Cutis

The Diagnosis: Acquired Lymphangioma Circumscriptum

A skin biopsy of the plaque on the right labium majus showed a proliferation of well-formed, dilated lymphatic vessels lined by benign-appearing endothelial cells in the papillary dermis (Figure). These findings were consistent with a diagnosis of acquired lymphangioma circumscriptum (ALC) in the setting of severe hidradenitis suppurativa (HS).

Image courtesy of Alicia Schnebelen, MD (Dallas, Texas).

Acquired lymphangioma circumscriptum (also known as acquired lymphangiectasia or secondary lymphangioma¹) is a rare skin finding resulting from chronic lymphatic obstruction that leads to dilated lymphatic vessels within the dermis.^2,3 There also is a distinct congenital form of lymphangioma circumscriptum caused by lymphatic malformations present at birth.^2,4 Acquired lymphangioma circumscriptum of the vulva is a rare phenomenon.³ Identified causes include radiation or surgery for carcinoma, solid gynecologic tumors, lymphadenectomy, Crohn disease, and tuberculosis and other infections, all of which can disrupt normal lymphatics to cause ALC.^2-4 Hidradenitis suppurativa is not a widely recognized cause of ALC; however, this phenomenon is reported in the literature. A long-standing history of severe HS complicated by lymphedema seems to precede the development of ALC in the reported cases, as in our patient.^5-7

Acquired lymphangioma circumscriptum of the vulva can appear in women of all ages as frog spawn or cobblestone papules or vesicles, sometimes with a hyperkeratotic or verrucous appearance.^2,4 Associated symptoms include serous drainage, edema, pruritus, and discomfort. The lesions may become eroded, which can predispose patients to secondary infections.^1,2 Acquired lymphangioma circumscriptum of the vulva can be difficult to diagnose, as the time interval between the initial cause and the appearance of skin findings can be years, leading to the misdiagnosis of ALC as other similar-appearing genital skin conditions such as squamous cell carcinoma or condyloma.^4,8 When misidentified as an infection, diagnosis can lead to substantial distress, abstinence from sexual activity, and unnecessary and painful treatments.

Skin biopsy is helpful in distinguishing ALC from other differential diagnoses such as condylomata acuminata, squamous cell carcinoma, and condyloma lata. Histopathology in ALC is notable for dilated lymphatic vessels filled with hypocellular fluid and lined with endothelial cells in the superficial dermis; the epidermis can appear hyperplastic, hyperkeratotic, or eroded.^3-5,9 These lymphatic vessels stain positively for CD31 and D2-40, markers for endothelial cells and lymphatic endothelium, respectively, and negative for CD34, a marker for vascular endothelium.^3,4,9 Features suggestive of condylomata acuminata such as rounded parakeratosis, hypergranulosis, and vacuolated keratinocytes⁹ are not present. The giant condyloma of Buschke-Löwenstein, a clinical variant of verrucous squamous cell carcinoma, also can present as a warty ulcerated papule or plaque in the genital region, but the characteristic rounded eosinophilic keratinocytes pushing down into the dermis⁹ are not seen in ALC. Secondary syphilis is associated with condyloma lata, which are verrucous or fleshy-appearing papules often coalescing into plaques located in the anogenital region. Pathologic features of secondary syphilis include vacuolar interface dermatitis and acanthosis with long slender rete ridges.⁹ Squamous cell carcinoma, which can arise from inflammation associated with long-standing HS, must be ruled out, as it is associated with a high risk of mortality in patients with HS.¹⁰

It is noteworthy to recognize the various, often confusing nomenclature used to describe cutaneous lymphatic conditions. The terms acquired lymphangioma circumscriptum, secondary lymphangioma, and lymphangiectasia are used interchangeably to describe dilated lymphatic vessels in the skin.¹ The term atypical vascular lesion refers to lymphectasias of the skin of the breast due to prior radiation therapy most often used in the treatment of breast carcinoma; clinically, these present as red-brown or flesh-colored papules or telangiectatic plaques on the breast.^11,12 Lymphedema also may occur alongside atypical vascular lesions, as prior radiation or surgical lymph node dissection can predispose patients to impaired lymphatic drainage.¹³ The lymphatic histopathologic subtype of atypical vascular lesions may appear similar to ALC; however, the vascular subtype will demonstrate collections of capillary-sized vessels and extravasated erythrocytes.^11,12 Unlike ALC, the benign nature of atypical vascular lesions has been questioned, as they may be associated with a small risk for progression to angiosarcoma.^11-13 It also is important to distinguish ALC from lymphangiomatosis, a generalized lymphatic anomaly that is characterized by extensive lymphatic malformations involving numerous internal organs, including the lungs and gastrointestinal tract. This condition is associated with notable morbidity and mortality.¹³

Although the suffix of the term lymphangioma suggests a neoplastic process, ALC is not a neoplasm and can be managed expectantly in many cases.^2,3,8 However, due to cosmetic appearance, pain, discomfort, and recurrent bacterial superinfections, many patients pursue treatment. Treatment options for ALC include sclerotherapy, electrocautery, radiofrequency or carbon dioxide laser ablation, and excision, though recurrence can arise.^3-5,7,8 Our patient elected to manage her asymptomatic ALC expectantly.

The Diagnosis: Acquired Lymphangioma Circumscriptum

A skin biopsy of the plaque on the right labium majus showed a proliferation of well-formed, dilated lymphatic vessels lined by benign-appearing endothelial cells in the papillary dermis (Figure). These findings were consistent with a diagnosis of acquired lymphangioma circumscriptum (ALC) in the setting of severe hidradenitis suppurativa (HS).

Image courtesy of Alicia Schnebelen, MD (Dallas, Texas).

