Bradley Monash, MD

DISCUSSION

KEY TEACHING POINTS

DISCUSSION

KEY TEACHING POINTS

DISCUSSION

KEY TEACHING POINTS

A previously healthy 11‐year‐old boy presented to the emergency department after referral from his pediatrician for 1 week of fevers. Seven days prior to admission he developed a fever to 40.8C, vomiting, and mild left knee pain. The vomiting resolved within 2 days. Five days prior to admission he developed a pruritic, pinpoint rash over his abdomen that resolved within 24 hours. He also developed red, cracked lips, redness of his tongue, redness surrounding his eyes, and slight swelling of his hands. Three days prior to admission his pediatrician noted a 1‐cm anterior cervical lymph node. His fevers occurred throughout each of the prior 7 days without a discernible pattern, and his mild knee pain persisted at the time of presentation.

This preteen has had high fevers for 1 week associated with arthralgia, pruritic rash, emesis, and oral mucosal erythema. His rash, lip and tongue erythema, and swollen hands are classic features of Kawasaki disease (KD), but he lacks the other characteristic physical examination findings. The diagnosis of KD requires fever for at least 5 days accompanied by 4 of the following 5 signs: polymorphous rash, oral mucous membrane changes, peripheral extremity changes such as swelling or skin desquamation, bilateral bulbar conjunctival injection, and cervical lymphadenopathy >1.5 cm in diameter. Children meeting fewer than 4 of these criteria may have an incomplete form of KD.

Because most patients with KD (80%) are under 5 years old, alternative diagnoses such as autoimmune illnesses or a hypersensitivity reaction should be considered. Travel, medication, and animal exposure histories may reveal clues to an infectious or drug‐induced etiology of his fever. Immunization status should be assessed, as measles is also associated with fever, rash, and mucosal changes. Arthralgia or arthritis may occur in KD, but these findings suggest the need to entertain other possibilities, including bone or joint infection, infective endocarditis, inflammatory bowel disease, juvenile idiopathic arthritis (JIA), or systemic lupus erythematosus (SLE).

The child's only past medical history was an episode of croup as an infant. There was no family history of autoimmune diseases. He was not taking any medications and had no known allergies. His immunizations were up to date, including measles, mumps, rubella, and varicella. He lived with his parents and his dog. He swam in fresh water during a trip to Maine 2 months earlier. Neither he nor his family recalled a tick bite. He had no exposure to raw meat or unpasteurized dairy products.

The travel to New England raises the possibility of Lyme disease, although a 2‐month interval between exposure and a high, prolonged fever would be very unusual. Knee arthralgia or arthritis is common in children with late‐stage Lyme disease, but can also be seen in early‐disseminated disease. The prior description of the rash is not suggestive of erythema chronicum migrans, which is seen in early‐stage Lyme disease.

C‐reactive protein (CRP) was 189 mg/L (normal <6.3 mg/L). An echocardiogram was normal. Intravenous immunoglobulin (IVIG) was administered for presumed KD, with immediate improvement of the periorbital erythema, tongue redness, and hand swelling. He was discharged the next day on aspirin with cardiology clinic follow‐up.

Improvement after IVIG supports the diagnosis of KD. It is typical to discharge KD patients from the hospital when they have been afebrile for 24 hours or when the CRP level has declined by approximately 50%.

Over the next 48 hours he felt unwell with high‐grade fevers, continued left knee pain, and new left hip pain. He was readmitted to the hospital. His temperature was 39.4C, respiratory rate was 22 breaths per minute, heart rate was 122 beats per minute, blood pressure was 103/50 mm Hg, and oxygen saturation was 100% while breathing ambient air. He appeared mildly uncomfortable. His conjunctivae were normal. His lips were dry, red, and cracked, and his tongue was red with prominent papillae. His neck was supple without lymphadenopathy. His lungs were clear to auscultation. His heart exam was without murmurs. His abdomen was soft, and the liver and spleen were not enlarged. He had no swelling or erythema of his joints; however, he experienced pain with range of motion of his left knee, and tenderness and restricted range of motion of his left hip. His neurologic exam was normal. There were no rashes.

He has persistent fever, tachycardia, and tachypnea, now without features of KD except oral mucosal changes including prominent tongue papillae consistent with a strawberry tongue. Continued or recurrent fever may suggest persistent KD with ongoing inflammation or the need to search for an alternative diagnoses. An echocardiogram should be repeated, as the coronary artery abnormalities in KD can evolve rapidly, particularly when inflammation persists. Additional findings may include decreased left ventricular function, mitral regurgitation, or pericardial effusion. A second dose of IVIG is necessary to control fever and inflammation in about 15% of patients with KD, although in this case IVIG should be withheld pending further evaluation.

Arthralgia occurs commonly in KD, whereas frank arthritis is less typical. Polyarticular or oligoarticular arthritis involving small or large joints (especially knee or ankle) affects 5% to 10% of patients. The severity of findings in his left hip warrants consideration of septic arthritis with pain referred to the knee; pelvic or femoral osteomyelitis; psoas abscess; or pyomyositis. Following basic lab tests, imaging of the left hip region is indicated.

Laboratory evaluation revealed: white blood cell (WBC) count 10,000/L (absolute neutrophil count 8,460/L, absolute lymphocyte count 530/L), hemoglobin 10.6 g/dL, platelet count 208,000/L, serum sodium 130 mmol/L, serum potassium 3.3 mmol/L, serum urea nitrogen 11 mg/dL, serum creatinine 0.54 mg/dL, aspartate transaminase (AST) 26 U/L, alanine transaminase (ALT) 31 U/L, albumin 1.7 g/dL, erythrocyte sedimentation rate (ESR) > 100 mm/h, and CRP 263 mg/L. No blast cells were seen on peripheral blood smear.

Hypoalbuminemia and markedly elevated inflammatory markers indicate an inflammatory condition that has been active for more than a week. Assessing ESR after IVIG therapy is not useful because exogenous globulins increase the ESR; however, CRP is useful to monitor inflammation and remains elevated here.

Incomplete KD is still possible. Hyponatremia, hypoalbuminemia, and anemia are all features of persistent KD, and have been utilized in several clinical scoring systems in Japan to identify KD patients at increased risk for developing coronary complications. A neoplastic process cannot be excluded, but does not appear likely based on the acuity of his presentation and peripheral blood smear review.

Upon readmission he received a second dose of 2 g/kg IVIG. He remained on aspirin and continued to have fevers. A repeat echocardiogram was normal. He had worsening pain in his left knee and hip with difficulty straightening his left leg. Physical examination was notable for tenderness to palpation over his left hip joint, refusal to bear weight, and resistance to passive range of motion. On hospital day 2, an ultrasound of his left hip and knee revealed a complex left hip effusion and small left knee effusion.

KD becomes less likely in the presence of persistent fevers after IVIG and a repeatedly normal echocardiogram. Worsening left leg symptoms including impaired hip extension with a complex hip effusion suggests an infectious process in or adjacent to the left hip, such as septic arthritis, myositis, or osteomyelitis of the pelvis or proximal femur. A complex hip effusion is less likely to be present with arthritis related to JIA or SLE. The patient needs an emergent hip aspiration and possibly magnetic resonance imaging (MRI) to evaluate adjacent structures.

Arthrotomy and open drainage of his left hip revealed purulent fluid with a WBC count of 49,000/L with 89% neutrophils and 2% lymphocytes. Gram stain was negative. A left knee aspirate demonstrated straw‐colored synovial fluid (which was not sent for cell counts). Bacterial, fungal, and acid‐fast bacilli cultures were requested from hip and knee aspirates. Intravenous ceftriaxone and vancomycin were administered.

The most likely organism in pediatric pyogenic arthritis is Staphylococcus aureus, but there is a long list of other potential pathogens, including Streptococcus pyogenes (group A streptococcus) and Streptococcus pneumoniae. Most pediatric patients with acute pyogenic arthritis have synovial fluid WBC counts in excess of 75,000 to 100,000/L. The protracted course and the initial lack of hip symptoms raise the possibility of a primary osteomyelitis of the femur (particularly the intracapsular portion of the femoral neck or head) or of the acetabulum, with subsequent extension into the hip joint. Pyogenic myositis involving muscle groups adjacent to the hip would be unlikely to spread into the hip space, but can lead to synovial irritation, characterized by sterile joint fluid and WBC counts that fall short of the usual numbers seen in septic arthritis. The blood supply to the femoral head can become compromised with prolonged inflammation and increased intracapsular pressure, resulting in aseptic necrosis.

All cultures from his hip and knee aspirations were sterile. He continued to have daily fevers and persistent tachycardia while receiving intravenous ceftriaxone and vancomycin. Additional testing was notable for: antinuclear antibody (ANA) 1:80, anti‐streptolysin O (ASO) titer 344 IU (normal <150 IU), AST and ALT within normal limits, ferritin 568 ng/mL (normal <322 ng/mL), and lactate dehydrogenase (LDH) 212 units/L (normal <257 units/L). Abdominal ultrasound revealed borderline hepatosplenomegaly. An ophthalmologic examination was normal.

On postoperative day 4 he developed left upper thigh swelling. An MRI showed rim‐enhancing juxta‐articular complex fluid collections surrounding the left femur with decreased marrow enhancement of the left proximal femur (Figure 1).

Figure 1

The limited rheumatologic evaluation is unrevealing; the ANA result is nondiagnostic and the ASO titer is normal for age. Laboratories generally report adult normal values for streptococcal antibodies regardless of the patient's age; children from ages 7 to 12 years are at their life peak frequency of group A streptococcal pharyngitis and typically have higher normal values of streptococcal antibodies, including ASO (up to about 480640 IU). The moderately elevated ferritin level is most likely an acute phase reactant and not high enough to suggest macrophage activation syndrome, which is unlikely with the normal AST, ALT, and LDH levels, the absence of significant splenomegaly, and the lack of cytopenias. Continued fever with progressive left upper thigh swelling point to osteomyelitis of the proximal femur, which may have initially ruptured into the hip and then infiltrated the femoral cortex and spread infection into the adjacent soft tissues. Surgical debridement is indicated.

The relative prevalence of methicillin‐sensitive S aureus (MSSA) and methicillin‐resistant S aureus vary widely with geography. MSSA strains are more likely to be highly toxigenic. The elaboration of 1 or more extracellular toxins could account for the patient's initial symptoms.

The patient was brought back to the operating room for drainage of the juxta‐articular fluid collections and a biopsy of his femur. The fluid collections were grossly purulent. His intraoperative cultures were positive for MSSA. The bone biopsy revealed necrotic tissue, acute inflammation, and bacterial colonies, consistent with acute osteomyelitis. Further testing of his S aureus isolate was positive for staphylococcal enterotoxin B. He completed a 4‐week course of oral clindamycin with subsequent normalization of his hip exam and inflammatory markers. At a follow‐up visit the patient was feeling better, but had developed skin peeling on the lateral aspects of his feet consistent with late sequelae of toxin‐mediated disease (Figure 2). Three months after discharge the patient had returned to his baseline activity level and remained asymptomatic.

Figure 2

COMMENTARY

The patient presented with a constellation of symptoms that was initially mistaken for incomplete KD until focal progression of his symptoms exposed an underlying femoral osteomyelitis with periarticular abscess formation. Bacterial cultures and subsequent toxin assay revealed an enterotoxin B‐producing strain of S aureus.

Certain staphylococcal strains secrete superantigens that may lead to the development of a systemic toxin‐mediated syndrome. Toxins elaborated by S aureus include toxic shock syndrome toxin‐1 (TSST‐1) and enterotoxins, which have been implicated in menstrual and nonmenstrual toxic shock syndromes.[1, 2] Enterotoxin B is a staphylococcal superantigen found in most strains of the USA400 clonal group, and has been frequently associated with skin and soft tissue infections.[3] Enterotoxin B production has been reported in nearly half of S aureus isolates from skin, soft tissue, and bone infections.[4]

Staphylococcal and streptococcal toxin‐mediated diseases can mimic vasculitis, systemic juvenile idiopathic arthritis, viral infections, and Stevens‐Johnson syndrome. Glossitis in toxin‐mediated syndromes manifests with a swollen, red tongue with overlying enlarged papillae, giving the appearance of a strawberry. Although pediatric providers often equate strawberry tongue, conjunctival injection, rash, and erythematous lips with KD, these findings are also seen in toxin‐mediated diseases, such as scarlet fever or staphylococcal toxic shock syndrome. Enterotoxin B mediated staphylococcal disease masquerading as KD has been reported in 2 cases: a 7‐month‐old boy with multifocal S aureus osteomyelitis and a 5‐year‐old boy with S aureus bacteremia. Both staphylococcal isolates produced enterotoxin B but were negative for other staphylococcus‐related toxins including TSST‐1.[5]

The Institute of Medicine (IOM) recently released its report Improving Diagnosis in Health Care, highlighting the under‐recognized quality and safety issue of diagnostic error.[6] The report uses the following broad and patient‐centered definition of diagnostic error: the failure to (a) establish an accurate and timely explanation of the patient's health problem(s) or (b) communicate that explanation to the patient. The IOM's conceptual model of diagnosis emphasizes the iterative nature of the diagnostic process, including the importance of generating a working diagnosis, gathering and incorporating new information in the reassessment of that diagnosis, and integrating treatment response into the formulation of the final diagnosis (Figure 3).

Figure 3

Even though the patient initially had several features consistent with KD, the increasing number of atypical features could have prompted the clinical team to reconsider their working diagnosis. The patient's age was atypical for KD, he had progressive knee and hip arthritis, and his fevers persisted after IVIG. An expanded differential should have included toxic shock syndrome; the resolution of conjunctival and mucosal injection and edema after IVIG may have been the result of antibodies in the IVIG preparation with neutralizing activity against superantigens. This antitoxin activity has established a role for IVIG in the management of staphylococcal toxic shock syndrome.[7] Ultimately, his imaging and surgical drainage revealed a focal staphylococcal toxin‐producing infectious source from which his fevers, rash, and mucosal and extremity changes emanated. This case reminds us that the more atypical a working diagnosis iseither in its presentation or treatment responsethe more readily clinicians should take it for another spin around the diagnostic wheel in search of a more suitable alternative.

KEY LEARNING POINTS

1. Staphylococcal toxin‐mediated disease may mimic KD, with common features including strawberry tongue, oral and conjunctival injection, and skin desquamation.
2. Improvement after treatment with IVIG is characteristic but not diagnostic of KD, and may be seen in toxin‐mediated disease.
3. KD may present with arthralgia or arthritis, but severe joint abnormalities warrant consideration of infectious and other autoimmune conditions.
4. The more atypical a working diagnosis iseither in its presentation or treatment responsethe more readily clinicians should gather, interpret, and integrate new information in search of a more suitable alternative.

Disclosure: Nothing to report.

A previously healthy 11‐year‐old boy presented to the emergency department after referral from his pediatrician for 1 week of fevers. Seven days prior to admission he developed a fever to 40.8C, vomiting, and mild left knee pain. The vomiting resolved within 2 days. Five days prior to admission he developed a pruritic, pinpoint rash over his abdomen that resolved within 24 hours. He also developed red, cracked lips, redness of his tongue, redness surrounding his eyes, and slight swelling of his hands. Three days prior to admission his pediatrician noted a 1‐cm anterior cervical lymph node. His fevers occurred throughout each of the prior 7 days without a discernible pattern, and his mild knee pain persisted at the time of presentation.

This preteen has had high fevers for 1 week associated with arthralgia, pruritic rash, emesis, and oral mucosal erythema. His rash, lip and tongue erythema, and swollen hands are classic features of Kawasaki disease (KD), but he lacks the other characteristic physical examination findings. The diagnosis of KD requires fever for at least 5 days accompanied by 4 of the following 5 signs: polymorphous rash, oral mucous membrane changes, peripheral extremity changes such as swelling or skin desquamation, bilateral bulbar conjunctival injection, and cervical lymphadenopathy >1.5 cm in diameter. Children meeting fewer than 4 of these criteria may have an incomplete form of KD.

Because most patients with KD (80%) are under 5 years old, alternative diagnoses such as autoimmune illnesses or a hypersensitivity reaction should be considered. Travel, medication, and animal exposure histories may reveal clues to an infectious or drug‐induced etiology of his fever. Immunization status should be assessed, as measles is also associated with fever, rash, and mucosal changes. Arthralgia or arthritis may occur in KD, but these findings suggest the need to entertain other possibilities, including bone or joint infection, infective endocarditis, inflammatory bowel disease, juvenile idiopathic arthritis (JIA), or systemic lupus erythematosus (SLE).

The child's only past medical history was an episode of croup as an infant. There was no family history of autoimmune diseases. He was not taking any medications and had no known allergies. His immunizations were up to date, including measles, mumps, rubella, and varicella. He lived with his parents and his dog. He swam in fresh water during a trip to Maine 2 months earlier. Neither he nor his family recalled a tick bite. He had no exposure to raw meat or unpasteurized dairy products.

The travel to New England raises the possibility of Lyme disease, although a 2‐month interval between exposure and a high, prolonged fever would be very unusual. Knee arthralgia or arthritis is common in children with late‐stage Lyme disease, but can also be seen in early‐disseminated disease. The prior description of the rash is not suggestive of erythema chronicum migrans, which is seen in early‐stage Lyme disease.

C‐reactive protein (CRP) was 189 mg/L (normal <6.3 mg/L). An echocardiogram was normal. Intravenous immunoglobulin (IVIG) was administered for presumed KD, with immediate improvement of the periorbital erythema, tongue redness, and hand swelling. He was discharged the next day on aspirin with cardiology clinic follow‐up.

Improvement after IVIG supports the diagnosis of KD. It is typical to discharge KD patients from the hospital when they have been afebrile for 24 hours or when the CRP level has declined by approximately 50%.

Over the next 48 hours he felt unwell with high‐grade fevers, continued left knee pain, and new left hip pain. He was readmitted to the hospital. His temperature was 39.4C, respiratory rate was 22 breaths per minute, heart rate was 122 beats per minute, blood pressure was 103/50 mm Hg, and oxygen saturation was 100% while breathing ambient air. He appeared mildly uncomfortable. His conjunctivae were normal. His lips were dry, red, and cracked, and his tongue was red with prominent papillae. His neck was supple without lymphadenopathy. His lungs were clear to auscultation. His heart exam was without murmurs. His abdomen was soft, and the liver and spleen were not enlarged. He had no swelling or erythema of his joints; however, he experienced pain with range of motion of his left knee, and tenderness and restricted range of motion of his left hip. His neurologic exam was normal. There were no rashes.

He has persistent fever, tachycardia, and tachypnea, now without features of KD except oral mucosal changes including prominent tongue papillae consistent with a strawberry tongue. Continued or recurrent fever may suggest persistent KD with ongoing inflammation or the need to search for an alternative diagnoses. An echocardiogram should be repeated, as the coronary artery abnormalities in KD can evolve rapidly, particularly when inflammation persists. Additional findings may include decreased left ventricular function, mitral regurgitation, or pericardial effusion. A second dose of IVIG is necessary to control fever and inflammation in about 15% of patients with KD, although in this case IVIG should be withheld pending further evaluation.

Arthralgia occurs commonly in KD, whereas frank arthritis is less typical. Polyarticular or oligoarticular arthritis involving small or large joints (especially knee or ankle) affects 5% to 10% of patients. The severity of findings in his left hip warrants consideration of septic arthritis with pain referred to the knee; pelvic or femoral osteomyelitis; psoas abscess; or pyomyositis. Following basic lab tests, imaging of the left hip region is indicated.

Laboratory evaluation revealed: white blood cell (WBC) count 10,000/L (absolute neutrophil count 8,460/L, absolute lymphocyte count 530/L), hemoglobin 10.6 g/dL, platelet count 208,000/L, serum sodium 130 mmol/L, serum potassium 3.3 mmol/L, serum urea nitrogen 11 mg/dL, serum creatinine 0.54 mg/dL, aspartate transaminase (AST) 26 U/L, alanine transaminase (ALT) 31 U/L, albumin 1.7 g/dL, erythrocyte sedimentation rate (ESR) > 100 mm/h, and CRP 263 mg/L. No blast cells were seen on peripheral blood smear.

Hypoalbuminemia and markedly elevated inflammatory markers indicate an inflammatory condition that has been active for more than a week. Assessing ESR after IVIG therapy is not useful because exogenous globulins increase the ESR; however, CRP is useful to monitor inflammation and remains elevated here.

Incomplete KD is still possible. Hyponatremia, hypoalbuminemia, and anemia are all features of persistent KD, and have been utilized in several clinical scoring systems in Japan to identify KD patients at increased risk for developing coronary complications. A neoplastic process cannot be excluded, but does not appear likely based on the acuity of his presentation and peripheral blood smear review.

Upon readmission he received a second dose of 2 g/kg IVIG. He remained on aspirin and continued to have fevers. A repeat echocardiogram was normal. He had worsening pain in his left knee and hip with difficulty straightening his left leg. Physical examination was notable for tenderness to palpation over his left hip joint, refusal to bear weight, and resistance to passive range of motion. On hospital day 2, an ultrasound of his left hip and knee revealed a complex left hip effusion and small left knee effusion.

KD becomes less likely in the presence of persistent fevers after IVIG and a repeatedly normal echocardiogram. Worsening left leg symptoms including impaired hip extension with a complex hip effusion suggests an infectious process in or adjacent to the left hip, such as septic arthritis, myositis, or osteomyelitis of the pelvis or proximal femur. A complex hip effusion is less likely to be present with arthritis related to JIA or SLE. The patient needs an emergent hip aspiration and possibly magnetic resonance imaging (MRI) to evaluate adjacent structures.

