Darius R. Mehregan, MD

To the Editor:

Palisaded neutrophilic granulomatous dermatitis (PNGD) is a rare disorder that often is associated with systemic disease. It has been shown to manifest in the presence of systemic lupus erythematosus; rheumatoid arthritis; Wegener granulomatosis; and other diseases, mainly autoimmune conditions. Interstitial granulomatous dermatitis (IGD) associated with arthritis was first described by Ackerman et al¹ in 1993. In 1994, IGD was placed among the spectrum of PNGD by Chu et al.² The disease entities included in the spectrum of PNGD of the immune complex disease are Churg-Strauss granuloma, cutaneous extravascular necrotizing granuloma, rheumatoid papules, superficial ulcerating rheumatoid necrobiosis, and IGD with arthritis.² It has been suggested that IGD has a distinct clinical presentation with associated histopathology, while others suggest it still is part of the PNGD spectrum.^2,3 We present 2 cases of granulomatous dermatitis and their findings related to IGD and PNGD.

A 58-year-old woman presented with recurrent painful lesions on the trunk, arms, and legs of 2 years’ duration. The lesions spontaneously resolved without scarring or hyperpigmentation but would recur in different areas on the trunk. She was diagnosed with rheumatoid arthritis following a recent autoimmune workup. At presentation, physical examination revealed tender erythematous edematous plaques on the bilateral upper back (Figure 1) and erythematous nodules on the bilateral upper arms. The patient previously had an antinuclear antibody titer of 1:320 with a speckled pattern. A repeat antinuclear antibody titer taken 1 year later was negative. Her rheumatoid factor initially was positive and remained positive upon repeat testing. Punch biopsies were performed for histologic evaluation of the lesions and immunofluorescence. Biopsies examined with hematoxylin and eosin stain revealed perivascular and interstitial mixed (lymphocytic, neutrophilic, eosinophilic) bottom-heavy inflammation with nuclear dust and basophilic degeneration of collagen (Figure 2). Immunofluorescence studies were negative. The patient deferred treatment.

A 74-year-old man presented with a rash on the flank and back with associated pruritus and occasional pain of 2 months’ duration. His primary care physician prescribed a course of cephalexin, but the rash did not improve. Review of systems was positive for intermittent swelling of the hands, feet, and lips, and negative for arthritis. His medical history included 2 episodes of rheumatic fever, one complicated by pneumonia. His medications included finasteride, simvastatin, bisoprolol-hydrochlorothiazide, aspirin, tiotropium, vitamin D, and fish oil. At presentation, physical examination revealed tender violaceous plaques with induration and central clearing distributed on the left side of the back, left side of the flank, and left axilla. The lesion on the axilla measured 30.0×3.5 cm and the lesions on the left side of the back measured 30.0×9.0 cm. The rims of the lesions were elevated and consistent with the rope sign (Figure 3). A punch biopsy of the lesion on the left axilla showed perivascular and interstitial infiltrate of lymphocytes, neutrophils, histiocytes, and eosinophils. There was no evidence of fibrin deposition in the blood vessels. Small areas of necrobiotic collagen surrounded by multinucleated giant cells and lymphocytes were noted (Figure 4). The rash improved spontaneously at the time of suture removal. No treatment was initiated.

Granulomatous dermatitis in the presence of an autoimmune disorder can present as IGD or PNGD. Both forms of granulomatous dermatitis are rare conditions and considered to be part of the same clinicopathological spectrum. These conditions can be difficult to distinguish clinically but are histologically unique.

Interstitial granulomatous dermatitis and PNGD can have a variable clinical expression. Palisaded neutrophilic granulomatous dermatitis generally presents as flesh-colored to erythematous papules or plaques, most commonly located on the upper arms. The lesions may have a central umbilication with perforation and ulceration.⁴ Interstitial granulomatous dermatitis most commonly presents as erythematous plaques and papules. The lesions are symmetric and asymptomatic. They most commonly appear on the trunk, axillae, buttocks, thighs, and groin. Subcutaneous linear cords (the rope sign) is a characteristic associated with IGD.^3,5 However, the rope sign also has been reported in a patient with PNGD with systemic lupus,⁶ which further demonstrates the overlapping spectrum of clinical expression seen in these 2 forms of granulomatous dermatitis. Therefore, a diagnosis cannot be made by clinical expression alone; histologic findings are needed for confirmation.

When differentiating IGD and PNGD histologically, it is important to keep in mind that these features exist on a spectrum and depend on the age of the lesion. Deposition of the immune complex around the dermal blood vessel initiates the pathogenesis. Early lesions of PNGD show a neutrophilic infiltrate, focal leukocytoclastic vasculitis, and dense nuclear dust. Developed lesions show zones of basophilic degenerated collagen surrounded by palisades of histiocytes, neutrophils, and nuclear debris.² The histologic pattern of IGD features smaller areas of palisading histiocytes surrounding foci of degenerated collagen. Neutrophils and eosinophils are seen among the degenerated collagen. There is no evidence of vasculitis and dermal mucin usually is absent.⁷

Palisaded neutrophilic granulomatous dermatitis has been reported to improve with systemic steroids and dapsone.⁸ The lesions may resolve spontaneously and with treatment of the underlying systemic disease. Similarly, IGD has been reported to resolve with systemic or topical steroids.^3,5 The rash in our patient with IGD (patient 2) spontaneously resolved. Our patient with PNGD (patient 1) reported that her lesions would spontaneously resolve, then recur. She had deferred all other treatment.

Some authors have disputed the spectrum that Chu et al² had determined in their study and proposed IGD is a separate entity from the PNGD spectrum. Verneuil et al⁹ stated that the clinical presentations in Chu et al’s² study (symmetric papules of the extremities) had not been reported in a patient with IGD. However, in a study of IGD by Peroni et al,³ 7 of 12 patients presented with symmetrical papules of the extremities. We believe that the spectrum proposed by Chu et al² still holds true.