Acquired lymphangioma circumscriptum (also known as acquired lymphangiectasia or secondary lymphangioma¹) is a rare skin finding resulting from chronic lymphatic obstruction that leads to dilated lymphatic vessels within the dermis.^2,3 There also is a distinct congenital form of lymphangioma circumscriptum caused by lymphatic malformations present at birth.^2,4 Acquired lymphangioma circumscriptum of the vulva is a rare phenomenon.³ Identified causes include radiation or surgery for carcinoma, solid gynecologic tumors, lymphadenectomy, Crohn disease, and tuberculosis and other infections, all of which can disrupt normal lymphatics to cause ALC.^2-4 Hidradenitis suppurativa is not a widely recognized cause of ALC; however, this phenomenon is reported in the literature. A long-standing history of severe HS complicated by lymphedema seems to precede the development of ALC in the reported cases, as in our patient.^5-7

Acquired lymphangioma circumscriptum of the vulva can appear in women of all ages as frog spawn or cobblestone papules or vesicles, sometimes with a hyperkeratotic or verrucous appearance.^2,4 Associated symptoms include serous drainage, edema, pruritus, and discomfort. The lesions may become eroded, which can predispose patients to secondary infections.^1,2 Acquired lymphangioma circumscriptum of the vulva can be difficult to diagnose, as the time interval between the initial cause and the appearance of skin findings can be years, leading to the misdiagnosis of ALC as other similar-appearing genital skin conditions such as squamous cell carcinoma or condyloma.^4,8 When misidentified as an infection, diagnosis can lead to substantial distress, abstinence from sexual activity, and unnecessary and painful treatments.

Skin biopsy is helpful in distinguishing ALC from other differential diagnoses such as condylomata acuminata, squamous cell carcinoma, and condyloma lata. Histopathology in ALC is notable for dilated lymphatic vessels filled with hypocellular fluid and lined with endothelial cells in the superficial dermis; the epidermis can appear hyperplastic, hyperkeratotic, or eroded.^3-5,9 These lymphatic vessels stain positively for CD31 and D2-40, markers for endothelial cells and lymphatic endothelium, respectively, and negative for CD34, a marker for vascular endothelium.^3,4,9 Features suggestive of condylomata acuminata such as rounded parakeratosis, hypergranulosis, and vacuolated keratinocytes⁹ are not present. The giant condyloma of Buschke-Löwenstein, a clinical variant of verrucous squamous cell carcinoma, also can present as a warty ulcerated papule or plaque in the genital region, but the characteristic rounded eosinophilic keratinocytes pushing down into the dermis⁹ are not seen in ALC. Secondary syphilis is associated with condyloma lata, which are verrucous or fleshy-appearing papules often coalescing into plaques located in the anogenital region. Pathologic features of secondary syphilis include vacuolar interface dermatitis and acanthosis with long slender rete ridges.⁹ Squamous cell carcinoma, which can arise from inflammation associated with long-standing HS, must be ruled out, as it is associated with a high risk of mortality in patients with HS.¹⁰

It is noteworthy to recognize the various, often confusing nomenclature used to describe cutaneous lymphatic conditions. The terms acquired lymphangioma circumscriptum, secondary lymphangioma, and lymphangiectasia are used interchangeably to describe dilated lymphatic vessels in the skin.¹ The term atypical vascular lesion refers to lymphectasias of the skin of the breast due to prior radiation therapy most often used in the treatment of breast carcinoma; clinically, these present as red-brown or flesh-colored papules or telangiectatic plaques on the breast.^11,12 Lymphedema also may occur alongside atypical vascular lesions, as prior radiation or surgical lymph node dissection can predispose patients to impaired lymphatic drainage.¹³ The lymphatic histopathologic subtype of atypical vascular lesions may appear similar to ALC; however, the vascular subtype will demonstrate collections of capillary-sized vessels and extravasated erythrocytes.^11,12 Unlike ALC, the benign nature of atypical vascular lesions has been questioned, as they may be associated with a small risk for progression to angiosarcoma.^11-13 It also is important to distinguish ALC from lymphangiomatosis, a generalized lymphatic anomaly that is characterized by extensive lymphatic malformations involving numerous internal organs, including the lungs and gastrointestinal tract. This condition is associated with notable morbidity and mortality.¹³

Although the suffix of the term lymphangioma suggests a neoplastic process, ALC is not a neoplasm and can be managed expectantly in many cases.^2,3,8 However, due to cosmetic appearance, pain, discomfort, and recurrent bacterial superinfections, many patients pursue treatment. Treatment options for ALC include sclerotherapy, electrocautery, radiofrequency or carbon dioxide laser ablation, and excision, though recurrence can arise.^3-5,7,8 Our patient elected to manage her asymptomatic ALC expectantly.

The Diagnosis: Acquired Lymphangioma Circumscriptum

A skin biopsy of the plaque on the right labium majus showed a proliferation of well-formed, dilated lymphatic vessels lined by benign-appearing endothelial cells in the papillary dermis (Figure). These findings were consistent with a diagnosis of acquired lymphangioma circumscriptum (ALC) in the setting of severe hidradenitis suppurativa (HS).

Image courtesy of Alicia Schnebelen, MD (Dallas, Texas).

Acquired lymphangioma circumscriptum (also known as acquired lymphangiectasia or secondary lymphangioma¹) is a rare skin finding resulting from chronic lymphatic obstruction that leads to dilated lymphatic vessels within the dermis.^2,3 There also is a distinct congenital form of lymphangioma circumscriptum caused by lymphatic malformations present at birth.^2,4 Acquired lymphangioma circumscriptum of the vulva is a rare phenomenon.³ Identified causes include radiation or surgery for carcinoma, solid gynecologic tumors, lymphadenectomy, Crohn disease, and tuberculosis and other infections, all of which can disrupt normal lymphatics to cause ALC.^2-4 Hidradenitis suppurativa is not a widely recognized cause of ALC; however, this phenomenon is reported in the literature. A long-standing history of severe HS complicated by lymphedema seems to precede the development of ALC in the reported cases, as in our patient.^5-7

Acquired lymphangioma circumscriptum of the vulva can appear in women of all ages as frog spawn or cobblestone papules or vesicles, sometimes with a hyperkeratotic or verrucous appearance.^2,4 Associated symptoms include serous drainage, edema, pruritus, and discomfort. The lesions may become eroded, which can predispose patients to secondary infections.^1,2 Acquired lymphangioma circumscriptum of the vulva can be difficult to diagnose, as the time interval between the initial cause and the appearance of skin findings can be years, leading to the misdiagnosis of ALC as other similar-appearing genital skin conditions such as squamous cell carcinoma or condyloma.^4,8 When misidentified as an infection, diagnosis can lead to substantial distress, abstinence from sexual activity, and unnecessary and painful treatments.

