer price index inflation data. The Medicare Administrative Contractors were used to group state pricing information by region in accordance with established US Census Bureau subdivisions (https://www.census.gov/programs-surveys/economic-census/guidance-geographies/levels.html). Weighted percentage change in reimbursement rate was calculated using physician (MD or DO) utilization (procedure volume) data available in the 2020 Physician and Other Practitioners Public Use File (https://data.cms.gov/provider-summary-by-type-of-service/medicare-physician-other-practitioners/medicare-physician-other-practitioners-by-provider-and-service). All descriptive statistics and visualization were generated using R software (v4.2.2)(R Development Core Team).

Table 1 provides physician utilization data and the corresponding number of Part B beneficiaries for phototherapy procedures in 2020. There were 65,045 services of actinotherapy provided to a total of 6855 unique Part B beneficiaries, 173,979 services of photochemotherapy by Goeckerman treatment or using petrolatum and UVB provided to 13,122 unique Part B beneficiaries, 2524 services of photochemotherapy using psoralen plus UVA provided to a total of 357 unique Part B beneficiaries, and 37 services of photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision provided to a total of 27 unique Part B beneficiaries.

On average (unweighted), phototherapy reimbursement rates in the North increased by 0.68% between 2010 and 2023 (Table 2). After weighting for 2020 physician utilization, the average change in reimbursement rate was +19.37%. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+31.45%)($98.12 to $128.98; compound annual growth rate [CAGR], +0.0213), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−12.76%)($126.09 to $109.97; CAGR, −0.0105). For CPT code 96900, the reported adjusted decrease in reimbursement was −11.68% ($30.21 to $26.68; CAGR, −0.0095), and for CPT code 96913, the reported adjusted decrease in reimbursement was −4.27% ($174.03 to $166.60; CAGR, −0.0034).

On average (unweighted), phototherapy reimbursement rates in the Midwest increased by 8.40% between 2010 and 2023 (Table 3). After weighting for 2020 physician utilization, the average change in reimbursement rate was +28.53%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+41.48%)($80.42 to $113.78; CAGR, +0.0270), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−6.14%)($103.28 to $97.03; CAGR, −0.0049). For CPT code 96900, the reported adjusted decrease in reimbursement was −4.73% ($24.69 to $23.52; CAGR, −0.0037), and for CPT code 96913, the reported adjusted increase in reimbursement was +2.99% ($142.72 to $146.99; CAGR, +0.0023).

On average (unweighted), phototherapy reimbursement rates in the South decreased by 2.62% between 2010 and 2023 (Table 4). After weighting for 2020 physician utilization, the average change in reimbursement rate was +15.41%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+27.26%)($90.40 to $115.04 USD; CAGR, +0.0187), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−15.50%)($116.08 to $98.09; CAGR, −0.0129). For CPT code 96900, the reported adjusted decrease in reimbursement was −15.06% ($28.02 to $23.80; CAGR, −0.0125), and for CPT code 96913, the reported adjusted decrease in reimbursement was −7.19% ($160.11 to $148.61; CAGR, −0.0057).

On average (unweighted), phototherapy reimbursement rates in the West increased by 27.53% between 2010 and 2023 (Table 5). After weighting for 2020 physician utilization, the average change in reimbursement rate was +51.16%. Reimbursement for all analyzed procedures increased in the western United States. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+66.56%)($80.84 to $134.65; CAGR, +0.0400), and CPT code 96912 reported the lowest adjusted increase in reimbursement (+10.64%)($103.88 to $114.93; CAGR, +0.0078). For CPT code 96900, the reported adjusted increase in reimbursement was 11.54% ($24.88 to $27.75; CAGR, +0.0084), and for CPT code 96913, the reported adjusted increase in reimbursement was 21.38% ($143.39 to $174.04; CAGR, +0.0150).

In this study evaluating geographical payment trends for phototherapy from 2010 to 2023, we demonstrated regional inconsistency in mean inflation-adjusted Medicare reimbursement rates. We found that all phototherapy procedures had increased reimbursement in the western United States, whereas all other regions reported cuts in reimbursement rates for at least half of the analyzed procedures. After adjusting for procedure utilization by physicians, weighted mean reimbursement for phototherapy increased in all US regions.

In a cross-sectional study that explored trends in the geographic distribution of dermatologists from 2012 to 2017, dermatologists in the northeastern and western United States were more likely to be located in higher-income zip codes, whereas dermatologists in the southern United States were more likely to be located in lower-income zip codes,⁷ suggesting that payment rate changes are not concordant with cost of living. Additionally, Lauck and colleagues⁸ observed that 75% of the top 20 most common procedures performed by dermatologists had decreased reimbursement (mean change, −10.8%) from 2011 to 2021. Other studies on Medicare reimbursement trends over the last 2 decades have reported major decreases within other specialties, suggesting that declining Medicare reimbursements are not unique to dermatology.^9,10 It is critical to monitor these developments, as the Centers for Medicare & Medicaid Services emphasized health care policy changes aimed at increasing reimbursements for evaluation and management services with compensatory payment cuts in billing for procedural services.¹¹

Mazmudar et al¹² previously reported a mean reimbursement decrease of −6.6% for laser/phototherapy procedures between 2007 and 2021, but these data did not include the heavily utilized Goeckerman treatment. Changes in reimbursement pose major ramifications for dermatologists—for practice size, scope, and longevity—as rates influence changes in commercial insurance reimbursements.¹³ Medicare plays a major role in the US health care system as the second largest expenditure¹⁴; indeed, between 2000 and 2015, Part B billing volume for phototherapy procedures increased 5% annually. However, phototherapy remains inaccessible in many locations due to unequal regional distribution of phototherapy clinics.⁶ Moreover, home phototherapy units are not yet widely utilized because of safety and efficacy concerns, lack of physician oversight, and difficulty obtaining insurance coverage.¹⁵ Acknowledgment and consideration of these geographical trends may persuasively allow policymakers, hospitals, and physicians to facilitate cost-effective phototherapy reimbursements that ensure continued access to quality and sustainable dermatologic care in the United States that tailor to regional needs.

In sum, this analysis reveals regional trends in Part B physician reimbursement for phototherapy procedures, with all US regions reporting a mean increase in phototherapy reimbursement after adjusting for utilization, albeit to varying degrees. Mean reimbursement for photochemotherapy by Goeckerman treatment or using petrolatum and UVB increased most among phototherapy procedures. Mean reimbursement for both actinotherapy and photochemotherapy using psoralen plus UVA decreased in all regions except the western United States.

Limitations include the restriction to Part B MPFS and the reliance on single-year (2020) physician utilization data to compute weighted changes in average reimbursement across a multiyear range, effectively restricting sweeping conclusions. Still, this study puts forth actionable insights for dermatologists and policymakers alike to appreciate and consider.

To the Editor:

Phototherapy regularly is utilized in the outpatient setting to address various skin pathologies, including atopic dermatitis, psoriasis, pruritus, vitiligo, and mycosis fungoides.^1,2 Phototherapy is broadly defined by the measured administration of nonionizing radiation within the UV range including wavelengths within the UVA (eg, psoralen sensitizer plus UVA-1) and UVB (eg, broadband UVB, narrowband UVB) spectrums.^1,3 Generally, the mechanism of action is derived from effects on inflammatory components of cutaneous disorders and the induction of apoptosis, both precipitating numerous downstream events.⁴

From 2015 to 2018, there were more than 1.3 million outpatient phototherapy visits in the United States, with the most common procedural indications being dermatitis not otherwise specified, atopic dermatitis, and pruritus.⁵ From 2000 to 2015, the quantity of phototherapy services billed to Medicare trended upwards by an average of 5% per year, increasing from 334,670 in the year 2000 to 692,093 in 2015.⁶ Therefore, an illustration of associated costs would be beneficial. Additionally, because total cost and physician reimbursement fluctuate from year to year, studies demonstrating overall trends can inform both US policymakers and physicians. There is a paucity of research on geographical trends for procedural reimbursements in dermatology for phototherapy. Understanding geographic trends of reimbursement could duly serve to optimize dermatologist practice patterns involving access to viable and quality care for patients seeking treatment as well as draw health policymakers’ attention to striking adjustments in physician fees. Therefore, in this study we aimed to illustrate the most recent regional payment trends in phototherapy procedures for Medicare B patients.

We queried the Centers for Medicare & Medicaid Services Medicare Physician Fee Schedule (MPFS) database (https://www.cms.gov/medicare/payment/fee-schedules/physician/lookup-tool) for the years 2010 to 2023 for Current Procedural Terminology (CPT) codes common to phototherapy procedures: actinotherapy (96900); photochemotherapy by Goeckerman treatment or using petrolatum and UVB (96910); photochemotherapy using psoralen plus UVA (96912); and photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision (96913). Nonfacility prices for these procedures were analyzed. For 2010, due to midyear alterations to Medicare reimbursement (owed to bills HR 3962 and HR 4872), the mean price data of MPFS files 2010A and 2010B were used. All dollar values were converted to January 2023 US dollars using corresponding consumer price index inflation data. The Medicare Administrative Contractors were used to group state pricing information by region in accordance with established US Census Bureau subdivisions (https://www.census.gov/programs-surveys/economic-census/guidance-geographies/levels.html). Weighted percentage change in reimbursement rate was calculated using physician (MD or DO) utilization (procedure volume) data available in the 2020 Physician and Other Practitioners Public Use File (https://data.cms.gov/provider-summary-by-type-of-service/medicare-physician-other-practitioners/medicare-physician-other-practitioners-by-provider-and-service). All descriptive statistics and visualization were generated using R software (v4.2.2)(R Development Core Team).

Table 1 provides physician utilization data and the corresponding number of Part B beneficiaries for phototherapy procedures in 2020. There were 65,045 services of actinotherapy provided to a total of 6855 unique Part B beneficiaries, 173,979 services of photochemotherapy by Goeckerman treatment or using petrolatum and UVB provided to 13,122 unique Part B beneficiaries, 2524 services of photochemotherapy using psoralen plus UVA provided to a total of 357 unique Part B beneficiaries, and 37 services of photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision provided to a total of 27 unique Part B beneficiaries.

On average (unweighted), phototherapy reimbursement rates in the North increased by 0.68% between 2010 and 2023 (Table 2). After weighting for 2020 physician utilization, the average change in reimbursement rate was +19.37%. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+31.45%)($98.12 to $128.98; compound annual growth rate [CAGR], +0.0213), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−12.76%)($126.09 to $109.97; CAGR, −0.0105). For CPT code 96900, the reported adjusted decrease in reimbursement was −11.68% ($30.21 to $26.68; CAGR, −0.0095), and for CPT code 96913, the reported adjusted decrease in reimbursement was −4.27% ($174.03 to $166.60; CAGR, −0.0034).

On average (unweighted), phototherapy reimbursement rates in the Midwest increased by 8.40% between 2010 and 2023 (Table 3). After weighting for 2020 physician utilization, the average change in reimbursement rate was +28.53%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+41.48%)($80.42 to $113.78; CAGR, +0.0270), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−6.14%)($103.28 to $97.03; CAGR, −0.0049). For CPT code 96900, the reported adjusted decrease in reimbursement was −4.73% ($24.69 to $23.52; CAGR, −0.0037), and for CPT code 96913, the reported adjusted increase in reimbursement was +2.99% ($142.72 to $146.99; CAGR, +0.0023).

On average (unweighted), phototherapy reimbursement rates in the South decreased by 2.62% between 2010 and 2023 (Table 4). After weighting for 2020 physician utilization, the average change in reimbursement rate was +15.41%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+27.26%)($90.40 to $115.04 USD; CAGR, +0.0187), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−15.50%)($116.08 to $98.09; CAGR, −0.0129). For CPT code 96900, the reported adjusted decrease in reimbursement was −15.06% ($28.02 to $23.80; CAGR, −0.0125), and for CPT code 96913, the reported adjusted decrease in reimbursement was −7.19% ($160.11 to $148.61; CAGR, −0.0057).

On average (unweighted), phototherapy reimbursement rates in the West increased by 27.53% between 2010 and 2023 (Table 5). After weighting for 2020 physician utilization, the average change in reimbursement rate was +51.16%. Reimbursement for all analyzed procedures increased in the western United States. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+66.56%)($80.84 to $134.65; CAGR, +0.0400), and CPT code 96912 reported the lowest adjusted increase in reimbursement (+10.64%)($103.88 to $114.93; CAGR, +0.0078). For CPT code 96900, the reported adjusted increase in reimbursement was 11.54% ($24.88 to $27.75; CAGR, +0.0084), and for CPT code 96913, the reported adjusted increase in reimbursement was 21.38% ($143.39 to $174.04; CAGR, +0.0150).

In this study evaluating geographical payment trends for phototherapy from 2010 to 2023, we demonstrated regional inconsistency in mean inflation-adjusted Medicare reimbursement rates. We found that all phototherapy procedures had increased reimbursement in the western United States, whereas all other regions reported cuts in reimbursement rates for at least half of the analyzed procedures. After adjusting for procedure utilization by physicians, weighted mean reimbursement for phototherapy increased in all US regions.

In a cross-sectional study that explored trends in the geographic distribution of dermatologists from 2012 to 2017, dermatologists in the northeastern and western United States were more likely to be located in higher-income zip codes, whereas dermatologists in the southern United States were more likely to be located in lower-income zip codes,⁷ suggesting that payment rate changes are not concordant with cost of living. Additionally, Lauck and colleagues⁸ observed that 75% of the top 20 most common procedures performed by dermatologists had decreased reimbursement (mean change, −10.8%) from 2011 to 2021. Other studies on Medicare reimbursement trends over the last 2 decades have reported major decreases within other specialties, suggesting that declining Medicare reimbursements are not unique to dermatology.^9,10 It is critical to monitor these developments, as the Centers for Medicare & Medicaid Services emphasized health care policy changes aimed at increasing reimbursements for evaluation and management services with compensatory payment cuts in billing for procedural services.¹¹

Mazmudar et al¹² previously reported a mean reimbursement decrease of −6.6% for laser/phototherapy procedures between 2007 and 2021, but these data did not include the heavily utilized Goeckerman treatment. Changes in reimbursement pose major ramifications for dermatologists—for practice size, scope, and longevity—as rates influence changes in commercial insurance reimbursements.¹³ Medicare plays a major role in the US health care system as the second largest expenditure¹⁴; indeed, between 2000 and 2015, Part B billing volume for phototherapy procedures increased 5% annually. However, phototherapy remains inaccessible in many locations due to unequal regional distribution of phototherapy clinics.⁶ Moreover, home phototherapy units are not yet widely utilized because of safety and efficacy concerns, lack of physician oversight, and difficulty obtaining insurance coverage.¹⁵ Acknowledgment and consideration of these geographical trends may persuasively allow policymakers, hospitals, and physicians to facilitate cost-effective phototherapy reimbursements that ensure continued access to quality and sustainable dermatologic care in the United States that tailor to regional needs.

In sum, this analysis reveals regional trends in Part B physician reimbursement for phototherapy procedures, with all US regions reporting a mean increase in phototherapy reimbursement after adjusting for utilization, albeit to varying degrees. Mean reimbursement for photochemotherapy by Goeckerman treatment or using petrolatum and UVB increased most among phototherapy procedures. Mean reimbursement for both actinotherapy and photochemotherapy using psoralen plus UVA decreased in all regions except the western United States.

Limitations include the restriction to Part B MPFS and the reliance on single-year (2020) physician utilization data to compute weighted changes in average reimbursement across a multiyear range, effectively restricting sweeping conclusions. Still, this study puts forth actionable insights for dermatologists and policymakers alike to appreciate and consider.

Hypertension is a major risk factor for heart disease, stroke, and kidney disease.The prevalence of hypertension increases with age, primarily due to age-related changes in arterial physiology.¹ For older adults, current guidelines regarding blood pressure (BP) treatment goals vary. The American Heart Association/American College of Cardiology 2017 clinical practice guidelines recommend a systolic BP (SBP) treatment goal of < 130 mm Hg for community-dwelling, ambulatory, noninstitutionalized adults aged ≥ 65 years; whereas the American College of Physicians/American Academy of Family Physicians recommend a goal of < 150 mm Hg for those aged ≥ 60 years without comorbidities and < 140 mm Hg for those with increased cardiovascular risk.^1-3 Regardless of the specific threshold, agreement that some degree of BP control even in those with advanced age improves outcomes.²

First-line therapy for uncomplicated hypertension includes thiazide diuretics, long-acting calcium channel blockers, and renin-angiotensin system inhibitors. When choosing between these options, it is recommended to engage in shared decision making and to consider the patient’s comorbidities. Among patients who are likely to require a second agent (eg, if initial BP is > 20/10 mm Hg above goal), it is recommended to begin both drugs at the same time, preferably benazepril plus amlodipine due to the reduction in cardiovascular events reported in the ACCOMPLISH trial.⁴ If BP remains elevated despite 2 agents at moderate to maximum doses, it is important to investigate for secondary hypertension causes and to explore medication adherence as possible etiologies of treatment failure. Older adults are often at higher risk of adverse drug events due to age-related changes in pharmacodynamics. Despite this, there are no guidelines for choosing between different classes of antihypertensives in this population. We present a case of thiazide-induced hyponatremia in an older adult and review the risks of thiazide use in this population.

Case Presentation

A man aged > 90 years was admitted to the hospital after a syncopal episode. His history was significant for hypertension, hyperlipidemia, and vitamin D deficiency. At the time, his home medications were amlodipine 5 mg daily, atorvastatin 40 mg daily, ergocalciferol 50,000 IU weekly, and polyethylene glycol 17 g daily as needed. His syncope workup was unremarkable and included negative orthostatic vital signs, normal serial troponins, an electrocardiogram without ischemic changes, normal serum creatinine, sodium, and glucose, and a head computed tomography without any acute abnormality. Throughout the patient’s hospital stay, he had multiple elevated SBP readings, including many > 200 mm Hg. On discharge, in addition to continuing his home medications, he was started on valsartan 20 mg daily and enrolled in a remote BP monitoring program.

Three weeks later, the patient was seen by their primary care practitioner for follow-up. He reported adherence to his antihypertensive regimen. However, his remote BP monitoring revealed persistently elevated BPs, with an average of 179/79 mm Hg, a high of 205/85 mm Hg, and a low of 150/67 mm Hg over the previous 7 days. Laboratory tests obtained at the visit were notable for serum sodium of 138 mmol/L and potassium of 4.1 mmol/L. His weight was 87 kg. Given persistently elevated BP readings, in addition to continuing his amlodipine 5 mg daily and valsartan 20 mg daily, he was started on hydrochlorothiazide 25 mg daily, with plans to repeat a basic metabolic panel in 2 weeks.

Two weeks later, he fell after getting out of his bed. On examination, he was noted to have dry mucous membranes, and although no formal delirium screening was performed, he was able to repeat the months of the year backward. Vital signs were notable for positive postural hypertension, and his laboratory tests revealed a normal serum creatinine, serum sodium of 117 mmol/L (reference range, 135-145 mmol/L), serum potassium of 3.2 mmol/L (reference range, 3.0-5.0 mmol/L), a low serum osmolality, and urine sodium of 35 mmol/L most consistent with hypovolemic hypoosmotic hyponatremia secondary to thiazide initiation. The patient’s hydrochlorothiazide was discontinued, and he was admitted to the hospital for close monitoring. His sodium levels gradually normalized over the next 2 weeks without any other intervention.

Discussion

Although thiazide diuretics are recommended as first-line therapy for uncomplicated hypertension, they are known to cause electrolyte abnormalities, including hypomagnesemia, hypokalemia, and hyponatremia.⁴ These metabolic derangements are more likely to occur in older adults. One study of adults aged ≥ 65 years found that at 9 months of follow-up, 14.3% of new thiazide users had developed a thiazide-related metabolic adverse event (hyponatremia < 135 mmol/L, hypokalemia < 3.5 mmol/L, and decrease in estimated glomerular filtration rate by > 25%) compared with 6.0% of nonusers (P < .001; number needed to harm [NNH] = 12).⁵ In addition, 3.8% of new thiazide users had an emergency department visit or were hospitalized for complications related to thiazides compared with only 2.0% of nonusers (P = .02; NNH = 56).⁵ Independent risk factors for thiazide-induced hyponatremia include high-comorbidity burden, low body weight, low-normal or unmeasured serum sodium, low potassium, and aged > 70 years.^5-7 Each 10-year increment in age is associated with a 2-fold increase in risk, suggesting that older adults are at a much higher risk for hyponatremia than their younger peers.⁶

Despite their designation as a first-line option for uncomplicated hypertension, thiazide diuretics may cause more harm than good in some older adults, especially those with additional risk factors for thiazide-induced hyponatremia. In this population, these adverse effects should be discussed before starting thiazides for the treatment of hypertension. If thiazides are initiated, they should be started at the lowest possible dose, and plans made to monitor bloodwork within 1 to 2 weeks of initiation or dose change and periodically thereafter while the patient remains on the therapy.

Medication Management in Older Adults

Due to the risks of medication use in older adults, the phrase “start low, go slow” is commonly used in geriatric medicine to describe the optimal method for initiation and up-titration of new medication with the hope of mitigating adverse drug events. In our case, we started valsartan at 20 mg daily—one-fourth the recommended initial dose. Although this strategy is reasonable to “start low,” we were not surprised to find that the patient’s BP did not markedly improve on such a low dose. The team could have increased the valsartan dose to a therapeutically efficacious dose before choosing to add another hypertensive agent. In alignment with geriatric prescribing principles, starting at the lowest possible dose of hydrochlorothiazide is recommended.⁵ However, the clinician started hydrochlorothiazide at 25 mg daily, potentially increasing this patient’s risk of electrolyte abnormalities and eventual fall.

Managing hypertension also invites a discussion of polypharmacy and medication adherence. Older adults are at risk of polypharmacy, defined as the prescription of 5 or more medications.⁸ Polypharmacy is associated with increased hospitalizations, higher costs of care for individuals and health care systems, increased risks of adverse drug events, medication nonadherence, and lower quality of life for patients.⁹ In some situations, the risks of polypharmacy may outweigh the benefits of using multiple antihypertensives with different mechanisms of action if patients can reach their BP goal on the maximum dose of a single agent. For patients taking multiple antihypertensives, it is important to routinely monitor BP and assess whether deprescribing is indicated. Cognitive impairment and decreased social support may affect medication adherence for older adults.⁶ Clinicians should be aware of strategies, such as medication reminders and pillboxes, to increase antihypertensive medication adherence. Polypills that contain 2 antihypertensives can be another tool used to manage older adults to increase adherence and decrease health care costs.¹⁰

A current strategy that encompasses discussing many, if not all, of these noted elements is the Institute for Healthcare Improvement’s Age-Friendly Health System. This framework uses evidence-based tools to provide care for older adults across all clinical settings and highlights the 4Ms: what matters, medication, mentation, and mobility.¹¹ Medication considers whether a medication is necessary, whether its use has benefits that outweigh the risks, and how it interacts with what matters, mentation, and mobility. In particular, what matters plays an important role in hypertension management in older adults given the recommended target BP differs, depending on which specialty organization guideline is followed. By better understanding what matters to patients, including their goals and priorities, clinicians can engage patients in shared decision making and provide individualized recommendations based on geriatric principles (eg, start low, go slow, principles of medication adherence) and patient comorbidities (eg, medical history and risk factors for hyponatremia) to help patients make a more informed choice about their antihypertensive treatment regimen (Figure).

Conclusions

This case illustrates the need for a specialized approach to hypertension management in older adults and the risks of thiazide diuretics in this population. Clinicians should consider BP goals, patient-specific factors, and principles of medication management in older adults. If initiating thiazide therapy, discuss the risks associated with use, start at the lowest possible dose, and monitor bloodwork within 1 to 2 weeks of initiation/dose change and periodically thereafter while the patient remains on the therapy to decrease the risk of adverse events. Finally, the Institute for Healthcare Improvement’s Age-Friendly Health System framework can be a useful when considering the addition of a new medication in an older adult’s treatment plan.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the New England Geriatrics Research, Education, and Clinical Center, Veterans Affairs Boston Healthcare System, and the Cincinnati VeteransAffairs Medical Center.