These 2 reports demonstrate the diverse presentation of IGD and PNGD. It is important for dermatologists to keep in mind the PNGD spectrum when a patient presents with granulomatous dermatitis in the presence of an autoimmune disorder.

To the Editor:

Palisaded neutrophilic granulomatous dermatitis (PNGD) is a rare disorder that often is associated with systemic disease. It has been shown to manifest in the presence of systemic lupus erythematosus; rheumatoid arthritis; Wegener granulomatosis; and other diseases, mainly autoimmune conditions. Interstitial granulomatous dermatitis (IGD) associated with arthritis was first described by Ackerman et al¹ in 1993. In 1994, IGD was placed among the spectrum of PNGD by Chu et al.² The disease entities included in the spectrum of PNGD of the immune complex disease are Churg-Strauss granuloma, cutaneous extravascular necrotizing granuloma, rheumatoid papules, superficial ulcerating rheumatoid necrobiosis, and IGD with arthritis.² It has been suggested that IGD has a distinct clinical presentation with associated histopathology, while others suggest it still is part of the PNGD spectrum.^2,3 We present 2 cases of granulomatous dermatitis and their findings related to IGD and PNGD.

A 58-year-old woman presented with recurrent painful lesions on the trunk, arms, and legs of 2 years’ duration. The lesions spontaneously resolved without scarring or hyperpigmentation but would recur in different areas on the trunk. She was diagnosed with rheumatoid arthritis following a recent autoimmune workup. At presentation, physical examination revealed tender erythematous edematous plaques on the bilateral upper back (Figure 1) and erythematous nodules on the bilateral upper arms. The patient previously had an antinuclear antibody titer of 1:320 with a speckled pattern. A repeat antinuclear antibody titer taken 1 year later was negative. Her rheumatoid factor initially was positive and remained positive upon repeat testing. Punch biopsies were performed for histologic evaluation of the lesions and immunofluorescence. Biopsies examined with hematoxylin and eosin stain revealed perivascular and interstitial mixed (lymphocytic, neutrophilic, eosinophilic) bottom-heavy inflammation with nuclear dust and basophilic degeneration of collagen (Figure 2). Immunofluorescence studies were negative. The patient deferred treatment.

A 74-year-old man presented with a rash on the flank and back with associated pruritus and occasional pain of 2 months’ duration. His primary care physician prescribed a course of cephalexin, but the rash did not improve. Review of systems was positive for intermittent swelling of the hands, feet, and lips, and negative for arthritis. His medical history included 2 episodes of rheumatic fever, one complicated by pneumonia. His medications included finasteride, simvastatin, bisoprolol-hydrochlorothiazide, aspirin, tiotropium, vitamin D, and fish oil. At presentation, physical examination revealed tender violaceous plaques with induration and central clearing distributed on the left side of the back, left side of the flank, and left axilla. The lesion on the axilla measured 30.0×3.5 cm and the lesions on the left side of the back measured 30.0×9.0 cm. The rims of the lesions were elevated and consistent with the rope sign (Figure 3). A punch biopsy of the lesion on the left axilla showed perivascular and interstitial infiltrate of lymphocytes, neutrophils, histiocytes, and eosinophils. There was no evidence of fibrin deposition in the blood vessels. Small areas of necrobiotic collagen surrounded by multinucleated giant cells and lymphocytes were noted (Figure 4). The rash improved spontaneously at the time of suture removal. No treatment was initiated.

Granulomatous dermatitis in the presence of an autoimmune disorder can present as IGD or PNGD. Both forms of granulomatous dermatitis are rare conditions and considered to be part of the same clinicopathological spectrum. These conditions can be difficult to distinguish clinically but are histologically unique.

Interstitial granulomatous dermatitis and PNGD can have a variable clinical expression. Palisaded neutrophilic granulomatous dermatitis generally presents as flesh-colored to erythematous papules or plaques, most commonly located on the upper arms. The lesions may have a central umbilication with perforation and ulceration.⁴ Interstitial granulomatous dermatitis most commonly presents as erythematous plaques and papules. The lesions are symmetric and asymptomatic. They most commonly appear on the trunk, axillae, buttocks, thighs, and groin. Subcutaneous linear cords (the rope sign) is a characteristic associated with IGD.^3,5 However, the rope sign also has been reported in a patient with PNGD with systemic lupus,⁶ which further demonstrates the overlapping spectrum of clinical expression seen in these 2 forms of granulomatous dermatitis. Therefore, a diagnosis cannot be made by clinical expression alone; histologic findings are needed for confirmation.

When differentiating IGD and PNGD histologically, it is important to keep in mind that these features exist on a spectrum and depend on the age of the lesion. Deposition of the immune complex around the dermal blood vessel initiates the pathogenesis. Early lesions of PNGD show a neutrophilic infiltrate, focal leukocytoclastic vasculitis, and dense nuclear dust. Developed lesions show zones of basophilic degenerated collagen surrounded by palisades of histiocytes, neutrophils, and nuclear debris.² The histologic pattern of IGD features smaller areas of palisading histiocytes surrounding foci of degenerated collagen. Neutrophils and eosinophils are seen among the degenerated collagen. There is no evidence of vasculitis and dermal mucin usually is absent.⁷

Palisaded neutrophilic granulomatous dermatitis has been reported to improve with systemic steroids and dapsone.⁸ The lesions may resolve spontaneously and with treatment of the underlying systemic disease. Similarly, IGD has been reported to resolve with systemic or topical steroids.^3,5 The rash in our patient with IGD (patient 2) spontaneously resolved. Our patient with PNGD (patient 1) reported that her lesions would spontaneously resolve, then recur. She had deferred all other treatment.

Some authors have disputed the spectrum that Chu et al² had determined in their study and proposed IGD is a separate entity from the PNGD spectrum. Verneuil et al⁹ stated that the clinical presentations in Chu et al’s² study (symmetric papules of the extremities) had not been reported in a patient with IGD. However, in a study of IGD by Peroni et al,³ 7 of 12 patients presented with symmetrical papules of the extremities. We believe that the spectrum proposed by Chu et al² still holds true.