Skin biopsy is helpful in distinguishing ALC from other differential diagnoses such as condylomata acuminata, squamous cell carcinoma, and condyloma lata. Histopathology in ALC is notable for dilated lymphatic vessels filled with hypocellular fluid and lined with endothelial cells in the superficial dermis; the epidermis can appear hyperplastic, hyperkeratotic, or eroded.^3-5,9 These lymphatic vessels stain positively for CD31 and D2-40, markers for endothelial cells and lymphatic endothelium, respectively, and negative for CD34, a marker for vascular endothelium.^3,4,9 Features suggestive of condylomata acuminata such as rounded parakeratosis, hypergranulosis, and vacuolated keratinocytes⁹ are not present. The giant condyloma of Buschke-Löwenstein, a clinical variant of verrucous squamous cell carcinoma, also can present as a warty ulcerated papule or plaque in the genital region, but the characteristic rounded eosinophilic keratinocytes pushing down into the dermis⁹ are not seen in ALC. Secondary syphilis is associated with condyloma lata, which are verrucous or fleshy-appearing papules often coalescing into plaques located in the anogenital region. Pathologic features of secondary syphilis include vacuolar interface dermatitis and acanthosis with long slender rete ridges.⁹ Squamous cell carcinoma, which can arise from inflammation associated with long-standing HS, must be ruled out, as it is associated with a high risk of mortality in patients with HS.¹⁰

It is noteworthy to recognize the various, often confusing nomenclature used to describe cutaneous lymphatic conditions. The terms acquired lymphangioma circumscriptum, secondary lymphangioma, and lymphangiectasia are used interchangeably to describe dilated lymphatic vessels in the skin.¹ The term atypical vascular lesion refers to lymphectasias of the skin of the breast due to prior radiation therapy most often used in the treatment of breast carcinoma; clinically, these present as red-brown or flesh-colored papules or telangiectatic plaques on the breast.^11,12 Lymphedema also may occur alongside atypical vascular lesions, as prior radiation or surgical lymph node dissection can predispose patients to impaired lymphatic drainage.¹³ The lymphatic histopathologic subtype of atypical vascular lesions may appear similar to ALC; however, the vascular subtype will demonstrate collections of capillary-sized vessels and extravasated erythrocytes.^11,12 Unlike ALC, the benign nature of atypical vascular lesions has been questioned, as they may be associated with a small risk for progression to angiosarcoma.^11-13 It also is important to distinguish ALC from lymphangiomatosis, a generalized lymphatic anomaly that is characterized by extensive lymphatic malformations involving numerous internal organs, including the lungs and gastrointestinal tract. This condition is associated with notable morbidity and mortality.¹³

Although the suffix of the term lymphangioma suggests a neoplastic process, ALC is not a neoplasm and can be managed expectantly in many cases.^2,3,8 However, due to cosmetic appearance, pain, discomfort, and recurrent bacterial superinfections, many patients pursue treatment. Treatment options for ALC include sclerotherapy, electrocautery, radiofrequency or carbon dioxide laser ablation, and excision, though recurrence can arise.^3-5,7,8 Our patient elected to manage her asymptomatic ALC expectantly.

Angioimmunoblastic T-cell lymphoma (AITL) is a rare and aggressive lymphoma arising from follicular T-helper cells that predominantly affects older adults and carries a 5-year overall survival rate of 32%.¹ Notably, as many as 50% of AITL patients present with a skin rash in addition to the more common but nonspecific acute-onset generalized lymphadenopathy, hepatosplenomegaly, and anemia.² At presentation, most AITL patients are already at an advanced (III/IV) stage of disease.

Formerly known as angioimmunoblastic lymphadenopathy with dysproteinemia, AITL was once considered a benign entity that carried a risk for malignant transformation. As more cases have been identified and explored, this entity has been recategorized as a frank lymphoma.³ Therefore, it is critical that AITL be diagnosed and treated as early as possible.

We present the case of a 65-year-old man with clinical features that resembled drug reaction with eosinophilia and systemic symptoms (DRESS syndrome). After extensive workup, he was found to have AITL. This atypical case highlights the importance of maintaining a flexible differential diagnosis in patients with a persistent rash that does not improve with appropriate drug withdrawal and therapy.

Case Report

A 65-year-old Filipino man whose medical history was notable for hepatitis B that had been treated with entecavir for years without issue was admitted to the internal medicine service with fever of unknown origin and malaise of approximately 6 weeks’ duration. Six days prior to admission and 5 days after completing courses of the antiviral oseltamivir phosphate and amoxicillin for an upper respiratory tract infection and sinusitis, he developed worsening of an intermittently pruritic rash of approximately 1 month's duration. The dermatology department was consulted the day of hospital admission for evaluation of the rash. Chronic home medications included entecavir, lisinopril/hydrochlorothiazide, amlodipine, atorvastatin, metformin, salsalate, and over-the-counter nonsteroidal anti-inflammatory drugs (NSAIDs) as needed.

Physical examination was notable for mild erythema and scale distributed across the entire face; mild facial edema; and a blanchable, nonconfluent, macular erythema distributed across the trunk and upper and proximal lower extremities (Figure). In addition, the patient displayed conjunctival injection, pitting edema of the hands, and bilateral cervical and inguinal lymphadenopathy.

Photographs courtesy of James Contestable, MD (Camp Lejeune, North Carolina).