Hypertension is a major risk factor for heart disease, stroke, and kidney disease.The prevalence of hypertension increases with age, primarily due to age-related changes in arterial physiology.¹ For older adults, current guidelines regarding blood pressure (BP) treatment goals vary. The American Heart Association/American College of Cardiology 2017 clinical practice guidelines recommend a systolic BP (SBP) treatment goal of < 130 mm Hg for community-dwelling, ambulatory, noninstitutionalized adults aged ≥ 65 years; whereas the American College of Physicians/American Academy of Family Physicians recommend a goal of < 150 mm Hg for those aged ≥ 60 years without comorbidities and < 140 mm Hg for those with increased cardiovascular risk.^1-3 Regardless of the specific threshold, agreement that some degree of BP control even in those with advanced age improves outcomes.²

First-line therapy for uncomplicated hypertension includes thiazide diuretics, long-acting calcium channel blockers, and renin-angiotensin system inhibitors. When choosing between these options, it is recommended to engage in shared decision making and to consider the patient’s comorbidities. Among patients who are likely to require a second agent (eg, if initial BP is > 20/10 mm Hg above goal), it is recommended to begin both drugs at the same time, preferably benazepril plus amlodipine due to the reduction in cardiovascular events reported in the ACCOMPLISH trial.⁴ If BP remains elevated despite 2 agents at moderate to maximum doses, it is important to investigate for secondary hypertension causes and to explore medication adherence as possible etiologies of treatment failure. Older adults are often at higher risk of adverse drug events due to age-related changes in pharmacodynamics. Despite this, there are no guidelines for choosing between different classes of antihypertensives in this population. We present a case of thiazide-induced hyponatremia in an older adult and review the risks of thiazide use in this population.

Case Presentation

A man aged > 90 years was admitted to the hospital after a syncopal episode. His history was significant for hypertension, hyperlipidemia, and vitamin D deficiency. At the time, his home medications were amlodipine 5 mg daily, atorvastatin 40 mg daily, ergocalciferol 50,000 IU weekly, and polyethylene glycol 17 g daily as needed. His syncope workup was unremarkable and included negative orthostatic vital signs, normal serial troponins, an electrocardiogram without ischemic changes, normal serum creatinine, sodium, and glucose, and a head computed tomography without any acute abnormality. Throughout the patient’s hospital stay, he had multiple elevated SBP readings, including many > 200 mm Hg. On discharge, in addition to continuing his home medications, he was started on valsartan 20 mg daily and enrolled in a remote BP monitoring program.

Three weeks later, the patient was seen by their primary care practitioner for follow-up. He reported adherence to his antihypertensive regimen. However, his remote BP monitoring revealed persistently elevated BPs, with an average of 179/79 mm Hg, a high of 205/85 mm Hg, and a low of 150/67 mm Hg over the previous 7 days. Laboratory tests obtained at the visit were notable for serum sodium of 138 mmol/L and potassium of 4.1 mmol/L. His weight was 87 kg. Given persistently elevated BP readings, in addition to continuing his amlodipine 5 mg daily and valsartan 20 mg daily, he was started on hydrochlorothiazide 25 mg daily, with plans to repeat a basic metabolic panel in 2 weeks.

Two weeks later, he fell after getting out of his bed. On examination, he was noted to have dry mucous membranes, and although no formal delirium screening was performed, he was able to repeat the months of the year backward. Vital signs were notable for positive postural hypertension, and his laboratory tests revealed a normal serum creatinine, serum sodium of 117 mmol/L (reference range, 135-145 mmol/L), serum potassium of 3.2 mmol/L (reference range, 3.0-5.0 mmol/L), a low serum osmolality, and urine sodium of 35 mmol/L most consistent with hypovolemic hypoosmotic hyponatremia secondary to thiazide initiation. The patient’s hydrochlorothiazide was discontinued, and he was admitted to the hospital for close monitoring. His sodium levels gradually normalized over the next 2 weeks without any other intervention.

Discussion

Although thiazide diuretics are recommended as first-line therapy for uncomplicated hypertension, they are known to cause electrolyte abnormalities, including hypomagnesemia, hypokalemia, and hyponatremia.⁴ These metabolic derangements are more likely to occur in older adults. One study of adults aged ≥ 65 years found that at 9 months of follow-up, 14.3% of new thiazide users had developed a thiazide-related metabolic adverse event (hyponatremia < 135 mmol/L, hypokalemia < 3.5 mmol/L, and decrease in estimated glomerular filtration rate by > 25%) compared with 6.0% of nonusers (P < .001; number needed to harm [NNH] = 12).⁵ In addition, 3.8% of new thiazide users had an emergency department visit or were hospitalized for complications related to thiazides compared with only 2.0% of nonusers (P = .02; NNH = 56).⁵ Independent risk factors for thiazide-induced hyponatremia include high-comorbidity burden, low body weight, low-normal or unmeasured serum sodium, low potassium, and aged > 70 years.^5-7 Each 10-year increment in age is associated with a 2-fold increase in risk, suggesting that older adults are at a much higher risk for hyponatremia than their younger peers.⁶

Despite their designation as a first-line option for uncomplicated hypertension, thiazide diuretics may cause more harm than good in some older adults, especially those with additional risk factors for thiazide-induced hyponatremia. In this population, these adverse effects should be discussed before starting thiazides for the treatment of hypertension. If thiazides are initiated, they should be started at the lowest possible dose, and plans made to monitor bloodwork within 1 to 2 weeks of initiation or dose change and periodically thereafter while the patient remains on the therapy.

Medication Management in Older Adults

Due to the risks of medication use in older adults, the phrase “start low, go slow” is commonly used in geriatric medicine to describe the optimal method for initiation and up-titration of new medication with the hope of mitigating adverse drug events. In our case, we started valsartan at 20 mg daily—one-fourth the recommended initial dose. Although this strategy is reasonable to “start low,” we were not surprised to find that the patient’s BP did not markedly improve on such a low dose. The team could have increased the valsartan dose to a therapeutically efficacious dose before choosing to add another hypertensive agent. In alignment with geriatric prescribing principles, starting at the lowest possible dose of hydrochlorothiazide is recommended.⁵ However, the clinician started hydrochlorothiazide at 25 mg daily, potentially increasing this patient’s risk of electrolyte abnormalities and eventual fall.

Managing hypertension also invites a discussion of polypharmacy and medication adherence. Older adults are at risk of polypharmacy, defined as the prescription of 5 or more medications.⁸ Polypharmacy is associated with increased hospitalizations, higher costs of care for individuals and health care systems, increased risks of adverse drug events, medication nonadherence, and lower quality of life for patients.⁹ In some situations, the risks of polypharmacy may outweigh the benefits of using multiple antihypertensives with different mechanisms of action if patients can reach their BP goal on the maximum dose of a single agent. For patients taking multiple antihypertensives, it is important to routinely monitor BP and assess whether deprescribing is indicated. Cognitive impairment and decreased social support may affect medication adherence for older adults.⁶ Clinicians should be aware of strategies, such as medication reminders and pillboxes, to increase antihypertensive medication adherence. Polypills that contain 2 antihypertensives can be another tool used to manage older adults to increase adherence and decrease health care costs.¹⁰

A current strategy that encompasses discussing many, if not all, of these noted elements is the Institute for Healthcare Improvement’s Age-Friendly Health System. This framework uses evidence-based tools to provide care for older adults across all clinical settings and highlights the 4Ms: what matters, medication, mentation, and mobility.¹¹ Medication considers whether a medication is necessary, whether its use has benefits that outweigh the risks, and how it interacts with what matters, mentation, and mobility. In particular, what matters plays an important role in hypertension management in older adults given the recommended target BP differs, depending on which specialty organization guideline is followed. By better understanding what matters to patients, including their goals and priorities, clinicians can engage patients in shared decision making and provide individualized recommendations based on geriatric principles (eg, start low, go slow, principles of medication adherence) and patient comorbidities (eg, medical history and risk factors for hyponatremia) to help patients make a more informed choice about their antihypertensive treatment regimen (Figure).

Conclusions

This case illustrates the need for a specialized approach to hypertension management in older adults and the risks of thiazide diuretics in this population. Clinicians should consider BP goals, patient-specific factors, and principles of medication management in older adults. If initiating thiazide therapy, discuss the risks associated with use, start at the lowest possible dose, and monitor bloodwork within 1 to 2 weeks of initiation/dose change and periodically thereafter while the patient remains on the therapy to decrease the risk of adverse events. Finally, the Institute for Healthcare Improvement’s Age-Friendly Health System framework can be a useful when considering the addition of a new medication in an older adult’s treatment plan.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the New England Geriatrics Research, Education, and Clinical Center, Veterans Affairs Boston Healthcare System, and the Cincinnati VeteransAffairs Medical Center.

Hypertension is a major risk factor for heart disease, stroke, and kidney disease.The prevalence of hypertension increases with age, primarily due to age-related changes in arterial physiology.¹ For older adults, current guidelines regarding blood pressure (BP) treatment goals vary. The American Heart Association/American College of Cardiology 2017 clinical practice guidelines recommend a systolic BP (SBP) treatment goal of < 130 mm Hg for community-dwelling, ambulatory, noninstitutionalized adults aged ≥ 65 years; whereas the American College of Physicians/American Academy of Family Physicians recommend a goal of < 150 mm Hg for those aged ≥ 60 years without comorbidities and < 140 mm Hg for those with increased cardiovascular risk.^1-3 Regardless of the specific threshold, agreement that some degree of BP control even in those with advanced age improves outcomes.²

First-line therapy for uncomplicated hypertension includes thiazide diuretics, long-acting calcium channel blockers, and renin-angiotensin system inhibitors. When choosing between these options, it is recommended to engage in shared decision making and to consider the patient’s comorbidities. Among patients who are likely to require a second agent (eg, if initial BP is > 20/10 mm Hg above goal), it is recommended to begin both drugs at the same time, preferably benazepril plus amlodipine due to the reduction in cardiovascular events reported in the ACCOMPLISH trial.⁴ If BP remains elevated despite 2 agents at moderate to maximum doses, it is important to investigate for secondary hypertension causes and to explore medication adherence as possible etiologies of treatment failure. Older adults are often at higher risk of adverse drug events due to age-related changes in pharmacodynamics. Despite this, there are no guidelines for choosing between different classes of antihypertensives in this population. We present a case of thiazide-induced hyponatremia in an older adult and review the risks of thiazide use in this population.

Case Presentation

A man aged > 90 years was admitted to the hospital after a syncopal episode. His history was significant for hypertension, hyperlipidemia, and vitamin D deficiency. At the time, his home medications were amlodipine 5 mg daily, atorvastatin 40 mg daily, ergocalciferol 50,000 IU weekly, and polyethylene glycol 17 g daily as needed. His syncope workup was unremarkable and included negative orthostatic vital signs, normal serial troponins, an electrocardiogram without ischemic changes, normal serum creatinine, sodium, and glucose, and a head computed tomography without any acute abnormality. Throughout the patient’s hospital stay, he had multiple elevated SBP readings, including many > 200 mm Hg. On discharge, in addition to continuing his home medications, he was started on valsartan 20 mg daily and enrolled in a remote BP monitoring program.

Three weeks later, the patient was seen by their primary care practitioner for follow-up. He reported adherence to his antihypertensive regimen. However, his remote BP monitoring revealed persistently elevated BPs, with an average of 179/79 mm Hg, a high of 205/85 mm Hg, and a low of 150/67 mm Hg over the previous 7 days. Laboratory tests obtained at the visit were notable for serum sodium of 138 mmol/L and potassium of 4.1 mmol/L. His weight was 87 kg. Given persistently elevated BP readings, in addition to continuing his amlodipine 5 mg daily and valsartan 20 mg daily, he was started on hydrochlorothiazide 25 mg daily, with plans to repeat a basic metabolic panel in 2 weeks.

Two weeks later, he fell after getting out of his bed. On examination, he was noted to have dry mucous membranes, and although no formal delirium screening was performed, he was able to repeat the months of the year backward. Vital signs were notable for positive postural hypertension, and his laboratory tests revealed a normal serum creatinine, serum sodium of 117 mmol/L (reference range, 135-145 mmol/L), serum potassium of 3.2 mmol/L (reference range, 3.0-5.0 mmol/L), a low serum osmolality, and urine sodium of 35 mmol/L most consistent with hypovolemic hypoosmotic hyponatremia secondary to thiazide initiation. The patient’s hydrochlorothiazide was discontinued, and he was admitted to the hospital for close monitoring. His sodium levels gradually normalized over the next 2 weeks without any other intervention.

Discussion

Although thiazide diuretics are recommended as first-line therapy for uncomplicated hypertension, they are known to cause electrolyte abnormalities, including hypomagnesemia, hypokalemia, and hyponatremia.⁴ These metabolic derangements are more likely to occur in older adults. One study of adults aged ≥ 65 years found that at 9 months of follow-up, 14.3% of new thiazide users had developed a thiazide-related metabolic adverse event (hyponatremia < 135 mmol/L, hypokalemia < 3.5 mmol/L, and decrease in estimated glomerular filtration rate by > 25%) compared with 6.0% of nonusers (P < .001; number needed to harm [NNH] = 12).⁵ In addition, 3.8% of new thiazide users had an emergency department visit or were hospitalized for complications related to thiazides compared with only 2.0% of nonusers (P = .02; NNH = 56).⁵ Independent risk factors for thiazide-induced hyponatremia include high-comorbidity burden, low body weight, low-normal or unmeasured serum sodium, low potassium, and aged > 70 years.^5-7 Each 10-year increment in age is associated with a 2-fold increase in risk, suggesting that older adults are at a much higher risk for hyponatremia than their younger peers.⁶

Despite their designation as a first-line option for uncomplicated hypertension, thiazide diuretics may cause more harm than good in some older adults, especially those with additional risk factors for thiazide-induced hyponatremia. In this population, these adverse effects should be discussed before starting thiazides for the treatment of hypertension. If thiazides are initiated, they should be started at the lowest possible dose, and plans made to monitor bloodwork within 1 to 2 weeks of initiation or dose change and periodically thereafter while the patient remains on the therapy.

Medication Management in Older Adults

Due to the risks of medication use in older adults, the phrase “start low, go slow” is commonly used in geriatric medicine to describe the optimal method for initiation and up-titration of new medication with the hope of mitigating adverse drug events. In our case, we started valsartan at 20 mg daily—one-fourth the recommended initial dose. Although this strategy is reasonable to “start low,” we were not surprised to find that the patient’s BP did not markedly improve on such a low dose. The team could have increased the valsartan dose to a therapeutically efficacious dose before choosing to add another hypertensive agent. In alignment with geriatric prescribing principles, starting at the lowest possible dose of hydrochlorothiazide is recommended.⁵ However, the clinician started hydrochlorothiazide at 25 mg daily, potentially increasing this patient’s risk of electrolyte abnormalities and eventual fall.

Managing hypertension also invites a discussion of polypharmacy and medication adherence. Older adults are at risk of polypharmacy, defined as the prescription of 5 or more medications.⁸ Polypharmacy is associated with increased hospitalizations, higher costs of care for individuals and health care systems, increased risks of adverse drug events, medication nonadherence, and lower quality of life for patients.⁹ In some situations, the risks of polypharmacy may outweigh the benefits of using multiple antihypertensives with different mechanisms of action if patients can reach their BP goal on the maximum dose of a single agent. For patients taking multiple antihypertensives, it is important to routinely monitor BP and assess whether deprescribing is indicated. Cognitive impairment and decreased social support may affect medication adherence for older adults.⁶ Clinicians should be aware of strategies, such as medication reminders and pillboxes, to increase antihypertensive medication adherence. Polypills that contain 2 antihypertensives can be another tool used to manage older adults to increase adherence and decrease health care costs.¹⁰

A current strategy that encompasses discussing many, if not all, of these noted elements is the Institute for Healthcare Improvement’s Age-Friendly Health System. This framework uses evidence-based tools to provide care for older adults across all clinical settings and highlights the 4Ms: what matters, medication, mentation, and mobility.¹¹ Medication considers whether a medication is necessary, whether its use has benefits that outweigh the risks, and how it interacts with what matters, mentation, and mobility. In particular, what matters plays an important role in hypertension management in older adults given the recommended target BP differs, depending on which specialty organization guideline is followed. By better understanding what matters to patients, including their goals and priorities, clinicians can engage patients in shared decision making and provide individualized recommendations based on geriatric principles (eg, start low, go slow, principles of medication adherence) and patient comorbidities (eg, medical history and risk factors for hyponatremia) to help patients make a more informed choice about their antihypertensive treatment regimen (Figure).

Conclusions

This case illustrates the need for a specialized approach to hypertension management in older adults and the risks of thiazide diuretics in this population. Clinicians should consider BP goals, patient-specific factors, and principles of medication management in older adults. If initiating thiazide therapy, discuss the risks associated with use, start at the lowest possible dose, and monitor bloodwork within 1 to 2 weeks of initiation/dose change and periodically thereafter while the patient remains on the therapy to decrease the risk of adverse events. Finally, the Institute for Healthcare Improvement’s Age-Friendly Health System framework can be a useful when considering the addition of a new medication in an older adult’s treatment plan.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the New England Geriatrics Research, Education, and Clinical Center, Veterans Affairs Boston Healthcare System, and the Cincinnati VeteransAffairs Medical Center.

Hypoglycemia and weight gain are well-known adverse effects that can result from insulin and sulfonylureas in patients with type 2 diabetes mellitus (T2DM).^1,2 Insulin and sulfonylurea medications can cause additional weight gain in patients who are overweight or obese, which can increase the burden of diabetes therapy with added medications, raise the risk of hypoglycemia complications, and raise atherosclerotic cardiovascular disease risk factors.³ Although increasing the insulin or sulfonylurea dose is an option health care practitioners or pharmacists have, this approach can increase the risk of hypoglycemia, especially in older adults, such as the veteran population, which could lead to complications, such as falls.²

Previous studies focusing on hypoglycemic events in patients with T2DM showed that glucagon-like peptide-1 (GLP-1) agonist monotherapy has a low incidence of a hypoglycemic events. However, when a GLP-1 agonist is combined with insulin or sulfonylureas, patients have an increased chance of a hypoglycemic event.^3-8 According to the prescribing information for semaglutide, 1.6% to 3.8% of patients on a GLP-1 agonist monotherapy reported a documented symptomatic hypoglycemic event (blood glucose ≤ 70 mg/dL), based on semaglutide dosing. ⁹ Patients on combination therapy of a GLP-1 agonist and basal insulin and a GLP-1 agonist and a sulfonylurea reported a documented symptomatic hypoglycemic event ranging from 16.7% to 29.8% and 17.3% to 24.4%, respectively.⁹ The incidences of hypoglycemia thus dramatically increase with combination therapy of a GLP-1 agonist plus insulin or a sulfonylurea.

When adding a GLP-1 agonist to insulin or a sulfonylurea, clinicians must be mindful of the increased risk of hypoglycemia. Per the warnings and precautions in the prescribing information of GLP-1 agonists, concomitant use with insulin or a sulfonylurea may increase the risk of hypoglycemia, and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 According to the American College of Cardiology guidelines, when starting a GLP-1 agonist, the insulin dose should be decreased by about 20% in patients with a well-controlled hemoglobin A_1c (HbA_1c).¹²

This study aimed to determine the percentage of patients who required dose reductions or discontinuations of insulin and sulfonylureas with the addition of a GLP-1 agonist. Understanding necessary dose reductions or discontinuations of these concomitant diabetes agents can assist pharmacists in preventing hypoglycemia and minimizing weight gain.

Methods

This clinical review was a single-center, retrospective chart review of patients prescribed a GLP-1 agonist while on insulin or a sulfonylurea between January 1, 2019, and September 30, 2022, at the Wilkes-Barre Veterans Affairs Medical Center (WBVAMC) in Pennsylvania and managed in a pharmacist-led patient aligned care team (PACT) clinic. It was determined by the US Department of Veterans Affairs Office of Research and Development that an institutional review board or other review committee approval was not needed for this nonresearch Veterans Health Administration quality assurance and improvement project. Patients aged ≥ 18 years were included in this study. Patients were excluded if they were not on insulin or a sulfonylurea when starting a GLP-1 agonist, started a GLP-1 agonist outside of the retrospective chart review dates, or were prescribed a GLP-1 agonist by anyone other than a pharmacist in their PACT clinic. This included if a GLP-1 agonist was prescribed by a primary care physician, endocrinologist, or someone outside the VA system.

The primary study outcomes were to determine the percentage of patients with a dose reduction of insulin or sulfonylurea and discontinuation of insulin or a sulfonylurea at intervals of 0 (baseline), 3, 6, and 12 months. Secondary outcomes included changes in HbA_1c and body weight measured at the same intervals of 0 (baseline), 3, 6, and 12 months.
Data were collected using the VA Computerized Patient Record System (CPRS) and stored in a locked spreadsheet. Descriptive statistics were used to analyze the data. Patient data included the number of patients on insulin or a sulfonylurea when initiating a GLP-1 agonist, the percentage of patients started on a certain GLP-1 agonist (dulaglutide, liraglutide, exenatide, and semaglutide), and the percentage of patients with a baseline HbA_1c of < 8%, 8% to 10%, and > 10%. The GLP-1 agonist formulary was adjusted during the time of this retrospective chart review. Patients who were not on semaglutide were switched over if they were on another GLP-1 agonist as semaglutide became the preferred GLP-1 agonist.

Patients were considered to have a dose reduction or discontinuation of insulin or a sulfonylurea if the dose or medication they were on decreased or was discontinued permanently within 12 months of starting a GLP-1 agonist. For example, if a patient who was administering 10 units of insulin daily was decreased to 8 but later increased back to 10, this was not counted as a dose reduction. If a patient discontinued insulin or a sulfonylurea and then restarted it within 12 months of initiating a GLP-1 agonist, this was not counted as a discontinuation.

This retrospective review included 136 patients; 96 patients taking insulin and 54 taking a sulfonylurea when they started a GLP-1 agonist. Fourteen patients were on both. Criteria for use, which are clinical criteria to determine if a patient is eligible for the use of a given medication, are used within the VA. The inclusion criteria for a patient initiating a GLP-1 agonist is that the patient must have atherosclerotic cardiovascular disease or chronic kidney disease with the patient receiving metformin (unless unable to use metformin) and empagliflozin (unless unable to use empagliflozin).

The baseline mean age and weight for the patient population in this retrospective chart review was 70.7 years and 238.2 lb, respectively. Ninety-six patients (70.6%) were started on semaglutide, 27 (19.9%) on dulaglutide, 12 (8.8%) on liraglutide, and 1 (0.7%) on exenatide. The mean HbA_1c when patients initiated a GLP-1 agonist was 8.6%. When starting a GLP-1 agonist, 34 patients (25.0%) had an HbA_1c < 8%, 89 (65.4%) had an HbA_1c between 8% to 10%, and 13 (9.6%) had an HbA_1c > 10% (Table).

For the primary results, 25 patients (26.0%) had a dose reduction of insulin when they started a GLP-1 agonist, and 55 patients (57.3%) had at least 1 insulin dose reduction within the year follow-up. Seven patients (13.0%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 16 patients (29.6%) had at least 1 dose reduction of a sulfonylurea within the year follow-up. Six patients (6.3%) discontinued insulin use when they initially started a GLP-1 agonist, and 14 patients (14.6%) discontinued insulin use within the year follow-up. Eleven patients (20.4%) discontinued sulfonylurea use when they initially started a GLP-1 agonist, and 21 patients (38.9%) discontinued sulfonylurea use within the year follow-up (Figure).

Fourteen patients were on both insulin and a sulfonylurea. Two patients (14.3%) had a dose reduction of insulin when they started a GLP-1 agonist, and 5 (35.7%) had ≥ 1 insulin dose reduction within the year follow-up. Three patients (21.4%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 6 (42.9%) had ≥ 1 dose reduction of a sulfonylurea within the year follow-up. Seven patients (50.0%) discontinued sulfonylurea and 3 (21.4%) discontinued insulin at any time throughout the year. The majority of the discontinuations were at the initial start of GLP-1 agonist therapy.

The mean HbA_1c for patients on GLP-1 agonist was 8.6% at baseline, 8.0% at 0 to 3 months, 7.6% at 3 to 6 months, and 7.5% at 12 months. Patients experienced a mean HbA_1c reduction of 1.1%. The mean weight when a GLP-1 agonist was started was 238.2 lb, 236.0 lb at 0 to 3 months, 223.8 lb at 3 to 6 months, and 224.3 lb after 12 months. Study participants lost a mean weight of 13.9 lb while on a GLP-1 agonist.