Arthrotomy and open drainage of his left hip revealed purulent fluid with a WBC count of 49,000/L with 89% neutrophils and 2% lymphocytes. Gram stain was negative. A left knee aspirate demonstrated straw‐colored synovial fluid (which was not sent for cell counts). Bacterial, fungal, and acid‐fast bacilli cultures were requested from hip and knee aspirates. Intravenous ceftriaxone and vancomycin were administered.

The most likely organism in pediatric pyogenic arthritis is Staphylococcus aureus, but there is a long list of other potential pathogens, including Streptococcus pyogenes (group A streptococcus) and Streptococcus pneumoniae. Most pediatric patients with acute pyogenic arthritis have synovial fluid WBC counts in excess of 75,000 to 100,000/L. The protracted course and the initial lack of hip symptoms raise the possibility of a primary osteomyelitis of the femur (particularly the intracapsular portion of the femoral neck or head) or of the acetabulum, with subsequent extension into the hip joint. Pyogenic myositis involving muscle groups adjacent to the hip would be unlikely to spread into the hip space, but can lead to synovial irritation, characterized by sterile joint fluid and WBC counts that fall short of the usual numbers seen in septic arthritis. The blood supply to the femoral head can become compromised with prolonged inflammation and increased intracapsular pressure, resulting in aseptic necrosis.

All cultures from his hip and knee aspirations were sterile. He continued to have daily fevers and persistent tachycardia while receiving intravenous ceftriaxone and vancomycin. Additional testing was notable for: antinuclear antibody (ANA) 1:80, anti‐streptolysin O (ASO) titer 344 IU (normal <150 IU), AST and ALT within normal limits, ferritin 568 ng/mL (normal <322 ng/mL), and lactate dehydrogenase (LDH) 212 units/L (normal <257 units/L). Abdominal ultrasound revealed borderline hepatosplenomegaly. An ophthalmologic examination was normal.

On postoperative day 4 he developed left upper thigh swelling. An MRI showed rim‐enhancing juxta‐articular complex fluid collections surrounding the left femur with decreased marrow enhancement of the left proximal femur (Figure 1).

Figure 1

The limited rheumatologic evaluation is unrevealing; the ANA result is nondiagnostic and the ASO titer is normal for age. Laboratories generally report adult normal values for streptococcal antibodies regardless of the patient's age; children from ages 7 to 12 years are at their life peak frequency of group A streptococcal pharyngitis and typically have higher normal values of streptococcal antibodies, including ASO (up to about 480640 IU). The moderately elevated ferritin level is most likely an acute phase reactant and not high enough to suggest macrophage activation syndrome, which is unlikely with the normal AST, ALT, and LDH levels, the absence of significant splenomegaly, and the lack of cytopenias. Continued fever with progressive left upper thigh swelling point to osteomyelitis of the proximal femur, which may have initially ruptured into the hip and then infiltrated the femoral cortex and spread infection into the adjacent soft tissues. Surgical debridement is indicated.

The relative prevalence of methicillin‐sensitive S aureus (MSSA) and methicillin‐resistant S aureus vary widely with geography. MSSA strains are more likely to be highly toxigenic. The elaboration of 1 or more extracellular toxins could account for the patient's initial symptoms.

The patient was brought back to the operating room for drainage of the juxta‐articular fluid collections and a biopsy of his femur. The fluid collections were grossly purulent. His intraoperative cultures were positive for MSSA. The bone biopsy revealed necrotic tissue, acute inflammation, and bacterial colonies, consistent with acute osteomyelitis. Further testing of his S aureus isolate was positive for staphylococcal enterotoxin B. He completed a 4‐week course of oral clindamycin with subsequent normalization of his hip exam and inflammatory markers. At a follow‐up visit the patient was feeling better, but had developed skin peeling on the lateral aspects of his feet consistent with late sequelae of toxin‐mediated disease (Figure 2). Three months after discharge the patient had returned to his baseline activity level and remained asymptomatic.

Figure 2

COMMENTARY

The patient presented with a constellation of symptoms that was initially mistaken for incomplete KD until focal progression of his symptoms exposed an underlying femoral osteomyelitis with periarticular abscess formation. Bacterial cultures and subsequent toxin assay revealed an enterotoxin B‐producing strain of S aureus.

Certain staphylococcal strains secrete superantigens that may lead to the development of a systemic toxin‐mediated syndrome. Toxins elaborated by S aureus include toxic shock syndrome toxin‐1 (TSST‐1) and enterotoxins, which have been implicated in menstrual and nonmenstrual toxic shock syndromes.[1, 2] Enterotoxin B is a staphylococcal superantigen found in most strains of the USA400 clonal group, and has been frequently associated with skin and soft tissue infections.[3] Enterotoxin B production has been reported in nearly half of S aureus isolates from skin, soft tissue, and bone infections.[4]

Staphylococcal and streptococcal toxin‐mediated diseases can mimic vasculitis, systemic juvenile idiopathic arthritis, viral infections, and Stevens‐Johnson syndrome. Glossitis in toxin‐mediated syndromes manifests with a swollen, red tongue with overlying enlarged papillae, giving the appearance of a strawberry. Although pediatric providers often equate strawberry tongue, conjunctival injection, rash, and erythematous lips with KD, these findings are also seen in toxin‐mediated diseases, such as scarlet fever or staphylococcal toxic shock syndrome. Enterotoxin B mediated staphylococcal disease masquerading as KD has been reported in 2 cases: a 7‐month‐old boy with multifocal S aureus osteomyelitis and a 5‐year‐old boy with S aureus bacteremia. Both staphylococcal isolates produced enterotoxin B but were negative for other staphylococcus‐related toxins including TSST‐1.[5]

The Institute of Medicine (IOM) recently released its report Improving Diagnosis in Health Care, highlighting the under‐recognized quality and safety issue of diagnostic error.[6] The report uses the following broad and patient‐centered definition of diagnostic error: the failure to (a) establish an accurate and timely explanation of the patient's health problem(s) or (b) communicate that explanation to the patient. The IOM's conceptual model of diagnosis emphasizes the iterative nature of the diagnostic process, including the importance of generating a working diagnosis, gathering and incorporating new information in the reassessment of that diagnosis, and integrating treatment response into the formulation of the final diagnosis (Figure 3).

Figure 3

Even though the patient initially had several features consistent with KD, the increasing number of atypical features could have prompted the clinical team to reconsider their working diagnosis. The patient's age was atypical for KD, he had progressive knee and hip arthritis, and his fevers persisted after IVIG. An expanded differential should have included toxic shock syndrome; the resolution of conjunctival and mucosal injection and edema after IVIG may have been the result of antibodies in the IVIG preparation with neutralizing activity against superantigens. This antitoxin activity has established a role for IVIG in the management of staphylococcal toxic shock syndrome.[7] Ultimately, his imaging and surgical drainage revealed a focal staphylococcal toxin‐producing infectious source from which his fevers, rash, and mucosal and extremity changes emanated. This case reminds us that the more atypical a working diagnosis iseither in its presentation or treatment responsethe more readily clinicians should take it for another spin around the diagnostic wheel in search of a more suitable alternative.

KEY LEARNING POINTS

1. Staphylococcal toxin‐mediated disease may mimic KD, with common features including strawberry tongue, oral and conjunctival injection, and skin desquamation.
2. Improvement after treatment with IVIG is characteristic but not diagnostic of KD, and may be seen in toxin‐mediated disease.
3. KD may present with arthralgia or arthritis, but severe joint abnormalities warrant consideration of infectious and other autoimmune conditions.
4. The more atypical a working diagnosis iseither in its presentation or treatment responsethe more readily clinicians should gather, interpret, and integrate new information in search of a more suitable alternative.

Disclosure: Nothing to report.

A previously healthy 11‐year‐old boy presented to the emergency department after referral from his pediatrician for 1 week of fevers. Seven days prior to admission he developed a fever to 40.8C, vomiting, and mild left knee pain. The vomiting resolved within 2 days. Five days prior to admission he developed a pruritic, pinpoint rash over his abdomen that resolved within 24 hours. He also developed red, cracked lips, redness of his tongue, redness surrounding his eyes, and slight swelling of his hands. Three days prior to admission his pediatrician noted a 1‐cm anterior cervical lymph node. His fevers occurred throughout each of the prior 7 days without a discernible pattern, and his mild knee pain persisted at the time of presentation.

This preteen has had high fevers for 1 week associated with arthralgia, pruritic rash, emesis, and oral mucosal erythema. His rash, lip and tongue erythema, and swollen hands are classic features of Kawasaki disease (KD), but he lacks the other characteristic physical examination findings. The diagnosis of KD requires fever for at least 5 days accompanied by 4 of the following 5 signs: polymorphous rash, oral mucous membrane changes, peripheral extremity changes such as swelling or skin desquamation, bilateral bulbar conjunctival injection, and cervical lymphadenopathy >1.5 cm in diameter. Children meeting fewer than 4 of these criteria may have an incomplete form of KD.

Because most patients with KD (80%) are under 5 years old, alternative diagnoses such as autoimmune illnesses or a hypersensitivity reaction should be considered. Travel, medication, and animal exposure histories may reveal clues to an infectious or drug‐induced etiology of his fever. Immunization status should be assessed, as measles is also associated with fever, rash, and mucosal changes. Arthralgia or arthritis may occur in KD, but these findings suggest the need to entertain other possibilities, including bone or joint infection, infective endocarditis, inflammatory bowel disease, juvenile idiopathic arthritis (JIA), or systemic lupus erythematosus (SLE).

The child's only past medical history was an episode of croup as an infant. There was no family history of autoimmune diseases. He was not taking any medications and had no known allergies. His immunizations were up to date, including measles, mumps, rubella, and varicella. He lived with his parents and his dog. He swam in fresh water during a trip to Maine 2 months earlier. Neither he nor his family recalled a tick bite. He had no exposure to raw meat or unpasteurized dairy products.

The travel to New England raises the possibility of Lyme disease, although a 2‐month interval between exposure and a high, prolonged fever would be very unusual. Knee arthralgia or arthritis is common in children with late‐stage Lyme disease, but can also be seen in early‐disseminated disease. The prior description of the rash is not suggestive of erythema chronicum migrans, which is seen in early‐stage Lyme disease.

C‐reactive protein (CRP) was 189 mg/L (normal <6.3 mg/L). An echocardiogram was normal. Intravenous immunoglobulin (IVIG) was administered for presumed KD, with immediate improvement of the periorbital erythema, tongue redness, and hand swelling. He was discharged the next day on aspirin with cardiology clinic follow‐up.

Improvement after IVIG supports the diagnosis of KD. It is typical to discharge KD patients from the hospital when they have been afebrile for 24 hours or when the CRP level has declined by approximately 50%.

Over the next 48 hours he felt unwell with high‐grade fevers, continued left knee pain, and new left hip pain. He was readmitted to the hospital. His temperature was 39.4C, respiratory rate was 22 breaths per minute, heart rate was 122 beats per minute, blood pressure was 103/50 mm Hg, and oxygen saturation was 100% while breathing ambient air. He appeared mildly uncomfortable. His conjunctivae were normal. His lips were dry, red, and cracked, and his tongue was red with prominent papillae. His neck was supple without lymphadenopathy. His lungs were clear to auscultation. His heart exam was without murmurs. His abdomen was soft, and the liver and spleen were not enlarged. He had no swelling or erythema of his joints; however, he experienced pain with range of motion of his left knee, and tenderness and restricted range of motion of his left hip. His neurologic exam was normal. There were no rashes.

He has persistent fever, tachycardia, and tachypnea, now without features of KD except oral mucosal changes including prominent tongue papillae consistent with a strawberry tongue. Continued or recurrent fever may suggest persistent KD with ongoing inflammation or the need to search for an alternative diagnoses. An echocardiogram should be repeated, as the coronary artery abnormalities in KD can evolve rapidly, particularly when inflammation persists. Additional findings may include decreased left ventricular function, mitral regurgitation, or pericardial effusion. A second dose of IVIG is necessary to control fever and inflammation in about 15% of patients with KD, although in this case IVIG should be withheld pending further evaluation.

Arthralgia occurs commonly in KD, whereas frank arthritis is less typical. Polyarticular or oligoarticular arthritis involving small or large joints (especially knee or ankle) affects 5% to 10% of patients. The severity of findings in his left hip warrants consideration of septic arthritis with pain referred to the knee; pelvic or femoral osteomyelitis; psoas abscess; or pyomyositis. Following basic lab tests, imaging of the left hip region is indicated.

Laboratory evaluation revealed: white blood cell (WBC) count 10,000/L (absolute neutrophil count 8,460/L, absolute lymphocyte count 530/L), hemoglobin 10.6 g/dL, platelet count 208,000/L, serum sodium 130 mmol/L, serum potassium 3.3 mmol/L, serum urea nitrogen 11 mg/dL, serum creatinine 0.54 mg/dL, aspartate transaminase (AST) 26 U/L, alanine transaminase (ALT) 31 U/L, albumin 1.7 g/dL, erythrocyte sedimentation rate (ESR) > 100 mm/h, and CRP 263 mg/L. No blast cells were seen on peripheral blood smear.

Hypoalbuminemia and markedly elevated inflammatory markers indicate an inflammatory condition that has been active for more than a week. Assessing ESR after IVIG therapy is not useful because exogenous globulins increase the ESR; however, CRP is useful to monitor inflammation and remains elevated here.

Incomplete KD is still possible. Hyponatremia, hypoalbuminemia, and anemia are all features of persistent KD, and have been utilized in several clinical scoring systems in Japan to identify KD patients at increased risk for developing coronary complications. A neoplastic process cannot be excluded, but does not appear likely based on the acuity of his presentation and peripheral blood smear review.

Upon readmission he received a second dose of 2 g/kg IVIG. He remained on aspirin and continued to have fevers. A repeat echocardiogram was normal. He had worsening pain in his left knee and hip with difficulty straightening his left leg. Physical examination was notable for tenderness to palpation over his left hip joint, refusal to bear weight, and resistance to passive range of motion. On hospital day 2, an ultrasound of his left hip and knee revealed a complex left hip effusion and small left knee effusion.

KD becomes less likely in the presence of persistent fevers after IVIG and a repeatedly normal echocardiogram. Worsening left leg symptoms including impaired hip extension with a complex hip effusion suggests an infectious process in or adjacent to the left hip, such as septic arthritis, myositis, or osteomyelitis of the pelvis or proximal femur. A complex hip effusion is less likely to be present with arthritis related to JIA or SLE. The patient needs an emergent hip aspiration and possibly magnetic resonance imaging (MRI) to evaluate adjacent structures.

Arthrotomy and open drainage of his left hip revealed purulent fluid with a WBC count of 49,000/L with 89% neutrophils and 2% lymphocytes. Gram stain was negative. A left knee aspirate demonstrated straw‐colored synovial fluid (which was not sent for cell counts). Bacterial, fungal, and acid‐fast bacilli cultures were requested from hip and knee aspirates. Intravenous ceftriaxone and vancomycin were administered.

The most likely organism in pediatric pyogenic arthritis is Staphylococcus aureus, but there is a long list of other potential pathogens, including Streptococcus pyogenes (group A streptococcus) and Streptococcus pneumoniae. Most pediatric patients with acute pyogenic arthritis have synovial fluid WBC counts in excess of 75,000 to 100,000/L. The protracted course and the initial lack of hip symptoms raise the possibility of a primary osteomyelitis of the femur (particularly the intracapsular portion of the femoral neck or head) or of the acetabulum, with subsequent extension into the hip joint. Pyogenic myositis involving muscle groups adjacent to the hip would be unlikely to spread into the hip space, but can lead to synovial irritation, characterized by sterile joint fluid and WBC counts that fall short of the usual numbers seen in septic arthritis. The blood supply to the femoral head can become compromised with prolonged inflammation and increased intracapsular pressure, resulting in aseptic necrosis.

All cultures from his hip and knee aspirations were sterile. He continued to have daily fevers and persistent tachycardia while receiving intravenous ceftriaxone and vancomycin. Additional testing was notable for: antinuclear antibody (ANA) 1:80, anti‐streptolysin O (ASO) titer 344 IU (normal <150 IU), AST and ALT within normal limits, ferritin 568 ng/mL (normal <322 ng/mL), and lactate dehydrogenase (LDH) 212 units/L (normal <257 units/L). Abdominal ultrasound revealed borderline hepatosplenomegaly. An ophthalmologic examination was normal.

On postoperative day 4 he developed left upper thigh swelling. An MRI showed rim‐enhancing juxta‐articular complex fluid collections surrounding the left femur with decreased marrow enhancement of the left proximal femur (Figure 1).

Figure 1

The limited rheumatologic evaluation is unrevealing; the ANA result is nondiagnostic and the ASO titer is normal for age. Laboratories generally report adult normal values for streptococcal antibodies regardless of the patient's age; children from ages 7 to 12 years are at their life peak frequency of group A streptococcal pharyngitis and typically have higher normal values of streptococcal antibodies, including ASO (up to about 480640 IU). The moderately elevated ferritin level is most likely an acute phase reactant and not high enough to suggest macrophage activation syndrome, which is unlikely with the normal AST, ALT, and LDH levels, the absence of significant splenomegaly, and the lack of cytopenias. Continued fever with progressive left upper thigh swelling point to osteomyelitis of the proximal femur, which may have initially ruptured into the hip and then infiltrated the femoral cortex and spread infection into the adjacent soft tissues. Surgical debridement is indicated.

The relative prevalence of methicillin‐sensitive S aureus (MSSA) and methicillin‐resistant S aureus vary widely with geography. MSSA strains are more likely to be highly toxigenic. The elaboration of 1 or more extracellular toxins could account for the patient's initial symptoms.

The patient was brought back to the operating room for drainage of the juxta‐articular fluid collections and a biopsy of his femur. The fluid collections were grossly purulent. His intraoperative cultures were positive for MSSA. The bone biopsy revealed necrotic tissue, acute inflammation, and bacterial colonies, consistent with acute osteomyelitis. Further testing of his S aureus isolate was positive for staphylococcal enterotoxin B. He completed a 4‐week course of oral clindamycin with subsequent normalization of his hip exam and inflammatory markers. At a follow‐up visit the patient was feeling better, but had developed skin peeling on the lateral aspects of his feet consistent with late sequelae of toxin‐mediated disease (Figure 2). Three months after discharge the patient had returned to his baseline activity level and remained asymptomatic.

Figure 2

COMMENTARY

The patient presented with a constellation of symptoms that was initially mistaken for incomplete KD until focal progression of his symptoms exposed an underlying femoral osteomyelitis with periarticular abscess formation. Bacterial cultures and subsequent toxin assay revealed an enterotoxin B‐producing strain of S aureus.

Certain staphylococcal strains secrete superantigens that may lead to the development of a systemic toxin‐mediated syndrome. Toxins elaborated by S aureus include toxic shock syndrome toxin‐1 (TSST‐1) and enterotoxins, which have been implicated in menstrual and nonmenstrual toxic shock syndromes.[1, 2] Enterotoxin B is a staphylococcal superantigen found in most strains of the USA400 clonal group, and has been frequently associated with skin and soft tissue infections.[3] Enterotoxin B production has been reported in nearly half of S aureus isolates from skin, soft tissue, and bone infections.[4]

Staphylococcal and streptococcal toxin‐mediated diseases can mimic vasculitis, systemic juvenile idiopathic arthritis, viral infections, and Stevens‐Johnson syndrome. Glossitis in toxin‐mediated syndromes manifests with a swollen, red tongue with overlying enlarged papillae, giving the appearance of a strawberry. Although pediatric providers often equate strawberry tongue, conjunctival injection, rash, and erythematous lips with KD, these findings are also seen in toxin‐mediated diseases, such as scarlet fever or staphylococcal toxic shock syndrome. Enterotoxin B mediated staphylococcal disease masquerading as KD has been reported in 2 cases: a 7‐month‐old boy with multifocal S aureus osteomyelitis and a 5‐year‐old boy with S aureus bacteremia. Both staphylococcal isolates produced enterotoxin B but were negative for other staphylococcus‐related toxins including TSST‐1.[5]

The Institute of Medicine (IOM) recently released its report Improving Diagnosis in Health Care, highlighting the under‐recognized quality and safety issue of diagnostic error.[6] The report uses the following broad and patient‐centered definition of diagnostic error: the failure to (a) establish an accurate and timely explanation of the patient's health problem(s) or (b) communicate that explanation to the patient. The IOM's conceptual model of diagnosis emphasizes the iterative nature of the diagnostic process, including the importance of generating a working diagnosis, gathering and incorporating new information in the reassessment of that diagnosis, and integrating treatment response into the formulation of the final diagnosis (Figure 3).

Figure 3

Even though the patient initially had several features consistent with KD, the increasing number of atypical features could have prompted the clinical team to reconsider their working diagnosis. The patient's age was atypical for KD, he had progressive knee and hip arthritis, and his fevers persisted after IVIG. An expanded differential should have included toxic shock syndrome; the resolution of conjunctival and mucosal injection and edema after IVIG may have been the result of antibodies in the IVIG preparation with neutralizing activity against superantigens. This antitoxin activity has established a role for IVIG in the management of staphylococcal toxic shock syndrome.[7] Ultimately, his imaging and surgical drainage revealed a focal staphylococcal toxin‐producing infectious source from which his fevers, rash, and mucosal and extremity changes emanated. This case reminds us that the more atypical a working diagnosis iseither in its presentation or treatment responsethe more readily clinicians should take it for another spin around the diagnostic wheel in search of a more suitable alternative.