These 2 reports demonstrate the diverse presentation of IGD and PNGD. It is important for dermatologists to keep in mind the PNGD spectrum when a patient presents with granulomatous dermatitis in the presence of an autoimmune disorder.

To the Editor:

Palisaded neutrophilic granulomatous dermatitis (PNGD) is a rare disorder that often is associated with systemic disease. It has been shown to manifest in the presence of systemic lupus erythematosus; rheumatoid arthritis; Wegener granulomatosis; and other diseases, mainly autoimmune conditions. Interstitial granulomatous dermatitis (IGD) associated with arthritis was first described by Ackerman et al¹ in 1993. In 1994, IGD was placed among the spectrum of PNGD by Chu et al.² The disease entities included in the spectrum of PNGD of the immune complex disease are Churg-Strauss granuloma, cutaneous extravascular necrotizing granuloma, rheumatoid papules, superficial ulcerating rheumatoid necrobiosis, and IGD with arthritis.² It has been suggested that IGD has a distinct clinical presentation with associated histopathology, while others suggest it still is part of the PNGD spectrum.^2,3 We present 2 cases of granulomatous dermatitis and their findings related to IGD and PNGD.

A 58-year-old woman presented with recurrent painful lesions on the trunk, arms, and legs of 2 years’ duration. The lesions spontaneously resolved without scarring or hyperpigmentation but would recur in different areas on the trunk. She was diagnosed with rheumatoid arthritis following a recent autoimmune workup. At presentation, physical examination revealed tender erythematous edematous plaques on the bilateral upper back (Figure 1) and erythematous nodules on the bilateral upper arms. The patient previously had an antinuclear antibody titer of 1:320 with a speckled pattern. A repeat antinuclear antibody titer taken 1 year later was negative. Her rheumatoid factor initially was positive and remained positive upon repeat testing. Punch biopsies were performed for histologic evaluation of the lesions and immunofluorescence. Biopsies examined with hematoxylin and eosin stain revealed perivascular and interstitial mixed (lymphocytic, neutrophilic, eosinophilic) bottom-heavy inflammation with nuclear dust and basophilic degeneration of collagen (Figure 2). Immunofluorescence studies were negative. The patient deferred treatment.

A 74-year-old man presented with a rash on the flank and back with associated pruritus and occasional pain of 2 months’ duration. His primary care physician prescribed a course of cephalexin, but the rash did not improve. Review of systems was positive for intermittent swelling of the hands, feet, and lips, and negative for arthritis. His medical history included 2 episodes of rheumatic fever, one complicated by pneumonia. His medications included finasteride, simvastatin, bisoprolol-hydrochlorothiazide, aspirin, tiotropium, vitamin D, and fish oil. At presentation, physical examination revealed tender violaceous plaques with induration and central clearing distributed on the left side of the back, left side of the flank, and left axilla. The lesion on the axilla measured 30.0×3.5 cm and the lesions on the left side of the back measured 30.0×9.0 cm. The rims of the lesions were elevated and consistent with the rope sign (Figure 3). A punch biopsy of the lesion on the left axilla showed perivascular and interstitial infiltrate of lymphocytes, neutrophils, histiocytes, and eosinophils. There was no evidence of fibrin deposition in the blood vessels. Small areas of necrobiotic collagen surrounded by multinucleated giant cells and lymphocytes were noted (Figure 4). The rash improved spontaneously at the time of suture removal. No treatment was initiated.

Granulomatous dermatitis in the presence of an autoimmune disorder can present as IGD or PNGD. Both forms of granulomatous dermatitis are rare conditions and considered to be part of the same clinicopathological spectrum. These conditions can be difficult to distinguish clinically but are histologically unique.

Interstitial granulomatous dermatitis and PNGD can have a variable clinical expression. Palisaded neutrophilic granulomatous dermatitis generally presents as flesh-colored to erythematous papules or plaques, most commonly located on the upper arms. The lesions may have a central umbilication with perforation and ulceration.⁴ Interstitial granulomatous dermatitis most commonly presents as erythematous plaques and papules. The lesions are symmetric and asymptomatic. They most commonly appear on the trunk, axillae, buttocks, thighs, and groin. Subcutaneous linear cords (the rope sign) is a characteristic associated with IGD.^3,5 However, the rope sign also has been reported in a patient with PNGD with systemic lupus,⁶ which further demonstrates the overlapping spectrum of clinical expression seen in these 2 forms of granulomatous dermatitis. Therefore, a diagnosis cannot be made by clinical expression alone; histologic findings are needed for confirmation.

When differentiating IGD and PNGD histologically, it is important to keep in mind that these features exist on a spectrum and depend on the age of the lesion. Deposition of the immune complex around the dermal blood vessel initiates the pathogenesis. Early lesions of PNGD show a neutrophilic infiltrate, focal leukocytoclastic vasculitis, and dense nuclear dust. Developed lesions show zones of basophilic degenerated collagen surrounded by palisades of histiocytes, neutrophils, and nuclear debris.² The histologic pattern of IGD features smaller areas of palisading histiocytes surrounding foci of degenerated collagen. Neutrophils and eosinophils are seen among the degenerated collagen. There is no evidence of vasculitis and dermal mucin usually is absent.⁷

Palisaded neutrophilic granulomatous dermatitis has been reported to improve with systemic steroids and dapsone.⁸ The lesions may resolve spontaneously and with treatment of the underlying systemic disease. Similarly, IGD has been reported to resolve with systemic or topical steroids.^3,5 The rash in our patient with IGD (patient 2) spontaneously resolved. Our patient with PNGD (patient 1) reported that her lesions would spontaneously resolve, then recur. She had deferred all other treatment.