Discussion

While this study did not examine why there were dose reductions or discontinuations, we can hypothesize that insulin or sulfonylureas were reduced or discontinued due to a myriad of reasons, such as prophylactic dosing per guidelines, patients having a hypoglycemic event, or pharmacists anticipating potential low blood glucose trends. Also, there could have been numerous reasons GLP-1 agonists were started in patients on insulin or a sulfonylurea, such as HbA_1c not being within goal range, cardiovascular benefits (reduce risk of stroke, heart attack, and death), weight loss, and renal protection, such as preventing albuminuria.^13,14

This retrospective chart review found a large proportion of patients had a dose reduction of insulin (57.3%) or sulfonylurea (29.6%). The percentage of patients with a dose reduction was potentially underestimated as patients were not counted if they discontinued insulin or sulfonylurea. Concomitant use of GLP-1 agonists with insulin or a sulfonylurea may increase the risk of hypoglycemia and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 The dose reductions in this study show that pharmacists within pharmacy-led PACT clinics monitor for or attempt to prevent hypoglycemia, which aligns with the prescribing information of GLP-1 agonists. While increasing the insulin or sulfonylurea dose is an option for patients, this approach can increase the risk of hypoglycemia, especially in an older population, like this one with a mean age > 70 years. The large proportions of patients with dose reductions or insulin and sulfonylurea discontinuations suggest that pharmacists may need to take a more cautious approach when initiating a GLP-1 agonist to prevent adverse health outcomes related to low blood sugar for older adults, such as falls and fractures.

Insulin was discontinued in 20.4% of patients and sulfonylurea was discontinued in 38.9% of patients within 12 months after starting a GLP-1 agonist. When a patient was on both insulin and a sulfonylurea, the percentage of patients who discontinued insulin (21.4%) or a sulfonylurea (50.0%) was higher compared with patients just on insulin (14.6%) or a sulfonylurea (38.9%) alone. Patients on both insulin and a sulfonylurea may need closer monitoring due to a higher incidence of discontinuations when these diabetes agents are administered in combination.

Limitations

Several limitations of this study should be considered when evaluating the results. This review was comprised of a mostly older, male population, which results in a low generalizability to organizations other than VA medical centers. In addition, this study only evaluated patients on a GLP-1 agonist followed in a pharmacist-led PACT clinic. This study excluded patients who were prescribed a GLP-1 agonist by an endocrinologist or a pharmacist at one of the community-based outpatient clinics affiliated with WBVAMC, or a pharmacist or clinician outside the VA. The sole focus of this study was patients in a pharmacist-led VAMC clinic. Not all patient data may have been included in the study. If a patient did not have an appointment at baseline, 3, 6, and 12 months or did not obtain laboratory tests, HbA_1c and weights were not recorded. Data were collected during the COVID-19 pandemic and in-person appointments were potentially switched to phone or video appointments. There were many instances during this chart review where a weight was not recorded at each time interval. Also, this study did not consider any other diabetes medications the patient was taking. There were many instances where the patient was taking metformin and/or sodium-glucose cotransporter-2 (SGLT-2) inhibitors. These medications along with diet could have affected the weight results as metformin is weight neutral and SGLT-2 inhibitors promote weight loss.¹⁵ Lastly, this study did not evaluate the amount of insulin reduced, only if there was a dose reduction or discontinuation of insulin and/or a sulfonylurea.

Conclusions

Dose reductions and a discontinuation of insulin or a sulfonylurea with the addition of a GLP-1 agonist may be needed. Patients on both insulin and a sulfonylurea may need closer monitoring due to the higher incidences of discontinuations compared with patients on just 1 of these agents. Dose reductions or discontinuations of these diabetic agents can promote positive patient outcomes, such as preventing hypoglycemia, minimizing weight gain, increasing weight loss, and reducing HbA_1c levels.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the Wilkes-Barre Veterans Affairs Medical Center in Pennsylvania.

Hypoglycemia and weight gain are well-known adverse effects that can result from insulin and sulfonylureas in patients with type 2 diabetes mellitus (T2DM).^1,2 Insulin and sulfonylurea medications can cause additional weight gain in patients who are overweight or obese, which can increase the burden of diabetes therapy with added medications, raise the risk of hypoglycemia complications, and raise atherosclerotic cardiovascular disease risk factors.³ Although increasing the insulin or sulfonylurea dose is an option health care practitioners or pharmacists have, this approach can increase the risk of hypoglycemia, especially in older adults, such as the veteran population, which could lead to complications, such as falls.²

Previous studies focusing on hypoglycemic events in patients with T2DM showed that glucagon-like peptide-1 (GLP-1) agonist monotherapy has a low incidence of a hypoglycemic events. However, when a GLP-1 agonist is combined with insulin or sulfonylureas, patients have an increased chance of a hypoglycemic event.^3-8 According to the prescribing information for semaglutide, 1.6% to 3.8% of patients on a GLP-1 agonist monotherapy reported a documented symptomatic hypoglycemic event (blood glucose ≤ 70 mg/dL), based on semaglutide dosing. ⁹ Patients on combination therapy of a GLP-1 agonist and basal insulin and a GLP-1 agonist and a sulfonylurea reported a documented symptomatic hypoglycemic event ranging from 16.7% to 29.8% and 17.3% to 24.4%, respectively.⁹ The incidences of hypoglycemia thus dramatically increase with combination therapy of a GLP-1 agonist plus insulin or a sulfonylurea.

When adding a GLP-1 agonist to insulin or a sulfonylurea, clinicians must be mindful of the increased risk of hypoglycemia. Per the warnings and precautions in the prescribing information of GLP-1 agonists, concomitant use with insulin or a sulfonylurea may increase the risk of hypoglycemia, and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 According to the American College of Cardiology guidelines, when starting a GLP-1 agonist, the insulin dose should be decreased by about 20% in patients with a well-controlled hemoglobin A_1c (HbA_1c).¹²

This study aimed to determine the percentage of patients who required dose reductions or discontinuations of insulin and sulfonylureas with the addition of a GLP-1 agonist. Understanding necessary dose reductions or discontinuations of these concomitant diabetes agents can assist pharmacists in preventing hypoglycemia and minimizing weight gain.

Methods

This clinical review was a single-center, retrospective chart review of patients prescribed a GLP-1 agonist while on insulin or a sulfonylurea between January 1, 2019, and September 30, 2022, at the Wilkes-Barre Veterans Affairs Medical Center (WBVAMC) in Pennsylvania and managed in a pharmacist-led patient aligned care team (PACT) clinic. It was determined by the US Department of Veterans Affairs Office of Research and Development that an institutional review board or other review committee approval was not needed for this nonresearch Veterans Health Administration quality assurance and improvement project. Patients aged ≥ 18 years were included in this study. Patients were excluded if they were not on insulin or a sulfonylurea when starting a GLP-1 agonist, started a GLP-1 agonist outside of the retrospective chart review dates, or were prescribed a GLP-1 agonist by anyone other than a pharmacist in their PACT clinic. This included if a GLP-1 agonist was prescribed by a primary care physician, endocrinologist, or someone outside the VA system.

The primary study outcomes were to determine the percentage of patients with a dose reduction of insulin or sulfonylurea and discontinuation of insulin or a sulfonylurea at intervals of 0 (baseline), 3, 6, and 12 months. Secondary outcomes included changes in HbA_1c and body weight measured at the same intervals of 0 (baseline), 3, 6, and 12 months.
Data were collected using the VA Computerized Patient Record System (CPRS) and stored in a locked spreadsheet. Descriptive statistics were used to analyze the data. Patient data included the number of patients on insulin or a sulfonylurea when initiating a GLP-1 agonist, the percentage of patients started on a certain GLP-1 agonist (dulaglutide, liraglutide, exenatide, and semaglutide), and the percentage of patients with a baseline HbA_1c of < 8%, 8% to 10%, and > 10%. The GLP-1 agonist formulary was adjusted during the time of this retrospective chart review. Patients who were not on semaglutide were switched over if they were on another GLP-1 agonist as semaglutide became the preferred GLP-1 agonist.

Patients were considered to have a dose reduction or discontinuation of insulin or a sulfonylurea if the dose or medication they were on decreased or was discontinued permanently within 12 months of starting a GLP-1 agonist. For example, if a patient who was administering 10 units of insulin daily was decreased to 8 but later increased back to 10, this was not counted as a dose reduction. If a patient discontinued insulin or a sulfonylurea and then restarted it within 12 months of initiating a GLP-1 agonist, this was not counted as a discontinuation.

This retrospective review included 136 patients; 96 patients taking insulin and 54 taking a sulfonylurea when they started a GLP-1 agonist. Fourteen patients were on both. Criteria for use, which are clinical criteria to determine if a patient is eligible for the use of a given medication, are used within the VA. The inclusion criteria for a patient initiating a GLP-1 agonist is that the patient must have atherosclerotic cardiovascular disease or chronic kidney disease with the patient receiving metformin (unless unable to use metformin) and empagliflozin (unless unable to use empagliflozin).

The baseline mean age and weight for the patient population in this retrospective chart review was 70.7 years and 238.2 lb, respectively. Ninety-six patients (70.6%) were started on semaglutide, 27 (19.9%) on dulaglutide, 12 (8.8%) on liraglutide, and 1 (0.7%) on exenatide. The mean HbA_1c when patients initiated a GLP-1 agonist was 8.6%. When starting a GLP-1 agonist, 34 patients (25.0%) had an HbA_1c < 8%, 89 (65.4%) had an HbA_1c between 8% to 10%, and 13 (9.6%) had an HbA_1c > 10% (Table).

For the primary results, 25 patients (26.0%) had a dose reduction of insulin when they started a GLP-1 agonist, and 55 patients (57.3%) had at least 1 insulin dose reduction within the year follow-up. Seven patients (13.0%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 16 patients (29.6%) had at least 1 dose reduction of a sulfonylurea within the year follow-up. Six patients (6.3%) discontinued insulin use when they initially started a GLP-1 agonist, and 14 patients (14.6%) discontinued insulin use within the year follow-up. Eleven patients (20.4%) discontinued sulfonylurea use when they initially started a GLP-1 agonist, and 21 patients (38.9%) discontinued sulfonylurea use within the year follow-up (Figure).

Fourteen patients were on both insulin and a sulfonylurea. Two patients (14.3%) had a dose reduction of insulin when they started a GLP-1 agonist, and 5 (35.7%) had ≥ 1 insulin dose reduction within the year follow-up. Three patients (21.4%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 6 (42.9%) had ≥ 1 dose reduction of a sulfonylurea within the year follow-up. Seven patients (50.0%) discontinued sulfonylurea and 3 (21.4%) discontinued insulin at any time throughout the year. The majority of the discontinuations were at the initial start of GLP-1 agonist therapy.

The mean HbA_1c for patients on GLP-1 agonist was 8.6% at baseline, 8.0% at 0 to 3 months, 7.6% at 3 to 6 months, and 7.5% at 12 months. Patients experienced a mean HbA_1c reduction of 1.1%. The mean weight when a GLP-1 agonist was started was 238.2 lb, 236.0 lb at 0 to 3 months, 223.8 lb at 3 to 6 months, and 224.3 lb after 12 months. Study participants lost a mean weight of 13.9 lb while on a GLP-1 agonist.

Discussion

While this study did not examine why there were dose reductions or discontinuations, we can hypothesize that insulin or sulfonylureas were reduced or discontinued due to a myriad of reasons, such as prophylactic dosing per guidelines, patients having a hypoglycemic event, or pharmacists anticipating potential low blood glucose trends. Also, there could have been numerous reasons GLP-1 agonists were started in patients on insulin or a sulfonylurea, such as HbA_1c not being within goal range, cardiovascular benefits (reduce risk of stroke, heart attack, and death), weight loss, and renal protection, such as preventing albuminuria.^13,14

This retrospective chart review found a large proportion of patients had a dose reduction of insulin (57.3%) or sulfonylurea (29.6%). The percentage of patients with a dose reduction was potentially underestimated as patients were not counted if they discontinued insulin or sulfonylurea. Concomitant use of GLP-1 agonists with insulin or a sulfonylurea may increase the risk of hypoglycemia and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 The dose reductions in this study show that pharmacists within pharmacy-led PACT clinics monitor for or attempt to prevent hypoglycemia, which aligns with the prescribing information of GLP-1 agonists. While increasing the insulin or sulfonylurea dose is an option for patients, this approach can increase the risk of hypoglycemia, especially in an older population, like this one with a mean age > 70 years. The large proportions of patients with dose reductions or insulin and sulfonylurea discontinuations suggest that pharmacists may need to take a more cautious approach when initiating a GLP-1 agonist to prevent adverse health outcomes related to low blood sugar for older adults, such as falls and fractures.

Insulin was discontinued in 20.4% of patients and sulfonylurea was discontinued in 38.9% of patients within 12 months after starting a GLP-1 agonist. When a patient was on both insulin and a sulfonylurea, the percentage of patients who discontinued insulin (21.4%) or a sulfonylurea (50.0%) was higher compared with patients just on insulin (14.6%) or a sulfonylurea (38.9%) alone. Patients on both insulin and a sulfonylurea may need closer monitoring due to a higher incidence of discontinuations when these diabetes agents are administered in combination.

Limitations

Several limitations of this study should be considered when evaluating the results. This review was comprised of a mostly older, male population, which results in a low generalizability to organizations other than VA medical centers. In addition, this study only evaluated patients on a GLP-1 agonist followed in a pharmacist-led PACT clinic. This study excluded patients who were prescribed a GLP-1 agonist by an endocrinologist or a pharmacist at one of the community-based outpatient clinics affiliated with WBVAMC, or a pharmacist or clinician outside the VA. The sole focus of this study was patients in a pharmacist-led VAMC clinic. Not all patient data may have been included in the study. If a patient did not have an appointment at baseline, 3, 6, and 12 months or did not obtain laboratory tests, HbA_1c and weights were not recorded. Data were collected during the COVID-19 pandemic and in-person appointments were potentially switched to phone or video appointments. There were many instances during this chart review where a weight was not recorded at each time interval. Also, this study did not consider any other diabetes medications the patient was taking. There were many instances where the patient was taking metformin and/or sodium-glucose cotransporter-2 (SGLT-2) inhibitors. These medications along with diet could have affected the weight results as metformin is weight neutral and SGLT-2 inhibitors promote weight loss.¹⁵ Lastly, this study did not evaluate the amount of insulin reduced, only if there was a dose reduction or discontinuation of insulin and/or a sulfonylurea.

Conclusions

Dose reductions and a discontinuation of insulin or a sulfonylurea with the addition of a GLP-1 agonist may be needed. Patients on both insulin and a sulfonylurea may need closer monitoring due to the higher incidences of discontinuations compared with patients on just 1 of these agents. Dose reductions or discontinuations of these diabetic agents can promote positive patient outcomes, such as preventing hypoglycemia, minimizing weight gain, increasing weight loss, and reducing HbA_1c levels.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the Wilkes-Barre Veterans Affairs Medical Center in Pennsylvania.

Hypoglycemia and weight gain are well-known adverse effects that can result from insulin and sulfonylureas in patients with type 2 diabetes mellitus (T2DM).^1,2 Insulin and sulfonylurea medications can cause additional weight gain in patients who are overweight or obese, which can increase the burden of diabetes therapy with added medications, raise the risk of hypoglycemia complications, and raise atherosclerotic cardiovascular disease risk factors.³ Although increasing the insulin or sulfonylurea dose is an option health care practitioners or pharmacists have, this approach can increase the risk of hypoglycemia, especially in older adults, such as the veteran population, which could lead to complications, such as falls.²

Previous studies focusing on hypoglycemic events in patients with T2DM showed that glucagon-like peptide-1 (GLP-1) agonist monotherapy has a low incidence of a hypoglycemic events. However, when a GLP-1 agonist is combined with insulin or sulfonylureas, patients have an increased chance of a hypoglycemic event.^3-8 According to the prescribing information for semaglutide, 1.6% to 3.8% of patients on a GLP-1 agonist monotherapy reported a documented symptomatic hypoglycemic event (blood glucose ≤ 70 mg/dL), based on semaglutide dosing. ⁹ Patients on combination therapy of a GLP-1 agonist and basal insulin and a GLP-1 agonist and a sulfonylurea reported a documented symptomatic hypoglycemic event ranging from 16.7% to 29.8% and 17.3% to 24.4%, respectively.⁹ The incidences of hypoglycemia thus dramatically increase with combination therapy of a GLP-1 agonist plus insulin or a sulfonylurea.

When adding a GLP-1 agonist to insulin or a sulfonylurea, clinicians must be mindful of the increased risk of hypoglycemia. Per the warnings and precautions in the prescribing information of GLP-1 agonists, concomitant use with insulin or a sulfonylurea may increase the risk of hypoglycemia, and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 According to the American College of Cardiology guidelines, when starting a GLP-1 agonist, the insulin dose should be decreased by about 20% in patients with a well-controlled hemoglobin A_1c (HbA_1c).¹²

This study aimed to determine the percentage of patients who required dose reductions or discontinuations of insulin and sulfonylureas with the addition of a GLP-1 agonist. Understanding necessary dose reductions or discontinuations of these concomitant diabetes agents can assist pharmacists in preventing hypoglycemia and minimizing weight gain.

Methods

This clinical review was a single-center, retrospective chart review of patients prescribed a GLP-1 agonist while on insulin or a sulfonylurea between January 1, 2019, and September 30, 2022, at the Wilkes-Barre Veterans Affairs Medical Center (WBVAMC) in Pennsylvania and managed in a pharmacist-led patient aligned care team (PACT) clinic. It was determined by the US Department of Veterans Affairs Office of Research and Development that an institutional review board or other review committee approval was not needed for this nonresearch Veterans Health Administration quality assurance and improvement project. Patients aged ≥ 18 years were included in this study. Patients were excluded if they were not on insulin or a sulfonylurea when starting a GLP-1 agonist, started a GLP-1 agonist outside of the retrospective chart review dates, or were prescribed a GLP-1 agonist by anyone other than a pharmacist in their PACT clinic. This included if a GLP-1 agonist was prescribed by a primary care physician, endocrinologist, or someone outside the VA system.

The primary study outcomes were to determine the percentage of patients with a dose reduction of insulin or sulfonylurea and discontinuation of insulin or a sulfonylurea at intervals of 0 (baseline), 3, 6, and 12 months. Secondary outcomes included changes in HbA_1c and body weight measured at the same intervals of 0 (baseline), 3, 6, and 12 months.
Data were collected using the VA Computerized Patient Record System (CPRS) and stored in a locked spreadsheet. Descriptive statistics were used to analyze the data. Patient data included the number of patients on insulin or a sulfonylurea when initiating a GLP-1 agonist, the percentage of patients started on a certain GLP-1 agonist (dulaglutide, liraglutide, exenatide, and semaglutide), and the percentage of patients with a baseline HbA_1c of < 8%, 8% to 10%, and > 10%. The GLP-1 agonist formulary was adjusted during the time of this retrospective chart review. Patients who were not on semaglutide were switched over if they were on another GLP-1 agonist as semaglutide became the preferred GLP-1 agonist.

Patients were considered to have a dose reduction or discontinuation of insulin or a sulfonylurea if the dose or medication they were on decreased or was discontinued permanently within 12 months of starting a GLP-1 agonist. For example, if a patient who was administering 10 units of insulin daily was decreased to 8 but later increased back to 10, this was not counted as a dose reduction. If a patient discontinued insulin or a sulfonylurea and then restarted it within 12 months of initiating a GLP-1 agonist, this was not counted as a discontinuation.

This retrospective review included 136 patients; 96 patients taking insulin and 54 taking a sulfonylurea when they started a GLP-1 agonist. Fourteen patients were on both. Criteria for use, which are clinical criteria to determine if a patient is eligible for the use of a given medication, are used within the VA. The inclusion criteria for a patient initiating a GLP-1 agonist is that the patient must have atherosclerotic cardiovascular disease or chronic kidney disease with the patient receiving metformin (unless unable to use metformin) and empagliflozin (unless unable to use empagliflozin).

The baseline mean age and weight for the patient population in this retrospective chart review was 70.7 years and 238.2 lb, respectively. Ninety-six patients (70.6%) were started on semaglutide, 27 (19.9%) on dulaglutide, 12 (8.8%) on liraglutide, and 1 (0.7%) on exenatide. The mean HbA_1c when patients initiated a GLP-1 agonist was 8.6%. When starting a GLP-1 agonist, 34 patients (25.0%) had an HbA_1c < 8%, 89 (65.4%) had an HbA_1c between 8% to 10%, and 13 (9.6%) had an HbA_1c > 10% (Table).

For the primary results, 25 patients (26.0%) had a dose reduction of insulin when they started a GLP-1 agonist, and 55 patients (57.3%) had at least 1 insulin dose reduction within the year follow-up. Seven patients (13.0%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 16 patients (29.6%) had at least 1 dose reduction of a sulfonylurea within the year follow-up. Six patients (6.3%) discontinued insulin use when they initially started a GLP-1 agonist, and 14 patients (14.6%) discontinued insulin use within the year follow-up. Eleven patients (20.4%) discontinued sulfonylurea use when they initially started a GLP-1 agonist, and 21 patients (38.9%) discontinued sulfonylurea use within the year follow-up (Figure).

Fourteen patients were on both insulin and a sulfonylurea. Two patients (14.3%) had a dose reduction of insulin when they started a GLP-1 agonist, and 5 (35.7%) had ≥ 1 insulin dose reduction within the year follow-up. Three patients (21.4%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 6 (42.9%) had ≥ 1 dose reduction of a sulfonylurea within the year follow-up. Seven patients (50.0%) discontinued sulfonylurea and 3 (21.4%) discontinued insulin at any time throughout the year. The majority of the discontinuations were at the initial start of GLP-1 agonist therapy.

The mean HbA_1c for patients on GLP-1 agonist was 8.6% at baseline, 8.0% at 0 to 3 months, 7.6% at 3 to 6 months, and 7.5% at 12 months. Patients experienced a mean HbA_1c reduction of 1.1%. The mean weight when a GLP-1 agonist was started was 238.2 lb, 236.0 lb at 0 to 3 months, 223.8 lb at 3 to 6 months, and 224.3 lb after 12 months. Study participants lost a mean weight of 13.9 lb while on a GLP-1 agonist.

Discussion

While this study did not examine why there were dose reductions or discontinuations, we can hypothesize that insulin or sulfonylureas were reduced or discontinued due to a myriad of reasons, such as prophylactic dosing per guidelines, patients having a hypoglycemic event, or pharmacists anticipating potential low blood glucose trends. Also, there could have been numerous reasons GLP-1 agonists were started in patients on insulin or a sulfonylurea, such as HbA_1c not being within goal range, cardiovascular benefits (reduce risk of stroke, heart attack, and death), weight loss, and renal protection, such as preventing albuminuria.^13,14

This retrospective chart review found a large proportion of patients had a dose reduction of insulin (57.3%) or sulfonylurea (29.6%). The percentage of patients with a dose reduction was potentially underestimated as patients were not counted if they discontinued insulin or sulfonylurea. Concomitant use of GLP-1 agonists with insulin or a sulfonylurea may increase the risk of hypoglycemia and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 The dose reductions in this study show that pharmacists within pharmacy-led PACT clinics monitor for or attempt to prevent hypoglycemia, which aligns with the prescribing information of GLP-1 agonists. While increasing the insulin or sulfonylurea dose is an option for patients, this approach can increase the risk of hypoglycemia, especially in an older population, like this one with a mean age > 70 years. The large proportions of patients with dose reductions or insulin and sulfonylurea discontinuations suggest that pharmacists may need to take a more cautious approach when initiating a GLP-1 agonist to prevent adverse health outcomes related to low blood sugar for older adults, such as falls and fractures.

Insulin was discontinued in 20.4% of patients and sulfonylurea was discontinued in 38.9% of patients within 12 months after starting a GLP-1 agonist. When a patient was on both insulin and a sulfonylurea, the percentage of patients who discontinued insulin (21.4%) or a sulfonylurea (50.0%) was higher compared with patients just on insulin (14.6%) or a sulfonylurea (38.9%) alone. Patients on both insulin and a sulfonylurea may need closer monitoring due to a higher incidence of discontinuations when these diabetes agents are administered in combination.

Limitations

Several limitations of this study should be considered when evaluating the results. This review was comprised of a mostly older, male population, which results in a low generalizability to organizations other than VA medical centers. In addition, this study only evaluated patients on a GLP-1 agonist followed in a pharmacist-led PACT clinic. This study excluded patients who were prescribed a GLP-1 agonist by an endocrinologist or a pharmacist at one of the community-based outpatient clinics affiliated with WBVAMC, or a pharmacist or clinician outside the VA. The sole focus of this study was patients in a pharmacist-led VAMC clinic. Not all patient data may have been included in the study. If a patient did not have an appointment at baseline, 3, 6, and 12 months or did not obtain laboratory tests, HbA_1c and weights were not recorded. Data were collected during the COVID-19 pandemic and in-person appointments were potentially switched to phone or video appointments. There were many instances during this chart review where a weight was not recorded at each time interval. Also, this study did not consider any other diabetes medications the patient was taking. There were many instances where the patient was taking metformin and/or sodium-glucose cotransporter-2 (SGLT-2) inhibitors. These medications along with diet could have affected the weight results as metformin is weight neutral and SGLT-2 inhibitors promote weight loss.¹⁵ Lastly, this study did not evaluate the amount of insulin reduced, only if there was a dose reduction or discontinuation of insulin and/or a sulfonylurea.