KEY LEARNING POINTS

1. Staphylococcal toxin‐mediated disease may mimic KD, with common features including strawberry tongue, oral and conjunctival injection, and skin desquamation.
2. Improvement after treatment with IVIG is characteristic but not diagnostic of KD, and may be seen in toxin‐mediated disease.
3. KD may present with arthralgia or arthritis, but severe joint abnormalities warrant consideration of infectious and other autoimmune conditions.
4. The more atypical a working diagnosis iseither in its presentation or treatment responsethe more readily clinicians should gather, interpret, and integrate new information in search of a more suitable alternative.

Disclosure: Nothing to report.

A 47‐year‐old paraplegic man presented to our emergency department (ED) with severe pain from his sacral decubitus ulcers. Earlier that day, he had left against medical advice from another area ED. He took no medications but noted that he had been on several medications during recent hospitalizations. He had smoked cocaine that morning. He reported 3 days of watery diarrhea. His triage temperature was 98.2F, his heart rate was 146 beats per minute, and his blood pressure was 155/84 mm Hg. He had multiple foul‐smelling, stage 4 sacral ulcers and a chronic indwelling urinary catheter. Laboratory results included: white blood cell (WBC) count 22,000/mm³ (neutrophils 17,900/mm³), hemoglobin 8.5 g/dL, platelet count 317,000/mm³, creatinine 4.2 mg/dL, and lactate 1.5 mg/dL.

The patient's tachycardia and leukocytosis suggest sepsis. Potential sources include soft tissue infection or osteomyelitis from his sacral ulcers, Clostridium difficile, or a urinary tract infection. Impaired visceral sensation from his spinal cord injury may dampen his response to an intra‐abdominal process, such as mesenteric ischemia or toxic megacolon. Records from other hospitals should be reviewed to assess the acuity of change in his WBC count, hemoglobin, and creatinine. His anemia may be from chronic inflammation (eg, osteomyelitis), renal insufficiency, hemolysis, or occult blood loss, including retroperitoneal and gastrointestinal sources. His kidney injury may be from tubular necrosis in the setting of sepsis or obstructive uropathy related to a neurogenic bladder.

Intravenous vancomycin and piperacillin‐tazobactam were initiated for presumed sepsis. The team requested records from his prior hospitalizations and received only his most recent ED documentation, which revealed that he had received single doses of vancomycin and piperacillin‐tazobactam for a similar presentation. Oral vancomycin and intravenous metronidazole were started when C difficile toxin testing returned positive. He appeared more comfortable and remained normotensive, although his sinus tachycardia persisted at approximately 130 beats per minute. Sixteen hours after admission, the patient went into pulseless electrical activity (PEA) arrest.

Potential contributors to his PEA and cardiovascular collapse are drug use (cocaine), alcohol withdrawal, infection, hypovolemia, myocardial ischemia, or heart failure. Severe hemorrhage, hyperkalemia, or acidosis from acute kidney injury and sepsis could also account for his cardiac arrest. His paraplegia and hospitalization raise the risk of venous thromboembolism, which can lead to PEA from pulmonary embolus and prolonged hypoxia.

Return of spontaneous circulation occurred following 6 minutes of cardiopulmonary resuscitation. Arterial blood gas during bag‐valve mask ventilation with 100% FiO₂ was: pH 7.00, paCO₂ 44 mm Hg, and paO₂ 209 mm Hg. Serum chemistries were: sodium 148 mmol/L, potassium 5.5 mmol/L, chloride 111 mmol/L, bicarbonate 11 mmol/L, blood urea nitrogen 78 mg/dL, creatinine 4.1 mg/dL, glucose 256 mg/dL, calcium 8.4 mg/dL, magnesium 1.7 mg/dL, and phosphorus 5.9 mg/dL. Venous lactate was >11 mmol/L. His WBC count was 28,000/mm³, hematocrit was 4.4%, and platelet count was 196,000/mm³. Hemoglobin was unmeasurable due to interfering substances. An electrocardiogram showed sinus tachycardia with diffuse ST segment depressions. Chest radiograph revealed a properly situated endotracheal tube, normal cardiomediastinal silhouette and right basilar atelectasis.

His profound anemia is the likely cause of his PEA arrest and severe lactic acidosis. Massive hemolysis is most likely given no overt evidence of bleeding to account for the precipitous fall in hematocrit. Hemolysis can result from disorders intrinsic or extrinsic to the red blood cell (RBC). Intrinsic defects are usually congenital and involve the membrane, hemoglobin, or metabolic enzymes within the RBC. Extrinsic hemolysis arises from processes that injure the RBC from the outside: antibodies, infections, and mechanical shearing.

A rapidly declining platelet count is seen in microangiopathic hemolytic conditions such as disseminated intravascular coagulation (DIC) or thrombotic thrombocytopenic purpura (TTP), where platelets are consumed along with RBCs; sepsis makes DIC more likely. Autoimmune hemolytic anemia (AIHA) is sometimes accompanied by immune thrombocytopenia. AIHA arises from antibodies that are idiopathic or produced in response to infection, autoimmune conditions (eg, systemic lupus erythematosus), lymphoproliferative disease, or drugs (eg, ‐lactam antibiotics). The antiphospholipid syndrome can lead to thrombocytopenia, hemolysis, and kidney injury. Devitalized tissue in his sacral ulcers may predispose the patient to infection with Clostridium perfringens, which can elaborate enzymes that trigger massive hemolysis.

Because automated hemoglobin measurement is performed by spectrophotometry (light absorption and scatter), high concentrations of poorly soluble autoantibodies can increase the turbidity of the sample and preclude the measurement of hemoglobin concentration. This could lead to the report of interfering substances.

Computed tomography of the chest, abdomen, and pelvis did not reveal internal bleeding. Laboratory studies during the resuscitation were: lactate dehydrogenase (LDH) 2,411 U/L (reference range: 88230 U/L), haptoglobin 25 mg/dL (reference range: 36195 mg/dL), and reticulocyte percentage 2.6%. Total bilirubin was 2.4 mg/dL (reference range: 0.11.2 mg/dL), with direct fraction 0.9 mg/dL (reference range: <0.3 mg/dL). Prothrombin time (PT) was 24.9 seconds (reference range: 11.815.2 seconds), international normalized ratio (INR) was 2.3, and activated partial thromboplastin time (aPTT) was 47.3 seconds (reference range: 22.233.0 seconds). Phlebotomy yielded icteric plasma. A peripheral blood smear revealed severe anemia, red cell agglutination, microspherocytes, and hemophagocytosis by neutrophils (Figure 1).

Figure 1

Low haptoglobin, elevated LDH, and hyperbilirubinemia confirm hemolysis. A more robust reticulocytosis is expected in the face of profound anemia, but the patient may also suffer from a concomitant hypoproliferative state (eg, nutritional deficiency). More likely, the rapidity of his decline outpaced the marrow's response, which can be delayed by days.

The most common cause of a combined elevation of the INR/PT and aPTT in a critically ill patient is DIC. Although no schistocytes were detected on the peripheral smear, they can be absent in up to 50% of DIC cases. TTP is associated with hemolytic anemia, kidney injury, and thrombocytopenia, but it generally does not cause coagulopathy.

The combination of red cell agglutination and hemophagocytosis suggests that the RBCs are coated with autoantibodies that cross‐link the cells and make them targets for phagocytosis by neutrophils in the circulation. This is distinct from the hemophagocytic syndrome, a rare immune activation syndrome characterized by macrophage phagocytosis of RBCs in the reticuloendothelial system. The blood smear also shows microspherocytes, which are seen in AIHA and hereditary spherocytosis.

Urinalysis showed large hemoglobin and 11 to 20 red blood cells per high‐power field. Serum creatine phosphokinase was 35 U/dL (reference range: 50388 U/dL). Four units of packed RBCs and 1 unit of fresh‐frozen plasma were transfused. Due to a rising creatinine level and declining urine output, continuous veno‐venous hemofiltration (CVVH) was initiated.

Acute tubular necrosis could result from sepsis, ischemic injury from DIC, hypotension during cardiac arrest, or heme pigment toxicity. Urine sediment should be reviewed for dysmorphic RBCs or RBC casts that would indicate glomerulonephritis (eg, from an underlying autoimmune process associated with AIHA).

Urine hemoglobin that is disproportionate to the degree of hematuria suggests hemoglobinuria, which in turn defines the hemolysis as intravascular. Processes that directly lyse RBCs in circulation via mechanical shearing, activation of complement, infection of the RBC, or enzymatic or oxidative destruction of the membrane cause intravascular hemolysis. Leading considerations include microangiopathy (eg, DIC, TTP), clostridial sepsis, and AIHA.

AIHA can be broadly classified as warm or cold. Warm AIHA is caused by immunoglobulin IgG antibodies that bind most avidly at body temperature. Because warm AIHA does not activate complement, patients present with evidence of extravascular hemolysis that is typically chronic and mild to moderate in severity. It does not typically cause the acute, fulminant, intravascular hemolytic condition seen here.

Cold AIHA is characterized by autoantibodies that bind at lower temperatures and comes in 2 forms: cold agglutinin disease and (rarely) paroxysmal cold hemoglobinuria (PCH). Cold agglutinins are most often IgM antibodies produced in response to infection (Mycoplasma pneumoniae, infectious mononucleosis), drugs, or a hematologic malignancy. These IgM antibodies bind RBCs, causing them to agglutinate, and fix complement (including C3) to the surface of RBCs when blood circulates to cooler parts of the body. This results in complement activation, formation of the membrane attack complex, and intravascular hemolysis when bound and activated complement is present in large numbers. Acute infection can increase the complement available for binding to the surface of RBCs. Through a slightly different mechanism, PCH causes intravascular hemolysis through direct IgG activation of complement fixed to the surface of RBCs. During a hemolytic episode the direct antibody test (DAT) is positive using anti‐C3 and negative for IgG.

Based on the patient's clinical evidence of intravascular hemolysis and a suspected autoimmune etiology, the leading diagnosis at this time is cold AIHA.

Due to coagulopathy and possible cold agglutinin disease, therapeutic hypothermia for neuroprotection was deferred. He continued on vancomycin, piperacillin‐tazobactam, and metronidazole. The DAT and direct IgG were strongly positive (3+), whereas the direct C3 was weakly positive (<1+). His serum free hemoglobin increased from 136.7 mg/dL to 223.8 mg/dL (normal: 0.06.9 mg/dL). His severe metabolic acidosis corrected with CVVH.

The DAT detects IgG or complement adherent to RBCs. This patient has tested positive for both IgG and C3, though much more strongly for IgG, suggesting an unusual ability of the patient's IgG to activate complement. The phenomenon of mixed AIHA, in which the patient has both warm‐ and cold‐reacting antibodies, is rare.

Regarding infections associated with AIHA, there is no cough or rash to suggest M pneumoniae, and there is no sore throat, fever, lymphadenopathy, splenomegaly, or atypical lymphocytosis to suggest infectious mononucleosis. He should be tested for human immunodeficiency virus, which is also associated with AIHA. His leukocytosis may raise suspicion for an underlying hematologic malignancy, but he does not have blasts, dysplastic leukocytes, or lymphocytosis on his peripheral blood smear. Systemic lupus erythematosus can be associated with AIHA, thrombocytopenia, and renal failure, but he lacks the more common clinical manifestations of rash, arthralgias, and fever.

Drug‐induced immune hemolytic anemia (DIIHA) can cause both the clinical and serologic profile of an AIHA, as seen here. DIIHA can be distinguished from mixed AIHA if hemolysis abates with discontinuation of an offending drug. His deterioration is temporally associated with drug administration at the time of admission. Cephalosporins and ‐lactams (e.g., piperacillin) are the most common causes of DIIHA, and ‐lactamases such as tazobactam have also been implicated. By exclusion of other causes, DIIHA secondary to piperacillin is most likely responsible for his massive intravascular hemolysis.

Medical records from 2 other hospitals arrived approximately 36 hours after the patient's initial presentation. Three weeks earlier, he had required RBC transfusions for a 6 g/dL hemoglobin reduction after receiving piperacillin‐tazobactam for sepsis. He had been treated with steroids for presumed warm AIHA and dialyzed for acute kidney injury attributed to hemoglobinuria. After 1 week in the intensive care unit, he left against medical advice. Two weeks later he presented to another ED, which immediately preceded his presentation to our hospital.

An antibody screen for drug‐dependent antibodies revealed antipiperacillin antibodies. Antitazobactam antibodies were not tested. Piperacillin‐tazobactam was discontinued, and plasmapheresis was initiated to decrease the amount of piperacillin in the blood. The patient's hemoglobin subsequently remained above 7.0 g/dL without RBC transfusions.

His renal function recovered and he completed antibiotic therapy for C difficile infection and for his pressure ulcers. However, he had sustained severe anoxic brain injury during his cardiac arrest and did not recover neurological function. In accordance with his family's wishes, he was discharged to a long‐term acute care hospital dependent on a ventilator.

COMMENTARY

This case illustrates a dramatic presentation of fulminant intravascular hemolysis secondary to piperacillin. The incidence of DIIHA is estimated to be 1 in 1 million.[1] Historically, methyldopa and high‐dose penicillin have been responsible for the majority of cases,[2] but in recent years complex penicillins, including piperacillin, and second‐ and third‐generation cephalosporins have been implicated.[3, 4] Cases of DIIHA are often underdiagnosed or misdiagnosed, as smoldering or less severe cases may not be recognized or are attributed to other causes.

A positive DAT, suggesting immunoglobulin and/or complement binding to RBCs, is the most reliable laboratory finding in DIIHA.[5] However, a positive DAT does not identify the source of the antigen and may result in misattribution of the immune hemolysis to autoimmunity rather than to a drug. Repeated or continued administration of the offending drug (as in this case) may perpetuate or worsen the hemolysis. Drug‐specific antibody tests may help to confirm the diagnosis, but these tests are complex and take significant time for specialized laboratories to run.

Severe hemolysis should be considered when a patient has a sudden and dramatic drop in his hemoglobin level in the absence of bleeding. Because DIIHA can be rapidly progressive, discontinuing a suspected culprit drug is the most important diagnostic and therapeutic measure. Typically, when an offending drug is stopped, the hemolysis stops as well. The time course over which this occurs depends on the rapidity of drug clearance.[4] Hemodialysis or plasmapheresis may be required in cases where the medication is renally excreted, particularly in cases of concomitant kidney injury. Evidence supporting corticosteroid use in DIIHA is limited, as the offending agent is usually discontinued by the time corticosteroids are initiated.[4]

This patient's DAT confirmed both IgG and complement activation, consistent with DIIHA caused by an immune complexlike reaction. This mechanism involves the antibody binding to a mixed epitope of the drug and a RBC membrane glycoprotein.[6] The offending drug was stopped only when review of his medical records established a clear temporal association between antibiotic administration and prior hemolysis.

The 2009 Health Information Technology for Economic and Clinical Health Act created an electronic health record (EHR) incentive program (meaningful use criteria).[7] By 2012, only 6% of hospitals met all of the stage 2 criteria, which include EHR interoperability across health systems.[8] The patient's preceding hemolytic event was described in records faxed by the outside hospitals, but without EHR interoperability, the treating clinicians did not have timely access to this information. Instead, the familiar manual process of obtaining outside records involving signed forms, phone calls, fax machines, and reams of paper progressed at its usual pace. Real‐time access to health records might have guided providers to select an alternative antibiotic regimen. Instead, a communication breakdown contributed to a catastrophic drug reaction and to this tragic patient outcome.

KEY TEACHING POINTS

1. In a patient presenting with acute hemolysis and a positive DAT, consider DIIHA.
2. Both piperacillin and tazobactam can cause a severe, complement‐mediated immune hemolytic anemia (DIIHA).
3. Drug‐induced antibodies are detected by direct antiglobulin testing, but a complete medication history is the key to diagnosis.
4. Management of drug‐induced hemolytic anemia involves immediate discontinuation of the culprit medication, supportive care, and potentially corticosteroids, plasmapheresis, and/or hemodialysis to expedite removal of the offending agent.
5. EHR interoperability may provide timely access to important health information across different hospitals, expedite health information exchange, and reduce adverse patient outcomes that stem from communication delays.

This case was submitted anonymously to AHRQ WebM&M on July 18, 2014, and was accepted on August 7, 2014. The case and WebM&M commentary were published online on October 26, 2015.[9] This separate commentary on the same case was later submitted to the Journal of Hospital Medicine on September 2, 2015, accepted on November 24, 2015, and published on January 22, 2016. The 2 publications are written by different authors, and although they reference the same case, they make different but valuable points.

A 47‐year‐old paraplegic man presented to our emergency department (ED) with severe pain from his sacral decubitus ulcers. Earlier that day, he had left against medical advice from another area ED. He took no medications but noted that he had been on several medications during recent hospitalizations. He had smoked cocaine that morning. He reported 3 days of watery diarrhea. His triage temperature was 98.2F, his heart rate was 146 beats per minute, and his blood pressure was 155/84 mm Hg. He had multiple foul‐smelling, stage 4 sacral ulcers and a chronic indwelling urinary catheter. Laboratory results included: white blood cell (WBC) count 22,000/mm³ (neutrophils 17,900/mm³), hemoglobin 8.5 g/dL, platelet count 317,000/mm³, creatinine 4.2 mg/dL, and lactate 1.5 mg/dL.

The patient's tachycardia and leukocytosis suggest sepsis. Potential sources include soft tissue infection or osteomyelitis from his sacral ulcers, Clostridium difficile, or a urinary tract infection. Impaired visceral sensation from his spinal cord injury may dampen his response to an intra‐abdominal process, such as mesenteric ischemia or toxic megacolon. Records from other hospitals should be reviewed to assess the acuity of change in his WBC count, hemoglobin, and creatinine. His anemia may be from chronic inflammation (eg, osteomyelitis), renal insufficiency, hemolysis, or occult blood loss, including retroperitoneal and gastrointestinal sources. His kidney injury may be from tubular necrosis in the setting of sepsis or obstructive uropathy related to a neurogenic bladder.

Intravenous vancomycin and piperacillin‐tazobactam were initiated for presumed sepsis. The team requested records from his prior hospitalizations and received only his most recent ED documentation, which revealed that he had received single doses of vancomycin and piperacillin‐tazobactam for a similar presentation. Oral vancomycin and intravenous metronidazole were started when C difficile toxin testing returned positive. He appeared more comfortable and remained normotensive, although his sinus tachycardia persisted at approximately 130 beats per minute. Sixteen hours after admission, the patient went into pulseless electrical activity (PEA) arrest.

Potential contributors to his PEA and cardiovascular collapse are drug use (cocaine), alcohol withdrawal, infection, hypovolemia, myocardial ischemia, or heart failure. Severe hemorrhage, hyperkalemia, or acidosis from acute kidney injury and sepsis could also account for his cardiac arrest. His paraplegia and hospitalization raise the risk of venous thromboembolism, which can lead to PEA from pulmonary embolus and prolonged hypoxia.

Return of spontaneous circulation occurred following 6 minutes of cardiopulmonary resuscitation. Arterial blood gas during bag‐valve mask ventilation with 100% FiO₂ was: pH 7.00, paCO₂ 44 mm Hg, and paO₂ 209 mm Hg. Serum chemistries were: sodium 148 mmol/L, potassium 5.5 mmol/L, chloride 111 mmol/L, bicarbonate 11 mmol/L, blood urea nitrogen 78 mg/dL, creatinine 4.1 mg/dL, glucose 256 mg/dL, calcium 8.4 mg/dL, magnesium 1.7 mg/dL, and phosphorus 5.9 mg/dL. Venous lactate was >11 mmol/L. His WBC count was 28,000/mm³, hematocrit was 4.4%, and platelet count was 196,000/mm³. Hemoglobin was unmeasurable due to interfering substances. An electrocardiogram showed sinus tachycardia with diffuse ST segment depressions. Chest radiograph revealed a properly situated endotracheal tube, normal cardiomediastinal silhouette and right basilar atelectasis.

His profound anemia is the likely cause of his PEA arrest and severe lactic acidosis. Massive hemolysis is most likely given no overt evidence of bleeding to account for the precipitous fall in hematocrit. Hemolysis can result from disorders intrinsic or extrinsic to the red blood cell (RBC). Intrinsic defects are usually congenital and involve the membrane, hemoglobin, or metabolic enzymes within the RBC. Extrinsic hemolysis arises from processes that injure the RBC from the outside: antibodies, infections, and mechanical shearing.

A rapidly declining platelet count is seen in microangiopathic hemolytic conditions such as disseminated intravascular coagulation (DIC) or thrombotic thrombocytopenic purpura (TTP), where platelets are consumed along with RBCs; sepsis makes DIC more likely. Autoimmune hemolytic anemia (AIHA) is sometimes accompanied by immune thrombocytopenia. AIHA arises from antibodies that are idiopathic or produced in response to infection, autoimmune conditions (eg, systemic lupus erythematosus), lymphoproliferative disease, or drugs (eg, ‐lactam antibiotics). The antiphospholipid syndrome can lead to thrombocytopenia, hemolysis, and kidney injury. Devitalized tissue in his sacral ulcers may predispose the patient to infection with Clostridium perfringens, which can elaborate enzymes that trigger massive hemolysis.