Some authors have disputed the spectrum that Chu et al² had determined in their study and proposed IGD is a separate entity from the PNGD spectrum. Verneuil et al⁹ stated that the clinical presentations in Chu et al’s² study (symmetric papules of the extremities) had not been reported in a patient with IGD. However, in a study of IGD by Peroni et al,³ 7 of 12 patients presented with symmetrical papules of the extremities. We believe that the spectrum proposed by Chu et al² still holds true.

These 2 reports demonstrate the diverse presentation of IGD and PNGD. It is important for dermatologists to keep in mind the PNGD spectrum when a patient presents with granulomatous dermatitis in the presence of an autoimmune disorder.

The term blueberry muffin rash historically was used to describe the cutaneous manifestations observed in congenital rubella. The term traditionally describes the result of a postnatal dermal extramedullary hematopoiesis. Today, TORCH (toxoplasmosis, other agents, rubella, cytomegalovirus, herpes) infections and plasma cell dyscrasias are all potential causes of extramedullary hematopoiesis. Herein, we present a unique case of a neonate born with a blueberry muffin rash secondary to extramedullary hematopoiesis induced by hereditary spherocytosis.

Case Report

The dermatology department was consulted to evaluate a 2-day-old male neonate born with a “rash.” The patient was born to a 34-year-old gravida 3, para 2, woman at 39 weeks’ gestation. The mother’s prenatal laboratory values were within reference range and ultrasounds were normal, and she was compliant with her prenatal care. She underwent a normal spontaneous vaginal delivery 3 hours after rupture of membranes without complication. The amniotic fluid and umbilical cord both were clear. There was no use of forceps or any other external aiding devices during the delivery. At the time of delivery, the consulting physician noted that the patient had “skin lesions from head to toe.”

The patient’s parents reported that the rash did not seem to cause any discomfort for the patient. In the 24 hours after birth, the parents reported that the erythema seemed to slightly fade. Physical examination revealed many scattered erythematous to violaceous, nonblanching papulonodules affecting the scalp (Figure 1), face, arms, hands (Figure 2A), back (Figure 2B), buttocks, legs, and feet. Some of the papulonodules were soft while others were firm and indurated. Several lesions had a yellowish hue with some overlying crust. There was no mucosal, genital, or ocular involvement. No erosions, ulcerations, petechiae, ecchymoses, or hepatosplenomegaly were noted on examination.

The patient was otherwise healthy with an Apgar score of 8/9 at 1 and 5 minutes. His birth weight, length, and head circumference were within normal limits. There was no evidence of ABO blood group or Rhesus factor incompatibility. His temperature, vital signs, laboratory values (including calcium level and TORCH titers, which included cytomegalovirus, rubella, toxoplasmosis, and herpes simplex virus), and review of systems all were within reference range. A bone survey of the skull, spine, ribs, arms, pelvis, legs, and feet was within normal limits.

The mother’s placenta was sent for pathology and revealed a lymphoplasmacytic chronic deciduitis and acute subchorionitis consistent with a nonspecific inflammatory response, unlikely to be from an infectious etiology.

A 4-mm punch biopsy was taken from the left thigh and revealed a predominately lymphocytic infiltrate with rare eosinophils and erythrocyte precursors (Figure 3). Immunohistochemical staining was performed showing that the majority of the lymphocytes represented T lymphocytes, which stained positive for CD45 and CD3 and negative for S-100, CD1a, CD30, and CD117. There were scattered CD34⁺ cells, and scattered cells stained positive for myeloperoxidase. No significant CD20 immunoreactivity was noted. There were scattered eosinophils and rare normoblasts but no megakaryocytes. A complete blood cell count (CBC) with differential and reticulocyte count was within reference range.

At 1-, 3-, 8-, 12-, and 28-week follow-up visits, the patient continued to grow and feed appropriately. No new lesions developed during this time, and the preexisting lesions continued to fade into slightly hyperpigmented patches without induration (Figure 4). At 6 months of age, a CBC performed at the time of an upper respiratory infection and otitis media revealed normocytic anemia with a hemoglobin level of 9.9 g/dL (reference range, 14.0–17.5 g/dL), a reticulocyte count of 0.8% (reference range, 0.5%–1.5%), and a lactate dehydrogenase level of 424 U/L (reference range, 100–200 U/L). All red blood cell (RBC) indices were within reference range. Flow cytometry, eosin-5-maleimide, and ektacytometry were performed with results consistent with mild hereditary spherocytosis.

Comment

Dermal extramedullary hematopoiesis is a normal component of embryologic development up until the fifth month of gestation.¹ The term blueberry muffin rash typically is used to describe the cutaneous manifestations of extramedullary hematopoiesis, which commonly is caused by a TORCH infection or hematologic dyscrasia.² It has been suggested that the term be expanded to include neoplastic processes (eg, neuroblastomas) and vascular processes (eg, multiple hemangiomas, blue rubber bleb nevus syndrome, glomangiomas, multifocal lymphangionendotheliomatosis), which although not associated with an extramedullary hematopoiesis, can clinically resemble a blueberry muffin rash.

Because of the potential for serious systemic complications, a cause must be sought for all newborns presenting with a blueberry muffin rash. Our patient’s lack of cardiovascular, otic, and ocular involvement combined with a negative TORCH screen and normal CBC strongly suggested against a TORCH infection. In addition, a normal bone survey and CBC, as well as a lack of petechiae, ecchymoses, and hepatosplenomegaly, were evidence against congenital leukemia.³ With the spontaneously resolving lesions and apparent clinical resolution, a bone marrow biopsy was not performed. The skin biopsy revealed negative staining for S-100 and CD1a, making the diagnosis of congenital self-healing reticulohistiocytosis unlikely. No panniculitis was noted and calcium levels were normal, ruling out subcutaneous fat necrosis of the newborn. The predominantly T-cell lymphocytic infiltrate demonstrated on skin biopsy led us to a differential diagnosis of aleukemic leukemia cutis versus idiopathic dermal extramedullary hematopoiesis; however, normocytic anemia was later identified when the patient’s hemoglobin level dropped to 9.9 g/dL. The abnormal eosin-5-maleimide and ektacytometry results unmasked a hereditary spherocytosis.