Laboratory tests revealed mild leukocytosis (11.6×10⁹/L, [reference range, 4.0–10.5×10⁹/L]), anemia (hemoglobin, 125 g/L (reference range, 138–170 g/L); hematocrit, 36.9%, [reference range, 40.0%–50.0%)], eosinophilia (1.07×10⁹/L [reference range, 0.00–0.70×10⁹/L)], hyponatremia, hypokalemia, and a mildly elevated creatinine level. Computed tomography and full-body positron-emission tomography (PET) scans during admission demonstrated diffuse lymphadenopathy. A skin biopsy from the left chest and a left inguinal lymph node biopsy also were performed.

Despite the lack of a clear medication trigger within the usual timeline for severe cutaneous drug-induced hypersensitivity reactions, DRESS syndrome was high on the differential diagnosis at the time of the initial presentation given the diffuse morbilliform eruption with pruritus, facial edema, eosinophilia, and lymphadenopathy.

Home medications were discontinued except for amlodipine, atorvastatin, and entecavir. The patient was treated symptomatically with topical steroids because it was believed that, if the clinical presentation represented DRESS syndrome, it was a mild variant that could be treated topically.⁴ His case was considered mild because of a lack of confirmed organ dysfunction and a mild protracted course.

After discharge following a 3-day inpatient stay, the patient was followed in the clinic weekly for 3 weeks without considerable change in the skin or laboratory findings. Discontinuation of entecavir was discussed and approved by his hepatologist.

Posthospitalization analysis of the punch biopsy specimen from the chest performed during the patient’s hospital stay revealed a superficial and deep dermal lymphoid infiltrate comprising CD3-, CD5-, and programmed cell death protein 1–positive cells with cytologic atypia in a perivascular distribution. Analysis of the lymph node biopsy specimen performed during the hospitalization showed effacement of the nodal architecture, a polymorphous lymphoid cell population with irregular nuclear contour, and abundant clear cytoplasm associated with high endothelial venules (HEVs). Cells of interest were positive for CD3, CD4, CD2, CD5, and CD7, with a subset staining positive for programmed cell death protein 1, inducible costimulator, CD10, and chemokine (C-X-C motif) ligand (CXCL) 13. CD21 demonstrated an expanded follicular dendritic cell meshwork in association with HEVs. Polymerase chain reaction revealed a clonal T-cell population. These findings of the skin and lymph node biopsies were consistent with AITL. Subsequent bone marrow biopsy with flow cytometry showed a normal CD4:CD8 ratio in T cells and no increase in natural killer cells.

Cyclophosphamide–hydroxydaunorubicin–Oncovin–prednisone (CHOP) chemotherapy was initiated; the patient completed a total of 6 cycles. He has had near resolution of the skin findings and is considered in remission based on a PET scan performed approximately 7 months after the initial presentation.

Comment

Angioimmunoblastic T-cell lymphoma is a rare peripheral T-cell lymphoma, part of a group of aggressive neoplasms that constitute approximately 15% of peripheral T-cell lymphomas and approximately 2% of non-Hodgkin lymphomas in adults worldwide.⁵ Cutaneous involvement occurs in approximately half of AITL cases and can be the first manifestation of disease.² Skin findings are largely nonspecific, ranging from simple morbilliform rashes to erythroderma, at times manifesting with purpura.

Given this variability in the presentation of AITL, early diagnosis is challenging in the absence of more specific signs and symptoms.² It can conceivably be mistaken for common entities such as viral exanthems or drug eruptions, depending on the history and context. DRESS syndrome, a T cell-mediated, delayed type-IV hypersensitivity drug reaction can present in a manner highly similar to that of AITL, with cutaneous involvement (diffuse morbilliform rash, fever, facial edema, and generalized lymphadenopathy) and variable systemic involvement. Laboratory findings of eosinophilia, atypical lymphocytes, and thrombocytopenia also might be seen in both entities.⁶ Furthermore, the AITL in our patient was accompanied by electrolyte disturbances that were concerning for syndrome of inappropriate antidiuretic hormone secretion, a rare complication of patients with DRESS syndrome complicated by encephalitis.^7,8

Our patient met 4 RegiSCAR criteria for DRESS syndrome, warranting high clinical suspicion for an offending drug.⁹ DRESS syndrome can be caused by numerous medications—most commonly anticonvulsants, sulfonamides, antibiotics, allopurinol, and NSAIDs. A review of our patient’s medication list identified NSAIDs (including salsalate), entecavir, and amoxicillin, as possible culpable medications. Notably, the only new addition to the patient’s regimen was amoxicillin, which did not fit the typical 2- to 8-week timeline for a DRESS syndrome nidus.¹⁰ Our patient’s fever began well before the antibiotic was initiated, and skin findings appeared within 1 week after the course of amoxicillin was completed. Although there is documented variability in the latency of onset of DRESS syndrome following administration of a culprit medication,¹¹ it is critical to maintain a broad differential diagnosis to allow for further diagnostic information to be obtained, especially when the medication timeline does not align with the clinical presentation.

DRESS syndrome is far more common than AITL. Similarities in their clinical presentation pose a substantial challenge and often cause a delay in the diagnosis of AITL, which is made by excisional tissue biopsy, most commonly of a lymph node, with assessment of morphology and immunophenotyping. Histologic assessment of tissue reveals a polymorphous infiltrate of variably sized atypical lymphocytes with prominent arborizing HEVs as well as expanded populations of follicular dendritic cells that can be detected by CD21 staining. Cells express CD3 and CD4, variably express BCL6 (B-cell lymphoma 6 antigen) and CD10, and also may have partial or complete loss of expression of a subset of pan T-cell antigens (CD2, CD3, CD5, and CD7).^12-18

The treatment approach to AITL mirrors that of other nodal peripheral T-cell lymphomas, including chemotherapy and consideration of autologous stem-cell transplantation. Recent prospective trials of CHOP and CHOP-like chemotherapy have reported 3-year event-free survival and overall survival rates of 50% and 68%, respectively.¹⁹ Novel chemotherapeutic targets and gene-expression profiling are being investigated as potential therapeutic avenues.²⁰

Conclusion

DRESS syndrome and AITL can have near-identical presentations. Clinicians should maintain a high index of suspicion for AITL in patients with presumed DRESS syndrome whose rash does not improve with appropriate drug withdrawal and steroid therapy or who lack a strong offending medication history. In such cases, skin and lymph node biopsies should be performed as early as possible to evaluate for AITL and so that appropriate therapy can be initiated.