Because automated hemoglobin measurement is performed by spectrophotometry (light absorption and scatter), high concentrations of poorly soluble autoantibodies can increase the turbidity of the sample and preclude the measurement of hemoglobin concentration. This could lead to the report of interfering substances.

Computed tomography of the chest, abdomen, and pelvis did not reveal internal bleeding. Laboratory studies during the resuscitation were: lactate dehydrogenase (LDH) 2,411 U/L (reference range: 88230 U/L), haptoglobin 25 mg/dL (reference range: 36195 mg/dL), and reticulocyte percentage 2.6%. Total bilirubin was 2.4 mg/dL (reference range: 0.11.2 mg/dL), with direct fraction 0.9 mg/dL (reference range: <0.3 mg/dL). Prothrombin time (PT) was 24.9 seconds (reference range: 11.815.2 seconds), international normalized ratio (INR) was 2.3, and activated partial thromboplastin time (aPTT) was 47.3 seconds (reference range: 22.233.0 seconds). Phlebotomy yielded icteric plasma. A peripheral blood smear revealed severe anemia, red cell agglutination, microspherocytes, and hemophagocytosis by neutrophils (Figure 1).

Figure 1

Low haptoglobin, elevated LDH, and hyperbilirubinemia confirm hemolysis. A more robust reticulocytosis is expected in the face of profound anemia, but the patient may also suffer from a concomitant hypoproliferative state (eg, nutritional deficiency). More likely, the rapidity of his decline outpaced the marrow's response, which can be delayed by days.

The most common cause of a combined elevation of the INR/PT and aPTT in a critically ill patient is DIC. Although no schistocytes were detected on the peripheral smear, they can be absent in up to 50% of DIC cases. TTP is associated with hemolytic anemia, kidney injury, and thrombocytopenia, but it generally does not cause coagulopathy.

The combination of red cell agglutination and hemophagocytosis suggests that the RBCs are coated with autoantibodies that cross‐link the cells and make them targets for phagocytosis by neutrophils in the circulation. This is distinct from the hemophagocytic syndrome, a rare immune activation syndrome characterized by macrophage phagocytosis of RBCs in the reticuloendothelial system. The blood smear also shows microspherocytes, which are seen in AIHA and hereditary spherocytosis.

Urinalysis showed large hemoglobin and 11 to 20 red blood cells per high‐power field. Serum creatine phosphokinase was 35 U/dL (reference range: 50388 U/dL). Four units of packed RBCs and 1 unit of fresh‐frozen plasma were transfused. Due to a rising creatinine level and declining urine output, continuous veno‐venous hemofiltration (CVVH) was initiated.

Acute tubular necrosis could result from sepsis, ischemic injury from DIC, hypotension during cardiac arrest, or heme pigment toxicity. Urine sediment should be reviewed for dysmorphic RBCs or RBC casts that would indicate glomerulonephritis (eg, from an underlying autoimmune process associated with AIHA).

Urine hemoglobin that is disproportionate to the degree of hematuria suggests hemoglobinuria, which in turn defines the hemolysis as intravascular. Processes that directly lyse RBCs in circulation via mechanical shearing, activation of complement, infection of the RBC, or enzymatic or oxidative destruction of the membrane cause intravascular hemolysis. Leading considerations include microangiopathy (eg, DIC, TTP), clostridial sepsis, and AIHA.

AIHA can be broadly classified as warm or cold. Warm AIHA is caused by immunoglobulin IgG antibodies that bind most avidly at body temperature. Because warm AIHA does not activate complement, patients present with evidence of extravascular hemolysis that is typically chronic and mild to moderate in severity. It does not typically cause the acute, fulminant, intravascular hemolytic condition seen here.

Cold AIHA is characterized by autoantibodies that bind at lower temperatures and comes in 2 forms: cold agglutinin disease and (rarely) paroxysmal cold hemoglobinuria (PCH). Cold agglutinins are most often IgM antibodies produced in response to infection (Mycoplasma pneumoniae, infectious mononucleosis), drugs, or a hematologic malignancy. These IgM antibodies bind RBCs, causing them to agglutinate, and fix complement (including C3) to the surface of RBCs when blood circulates to cooler parts of the body. This results in complement activation, formation of the membrane attack complex, and intravascular hemolysis when bound and activated complement is present in large numbers. Acute infection can increase the complement available for binding to the surface of RBCs. Through a slightly different mechanism, PCH causes intravascular hemolysis through direct IgG activation of complement fixed to the surface of RBCs. During a hemolytic episode the direct antibody test (DAT) is positive using anti‐C3 and negative for IgG.

Based on the patient's clinical evidence of intravascular hemolysis and a suspected autoimmune etiology, the leading diagnosis at this time is cold AIHA.

Due to coagulopathy and possible cold agglutinin disease, therapeutic hypothermia for neuroprotection was deferred. He continued on vancomycin, piperacillin‐tazobactam, and metronidazole. The DAT and direct IgG were strongly positive (3+), whereas the direct C3 was weakly positive (<1+). His serum free hemoglobin increased from 136.7 mg/dL to 223.8 mg/dL (normal: 0.06.9 mg/dL). His severe metabolic acidosis corrected with CVVH.

The DAT detects IgG or complement adherent to RBCs. This patient has tested positive for both IgG and C3, though much more strongly for IgG, suggesting an unusual ability of the patient's IgG to activate complement. The phenomenon of mixed AIHA, in which the patient has both warm‐ and cold‐reacting antibodies, is rare.

Regarding infections associated with AIHA, there is no cough or rash to suggest M pneumoniae, and there is no sore throat, fever, lymphadenopathy, splenomegaly, or atypical lymphocytosis to suggest infectious mononucleosis. He should be tested for human immunodeficiency virus, which is also associated with AIHA. His leukocytosis may raise suspicion for an underlying hematologic malignancy, but he does not have blasts, dysplastic leukocytes, or lymphocytosis on his peripheral blood smear. Systemic lupus erythematosus can be associated with AIHA, thrombocytopenia, and renal failure, but he lacks the more common clinical manifestations of rash, arthralgias, and fever.

Drug‐induced immune hemolytic anemia (DIIHA) can cause both the clinical and serologic profile of an AIHA, as seen here. DIIHA can be distinguished from mixed AIHA if hemolysis abates with discontinuation of an offending drug. His deterioration is temporally associated with drug administration at the time of admission. Cephalosporins and ‐lactams (e.g., piperacillin) are the most common causes of DIIHA, and ‐lactamases such as tazobactam have also been implicated. By exclusion of other causes, DIIHA secondary to piperacillin is most likely responsible for his massive intravascular hemolysis.

Medical records from 2 other hospitals arrived approximately 36 hours after the patient's initial presentation. Three weeks earlier, he had required RBC transfusions for a 6 g/dL hemoglobin reduction after receiving piperacillin‐tazobactam for sepsis. He had been treated with steroids for presumed warm AIHA and dialyzed for acute kidney injury attributed to hemoglobinuria. After 1 week in the intensive care unit, he left against medical advice. Two weeks later he presented to another ED, which immediately preceded his presentation to our hospital.

An antibody screen for drug‐dependent antibodies revealed antipiperacillin antibodies. Antitazobactam antibodies were not tested. Piperacillin‐tazobactam was discontinued, and plasmapheresis was initiated to decrease the amount of piperacillin in the blood. The patient's hemoglobin subsequently remained above 7.0 g/dL without RBC transfusions.

His renal function recovered and he completed antibiotic therapy for C difficile infection and for his pressure ulcers. However, he had sustained severe anoxic brain injury during his cardiac arrest and did not recover neurological function. In accordance with his family's wishes, he was discharged to a long‐term acute care hospital dependent on a ventilator.

COMMENTARY

This case illustrates a dramatic presentation of fulminant intravascular hemolysis secondary to piperacillin. The incidence of DIIHA is estimated to be 1 in 1 million.[1] Historically, methyldopa and high‐dose penicillin have been responsible for the majority of cases,[2] but in recent years complex penicillins, including piperacillin, and second‐ and third‐generation cephalosporins have been implicated.[3, 4] Cases of DIIHA are often underdiagnosed or misdiagnosed, as smoldering or less severe cases may not be recognized or are attributed to other causes.

A positive DAT, suggesting immunoglobulin and/or complement binding to RBCs, is the most reliable laboratory finding in DIIHA.[5] However, a positive DAT does not identify the source of the antigen and may result in misattribution of the immune hemolysis to autoimmunity rather than to a drug. Repeated or continued administration of the offending drug (as in this case) may perpetuate or worsen the hemolysis. Drug‐specific antibody tests may help to confirm the diagnosis, but these tests are complex and take significant time for specialized laboratories to run.

Severe hemolysis should be considered when a patient has a sudden and dramatic drop in his hemoglobin level in the absence of bleeding. Because DIIHA can be rapidly progressive, discontinuing a suspected culprit drug is the most important diagnostic and therapeutic measure. Typically, when an offending drug is stopped, the hemolysis stops as well. The time course over which this occurs depends on the rapidity of drug clearance.[4] Hemodialysis or plasmapheresis may be required in cases where the medication is renally excreted, particularly in cases of concomitant kidney injury. Evidence supporting corticosteroid use in DIIHA is limited, as the offending agent is usually discontinued by the time corticosteroids are initiated.[4]

This patient's DAT confirmed both IgG and complement activation, consistent with DIIHA caused by an immune complexlike reaction. This mechanism involves the antibody binding to a mixed epitope of the drug and a RBC membrane glycoprotein.[6] The offending drug was stopped only when review of his medical records established a clear temporal association between antibiotic administration and prior hemolysis.

The 2009 Health Information Technology for Economic and Clinical Health Act created an electronic health record (EHR) incentive program (meaningful use criteria).[7] By 2012, only 6% of hospitals met all of the stage 2 criteria, which include EHR interoperability across health systems.[8] The patient's preceding hemolytic event was described in records faxed by the outside hospitals, but without EHR interoperability, the treating clinicians did not have timely access to this information. Instead, the familiar manual process of obtaining outside records involving signed forms, phone calls, fax machines, and reams of paper progressed at its usual pace. Real‐time access to health records might have guided providers to select an alternative antibiotic regimen. Instead, a communication breakdown contributed to a catastrophic drug reaction and to this tragic patient outcome.

KEY TEACHING POINTS

1. In a patient presenting with acute hemolysis and a positive DAT, consider DIIHA.
2. Both piperacillin and tazobactam can cause a severe, complement‐mediated immune hemolytic anemia (DIIHA).
3. Drug‐induced antibodies are detected by direct antiglobulin testing, but a complete medication history is the key to diagnosis.
4. Management of drug‐induced hemolytic anemia involves immediate discontinuation of the culprit medication, supportive care, and potentially corticosteroids, plasmapheresis, and/or hemodialysis to expedite removal of the offending agent.
5. EHR interoperability may provide timely access to important health information across different hospitals, expedite health information exchange, and reduce adverse patient outcomes that stem from communication delays.

This case was submitted anonymously to AHRQ WebM&M on July 18, 2014, and was accepted on August 7, 2014. The case and WebM&M commentary were published online on October 26, 2015.[9] This separate commentary on the same case was later submitted to the Journal of Hospital Medicine on September 2, 2015, accepted on November 24, 2015, and published on January 22, 2016. The 2 publications are written by different authors, and although they reference the same case, they make different but valuable points.

A 47‐year‐old paraplegic man presented to our emergency department (ED) with severe pain from his sacral decubitus ulcers. Earlier that day, he had left against medical advice from another area ED. He took no medications but noted that he had been on several medications during recent hospitalizations. He had smoked cocaine that morning. He reported 3 days of watery diarrhea. His triage temperature was 98.2F, his heart rate was 146 beats per minute, and his blood pressure was 155/84 mm Hg. He had multiple foul‐smelling, stage 4 sacral ulcers and a chronic indwelling urinary catheter. Laboratory results included: white blood cell (WBC) count 22,000/mm³ (neutrophils 17,900/mm³), hemoglobin 8.5 g/dL, platelet count 317,000/mm³, creatinine 4.2 mg/dL, and lactate 1.5 mg/dL.

The patient's tachycardia and leukocytosis suggest sepsis. Potential sources include soft tissue infection or osteomyelitis from his sacral ulcers, Clostridium difficile, or a urinary tract infection. Impaired visceral sensation from his spinal cord injury may dampen his response to an intra‐abdominal process, such as mesenteric ischemia or toxic megacolon. Records from other hospitals should be reviewed to assess the acuity of change in his WBC count, hemoglobin, and creatinine. His anemia may be from chronic inflammation (eg, osteomyelitis), renal insufficiency, hemolysis, or occult blood loss, including retroperitoneal and gastrointestinal sources. His kidney injury may be from tubular necrosis in the setting of sepsis or obstructive uropathy related to a neurogenic bladder.

Intravenous vancomycin and piperacillin‐tazobactam were initiated for presumed sepsis. The team requested records from his prior hospitalizations and received only his most recent ED documentation, which revealed that he had received single doses of vancomycin and piperacillin‐tazobactam for a similar presentation. Oral vancomycin and intravenous metronidazole were started when C difficile toxin testing returned positive. He appeared more comfortable and remained normotensive, although his sinus tachycardia persisted at approximately 130 beats per minute. Sixteen hours after admission, the patient went into pulseless electrical activity (PEA) arrest.

Potential contributors to his PEA and cardiovascular collapse are drug use (cocaine), alcohol withdrawal, infection, hypovolemia, myocardial ischemia, or heart failure. Severe hemorrhage, hyperkalemia, or acidosis from acute kidney injury and sepsis could also account for his cardiac arrest. His paraplegia and hospitalization raise the risk of venous thromboembolism, which can lead to PEA from pulmonary embolus and prolonged hypoxia.

Return of spontaneous circulation occurred following 6 minutes of cardiopulmonary resuscitation. Arterial blood gas during bag‐valve mask ventilation with 100% FiO₂ was: pH 7.00, paCO₂ 44 mm Hg, and paO₂ 209 mm Hg. Serum chemistries were: sodium 148 mmol/L, potassium 5.5 mmol/L, chloride 111 mmol/L, bicarbonate 11 mmol/L, blood urea nitrogen 78 mg/dL, creatinine 4.1 mg/dL, glucose 256 mg/dL, calcium 8.4 mg/dL, magnesium 1.7 mg/dL, and phosphorus 5.9 mg/dL. Venous lactate was >11 mmol/L. His WBC count was 28,000/mm³, hematocrit was 4.4%, and platelet count was 196,000/mm³. Hemoglobin was unmeasurable due to interfering substances. An electrocardiogram showed sinus tachycardia with diffuse ST segment depressions. Chest radiograph revealed a properly situated endotracheal tube, normal cardiomediastinal silhouette and right basilar atelectasis.

His profound anemia is the likely cause of his PEA arrest and severe lactic acidosis. Massive hemolysis is most likely given no overt evidence of bleeding to account for the precipitous fall in hematocrit. Hemolysis can result from disorders intrinsic or extrinsic to the red blood cell (RBC). Intrinsic defects are usually congenital and involve the membrane, hemoglobin, or metabolic enzymes within the RBC. Extrinsic hemolysis arises from processes that injure the RBC from the outside: antibodies, infections, and mechanical shearing.

A rapidly declining platelet count is seen in microangiopathic hemolytic conditions such as disseminated intravascular coagulation (DIC) or thrombotic thrombocytopenic purpura (TTP), where platelets are consumed along with RBCs; sepsis makes DIC more likely. Autoimmune hemolytic anemia (AIHA) is sometimes accompanied by immune thrombocytopenia. AIHA arises from antibodies that are idiopathic or produced in response to infection, autoimmune conditions (eg, systemic lupus erythematosus), lymphoproliferative disease, or drugs (eg, ‐lactam antibiotics). The antiphospholipid syndrome can lead to thrombocytopenia, hemolysis, and kidney injury. Devitalized tissue in his sacral ulcers may predispose the patient to infection with Clostridium perfringens, which can elaborate enzymes that trigger massive hemolysis.

Because automated hemoglobin measurement is performed by spectrophotometry (light absorption and scatter), high concentrations of poorly soluble autoantibodies can increase the turbidity of the sample and preclude the measurement of hemoglobin concentration. This could lead to the report of interfering substances.

Computed tomography of the chest, abdomen, and pelvis did not reveal internal bleeding. Laboratory studies during the resuscitation were: lactate dehydrogenase (LDH) 2,411 U/L (reference range: 88230 U/L), haptoglobin 25 mg/dL (reference range: 36195 mg/dL), and reticulocyte percentage 2.6%. Total bilirubin was 2.4 mg/dL (reference range: 0.11.2 mg/dL), with direct fraction 0.9 mg/dL (reference range: <0.3 mg/dL). Prothrombin time (PT) was 24.9 seconds (reference range: 11.815.2 seconds), international normalized ratio (INR) was 2.3, and activated partial thromboplastin time (aPTT) was 47.3 seconds (reference range: 22.233.0 seconds). Phlebotomy yielded icteric plasma. A peripheral blood smear revealed severe anemia, red cell agglutination, microspherocytes, and hemophagocytosis by neutrophils (Figure 1).

Figure 1

Low haptoglobin, elevated LDH, and hyperbilirubinemia confirm hemolysis. A more robust reticulocytosis is expected in the face of profound anemia, but the patient may also suffer from a concomitant hypoproliferative state (eg, nutritional deficiency). More likely, the rapidity of his decline outpaced the marrow's response, which can be delayed by days.

The most common cause of a combined elevation of the INR/PT and aPTT in a critically ill patient is DIC. Although no schistocytes were detected on the peripheral smear, they can be absent in up to 50% of DIC cases. TTP is associated with hemolytic anemia, kidney injury, and thrombocytopenia, but it generally does not cause coagulopathy.

The combination of red cell agglutination and hemophagocytosis suggests that the RBCs are coated with autoantibodies that cross‐link the cells and make them targets for phagocytosis by neutrophils in the circulation. This is distinct from the hemophagocytic syndrome, a rare immune activation syndrome characterized by macrophage phagocytosis of RBCs in the reticuloendothelial system. The blood smear also shows microspherocytes, which are seen in AIHA and hereditary spherocytosis.

Urinalysis showed large hemoglobin and 11 to 20 red blood cells per high‐power field. Serum creatine phosphokinase was 35 U/dL (reference range: 50388 U/dL). Four units of packed RBCs and 1 unit of fresh‐frozen plasma were transfused. Due to a rising creatinine level and declining urine output, continuous veno‐venous hemofiltration (CVVH) was initiated.

Acute tubular necrosis could result from sepsis, ischemic injury from DIC, hypotension during cardiac arrest, or heme pigment toxicity. Urine sediment should be reviewed for dysmorphic RBCs or RBC casts that would indicate glomerulonephritis (eg, from an underlying autoimmune process associated with AIHA).

Urine hemoglobin that is disproportionate to the degree of hematuria suggests hemoglobinuria, which in turn defines the hemolysis as intravascular. Processes that directly lyse RBCs in circulation via mechanical shearing, activation of complement, infection of the RBC, or enzymatic or oxidative destruction of the membrane cause intravascular hemolysis. Leading considerations include microangiopathy (eg, DIC, TTP), clostridial sepsis, and AIHA.

AIHA can be broadly classified as warm or cold. Warm AIHA is caused by immunoglobulin IgG antibodies that bind most avidly at body temperature. Because warm AIHA does not activate complement, patients present with evidence of extravascular hemolysis that is typically chronic and mild to moderate in severity. It does not typically cause the acute, fulminant, intravascular hemolytic condition seen here.

Cold AIHA is characterized by autoantibodies that bind at lower temperatures and comes in 2 forms: cold agglutinin disease and (rarely) paroxysmal cold hemoglobinuria (PCH). Cold agglutinins are most often IgM antibodies produced in response to infection (Mycoplasma pneumoniae, infectious mononucleosis), drugs, or a hematologic malignancy. These IgM antibodies bind RBCs, causing them to agglutinate, and fix complement (including C3) to the surface of RBCs when blood circulates to cooler parts of the body. This results in complement activation, formation of the membrane attack complex, and intravascular hemolysis when bound and activated complement is present in large numbers. Acute infection can increase the complement available for binding to the surface of RBCs. Through a slightly different mechanism, PCH causes intravascular hemolysis through direct IgG activation of complement fixed to the surface of RBCs. During a hemolytic episode the direct antibody test (DAT) is positive using anti‐C3 and negative for IgG.

Based on the patient's clinical evidence of intravascular hemolysis and a suspected autoimmune etiology, the leading diagnosis at this time is cold AIHA.

Due to coagulopathy and possible cold agglutinin disease, therapeutic hypothermia for neuroprotection was deferred. He continued on vancomycin, piperacillin‐tazobactam, and metronidazole. The DAT and direct IgG were strongly positive (3+), whereas the direct C3 was weakly positive (<1+). His serum free hemoglobin increased from 136.7 mg/dL to 223.8 mg/dL (normal: 0.06.9 mg/dL). His severe metabolic acidosis corrected with CVVH.

The DAT detects IgG or complement adherent to RBCs. This patient has tested positive for both IgG and C3, though much more strongly for IgG, suggesting an unusual ability of the patient's IgG to activate complement. The phenomenon of mixed AIHA, in which the patient has both warm‐ and cold‐reacting antibodies, is rare.

Regarding infections associated with AIHA, there is no cough or rash to suggest M pneumoniae, and there is no sore throat, fever, lymphadenopathy, splenomegaly, or atypical lymphocytosis to suggest infectious mononucleosis. He should be tested for human immunodeficiency virus, which is also associated with AIHA. His leukocytosis may raise suspicion for an underlying hematologic malignancy, but he does not have blasts, dysplastic leukocytes, or lymphocytosis on his peripheral blood smear. Systemic lupus erythematosus can be associated with AIHA, thrombocytopenia, and renal failure, but he lacks the more common clinical manifestations of rash, arthralgias, and fever.