Hereditary spherocytosis typically is inherited in an autosomal-dominant manner and may be caused by mutations in ankyrin-1, band 3, spectrin, or protein 4.2 on the erythrocyte membrane. It is the third leading cause of hemolytic anemia in newborns and the leading cause of direct Coombs-negative hemolytic anemia requiring blood transfusion in neonates. It is most common in neonates of Northern European ancestry, affecting 1 in every 1000 to 2000 births.⁴ Presentation may range from asymptomatic to severe anemia with hydrops fetalis. Most neonates have an elevated mean corpuscular hemoglobin and low mean corpuscular volume. Acute illness may cause hemolytic or aplastic crises, possibly explaining our patient’s normocytic anemia discovered on a CBC during an episode of an upper respiratory infection and otitis media.

Treatment options for hereditary spherocytosis include phototherapy for jaundiced neonates, folate supplementation, packed erythrocyte transfusions for symptomatic anemia, and recombinant erythropoietin in neonates.⁴ Splenectomy is curative for the majority of patients and requires immunization against Streptococcus pneumoniae, Haemophilus influenzae type b, and Neisseria meningitidis several weeks preoperatively. Patients with symptomatic gallstones may be treated with cholecystectomy at the time of splenectomy or by laparoscopic cholecystectomy, endoscopic sphincterotomy, cholecystostomy, or extracorporeal cholecystolithotripsy.⁵

Although a PubMed search of articles indexed for MEDLINE using the terms dermal hematopoiesis, extramedullary hematopoiesis, hereditary spherocytosis, and blueberry muffin rash yielded only 1 other known case of blueberry muffin rash caused by hereditary spherocytosis,⁶ other case reports demonstrate extramedullary hematopoiesis in hereditary spherocytosis patients in locations other than the skin. Calhoun et al⁷ described a case of a 9-year-old boy with hereditary spherocytosis who presented with jaundice. Pathologic examination revealed a 5-cm suprarenal mass demonstrating extramedullary hematopoiesis.⁷ A case reported by Xiros et al⁸ described a 64-year-old man with a history of hereditary spherocytosis who presented with hemothorax from paravertebral extramedullary hematopoiesis. De Backer et al⁹ reported a case of a 60-year-old man diagnosed with hereditary spherocytosis after an abnormal CBC who was subsequently found to have paravertebral masses containing extramedullary hematopoiesis.

There is one known case of a blueberry muffin rash caused by hereditary spherocytosis.⁶ A female neonate was born at 38 weeks’ gestation with multiple petechiae and faint purpuric papules. Initial complications included intracranial ventricular hemorrhage, hyperbilirubinemia, and anemia requiring blood transfusions on the first day of life. TORCH titers were negative and a skin biopsy demonstrated a diffuse infiltrate of mature RBCs, normoblasts, and pronormoblasts in the reticular dermis. She was healthy until 3 months of age when she had several days of vomiting and diarrhea. Laboratory workup revealed a hematocrit level of 20.5% (reference range, 41%–50%); a reticulocyte count of 22.6% (reference range, 0.5%–1.5%); and a peripheral blood smear demonstrating polychromatophilia, anisocytosis, and spherocytosis. She was then diagnosed with hereditary spherocytosis.⁶

Hereditary spherocytosis is a known, albeit rare, cause of extramedullary hematopoiesis presenting as blueberry muffin rash. Patients with mild hereditary spherocytosis may have a compensated hemolysis without anemia or spherocytes on peripheral smear, which may explain the lack of severe hemolytic anemia or RBC-predominant pathology in our patient.⁵ Argyle and Zone⁶ proposed that severe hemolysis and hypoxia were the cause of extramedullary hematopoiesis in their patient. Because our patient did not experience a notable hemolytic episode until he had an upper respiratory infection and otitis media at 6 months of age, the pathophysiology is less clear; a compensated hemolytic process may underlie the extramedullary hematopoiesis and normal RBC indices.

Regardless of the precise cause of extramedullary hematopoiesis in our patient, this case of a T lymphocyte–dominant cutaneous infiltrate in a patient with mild hereditary spherocytosis is exceptionally rare and leads us to consider that perhaps there are causes of this pathology that are unknown to us.

The term blueberry muffin rash historically was used to describe the cutaneous manifestations observed in congenital rubella. The term traditionally describes the result of a postnatal dermal extramedullary hematopoiesis. Today, TORCH (toxoplasmosis, other agents, rubella, cytomegalovirus, herpes) infections and plasma cell dyscrasias are all potential causes of extramedullary hematopoiesis. Herein, we present a unique case of a neonate born with a blueberry muffin rash secondary to extramedullary hematopoiesis induced by hereditary spherocytosis.

Case Report

The dermatology department was consulted to evaluate a 2-day-old male neonate born with a “rash.” The patient was born to a 34-year-old gravida 3, para 2, woman at 39 weeks’ gestation. The mother’s prenatal laboratory values were within reference range and ultrasounds were normal, and she was compliant with her prenatal care. She underwent a normal spontaneous vaginal delivery 3 hours after rupture of membranes without complication. The amniotic fluid and umbilical cord both were clear. There was no use of forceps or any other external aiding devices during the delivery. At the time of delivery, the consulting physician noted that the patient had “skin lesions from head to toe.”

The patient’s parents reported that the rash did not seem to cause any discomfort for the patient. In the 24 hours after birth, the parents reported that the erythema seemed to slightly fade. Physical examination revealed many scattered erythematous to violaceous, nonblanching papulonodules affecting the scalp (Figure 1), face, arms, hands (Figure 2A), back (Figure 2B), buttocks, legs, and feet. Some of the papulonodules were soft while others were firm and indurated. Several lesions had a yellowish hue with some overlying crust. There was no mucosal, genital, or ocular involvement. No erosions, ulcerations, petechiae, ecchymoses, or hepatosplenomegaly were noted on examination.