Angioimmunoblastic T-cell lymphoma (AITL) is a rare and aggressive lymphoma arising from follicular T-helper cells that predominantly affects older adults and carries a 5-year overall survival rate of 32%.¹ Notably, as many as 50% of AITL patients present with a skin rash in addition to the more common but nonspecific acute-onset generalized lymphadenopathy, hepatosplenomegaly, and anemia.² At presentation, most AITL patients are already at an advanced (III/IV) stage of disease.

Formerly known as angioimmunoblastic lymphadenopathy with dysproteinemia, AITL was once considered a benign entity that carried a risk for malignant transformation. As more cases have been identified and explored, this entity has been recategorized as a frank lymphoma.³ Therefore, it is critical that AITL be diagnosed and treated as early as possible.

We present the case of a 65-year-old man with clinical features that resembled drug reaction with eosinophilia and systemic symptoms (DRESS syndrome). After extensive workup, he was found to have AITL. This atypical case highlights the importance of maintaining a flexible differential diagnosis in patients with a persistent rash that does not improve with appropriate drug withdrawal and therapy.

Case Report

A 65-year-old Filipino man whose medical history was notable for hepatitis B that had been treated with entecavir for years without issue was admitted to the internal medicine service with fever of unknown origin and malaise of approximately 6 weeks’ duration. Six days prior to admission and 5 days after completing courses of the antiviral oseltamivir phosphate and amoxicillin for an upper respiratory tract infection and sinusitis, he developed worsening of an intermittently pruritic rash of approximately 1 month's duration. The dermatology department was consulted the day of hospital admission for evaluation of the rash. Chronic home medications included entecavir, lisinopril/hydrochlorothiazide, amlodipine, atorvastatin, metformin, salsalate, and over-the-counter nonsteroidal anti-inflammatory drugs (NSAIDs) as needed.

Physical examination was notable for mild erythema and scale distributed across the entire face; mild facial edema; and a blanchable, nonconfluent, macular erythema distributed across the trunk and upper and proximal lower extremities (Figure). In addition, the patient displayed conjunctival injection, pitting edema of the hands, and bilateral cervical and inguinal lymphadenopathy.

Photographs courtesy of James Contestable, MD (Camp Lejeune, North Carolina).

Laboratory tests revealed mild leukocytosis (11.6×10⁹/L, [reference range, 4.0–10.5×10⁹/L]), anemia (hemoglobin, 125 g/L (reference range, 138–170 g/L); hematocrit, 36.9%, [reference range, 40.0%–50.0%)], eosinophilia (1.07×10⁹/L [reference range, 0.00–0.70×10⁹/L)], hyponatremia, hypokalemia, and a mildly elevated creatinine level. Computed tomography and full-body positron-emission tomography (PET) scans during admission demonstrated diffuse lymphadenopathy. A skin biopsy from the left chest and a left inguinal lymph node biopsy also were performed.

Despite the lack of a clear medication trigger within the usual timeline for severe cutaneous drug-induced hypersensitivity reactions, DRESS syndrome was high on the differential diagnosis at the time of the initial presentation given the diffuse morbilliform eruption with pruritus, facial edema, eosinophilia, and lymphadenopathy.

Home medications were discontinued except for amlodipine, atorvastatin, and entecavir. The patient was treated symptomatically with topical steroids because it was believed that, if the clinical presentation represented DRESS syndrome, it was a mild variant that could be treated topically.⁴ His case was considered mild because of a lack of confirmed organ dysfunction and a mild protracted course.

After discharge following a 3-day inpatient stay, the patient was followed in the clinic weekly for 3 weeks without considerable change in the skin or laboratory findings. Discontinuation of entecavir was discussed and approved by his hepatologist.

Posthospitalization analysis of the punch biopsy specimen from the chest performed during the patient’s hospital stay revealed a superficial and deep dermal lymphoid infiltrate comprising CD3-, CD5-, and programmed cell death protein 1–positive cells with cytologic atypia in a perivascular distribution. Analysis of the lymph node biopsy specimen performed during the hospitalization showed effacement of the nodal architecture, a polymorphous lymphoid cell population with irregular nuclear contour, and abundant clear cytoplasm associated with high endothelial venules (HEVs). Cells of interest were positive for CD3, CD4, CD2, CD5, and CD7, with a subset staining positive for programmed cell death protein 1, inducible costimulator, CD10, and chemokine (C-X-C motif) ligand (CXCL) 13. CD21 demonstrated an expanded follicular dendritic cell meshwork in association with HEVs. Polymerase chain reaction revealed a clonal T-cell population. These findings of the skin and lymph node biopsies were consistent with AITL. Subsequent bone marrow biopsy with flow cytometry showed a normal CD4:CD8 ratio in T cells and no increase in natural killer cells.

Cyclophosphamide–hydroxydaunorubicin–Oncovin–prednisone (CHOP) chemotherapy was initiated; the patient completed a total of 6 cycles. He has had near resolution of the skin findings and is considered in remission based on a PET scan performed approximately 7 months after the initial presentation.

Comment

Angioimmunoblastic T-cell lymphoma is a rare peripheral T-cell lymphoma, part of a group of aggressive neoplasms that constitute approximately 15% of peripheral T-cell lymphomas and approximately 2% of non-Hodgkin lymphomas in adults worldwide.⁵ Cutaneous involvement occurs in approximately half of AITL cases and can be the first manifestation of disease.² Skin findings are largely nonspecific, ranging from simple morbilliform rashes to erythroderma, at times manifesting with purpura.