Conclusions

Dose reductions and a discontinuation of insulin or a sulfonylurea with the addition of a GLP-1 agonist may be needed. Patients on both insulin and a sulfonylurea may need closer monitoring due to the higher incidences of discontinuations compared with patients on just 1 of these agents. Dose reductions or discontinuations of these diabetic agents can promote positive patient outcomes, such as preventing hypoglycemia, minimizing weight gain, increasing weight loss, and reducing HbA_1c levels.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the Wilkes-Barre Veterans Affairs Medical Center in Pennsylvania.

Competency in paracentesis is an important procedural skill for medical practitioners caring for patients with decompensated liver cirrhosis. Paracentesis is performed to drain ascitic fluid for both diagnosis and/or therapeutic purposes.¹ While this procedure can be performed without the use of ultrasound, it is preferable to use ultrasound to identify an area of fluid that is away from dangerous anatomy including bowel loops, the liver, and spleen. After prepping the area, lidocaine is administered locally. A catheter is then inserted until fluid begins flowing freely. The catheter is connected to a suction canister or collection kit, and the patient is monitored until the flow ceases. Samples can be sent for analysis to determine the etiology of ascites, identify concerns for infection, and more.

Paracentesis is a very common procedure. Barsuk and colleagues noted that between 2010 and 2012, 97,577 procedures were performed across 120 academic medical centers and 290 affiliated hospitals.² Patients undergo paracentesis in a variety of settings including the emergency department, inpatient hospitalizations, and clinics. Some patients may require only 1 paracentesis procedure while others may require it regularly.

Due to the rising need for paracentesis in the Central Texas Veterans Affairs Hospital (CTVAH) in Temple, a paracentesis clinic was started in February 2018. The goal of the paracentesis clinic was multifocal—to reduce hospital admissions, improve access to regularly scheduled procedures, decrease wait times, and increase patient satisfaction.³ Through the CTVAH affiliation with the Texas A&M internal medicine residency program, the paracentesis clinic started involving and training residents on this procedure. Up to 3 residents are on weekly rotation and can perform up to 6 paracentesis procedures in a week. The purpose of this article was to evaluate resident competency in paracentesis after completion of the paracentesis clinic.

Methods

The paracentesis clinic schedules up to 3 patients on Tuesdays and Thursdays between 8 am and noon. All the necessary equipment is readily available and includes the paracentesis kit, lidocaine, sterile gloves, ultrasound, and albumin if needed. Because this procedure is performed at the hospital, direct access to the emergency department is available. Residents are scheduled weekly. Up to 2 residents are scheduled for the paracentesis clinic during their dedicated clinic week. They are expected to practice obtaining consent, performing the procedure, and documenting the encounter under staff supervision. Additionally, 1 or 2 residents participate in the paracentesis clinic as part of an ultrasound elective twice per year. In this elective, they practice ultrasound skills using a simulation and translate that information in the paracentesis clinic while identifying anatomy and performing the paracentesis procedure under staff supervision.

A survey was sent via email to all categorical internal medicine residents across all 3 program years at the time of data collection. Competency for paracentesis sign-off was defined as completing and logging 5 procedures supervised by a competent physician who confirmed that all portions of the procedure were performed correctly. Residents were also asked to answer questions on a scale from 1 to 10, with 1 representing no confidence and 10 representing strong confidence to practice independently (Table).

We also evaluated the number of procedures performed by internal medicine residents 3 years before the clinic was started in 2015 up to the completion of 2022. The numbers were obtained by examining procedural log data for each year for all internal medicine residents.

Results

Thirty-three residents completed the survey: 10 first-year internal medicine residents (PGY1), 12 second-year residents (PGY2), and 11 third-year residents (PGY3). The mean participation was 4.8 paracentesis sessions per person for the duration of the study. The range of paracentesis procedures performed varied based on PGY year: PGY1s performed 1 to > 10 procedures, PGY2s performed 2 to > 10 procedures, and PGY3s performed 5 to > 10 procedures. Thirty-six percent of residents completed > 10 procedures in the paracentesis clinic; 82% of PGY3s had completed > 10 procedures by December of their third year. Twenty-six residents (79%) were credentialed to perform paracentesis procedures independently after performing > 5 procedures, and 7 residents were not yet cleared for procedural independence.

In the survey, residents rated their comfort with performing paracentesis procedures independently at a mean of 7.9. The mean comfort reported by PGY1s was 7.2, PGY2s was 7.3, and PGY3s was 9.3. Residents also rated their opinion on whether or not the paracentesis clinic adequately prepared them for paracentesis procedural independence; the mean was 8.9 across all residents.

The total number of procedures performed by residents at CTVAH also increased. Starting in 2015, 3 years before the clinic was started, 38 procedures were performed by internal medicine residents, followed by 72 procedures in 2016; 76 in 2017; 58 in 2018; 94 in 2019; 88 in 2020; 136 in 2021; and 188 in 2022.

Paracentesis is a simple but invasive procedure to relieve ascites, often relieving patients’ symptoms, preventing hospital admission, and increasing patient satisfaction.⁴ The CTVAH does not have the capacity to perform outpatient paracentesis effectively in its emergency or radiology departments. Furthermore, the use of the emergency or radiology departments for routine paracentesis may not be feasible due to the acuity of care being provided, as these procedures can be time consuming and can draw away critical resources and time from patients that need emergent care. The paracentesis clinic was then formed to provide veterans access to the procedural care they need, while also preparing residents to ably and confidently perform the procedure independently.

Based on our study, most residents were cleared to independently perform paracentesis procedures across all 3 years, with 79% of residents having completed the required 5 supervised procedures to independently practice. A study assessing unsupervised practice standards showed that paracentesis skill declines as soon as 3 months after training. However, retraining was shown to potentially interrupt this skill decline.⁵ Studies have shown that procedure-driven electives or services significantly improved paracentesis certification rates and total logged procedures, with minimal funding or scheduling changes required.⁶ Our clinic showed a significant increase in the number of procedures logged starting with the minimum of 38 procedures in 2015 and ending with 188 procedures logged at the end of 2022.

By allowing residents to routinely return to the paracentesis clinic across all 3 years, residents were more likely to feel comfortable independently performing the procedure, with residents reporting a mean comfort score of 7.9. The spaced repetition and ability to work with the clinic during elective time allows regular opportunities to undergo supervised training in a controlled environment and created scheduled retraining opportunities. Future studies should evaluate residents prior to each paracentesis clinic to ascertain if skill decline is occurring at a slower rate.

The inpatient effect of the clinic is also multifocal. Pham and colleagues showed that integrating paracentesis into timely training can reduce paracentesis delay and delays in care.⁷ By increasing the volume of procedures each resident performs and creating a sense of confidence amongst residents, the clinic increases the number of residents able and willing to perform inpatient procedures, thus reducing the number of unnecessary consultations and hospital resources. One of the reasons the paracentesis clinic was started was to allow patients to have scheduled times to remove fluid from their abdomen, thus cutting down on emergency department procedures and unnecessary admissions. Additionally, the benefits of early paracentesis procedural performance by residents and internal medicine physicians have been demonstrated in the literature. A study by Gaetano and colleagues noted that patients undergoing early paracentesis had reduced mortality of 5.5% vs 7.5% in those undergoing late paracentesis.⁸ This study also showed the in-hospital mortality rate was decreased with paracentesis (6.3%) vs without paracentesis (8.9%).⁸ By offering residents a chance to participate in the clinic, we have shown that regular opportunities to perform paracentesis may increase the number of physicians capable of independently practicing, improve procedural competency, and improve patient access to this procedure.

Limitations

Our study was not free of bias and has potential weaknesses. The survey was sent to all current residents who have participated in the paracentesis clinic, but not every resident filled out the survey (55% of all residents across 3 years completed the survey, 68.7% who had done clinic that year completed the survey). There is a possibility that those not signed off avoided doing the survey, but we are unable to confirm this. The survey also depended on resident recall of the number of paracenteses completed or looking at their procedure log. It is possible that some procedures were not documented, changing the true number. Additionally, rating comfortability with procedures is subjective, which may also create a source of potential weakness. Future projects should include a baseline survey for residents, followed by a repeat survey a year later to show changes from baseline competency.

Conclusions

A dedicated paracentesis clinic with internal medicine resident involvement may increase resident paracentesis procedural independence, the number of procedures available and performed, and procedural comfort level.

Competency in paracentesis is an important procedural skill for medical practitioners caring for patients with decompensated liver cirrhosis. Paracentesis is performed to drain ascitic fluid for both diagnosis and/or therapeutic purposes.¹ While this procedure can be performed without the use of ultrasound, it is preferable to use ultrasound to identify an area of fluid that is away from dangerous anatomy including bowel loops, the liver, and spleen. After prepping the area, lidocaine is administered locally. A catheter is then inserted until fluid begins flowing freely. The catheter is connected to a suction canister or collection kit, and the patient is monitored until the flow ceases. Samples can be sent for analysis to determine the etiology of ascites, identify concerns for infection, and more.

Paracentesis is a very common procedure. Barsuk and colleagues noted that between 2010 and 2012, 97,577 procedures were performed across 120 academic medical centers and 290 affiliated hospitals.² Patients undergo paracentesis in a variety of settings including the emergency department, inpatient hospitalizations, and clinics. Some patients may require only 1 paracentesis procedure while others may require it regularly.

Due to the rising need for paracentesis in the Central Texas Veterans Affairs Hospital (CTVAH) in Temple, a paracentesis clinic was started in February 2018. The goal of the paracentesis clinic was multifocal—to reduce hospital admissions, improve access to regularly scheduled procedures, decrease wait times, and increase patient satisfaction.³ Through the CTVAH affiliation with the Texas A&M internal medicine residency program, the paracentesis clinic started involving and training residents on this procedure. Up to 3 residents are on weekly rotation and can perform up to 6 paracentesis procedures in a week. The purpose of this article was to evaluate resident competency in paracentesis after completion of the paracentesis clinic.

Methods

The paracentesis clinic schedules up to 3 patients on Tuesdays and Thursdays between 8 am and noon. All the necessary equipment is readily available and includes the paracentesis kit, lidocaine, sterile gloves, ultrasound, and albumin if needed. Because this procedure is performed at the hospital, direct access to the emergency department is available. Residents are scheduled weekly. Up to 2 residents are scheduled for the paracentesis clinic during their dedicated clinic week. They are expected to practice obtaining consent, performing the procedure, and documenting the encounter under staff supervision. Additionally, 1 or 2 residents participate in the paracentesis clinic as part of an ultrasound elective twice per year. In this elective, they practice ultrasound skills using a simulation and translate that information in the paracentesis clinic while identifying anatomy and performing the paracentesis procedure under staff supervision.

A survey was sent via email to all categorical internal medicine residents across all 3 program years at the time of data collection. Competency for paracentesis sign-off was defined as completing and logging 5 procedures supervised by a competent physician who confirmed that all portions of the procedure were performed correctly. Residents were also asked to answer questions on a scale from 1 to 10, with 1 representing no confidence and 10 representing strong confidence to practice independently (Table).

We also evaluated the number of procedures performed by internal medicine residents 3 years before the clinic was started in 2015 up to the completion of 2022. The numbers were obtained by examining procedural log data for each year for all internal medicine residents.

Results

Thirty-three residents completed the survey: 10 first-year internal medicine residents (PGY1), 12 second-year residents (PGY2), and 11 third-year residents (PGY3). The mean participation was 4.8 paracentesis sessions per person for the duration of the study. The range of paracentesis procedures performed varied based on PGY year: PGY1s performed 1 to > 10 procedures, PGY2s performed 2 to > 10 procedures, and PGY3s performed 5 to > 10 procedures. Thirty-six percent of residents completed > 10 procedures in the paracentesis clinic; 82% of PGY3s had completed > 10 procedures by December of their third year. Twenty-six residents (79%) were credentialed to perform paracentesis procedures independently after performing > 5 procedures, and 7 residents were not yet cleared for procedural independence.

In the survey, residents rated their comfort with performing paracentesis procedures independently at a mean of 7.9. The mean comfort reported by PGY1s was 7.2, PGY2s was 7.3, and PGY3s was 9.3. Residents also rated their opinion on whether or not the paracentesis clinic adequately prepared them for paracentesis procedural independence; the mean was 8.9 across all residents.

The total number of procedures performed by residents at CTVAH also increased. Starting in 2015, 3 years before the clinic was started, 38 procedures were performed by internal medicine residents, followed by 72 procedures in 2016; 76 in 2017; 58 in 2018; 94 in 2019; 88 in 2020; 136 in 2021; and 188 in 2022.

Paracentesis is a simple but invasive procedure to relieve ascites, often relieving patients’ symptoms, preventing hospital admission, and increasing patient satisfaction.⁴ The CTVAH does not have the capacity to perform outpatient paracentesis effectively in its emergency or radiology departments. Furthermore, the use of the emergency or radiology departments for routine paracentesis may not be feasible due to the acuity of care being provided, as these procedures can be time consuming and can draw away critical resources and time from patients that need emergent care. The paracentesis clinic was then formed to provide veterans access to the procedural care they need, while also preparing residents to ably and confidently perform the procedure independently.

Based on our study, most residents were cleared to independently perform paracentesis procedures across all 3 years, with 79% of residents having completed the required 5 supervised procedures to independently practice. A study assessing unsupervised practice standards showed that paracentesis skill declines as soon as 3 months after training. However, retraining was shown to potentially interrupt this skill decline.⁵ Studies have shown that procedure-driven electives or services significantly improved paracentesis certification rates and total logged procedures, with minimal funding or scheduling changes required.⁶ Our clinic showed a significant increase in the number of procedures logged starting with the minimum of 38 procedures in 2015 and ending with 188 procedures logged at the end of 2022.

By allowing residents to routinely return to the paracentesis clinic across all 3 years, residents were more likely to feel comfortable independently performing the procedure, with residents reporting a mean comfort score of 7.9. The spaced repetition and ability to work with the clinic during elective time allows regular opportunities to undergo supervised training in a controlled environment and created scheduled retraining opportunities. Future studies should evaluate residents prior to each paracentesis clinic to ascertain if skill decline is occurring at a slower rate.

The inpatient effect of the clinic is also multifocal. Pham and colleagues showed that integrating paracentesis into timely training can reduce paracentesis delay and delays in care.⁷ By increasing the volume of procedures each resident performs and creating a sense of confidence amongst residents, the clinic increases the number of residents able and willing to perform inpatient procedures, thus reducing the number of unnecessary consultations and hospital resources. One of the reasons the paracentesis clinic was started was to allow patients to have scheduled times to remove fluid from their abdomen, thus cutting down on emergency department procedures and unnecessary admissions. Additionally, the benefits of early paracentesis procedural performance by residents and internal medicine physicians have been demonstrated in the literature. A study by Gaetano and colleagues noted that patients undergoing early paracentesis had reduced mortality of 5.5% vs 7.5% in those undergoing late paracentesis.⁸ This study also showed the in-hospital mortality rate was decreased with paracentesis (6.3%) vs without paracentesis (8.9%).⁸ By offering residents a chance to participate in the clinic, we have shown that regular opportunities to perform paracentesis may increase the number of physicians capable of independently practicing, improve procedural competency, and improve patient access to this procedure.

Limitations

Our study was not free of bias and has potential weaknesses. The survey was sent to all current residents who have participated in the paracentesis clinic, but not every resident filled out the survey (55% of all residents across 3 years completed the survey, 68.7% who had done clinic that year completed the survey). There is a possibility that those not signed off avoided doing the survey, but we are unable to confirm this. The survey also depended on resident recall of the number of paracenteses completed or looking at their procedure log. It is possible that some procedures were not documented, changing the true number. Additionally, rating comfortability with procedures is subjective, which may also create a source of potential weakness. Future projects should include a baseline survey for residents, followed by a repeat survey a year later to show changes from baseline competency.

Conclusions

A dedicated paracentesis clinic with internal medicine resident involvement may increase resident paracentesis procedural independence, the number of procedures available and performed, and procedural comfort level.

Competency in paracentesis is an important procedural skill for medical practitioners caring for patients with decompensated liver cirrhosis. Paracentesis is performed to drain ascitic fluid for both diagnosis and/or therapeutic purposes.¹ While this procedure can be performed without the use of ultrasound, it is preferable to use ultrasound to identify an area of fluid that is away from dangerous anatomy including bowel loops, the liver, and spleen. After prepping the area, lidocaine is administered locally. A catheter is then inserted until fluid begins flowing freely. The catheter is connected to a suction canister or collection kit, and the patient is monitored until the flow ceases. Samples can be sent for analysis to determine the etiology of ascites, identify concerns for infection, and more.

Paracentesis is a very common procedure. Barsuk and colleagues noted that between 2010 and 2012, 97,577 procedures were performed across 120 academic medical centers and 290 affiliated hospitals.² Patients undergo paracentesis in a variety of settings including the emergency department, inpatient hospitalizations, and clinics. Some patients may require only 1 paracentesis procedure while others may require it regularly.

Due to the rising need for paracentesis in the Central Texas Veterans Affairs Hospital (CTVAH) in Temple, a paracentesis clinic was started in February 2018. The goal of the paracentesis clinic was multifocal—to reduce hospital admissions, improve access to regularly scheduled procedures, decrease wait times, and increase patient satisfaction.³ Through the CTVAH affiliation with the Texas A&M internal medicine residency program, the paracentesis clinic started involving and training residents on this procedure. Up to 3 residents are on weekly rotation and can perform up to 6 paracentesis procedures in a week. The purpose of this article was to evaluate resident competency in paracentesis after completion of the paracentesis clinic.

Methods

The paracentesis clinic schedules up to 3 patients on Tuesdays and Thursdays between 8 am and noon. All the necessary equipment is readily available and includes the paracentesis kit, lidocaine, sterile gloves, ultrasound, and albumin if needed. Because this procedure is performed at the hospital, direct access to the emergency department is available. Residents are scheduled weekly. Up to 2 residents are scheduled for the paracentesis clinic during their dedicated clinic week. They are expected to practice obtaining consent, performing the procedure, and documenting the encounter under staff supervision. Additionally, 1 or 2 residents participate in the paracentesis clinic as part of an ultrasound elective twice per year. In this elective, they practice ultrasound skills using a simulation and translate that information in the paracentesis clinic while identifying anatomy and performing the paracentesis procedure under staff supervision.

A survey was sent via email to all categorical internal medicine residents across all 3 program years at the time of data collection. Competency for paracentesis sign-off was defined as completing and logging 5 procedures supervised by a competent physician who confirmed that all portions of the procedure were performed correctly. Residents were also asked to answer questions on a scale from 1 to 10, with 1 representing no confidence and 10 representing strong confidence to practice independently (Table).

We also evaluated the number of procedures performed by internal medicine residents 3 years before the clinic was started in 2015 up to the completion of 2022. The numbers were obtained by examining procedural log data for each year for all internal medicine residents.

Results

Thirty-three residents completed the survey: 10 first-year internal medicine residents (PGY1), 12 second-year residents (PGY2), and 11 third-year residents (PGY3). The mean participation was 4.8 paracentesis sessions per person for the duration of the study. The range of paracentesis procedures performed varied based on PGY year: PGY1s performed 1 to > 10 procedures, PGY2s performed 2 to > 10 procedures, and PGY3s performed 5 to > 10 procedures. Thirty-six percent of residents completed > 10 procedures in the paracentesis clinic; 82% of PGY3s had completed > 10 procedures by December of their third year. Twenty-six residents (79%) were credentialed to perform paracentesis procedures independently after performing > 5 procedures, and 7 residents were not yet cleared for procedural independence.

In the survey, residents rated their comfort with performing paracentesis procedures independently at a mean of 7.9. The mean comfort reported by PGY1s was 7.2, PGY2s was 7.3, and PGY3s was 9.3. Residents also rated their opinion on whether or not the paracentesis clinic adequately prepared them for paracentesis procedural independence; the mean was 8.9 across all residents.

The total number of procedures performed by residents at CTVAH also increased. Starting in 2015, 3 years before the clinic was started, 38 procedures were performed by internal medicine residents, followed by 72 procedures in 2016; 76 in 2017; 58 in 2018; 94 in 2019; 88 in 2020; 136 in 2021; and 188 in 2022.

Paracentesis is a simple but invasive procedure to relieve ascites, often relieving patients’ symptoms, preventing hospital admission, and increasing patient satisfaction.⁴ The CTVAH does not have the capacity to perform outpatient paracentesis effectively in its emergency or radiology departments. Furthermore, the use of the emergency or radiology departments for routine paracentesis may not be feasible due to the acuity of care being provided, as these procedures can be time consuming and can draw away critical resources and time from patients that need emergent care. The paracentesis clinic was then formed to provide veterans access to the procedural care they need, while also preparing residents to ably and confidently perform the procedure independently.

Based on our study, most residents were cleared to independently perform paracentesis procedures across all 3 years, with 79% of residents having completed the required 5 supervised procedures to independently practice. A study assessing unsupervised practice standards showed that paracentesis skill declines as soon as 3 months after training. However, retraining was shown to potentially interrupt this skill decline.⁵ Studies have shown that procedure-driven electives or services significantly improved paracentesis certification rates and total logged procedures, with minimal funding or scheduling changes required.⁶ Our clinic showed a significant increase in the number of procedures logged starting with the minimum of 38 procedures in 2015 and ending with 188 procedures logged at the end of 2022.

By allowing residents to routinely return to the paracentesis clinic across all 3 years, residents were more likely to feel comfortable independently performing the procedure, with residents reporting a mean comfort score of 7.9. The spaced repetition and ability to work with the clinic during elective time allows regular opportunities to undergo supervised training in a controlled environment and created scheduled retraining opportunities. Future studies should evaluate residents prior to each paracentesis clinic to ascertain if skill decline is occurring at a slower rate.

The inpatient effect of the clinic is also multifocal. Pham and colleagues showed that integrating paracentesis into timely training can reduce paracentesis delay and delays in care.⁷ By increasing the volume of procedures each resident performs and creating a sense of confidence amongst residents, the clinic increases the number of residents able and willing to perform inpatient procedures, thus reducing the number of unnecessary consultations and hospital resources. One of the reasons the paracentesis clinic was started was to allow patients to have scheduled times to remove fluid from their abdomen, thus cutting down on emergency department procedures and unnecessary admissions. Additionally, the benefits of early paracentesis procedural performance by residents and internal medicine physicians have been demonstrated in the literature. A study by Gaetano and colleagues noted that patients undergoing early paracentesis had reduced mortality of 5.5% vs 7.5% in those undergoing late paracentesis.⁸ This study also showed the in-hospital mortality rate was decreased with paracentesis (6.3%) vs without paracentesis (8.9%).⁸ By offering residents a chance to participate in the clinic, we have shown that regular opportunities to perform paracentesis may increase the number of physicians capable of independently practicing, improve procedural competency, and improve patient access to this procedure.

Limitations

Our study was not free of bias and has potential weaknesses. The survey was sent to all current residents who have participated in the paracentesis clinic, but not every resident filled out the survey (55% of all residents across 3 years completed the survey, 68.7% who had done clinic that year completed the survey). There is a possibility that those not signed off avoided doing the survey, but we are unable to confirm this. The survey also depended on resident recall of the number of paracenteses completed or looking at their procedure log. It is possible that some procedures were not documented, changing the true number. Additionally, rating comfortability with procedures is subjective, which may also create a source of potential weakness. Future projects should include a baseline survey for residents, followed by a repeat survey a year later to show changes from baseline competency.

Conclusions

A dedicated paracentesis clinic with internal medicine resident involvement may increase resident paracentesis procedural independence, the number of procedures available and performed, and procedural comfort level.