The patient was otherwise healthy with an Apgar score of 8/9 at 1 and 5 minutes. His birth weight, length, and head circumference were within normal limits. There was no evidence of ABO blood group or Rhesus factor incompatibility. His temperature, vital signs, laboratory values (including calcium level and TORCH titers, which included cytomegalovirus, rubella, toxoplasmosis, and herpes simplex virus), and review of systems all were within reference range. A bone survey of the skull, spine, ribs, arms, pelvis, legs, and feet was within normal limits.

The mother’s placenta was sent for pathology and revealed a lymphoplasmacytic chronic deciduitis and acute subchorionitis consistent with a nonspecific inflammatory response, unlikely to be from an infectious etiology.

A 4-mm punch biopsy was taken from the left thigh and revealed a predominately lymphocytic infiltrate with rare eosinophils and erythrocyte precursors (Figure 3). Immunohistochemical staining was performed showing that the majority of the lymphocytes represented T lymphocytes, which stained positive for CD45 and CD3 and negative for S-100, CD1a, CD30, and CD117. There were scattered CD34⁺ cells, and scattered cells stained positive for myeloperoxidase. No significant CD20 immunoreactivity was noted. There were scattered eosinophils and rare normoblasts but no megakaryocytes. A complete blood cell count (CBC) with differential and reticulocyte count was within reference range.

At 1-, 3-, 8-, 12-, and 28-week follow-up visits, the patient continued to grow and feed appropriately. No new lesions developed during this time, and the preexisting lesions continued to fade into slightly hyperpigmented patches without induration (Figure 4). At 6 months of age, a CBC performed at the time of an upper respiratory infection and otitis media revealed normocytic anemia with a hemoglobin level of 9.9 g/dL (reference range, 14.0–17.5 g/dL), a reticulocyte count of 0.8% (reference range, 0.5%–1.5%), and a lactate dehydrogenase level of 424 U/L (reference range, 100–200 U/L). All red blood cell (RBC) indices were within reference range. Flow cytometry, eosin-5-maleimide, and ektacytometry were performed with results consistent with mild hereditary spherocytosis.

Comment

Dermal extramedullary hematopoiesis is a normal component of embryologic development up until the fifth month of gestation.¹ The term blueberry muffin rash typically is used to describe the cutaneous manifestations of extramedullary hematopoiesis, which commonly is caused by a TORCH infection or hematologic dyscrasia.² It has been suggested that the term be expanded to include neoplastic processes (eg, neuroblastomas) and vascular processes (eg, multiple hemangiomas, blue rubber bleb nevus syndrome, glomangiomas, multifocal lymphangionendotheliomatosis), which although not associated with an extramedullary hematopoiesis, can clinically resemble a blueberry muffin rash.

Because of the potential for serious systemic complications, a cause must be sought for all newborns presenting with a blueberry muffin rash. Our patient’s lack of cardiovascular, otic, and ocular involvement combined with a negative TORCH screen and normal CBC strongly suggested against a TORCH infection. In addition, a normal bone survey and CBC, as well as a lack of petechiae, ecchymoses, and hepatosplenomegaly, were evidence against congenital leukemia.³ With the spontaneously resolving lesions and apparent clinical resolution, a bone marrow biopsy was not performed. The skin biopsy revealed negative staining for S-100 and CD1a, making the diagnosis of congenital self-healing reticulohistiocytosis unlikely. No panniculitis was noted and calcium levels were normal, ruling out subcutaneous fat necrosis of the newborn. The predominantly T-cell lymphocytic infiltrate demonstrated on skin biopsy led us to a differential diagnosis of aleukemic leukemia cutis versus idiopathic dermal extramedullary hematopoiesis; however, normocytic anemia was later identified when the patient’s hemoglobin level dropped to 9.9 g/dL. The abnormal eosin-5-maleimide and ektacytometry results unmasked a hereditary spherocytosis.

Hereditary spherocytosis typically is inherited in an autosomal-dominant manner and may be caused by mutations in ankyrin-1, band 3, spectrin, or protein 4.2 on the erythrocyte membrane. It is the third leading cause of hemolytic anemia in newborns and the leading cause of direct Coombs-negative hemolytic anemia requiring blood transfusion in neonates. It is most common in neonates of Northern European ancestry, affecting 1 in every 1000 to 2000 births.⁴ Presentation may range from asymptomatic to severe anemia with hydrops fetalis. Most neonates have an elevated mean corpuscular hemoglobin and low mean corpuscular volume. Acute illness may cause hemolytic or aplastic crises, possibly explaining our patient’s normocytic anemia discovered on a CBC during an episode of an upper respiratory infection and otitis media.

Treatment options for hereditary spherocytosis include phototherapy for jaundiced neonates, folate supplementation, packed erythrocyte transfusions for symptomatic anemia, and recombinant erythropoietin in neonates.⁴ Splenectomy is curative for the majority of patients and requires immunization against Streptococcus pneumoniae, Haemophilus influenzae type b, and Neisseria meningitidis several weeks preoperatively. Patients with symptomatic gallstones may be treated with cholecystectomy at the time of splenectomy or by laparoscopic cholecystectomy, endoscopic sphincterotomy, cholecystostomy, or extracorporeal cholecystolithotripsy.⁵

Although a PubMed search of articles indexed for MEDLINE using the terms dermal hematopoiesis, extramedullary hematopoiesis, hereditary spherocytosis, and blueberry muffin rash yielded only 1 other known case of blueberry muffin rash caused by hereditary spherocytosis,⁶ other case reports demonstrate extramedullary hematopoiesis in hereditary spherocytosis patients in locations other than the skin. Calhoun et al⁷ described a case of a 9-year-old boy with hereditary spherocytosis who presented with jaundice. Pathologic examination revealed a 5-cm suprarenal mass demonstrating extramedullary hematopoiesis.⁷ A case reported by Xiros et al⁸ described a 64-year-old man with a history of hereditary spherocytosis who presented with hemothorax from paravertebral extramedullary hematopoiesis. De Backer et al⁹ reported a case of a 60-year-old man diagnosed with hereditary spherocytosis after an abnormal CBC who was subsequently found to have paravertebral masses containing extramedullary hematopoiesis.