Given this variability in the presentation of AITL, early diagnosis is challenging in the absence of more specific signs and symptoms.² It can conceivably be mistaken for common entities such as viral exanthems or drug eruptions, depending on the history and context. DRESS syndrome, a T cell-mediated, delayed type-IV hypersensitivity drug reaction can present in a manner highly similar to that of AITL, with cutaneous involvement (diffuse morbilliform rash, fever, facial edema, and generalized lymphadenopathy) and variable systemic involvement. Laboratory findings of eosinophilia, atypical lymphocytes, and thrombocytopenia also might be seen in both entities.⁶ Furthermore, the AITL in our patient was accompanied by electrolyte disturbances that were concerning for syndrome of inappropriate antidiuretic hormone secretion, a rare complication of patients with DRESS syndrome complicated by encephalitis.^7,8

Our patient met 4 RegiSCAR criteria for DRESS syndrome, warranting high clinical suspicion for an offending drug.⁹ DRESS syndrome can be caused by numerous medications—most commonly anticonvulsants, sulfonamides, antibiotics, allopurinol, and NSAIDs. A review of our patient’s medication list identified NSAIDs (including salsalate), entecavir, and amoxicillin, as possible culpable medications. Notably, the only new addition to the patient’s regimen was amoxicillin, which did not fit the typical 2- to 8-week timeline for a DRESS syndrome nidus.¹⁰ Our patient’s fever began well before the antibiotic was initiated, and skin findings appeared within 1 week after the course of amoxicillin was completed. Although there is documented variability in the latency of onset of DRESS syndrome following administration of a culprit medication,¹¹ it is critical to maintain a broad differential diagnosis to allow for further diagnostic information to be obtained, especially when the medication timeline does not align with the clinical presentation.

DRESS syndrome is far more common than AITL. Similarities in their clinical presentation pose a substantial challenge and often cause a delay in the diagnosis of AITL, which is made by excisional tissue biopsy, most commonly of a lymph node, with assessment of morphology and immunophenotyping. Histologic assessment of tissue reveals a polymorphous infiltrate of variably sized atypical lymphocytes with prominent arborizing HEVs as well as expanded populations of follicular dendritic cells that can be detected by CD21 staining. Cells express CD3 and CD4, variably express BCL6 (B-cell lymphoma 6 antigen) and CD10, and also may have partial or complete loss of expression of a subset of pan T-cell antigens (CD2, CD3, CD5, and CD7).^12-18

The treatment approach to AITL mirrors that of other nodal peripheral T-cell lymphomas, including chemotherapy and consideration of autologous stem-cell transplantation. Recent prospective trials of CHOP and CHOP-like chemotherapy have reported 3-year event-free survival and overall survival rates of 50% and 68%, respectively.¹⁹ Novel chemotherapeutic targets and gene-expression profiling are being investigated as potential therapeutic avenues.²⁰

Conclusion

DRESS syndrome and AITL can have near-identical presentations. Clinicians should maintain a high index of suspicion for AITL in patients with presumed DRESS syndrome whose rash does not improve with appropriate drug withdrawal and steroid therapy or who lack a strong offending medication history. In such cases, skin and lymph node biopsies should be performed as early as possible to evaluate for AITL and so that appropriate therapy can be initiated.

Angioimmunoblastic T-cell lymphoma (AITL) is a rare and aggressive lymphoma arising from follicular T-helper cells that predominantly affects older adults and carries a 5-year overall survival rate of 32%.¹ Notably, as many as 50% of AITL patients present with a skin rash in addition to the more common but nonspecific acute-onset generalized lymphadenopathy, hepatosplenomegaly, and anemia.² At presentation, most AITL patients are already at an advanced (III/IV) stage of disease.

Formerly known as angioimmunoblastic lymphadenopathy with dysproteinemia, AITL was once considered a benign entity that carried a risk for malignant transformation. As more cases have been identified and explored, this entity has been recategorized as a frank lymphoma.³ Therefore, it is critical that AITL be diagnosed and treated as early as possible.

We present the case of a 65-year-old man with clinical features that resembled drug reaction with eosinophilia and systemic symptoms (DRESS syndrome). After extensive workup, he was found to have AITL. This atypical case highlights the importance of maintaining a flexible differential diagnosis in patients with a persistent rash that does not improve with appropriate drug withdrawal and therapy.

Case Report

A 65-year-old Filipino man whose medical history was notable for hepatitis B that had been treated with entecavir for years without issue was admitted to the internal medicine service with fever of unknown origin and malaise of approximately 6 weeks’ duration. Six days prior to admission and 5 days after completing courses of the antiviral oseltamivir phosphate and amoxicillin for an upper respiratory tract infection and sinusitis, he developed worsening of an intermittently pruritic rash of approximately 1 month's duration. The dermatology department was consulted the day of hospital admission for evaluation of the rash. Chronic home medications included entecavir, lisinopril/hydrochlorothiazide, amlodipine, atorvastatin, metformin, salsalate, and over-the-counter nonsteroidal anti-inflammatory drugs (NSAIDs) as needed.

Physical examination was notable for mild erythema and scale distributed across the entire face; mild facial edema; and a blanchable, nonconfluent, macular erythema distributed across the trunk and upper and proximal lower extremities (Figure). In addition, the patient displayed conjunctival injection, pitting edema of the hands, and bilateral cervical and inguinal lymphadenopathy.

Photographs courtesy of James Contestable, MD (Camp Lejeune, North Carolina).

Laboratory tests revealed mild leukocytosis (11.6×10⁹/L, [reference range, 4.0–10.5×10⁹/L]), anemia (hemoglobin, 125 g/L (reference range, 138–170 g/L); hematocrit, 36.9%, [reference range, 40.0%–50.0%)], eosinophilia (1.07×10⁹/L [reference range, 0.00–0.70×10⁹/L)], hyponatremia, hypokalemia, and a mildly elevated creatinine level. Computed tomography and full-body positron-emission tomography (PET) scans during admission demonstrated diffuse lymphadenopathy. A skin biopsy from the left chest and a left inguinal lymph node biopsy also were performed.