Piperacillin/tazobactam (PTZ) is a combination IV antibiotic comprised of the semisynthetic antipseudomonal β-lactam, piperacillin sodium, and the β-lactamase inhibitor, tazobactam sodium.¹ PTZ is extensively prescribed in the hospital setting for a multitude of infections including but not limited to the US Food and Drug Administration–approved indications: intra-abdominal infection, skin and skin structure infection (SSTI), urinary tract infection (UTI), and pneumonia. Given its broad spectrum of activity and relative safety profile, PTZ is a mainstay of many empiric IV antibiotic regimens. The primary elimination pathway for PTZ is renal excretion, and dosage adjustments are recommended with reduced creatinine clearance. Additionally, PTZ use has been associated with acute renal injury and delayed renal recovery.^1-3

There are various mechanisms through which medications can contribute to acute decomopensated heart failure (ADHF).⁴ These mechanisms include direct cardiotoxicity; negative inotropic, lusitropic, or chronotropic effects; exacerbating hypertension; sodium loading; and drug-drug interactions that limit the benefits of heart failure (HF) medications. One potentially overlooked constituent of PTZ is the sodium content, with the standard formulation containing 65 mg of sodium per gram of piperacillin.^1-3 Furthermore, PTZ must be diluted in 50 to 150 mL of diluent, commonly 0.9% sodium chloride, which can contribute an additional 177 to 531 mg of sodium per dose. PTZ prescribing information advises caution for use in patients with decreased renal, hepatic, and/or cardiac function and notes that geriatric patients, particularly with HF, may be at risk of impaired natriuresis in the setting of large sodium doses.

It is estimated that roughly 6.2 million adults in the United States have HF and prevalence continues to rise.^5,6 Mortality rates after hospitalization due to HF are 20% to 25% at 1 year. Health care expenditures for the management of HF surpass $30 billion per year in the US, with most of this cost attributed to hospitalizations. Consequently, it is important to continue to identify and practice preventative strategies when managing patients with HF.

Methods

This single-center, retrospective, cohort study was conducted at James H. Quillen Veterans Affairs Medical Center (JHQVAMC) in Mountain Home, Tennessee, a 174-bed tertiary medical center. The purpose of this study was to compare the incidence of ADHF in patients who received PTZ vs cefepime (CFP). This project was reviewed by the JHQVAMC Institutional Review Board and deemed exempt as a clinical process improvement operations activity.

The antimicrobial stewardship team at JHQVAMC reviewed the use of PTZ in veterans between January 1, 2018, to December 31, 2019, and compared baseline demographics, history of HF, and outcomes in patients receiving analogous broad-spectrum empiric antibiotic therapy with CFP. Patients were included if they received at least 24 hours of PTZ or CFP. Patients were excluded if they were diagnosed with ADHF before initiation of antibiotic therapy. Patients with ADHF were identified by clinical diagnosis of ADHF documented by the treating clinician and reaffirmed by the study clinician during retrospective chart review. Clinical information used to determine ADHF included clinical presentation, imaging (ie, chest X-ray, echocardiograms), and laboratory parameters, such as B-type natriuretic peptide. The primary endpoint of this study was the incidence of ADHF during the current hospitalization. Secondary endpoints included the length of hospital stay, hospital readmission, and overall mortality. Patient chart reviews were performed using the JHQVAMC Computerized Patient Record System (CPRS).

Statistical Analysis

Analysis was conducted with R Software. Pearson χ² and t tests were used to compare baseline demographics, length of stay, readmission, and mortality. Significance used was α = .05.

A retrospective chart review was performed on 389 veterans. Of the 389, 204 patients received at least 24 hours of PTZ, and 185 patients received CFP. The mean age in both groups was 75 years. Patients in the PTZ group were more likely to have been admitted with the diagnosis of pneumonia (105 vs 49, P < .001). However, 29 patients (15.7%) in the CFP group were admitted with a UTI diagnosis compared with 6 patients (2.9%) in the PTZ group (P < .001) and 62 patients (33.5%) in the CFP group were admitted with a SSTI diagnosis compared with 48 patients (23.5%) in the PTZ group (P = .03). Otherwise, there were no differences between other admitting diagnoses. Additionally, there was no difference in prior history of HF between groups (Table 1).

Twenty-five patients (12.3%) in the PTZ group and 4 patients (2.2%) in the CFP group were subsequently diagnosed with ADHF (P < .001). Hospital readmissions due to HF were higher in the PTZ group compared with the CFP group (11 vs 2, P = .02). Hospital readmission due to other causes was not significantly different between groups. Hospital readmission due to infection occurred in 18 patients who received PTZ and 25 who received CFP (8.8% vs 13.5%, P = .14). Hospital readmission due to any other indication occurred in 24 patients who received PTZ and 24 who received CFP (11.8% vs 13.0%, P = .72). There was no statistically significant difference in all-cause mortality during the associated admission or within 6 months of discharge between groups, with 59 total deaths in the PTZ group and 50 in the CFP group (28.9% vs 27.0%, P = .63).

There was no difference in length of stay outcomes between patients receiving PTZ compared with CFP. Twenty-eight patients in the PTZ group and 20 in the CFP group had a length of stay duration of < 3 days (13.7% vs 10.8%, P = .46). Seventy-three patients in the PTZ group and 76 in the CFP group had a length of stay duration of 4 to 6 days (36.3% vs 41.1%, P = .28). One hundred three patients in the PTZ group and 89 in the CFP group had a length of stay duration ≥ 7 days (50.5% vs 48.1%, P = .78). Table 2 includes a complete overview of primary and secondary endpoint results.

Discussion

The American Heart Association (AHA) lists PTZ as a medication that may cause or exacerbate HF, though no studies have identified a clear association between PTZ use and ADHF.⁴ Sodium restriction is consistently recommended as an important strategy for the prevention of ADHF. Accordingly, PTZ prescribing information and the AHA advise careful consideration with PTZ use in this patient population.^1,4

The specific mechanism responsible for the association of PTZ with cardiac-related adverse outcomes is unclear. It is easy to presume that the sodium content of PTZ is solely responsible; however, other antibiotic regimens not included as agents of concern by the AHA, such as meropenem, can approach similar overall daily sodium amounts.^4,7 Additionally, total sodium and volume can also be contributed by various IV medications and fluids. This study did not evaluate total sodium intake from all sources, but it is notable that this study identified a possible trend toward the risk of ADHF with PTZ use in a routine practice environment. It is reasonable to postulate additional intrinsic properties of PTZ may be contributing to the development of ADHF, such as its association with renal injury possibly resulting in increased fluid retainment and subsequent fluid volume overload.^1,2,4 Other hypothesized mechanisms may include those previously described, such as direct myocardial toxicity; negative inotropic, lusitropic, or chronotropic effects; exacerbating hypertension; and drug-drug interactions that limit the benefits of HF medications, although these have not been overtly associated with PTZ in the literature to date.^4,8

ADHF can present similarly to other acute pulmonary conditions, including pneumonia.^9,10 It is important to acknowledge the challenge this creates for diagnosticians to differentiate between these conditions rapidly and precisely. As a result, this patient population is likely at increased risk of IV antibiotic exposure. Other studies have identified worse outcomes in patients who receive potentially unwarranted IV antibiotics in patients with ADHF.^9,10 The results of this study further emphasize the importance of careful considerate antibiotic selection and overall avoidance of unnecessary antibiotic exposure to limit potential adverse outcomes.

Limitations

There are various limitations to this study. Firstly, no women were included due to the predominantly male population within the Veterans Health Administration system. Secondly, this study was retrospective in design and was therefore limited to the completeness and accuracy of the available data collected. Additionally, this study evaluated any ADHF episode during the associated hospitalization as the primary endpoint. The time to diagnosis of ADHF in relation to PTZ initiation was not evaluated, which may have helped better elucidate this possible association. Furthermore, while a significant statistical difference was identified, the smaller sample size may have limited the ability to accurately identify differences in lower event rate outcomes.

Conclusions

This study identifies an association between PTZ use and significant cardiac-related adverse outcomes, including increased incidence of ADHF and readmission due to HF exacerbation. While more research is needed to define the exact mechanisms by which PTZ may precipitate acute decompensation in patients with HF, it is judicious to consider close monitoring or the avoidance of PTZ when appropriate antibiotic alternatives are available in patients with a known history of HF.

Piperacillin/tazobactam (PTZ) is a combination IV antibiotic comprised of the semisynthetic antipseudomonal β-lactam, piperacillin sodium, and the β-lactamase inhibitor, tazobactam sodium.¹ PTZ is extensively prescribed in the hospital setting for a multitude of infections including but not limited to the US Food and Drug Administration–approved indications: intra-abdominal infection, skin and skin structure infection (SSTI), urinary tract infection (UTI), and pneumonia. Given its broad spectrum of activity and relative safety profile, PTZ is a mainstay of many empiric IV antibiotic regimens. The primary elimination pathway for PTZ is renal excretion, and dosage adjustments are recommended with reduced creatinine clearance. Additionally, PTZ use has been associated with acute renal injury and delayed renal recovery.^1-3

There are various mechanisms through which medications can contribute to acute decomopensated heart failure (ADHF).⁴ These mechanisms include direct cardiotoxicity; negative inotropic, lusitropic, or chronotropic effects; exacerbating hypertension; sodium loading; and drug-drug interactions that limit the benefits of heart failure (HF) medications. One potentially overlooked constituent of PTZ is the sodium content, with the standard formulation containing 65 mg of sodium per gram of piperacillin.^1-3 Furthermore, PTZ must be diluted in 50 to 150 mL of diluent, commonly 0.9% sodium chloride, which can contribute an additional 177 to 531 mg of sodium per dose. PTZ prescribing information advises caution for use in patients with decreased renal, hepatic, and/or cardiac function and notes that geriatric patients, particularly with HF, may be at risk of impaired natriuresis in the setting of large sodium doses.

It is estimated that roughly 6.2 million adults in the United States have HF and prevalence continues to rise.^5,6 Mortality rates after hospitalization due to HF are 20% to 25% at 1 year. Health care expenditures for the management of HF surpass $30 billion per year in the US, with most of this cost attributed to hospitalizations. Consequently, it is important to continue to identify and practice preventative strategies when managing patients with HF.

Methods

This single-center, retrospective, cohort study was conducted at James H. Quillen Veterans Affairs Medical Center (JHQVAMC) in Mountain Home, Tennessee, a 174-bed tertiary medical center. The purpose of this study was to compare the incidence of ADHF in patients who received PTZ vs cefepime (CFP). This project was reviewed by the JHQVAMC Institutional Review Board and deemed exempt as a clinical process improvement operations activity.

The antimicrobial stewardship team at JHQVAMC reviewed the use of PTZ in veterans between January 1, 2018, to December 31, 2019, and compared baseline demographics, history of HF, and outcomes in patients receiving analogous broad-spectrum empiric antibiotic therapy with CFP. Patients were included if they received at least 24 hours of PTZ or CFP. Patients were excluded if they were diagnosed with ADHF before initiation of antibiotic therapy. Patients with ADHF were identified by clinical diagnosis of ADHF documented by the treating clinician and reaffirmed by the study clinician during retrospective chart review. Clinical information used to determine ADHF included clinical presentation, imaging (ie, chest X-ray, echocardiograms), and laboratory parameters, such as B-type natriuretic peptide. The primary endpoint of this study was the incidence of ADHF during the current hospitalization. Secondary endpoints included the length of hospital stay, hospital readmission, and overall mortality. Patient chart reviews were performed using the JHQVAMC Computerized Patient Record System (CPRS).

Statistical Analysis

Analysis was conducted with R Software. Pearson χ² and t tests were used to compare baseline demographics, length of stay, readmission, and mortality. Significance used was α = .05.

A retrospective chart review was performed on 389 veterans. Of the 389, 204 patients received at least 24 hours of PTZ, and 185 patients received CFP. The mean age in both groups was 75 years. Patients in the PTZ group were more likely to have been admitted with the diagnosis of pneumonia (105 vs 49, P < .001). However, 29 patients (15.7%) in the CFP group were admitted with a UTI diagnosis compared with 6 patients (2.9%) in the PTZ group (P < .001) and 62 patients (33.5%) in the CFP group were admitted with a SSTI diagnosis compared with 48 patients (23.5%) in the PTZ group (P = .03). Otherwise, there were no differences between other admitting diagnoses. Additionally, there was no difference in prior history of HF between groups (Table 1).

Twenty-five patients (12.3%) in the PTZ group and 4 patients (2.2%) in the CFP group were subsequently diagnosed with ADHF (P < .001). Hospital readmissions due to HF were higher in the PTZ group compared with the CFP group (11 vs 2, P = .02). Hospital readmission due to other causes was not significantly different between groups. Hospital readmission due to infection occurred in 18 patients who received PTZ and 25 who received CFP (8.8% vs 13.5%, P = .14). Hospital readmission due to any other indication occurred in 24 patients who received PTZ and 24 who received CFP (11.8% vs 13.0%, P = .72). There was no statistically significant difference in all-cause mortality during the associated admission or within 6 months of discharge between groups, with 59 total deaths in the PTZ group and 50 in the CFP group (28.9% vs 27.0%, P = .63).

There was no difference in length of stay outcomes between patients receiving PTZ compared with CFP. Twenty-eight patients in the PTZ group and 20 in the CFP group had a length of stay duration of < 3 days (13.7% vs 10.8%, P = .46). Seventy-three patients in the PTZ group and 76 in the CFP group had a length of stay duration of 4 to 6 days (36.3% vs 41.1%, P = .28). One hundred three patients in the PTZ group and 89 in the CFP group had a length of stay duration ≥ 7 days (50.5% vs 48.1%, P = .78). Table 2 includes a complete overview of primary and secondary endpoint results.

Discussion

The American Heart Association (AHA) lists PTZ as a medication that may cause or exacerbate HF, though no studies have identified a clear association between PTZ use and ADHF.⁴ Sodium restriction is consistently recommended as an important strategy for the prevention of ADHF. Accordingly, PTZ prescribing information and the AHA advise careful consideration with PTZ use in this patient population.^1,4

The specific mechanism responsible for the association of PTZ with cardiac-related adverse outcomes is unclear. It is easy to presume that the sodium content of PTZ is solely responsible; however, other antibiotic regimens not included as agents of concern by the AHA, such as meropenem, can approach similar overall daily sodium amounts.^4,7 Additionally, total sodium and volume can also be contributed by various IV medications and fluids. This study did not evaluate total sodium intake from all sources, but it is notable that this study identified a possible trend toward the risk of ADHF with PTZ use in a routine practice environment. It is reasonable to postulate additional intrinsic properties of PTZ may be contributing to the development of ADHF, such as its association with renal injury possibly resulting in increased fluid retainment and subsequent fluid volume overload.^1,2,4 Other hypothesized mechanisms may include those previously described, such as direct myocardial toxicity; negative inotropic, lusitropic, or chronotropic effects; exacerbating hypertension; and drug-drug interactions that limit the benefits of HF medications, although these have not been overtly associated with PTZ in the literature to date.^4,8

ADHF can present similarly to other acute pulmonary conditions, including pneumonia.^9,10 It is important to acknowledge the challenge this creates for diagnosticians to differentiate between these conditions rapidly and precisely. As a result, this patient population is likely at increased risk of IV antibiotic exposure. Other studies have identified worse outcomes in patients who receive potentially unwarranted IV antibiotics in patients with ADHF.^9,10 The results of this study further emphasize the importance of careful considerate antibiotic selection and overall avoidance of unnecessary antibiotic exposure to limit potential adverse outcomes.

Limitations

There are various limitations to this study. Firstly, no women were included due to the predominantly male population within the Veterans Health Administration system. Secondly, this study was retrospective in design and was therefore limited to the completeness and accuracy of the available data collected. Additionally, this study evaluated any ADHF episode during the associated hospitalization as the primary endpoint. The time to diagnosis of ADHF in relation to PTZ initiation was not evaluated, which may have helped better elucidate this possible association. Furthermore, while a significant statistical difference was identified, the smaller sample size may have limited the ability to accurately identify differences in lower event rate outcomes.

Conclusions

This study identifies an association between PTZ use and significant cardiac-related adverse outcomes, including increased incidence of ADHF and readmission due to HF exacerbation. While more research is needed to define the exact mechanisms by which PTZ may precipitate acute decompensation in patients with HF, it is judicious to consider close monitoring or the avoidance of PTZ when appropriate antibiotic alternatives are available in patients with a known history of HF.

Piperacillin/tazobactam (PTZ) is a combination IV antibiotic comprised of the semisynthetic antipseudomonal β-lactam, piperacillin sodium, and the β-lactamase inhibitor, tazobactam sodium.¹ PTZ is extensively prescribed in the hospital setting for a multitude of infections including but not limited to the US Food and Drug Administration–approved indications: intra-abdominal infection, skin and skin structure infection (SSTI), urinary tract infection (UTI), and pneumonia. Given its broad spectrum of activity and relative safety profile, PTZ is a mainstay of many empiric IV antibiotic regimens. The primary elimination pathway for PTZ is renal excretion, and dosage adjustments are recommended with reduced creatinine clearance. Additionally, PTZ use has been associated with acute renal injury and delayed renal recovery.^1-3

There are various mechanisms through which medications can contribute to acute decomopensated heart failure (ADHF).⁴ These mechanisms include direct cardiotoxicity; negative inotropic, lusitropic, or chronotropic effects; exacerbating hypertension; sodium loading; and drug-drug interactions that limit the benefits of heart failure (HF) medications. One potentially overlooked constituent of PTZ is the sodium content, with the standard formulation containing 65 mg of sodium per gram of piperacillin.^1-3 Furthermore, PTZ must be diluted in 50 to 150 mL of diluent, commonly 0.9% sodium chloride, which can contribute an additional 177 to 531 mg of sodium per dose. PTZ prescribing information advises caution for use in patients with decreased renal, hepatic, and/or cardiac function and notes that geriatric patients, particularly with HF, may be at risk of impaired natriuresis in the setting of large sodium doses.

It is estimated that roughly 6.2 million adults in the United States have HF and prevalence continues to rise.^5,6 Mortality rates after hospitalization due to HF are 20% to 25% at 1 year. Health care expenditures for the management of HF surpass $30 billion per year in the US, with most of this cost attributed to hospitalizations. Consequently, it is important to continue to identify and practice preventative strategies when managing patients with HF.

Methods

This single-center, retrospective, cohort study was conducted at James H. Quillen Veterans Affairs Medical Center (JHQVAMC) in Mountain Home, Tennessee, a 174-bed tertiary medical center. The purpose of this study was to compare the incidence of ADHF in patients who received PTZ vs cefepime (CFP). This project was reviewed by the JHQVAMC Institutional Review Board and deemed exempt as a clinical process improvement operations activity.

The antimicrobial stewardship team at JHQVAMC reviewed the use of PTZ in veterans between January 1, 2018, to December 31, 2019, and compared baseline demographics, history of HF, and outcomes in patients receiving analogous broad-spectrum empiric antibiotic therapy with CFP. Patients were included if they received at least 24 hours of PTZ or CFP. Patients were excluded if they were diagnosed with ADHF before initiation of antibiotic therapy. Patients with ADHF were identified by clinical diagnosis of ADHF documented by the treating clinician and reaffirmed by the study clinician during retrospective chart review. Clinical information used to determine ADHF included clinical presentation, imaging (ie, chest X-ray, echocardiograms), and laboratory parameters, such as B-type natriuretic peptide. The primary endpoint of this study was the incidence of ADHF during the current hospitalization. Secondary endpoints included the length of hospital stay, hospital readmission, and overall mortality. Patient chart reviews were performed using the JHQVAMC Computerized Patient Record System (CPRS).

Statistical Analysis

Analysis was conducted with R Software. Pearson χ² and t tests were used to compare baseline demographics, length of stay, readmission, and mortality. Significance used was α = .05.

A retrospective chart review was performed on 389 veterans. Of the 389, 204 patients received at least 24 hours of PTZ, and 185 patients received CFP. The mean age in both groups was 75 years. Patients in the PTZ group were more likely to have been admitted with the diagnosis of pneumonia (105 vs 49, P < .001). However, 29 patients (15.7%) in the CFP group were admitted with a UTI diagnosis compared with 6 patients (2.9%) in the PTZ group (P < .001) and 62 patients (33.5%) in the CFP group were admitted with a SSTI diagnosis compared with 48 patients (23.5%) in the PTZ group (P = .03). Otherwise, there were no differences between other admitting diagnoses. Additionally, there was no difference in prior history of HF between groups (Table 1).

Twenty-five patients (12.3%) in the PTZ group and 4 patients (2.2%) in the CFP group were subsequently diagnosed with ADHF (P < .001). Hospital readmissions due to HF were higher in the PTZ group compared with the CFP group (11 vs 2, P = .02). Hospital readmission due to other causes was not significantly different between groups. Hospital readmission due to infection occurred in 18 patients who received PTZ and 25 who received CFP (8.8% vs 13.5%, P = .14). Hospital readmission due to any other indication occurred in 24 patients who received PTZ and 24 who received CFP (11.8% vs 13.0%, P = .72). There was no statistically significant difference in all-cause mortality during the associated admission or within 6 months of discharge between groups, with 59 total deaths in the PTZ group and 50 in the CFP group (28.9% vs 27.0%, P = .63).

There was no difference in length of stay outcomes between patients receiving PTZ compared with CFP. Twenty-eight patients in the PTZ group and 20 in the CFP group had a length of stay duration of < 3 days (13.7% vs 10.8%, P = .46). Seventy-three patients in the PTZ group and 76 in the CFP group had a length of stay duration of 4 to 6 days (36.3% vs 41.1%, P = .28). One hundred three patients in the PTZ group and 89 in the CFP group had a length of stay duration ≥ 7 days (50.5% vs 48.1%, P = .78). Table 2 includes a complete overview of primary and secondary endpoint results.

Discussion

The American Heart Association (AHA) lists PTZ as a medication that may cause or exacerbate HF, though no studies have identified a clear association between PTZ use and ADHF.⁴ Sodium restriction is consistently recommended as an important strategy for the prevention of ADHF. Accordingly, PTZ prescribing information and the AHA advise careful consideration with PTZ use in this patient population.^1,4

The specific mechanism responsible for the association of PTZ with cardiac-related adverse outcomes is unclear. It is easy to presume that the sodium content of PTZ is solely responsible; however, other antibiotic regimens not included as agents of concern by the AHA, such as meropenem, can approach similar overall daily sodium amounts.^4,7 Additionally, total sodium and volume can also be contributed by various IV medications and fluids. This study did not evaluate total sodium intake from all sources, but it is notable that this study identified a possible trend toward the risk of ADHF with PTZ use in a routine practice environment. It is reasonable to postulate additional intrinsic properties of PTZ may be contributing to the development of ADHF, such as its association with renal injury possibly resulting in increased fluid retainment and subsequent fluid volume overload.^1,2,4 Other hypothesized mechanisms may include those previously described, such as direct myocardial toxicity; negative inotropic, lusitropic, or chronotropic effects; exacerbating hypertension; and drug-drug interactions that limit the benefits of HF medications, although these have not been overtly associated with PTZ in the literature to date.^4,8

ADHF can present similarly to other acute pulmonary conditions, including pneumonia.^9,10 It is important to acknowledge the challenge this creates for diagnosticians to differentiate between these conditions rapidly and precisely. As a result, this patient population is likely at increased risk of IV antibiotic exposure. Other studies have identified worse outcomes in patients who receive potentially unwarranted IV antibiotics in patients with ADHF.^9,10 The results of this study further emphasize the importance of careful considerate antibiotic selection and overall avoidance of unnecessary antibiotic exposure to limit potential adverse outcomes.

Limitations

There are various limitations to this study. Firstly, no women were included due to the predominantly male population within the Veterans Health Administration system. Secondly, this study was retrospective in design and was therefore limited to the completeness and accuracy of the available data collected. Additionally, this study evaluated any ADHF episode during the associated hospitalization as the primary endpoint. The time to diagnosis of ADHF in relation to PTZ initiation was not evaluated, which may have helped better elucidate this possible association. Furthermore, while a significant statistical difference was identified, the smaller sample size may have limited the ability to accurately identify differences in lower event rate outcomes.

Conclusions

This study identifies an association between PTZ use and significant cardiac-related adverse outcomes, including increased incidence of ADHF and readmission due to HF exacerbation. While more research is needed to define the exact mechanisms by which PTZ may precipitate acute decompensation in patients with HF, it is judicious to consider close monitoring or the avoidance of PTZ when appropriate antibiotic alternatives are available in patients with a known history of HF.