Drug‐induced immune hemolytic anemia (DIIHA) can cause both the clinical and serologic profile of an AIHA, as seen here. DIIHA can be distinguished from mixed AIHA if hemolysis abates with discontinuation of an offending drug. His deterioration is temporally associated with drug administration at the time of admission. Cephalosporins and ‐lactams (e.g., piperacillin) are the most common causes of DIIHA, and ‐lactamases such as tazobactam have also been implicated. By exclusion of other causes, DIIHA secondary to piperacillin is most likely responsible for his massive intravascular hemolysis.

Medical records from 2 other hospitals arrived approximately 36 hours after the patient's initial presentation. Three weeks earlier, he had required RBC transfusions for a 6 g/dL hemoglobin reduction after receiving piperacillin‐tazobactam for sepsis. He had been treated with steroids for presumed warm AIHA and dialyzed for acute kidney injury attributed to hemoglobinuria. After 1 week in the intensive care unit, he left against medical advice. Two weeks later he presented to another ED, which immediately preceded his presentation to our hospital.

An antibody screen for drug‐dependent antibodies revealed antipiperacillin antibodies. Antitazobactam antibodies were not tested. Piperacillin‐tazobactam was discontinued, and plasmapheresis was initiated to decrease the amount of piperacillin in the blood. The patient's hemoglobin subsequently remained above 7.0 g/dL without RBC transfusions.

His renal function recovered and he completed antibiotic therapy for C difficile infection and for his pressure ulcers. However, he had sustained severe anoxic brain injury during his cardiac arrest and did not recover neurological function. In accordance with his family's wishes, he was discharged to a long‐term acute care hospital dependent on a ventilator.

COMMENTARY

This case illustrates a dramatic presentation of fulminant intravascular hemolysis secondary to piperacillin. The incidence of DIIHA is estimated to be 1 in 1 million.[1] Historically, methyldopa and high‐dose penicillin have been responsible for the majority of cases,[2] but in recent years complex penicillins, including piperacillin, and second‐ and third‐generation cephalosporins have been implicated.[3, 4] Cases of DIIHA are often underdiagnosed or misdiagnosed, as smoldering or less severe cases may not be recognized or are attributed to other causes.

A positive DAT, suggesting immunoglobulin and/or complement binding to RBCs, is the most reliable laboratory finding in DIIHA.[5] However, a positive DAT does not identify the source of the antigen and may result in misattribution of the immune hemolysis to autoimmunity rather than to a drug. Repeated or continued administration of the offending drug (as in this case) may perpetuate or worsen the hemolysis. Drug‐specific antibody tests may help to confirm the diagnosis, but these tests are complex and take significant time for specialized laboratories to run.

Severe hemolysis should be considered when a patient has a sudden and dramatic drop in his hemoglobin level in the absence of bleeding. Because DIIHA can be rapidly progressive, discontinuing a suspected culprit drug is the most important diagnostic and therapeutic measure. Typically, when an offending drug is stopped, the hemolysis stops as well. The time course over which this occurs depends on the rapidity of drug clearance.[4] Hemodialysis or plasmapheresis may be required in cases where the medication is renally excreted, particularly in cases of concomitant kidney injury. Evidence supporting corticosteroid use in DIIHA is limited, as the offending agent is usually discontinued by the time corticosteroids are initiated.[4]

This patient's DAT confirmed both IgG and complement activation, consistent with DIIHA caused by an immune complexlike reaction. This mechanism involves the antibody binding to a mixed epitope of the drug and a RBC membrane glycoprotein.[6] The offending drug was stopped only when review of his medical records established a clear temporal association between antibiotic administration and prior hemolysis.

The 2009 Health Information Technology for Economic and Clinical Health Act created an electronic health record (EHR) incentive program (meaningful use criteria).[7] By 2012, only 6% of hospitals met all of the stage 2 criteria, which include EHR interoperability across health systems.[8] The patient's preceding hemolytic event was described in records faxed by the outside hospitals, but without EHR interoperability, the treating clinicians did not have timely access to this information. Instead, the familiar manual process of obtaining outside records involving signed forms, phone calls, fax machines, and reams of paper progressed at its usual pace. Real‐time access to health records might have guided providers to select an alternative antibiotic regimen. Instead, a communication breakdown contributed to a catastrophic drug reaction and to this tragic patient outcome.

KEY TEACHING POINTS

1. In a patient presenting with acute hemolysis and a positive DAT, consider DIIHA.
2. Both piperacillin and tazobactam can cause a severe, complement‐mediated immune hemolytic anemia (DIIHA).
3. Drug‐induced antibodies are detected by direct antiglobulin testing, but a complete medication history is the key to diagnosis.
4. Management of drug‐induced hemolytic anemia involves immediate discontinuation of the culprit medication, supportive care, and potentially corticosteroids, plasmapheresis, and/or hemodialysis to expedite removal of the offending agent.
5. EHR interoperability may provide timely access to important health information across different hospitals, expedite health information exchange, and reduce adverse patient outcomes that stem from communication delays.

This case was submitted anonymously to AHRQ WebM&M on July 18, 2014, and was accepted on August 7, 2014. The case and WebM&M commentary were published online on October 26, 2015.[9] This separate commentary on the same case was later submitted to the Journal of Hospital Medicine on September 2, 2015, accepted on November 24, 2015, and published on January 22, 2016. The 2 publications are written by different authors, and although they reference the same case, they make different but valuable points.

A 59‐year‐old man with a history of hypertension was admitted with 6 months of shortness of breath, night sweats, and debilitating fatigue. His symptoms were initially mild and would persist for weeks at a time, after which he would feel better for several days. Over the 2 weeks prior to admission his symptoms had progressed, and he had become dyspneic with minimal exertion.

Progressive dyspnea has a broad differential that includes diseases of the heart (eg, congestive heart failure, aortic stenosis, constrictive pericarditis), lung (eg, chronic obstructive pulmonary disease, interstitial lung disease, pulmonary hypertension, pleural effusion), and blood (eg, anemia).

Night sweats suggest an inflammatory condition, but do not help prioritize infection, malignancy, or autoimmunity. Any of those conditions can be relapsing and remitting, at least in their early phases, but the return to normalcy raises the possibility of hypersensitivity pneumonitis from a periodic exposure.

The 6‐month duration makes typical bacterial and viral infections less likely and suggests indolent infections such as mycobacteria, fungi, or human immunodeficiency virus. Lymphoma or chronic leukemia could cause dyspnea through pleural or pulmonary involvement or from anemia. Autoimmune conditions such as systemic lupus erythematosus or adult Still's disease could also present with this course.

On admission, he described progressive orthopnea, lower extremity edema, and a 15‐lb weight gain. He denied chest pain or palpitations. His symptoms did not correlate with environmental or occupational exposures. He had been diagnosed with essential hypertension a few years earlier but was not taking any medications. He worked as an editor for a newspaper and had traveled throughout California. He never used tobacco and drank alcohol in moderation. He previously smoked marijuana. His father died of Alzheimer's disease, and his mother and 2 siblings were healthy.

Orthopnea, lower extremity edema, and weight gain suggest volume overload, which can result from heart failure, cirrhosis, renal failure, or nephrotic syndrome. The untreated hypertension is a principal risk factor for heart failure. Subacute bacterial endocarditis is an important consideration in a patient with suspected heart failure and night sweats. Travel through the central valley of California may have exposed him to coccidiodomycosis, which can cause chronic pulmonary and extrapulmonary infection.

Physical examination revealed a chronically ill‐appearing man in mild respiratory distress. His temperature was 37.2C, heart rate was 83 bpm, and blood pressure was 168/81 mm Hg. His oxygen saturation was 97% with a respiratory rate of 17 while breathing ambient air. Bilateral chemosis was present. He had crackles at the lung bases. There was a 2/6 systolic murmur loudest at the left lower sternal border with apical radiation. His jugular venous pressure was 2 cm above the sternal angle at 45. He had mild pitting edema of both lower extremities. His abdomen was soft and nondistended. He demonstrated full range of motion of all extremities and had no rashes. He was alert and oriented to person, place, and time. There were no cranial nerve deficits. His strength, sensation and coordination were intact, and he had a normal gait.

Chemosis (conjunctival edema) usually represents conjunctival irritation from an allergic, infectious, or toxic process. It can also be seen in cases of increased ophthalmic venous pressure such as hyperthyroid ophthalmopathy, superior vena cava syndrome, or carotid‐cavernous sinus fistula. The crackles, weight gain, borderline jugular venous distention, and edema suggest some systemic volume overload, but not enough to produce chemosis.

The location and timing of the murmur suggests regurgitation through the mitral or tricuspid valve, a ventricular septal defect, or hypertrophic cardiomyopathy. Tricuspid regurgitation may indicate pulmonary hypertension with right ventricular failure. Despite the absence of fever, subacute bacterial endocarditis remains a concern.

Laboratory evaluation revealed a white blood cell count of 9600/L, hemoglobin of 8.7 g/dL, and platelet count of 522,000/L. Mean corpuscular volume was 88 fL. Serum chemistries were normal; serum creatinine was 1.2 mg/dL. Serum albumin was 2.6 g/dL. A urinalysis was normal. An electrocardiogram demonstrated normal sinus rhythm and left ventricular hypertrophy (LVH). A chest x‐ray revealed interstitial edema and small bilateral pleural effusions. A transthoracic echocardiogram demonstrated normal left ventricular systolic function, an ejection fraction of 65%, mild LVH, and mild diastolic dysfunction. Mild mitral regurgitation, a mildly dilated left atrium, and a minimal pericardial effusion were also noted. A renal ultrasound revealed an atrophic left kidney without arterial flow. He was treated with diuretics for presumed heart failure related to diastolic dysfunction. His dyspnea partially improved, and he was discharged.

Heart failure with preserved ejection fraction may be contributing to his dyspnea but is unlikely to be entirely explanatory given the laboratory abnormalities. The absence of valvular vegetations on transthoracic echocardiogram lowers the probability of bacterial endocarditis. The interstitial pulmonary markings may represent pulmonary edema but alternatively could reflect interstitial lung disease, lymphangitic spread of cancer, infection (eg, Pneumocystis jiroveci), or diffuse alveolar hemorrhage.

Anemia may also be contributing to his dyspnea. There is no evidence of bleeding on history, examination, or imaging. Hemolysis is unlikely given the absence of jaundice, splenomegaly, or a known predisposing condition. The normocytic anemia may also arise from chronic inflammation. Severe anemia can cause high output heart failure, but usually the hemoglobin level is much lower and the echocardiogram would have suggestive findings. Thrombocytosis suggests inflammation, a primary myeloproliferative disorder, or severe iron deficiency (not suspected here). His hypoalbuminemia is further evidence of chronic inflammation especially in the absence of nephropathy, hepatopathy, or a protein‐losing enteropathy.

An atrophic kidney may be congenital or result from long‐standing unilateral renal ischemia, infection, or obstruction. Diminished arterial flow in a middle‐aged man with hypertension may simply reflect atherosclerotic renal artery stenosis, but mass effect within the left renal artery from thrombus, infection, or cancer cannot be ruled out.

Four weeks later he was readmitted for progressive dyspnea and persistent night sweats. He was afebrile, fatigued, and in marked respiratory distress. The remainder of his physical examination was unchanged. Laboratory evaluation revealed a white blood cell count of 20,000/L with neutrophilic predominance, hemoglobin of 11 g/dL, and platelet count of 614,000/L. Creatinine was 1.4 mg/dL. Erythrocyte sedimentation rate (ESR) was greater than 100 mm/h, and C‐reactive protein (CRP) was 44 mg/L. Blood cultures were negative. Chest x‐ray (Figure 1) revealed persistent interstitial edema and increased bilateral pleural effusions.

Figure 1

Although clinical and radiologic features continue to suggest heart failure, the marked respiratory distress and persistent chest x‐ray abnormalities imply that a superimposed process is affecting the lungs. The night sweats, neutrophilia, and elevated ESR and CRP strongly suggest an inflammatory state from infection, malignancy, or autoimmunity.

A computed tomography (CT) scan of the lungs would help assess for interstitial lung disease, lymphangitic carcinomatosis and septic emboli. Blood cultures should be repeated to definitively exclude subacute endocarditis. A peripheral blood smear is needed to evaluate for hematologic malignancy. Finally, human immunodeficiency virus antibody testing is indicated.

CT of the abdomen and pelvis demonstrated left renal artery stenosis, an atrophic left kidney, right kidney edema with mild perinephric stranding, and mild‐to‐moderate right hydroureter without an obstructing mass or calculus. There was mild splenomegaly and mesenteric lymphadenopathy up to 3 cm in diameter. The distal thoracic and suprarenal abdominal aorta had crescentic high‐density wall thickening. There were small sclerotic densities of the proximal femora, pelvic girdle, and thoracolumbar spine (Figure 2). Contrast chest CT demonstrated severe wall thickening of his entire thoracic aorta. There was also cardiomegaly, mild interlobular septal thickening, small bilateral pleural effusions, a 3.2‐cm right upper lobe paratracheal lymph node, and nodular pleural thickening (Figure 3).

Figure 2

Figure 3

Diffuse aortopathy is caused by inflammatory, infectious, or infiltrative processes. Large vessel vasculitides such as Behet's disease, giant cell arteritis, and Takayasu's arteritis are unlikely, as the patient lacks the associated clinical findings or epidemiology. Imaging does not reveal preexisting aortic pathology, such as an aneurysm or atherosclerotic plaque, which could predispose him to bacterial endovascular infection.

Urinary system dilation without an obvious obstruction could be explained by retroperitoneal fibrosis. Generalized lymphadenopathy, (suspected) retroperitoneal fibrosis, sclerotic bone lesions, and cardiopulmonary disease collectively suggest a widespread infiltrative process. Lymphoma may lead to lymphadenopathy and bone lesions but would not explain the aortopathy. He lacks risk factors for infections like tuberculosis or tertiary syphilis, a well‐known cause of aortopathy in the past.

Widespread multisystem involvement invites consideration of nonmalignant, noninfectious infiltrative disorders such as immunoglobulin G4‐related disease (IgG4‐RD), histiocytoses such as Erdheim‐Chester disease (ECD), systemic mastocytosis (SM), and sarcoidosis. ECD is a disorder of non‐Langerhans histiocytes that infiltrate the aorta, bones, retroperitoneum, lungs, myocardium, and periorbital structures. Perinephric stranding is sometimes seen in this condition. The lymphoplasmacytes in IgG4‐RD and noncaseating granulomas of sarcoidosis infiltrate many of the same organs. Common sites infiltrated by mast cells in SM include the bone and lymph nodes. Among these diseases, ECD and IgG4‐RD more commonly manifest with aortic and retroperitoneal infiltration and thus are prioritized on this differential diagnosis.

A positron emission tomography (PET) scan revealed abnormal fluordeoxyglucose uptake involving the thoracic aorta, right apical pleural surface, perinephric soft tissue, and various marrow spaces. Core needle biopsy of a sclerotic lesion in the right ischium demonstrated focal marrow replacement by a fibrohistiocytic process. No malignant cells or pathogenic organisms were identified. Biopsy of the right kidney revealed chronic interstitial nephritis with features of megalocytic interstitial nephritis (histiocytic inflammation) and arteriolar nephrosclerosis. A transbronchial biopsy demonstrated alveolar tissue with focal intra‐alveolar hemorrhage and organization, but no malignancy, atypia, or pathogenic organisms.

The biopsy results do not support infection, lymphoma, or carcinoma. The absence of noncaseating granulomas and mastocytes on multiple biopsies essentially rules out sarcoidosis and SM, respectively. None of the characteristic pathologic features of IgG4‐RDlymphoplasmacytic infiltrate, obliterative phlebitis, and fibrosiswere observed. The pulmonary pathology points to injury, but not the underlying cause. The bone and kidney tissue samples reveal histiocytic infiltration.

The abnormalities of the aorta, bone, lung, kidney, and retroperitoneum can be explained by the diffuse histiocytic involvement seen in ECD. The perinephric stranding detected on CT and perinephric inflammation on the PET scan may reflect the hairy kidney of ECD, which is a result of histiocytic infiltration. It is possible that the chemosis relates to exophthalmos from histiocytic orbital infiltrates. Sensitivity for detecting orbital pathology on a PET scan is limited because of the high signal from the adjacent brain.

ECD should be distinguished from Langerhans cell histiocytosis (LCH) by immunohistologic staining and microscopic characteristics of the histiocytes. LCH usually does not involve the aorta, and it more commonly involves the skin.

Serum IgG4 was within normal limits, and immunohistochemical staining of pathology specimens for IgG4 was negative. The BRAF V600E mutation, which is present in the majority of patients with ECD, was detected in a subsequent right perinephric biopsy specimen. The patient was diagnosed with ECD.

Prednisone and pegylated interferon‐ led to a rapid improvement in his symptoms. As the prednisone was tapered, he developed bilateral periorbital swelling. Magnetic resonance imaging (MRI) revealed well‐circumscribed, intraorbital soft tissue masses with partial encasement of his optic nerves and superior ophthalmic veins, as well as infiltration of his transverse sinuses, consistent with intracranial manifestations of ECD. There was no evidence of pituitary, hypothalamic, or other brain parenchymal infiltration. His dyspnea, night sweats, and hypertension improved; however, 3 months into therapy he developed an extensive rash. Interferon was discontinued. Vemurafinib, a serine kinase inhibitor that targets the BRAF mutation, was prescribed with subsequent resolution of the rash.

COMMENTARY

This patient suffered from a chronic, progressive, inflammatory illness. Although the disease initially appeared to be confined to the heart and lungs, laboratory testing signaled a more systemic condition, and subsequent imaging demonstrated involvement of a disparate group of organs. Subacute disease processes with elevated markers of inflammation and diffuse organ involvement often fall into 1 of 3 categories: infectious, autoimmune, or neoplastic. The histiocytoses inhabit a fourth and easily overlooked category that can be described as infiltrative. Infiltrative diseases are a heterogeneous group of conditions that cause illness when cells or substances not normally found in tissues lead to organ dysfunction.

Although traditional teaching has focused on sarcoidosis, amyloidosis, and hemochromatosis as the primary representatives of this category, the medical literature describes a number of other infiltrative disease processes. IgG4‐RD is a fibroinflammatory disorder characterized by space‐occupying lesions, a lymphoplasmacytic infiltrate of IgG4‐positive plasma cells, and storiform (matted and irregularly whorled microscopic pattern) fibrosis.[1] IgG4‐RD, like sarcoidosis, blurs the categorical line between infiltrative and autoimmune diseases. Other infiltrative cellular disorders, such as histiocytosis and mastocytosis, exist on a spectrum between monoclonal proliferation and neoplastic invasion.

The histiocytoses represent a diverse group of disorders with an evolving nomenclature, characterized by localized or diffuse infiltration of macrophages, monocytes, and dendritic cells (Table 1). ECD is a rare, non‐Langerhans histiocytosis characterized by excessive recruitment and activation of histiocytes through kinase signaling pathways.[2, 3] Immunohistochemical staining for CD68, CD163, and Factor XIIIa, with lack of staining for CD1a, S100, and CD207, supports the diagnosis.[3] Mutations in the BRAF V600E gene (a protein kinase involved in cell proliferation) represent the most likely etiology of this overactivation. An estimated 38% to 100% of patients with ECD harbor this mutation, with detection rates influenced by the sensitivity of testing techniques.[3] The serine kinase inhibitor vemurafinib targets this mutation, and early experience with this agent in ECD demonstrates encouraging results.[4]

ECD presents heterogeneously, occurring most commonly between the ages of 40 and 70 years. Nonspecific symptoms include weakness, fatigue, fever, chills, weight loss, and night sweats. Typical sites of involvement include the bone, central nervous system, cardiovascular system, lungs, and retroperitoneum. Bone involvement is nearly universal, and bone pain is the most common presenting symptom. Symmetric diaphyseal and metaphyseal osteosclerotic lesions may be seen on x‐rays, bone scan, PET, CT, and MRI.[3] Approximately 50% of patients have extraskeletal involvement at diagnosis.[5] Neurologic manifestations may result from invasion of histiocytes into the facial bones, orbits, meninges, and intracranial vessels, as eventually developed in this patient. Diabetes insipidus is the most common neurologic manifestation of ECD, followed by exophthalmos, cerebellar ataxia, panhypopituitarism, and papilledema.[6, 7] Approximately 75% of patients eventually suffer from cardiovascular disease, including hypertension, congestive heart failure, acute myocardial infarction, valvular dysfunction, pericardial infiltration, and cardiac tamponade.[8] Vascular involvement includes perivascular infiltration and periaortic fibrosis, resulting in the coated aorta seen in 20% of patients with ECD.[3] Pulmonary manifestations of ECD include interstitial, pleural, and consolidative lung disease. A review of high‐resolution chest CTs of patients with ECD demonstrated that greater than half had evidence of parenchymal lung disease, with interlobular septal thickening being the most common finding.[9] Infiltration and fibrosis of retroperitoneal structures is common. Infiltration of perinephric fat creates irregular renal borders, appearing radiographically as hairy kidneys.