Despite the lack of a clear medication trigger within the usual timeline for severe cutaneous drug-induced hypersensitivity reactions, DRESS syndrome was high on the differential diagnosis at the time of the initial presentation given the diffuse morbilliform eruption with pruritus, facial edema, eosinophilia, and lymphadenopathy.

Home medications were discontinued except for amlodipine, atorvastatin, and entecavir. The patient was treated symptomatically with topical steroids because it was believed that, if the clinical presentation represented DRESS syndrome, it was a mild variant that could be treated topically.⁴ His case was considered mild because of a lack of confirmed organ dysfunction and a mild protracted course.

After discharge following a 3-day inpatient stay, the patient was followed in the clinic weekly for 3 weeks without considerable change in the skin or laboratory findings. Discontinuation of entecavir was discussed and approved by his hepatologist.

Posthospitalization analysis of the punch biopsy specimen from the chest performed during the patient’s hospital stay revealed a superficial and deep dermal lymphoid infiltrate comprising CD3-, CD5-, and programmed cell death protein 1–positive cells with cytologic atypia in a perivascular distribution. Analysis of the lymph node biopsy specimen performed during the hospitalization showed effacement of the nodal architecture, a polymorphous lymphoid cell population with irregular nuclear contour, and abundant clear cytoplasm associated with high endothelial venules (HEVs). Cells of interest were positive for CD3, CD4, CD2, CD5, and CD7, with a subset staining positive for programmed cell death protein 1, inducible costimulator, CD10, and chemokine (C-X-C motif) ligand (CXCL) 13. CD21 demonstrated an expanded follicular dendritic cell meshwork in association with HEVs. Polymerase chain reaction revealed a clonal T-cell population. These findings of the skin and lymph node biopsies were consistent with AITL. Subsequent bone marrow biopsy with flow cytometry showed a normal CD4:CD8 ratio in T cells and no increase in natural killer cells.

Cyclophosphamide–hydroxydaunorubicin–Oncovin–prednisone (CHOP) chemotherapy was initiated; the patient completed a total of 6 cycles. He has had near resolution of the skin findings and is considered in remission based on a PET scan performed approximately 7 months after the initial presentation.

Comment

Angioimmunoblastic T-cell lymphoma is a rare peripheral T-cell lymphoma, part of a group of aggressive neoplasms that constitute approximately 15% of peripheral T-cell lymphomas and approximately 2% of non-Hodgkin lymphomas in adults worldwide.⁵ Cutaneous involvement occurs in approximately half of AITL cases and can be the first manifestation of disease.² Skin findings are largely nonspecific, ranging from simple morbilliform rashes to erythroderma, at times manifesting with purpura.

Given this variability in the presentation of AITL, early diagnosis is challenging in the absence of more specific signs and symptoms.² It can conceivably be mistaken for common entities such as viral exanthems or drug eruptions, depending on the history and context. DRESS syndrome, a T cell-mediated, delayed type-IV hypersensitivity drug reaction can present in a manner highly similar to that of AITL, with cutaneous involvement (diffuse morbilliform rash, fever, facial edema, and generalized lymphadenopathy) and variable systemic involvement. Laboratory findings of eosinophilia, atypical lymphocytes, and thrombocytopenia also might be seen in both entities.⁶ Furthermore, the AITL in our patient was accompanied by electrolyte disturbances that were concerning for syndrome of inappropriate antidiuretic hormone secretion, a rare complication of patients with DRESS syndrome complicated by encephalitis.^7,8

Our patient met 4 RegiSCAR criteria for DRESS syndrome, warranting high clinical suspicion for an offending drug.⁹ DRESS syndrome can be caused by numerous medications—most commonly anticonvulsants, sulfonamides, antibiotics, allopurinol, and NSAIDs. A review of our patient’s medication list identified NSAIDs (including salsalate), entecavir, and amoxicillin, as possible culpable medications. Notably, the only new addition to the patient’s regimen was amoxicillin, which did not fit the typical 2- to 8-week timeline for a DRESS syndrome nidus.¹⁰ Our patient’s fever began well before the antibiotic was initiated, and skin findings appeared within 1 week after the course of amoxicillin was completed. Although there is documented variability in the latency of onset of DRESS syndrome following administration of a culprit medication,¹¹ it is critical to maintain a broad differential diagnosis to allow for further diagnostic information to be obtained, especially when the medication timeline does not align with the clinical presentation.

DRESS syndrome is far more common than AITL. Similarities in their clinical presentation pose a substantial challenge and often cause a delay in the diagnosis of AITL, which is made by excisional tissue biopsy, most commonly of a lymph node, with assessment of morphology and immunophenotyping. Histologic assessment of tissue reveals a polymorphous infiltrate of variably sized atypical lymphocytes with prominent arborizing HEVs as well as expanded populations of follicular dendritic cells that can be detected by CD21 staining. Cells express CD3 and CD4, variably express BCL6 (B-cell lymphoma 6 antigen) and CD10, and also may have partial or complete loss of expression of a subset of pan T-cell antigens (CD2, CD3, CD5, and CD7).^12-18

The treatment approach to AITL mirrors that of other nodal peripheral T-cell lymphomas, including chemotherapy and consideration of autologous stem-cell transplantation. Recent prospective trials of CHOP and CHOP-like chemotherapy have reported 3-year event-free survival and overall survival rates of 50% and 68%, respectively.¹⁹ Novel chemotherapeutic targets and gene-expression profiling are being investigated as potential therapeutic avenues.²⁰

Conclusion

DRESS syndrome and AITL can have near-identical presentations. Clinicians should maintain a high index of suspicion for AITL in patients with presumed DRESS syndrome whose rash does not improve with appropriate drug withdrawal and steroid therapy or who lack a strong offending medication history. In such cases, skin and lymph node biopsies should be performed as early as possible to evaluate for AITL and so that appropriate therapy can be initiated.