Artificial intelligence (AI) has lagged in health care but has considerable potential to improve quality, safety, clinician experience, and access to care. It is being tested in areas like billing, hospital operations, and preventing adverse events (eg, sepsis mortality) with some early success. However, there are still many barriers preventing the widespread use of AI, such as data problems, mismatched rewards, and workplace obstacles. Innovative projects, partnerships, better rewards, and more investment could remove barriers. Implemented reliably and safely, AI can add to what clinicians know, help them work faster, cut costs, and, most importantly, improve patient care.¹

AI can potentially bring several clinical benefits, such as reducing the administrative strain on clinicians and granting them more time for direct patient care. It can also improve diagnostic accuracy by analyzing patient data and diagnostic images, providing differential diagnoses, and increasing access to care by providing medical information and essential online services to patients.²

High Reliability Organizations

High reliability health care organizations have considerable experience safely launching new programs. For example, the Patient Safety Adoption Framework gives practical tips for smoothly rolling out safety initiatives (Table 1). Developed with experts and diverse views, this framework has 5 key areas: leadership, culture and context, process, measurement, and person-centeredness. These address adoption problems, guide leaders step-by-step, and focus on leadership buy-in, safety culture, cooperation, and local customization. Checklists and tools make it systematic to go from ideas to action on patient safety.³

Leadership involves establishing organizational commitment behind new safety programs. This visible commitment signals importance and priorities to others. Leaders model desired behaviors and language around safety, allocate resources, remove obstacles, and keep initiatives energized over time through consistent messaging.⁴ Culture and context recognizes that safety culture differs across units and facilities. Local input tailors programs to fit and examines strengths to build on, like psychological safety. Surveys gauge the existing culture and its need for change. Process details how to plan, design, test, implement, and improve new safety practices and provides a phased roadmap from idea to results. Measurement collects data to drive improvement and show impact. Metrics track progress and allow benchmarking. Person-centeredness puts patients first in safety efforts through participation, education, and transparency.

The Veterans Health Administration piloted a comprehensive high reliability hospital (HRH) model. Over 3 years, the Veterans Health Administration focused on leadership, culture, and process improvement at a hospital. After initiating the model, the pilot hospital improved its safety culture, reported more minor safety issues, and reduced deaths and complications better than other hospitals. The high-reliability approach successfully instilled principles and improved culture and outcomes. The HRH model is set to be expanded to 18 more US Department of Veterans Affairs (VA) sites for further evaluation across diverse settings.⁵

Trustworthy AI Framework

AI systems are growing more powerful and widespread, including in health care. Unfortunately, irresponsible AI can introduce new harm. ChatGPT and other large language models, for example, sometimes are known to provide erroneous information in a compelling way. Clinicians and patients who use such programs can act on such information, which would lead to unforeseen negative consequences. Several frameworks on ethical AI have come from governmental groups.^6-9 In 2023, the VA National AI Institute suggested a Trustworthy AI Framework based on core principles tailored for federal health care. The framework has 6 key principles: purposeful, effective and safe, secure and private, fair and equitable, transparent and explainable, and accountable and monitored (Table 2).¹⁰

First, AI must clearly help veterans while minimizing risks. To ensure purpose, the VA will assess patient and clinician needs and design AI that targets meaningful problems to avoid scope creep or feature bloat. For example, adding new features to the AI software after release can clutter and complicate the interface, making it difficult to use. Rigorous testing will confirm that AI meets intent prior to deployment. Second, AI is designed and checked for effectiveness, safety, and reliability. The VA pledges to monitor AI’s impact to ensure it performs as expected without unintended consequences. Algorithms will be stress tested across representative datasets and approval processes will screen for safety issues. Third, AI models are secured from vulnerabilities and misuse. Technical controls will prevent unauthorized access or changes to AI systems. Audits will check for appropriate internal usage per policies. Continual patches and upgrades will maintain security. Fourth, the VA manages AI for fairness, avoiding bias. They will proactively assess datasets and algorithms for potential biases based on protected attributes like race, gender, or age. Biased outputs will be addressed through techniques such as data augmentation, reweighting, and algorithm tweaks. Fifth, transparency explains AI’s role in care. Documentation will detail an AI system’s data sources, methodology, testing, limitations, and integration with clinical workflows. Clinicians and patients will receive education on interpreting AI outputs. Finally, the VA pledges to closely monitor AI systems to sustain trust. The VA will establish oversight processes to quickly identify any declines in reliability or unfair impacts on subgroups. AI models will be retrained as needed based on incoming data patterns.

Each Trustworthy AI Framework principle connects to others in existing frameworks. The purpose principle aligns with human-centric AI focused on benefits. Effectiveness and safety link to technical robustness and risk management principles. Security maps to privacy protection principles. Fairness connects to principles of avoiding bias and discrimination. Transparency corresponds with accountable and explainable AI. Monitoring and accountability tie back to governance principles. Overall, the VA framework aims to guide ethical AI based on context. It offers a model for managing risks and building trust in health care AI.

Combining VA principles with high-reliability safety principles can ensure that AI benefits veterans. The leadership and culture aspects will drive commitment to trustworthy AI practices. Leaders will communicate the importance of responsible AI through words and actions. Culture surveys can assess baseline awareness of AI ethics issues to target education. AI security and fairness will be emphasized as safety critical. The process aspect will institute policies and procedures to uphold AI principles through the project lifecycle. For example, structured testing processes will validate safety. Measurement will collect data on principles like transparency and fairness. Dashboards can track metrics like explainability and biases. A patient-centered approach will incorporate veteran perspectives on AI through participatory design and advisory councils. They can give input on AI explainability and potential biases based on their diverse backgrounds.

Conclusions

Joint principles will lead to successful AI that improves care while proactively managing risks. Involve leaders to stress the necessity of eliminating biases. Build security into the AI development process. Co-design AI transparency features with end users. Closely monitor the impact of AI across safety, fairness, and other principles. Adhering to both Trustworthy AI and high reliability organizations principles will earn veterans’ confidence. Health care organizations like the VA can integrate ethical AI safely via established frameworks. With responsible design and implementation, AI’s potential to enhance care quality, safety, and access can be realized.

Acknowledgments

We would like to acknowledge Joshua Mueller, Theo Tiffney, John Zachary, and Gil Alterovitz for their excellent work creating the VA Trustworthy Principles. This material is the result of work supported by resources and the use of facilities at the James A. Haley Veterans’ Hospital.

Artificial intelligence (AI) has lagged in health care but has considerable potential to improve quality, safety, clinician experience, and access to care. It is being tested in areas like billing, hospital operations, and preventing adverse events (eg, sepsis mortality) with some early success. However, there are still many barriers preventing the widespread use of AI, such as data problems, mismatched rewards, and workplace obstacles. Innovative projects, partnerships, better rewards, and more investment could remove barriers. Implemented reliably and safely, AI can add to what clinicians know, help them work faster, cut costs, and, most importantly, improve patient care.¹

AI can potentially bring several clinical benefits, such as reducing the administrative strain on clinicians and granting them more time for direct patient care. It can also improve diagnostic accuracy by analyzing patient data and diagnostic images, providing differential diagnoses, and increasing access to care by providing medical information and essential online services to patients.²

High Reliability Organizations

High reliability health care organizations have considerable experience safely launching new programs. For example, the Patient Safety Adoption Framework gives practical tips for smoothly rolling out safety initiatives (Table 1). Developed with experts and diverse views, this framework has 5 key areas: leadership, culture and context, process, measurement, and person-centeredness. These address adoption problems, guide leaders step-by-step, and focus on leadership buy-in, safety culture, cooperation, and local customization. Checklists and tools make it systematic to go from ideas to action on patient safety.³

Leadership involves establishing organizational commitment behind new safety programs. This visible commitment signals importance and priorities to others. Leaders model desired behaviors and language around safety, allocate resources, remove obstacles, and keep initiatives energized over time through consistent messaging.⁴ Culture and context recognizes that safety culture differs across units and facilities. Local input tailors programs to fit and examines strengths to build on, like psychological safety. Surveys gauge the existing culture and its need for change. Process details how to plan, design, test, implement, and improve new safety practices and provides a phased roadmap from idea to results. Measurement collects data to drive improvement and show impact. Metrics track progress and allow benchmarking. Person-centeredness puts patients first in safety efforts through participation, education, and transparency.

The Veterans Health Administration piloted a comprehensive high reliability hospital (HRH) model. Over 3 years, the Veterans Health Administration focused on leadership, culture, and process improvement at a hospital. After initiating the model, the pilot hospital improved its safety culture, reported more minor safety issues, and reduced deaths and complications better than other hospitals. The high-reliability approach successfully instilled principles and improved culture and outcomes. The HRH model is set to be expanded to 18 more US Department of Veterans Affairs (VA) sites for further evaluation across diverse settings.⁵

Trustworthy AI Framework

AI systems are growing more powerful and widespread, including in health care. Unfortunately, irresponsible AI can introduce new harm. ChatGPT and other large language models, for example, sometimes are known to provide erroneous information in a compelling way. Clinicians and patients who use such programs can act on such information, which would lead to unforeseen negative consequences. Several frameworks on ethical AI have come from governmental groups.^6-9 In 2023, the VA National AI Institute suggested a Trustworthy AI Framework based on core principles tailored for federal health care. The framework has 6 key principles: purposeful, effective and safe, secure and private, fair and equitable, transparent and explainable, and accountable and monitored (Table 2).¹⁰

First, AI must clearly help veterans while minimizing risks. To ensure purpose, the VA will assess patient and clinician needs and design AI that targets meaningful problems to avoid scope creep or feature bloat. For example, adding new features to the AI software after release can clutter and complicate the interface, making it difficult to use. Rigorous testing will confirm that AI meets intent prior to deployment. Second, AI is designed and checked for effectiveness, safety, and reliability. The VA pledges to monitor AI’s impact to ensure it performs as expected without unintended consequences. Algorithms will be stress tested across representative datasets and approval processes will screen for safety issues. Third, AI models are secured from vulnerabilities and misuse. Technical controls will prevent unauthorized access or changes to AI systems. Audits will check for appropriate internal usage per policies. Continual patches and upgrades will maintain security. Fourth, the VA manages AI for fairness, avoiding bias. They will proactively assess datasets and algorithms for potential biases based on protected attributes like race, gender, or age. Biased outputs will be addressed through techniques such as data augmentation, reweighting, and algorithm tweaks. Fifth, transparency explains AI’s role in care. Documentation will detail an AI system’s data sources, methodology, testing, limitations, and integration with clinical workflows. Clinicians and patients will receive education on interpreting AI outputs. Finally, the VA pledges to closely monitor AI systems to sustain trust. The VA will establish oversight processes to quickly identify any declines in reliability or unfair impacts on subgroups. AI models will be retrained as needed based on incoming data patterns.

Each Trustworthy AI Framework principle connects to others in existing frameworks. The purpose principle aligns with human-centric AI focused on benefits. Effectiveness and safety link to technical robustness and risk management principles. Security maps to privacy protection principles. Fairness connects to principles of avoiding bias and discrimination. Transparency corresponds with accountable and explainable AI. Monitoring and accountability tie back to governance principles. Overall, the VA framework aims to guide ethical AI based on context. It offers a model for managing risks and building trust in health care AI.

Combining VA principles with high-reliability safety principles can ensure that AI benefits veterans. The leadership and culture aspects will drive commitment to trustworthy AI practices. Leaders will communicate the importance of responsible AI through words and actions. Culture surveys can assess baseline awareness of AI ethics issues to target education. AI security and fairness will be emphasized as safety critical. The process aspect will institute policies and procedures to uphold AI principles through the project lifecycle. For example, structured testing processes will validate safety. Measurement will collect data on principles like transparency and fairness. Dashboards can track metrics like explainability and biases. A patient-centered approach will incorporate veteran perspectives on AI through participatory design and advisory councils. They can give input on AI explainability and potential biases based on their diverse backgrounds.

Conclusions

Joint principles will lead to successful AI that improves care while proactively managing risks. Involve leaders to stress the necessity of eliminating biases. Build security into the AI development process. Co-design AI transparency features with end users. Closely monitor the impact of AI across safety, fairness, and other principles. Adhering to both Trustworthy AI and high reliability organizations principles will earn veterans’ confidence. Health care organizations like the VA can integrate ethical AI safely via established frameworks. With responsible design and implementation, AI’s potential to enhance care quality, safety, and access can be realized.

Acknowledgments

We would like to acknowledge Joshua Mueller, Theo Tiffney, John Zachary, and Gil Alterovitz for their excellent work creating the VA Trustworthy Principles. This material is the result of work supported by resources and the use of facilities at the James A. Haley Veterans’ Hospital.

Artificial intelligence (AI) has lagged in health care but has considerable potential to improve quality, safety, clinician experience, and access to care. It is being tested in areas like billing, hospital operations, and preventing adverse events (eg, sepsis mortality) with some early success. However, there are still many barriers preventing the widespread use of AI, such as data problems, mismatched rewards, and workplace obstacles. Innovative projects, partnerships, better rewards, and more investment could remove barriers. Implemented reliably and safely, AI can add to what clinicians know, help them work faster, cut costs, and, most importantly, improve patient care.¹

AI can potentially bring several clinical benefits, such as reducing the administrative strain on clinicians and granting them more time for direct patient care. It can also improve diagnostic accuracy by analyzing patient data and diagnostic images, providing differential diagnoses, and increasing access to care by providing medical information and essential online services to patients.²

High Reliability Organizations

High reliability health care organizations have considerable experience safely launching new programs. For example, the Patient Safety Adoption Framework gives practical tips for smoothly rolling out safety initiatives (Table 1). Developed with experts and diverse views, this framework has 5 key areas: leadership, culture and context, process, measurement, and person-centeredness. These address adoption problems, guide leaders step-by-step, and focus on leadership buy-in, safety culture, cooperation, and local customization. Checklists and tools make it systematic to go from ideas to action on patient safety.³

Leadership involves establishing organizational commitment behind new safety programs. This visible commitment signals importance and priorities to others. Leaders model desired behaviors and language around safety, allocate resources, remove obstacles, and keep initiatives energized over time through consistent messaging.⁴ Culture and context recognizes that safety culture differs across units and facilities. Local input tailors programs to fit and examines strengths to build on, like psychological safety. Surveys gauge the existing culture and its need for change. Process details how to plan, design, test, implement, and improve new safety practices and provides a phased roadmap from idea to results. Measurement collects data to drive improvement and show impact. Metrics track progress and allow benchmarking. Person-centeredness puts patients first in safety efforts through participation, education, and transparency.

The Veterans Health Administration piloted a comprehensive high reliability hospital (HRH) model. Over 3 years, the Veterans Health Administration focused on leadership, culture, and process improvement at a hospital. After initiating the model, the pilot hospital improved its safety culture, reported more minor safety issues, and reduced deaths and complications better than other hospitals. The high-reliability approach successfully instilled principles and improved culture and outcomes. The HRH model is set to be expanded to 18 more US Department of Veterans Affairs (VA) sites for further evaluation across diverse settings.⁵

Trustworthy AI Framework

AI systems are growing more powerful and widespread, including in health care. Unfortunately, irresponsible AI can introduce new harm. ChatGPT and other large language models, for example, sometimes are known to provide erroneous information in a compelling way. Clinicians and patients who use such programs can act on such information, which would lead to unforeseen negative consequences. Several frameworks on ethical AI have come from governmental groups.^6-9 In 2023, the VA National AI Institute suggested a Trustworthy AI Framework based on core principles tailored for federal health care. The framework has 6 key principles: purposeful, effective and safe, secure and private, fair and equitable, transparent and explainable, and accountable and monitored (Table 2).¹⁰

First, AI must clearly help veterans while minimizing risks. To ensure purpose, the VA will assess patient and clinician needs and design AI that targets meaningful problems to avoid scope creep or feature bloat. For example, adding new features to the AI software after release can clutter and complicate the interface, making it difficult to use. Rigorous testing will confirm that AI meets intent prior to deployment. Second, AI is designed and checked for effectiveness, safety, and reliability. The VA pledges to monitor AI’s impact to ensure it performs as expected without unintended consequences. Algorithms will be stress tested across representative datasets and approval processes will screen for safety issues. Third, AI models are secured from vulnerabilities and misuse. Technical controls will prevent unauthorized access or changes to AI systems. Audits will check for appropriate internal usage per policies. Continual patches and upgrades will maintain security. Fourth, the VA manages AI for fairness, avoiding bias. They will proactively assess datasets and algorithms for potential biases based on protected attributes like race, gender, or age. Biased outputs will be addressed through techniques such as data augmentation, reweighting, and algorithm tweaks. Fifth, transparency explains AI’s role in care. Documentation will detail an AI system’s data sources, methodology, testing, limitations, and integration with clinical workflows. Clinicians and patients will receive education on interpreting AI outputs. Finally, the VA pledges to closely monitor AI systems to sustain trust. The VA will establish oversight processes to quickly identify any declines in reliability or unfair impacts on subgroups. AI models will be retrained as needed based on incoming data patterns.

Each Trustworthy AI Framework principle connects to others in existing frameworks. The purpose principle aligns with human-centric AI focused on benefits. Effectiveness and safety link to technical robustness and risk management principles. Security maps to privacy protection principles. Fairness connects to principles of avoiding bias and discrimination. Transparency corresponds with accountable and explainable AI. Monitoring and accountability tie back to governance principles. Overall, the VA framework aims to guide ethical AI based on context. It offers a model for managing risks and building trust in health care AI.

Combining VA principles with high-reliability safety principles can ensure that AI benefits veterans. The leadership and culture aspects will drive commitment to trustworthy AI practices. Leaders will communicate the importance of responsible AI through words and actions. Culture surveys can assess baseline awareness of AI ethics issues to target education. AI security and fairness will be emphasized as safety critical. The process aspect will institute policies and procedures to uphold AI principles through the project lifecycle. For example, structured testing processes will validate safety. Measurement will collect data on principles like transparency and fairness. Dashboards can track metrics like explainability and biases. A patient-centered approach will incorporate veteran perspectives on AI through participatory design and advisory councils. They can give input on AI explainability and potential biases based on their diverse backgrounds.

Conclusions

Joint principles will lead to successful AI that improves care while proactively managing risks. Involve leaders to stress the necessity of eliminating biases. Build security into the AI development process. Co-design AI transparency features with end users. Closely monitor the impact of AI across safety, fairness, and other principles. Adhering to both Trustworthy AI and high reliability organizations principles will earn veterans’ confidence. Health care organizations like the VA can integrate ethical AI safely via established frameworks. With responsible design and implementation, AI’s potential to enhance care quality, safety, and access can be realized.

Acknowledgments

We would like to acknowledge Joshua Mueller, Theo Tiffney, John Zachary, and Gil Alterovitz for their excellent work creating the VA Trustworthy Principles. This material is the result of work supported by resources and the use of facilities at the James A. Haley Veterans’ Hospital.

To the Editor:

A 14-year-old Black adolescent girl presented with episodic, painful, edematous plaques that occurred symmetrically on the arms and legs of 5 years’ duration. The plaques evolved into hyperpigmented patches within 24 to 48 hours before eventually resolving. Fatigue, headache, arthralgias of the arms and legs, chest pain, abdominal pain, nausea, and vomiting variably accompanied these episodes.

Prior to visiting our clinic, the patient had been seen by numerous specialists. A review of her medical records revealed an initial diagnosis of Henoch-Schönlein purpura (HSP), then urticarial vasculitis. She had been treated with antihistamines, topical and systemic steroids, hydroxychloroquine, mycophenolate mofetil, dapsone, azathioprine, and gabapentin. All treatments were ineffectual. She underwent extensive diagnostic testing and imaging, which were normal or noncontributory, including type I allergy testing; multiple exhaustive batteries of hematologic testing; and computed tomography/magnetic resonance imaging/magnetic resonance angiography of the brain, chest, abdomen, and pelvic region. Biopsies from symptomatic segments of the gastrointestinal tract were normal.

Chronic treatment with systemic steroids over 9 months resulted in gastritis and an episode of hematemesis requiring emergent hospitalization. A lengthy multidisciplinary evaluation was conducted at the patient’s local community hospital; the team concluded that she had an urticarial-type rash with accompanying symptoms that did not have an autoimmune, rheumatologic, or inflammatory basis.

The patient’s medical history was remarkable for recent-onset panic attacks. Her family medical history was noncontributory. Physical examination revealed multiple violaceous hyperpigmented patches diffusely located on the proximal upper arms (Figure 1). There were no additional findings on physical examination.

Punch biopsies were performed on lesional areas of the arm. Histopathology indicated a mild superficial perivascular dermal mixed infiltrate and extravasated erythrocytes (Figure 2). Direct immunofluorescence (DIF) testing was negative for vasculitis. Immunohistochemical stains for CD117 and tryptase demonstrated a slight increase in the number of dermal mast cells; however, the increase was not sufficient to diagnose cutaneous mastocytosis, which was in the differential. We proposed a diagnosis of psychogenic purpura (PP)(also known as Gardner-Diamond syndrome). She was treated with gabapentin, a selective serotonin reuptake inhibitor, and cognitive therapy. Unfortunately, after starting therapy the patient was lost to follow-up.

Psychogenic purpura is a rare vasculopathy of unknown etiology that may be a special form of factitious disorder.^1,2 In one study, PP occurred predominantly in females aged 15 to 66 years, with a median onset age of 33 years.³ A prodrome of localized itching, burning, and/or pain precedes the development of edematous plaques. The plaques evolve into painful ecchymoses within 1 to 2 days and resolve in 10 days or fewer without treatment. Lesions most commonly occur on the extremities but may occur anywhere on the body. The most common associated finding is an underlying depressive disorder. Episodes may be accompanied by headache, dizziness, fatigue, fever, arthralgia, nausea, vomiting, abdominal pain, menstrual irregularities, myalgia, and urologic conditions.

In 1955, Gardner and Diamond⁴ described the first cases of PP in 4 female patients at Peter Bent Brigham Hospital in Boston, Massachusetts. The investigators were able to replicate the painful ecchymoses with intradermal injection of the patient’s own erythrocytes into the skin. They proposed that the underlying pathogenesis involved autosensitization to erythrocyte stroma.⁴ Since then, others have suggested that the pathogenesis may include autosensitization to erythrocyte phosphatidylserine, tonus dysregulation of venous capillaries, abnormal endothelial fibrin synthesis, and capillary wall instability.^5-7 

Histopathology typically reveals superficial and deep perivascular inflammation with extravasated erythrocytes. Direct immunofluorescence is negative for vasculitis.⁸ Diagnostics and laboratory findings for underlying systemic illness are negative or noncontributory. Cutaneous injection of 1 mL of the patient’s own washed erythrocytes may result in the formation of the characteristic painful plaques within 24 hours; however, this test is limited by lack of standardization and low sensitivity.³

Psychogenic purpura may share clinical features with cutaneous small vessel vasculitis, such as HSP or urticarial vasculitis. Some of the findings that our patient was experiencing, including purpura, arthralgia, and abdominal pain, are associated with HSP. However, HSP typically is self-limiting and classically features palpable purpura distributed across the lower extremities and buttocks. Histopathology demonstrates the classic findings of leukocytoclastic vasculitis; DIF typically is positive for perivascular IgA and C3 deposition. Increased serum IgA may be present.⁹ Urticarial vasculitis appears as erythematous indurated wheals that favor a proximal extremity and truncal distribution. They characteristically last longer than 24 hours, are frequently associated with nonprodromal pain or burning, and resolve with hyperpigmentation. Arthralgia and gastrointestinal, renal, pulmonary, cardiac, and neurologic symptoms may be present, especially in patients with low complement levels.¹⁰ Skin biopsy demonstrates leukocytoclasia that must be accompanied by vessel wall necrosis. Fibrinoid deposition, erythrocyte extravasation, or perivascular inflammation may be present. In 70% of cases revealing perivascular immunoglobulin, C3, and fibrinogen deposition, DIF is positive. Serum C1q autoantibody may be associated with the hypocomplementemic form.¹⁰

The classic histopathologic findings in leukocytoclastic vasculitis include transmural neutrophilic infiltration of the walls of small vessels, fibrinoid necrosis of vessel walls, leukocytoclasia, extravasated erythrocytes, and signs of endothelial cell damage.⁹ A prior punch biopsy in this patient demonstrated rare neutrophilic nuclear debris within the vessel walls without fibrin deposition. Although the presence of nuclear debris and extravasated erythrocytes could be compatible with a manifestation of urticarial vasculitis, the lack of direct evidence of vessel wall necrosis combined with subsequent biopsies unequivocally ruled out cutaneous small vessel vasculitis in our patient.

Psychogenic purpura has been reported to occur frequently in the background of psycho-emotional distress. In 1989, Ratnoff¹¹ noted that many of the patients he was treating at the University Hospitals of Cleveland, Ohio, had a depressive syndrome. A review of patients treated at the Mayo Clinic in Rochester, Minnesota, illustrated concomitant psychiatric illnesses in 41 of 76 (54%) patients treated for PP, most commonly depressive, personality, and anxiety disorders.³

There is no consensus on therapy for PP. Treatment is based on providing symptomatic relief and relieving underlying psychiatric distress. Block et al¹² found the use of selective serotonin reuptake inhibitors, tricyclic antidepressants, and psychotherapy to be successful in improving symptoms and reducing lesions at follow-up visits.