Arriving at the diagnosis in this case proved to be challenging because the early presentation was consistent with congestive heart failure. As the patient's conditioned deteriorated, imaging suggested multisystem involvement. It was the extensive aortopathy in particularnot the less specific bone, kidney, lymph node, or pulmonary findingsthat allowed the clinicians to hone the extensive differential diagnosis. The coated aorta is a finding that has been strongly associated with ECD; few other conditions coat the aorta in a similar fashion.[10] In most mysteries, the perpetrator's coat conceals his identity; however, in this story the coat gave it away.

KEY LEARNING POINTS

1. Subacute, inflammatory, multiorgan disease is usually explained by 3 categoriesinfection, autoimmunity, and neoplasiabut a fourth category, infiltrative disorders, sometimes warrants consideration.
2. ECD presents heterogeneously, ranging from localized disease to widespread organ infiltration. The classic presentation includes bone pain, diabetes insipidus, and exophthalmos.
3. Characteristic radiological findings that suggest ECD include long bone osteosclerosis, a coated aorta from periaortic infiltration, and hairy kidneys from perinephric infiltration.

Disclosure

Nothing to report

A 59‐year‐old man with a history of hypertension was admitted with 6 months of shortness of breath, night sweats, and debilitating fatigue. His symptoms were initially mild and would persist for weeks at a time, after which he would feel better for several days. Over the 2 weeks prior to admission his symptoms had progressed, and he had become dyspneic with minimal exertion.

Progressive dyspnea has a broad differential that includes diseases of the heart (eg, congestive heart failure, aortic stenosis, constrictive pericarditis), lung (eg, chronic obstructive pulmonary disease, interstitial lung disease, pulmonary hypertension, pleural effusion), and blood (eg, anemia).

Night sweats suggest an inflammatory condition, but do not help prioritize infection, malignancy, or autoimmunity. Any of those conditions can be relapsing and remitting, at least in their early phases, but the return to normalcy raises the possibility of hypersensitivity pneumonitis from a periodic exposure.

The 6‐month duration makes typical bacterial and viral infections less likely and suggests indolent infections such as mycobacteria, fungi, or human immunodeficiency virus. Lymphoma or chronic leukemia could cause dyspnea through pleural or pulmonary involvement or from anemia. Autoimmune conditions such as systemic lupus erythematosus or adult Still's disease could also present with this course.

On admission, he described progressive orthopnea, lower extremity edema, and a 15‐lb weight gain. He denied chest pain or palpitations. His symptoms did not correlate with environmental or occupational exposures. He had been diagnosed with essential hypertension a few years earlier but was not taking any medications. He worked as an editor for a newspaper and had traveled throughout California. He never used tobacco and drank alcohol in moderation. He previously smoked marijuana. His father died of Alzheimer's disease, and his mother and 2 siblings were healthy.

Orthopnea, lower extremity edema, and weight gain suggest volume overload, which can result from heart failure, cirrhosis, renal failure, or nephrotic syndrome. The untreated hypertension is a principal risk factor for heart failure. Subacute bacterial endocarditis is an important consideration in a patient with suspected heart failure and night sweats. Travel through the central valley of California may have exposed him to coccidiodomycosis, which can cause chronic pulmonary and extrapulmonary infection.

Physical examination revealed a chronically ill‐appearing man in mild respiratory distress. His temperature was 37.2C, heart rate was 83 bpm, and blood pressure was 168/81 mm Hg. His oxygen saturation was 97% with a respiratory rate of 17 while breathing ambient air. Bilateral chemosis was present. He had crackles at the lung bases. There was a 2/6 systolic murmur loudest at the left lower sternal border with apical radiation. His jugular venous pressure was 2 cm above the sternal angle at 45. He had mild pitting edema of both lower extremities. His abdomen was soft and nondistended. He demonstrated full range of motion of all extremities and had no rashes. He was alert and oriented to person, place, and time. There were no cranial nerve deficits. His strength, sensation and coordination were intact, and he had a normal gait.

Chemosis (conjunctival edema) usually represents conjunctival irritation from an allergic, infectious, or toxic process. It can also be seen in cases of increased ophthalmic venous pressure such as hyperthyroid ophthalmopathy, superior vena cava syndrome, or carotid‐cavernous sinus fistula. The crackles, weight gain, borderline jugular venous distention, and edema suggest some systemic volume overload, but not enough to produce chemosis.

The location and timing of the murmur suggests regurgitation through the mitral or tricuspid valve, a ventricular septal defect, or hypertrophic cardiomyopathy. Tricuspid regurgitation may indicate pulmonary hypertension with right ventricular failure. Despite the absence of fever, subacute bacterial endocarditis remains a concern.

Laboratory evaluation revealed a white blood cell count of 9600/L, hemoglobin of 8.7 g/dL, and platelet count of 522,000/L. Mean corpuscular volume was 88 fL. Serum chemistries were normal; serum creatinine was 1.2 mg/dL. Serum albumin was 2.6 g/dL. A urinalysis was normal. An electrocardiogram demonstrated normal sinus rhythm and left ventricular hypertrophy (LVH). A chest x‐ray revealed interstitial edema and small bilateral pleural effusions. A transthoracic echocardiogram demonstrated normal left ventricular systolic function, an ejection fraction of 65%, mild LVH, and mild diastolic dysfunction. Mild mitral regurgitation, a mildly dilated left atrium, and a minimal pericardial effusion were also noted. A renal ultrasound revealed an atrophic left kidney without arterial flow. He was treated with diuretics for presumed heart failure related to diastolic dysfunction. His dyspnea partially improved, and he was discharged.

Heart failure with preserved ejection fraction may be contributing to his dyspnea but is unlikely to be entirely explanatory given the laboratory abnormalities. The absence of valvular vegetations on transthoracic echocardiogram lowers the probability of bacterial endocarditis. The interstitial pulmonary markings may represent pulmonary edema but alternatively could reflect interstitial lung disease, lymphangitic spread of cancer, infection (eg, Pneumocystis jiroveci), or diffuse alveolar hemorrhage.

Anemia may also be contributing to his dyspnea. There is no evidence of bleeding on history, examination, or imaging. Hemolysis is unlikely given the absence of jaundice, splenomegaly, or a known predisposing condition. The normocytic anemia may also arise from chronic inflammation. Severe anemia can cause high output heart failure, but usually the hemoglobin level is much lower and the echocardiogram would have suggestive findings. Thrombocytosis suggests inflammation, a primary myeloproliferative disorder, or severe iron deficiency (not suspected here). His hypoalbuminemia is further evidence of chronic inflammation especially in the absence of nephropathy, hepatopathy, or a protein‐losing enteropathy.

An atrophic kidney may be congenital or result from long‐standing unilateral renal ischemia, infection, or obstruction. Diminished arterial flow in a middle‐aged man with hypertension may simply reflect atherosclerotic renal artery stenosis, but mass effect within the left renal artery from thrombus, infection, or cancer cannot be ruled out.

Four weeks later he was readmitted for progressive dyspnea and persistent night sweats. He was afebrile, fatigued, and in marked respiratory distress. The remainder of his physical examination was unchanged. Laboratory evaluation revealed a white blood cell count of 20,000/L with neutrophilic predominance, hemoglobin of 11 g/dL, and platelet count of 614,000/L. Creatinine was 1.4 mg/dL. Erythrocyte sedimentation rate (ESR) was greater than 100 mm/h, and C‐reactive protein (CRP) was 44 mg/L. Blood cultures were negative. Chest x‐ray (Figure 1) revealed persistent interstitial edema and increased bilateral pleural effusions.

Figure 1

Although clinical and radiologic features continue to suggest heart failure, the marked respiratory distress and persistent chest x‐ray abnormalities imply that a superimposed process is affecting the lungs. The night sweats, neutrophilia, and elevated ESR and CRP strongly suggest an inflammatory state from infection, malignancy, or autoimmunity.

A computed tomography (CT) scan of the lungs would help assess for interstitial lung disease, lymphangitic carcinomatosis and septic emboli. Blood cultures should be repeated to definitively exclude subacute endocarditis. A peripheral blood smear is needed to evaluate for hematologic malignancy. Finally, human immunodeficiency virus antibody testing is indicated.

CT of the abdomen and pelvis demonstrated left renal artery stenosis, an atrophic left kidney, right kidney edema with mild perinephric stranding, and mild‐to‐moderate right hydroureter without an obstructing mass or calculus. There was mild splenomegaly and mesenteric lymphadenopathy up to 3 cm in diameter. The distal thoracic and suprarenal abdominal aorta had crescentic high‐density wall thickening. There were small sclerotic densities of the proximal femora, pelvic girdle, and thoracolumbar spine (Figure 2). Contrast chest CT demonstrated severe wall thickening of his entire thoracic aorta. There was also cardiomegaly, mild interlobular septal thickening, small bilateral pleural effusions, a 3.2‐cm right upper lobe paratracheal lymph node, and nodular pleural thickening (Figure 3).

Figure 2

Figure 3

Diffuse aortopathy is caused by inflammatory, infectious, or infiltrative processes. Large vessel vasculitides such as Behet's disease, giant cell arteritis, and Takayasu's arteritis are unlikely, as the patient lacks the associated clinical findings or epidemiology. Imaging does not reveal preexisting aortic pathology, such as an aneurysm or atherosclerotic plaque, which could predispose him to bacterial endovascular infection.

Urinary system dilation without an obvious obstruction could be explained by retroperitoneal fibrosis. Generalized lymphadenopathy, (suspected) retroperitoneal fibrosis, sclerotic bone lesions, and cardiopulmonary disease collectively suggest a widespread infiltrative process. Lymphoma may lead to lymphadenopathy and bone lesions but would not explain the aortopathy. He lacks risk factors for infections like tuberculosis or tertiary syphilis, a well‐known cause of aortopathy in the past.

Widespread multisystem involvement invites consideration of nonmalignant, noninfectious infiltrative disorders such as immunoglobulin G4‐related disease (IgG4‐RD), histiocytoses such as Erdheim‐Chester disease (ECD), systemic mastocytosis (SM), and sarcoidosis. ECD is a disorder of non‐Langerhans histiocytes that infiltrate the aorta, bones, retroperitoneum, lungs, myocardium, and periorbital structures. Perinephric stranding is sometimes seen in this condition. The lymphoplasmacytes in IgG4‐RD and noncaseating granulomas of sarcoidosis infiltrate many of the same organs. Common sites infiltrated by mast cells in SM include the bone and lymph nodes. Among these diseases, ECD and IgG4‐RD more commonly manifest with aortic and retroperitoneal infiltration and thus are prioritized on this differential diagnosis.

A positron emission tomography (PET) scan revealed abnormal fluordeoxyglucose uptake involving the thoracic aorta, right apical pleural surface, perinephric soft tissue, and various marrow spaces. Core needle biopsy of a sclerotic lesion in the right ischium demonstrated focal marrow replacement by a fibrohistiocytic process. No malignant cells or pathogenic organisms were identified. Biopsy of the right kidney revealed chronic interstitial nephritis with features of megalocytic interstitial nephritis (histiocytic inflammation) and arteriolar nephrosclerosis. A transbronchial biopsy demonstrated alveolar tissue with focal intra‐alveolar hemorrhage and organization, but no malignancy, atypia, or pathogenic organisms.

The biopsy results do not support infection, lymphoma, or carcinoma. The absence of noncaseating granulomas and mastocytes on multiple biopsies essentially rules out sarcoidosis and SM, respectively. None of the characteristic pathologic features of IgG4‐RDlymphoplasmacytic infiltrate, obliterative phlebitis, and fibrosiswere observed. The pulmonary pathology points to injury, but not the underlying cause. The bone and kidney tissue samples reveal histiocytic infiltration.

The abnormalities of the aorta, bone, lung, kidney, and retroperitoneum can be explained by the diffuse histiocytic involvement seen in ECD. The perinephric stranding detected on CT and perinephric inflammation on the PET scan may reflect the hairy kidney of ECD, which is a result of histiocytic infiltration. It is possible that the chemosis relates to exophthalmos from histiocytic orbital infiltrates. Sensitivity for detecting orbital pathology on a PET scan is limited because of the high signal from the adjacent brain.

ECD should be distinguished from Langerhans cell histiocytosis (LCH) by immunohistologic staining and microscopic characteristics of the histiocytes. LCH usually does not involve the aorta, and it more commonly involves the skin.

Serum IgG4 was within normal limits, and immunohistochemical staining of pathology specimens for IgG4 was negative. The BRAF V600E mutation, which is present in the majority of patients with ECD, was detected in a subsequent right perinephric biopsy specimen. The patient was diagnosed with ECD.

Prednisone and pegylated interferon‐ led to a rapid improvement in his symptoms. As the prednisone was tapered, he developed bilateral periorbital swelling. Magnetic resonance imaging (MRI) revealed well‐circumscribed, intraorbital soft tissue masses with partial encasement of his optic nerves and superior ophthalmic veins, as well as infiltration of his transverse sinuses, consistent with intracranial manifestations of ECD. There was no evidence of pituitary, hypothalamic, or other brain parenchymal infiltration. His dyspnea, night sweats, and hypertension improved; however, 3 months into therapy he developed an extensive rash. Interferon was discontinued. Vemurafinib, a serine kinase inhibitor that targets the BRAF mutation, was prescribed with subsequent resolution of the rash.

COMMENTARY

This patient suffered from a chronic, progressive, inflammatory illness. Although the disease initially appeared to be confined to the heart and lungs, laboratory testing signaled a more systemic condition, and subsequent imaging demonstrated involvement of a disparate group of organs. Subacute disease processes with elevated markers of inflammation and diffuse organ involvement often fall into 1 of 3 categories: infectious, autoimmune, or neoplastic. The histiocytoses inhabit a fourth and easily overlooked category that can be described as infiltrative. Infiltrative diseases are a heterogeneous group of conditions that cause illness when cells or substances not normally found in tissues lead to organ dysfunction.

Although traditional teaching has focused on sarcoidosis, amyloidosis, and hemochromatosis as the primary representatives of this category, the medical literature describes a number of other infiltrative disease processes. IgG4‐RD is a fibroinflammatory disorder characterized by space‐occupying lesions, a lymphoplasmacytic infiltrate of IgG4‐positive plasma cells, and storiform (matted and irregularly whorled microscopic pattern) fibrosis.[1] IgG4‐RD, like sarcoidosis, blurs the categorical line between infiltrative and autoimmune diseases. Other infiltrative cellular disorders, such as histiocytosis and mastocytosis, exist on a spectrum between monoclonal proliferation and neoplastic invasion.

The histiocytoses represent a diverse group of disorders with an evolving nomenclature, characterized by localized or diffuse infiltration of macrophages, monocytes, and dendritic cells (Table 1). ECD is a rare, non‐Langerhans histiocytosis characterized by excessive recruitment and activation of histiocytes through kinase signaling pathways.[2, 3] Immunohistochemical staining for CD68, CD163, and Factor XIIIa, with lack of staining for CD1a, S100, and CD207, supports the diagnosis.[3] Mutations in the BRAF V600E gene (a protein kinase involved in cell proliferation) represent the most likely etiology of this overactivation. An estimated 38% to 100% of patients with ECD harbor this mutation, with detection rates influenced by the sensitivity of testing techniques.[3] The serine kinase inhibitor vemurafinib targets this mutation, and early experience with this agent in ECD demonstrates encouraging results.[4]

ECD presents heterogeneously, occurring most commonly between the ages of 40 and 70 years. Nonspecific symptoms include weakness, fatigue, fever, chills, weight loss, and night sweats. Typical sites of involvement include the bone, central nervous system, cardiovascular system, lungs, and retroperitoneum. Bone involvement is nearly universal, and bone pain is the most common presenting symptom. Symmetric diaphyseal and metaphyseal osteosclerotic lesions may be seen on x‐rays, bone scan, PET, CT, and MRI.[3] Approximately 50% of patients have extraskeletal involvement at diagnosis.[5] Neurologic manifestations may result from invasion of histiocytes into the facial bones, orbits, meninges, and intracranial vessels, as eventually developed in this patient. Diabetes insipidus is the most common neurologic manifestation of ECD, followed by exophthalmos, cerebellar ataxia, panhypopituitarism, and papilledema.[6, 7] Approximately 75% of patients eventually suffer from cardiovascular disease, including hypertension, congestive heart failure, acute myocardial infarction, valvular dysfunction, pericardial infiltration, and cardiac tamponade.[8] Vascular involvement includes perivascular infiltration and periaortic fibrosis, resulting in the coated aorta seen in 20% of patients with ECD.[3] Pulmonary manifestations of ECD include interstitial, pleural, and consolidative lung disease. A review of high‐resolution chest CTs of patients with ECD demonstrated that greater than half had evidence of parenchymal lung disease, with interlobular septal thickening being the most common finding.[9] Infiltration and fibrosis of retroperitoneal structures is common. Infiltration of perinephric fat creates irregular renal borders, appearing radiographically as hairy kidneys.

Arriving at the diagnosis in this case proved to be challenging because the early presentation was consistent with congestive heart failure. As the patient's conditioned deteriorated, imaging suggested multisystem involvement. It was the extensive aortopathy in particularnot the less specific bone, kidney, lymph node, or pulmonary findingsthat allowed the clinicians to hone the extensive differential diagnosis. The coated aorta is a finding that has been strongly associated with ECD; few other conditions coat the aorta in a similar fashion.[10] In most mysteries, the perpetrator's coat conceals his identity; however, in this story the coat gave it away.

KEY LEARNING POINTS

1. Subacute, inflammatory, multiorgan disease is usually explained by 3 categoriesinfection, autoimmunity, and neoplasiabut a fourth category, infiltrative disorders, sometimes warrants consideration.
2. ECD presents heterogeneously, ranging from localized disease to widespread organ infiltration. The classic presentation includes bone pain, diabetes insipidus, and exophthalmos.
3. Characteristic radiological findings that suggest ECD include long bone osteosclerosis, a coated aorta from periaortic infiltration, and hairy kidneys from perinephric infiltration.

Disclosure

Nothing to report

A 59‐year‐old man with a history of hypertension was admitted with 6 months of shortness of breath, night sweats, and debilitating fatigue. His symptoms were initially mild and would persist for weeks at a time, after which he would feel better for several days. Over the 2 weeks prior to admission his symptoms had progressed, and he had become dyspneic with minimal exertion.

Progressive dyspnea has a broad differential that includes diseases of the heart (eg, congestive heart failure, aortic stenosis, constrictive pericarditis), lung (eg, chronic obstructive pulmonary disease, interstitial lung disease, pulmonary hypertension, pleural effusion), and blood (eg, anemia).

Night sweats suggest an inflammatory condition, but do not help prioritize infection, malignancy, or autoimmunity. Any of those conditions can be relapsing and remitting, at least in their early phases, but the return to normalcy raises the possibility of hypersensitivity pneumonitis from a periodic exposure.

The 6‐month duration makes typical bacterial and viral infections less likely and suggests indolent infections such as mycobacteria, fungi, or human immunodeficiency virus. Lymphoma or chronic leukemia could cause dyspnea through pleural or pulmonary involvement or from anemia. Autoimmune conditions such as systemic lupus erythematosus or adult Still's disease could also present with this course.

On admission, he described progressive orthopnea, lower extremity edema, and a 15‐lb weight gain. He denied chest pain or palpitations. His symptoms did not correlate with environmental or occupational exposures. He had been diagnosed with essential hypertension a few years earlier but was not taking any medications. He worked as an editor for a newspaper and had traveled throughout California. He never used tobacco and drank alcohol in moderation. He previously smoked marijuana. His father died of Alzheimer's disease, and his mother and 2 siblings were healthy.

Orthopnea, lower extremity edema, and weight gain suggest volume overload, which can result from heart failure, cirrhosis, renal failure, or nephrotic syndrome. The untreated hypertension is a principal risk factor for heart failure. Subacute bacterial endocarditis is an important consideration in a patient with suspected heart failure and night sweats. Travel through the central valley of California may have exposed him to coccidiodomycosis, which can cause chronic pulmonary and extrapulmonary infection.

Physical examination revealed a chronically ill‐appearing man in mild respiratory distress. His temperature was 37.2C, heart rate was 83 bpm, and blood pressure was 168/81 mm Hg. His oxygen saturation was 97% with a respiratory rate of 17 while breathing ambient air. Bilateral chemosis was present. He had crackles at the lung bases. There was a 2/6 systolic murmur loudest at the left lower sternal border with apical radiation. His jugular venous pressure was 2 cm above the sternal angle at 45. He had mild pitting edema of both lower extremities. His abdomen was soft and nondistended. He demonstrated full range of motion of all extremities and had no rashes. He was alert and oriented to person, place, and time. There were no cranial nerve deficits. His strength, sensation and coordination were intact, and he had a normal gait.

Chemosis (conjunctival edema) usually represents conjunctival irritation from an allergic, infectious, or toxic process. It can also be seen in cases of increased ophthalmic venous pressure such as hyperthyroid ophthalmopathy, superior vena cava syndrome, or carotid‐cavernous sinus fistula. The crackles, weight gain, borderline jugular venous distention, and edema suggest some systemic volume overload, but not enough to produce chemosis.

The location and timing of the murmur suggests regurgitation through the mitral or tricuspid valve, a ventricular septal defect, or hypertrophic cardiomyopathy. Tricuspid regurgitation may indicate pulmonary hypertension with right ventricular failure. Despite the absence of fever, subacute bacterial endocarditis remains a concern.