There is one known case of a blueberry muffin rash caused by hereditary spherocytosis.⁶ A female neonate was born at 38 weeks’ gestation with multiple petechiae and faint purpuric papules. Initial complications included intracranial ventricular hemorrhage, hyperbilirubinemia, and anemia requiring blood transfusions on the first day of life. TORCH titers were negative and a skin biopsy demonstrated a diffuse infiltrate of mature RBCs, normoblasts, and pronormoblasts in the reticular dermis. She was healthy until 3 months of age when she had several days of vomiting and diarrhea. Laboratory workup revealed a hematocrit level of 20.5% (reference range, 41%–50%); a reticulocyte count of 22.6% (reference range, 0.5%–1.5%); and a peripheral blood smear demonstrating polychromatophilia, anisocytosis, and spherocytosis. She was then diagnosed with hereditary spherocytosis.⁶

Hereditary spherocytosis is a known, albeit rare, cause of extramedullary hematopoiesis presenting as blueberry muffin rash. Patients with mild hereditary spherocytosis may have a compensated hemolysis without anemia or spherocytes on peripheral smear, which may explain the lack of severe hemolytic anemia or RBC-predominant pathology in our patient.⁵ Argyle and Zone⁶ proposed that severe hemolysis and hypoxia were the cause of extramedullary hematopoiesis in their patient. Because our patient did not experience a notable hemolytic episode until he had an upper respiratory infection and otitis media at 6 months of age, the pathophysiology is less clear; a compensated hemolytic process may underlie the extramedullary hematopoiesis and normal RBC indices.

Regardless of the precise cause of extramedullary hematopoiesis in our patient, this case of a T lymphocyte–dominant cutaneous infiltrate in a patient with mild hereditary spherocytosis is exceptionally rare and leads us to consider that perhaps there are causes of this pathology that are unknown to us.

The term blueberry muffin rash historically was used to describe the cutaneous manifestations observed in congenital rubella. The term traditionally describes the result of a postnatal dermal extramedullary hematopoiesis. Today, TORCH (toxoplasmosis, other agents, rubella, cytomegalovirus, herpes) infections and plasma cell dyscrasias are all potential causes of extramedullary hematopoiesis. Herein, we present a unique case of a neonate born with a blueberry muffin rash secondary to extramedullary hematopoiesis induced by hereditary spherocytosis.

Case Report

The dermatology department was consulted to evaluate a 2-day-old male neonate born with a “rash.” The patient was born to a 34-year-old gravida 3, para 2, woman at 39 weeks’ gestation. The mother’s prenatal laboratory values were within reference range and ultrasounds were normal, and she was compliant with her prenatal care. She underwent a normal spontaneous vaginal delivery 3 hours after rupture of membranes without complication. The amniotic fluid and umbilical cord both were clear. There was no use of forceps or any other external aiding devices during the delivery. At the time of delivery, the consulting physician noted that the patient had “skin lesions from head to toe.”

The patient’s parents reported that the rash did not seem to cause any discomfort for the patient. In the 24 hours after birth, the parents reported that the erythema seemed to slightly fade. Physical examination revealed many scattered erythematous to violaceous, nonblanching papulonodules affecting the scalp (Figure 1), face, arms, hands (Figure 2A), back (Figure 2B), buttocks, legs, and feet. Some of the papulonodules were soft while others were firm and indurated. Several lesions had a yellowish hue with some overlying crust. There was no mucosal, genital, or ocular involvement. No erosions, ulcerations, petechiae, ecchymoses, or hepatosplenomegaly were noted on examination.

The patient was otherwise healthy with an Apgar score of 8/9 at 1 and 5 minutes. His birth weight, length, and head circumference were within normal limits. There was no evidence of ABO blood group or Rhesus factor incompatibility. His temperature, vital signs, laboratory values (including calcium level and TORCH titers, which included cytomegalovirus, rubella, toxoplasmosis, and herpes simplex virus), and review of systems all were within reference range. A bone survey of the skull, spine, ribs, arms, pelvis, legs, and feet was within normal limits.

The mother’s placenta was sent for pathology and revealed a lymphoplasmacytic chronic deciduitis and acute subchorionitis consistent with a nonspecific inflammatory response, unlikely to be from an infectious etiology.

A 4-mm punch biopsy was taken from the left thigh and revealed a predominately lymphocytic infiltrate with rare eosinophils and erythrocyte precursors (Figure 3). Immunohistochemical staining was performed showing that the majority of the lymphocytes represented T lymphocytes, which stained positive for CD45 and CD3 and negative for S-100, CD1a, CD30, and CD117. There were scattered CD34⁺ cells, and scattered cells stained positive for myeloperoxidase. No significant CD20 immunoreactivity was noted. There were scattered eosinophils and rare normoblasts but no megakaryocytes. A complete blood cell count (CBC) with differential and reticulocyte count was within reference range.

At 1-, 3-, 8-, 12-, and 28-week follow-up visits, the patient continued to grow and feed appropriately. No new lesions developed during this time, and the preexisting lesions continued to fade into slightly hyperpigmented patches without induration (Figure 4). At 6 months of age, a CBC performed at the time of an upper respiratory infection and otitis media revealed normocytic anemia with a hemoglobin level of 9.9 g/dL (reference range, 14.0–17.5 g/dL), a reticulocyte count of 0.8% (reference range, 0.5%–1.5%), and a lactate dehydrogenase level of 424 U/L (reference range, 100–200 U/L). All red blood cell (RBC) indices were within reference range. Flow cytometry, eosin-5-maleimide, and ektacytometry were performed with results consistent with mild hereditary spherocytosis.