To the Editor:

A 14-year-old Black adolescent girl presented with episodic, painful, edematous plaques that occurred symmetrically on the arms and legs of 5 years’ duration. The plaques evolved into hyperpigmented patches within 24 to 48 hours before eventually resolving. Fatigue, headache, arthralgias of the arms and legs, chest pain, abdominal pain, nausea, and vomiting variably accompanied these episodes.

Prior to visiting our clinic, the patient had been seen by numerous specialists. A review of her medical records revealed an initial diagnosis of Henoch-Schönlein purpura (HSP), then urticarial vasculitis. She had been treated with antihistamines, topical and systemic steroids, hydroxychloroquine, mycophenolate mofetil, dapsone, azathioprine, and gabapentin. All treatments were ineffectual. She underwent extensive diagnostic testing and imaging, which were normal or noncontributory, including type I allergy testing; multiple exhaustive batteries of hematologic testing; and computed tomography/magnetic resonance imaging/magnetic resonance angiography of the brain, chest, abdomen, and pelvic region. Biopsies from symptomatic segments of the gastrointestinal tract were normal.

Chronic treatment with systemic steroids over 9 months resulted in gastritis and an episode of hematemesis requiring emergent hospitalization. A lengthy multidisciplinary evaluation was conducted at the patient’s local community hospital; the team concluded that she had an urticarial-type rash with accompanying symptoms that did not have an autoimmune, rheumatologic, or inflammatory basis.

The patient’s medical history was remarkable for recent-onset panic attacks. Her family medical history was noncontributory. Physical examination revealed multiple violaceous hyperpigmented patches diffusely located on the proximal upper arms (Figure 1). There were no additional findings on physical examination.

Punch biopsies were performed on lesional areas of the arm. Histopathology indicated a mild superficial perivascular dermal mixed infiltrate and extravasated erythrocytes (Figure 2). Direct immunofluorescence (DIF) testing was negative for vasculitis. Immunohistochemical stains for CD117 and tryptase demonstrated a slight increase in the number of dermal mast cells; however, the increase was not sufficient to diagnose cutaneous mastocytosis, which was in the differential. We proposed a diagnosis of psychogenic purpura (PP)(also known as Gardner-Diamond syndrome). She was treated with gabapentin, a selective serotonin reuptake inhibitor, and cognitive therapy. Unfortunately, after starting therapy the patient was lost to follow-up.

Psychogenic purpura is a rare vasculopathy of unknown etiology that may be a special form of factitious disorder.^1,2 In one study, PP occurred predominantly in females aged 15 to 66 years, with a median onset age of 33 years.³ A prodrome of localized itching, burning, and/or pain precedes the development of edematous plaques. The plaques evolve into painful ecchymoses within 1 to 2 days and resolve in 10 days or fewer without treatment. Lesions most commonly occur on the extremities but may occur anywhere on the body. The most common associated finding is an underlying depressive disorder. Episodes may be accompanied by headache, dizziness, fatigue, fever, arthralgia, nausea, vomiting, abdominal pain, menstrual irregularities, myalgia, and urologic conditions.

In 1955, Gardner and Diamond⁴ described the first cases of PP in 4 female patients at Peter Bent Brigham Hospital in Boston, Massachusetts. The investigators were able to replicate the painful ecchymoses with intradermal injection of the patient’s own erythrocytes into the skin. They proposed that the underlying pathogenesis involved autosensitization to erythrocyte stroma.⁴ Since then, others have suggested that the pathogenesis may include autosensitization to erythrocyte phosphatidylserine, tonus dysregulation of venous capillaries, abnormal endothelial fibrin synthesis, and capillary wall instability.^5-7 

Histopathology typically reveals superficial and deep perivascular inflammation with extravasated erythrocytes. Direct immunofluorescence is negative for vasculitis.⁸ Diagnostics and laboratory findings for underlying systemic illness are negative or noncontributory. Cutaneous injection of 1 mL of the patient’s own washed erythrocytes may result in the formation of the characteristic painful plaques within 24 hours; however, this test is limited by lack of standardization and low sensitivity.³

Psychogenic purpura may share clinical features with cutaneous small vessel vasculitis, such as HSP or urticarial vasculitis. Some of the findings that our patient was experiencing, including purpura, arthralgia, and abdominal pain, are associated with HSP. However, HSP typically is self-limiting and classically features palpable purpura distributed across the lower extremities and buttocks. Histopathology demonstrates the classic findings of leukocytoclastic vasculitis; DIF typically is positive for perivascular IgA and C3 deposition. Increased serum IgA may be present.⁹ Urticarial vasculitis appears as erythematous indurated wheals that favor a proximal extremity and truncal distribution. They characteristically last longer than 24 hours, are frequently associated with nonprodromal pain or burning, and resolve with hyperpigmentation. Arthralgia and gastrointestinal, renal, pulmonary, cardiac, and neurologic symptoms may be present, especially in patients with low complement levels.¹⁰ Skin biopsy demonstrates leukocytoclasia that must be accompanied by vessel wall necrosis. Fibrinoid deposition, erythrocyte extravasation, or perivascular inflammation may be present. In 70% of cases revealing perivascular immunoglobulin, C3, and fibrinogen deposition, DIF is positive. Serum C1q autoantibody may be associated with the hypocomplementemic form.¹⁰

The classic histopathologic findings in leukocytoclastic vasculitis include transmural neutrophilic infiltration of the walls of small vessels, fibrinoid necrosis of vessel walls, leukocytoclasia, extravasated erythrocytes, and signs of endothelial cell damage.⁹ A prior punch biopsy in this patient demonstrated rare neutrophilic nuclear debris within the vessel walls without fibrin deposition. Although the presence of nuclear debris and extravasated erythrocytes could be compatible with a manifestation of urticarial vasculitis, the lack of direct evidence of vessel wall necrosis combined with subsequent biopsies unequivocally ruled out cutaneous small vessel vasculitis in our patient.

Psychogenic purpura has been reported to occur frequently in the background of psycho-emotional distress. In 1989, Ratnoff¹¹ noted that many of the patients he was treating at the University Hospitals of Cleveland, Ohio, had a depressive syndrome. A review of patients treated at the Mayo Clinic in Rochester, Minnesota, illustrated concomitant psychiatric illnesses in 41 of 76 (54%) patients treated for PP, most commonly depressive, personality, and anxiety disorders.³

There is no consensus on therapy for PP. Treatment is based on providing symptomatic relief and relieving underlying psychiatric distress. Block et al¹² found the use of selective serotonin reuptake inhibitors, tricyclic antidepressants, and psychotherapy to be successful in improving symptoms and reducing lesions at follow-up visits.

To the Editor:

A 14-year-old Black adolescent girl presented with episodic, painful, edematous plaques that occurred symmetrically on the arms and legs of 5 years’ duration. The plaques evolved into hyperpigmented patches within 24 to 48 hours before eventually resolving. Fatigue, headache, arthralgias of the arms and legs, chest pain, abdominal pain, nausea, and vomiting variably accompanied these episodes.

Prior to visiting our clinic, the patient had been seen by numerous specialists. A review of her medical records revealed an initial diagnosis of Henoch-Schönlein purpura (HSP), then urticarial vasculitis. She had been treated with antihistamines, topical and systemic steroids, hydroxychloroquine, mycophenolate mofetil, dapsone, azathioprine, and gabapentin. All treatments were ineffectual. She underwent extensive diagnostic testing and imaging, which were normal or noncontributory, including type I allergy testing; multiple exhaustive batteries of hematologic testing; and computed tomography/magnetic resonance imaging/magnetic resonance angiography of the brain, chest, abdomen, and pelvic region. Biopsies from symptomatic segments of the gastrointestinal tract were normal.

Chronic treatment with systemic steroids over 9 months resulted in gastritis and an episode of hematemesis requiring emergent hospitalization. A lengthy multidisciplinary evaluation was conducted at the patient’s local community hospital; the team concluded that she had an urticarial-type rash with accompanying symptoms that did not have an autoimmune, rheumatologic, or inflammatory basis.

The patient’s medical history was remarkable for recent-onset panic attacks. Her family medical history was noncontributory. Physical examination revealed multiple violaceous hyperpigmented patches diffusely located on the proximal upper arms (Figure 1). There were no additional findings on physical examination.

Punch biopsies were performed on lesional areas of the arm. Histopathology indicated a mild superficial perivascular dermal mixed infiltrate and extravasated erythrocytes (Figure 2). Direct immunofluorescence (DIF) testing was negative for vasculitis. Immunohistochemical stains for CD117 and tryptase demonstrated a slight increase in the number of dermal mast cells; however, the increase was not sufficient to diagnose cutaneous mastocytosis, which was in the differential. We proposed a diagnosis of psychogenic purpura (PP)(also known as Gardner-Diamond syndrome). She was treated with gabapentin, a selective serotonin reuptake inhibitor, and cognitive therapy. Unfortunately, after starting therapy the patient was lost to follow-up.

Psychogenic purpura is a rare vasculopathy of unknown etiology that may be a special form of factitious disorder.^1,2 In one study, PP occurred predominantly in females aged 15 to 66 years, with a median onset age of 33 years.³ A prodrome of localized itching, burning, and/or pain precedes the development of edematous plaques. The plaques evolve into painful ecchymoses within 1 to 2 days and resolve in 10 days or fewer without treatment. Lesions most commonly occur on the extremities but may occur anywhere on the body. The most common associated finding is an underlying depressive disorder. Episodes may be accompanied by headache, dizziness, fatigue, fever, arthralgia, nausea, vomiting, abdominal pain, menstrual irregularities, myalgia, and urologic conditions.

In 1955, Gardner and Diamond⁴ described the first cases of PP in 4 female patients at Peter Bent Brigham Hospital in Boston, Massachusetts. The investigators were able to replicate the painful ecchymoses with intradermal injection of the patient’s own erythrocytes into the skin. They proposed that the underlying pathogenesis involved autosensitization to erythrocyte stroma.⁴ Since then, others have suggested that the pathogenesis may include autosensitization to erythrocyte phosphatidylserine, tonus dysregulation of venous capillaries, abnormal endothelial fibrin synthesis, and capillary wall instability.^5-7 

Histopathology typically reveals superficial and deep perivascular inflammation with extravasated erythrocytes. Direct immunofluorescence is negative for vasculitis.⁸ Diagnostics and laboratory findings for underlying systemic illness are negative or noncontributory. Cutaneous injection of 1 mL of the patient’s own washed erythrocytes may result in the formation of the characteristic painful plaques within 24 hours; however, this test is limited by lack of standardization and low sensitivity.³

Psychogenic purpura may share clinical features with cutaneous small vessel vasculitis, such as HSP or urticarial vasculitis. Some of the findings that our patient was experiencing, including purpura, arthralgia, and abdominal pain, are associated with HSP. However, HSP typically is self-limiting and classically features palpable purpura distributed across the lower extremities and buttocks. Histopathology demonstrates the classic findings of leukocytoclastic vasculitis; DIF typically is positive for perivascular IgA and C3 deposition. Increased serum IgA may be present.⁹ Urticarial vasculitis appears as erythematous indurated wheals that favor a proximal extremity and truncal distribution. They characteristically last longer than 24 hours, are frequently associated with nonprodromal pain or burning, and resolve with hyperpigmentation. Arthralgia and gastrointestinal, renal, pulmonary, cardiac, and neurologic symptoms may be present, especially in patients with low complement levels.¹⁰ Skin biopsy demonstrates leukocytoclasia that must be accompanied by vessel wall necrosis. Fibrinoid deposition, erythrocyte extravasation, or perivascular inflammation may be present. In 70% of cases revealing perivascular immunoglobulin, C3, and fibrinogen deposition, DIF is positive. Serum C1q autoantibody may be associated with the hypocomplementemic form.¹⁰

The classic histopathologic findings in leukocytoclastic vasculitis include transmural neutrophilic infiltration of the walls of small vessels, fibrinoid necrosis of vessel walls, leukocytoclasia, extravasated erythrocytes, and signs of endothelial cell damage.⁹ A prior punch biopsy in this patient demonstrated rare neutrophilic nuclear debris within the vessel walls without fibrin deposition. Although the presence of nuclear debris and extravasated erythrocytes could be compatible with a manifestation of urticarial vasculitis, the lack of direct evidence of vessel wall necrosis combined with subsequent biopsies unequivocally ruled out cutaneous small vessel vasculitis in our patient.

Psychogenic purpura has been reported to occur frequently in the background of psycho-emotional distress. In 1989, Ratnoff¹¹ noted that many of the patients he was treating at the University Hospitals of Cleveland, Ohio, had a depressive syndrome. A review of patients treated at the Mayo Clinic in Rochester, Minnesota, illustrated concomitant psychiatric illnesses in 41 of 76 (54%) patients treated for PP, most commonly depressive, personality, and anxiety disorders.³

There is no consensus on therapy for PP. Treatment is based on providing symptomatic relief and relieving underlying psychiatric distress. Block et al¹² found the use of selective serotonin reuptake inhibitors, tricyclic antidepressants, and psychotherapy to be successful in improving symptoms and reducing lesions at follow-up visits.

Characteristics

Rhipicephalus ticks belong to the Ixodidae family of hard-bodied ticks. They are large and teardrop shaped with an inornate scutum (hard dorsal plate) and relatively short mouthparts attached at a hexagonal basis capitulum (base of the head to which mouthparts are attached)(Figure).¹ Widely spaced eyes and festoons also are present. The first pair of coxae—attachment base for the first pair of legs—are characteristically bifid; males have a pair of sclerotized adanal plates on the ventral surface adjacent to the anus as well as accessory adanal shields.²Rhipicephalus (formerly Boophilus) microplus (the so-called cattle tick) is a newly added species; it lacks posterior festoons, and the anal groove is absent.³

Almost all Rhipicephalus ticks, except for R microplus, are 3-host ticks in which a single blood meal is consumed from a vertebrate host at each active life stage—larva, nymph, and adult—to complete development.^4,5 In contrast to most ixodid ticks, which are exophilic (living outside of human habitation), the Rhipicephalus sanguineus sensu lato species (the brown dog tick) is highly endophilic (adapted to indoor living) and often can be found hidden in cracks and crevices of walls in homes and peridomestic structures.⁶ It is predominately monotropic (all developmental stages feed on the same host species) and has a strong host preference for dogs, though it occasionally feeds on other hosts (eg, humans).⁷ Although most common in tropical and subtropical climates, they can be found anywhere there are dogs due to their ability to colonize indoor dwellings.⁸ In contrast, R microplus ticks have a predilection for cattle and livestock rather than humans, posing a notable concern to livestock worldwide. Infestation results in transmission of disease-causing pathogens, such as Babesia and Anaplasma species, which costs the cattle industry billions of dollars annually.⁹

Clinical Manifestations and Treatment

Tick bites usually manifest as intensely pruritic, erythematous papules at the site of tick attachment due to a local type IV hypersensitivity reaction to antigens in the tick’s saliva. This reaction can be long-lasting. In addition to pruritic papules following a bite, an attached tick can be mistaken for a skin neoplasm or nevus. Given that ticks are small, especially during the larval stage, dermoscopy may be helpful in making a diagnosis.¹⁰ Symptomatic relief usually can be achieved with topical antipruritics or oral antihistamines.

Of public health concern, brown dog ticks are important vectors of Rickettsia rickettsii (the causative organism of Rocky Mountain spotted fever [RMSF]) in the Western hemisphere, and Rickettsia conorii (the causative organism of Mediterranean spotted fever [MSF][also known as Boutonneuse fever]) in the Eastern hemisphere.¹¹ Bites by ticks carrying rickettsial disease classically manifest with early symptoms of fever, headache, and myalgia, followed by a rash or by a localized eschar or tache noire (a black, necrotic, scabbed lesion) that represents direct endothelial invasion and vascular damage by Rickettsia.¹² Rocky Mountain spotted fever and MSF are more prevalent during summer, likely due, in part, to the combination of increased outdoor activity and a higher rate of tick-questing (host-seeking) behavior in warmer climates.^4,7

Rocky Mountain Spotted Fever—Dermacentor variabilis is the primary vector of RMSF in the southeastern United States; Dermacentor andersoni is the major vector of RMSF in Rocky Mountain states. Rhipicephalus sanguineus sensu lato is an important vector of RMSF in the southwestern United States, Mexico, and Central America.^11,13

Early symptoms of RMSF are nonspecific and can include fever, headache, arthralgia, myalgia, and malaise. Gastrointestinal tract symptoms (eg, nausea, vomiting, anorexia) may occur; notable abdominal pain occurs in some patients, particularly children. A characteristic petechial rash occurs in as many as 90% of patients, typically at the third to fifth day of illness, and classically begins on the wrists and ankles, with progression to the palms and soles before spreading centripetally to the arms, legs, and trunk.¹⁴ An eschar at the inoculation site is uncommon in RMSF; when present, it is more suggestive of MSF.¹⁵

The classic triad of fever, headache, and rash is present in 3% of patients during the first 3 days after a tick bite and in 60% to 70% within 2 weeks.¹⁶ A rash often is absent when patients first seek medical attention and may not develop (absent in 9% to 12% of cases; so-called spotless RMSF). Therefore, absence of rash should not be a reason to withhold treatment.¹⁶ Empiric treatment with doxycycline should be started promptly for all suspected cases of RMSF because of the rapid progression of disease and an increased risk for morbidity and mortality with delayed diagnosis.

Patients do not become antibody positive until 7 to 10 days after symptoms begin; therefore, treatment should not be delayed while awaiting serologic test results. The case fatality rate in the United States is estimated to be 5% to 10% overall and as high as 40% to 50% among patients who are not treated until day 8 or 9 of illness.¹⁷

Cutaneous complications include skin necrosis and gangrene due to continuous tissue damage in severe cases.¹⁶ Severe infection also may manifest with signs of multiorgan system damage, including altered mental status, cerebral edema, meningismus, transient deafness, myocarditis, pulmonary hemorrhage and edema, conjunctivitis, retinal abnormalities, and acute renal failure.^14,16 Risk factors for more severe illness include delayed treatment, age 40 years or older or younger than 10 years, and underlying medical conditions such as alcoholic liver disease and glucose-6-phosphate dehydrogenase deficiency. However, even some healthy young patients die of this disease.¹⁷

Mediterranean Spotted Fever—Rhipicephalus sanguineus sensu lato is the primary vector of MSF, which is prevalent in areas adjacent to the Mediterranean Sea, including southern Europe, Africa, and Central Asia; Sicily is the most highly affected region.¹⁸ Findings with MSF are nearly identical to those of RMSF, except that tache noire is more common, present in as many as 70% of cases at the site of the inoculating tick bite, and MSF typically follows a less severe clinical course.¹² Similar to other rickettsial diseases, the pathogenesis of MSF involves direct injury to vascular endothelial cells, causing a vasculitis that is responsible for the clinical abnormalities observed.

Patients with severe MSF experience complications similar to severe RMSF, including neurologic manifestations and multiorgan damage.¹⁸ Risk factors include advanced age, immunocompromised state, cardiac disease, chronic alcoholism, diabetes mellitus, glucose-6-phosphate dehydrogenase deficiency, respiratory insufficiency, and delayed treatment.¹⁸

Treatment—For all spotted fever group rickettsial infections, doxycycline is the treatment of choice for all patients, including children and pregnant women. Treatment should be started without delay; recommended dosages are 100 mg twice daily for children weighing more than 45 kg and adults, and 2.2 mg/kg twice daily for children weighing 45 kg or less.¹²

Rhipicephalus tick bites rarely can result in paralysis; however, Dermacentor ticks are responsible for most cases of tick-related paralysis in North America. Other pathogens proven or reputed to be transmitted by Rhipicephalus sanguineus sensu lato with zoonotic potential include but are not limited to Rickettsia massiliae, Coxiella burnetti, Anaplasma platys, Leishmania infantum, and Crimean-Congo hemorrhagic fever virus (Nairovirus).¹⁹

Environmental Treatment and Prevention

The most effective way to prevent tick-borne illness is avoidance of tick bites. Primary prevention methods include vector control, use of repellents (eg, N,N-diethyl-meta-toluamide [DEET]), picaridin, permethrin), avoidance of areas with a high tick burden, use of protective clothing, and detection and removal of ticks as soon as possible.

Environmental and veterinary controls also are important methods of tick-bite prevention. A veterinarian can recommend a variety of agents for dogs and cats that prevent attachment of ticks. Environmental controls include synthetic or natural product-based chemical acaricides and nonchemical methods, such as landscape management (eg, sealing cracks and crevices in homes and controlling tall grasses, weeds, and leaf debris) to minimize potential tick habitat.²⁰ Secondary prevention includes antibiotics for prophylaxis or for treatment of tick-borne disease, when indicated.

Numerous tick repellents are available commercially; others are being studied. DEET, the most widely used topical repellent, has a broad spectrum of activity against many tick species.²¹ In addition, DEET has a well-known safety and toxicity profile, with rare adverse effects, and is safe for use in pregnant women and children older than 2 years. Alternative repellents, such as those containing picaridin, ethyl butylacetylaminopropionate (IR3535 [Merck]), oil of lemon eucalyptus, and 2-undecanone can be effective; some show efficacy comparable to that of DEET.²² Permethrin, a synthetic pyrethroid, is a highly efficacious tick repellent and insecticide, especially when used in conjunction with a topical repellent such as DEET. Unlike topically applied repellents, permethrin spray is applied to fabric (eg, clothing, shoes, bed nets, camping gear), not to skin.

Indiscriminate use of acaricides worldwide has led to increasing selection of acaricide resistance in Rhipicephalus tick species, which is especially true with the use of acaricides in controlling R microplus livestock infestations; several tick populations now show resistance to all major classes of these compounds.^23-25 For that reason, there has been an increasing effort to develop new chemical and nonchemical approaches to tick control that are more environmentally sustainable and strategies to minimize development and progression of resistance such as rotation of acaricides; reducing the frequency of their application; use of pesticide mixtures, synergists, or both; and increasing use of nonacaricidal methods of control.²⁶

Prompt removal of ticks is important for preventing the transmission of tick-borne disease. Proper removal involves rubbing the tick in a circular motion with a moist gauze pad or using fine-tipped tweezers to grasp the tick as close to the skin surface as possible and pulling upward with a steady pressure.^17,27 It is important not to jerk, twist, squeeze, smash, or burn the tick, as this can result in insufficient removal of mouthparts or spread contaminated tick fluids to mucous membranes, increasing the risk for infection. Application of petroleum jelly or nail polish to aid in tick removal have not been shown to be effective and are not recommended.^16,28

Characteristics

Rhipicephalus ticks belong to the Ixodidae family of hard-bodied ticks. They are large and teardrop shaped with an inornate scutum (hard dorsal plate) and relatively short mouthparts attached at a hexagonal basis capitulum (base of the head to which mouthparts are attached)(Figure).¹ Widely spaced eyes and festoons also are present. The first pair of coxae—attachment base for the first pair of legs—are characteristically bifid; males have a pair of sclerotized adanal plates on the ventral surface adjacent to the anus as well as accessory adanal shields.²Rhipicephalus (formerly Boophilus) microplus (the so-called cattle tick) is a newly added species; it lacks posterior festoons, and the anal groove is absent.³

Almost all Rhipicephalus ticks, except for R microplus, are 3-host ticks in which a single blood meal is consumed from a vertebrate host at each active life stage—larva, nymph, and adult—to complete development.^4,5 In contrast to most ixodid ticks, which are exophilic (living outside of human habitation), the Rhipicephalus sanguineus sensu lato species (the brown dog tick) is highly endophilic (adapted to indoor living) and often can be found hidden in cracks and crevices of walls in homes and peridomestic structures.⁶ It is predominately monotropic (all developmental stages feed on the same host species) and has a strong host preference for dogs, though it occasionally feeds on other hosts (eg, humans).⁷ Although most common in tropical and subtropical climates, they can be found anywhere there are dogs due to their ability to colonize indoor dwellings.⁸ In contrast, R microplus ticks have a predilection for cattle and livestock rather than humans, posing a notable concern to livestock worldwide. Infestation results in transmission of disease-causing pathogens, such as Babesia and Anaplasma species, which costs the cattle industry billions of dollars annually.⁹

Clinical Manifestations and Treatment

Tick bites usually manifest as intensely pruritic, erythematous papules at the site of tick attachment due to a local type IV hypersensitivity reaction to antigens in the tick’s saliva. This reaction can be long-lasting. In addition to pruritic papules following a bite, an attached tick can be mistaken for a skin neoplasm or nevus. Given that ticks are small, especially during the larval stage, dermoscopy may be helpful in making a diagnosis.¹⁰ Symptomatic relief usually can be achieved with topical antipruritics or oral antihistamines.