Laboratory evaluation revealed a white blood cell count of 9600/L, hemoglobin of 8.7 g/dL, and platelet count of 522,000/L. Mean corpuscular volume was 88 fL. Serum chemistries were normal; serum creatinine was 1.2 mg/dL. Serum albumin was 2.6 g/dL. A urinalysis was normal. An electrocardiogram demonstrated normal sinus rhythm and left ventricular hypertrophy (LVH). A chest x‐ray revealed interstitial edema and small bilateral pleural effusions. A transthoracic echocardiogram demonstrated normal left ventricular systolic function, an ejection fraction of 65%, mild LVH, and mild diastolic dysfunction. Mild mitral regurgitation, a mildly dilated left atrium, and a minimal pericardial effusion were also noted. A renal ultrasound revealed an atrophic left kidney without arterial flow. He was treated with diuretics for presumed heart failure related to diastolic dysfunction. His dyspnea partially improved, and he was discharged.

Heart failure with preserved ejection fraction may be contributing to his dyspnea but is unlikely to be entirely explanatory given the laboratory abnormalities. The absence of valvular vegetations on transthoracic echocardiogram lowers the probability of bacterial endocarditis. The interstitial pulmonary markings may represent pulmonary edema but alternatively could reflect interstitial lung disease, lymphangitic spread of cancer, infection (eg, Pneumocystis jiroveci), or diffuse alveolar hemorrhage.

Anemia may also be contributing to his dyspnea. There is no evidence of bleeding on history, examination, or imaging. Hemolysis is unlikely given the absence of jaundice, splenomegaly, or a known predisposing condition. The normocytic anemia may also arise from chronic inflammation. Severe anemia can cause high output heart failure, but usually the hemoglobin level is much lower and the echocardiogram would have suggestive findings. Thrombocytosis suggests inflammation, a primary myeloproliferative disorder, or severe iron deficiency (not suspected here). His hypoalbuminemia is further evidence of chronic inflammation especially in the absence of nephropathy, hepatopathy, or a protein‐losing enteropathy.

An atrophic kidney may be congenital or result from long‐standing unilateral renal ischemia, infection, or obstruction. Diminished arterial flow in a middle‐aged man with hypertension may simply reflect atherosclerotic renal artery stenosis, but mass effect within the left renal artery from thrombus, infection, or cancer cannot be ruled out.

Four weeks later he was readmitted for progressive dyspnea and persistent night sweats. He was afebrile, fatigued, and in marked respiratory distress. The remainder of his physical examination was unchanged. Laboratory evaluation revealed a white blood cell count of 20,000/L with neutrophilic predominance, hemoglobin of 11 g/dL, and platelet count of 614,000/L. Creatinine was 1.4 mg/dL. Erythrocyte sedimentation rate (ESR) was greater than 100 mm/h, and C‐reactive protein (CRP) was 44 mg/L. Blood cultures were negative. Chest x‐ray (Figure 1) revealed persistent interstitial edema and increased bilateral pleural effusions.

Figure 1

Although clinical and radiologic features continue to suggest heart failure, the marked respiratory distress and persistent chest x‐ray abnormalities imply that a superimposed process is affecting the lungs. The night sweats, neutrophilia, and elevated ESR and CRP strongly suggest an inflammatory state from infection, malignancy, or autoimmunity.

A computed tomography (CT) scan of the lungs would help assess for interstitial lung disease, lymphangitic carcinomatosis and septic emboli. Blood cultures should be repeated to definitively exclude subacute endocarditis. A peripheral blood smear is needed to evaluate for hematologic malignancy. Finally, human immunodeficiency virus antibody testing is indicated.

CT of the abdomen and pelvis demonstrated left renal artery stenosis, an atrophic left kidney, right kidney edema with mild perinephric stranding, and mild‐to‐moderate right hydroureter without an obstructing mass or calculus. There was mild splenomegaly and mesenteric lymphadenopathy up to 3 cm in diameter. The distal thoracic and suprarenal abdominal aorta had crescentic high‐density wall thickening. There were small sclerotic densities of the proximal femora, pelvic girdle, and thoracolumbar spine (Figure 2). Contrast chest CT demonstrated severe wall thickening of his entire thoracic aorta. There was also cardiomegaly, mild interlobular septal thickening, small bilateral pleural effusions, a 3.2‐cm right upper lobe paratracheal lymph node, and nodular pleural thickening (Figure 3).

Figure 2

Figure 3

Diffuse aortopathy is caused by inflammatory, infectious, or infiltrative processes. Large vessel vasculitides such as Behet's disease, giant cell arteritis, and Takayasu's arteritis are unlikely, as the patient lacks the associated clinical findings or epidemiology. Imaging does not reveal preexisting aortic pathology, such as an aneurysm or atherosclerotic plaque, which could predispose him to bacterial endovascular infection.

Urinary system dilation without an obvious obstruction could be explained by retroperitoneal fibrosis. Generalized lymphadenopathy, (suspected) retroperitoneal fibrosis, sclerotic bone lesions, and cardiopulmonary disease collectively suggest a widespread infiltrative process. Lymphoma may lead to lymphadenopathy and bone lesions but would not explain the aortopathy. He lacks risk factors for infections like tuberculosis or tertiary syphilis, a well‐known cause of aortopathy in the past.

Widespread multisystem involvement invites consideration of nonmalignant, noninfectious infiltrative disorders such as immunoglobulin G4‐related disease (IgG4‐RD), histiocytoses such as Erdheim‐Chester disease (ECD), systemic mastocytosis (SM), and sarcoidosis. ECD is a disorder of non‐Langerhans histiocytes that infiltrate the aorta, bones, retroperitoneum, lungs, myocardium, and periorbital structures. Perinephric stranding is sometimes seen in this condition. The lymphoplasmacytes in IgG4‐RD and noncaseating granulomas of sarcoidosis infiltrate many of the same organs. Common sites infiltrated by mast cells in SM include the bone and lymph nodes. Among these diseases, ECD and IgG4‐RD more commonly manifest with aortic and retroperitoneal infiltration and thus are prioritized on this differential diagnosis.

A positron emission tomography (PET) scan revealed abnormal fluordeoxyglucose uptake involving the thoracic aorta, right apical pleural surface, perinephric soft tissue, and various marrow spaces. Core needle biopsy of a sclerotic lesion in the right ischium demonstrated focal marrow replacement by a fibrohistiocytic process. No malignant cells or pathogenic organisms were identified. Biopsy of the right kidney revealed chronic interstitial nephritis with features of megalocytic interstitial nephritis (histiocytic inflammation) and arteriolar nephrosclerosis. A transbronchial biopsy demonstrated alveolar tissue with focal intra‐alveolar hemorrhage and organization, but no malignancy, atypia, or pathogenic organisms.

The biopsy results do not support infection, lymphoma, or carcinoma. The absence of noncaseating granulomas and mastocytes on multiple biopsies essentially rules out sarcoidosis and SM, respectively. None of the characteristic pathologic features of IgG4‐RDlymphoplasmacytic infiltrate, obliterative phlebitis, and fibrosiswere observed. The pulmonary pathology points to injury, but not the underlying cause. The bone and kidney tissue samples reveal histiocytic infiltration.

The abnormalities of the aorta, bone, lung, kidney, and retroperitoneum can be explained by the diffuse histiocytic involvement seen in ECD. The perinephric stranding detected on CT and perinephric inflammation on the PET scan may reflect the hairy kidney of ECD, which is a result of histiocytic infiltration. It is possible that the chemosis relates to exophthalmos from histiocytic orbital infiltrates. Sensitivity for detecting orbital pathology on a PET scan is limited because of the high signal from the adjacent brain.

ECD should be distinguished from Langerhans cell histiocytosis (LCH) by immunohistologic staining and microscopic characteristics of the histiocytes. LCH usually does not involve the aorta, and it more commonly involves the skin.

Serum IgG4 was within normal limits, and immunohistochemical staining of pathology specimens for IgG4 was negative. The BRAF V600E mutation, which is present in the majority of patients with ECD, was detected in a subsequent right perinephric biopsy specimen. The patient was diagnosed with ECD.

Prednisone and pegylated interferon‐ led to a rapid improvement in his symptoms. As the prednisone was tapered, he developed bilateral periorbital swelling. Magnetic resonance imaging (MRI) revealed well‐circumscribed, intraorbital soft tissue masses with partial encasement of his optic nerves and superior ophthalmic veins, as well as infiltration of his transverse sinuses, consistent with intracranial manifestations of ECD. There was no evidence of pituitary, hypothalamic, or other brain parenchymal infiltration. His dyspnea, night sweats, and hypertension improved; however, 3 months into therapy he developed an extensive rash. Interferon was discontinued. Vemurafinib, a serine kinase inhibitor that targets the BRAF mutation, was prescribed with subsequent resolution of the rash.

COMMENTARY

This patient suffered from a chronic, progressive, inflammatory illness. Although the disease initially appeared to be confined to the heart and lungs, laboratory testing signaled a more systemic condition, and subsequent imaging demonstrated involvement of a disparate group of organs. Subacute disease processes with elevated markers of inflammation and diffuse organ involvement often fall into 1 of 3 categories: infectious, autoimmune, or neoplastic. The histiocytoses inhabit a fourth and easily overlooked category that can be described as infiltrative. Infiltrative diseases are a heterogeneous group of conditions that cause illness when cells or substances not normally found in tissues lead to organ dysfunction.

Although traditional teaching has focused on sarcoidosis, amyloidosis, and hemochromatosis as the primary representatives of this category, the medical literature describes a number of other infiltrative disease processes. IgG4‐RD is a fibroinflammatory disorder characterized by space‐occupying lesions, a lymphoplasmacytic infiltrate of IgG4‐positive plasma cells, and storiform (matted and irregularly whorled microscopic pattern) fibrosis.[1] IgG4‐RD, like sarcoidosis, blurs the categorical line between infiltrative and autoimmune diseases. Other infiltrative cellular disorders, such as histiocytosis and mastocytosis, exist on a spectrum between monoclonal proliferation and neoplastic invasion.

The histiocytoses represent a diverse group of disorders with an evolving nomenclature, characterized by localized or diffuse infiltration of macrophages, monocytes, and dendritic cells (Table 1). ECD is a rare, non‐Langerhans histiocytosis characterized by excessive recruitment and activation of histiocytes through kinase signaling pathways.[2, 3] Immunohistochemical staining for CD68, CD163, and Factor XIIIa, with lack of staining for CD1a, S100, and CD207, supports the diagnosis.[3] Mutations in the BRAF V600E gene (a protein kinase involved in cell proliferation) represent the most likely etiology of this overactivation. An estimated 38% to 100% of patients with ECD harbor this mutation, with detection rates influenced by the sensitivity of testing techniques.[3] The serine kinase inhibitor vemurafinib targets this mutation, and early experience with this agent in ECD demonstrates encouraging results.[4]

ECD presents heterogeneously, occurring most commonly between the ages of 40 and 70 years. Nonspecific symptoms include weakness, fatigue, fever, chills, weight loss, and night sweats. Typical sites of involvement include the bone, central nervous system, cardiovascular system, lungs, and retroperitoneum. Bone involvement is nearly universal, and bone pain is the most common presenting symptom. Symmetric diaphyseal and metaphyseal osteosclerotic lesions may be seen on x‐rays, bone scan, PET, CT, and MRI.[3] Approximately 50% of patients have extraskeletal involvement at diagnosis.[5] Neurologic manifestations may result from invasion of histiocytes into the facial bones, orbits, meninges, and intracranial vessels, as eventually developed in this patient. Diabetes insipidus is the most common neurologic manifestation of ECD, followed by exophthalmos, cerebellar ataxia, panhypopituitarism, and papilledema.[6, 7] Approximately 75% of patients eventually suffer from cardiovascular disease, including hypertension, congestive heart failure, acute myocardial infarction, valvular dysfunction, pericardial infiltration, and cardiac tamponade.[8] Vascular involvement includes perivascular infiltration and periaortic fibrosis, resulting in the coated aorta seen in 20% of patients with ECD.[3] Pulmonary manifestations of ECD include interstitial, pleural, and consolidative lung disease. A review of high‐resolution chest CTs of patients with ECD demonstrated that greater than half had evidence of parenchymal lung disease, with interlobular septal thickening being the most common finding.[9] Infiltration and fibrosis of retroperitoneal structures is common. Infiltration of perinephric fat creates irregular renal borders, appearing radiographically as hairy kidneys.

Arriving at the diagnosis in this case proved to be challenging because the early presentation was consistent with congestive heart failure. As the patient's conditioned deteriorated, imaging suggested multisystem involvement. It was the extensive aortopathy in particularnot the less specific bone, kidney, lymph node, or pulmonary findingsthat allowed the clinicians to hone the extensive differential diagnosis. The coated aorta is a finding that has been strongly associated with ECD; few other conditions coat the aorta in a similar fashion.[10] In most mysteries, the perpetrator's coat conceals his identity; however, in this story the coat gave it away.

KEY LEARNING POINTS

1. Subacute, inflammatory, multiorgan disease is usually explained by 3 categoriesinfection, autoimmunity, and neoplasiabut a fourth category, infiltrative disorders, sometimes warrants consideration.
2. ECD presents heterogeneously, ranging from localized disease to widespread organ infiltration. The classic presentation includes bone pain, diabetes insipidus, and exophthalmos.
3. Characteristic radiological findings that suggest ECD include long bone osteosclerosis, a coated aorta from periaortic infiltration, and hairy kidneys from perinephric infiltration.

Disclosure

Nothing to report

Hand hygiene is a proven and guideline‐recommended safety practice, although clinicians and particularly physicians are unreliable at performing it.[1] Like hands, stethoscopes can carry pathogens from patient to patient. In 1 study, stethoscopes were as likely to be contaminated after use with methicillin‐resistant Staphylococcus aureuspositive patients as the provider's hands.[2] Furthermore, like hands, stethoscopes can be effectively decolonized with alcohol.[3, 4] However, although hand hygiene rates have been extensively studied,[1] and hand hygiene has been linked to reductions in nosocomial infection,[5] stethoscope hygiene is less well studied and emphasized less by guidelines.[6] Several surveys have documented low self‐reported compliance with stethoscope hygiene.[7, 8, 9, 10] Of 150 healthcare workers, 48% reported stethoscope hygiene between daily and weekly, 37% did stethoscope hygiene monthly, and 7% did stethoscope hygiene annually or never.[8] Of 1401 doctors asked about their stethoscope hygiene beliefs and practices, 76% believed that stethoscopes could transmit infection, but only 24% reported cleaning their scopes regularly.[9] Moreover, of 308 students, 22% had never done stethoscope hygiene, and <4% did it consistently.[10] However, we were unable to find any data on observed rates of stethoscope hygiene. Thus, we observed student and trainee physician stethoscope hygiene performance during hospital medicine rotations as part of the baseline data‐collection phase of a quality‐improvement effort linked to hand hygiene efforts.

METHODS

Attending hospitalists (I.H.J., B.M., and A.A.) and 1 graduate assistant (J.W.) at 3 sites observed stethoscope hygiene opportunities over an 11‐month period. Stethoscope hygiene was counted as performed if a patient‐specific stethoscope was used in an isolation room, or if any type of cleaning (alcohol gel, alcohol wipe, or cleansing cloth) was performed on a stethoscope carried out of the room. Observers also recorded whether stethoscope hygiene opportunities occurred in isolation rooms or nonisolation rooms, and noted if stethoscope hygiene was obviously triggered by an attending's stethoscope hygiene behavior (eg, a trainee asked an attending why he performed stethoscope hygiene, then performed it him or herself). Trainees were not aware that their stethoscope hygiene behaviors were being recorded.

RESULTS

We observed 352 opportunities for stethoscope hygiene, in which doctors or students used stethoscope hygiene in 58 encounters (16%). Twenty of the 58 stethoscope hygiene events occurred only after a trainee observed an attending physician perform stethoscope hygiene. Eliminating stethoscope hygiene events that were triggered by attending physicians, stethoscope hygiene was performed in 38 of 332 opportunities (11%). There was a significant difference between the rate of stethoscope hygiene performed in isolation versus nonisolation rooms: 24/29 (82.7%) versus 14 of 303 (4.6%) (P<0.001 by Pearson ² statistic). In isolation room stethoscope hygiene, in which the type of hygiene was recorded, 18 of 20 (90%) involved use of an isolation stethoscope, and 2 of 20 (10%) involved cleaning of a personal stethoscope.

DISCUSSION

Stethoscope hygiene is rarely performed by trainees. Stethoscope hygiene performance depends on the isolation status of the patient, with more than 80% performance in isolated patients and <5% in nonisolated patients.

Although little is known about the rate of infection related to stethoscopes, colonization of stethoscopes with nosocomial bacteria is well described.[2] Transmission of pathogens from patient to patient by stethoscopes could undermine the benefits of hand hygiene programs, as patients are commonly exposed to unclean stethoscopes.

Our observations are limited by several factors. We used a convenience sample of general medicine trainee behavior at academic medical centers; the behavior of attending physicians, ancillary staff, and nonacademic physicians may be different. Moreover, attending behavior may have prompted more episodes of stethoscope hygiene performance than we recorded, because we only noted when stethoscope hygiene was clearly related to attending behavior. The very low rate of stethoscope hygiene after contact with nonisolation patients represents a current and potentially serious safety threat. Future research might be able to quantify the risk associated with uncleaned stethoscopes or demonstrate the effectiveness of stethoscope hygiene programs. The effect of modeling on hand hygiene and stethoscope hygiene[10, 11] and on stethoscope hygiene in our data suggests a method for improving stethoscope hygiene.

Hand hygiene is a proven and guideline‐recommended safety practice, although clinicians and particularly physicians are unreliable at performing it.[1] Like hands, stethoscopes can carry pathogens from patient to patient. In 1 study, stethoscopes were as likely to be contaminated after use with methicillin‐resistant Staphylococcus aureuspositive patients as the provider's hands.[2] Furthermore, like hands, stethoscopes can be effectively decolonized with alcohol.[3, 4] However, although hand hygiene rates have been extensively studied,[1] and hand hygiene has been linked to reductions in nosocomial infection,[5] stethoscope hygiene is less well studied and emphasized less by guidelines.[6] Several surveys have documented low self‐reported compliance with stethoscope hygiene.[7, 8, 9, 10] Of 150 healthcare workers, 48% reported stethoscope hygiene between daily and weekly, 37% did stethoscope hygiene monthly, and 7% did stethoscope hygiene annually or never.[8] Of 1401 doctors asked about their stethoscope hygiene beliefs and practices, 76% believed that stethoscopes could transmit infection, but only 24% reported cleaning their scopes regularly.[9] Moreover, of 308 students, 22% had never done stethoscope hygiene, and <4% did it consistently.[10] However, we were unable to find any data on observed rates of stethoscope hygiene. Thus, we observed student and trainee physician stethoscope hygiene performance during hospital medicine rotations as part of the baseline data‐collection phase of a quality‐improvement effort linked to hand hygiene efforts.

METHODS

Attending hospitalists (I.H.J., B.M., and A.A.) and 1 graduate assistant (J.W.) at 3 sites observed stethoscope hygiene opportunities over an 11‐month period. Stethoscope hygiene was counted as performed if a patient‐specific stethoscope was used in an isolation room, or if any type of cleaning (alcohol gel, alcohol wipe, or cleansing cloth) was performed on a stethoscope carried out of the room. Observers also recorded whether stethoscope hygiene opportunities occurred in isolation rooms or nonisolation rooms, and noted if stethoscope hygiene was obviously triggered by an attending's stethoscope hygiene behavior (eg, a trainee asked an attending why he performed stethoscope hygiene, then performed it him or herself). Trainees were not aware that their stethoscope hygiene behaviors were being recorded.

RESULTS

We observed 352 opportunities for stethoscope hygiene, in which doctors or students used stethoscope hygiene in 58 encounters (16%). Twenty of the 58 stethoscope hygiene events occurred only after a trainee observed an attending physician perform stethoscope hygiene. Eliminating stethoscope hygiene events that were triggered by attending physicians, stethoscope hygiene was performed in 38 of 332 opportunities (11%). There was a significant difference between the rate of stethoscope hygiene performed in isolation versus nonisolation rooms: 24/29 (82.7%) versus 14 of 303 (4.6%) (P<0.001 by Pearson ² statistic). In isolation room stethoscope hygiene, in which the type of hygiene was recorded, 18 of 20 (90%) involved use of an isolation stethoscope, and 2 of 20 (10%) involved cleaning of a personal stethoscope.

DISCUSSION

Stethoscope hygiene is rarely performed by trainees. Stethoscope hygiene performance depends on the isolation status of the patient, with more than 80% performance in isolated patients and <5% in nonisolated patients.

Although little is known about the rate of infection related to stethoscopes, colonization of stethoscopes with nosocomial bacteria is well described.[2] Transmission of pathogens from patient to patient by stethoscopes could undermine the benefits of hand hygiene programs, as patients are commonly exposed to unclean stethoscopes.

Our observations are limited by several factors. We used a convenience sample of general medicine trainee behavior at academic medical centers; the behavior of attending physicians, ancillary staff, and nonacademic physicians may be different. Moreover, attending behavior may have prompted more episodes of stethoscope hygiene performance than we recorded, because we only noted when stethoscope hygiene was clearly related to attending behavior. The very low rate of stethoscope hygiene after contact with nonisolation patients represents a current and potentially serious safety threat. Future research might be able to quantify the risk associated with uncleaned stethoscopes or demonstrate the effectiveness of stethoscope hygiene programs. The effect of modeling on hand hygiene and stethoscope hygiene[10, 11] and on stethoscope hygiene in our data suggests a method for improving stethoscope hygiene.