Comment

Dermal extramedullary hematopoiesis is a normal component of embryologic development up until the fifth month of gestation.¹ The term blueberry muffin rash typically is used to describe the cutaneous manifestations of extramedullary hematopoiesis, which commonly is caused by a TORCH infection or hematologic dyscrasia.² It has been suggested that the term be expanded to include neoplastic processes (eg, neuroblastomas) and vascular processes (eg, multiple hemangiomas, blue rubber bleb nevus syndrome, glomangiomas, multifocal lymphangionendotheliomatosis), which although not associated with an extramedullary hematopoiesis, can clinically resemble a blueberry muffin rash.

Because of the potential for serious systemic complications, a cause must be sought for all newborns presenting with a blueberry muffin rash. Our patient’s lack of cardiovascular, otic, and ocular involvement combined with a negative TORCH screen and normal CBC strongly suggested against a TORCH infection. In addition, a normal bone survey and CBC, as well as a lack of petechiae, ecchymoses, and hepatosplenomegaly, were evidence against congenital leukemia.³ With the spontaneously resolving lesions and apparent clinical resolution, a bone marrow biopsy was not performed. The skin biopsy revealed negative staining for S-100 and CD1a, making the diagnosis of congenital self-healing reticulohistiocytosis unlikely. No panniculitis was noted and calcium levels were normal, ruling out subcutaneous fat necrosis of the newborn. The predominantly T-cell lymphocytic infiltrate demonstrated on skin biopsy led us to a differential diagnosis of aleukemic leukemia cutis versus idiopathic dermal extramedullary hematopoiesis; however, normocytic anemia was later identified when the patient’s hemoglobin level dropped to 9.9 g/dL. The abnormal eosin-5-maleimide and ektacytometry results unmasked a hereditary spherocytosis.

Hereditary spherocytosis typically is inherited in an autosomal-dominant manner and may be caused by mutations in ankyrin-1, band 3, spectrin, or protein 4.2 on the erythrocyte membrane. It is the third leading cause of hemolytic anemia in newborns and the leading cause of direct Coombs-negative hemolytic anemia requiring blood transfusion in neonates. It is most common in neonates of Northern European ancestry, affecting 1 in every 1000 to 2000 births.⁴ Presentation may range from asymptomatic to severe anemia with hydrops fetalis. Most neonates have an elevated mean corpuscular hemoglobin and low mean corpuscular volume. Acute illness may cause hemolytic or aplastic crises, possibly explaining our patient’s normocytic anemia discovered on a CBC during an episode of an upper respiratory infection and otitis media.

Treatment options for hereditary spherocytosis include phototherapy for jaundiced neonates, folate supplementation, packed erythrocyte transfusions for symptomatic anemia, and recombinant erythropoietin in neonates.⁴ Splenectomy is curative for the majority of patients and requires immunization against Streptococcus pneumoniae, Haemophilus influenzae type b, and Neisseria meningitidis several weeks preoperatively. Patients with symptomatic gallstones may be treated with cholecystectomy at the time of splenectomy or by laparoscopic cholecystectomy, endoscopic sphincterotomy, cholecystostomy, or extracorporeal cholecystolithotripsy.⁵

Although a PubMed search of articles indexed for MEDLINE using the terms dermal hematopoiesis, extramedullary hematopoiesis, hereditary spherocytosis, and blueberry muffin rash yielded only 1 other known case of blueberry muffin rash caused by hereditary spherocytosis,⁶ other case reports demonstrate extramedullary hematopoiesis in hereditary spherocytosis patients in locations other than the skin. Calhoun et al⁷ described a case of a 9-year-old boy with hereditary spherocytosis who presented with jaundice. Pathologic examination revealed a 5-cm suprarenal mass demonstrating extramedullary hematopoiesis.⁷ A case reported by Xiros et al⁸ described a 64-year-old man with a history of hereditary spherocytosis who presented with hemothorax from paravertebral extramedullary hematopoiesis. De Backer et al⁹ reported a case of a 60-year-old man diagnosed with hereditary spherocytosis after an abnormal CBC who was subsequently found to have paravertebral masses containing extramedullary hematopoiesis.

There is one known case of a blueberry muffin rash caused by hereditary spherocytosis.⁶ A female neonate was born at 38 weeks’ gestation with multiple petechiae and faint purpuric papules. Initial complications included intracranial ventricular hemorrhage, hyperbilirubinemia, and anemia requiring blood transfusions on the first day of life. TORCH titers were negative and a skin biopsy demonstrated a diffuse infiltrate of mature RBCs, normoblasts, and pronormoblasts in the reticular dermis. She was healthy until 3 months of age when she had several days of vomiting and diarrhea. Laboratory workup revealed a hematocrit level of 20.5% (reference range, 41%–50%); a reticulocyte count of 22.6% (reference range, 0.5%–1.5%); and a peripheral blood smear demonstrating polychromatophilia, anisocytosis, and spherocytosis. She was then diagnosed with hereditary spherocytosis.⁶

Hereditary spherocytosis is a known, albeit rare, cause of extramedullary hematopoiesis presenting as blueberry muffin rash. Patients with mild hereditary spherocytosis may have a compensated hemolysis without anemia or spherocytes on peripheral smear, which may explain the lack of severe hemolytic anemia or RBC-predominant pathology in our patient.⁵ Argyle and Zone⁶ proposed that severe hemolysis and hypoxia were the cause of extramedullary hematopoiesis in their patient. Because our patient did not experience a notable hemolytic episode until he had an upper respiratory infection and otitis media at 6 months of age, the pathophysiology is less clear; a compensated hemolytic process may underlie the extramedullary hematopoiesis and normal RBC indices.

Regardless of the precise cause of extramedullary hematopoiesis in our patient, this case of a T lymphocyte–dominant cutaneous infiltrate in a patient with mild hereditary spherocytosis is exceptionally rare and leads us to consider that perhaps there are causes of this pathology that are unknown to us.