Of public health concern, brown dog ticks are important vectors of Rickettsia rickettsii (the causative organism of Rocky Mountain spotted fever [RMSF]) in the Western hemisphere, and Rickettsia conorii (the causative organism of Mediterranean spotted fever [MSF][also known as Boutonneuse fever]) in the Eastern hemisphere.¹¹ Bites by ticks carrying rickettsial disease classically manifest with early symptoms of fever, headache, and myalgia, followed by a rash or by a localized eschar or tache noire (a black, necrotic, scabbed lesion) that represents direct endothelial invasion and vascular damage by Rickettsia.¹² Rocky Mountain spotted fever and MSF are more prevalent during summer, likely due, in part, to the combination of increased outdoor activity and a higher rate of tick-questing (host-seeking) behavior in warmer climates.^4,7

Rocky Mountain Spotted Fever—Dermacentor variabilis is the primary vector of RMSF in the southeastern United States; Dermacentor andersoni is the major vector of RMSF in Rocky Mountain states. Rhipicephalus sanguineus sensu lato is an important vector of RMSF in the southwestern United States, Mexico, and Central America.^11,13

Early symptoms of RMSF are nonspecific and can include fever, headache, arthralgia, myalgia, and malaise. Gastrointestinal tract symptoms (eg, nausea, vomiting, anorexia) may occur; notable abdominal pain occurs in some patients, particularly children. A characteristic petechial rash occurs in as many as 90% of patients, typically at the third to fifth day of illness, and classically begins on the wrists and ankles, with progression to the palms and soles before spreading centripetally to the arms, legs, and trunk.¹⁴ An eschar at the inoculation site is uncommon in RMSF; when present, it is more suggestive of MSF.¹⁵

The classic triad of fever, headache, and rash is present in 3% of patients during the first 3 days after a tick bite and in 60% to 70% within 2 weeks.¹⁶ A rash often is absent when patients first seek medical attention and may not develop (absent in 9% to 12% of cases; so-called spotless RMSF). Therefore, absence of rash should not be a reason to withhold treatment.¹⁶ Empiric treatment with doxycycline should be started promptly for all suspected cases of RMSF because of the rapid progression of disease and an increased risk for morbidity and mortality with delayed diagnosis.

Patients do not become antibody positive until 7 to 10 days after symptoms begin; therefore, treatment should not be delayed while awaiting serologic test results. The case fatality rate in the United States is estimated to be 5% to 10% overall and as high as 40% to 50% among patients who are not treated until day 8 or 9 of illness.¹⁷

Cutaneous complications include skin necrosis and gangrene due to continuous tissue damage in severe cases.¹⁶ Severe infection also may manifest with signs of multiorgan system damage, including altered mental status, cerebral edema, meningismus, transient deafness, myocarditis, pulmonary hemorrhage and edema, conjunctivitis, retinal abnormalities, and acute renal failure.^14,16 Risk factors for more severe illness include delayed treatment, age 40 years or older or younger than 10 years, and underlying medical conditions such as alcoholic liver disease and glucose-6-phosphate dehydrogenase deficiency. However, even some healthy young patients die of this disease.¹⁷

Mediterranean Spotted Fever—Rhipicephalus sanguineus sensu lato is the primary vector of MSF, which is prevalent in areas adjacent to the Mediterranean Sea, including southern Europe, Africa, and Central Asia; Sicily is the most highly affected region.¹⁸ Findings with MSF are nearly identical to those of RMSF, except that tache noire is more common, present in as many as 70% of cases at the site of the inoculating tick bite, and MSF typically follows a less severe clinical course.¹² Similar to other rickettsial diseases, the pathogenesis of MSF involves direct injury to vascular endothelial cells, causing a vasculitis that is responsible for the clinical abnormalities observed.

Patients with severe MSF experience complications similar to severe RMSF, including neurologic manifestations and multiorgan damage.¹⁸ Risk factors include advanced age, immunocompromised state, cardiac disease, chronic alcoholism, diabetes mellitus, glucose-6-phosphate dehydrogenase deficiency, respiratory insufficiency, and delayed treatment.¹⁸

Treatment—For all spotted fever group rickettsial infections, doxycycline is the treatment of choice for all patients, including children and pregnant women. Treatment should be started without delay; recommended dosages are 100 mg twice daily for children weighing more than 45 kg and adults, and 2.2 mg/kg twice daily for children weighing 45 kg or less.¹²

Rhipicephalus tick bites rarely can result in paralysis; however, Dermacentor ticks are responsible for most cases of tick-related paralysis in North America. Other pathogens proven or reputed to be transmitted by Rhipicephalus sanguineus sensu lato with zoonotic potential include but are not limited to Rickettsia massiliae, Coxiella burnetti, Anaplasma platys, Leishmania infantum, and Crimean-Congo hemorrhagic fever virus (Nairovirus).¹⁹

Environmental Treatment and Prevention

The most effective way to prevent tick-borne illness is avoidance of tick bites. Primary prevention methods include vector control, use of repellents (eg, N,N-diethyl-meta-toluamide [DEET]), picaridin, permethrin), avoidance of areas with a high tick burden, use of protective clothing, and detection and removal of ticks as soon as possible.

Environmental and veterinary controls also are important methods of tick-bite prevention. A veterinarian can recommend a variety of agents for dogs and cats that prevent attachment of ticks. Environmental controls include synthetic or natural product-based chemical acaricides and nonchemical methods, such as landscape management (eg, sealing cracks and crevices in homes and controlling tall grasses, weeds, and leaf debris) to minimize potential tick habitat.²⁰ Secondary prevention includes antibiotics for prophylaxis or for treatment of tick-borne disease, when indicated.

Numerous tick repellents are available commercially; others are being studied. DEET, the most widely used topical repellent, has a broad spectrum of activity against many tick species.²¹ In addition, DEET has a well-known safety and toxicity profile, with rare adverse effects, and is safe for use in pregnant women and children older than 2 years. Alternative repellents, such as those containing picaridin, ethyl butylacetylaminopropionate (IR3535 [Merck]), oil of lemon eucalyptus, and 2-undecanone can be effective; some show efficacy comparable to that of DEET.²² Permethrin, a synthetic pyrethroid, is a highly efficacious tick repellent and insecticide, especially when used in conjunction with a topical repellent such as DEET. Unlike topically applied repellents, permethrin spray is applied to fabric (eg, clothing, shoes, bed nets, camping gear), not to skin.

Indiscriminate use of acaricides worldwide has led to increasing selection of acaricide resistance in Rhipicephalus tick species, which is especially true with the use of acaricides in controlling R microplus livestock infestations; several tick populations now show resistance to all major classes of these compounds.^23-25 For that reason, there has been an increasing effort to develop new chemical and nonchemical approaches to tick control that are more environmentally sustainable and strategies to minimize development and progression of resistance such as rotation of acaricides; reducing the frequency of their application; use of pesticide mixtures, synergists, or both; and increasing use of nonacaricidal methods of control.²⁶

Prompt removal of ticks is important for preventing the transmission of tick-borne disease. Proper removal involves rubbing the tick in a circular motion with a moist gauze pad or using fine-tipped tweezers to grasp the tick as close to the skin surface as possible and pulling upward with a steady pressure.^17,27 It is important not to jerk, twist, squeeze, smash, or burn the tick, as this can result in insufficient removal of mouthparts or spread contaminated tick fluids to mucous membranes, increasing the risk for infection. Application of petroleum jelly or nail polish to aid in tick removal have not been shown to be effective and are not recommended.^16,28

Characteristics

Rhipicephalus ticks belong to the Ixodidae family of hard-bodied ticks. They are large and teardrop shaped with an inornate scutum (hard dorsal plate) and relatively short mouthparts attached at a hexagonal basis capitulum (base of the head to which mouthparts are attached)(Figure).¹ Widely spaced eyes and festoons also are present. The first pair of coxae—attachment base for the first pair of legs—are characteristically bifid; males have a pair of sclerotized adanal plates on the ventral surface adjacent to the anus as well as accessory adanal shields.²Rhipicephalus (formerly Boophilus) microplus (the so-called cattle tick) is a newly added species; it lacks posterior festoons, and the anal groove is absent.³

Almost all Rhipicephalus ticks, except for R microplus, are 3-host ticks in which a single blood meal is consumed from a vertebrate host at each active life stage—larva, nymph, and adult—to complete development.^4,5 In contrast to most ixodid ticks, which are exophilic (living outside of human habitation), the Rhipicephalus sanguineus sensu lato species (the brown dog tick) is highly endophilic (adapted to indoor living) and often can be found hidden in cracks and crevices of walls in homes and peridomestic structures.⁶ It is predominately monotropic (all developmental stages feed on the same host species) and has a strong host preference for dogs, though it occasionally feeds on other hosts (eg, humans).⁷ Although most common in tropical and subtropical climates, they can be found anywhere there are dogs due to their ability to colonize indoor dwellings.⁸ In contrast, R microplus ticks have a predilection for cattle and livestock rather than humans, posing a notable concern to livestock worldwide. Infestation results in transmission of disease-causing pathogens, such as Babesia and Anaplasma species, which costs the cattle industry billions of dollars annually.⁹

Clinical Manifestations and Treatment

Tick bites usually manifest as intensely pruritic, erythematous papules at the site of tick attachment due to a local type IV hypersensitivity reaction to antigens in the tick’s saliva. This reaction can be long-lasting. In addition to pruritic papules following a bite, an attached tick can be mistaken for a skin neoplasm or nevus. Given that ticks are small, especially during the larval stage, dermoscopy may be helpful in making a diagnosis.¹⁰ Symptomatic relief usually can be achieved with topical antipruritics or oral antihistamines.

Of public health concern, brown dog ticks are important vectors of Rickettsia rickettsii (the causative organism of Rocky Mountain spotted fever [RMSF]) in the Western hemisphere, and Rickettsia conorii (the causative organism of Mediterranean spotted fever [MSF][also known as Boutonneuse fever]) in the Eastern hemisphere.¹¹ Bites by ticks carrying rickettsial disease classically manifest with early symptoms of fever, headache, and myalgia, followed by a rash or by a localized eschar or tache noire (a black, necrotic, scabbed lesion) that represents direct endothelial invasion and vascular damage by Rickettsia.¹² Rocky Mountain spotted fever and MSF are more prevalent during summer, likely due, in part, to the combination of increased outdoor activity and a higher rate of tick-questing (host-seeking) behavior in warmer climates.^4,7

Rocky Mountain Spotted Fever—Dermacentor variabilis is the primary vector of RMSF in the southeastern United States; Dermacentor andersoni is the major vector of RMSF in Rocky Mountain states. Rhipicephalus sanguineus sensu lato is an important vector of RMSF in the southwestern United States, Mexico, and Central America.^11,13

Early symptoms of RMSF are nonspecific and can include fever, headache, arthralgia, myalgia, and malaise. Gastrointestinal tract symptoms (eg, nausea, vomiting, anorexia) may occur; notable abdominal pain occurs in some patients, particularly children. A characteristic petechial rash occurs in as many as 90% of patients, typically at the third to fifth day of illness, and classically begins on the wrists and ankles, with progression to the palms and soles before spreading centripetally to the arms, legs, and trunk.¹⁴ An eschar at the inoculation site is uncommon in RMSF; when present, it is more suggestive of MSF.¹⁵

The classic triad of fever, headache, and rash is present in 3% of patients during the first 3 days after a tick bite and in 60% to 70% within 2 weeks.¹⁶ A rash often is absent when patients first seek medical attention and may not develop (absent in 9% to 12% of cases; so-called spotless RMSF). Therefore, absence of rash should not be a reason to withhold treatment.¹⁶ Empiric treatment with doxycycline should be started promptly for all suspected cases of RMSF because of the rapid progression of disease and an increased risk for morbidity and mortality with delayed diagnosis.

Patients do not become antibody positive until 7 to 10 days after symptoms begin; therefore, treatment should not be delayed while awaiting serologic test results. The case fatality rate in the United States is estimated to be 5% to 10% overall and as high as 40% to 50% among patients who are not treated until day 8 or 9 of illness.¹⁷

Cutaneous complications include skin necrosis and gangrene due to continuous tissue damage in severe cases.¹⁶ Severe infection also may manifest with signs of multiorgan system damage, including altered mental status, cerebral edema, meningismus, transient deafness, myocarditis, pulmonary hemorrhage and edema, conjunctivitis, retinal abnormalities, and acute renal failure.^14,16 Risk factors for more severe illness include delayed treatment, age 40 years or older or younger than 10 years, and underlying medical conditions such as alcoholic liver disease and glucose-6-phosphate dehydrogenase deficiency. However, even some healthy young patients die of this disease.¹⁷

Mediterranean Spotted Fever—Rhipicephalus sanguineus sensu lato is the primary vector of MSF, which is prevalent in areas adjacent to the Mediterranean Sea, including southern Europe, Africa, and Central Asia; Sicily is the most highly affected region.¹⁸ Findings with MSF are nearly identical to those of RMSF, except that tache noire is more common, present in as many as 70% of cases at the site of the inoculating tick bite, and MSF typically follows a less severe clinical course.¹² Similar to other rickettsial diseases, the pathogenesis of MSF involves direct injury to vascular endothelial cells, causing a vasculitis that is responsible for the clinical abnormalities observed.

Patients with severe MSF experience complications similar to severe RMSF, including neurologic manifestations and multiorgan damage.¹⁸ Risk factors include advanced age, immunocompromised state, cardiac disease, chronic alcoholism, diabetes mellitus, glucose-6-phosphate dehydrogenase deficiency, respiratory insufficiency, and delayed treatment.¹⁸

Treatment—For all spotted fever group rickettsial infections, doxycycline is the treatment of choice for all patients, including children and pregnant women. Treatment should be started without delay; recommended dosages are 100 mg twice daily for children weighing more than 45 kg and adults, and 2.2 mg/kg twice daily for children weighing 45 kg or less.¹²

Rhipicephalus tick bites rarely can result in paralysis; however, Dermacentor ticks are responsible for most cases of tick-related paralysis in North America. Other pathogens proven or reputed to be transmitted by Rhipicephalus sanguineus sensu lato with zoonotic potential include but are not limited to Rickettsia massiliae, Coxiella burnetti, Anaplasma platys, Leishmania infantum, and Crimean-Congo hemorrhagic fever virus (Nairovirus).¹⁹

Environmental Treatment and Prevention

The most effective way to prevent tick-borne illness is avoidance of tick bites. Primary prevention methods include vector control, use of repellents (eg, N,N-diethyl-meta-toluamide [DEET]), picaridin, permethrin), avoidance of areas with a high tick burden, use of protective clothing, and detection and removal of ticks as soon as possible.

Environmental and veterinary controls also are important methods of tick-bite prevention. A veterinarian can recommend a variety of agents for dogs and cats that prevent attachment of ticks. Environmental controls include synthetic or natural product-based chemical acaricides and nonchemical methods, such as landscape management (eg, sealing cracks and crevices in homes and controlling tall grasses, weeds, and leaf debris) to minimize potential tick habitat.²⁰ Secondary prevention includes antibiotics for prophylaxis or for treatment of tick-borne disease, when indicated.

Numerous tick repellents are available commercially; others are being studied. DEET, the most widely used topical repellent, has a broad spectrum of activity against many tick species.²¹ In addition, DEET has a well-known safety and toxicity profile, with rare adverse effects, and is safe for use in pregnant women and children older than 2 years. Alternative repellents, such as those containing picaridin, ethyl butylacetylaminopropionate (IR3535 [Merck]), oil of lemon eucalyptus, and 2-undecanone can be effective; some show efficacy comparable to that of DEET.²² Permethrin, a synthetic pyrethroid, is a highly efficacious tick repellent and insecticide, especially when used in conjunction with a topical repellent such as DEET. Unlike topically applied repellents, permethrin spray is applied to fabric (eg, clothing, shoes, bed nets, camping gear), not to skin.

Indiscriminate use of acaricides worldwide has led to increasing selection of acaricide resistance in Rhipicephalus tick species, which is especially true with the use of acaricides in controlling R microplus livestock infestations; several tick populations now show resistance to all major classes of these compounds.^23-25 For that reason, there has been an increasing effort to develop new chemical and nonchemical approaches to tick control that are more environmentally sustainable and strategies to minimize development and progression of resistance such as rotation of acaricides; reducing the frequency of their application; use of pesticide mixtures, synergists, or both; and increasing use of nonacaricidal methods of control.²⁶

Prompt removal of ticks is important for preventing the transmission of tick-borne disease. Proper removal involves rubbing the tick in a circular motion with a moist gauze pad or using fine-tipped tweezers to grasp the tick as close to the skin surface as possible and pulling upward with a steady pressure.^17,27 It is important not to jerk, twist, squeeze, smash, or burn the tick, as this can result in insufficient removal of mouthparts or spread contaminated tick fluids to mucous membranes, increasing the risk for infection. Application of petroleum jelly or nail polish to aid in tick removal have not been shown to be effective and are not recommended.^16,28

Another study investigated the persistence of targeted therapies for PsA in women compared with men. In a nationwide cohort study using administrative information from French health insurance, the study looked at 14,778 patients (57% women) with PsA who were new users of targeted therapies. The study showed that women had 20%-40% lower treatment persistence rates than men for tumour necrosis factor (TNF) inhibitors (adjusted hazard ratio [aHR] 1.4; 99% CI 1.3-1.5) and interleukin (IL)-17 inhibitors (aHR 1.2; 99% CI 1.1-1.3). However, the treatment persistence between both sexes was comparable for IL-12/23 inhibitor (aHR 1.1; 99% CI 0.9-1.3), IL-23 inhibitor (aHR 1.1; 99% CI 0.7-1.5), and Janus kinase (JAK) inhibitor (aHR 1.2; 99% CI 0.9-1.6) therapies. The paradigm that women have lower treatment persistence is based on studies done primarily in patients treated with TNF inhibitors. This study and a few other recent studies challenge this paradigm by indicating that other targeted therapies, especially JAK inhibitors, may not have lower persistence in women. Sex should be taken into consideration while choosing and counseling women about PsA therapies.

There are few studies on exercise and its impact on PsA. Functional training (FT) and resistance training (RT) may improve functional capacity and quality of life of patients with PsA. The safety of exercise is also not known, given that (micro)trauma is a risk factor for PsA. To evaluate this, Silva and colleagues conducted a 12-week, single-blind trial including 41 patients with PsA who were randomly assigned to undergo FT with elastic bands or RT with weight machines. They demonstrated that FT and RT led to similar improvements in functional capacity measured by the Bath Ankylosing Spondylitis Functional Index (P = .919), functional status measured by the Health Assessment Questionnaire for Spondyloarthritis (P = .932), disease activity measured by the Bath Ankylosing Spondylitis Disease Activity Index (P = .700), and muscle strength. No adverse events occurred in either group. Thus, FT and RT improved functional capacity, functional status, disease activity, and muscle strength to a comparable extent in patients with PsA, with no adverse events. Both modalities may be recommended for PsA patients.

Finally, a cross-sectional study that included 503 patients with PsA, of whom 160 patients underwent treatment escalation, evaluated whether the patient-reported outcome (PsA Impact of Disease questionnaire [PsAID-12]) affected treatment decisions by the treating rheumatologist. Coyle and colleagues demonstrated that although PsAID-12 scores were higher in patients who did vs did not have a treatment escalation, physicians relied more on their assessment of disease activity rather than the PsAID-12 scores when making treatment-related decisions. Of note, physicians also reported that PsAID-12 scores influenced treatment reduction decisions.

Another study investigated the persistence of targeted therapies for PsA in women compared with men. In a nationwide cohort study using administrative information from French health insurance, the study looked at 14,778 patients (57% women) with PsA who were new users of targeted therapies. The study showed that women had 20%-40% lower treatment persistence rates than men for tumour necrosis factor (TNF) inhibitors (adjusted hazard ratio [aHR] 1.4; 99% CI 1.3-1.5) and interleukin (IL)-17 inhibitors (aHR 1.2; 99% CI 1.1-1.3). However, the treatment persistence between both sexes was comparable for IL-12/23 inhibitor (aHR 1.1; 99% CI 0.9-1.3), IL-23 inhibitor (aHR 1.1; 99% CI 0.7-1.5), and Janus kinase (JAK) inhibitor (aHR 1.2; 99% CI 0.9-1.6) therapies. The paradigm that women have lower treatment persistence is based on studies done primarily in patients treated with TNF inhibitors. This study and a few other recent studies challenge this paradigm by indicating that other targeted therapies, especially JAK inhibitors, may not have lower persistence in women. Sex should be taken into consideration while choosing and counseling women about PsA therapies.

There are few studies on exercise and its impact on PsA. Functional training (FT) and resistance training (RT) may improve functional capacity and quality of life of patients with PsA. The safety of exercise is also not known, given that (micro)trauma is a risk factor for PsA. To evaluate this, Silva and colleagues conducted a 12-week, single-blind trial including 41 patients with PsA who were randomly assigned to undergo FT with elastic bands or RT with weight machines. They demonstrated that FT and RT led to similar improvements in functional capacity measured by the Bath Ankylosing Spondylitis Functional Index (P = .919), functional status measured by the Health Assessment Questionnaire for Spondyloarthritis (P = .932), disease activity measured by the Bath Ankylosing Spondylitis Disease Activity Index (P = .700), and muscle strength. No adverse events occurred in either group. Thus, FT and RT improved functional capacity, functional status, disease activity, and muscle strength to a comparable extent in patients with PsA, with no adverse events. Both modalities may be recommended for PsA patients.

Finally, a cross-sectional study that included 503 patients with PsA, of whom 160 patients underwent treatment escalation, evaluated whether the patient-reported outcome (PsA Impact of Disease questionnaire [PsAID-12]) affected treatment decisions by the treating rheumatologist. Coyle and colleagues demonstrated that although PsAID-12 scores were higher in patients who did vs did not have a treatment escalation, physicians relied more on their assessment of disease activity rather than the PsAID-12 scores when making treatment-related decisions. Of note, physicians also reported that PsAID-12 scores influenced treatment reduction decisions.

Another study investigated the persistence of targeted therapies for PsA in women compared with men. In a nationwide cohort study using administrative information from French health insurance, the study looked at 14,778 patients (57% women) with PsA who were new users of targeted therapies. The study showed that women had 20%-40% lower treatment persistence rates than men for tumour necrosis factor (TNF) inhibitors (adjusted hazard ratio [aHR] 1.4; 99% CI 1.3-1.5) and interleukin (IL)-17 inhibitors (aHR 1.2; 99% CI 1.1-1.3). However, the treatment persistence between both sexes was comparable for IL-12/23 inhibitor (aHR 1.1; 99% CI 0.9-1.3), IL-23 inhibitor (aHR 1.1; 99% CI 0.7-1.5), and Janus kinase (JAK) inhibitor (aHR 1.2; 99% CI 0.9-1.6) therapies. The paradigm that women have lower treatment persistence is based on studies done primarily in patients treated with TNF inhibitors. This study and a few other recent studies challenge this paradigm by indicating that other targeted therapies, especially JAK inhibitors, may not have lower persistence in women. Sex should be taken into consideration while choosing and counseling women about PsA therapies.

There are few studies on exercise and its impact on PsA. Functional training (FT) and resistance training (RT) may improve functional capacity and quality of life of patients with PsA. The safety of exercise is also not known, given that (micro)trauma is a risk factor for PsA. To evaluate this, Silva and colleagues conducted a 12-week, single-blind trial including 41 patients with PsA who were randomly assigned to undergo FT with elastic bands or RT with weight machines. They demonstrated that FT and RT led to similar improvements in functional capacity measured by the Bath Ankylosing Spondylitis Functional Index (P = .919), functional status measured by the Health Assessment Questionnaire for Spondyloarthritis (P = .932), disease activity measured by the Bath Ankylosing Spondylitis Disease Activity Index (P = .700), and muscle strength. No adverse events occurred in either group. Thus, FT and RT improved functional capacity, functional status, disease activity, and muscle strength to a comparable extent in patients with PsA, with no adverse events. Both modalities may be recommended for PsA patients.

Finally, a cross-sectional study that included 503 patients with PsA, of whom 160 patients underwent treatment escalation, evaluated whether the patient-reported outcome (PsA Impact of Disease questionnaire [PsAID-12]) affected treatment decisions by the treating rheumatologist. Coyle and colleagues demonstrated that although PsAID-12 scores were higher in patients who did vs did not have a treatment escalation, physicians relied more on their assessment of disease activity rather than the PsAID-12 scores when making treatment-related decisions. Of note, physicians also reported that PsAID-12 scores influenced treatment reduction decisions.