Hand hygiene is a proven and guideline‐recommended safety practice, although clinicians and particularly physicians are unreliable at performing it.[1] Like hands, stethoscopes can carry pathogens from patient to patient. In 1 study, stethoscopes were as likely to be contaminated after use with methicillin‐resistant Staphylococcus aureuspositive patients as the provider's hands.[2] Furthermore, like hands, stethoscopes can be effectively decolonized with alcohol.[3, 4] However, although hand hygiene rates have been extensively studied,[1] and hand hygiene has been linked to reductions in nosocomial infection,[5] stethoscope hygiene is less well studied and emphasized less by guidelines.[6] Several surveys have documented low self‐reported compliance with stethoscope hygiene.[7, 8, 9, 10] Of 150 healthcare workers, 48% reported stethoscope hygiene between daily and weekly, 37% did stethoscope hygiene monthly, and 7% did stethoscope hygiene annually or never.[8] Of 1401 doctors asked about their stethoscope hygiene beliefs and practices, 76% believed that stethoscopes could transmit infection, but only 24% reported cleaning their scopes regularly.[9] Moreover, of 308 students, 22% had never done stethoscope hygiene, and <4% did it consistently.[10] However, we were unable to find any data on observed rates of stethoscope hygiene. Thus, we observed student and trainee physician stethoscope hygiene performance during hospital medicine rotations as part of the baseline data‐collection phase of a quality‐improvement effort linked to hand hygiene efforts.

METHODS

Attending hospitalists (I.H.J., B.M., and A.A.) and 1 graduate assistant (J.W.) at 3 sites observed stethoscope hygiene opportunities over an 11‐month period. Stethoscope hygiene was counted as performed if a patient‐specific stethoscope was used in an isolation room, or if any type of cleaning (alcohol gel, alcohol wipe, or cleansing cloth) was performed on a stethoscope carried out of the room. Observers also recorded whether stethoscope hygiene opportunities occurred in isolation rooms or nonisolation rooms, and noted if stethoscope hygiene was obviously triggered by an attending's stethoscope hygiene behavior (eg, a trainee asked an attending why he performed stethoscope hygiene, then performed it him or herself). Trainees were not aware that their stethoscope hygiene behaviors were being recorded.

RESULTS

We observed 352 opportunities for stethoscope hygiene, in which doctors or students used stethoscope hygiene in 58 encounters (16%). Twenty of the 58 stethoscope hygiene events occurred only after a trainee observed an attending physician perform stethoscope hygiene. Eliminating stethoscope hygiene events that were triggered by attending physicians, stethoscope hygiene was performed in 38 of 332 opportunities (11%). There was a significant difference between the rate of stethoscope hygiene performed in isolation versus nonisolation rooms: 24/29 (82.7%) versus 14 of 303 (4.6%) (P<0.001 by Pearson ² statistic). In isolation room stethoscope hygiene, in which the type of hygiene was recorded, 18 of 20 (90%) involved use of an isolation stethoscope, and 2 of 20 (10%) involved cleaning of a personal stethoscope.

DISCUSSION

Stethoscope hygiene is rarely performed by trainees. Stethoscope hygiene performance depends on the isolation status of the patient, with more than 80% performance in isolated patients and <5% in nonisolated patients.

Although little is known about the rate of infection related to stethoscopes, colonization of stethoscopes with nosocomial bacteria is well described.[2] Transmission of pathogens from patient to patient by stethoscopes could undermine the benefits of hand hygiene programs, as patients are commonly exposed to unclean stethoscopes.

Our observations are limited by several factors. We used a convenience sample of general medicine trainee behavior at academic medical centers; the behavior of attending physicians, ancillary staff, and nonacademic physicians may be different. Moreover, attending behavior may have prompted more episodes of stethoscope hygiene performance than we recorded, because we only noted when stethoscope hygiene was clearly related to attending behavior. The very low rate of stethoscope hygiene after contact with nonisolation patients represents a current and potentially serious safety threat. Future research might be able to quantify the risk associated with uncleaned stethoscopes or demonstrate the effectiveness of stethoscope hygiene programs. The effect of modeling on hand hygiene and stethoscope hygiene[10, 11] and on stethoscope hygiene in our data suggests a method for improving stethoscope hygiene.

Hospitalists are increasingly responsible for educating students and housestaff in internal medicine.[1] Because the quality of teaching is an important factor in learning,[2, 3, 4] leaders in medical education have expressed concern over the rapid shift of teaching responsibilities to this new group of educators.[5, 6, 7, 8] Moreover, recent changes in duty hour restrictions have strained both student and resident education,[9, 10] necessitating the optimization of inpatient teaching.[11, 12] Many hospitalists have recently finished residency and have not had formal training in clinical teaching. Collectively, most hospital medicine groups are early in their careers, have significant clinical obligations,[13] and may not have the bandwidth or expertise to provide faculty development for improving clinical teaching.

Rationally designed and theoretically sound faculty development to improve inpatient clinical teaching is required to meet this challenge. There are a limited number of reports describing faculty development focused on strengthening the teaching of hospitalists, and only 3 utilized direct observation and feedback, 1 of which involved peer observation in the clinical setting.[14, 15, 16] This 2011 report described a narrative method of peer observation and feedback but did not assess for efficacy of the program.[16] To our knowledge, there have been no studies of structured peer observation and feedback to optimize hospitalist attendings' teaching which have evaluated the efficacy of the intervention.

We developed a faculty development program based on peer observation and feedback based on actual teaching practices, using structured feedback anchored in validated and observable measures of effective teaching. We hypothesized that participation in the program would increase confidence in key teaching skills, increase confidence in the ability to give and receive peer feedback, and strengthen attitudes toward peer observation and feedback.

METHODS

Intervention

Using a modified Delphi approach, 5 medical education experts selected the 10 items that are most easily observable and salient to formal attending teaching rounds from the SFDP‐26 teaching assessment tool. A structured observation form was created, which included a checklist of the 10 selected items, space for note taking, and a template for narrative feedback (Figure 1).

Figure 1

We introduced the SFDP framework during a 2‐hour initial training session. Participants watched videos of teaching, learned to identify the 10 selected teaching behaviors, developed appropriate constructive and reinforcing comments, and practiced giving and receiving peer feedback.

Dyads were created on the basis of predetermined attending schedules. Participants were asked to observe and be observed twice during attending teaching rounds over the course of the academic year. Attending teaching rounds were defined as any preplanned didactic activity for ward teams. The structured observation forms were returned to the study coordinators after the observer had given feedback to the presenter. A copy of the feedback without the observer's notes was also given to each speaker. At the midpoint of the academic year, a refresher session was offered to reinforce those teaching behaviors that were the least frequently performed to date. All participants received a $50.00 Amazon.com gift card, and additional gift card incentives were offered to the dyads that first completed both observations.

RESULTS

DISCUSSION

Previous studies have shown that teaching skills are unlikely to improve without feedback,[28, 29, 30] yet feedback for hospitalists is usually limited to summative, end‐rotation evaluations from learners, disconnected from the teaching encounter. Our theory‐based, rationally designed peer observation and feedback program resulted in increased confidence in the ability to give feedback, receive feedback, and teach effectively. Participation did not result in negative attitudes toward giving and receiving feedback from colleagues. Participants self‐reported increased performance of important teaching behaviors. Most participants rated the program very highly, and endorsed improved teaching skills as a result of the program.

Our experience provides several lessons for other groups considering the implementation of peer feedback to strengthen teaching. First, we suggest that hospitalist groups may expect variable degrees of participation in a voluntary peer feedback program. In our program, 41% of eligible attendings did not participate. We did not specifically investigate why; we speculate that they may not have had the time, believed that their teaching skills were already strong, or they may have been daunted at the idea of peer review. It is also possible that participants were a self‐selected group who were the most motivated to strengthen their teaching. Second, we note the steep decline in the number of observations in the second half of the year. Informal assessment for reasons for the drop‐off suggested that after initial enthusiasm for the program, navigating the logistics of observing the same peer in the second half of the year proved to be prohibitive to many participants. Therefore, future versions of peer feedback programs may benefit from removing the dyad requirement and encouraging all participants to observe one another whenever possible.

With these lessons in mind, we believe that a peer observation program could be implemented by other hospital medicine groups. The program does not require extensive content expertise or senior faculty but does require engaged leadership and interested and motivated faculty. Groups could identify an individual in their group with an interest in clinical teaching who could then be responsible for creating the training session (materials available upon request). We believe that with only a small upfront investment, most hospital medicine groups could use this as a model to build a peer observation program aimed at improving clinical teaching.

Our study has several limitations. As noted above, our participation rate was 59%, and the number of participating attendings declined through the year. We did not examine whether our program resulted in advances in the knowledge, skills, or attitudes of the learners; because each attending teaching session was unique, it was not possible to measure changes in learner knowledge. Our primary outcome measures relied on self‐assessment rather than higher order and more objective measures of teaching efficacy. Furthermore, our results may not be generalizable to other programs, given the heterogeneity in service structures and teaching practices across the country. This was an uncontrolled study; some of the outcomes may have naturally occurred independent of the intervention due to the natural evolution of clinical teaching. As with any educational intervention that integrates multiple strategies, we are not able to discern if the improved outcomes were the result of the initial didactic sessions, the refresher sessions, or the peer feedback itself. Serial assessments of frequency of teaching behaviors were not done due to the low number of observations in the second half of the program. Finally, our 10‐item tool derived from the validated SFDP‐26 tool is not itself a validated assessment of teaching.

We acknowledge that the increased confidence seen in our participants does not necessarily predict improved performance. Although increased confidence in core skills is a necessary step that can lead to changes in behavior, further studies are needed to determine whether the increase in faculty confidence that results from peer observation and feedback translates into improved educational outcomes.

The pressure on hospitalists to be excellent teachers is here to stay. Resources to train these faculty are scarce, yet we must prioritize faculty development in teaching to optimize the training of future physicians. Our data illustrate the benefits of peer observation and feedback. Hospitalist programs should consider this option in addressing the professional development needs of their faculty.

Hospitalists are increasingly responsible for educating students and housestaff in internal medicine.[1] Because the quality of teaching is an important factor in learning,[2, 3, 4] leaders in medical education have expressed concern over the rapid shift of teaching responsibilities to this new group of educators.[5, 6, 7, 8] Moreover, recent changes in duty hour restrictions have strained both student and resident education,[9, 10] necessitating the optimization of inpatient teaching.[11, 12] Many hospitalists have recently finished residency and have not had formal training in clinical teaching. Collectively, most hospital medicine groups are early in their careers, have significant clinical obligations,[13] and may not have the bandwidth or expertise to provide faculty development for improving clinical teaching.

Rationally designed and theoretically sound faculty development to improve inpatient clinical teaching is required to meet this challenge. There are a limited number of reports describing faculty development focused on strengthening the teaching of hospitalists, and only 3 utilized direct observation and feedback, 1 of which involved peer observation in the clinical setting.[14, 15, 16] This 2011 report described a narrative method of peer observation and feedback but did not assess for efficacy of the program.[16] To our knowledge, there have been no studies of structured peer observation and feedback to optimize hospitalist attendings' teaching which have evaluated the efficacy of the intervention.

We developed a faculty development program based on peer observation and feedback based on actual teaching practices, using structured feedback anchored in validated and observable measures of effective teaching. We hypothesized that participation in the program would increase confidence in key teaching skills, increase confidence in the ability to give and receive peer feedback, and strengthen attitudes toward peer observation and feedback.

METHODS

Intervention

Using a modified Delphi approach, 5 medical education experts selected the 10 items that are most easily observable and salient to formal attending teaching rounds from the SFDP‐26 teaching assessment tool. A structured observation form was created, which included a checklist of the 10 selected items, space for note taking, and a template for narrative feedback (Figure 1).

Figure 1

We introduced the SFDP framework during a 2‐hour initial training session. Participants watched videos of teaching, learned to identify the 10 selected teaching behaviors, developed appropriate constructive and reinforcing comments, and practiced giving and receiving peer feedback.

Dyads were created on the basis of predetermined attending schedules. Participants were asked to observe and be observed twice during attending teaching rounds over the course of the academic year. Attending teaching rounds were defined as any preplanned didactic activity for ward teams. The structured observation forms were returned to the study coordinators after the observer had given feedback to the presenter. A copy of the feedback without the observer's notes was also given to each speaker. At the midpoint of the academic year, a refresher session was offered to reinforce those teaching behaviors that were the least frequently performed to date. All participants received a $50.00 Amazon.com gift card, and additional gift card incentives were offered to the dyads that first completed both observations.

RESULTS

DISCUSSION

Previous studies have shown that teaching skills are unlikely to improve without feedback,[28, 29, 30] yet feedback for hospitalists is usually limited to summative, end‐rotation evaluations from learners, disconnected from the teaching encounter. Our theory‐based, rationally designed peer observation and feedback program resulted in increased confidence in the ability to give feedback, receive feedback, and teach effectively. Participation did not result in negative attitudes toward giving and receiving feedback from colleagues. Participants self‐reported increased performance of important teaching behaviors. Most participants rated the program very highly, and endorsed improved teaching skills as a result of the program.

Our experience provides several lessons for other groups considering the implementation of peer feedback to strengthen teaching. First, we suggest that hospitalist groups may expect variable degrees of participation in a voluntary peer feedback program. In our program, 41% of eligible attendings did not participate. We did not specifically investigate why; we speculate that they may not have had the time, believed that their teaching skills were already strong, or they may have been daunted at the idea of peer review. It is also possible that participants were a self‐selected group who were the most motivated to strengthen their teaching. Second, we note the steep decline in the number of observations in the second half of the year. Informal assessment for reasons for the drop‐off suggested that after initial enthusiasm for the program, navigating the logistics of observing the same peer in the second half of the year proved to be prohibitive to many participants. Therefore, future versions of peer feedback programs may benefit from removing the dyad requirement and encouraging all participants to observe one another whenever possible.

With these lessons in mind, we believe that a peer observation program could be implemented by other hospital medicine groups. The program does not require extensive content expertise or senior faculty but does require engaged leadership and interested and motivated faculty. Groups could identify an individual in their group with an interest in clinical teaching who could then be responsible for creating the training session (materials available upon request). We believe that with only a small upfront investment, most hospital medicine groups could use this as a model to build a peer observation program aimed at improving clinical teaching.

Our study has several limitations. As noted above, our participation rate was 59%, and the number of participating attendings declined through the year. We did not examine whether our program resulted in advances in the knowledge, skills, or attitudes of the learners; because each attending teaching session was unique, it was not possible to measure changes in learner knowledge. Our primary outcome measures relied on self‐assessment rather than higher order and more objective measures of teaching efficacy. Furthermore, our results may not be generalizable to other programs, given the heterogeneity in service structures and teaching practices across the country. This was an uncontrolled study; some of the outcomes may have naturally occurred independent of the intervention due to the natural evolution of clinical teaching. As with any educational intervention that integrates multiple strategies, we are not able to discern if the improved outcomes were the result of the initial didactic sessions, the refresher sessions, or the peer feedback itself. Serial assessments of frequency of teaching behaviors were not done due to the low number of observations in the second half of the program. Finally, our 10‐item tool derived from the validated SFDP‐26 tool is not itself a validated assessment of teaching.

We acknowledge that the increased confidence seen in our participants does not necessarily predict improved performance. Although increased confidence in core skills is a necessary step that can lead to changes in behavior, further studies are needed to determine whether the increase in faculty confidence that results from peer observation and feedback translates into improved educational outcomes.

The pressure on hospitalists to be excellent teachers is here to stay. Resources to train these faculty are scarce, yet we must prioritize faculty development in teaching to optimize the training of future physicians. Our data illustrate the benefits of peer observation and feedback. Hospitalist programs should consider this option in addressing the professional development needs of their faculty.

Hospitalists are increasingly responsible for educating students and housestaff in internal medicine.[1] Because the quality of teaching is an important factor in learning,[2, 3, 4] leaders in medical education have expressed concern over the rapid shift of teaching responsibilities to this new group of educators.[5, 6, 7, 8] Moreover, recent changes in duty hour restrictions have strained both student and resident education,[9, 10] necessitating the optimization of inpatient teaching.[11, 12] Many hospitalists have recently finished residency and have not had formal training in clinical teaching. Collectively, most hospital medicine groups are early in their careers, have significant clinical obligations,[13] and may not have the bandwidth or expertise to provide faculty development for improving clinical teaching.

Rationally designed and theoretically sound faculty development to improve inpatient clinical teaching is required to meet this challenge. There are a limited number of reports describing faculty development focused on strengthening the teaching of hospitalists, and only 3 utilized direct observation and feedback, 1 of which involved peer observation in the clinical setting.[14, 15, 16] This 2011 report described a narrative method of peer observation and feedback but did not assess for efficacy of the program.[16] To our knowledge, there have been no studies of structured peer observation and feedback to optimize hospitalist attendings' teaching which have evaluated the efficacy of the intervention.

We developed a faculty development program based on peer observation and feedback based on actual teaching practices, using structured feedback anchored in validated and observable measures of effective teaching. We hypothesized that participation in the program would increase confidence in key teaching skills, increase confidence in the ability to give and receive peer feedback, and strengthen attitudes toward peer observation and feedback.

METHODS

Intervention

Using a modified Delphi approach, 5 medical education experts selected the 10 items that are most easily observable and salient to formal attending teaching rounds from the SFDP‐26 teaching assessment tool. A structured observation form was created, which included a checklist of the 10 selected items, space for note taking, and a template for narrative feedback (Figure 1).

Figure 1

We introduced the SFDP framework during a 2‐hour initial training session. Participants watched videos of teaching, learned to identify the 10 selected teaching behaviors, developed appropriate constructive and reinforcing comments, and practiced giving and receiving peer feedback.

Dyads were created on the basis of predetermined attending schedules. Participants were asked to observe and be observed twice during attending teaching rounds over the course of the academic year. Attending teaching rounds were defined as any preplanned didactic activity for ward teams. The structured observation forms were returned to the study coordinators after the observer had given feedback to the presenter. A copy of the feedback without the observer's notes was also given to each speaker. At the midpoint of the academic year, a refresher session was offered to reinforce those teaching behaviors that were the least frequently performed to date. All participants received a $50.00 Amazon.com gift card, and additional gift card incentives were offered to the dyads that first completed both observations.

RESULTS

DISCUSSION

Previous studies have shown that teaching skills are unlikely to improve without feedback,[28, 29, 30] yet feedback for hospitalists is usually limited to summative, end‐rotation evaluations from learners, disconnected from the teaching encounter. Our theory‐based, rationally designed peer observation and feedback program resulted in increased confidence in the ability to give feedback, receive feedback, and teach effectively. Participation did not result in negative attitudes toward giving and receiving feedback from colleagues. Participants self‐reported increased performance of important teaching behaviors. Most participants rated the program very highly, and endorsed improved teaching skills as a result of the program.

Our experience provides several lessons for other groups considering the implementation of peer feedback to strengthen teaching. First, we suggest that hospitalist groups may expect variable degrees of participation in a voluntary peer feedback program. In our program, 41% of eligible attendings did not participate. We did not specifically investigate why; we speculate that they may not have had the time, believed that their teaching skills were already strong, or they may have been daunted at the idea of peer review. It is also possible that participants were a self‐selected group who were the most motivated to strengthen their teaching. Second, we note the steep decline in the number of observations in the second half of the year. Informal assessment for reasons for the drop‐off suggested that after initial enthusiasm for the program, navigating the logistics of observing the same peer in the second half of the year proved to be prohibitive to many participants. Therefore, future versions of peer feedback programs may benefit from removing the dyad requirement and encouraging all participants to observe one another whenever possible.

With these lessons in mind, we believe that a peer observation program could be implemented by other hospital medicine groups. The program does not require extensive content expertise or senior faculty but does require engaged leadership and interested and motivated faculty. Groups could identify an individual in their group with an interest in clinical teaching who could then be responsible for creating the training session (materials available upon request). We believe that with only a small upfront investment, most hospital medicine groups could use this as a model to build a peer observation program aimed at improving clinical teaching.

Our study has several limitations. As noted above, our participation rate was 59%, and the number of participating attendings declined through the year. We did not examine whether our program resulted in advances in the knowledge, skills, or attitudes of the learners; because each attending teaching session was unique, it was not possible to measure changes in learner knowledge. Our primary outcome measures relied on self‐assessment rather than higher order and more objective measures of teaching efficacy. Furthermore, our results may not be generalizable to other programs, given the heterogeneity in service structures and teaching practices across the country. This was an uncontrolled study; some of the outcomes may have naturally occurred independent of the intervention due to the natural evolution of clinical teaching. As with any educational intervention that integrates multiple strategies, we are not able to discern if the improved outcomes were the result of the initial didactic sessions, the refresher sessions, or the peer feedback itself. Serial assessments of frequency of teaching behaviors were not done due to the low number of observations in the second half of the program. Finally, our 10‐item tool derived from the validated SFDP‐26 tool is not itself a validated assessment of teaching.

We acknowledge that the increased confidence seen in our participants does not necessarily predict improved performance. Although increased confidence in core skills is a necessary step that can lead to changes in behavior, further studies are needed to determine whether the increase in faculty confidence that results from peer observation and feedback translates into improved educational outcomes.

The pressure on hospitalists to be excellent teachers is here to stay. Resources to train these faculty are scarce, yet we must prioritize faculty development in teaching to optimize the training of future physicians. Our data illustrate the benefits of peer observation and feedback. Hospitalist programs should consider this option in addressing the professional development needs of their faculty.