To the Editor:

Dermatologists sometimes are consulted in the inpatient setting to rule out possible skin cancer. This scenario provides an opportunity to facilitate the diagnosis and treatment of cutaneous malignancy, often in patients who might not have sought regular outpatient dermatology care. Few studies have described the outcomes of inpatient biopsies to identify skin cancer.^1,2

Seeking to better understand the nature of these patient encounters, we reviewed all consultations at a medical center for which the referring physician suspected skin cancer rather than only those lesions that were biopsied by the dermatologist. We also collected data about subsequent treatment to better understand the outcomes of these patient encounters.

We conducted a retrospective review of inpatient dermatology referrals at an academic-affiliated tertiary medical center. We identified all patients who were provided with an inpatient dermatology consultation for suspected skin cancer or what was identified as a “skin lesion” between July 1, 2013, and July 1, 2019. We collected information on each patient’s sex, age at time of consultation, and race, as well as the specialty of the referring provider, lesion location, maximum diameter of the lesion, whether a biopsy was performed, where the biopsy was performed (inpatient or outpatient setting), clinical diagnosis, histopathologic diagnosis, and subsequent treatment.

The institutional review board at Eastern Virginia Medical School (Norfolk, Virginia) approved this study, and all protocol conformed to the ethical guidelines of the Declaration of Helsinki.

Thirty-eight patients met the inclusion criteria. Their characteristics are listed in the Table. Consultations for possible skin cancer accounted for 4% (38/950) of all inpatient dermatology consultations over the study period. Outcomes of the referrals are shown in the Figure. Consultations were received from 12 different physician specialties.

In the 38 patients, 47 lesions were identified; most (66% [31/47]) were on the head and neck. Twenty of 38 patients were found to have at least 1 biopsy-confirmed cutaneous malignancy (23 total tumors). Of those 23 identified malignancies, 10 were basal cell carcinoma, 11 squamous cell carcinoma, 1 malignant melanoma, and 1 anaplastic T-cell lymphoma. Of note, 17 of 23 (74%) identified cutaneous malignancies were 2.0 cm in diameter at biopsy or larger. Subsequently performed treatments for these patients included wide local excision (n=3), Mohs micrographic surgery (n=5), radiation therapy (n=3), topical fluorouracil (n=1), electrodesiccation and curettage (n=4), and chemotherapy or immunotherapy (n=2). Two patients who were diagnosed with skin cancer died of unrelated causes before treatment was completed.

In 10 of 38 patients, only nonmalignant entities were diagnosed, including seborrheic keratosis (n=6), benign melanocytic nevus (n=1), epidermal inclusion cyst (n=1), actinic keratosis (n=1), and radiation-induced necrosis (n=1). Of the 8 remaining patients, 4 were ultimately lost to follow-up before planned outpatient biopsy could be completed; 1 opted to follow up for biopsy at an unaffiliated outpatient dermatology provider. For 2 patients, the decision was made to forgo biopsy despite clinical suspicion of skin cancer because of overall poor health status, and 1 additional patient died before a planned outpatient biopsy could be performed.

In summary, approximately half of the inpatient dermatology consultations for suspected cutaneous malignancy resulted in a diagnosis of skin cancer. The patients in this population were admitted for a range of diagnoses, most unrelated to their cutaneous malignancy, suggesting that the inpatient setting offers the opportunity for physicians in a variety of specialties to help identify skin cancer that might otherwise be unaddressed and then facilitate management, whether ultimately in an inpatient or outpatient setting.

In many of these cases, it might be most appropriate to arrange subsequent outpatient dermatology follow-up after hospitalization, rather than making an inpatient consultation, as these situations usually are nonurgent and not directly related to hospitalization. However, in cases in which the lesion is directly related to admission, the lesion is advanced, there is concern for metastatic disease, or extenuating circumstances make outpatient follow-up difficult, inpatient dermatology consultation may be reasonable. There sometimes can be compelling reasons to expedite diagnosis and treatment as an inpatient.

In hospitalized, medically complex patients, in whom a new cutaneous malignancy is identified, dermatologists should discuss the situation thoughtfully with the patient, the patient’s family (when appropriate), and other physicians on the treatment team to determine the most appropriate course of action. In some cases, the most appropriate course might be to delay biopsy or treatment until the outpatient setting or to even defer further action completely when the prognosis is very limited. Consulting dermatologists must be mindful of patients’ overall medical situation in planning care for a cutaneous malignancy in these inpatient situations.

This study also highlights the surprising number of large-diameter, high-risk tumors identified in these scenarios. Limitations of this study include a relatively small sample size from a single facility that might not be representative of other practice settings and locations. Future multicenter studies could further explore the impact of inpatient dermatologic consultation on the diagnosis and management of skin cancer.

To the Editor:

Dermatologists sometimes are consulted in the inpatient setting to rule out possible skin cancer. This scenario provides an opportunity to facilitate the diagnosis and treatment of cutaneous malignancy, often in patients who might not have sought regular outpatient dermatology care. Few studies have described the outcomes of inpatient biopsies to identify skin cancer.^1,2

Seeking to better understand the nature of these patient encounters, we reviewed all consultations at a medical center for which the referring physician suspected skin cancer rather than only those lesions that were biopsied by the dermatologist. We also collected data about subsequent treatment to better understand the outcomes of these patient encounters.

We conducted a retrospective review of inpatient dermatology referrals at an academic-affiliated tertiary medical center. We identified all patients who were provided with an inpatient dermatology consultation for suspected skin cancer or what was identified as a “skin lesion” between July 1, 2013, and July 1, 2019. We collected information on each patient’s sex, age at time of consultation, and race, as well as the specialty of the referring provider, lesion location, maximum diameter of the lesion, whether a biopsy was performed, where the biopsy was performed (inpatient or outpatient setting), clinical diagnosis, histopathologic diagnosis, and subsequent treatment.

The institutional review board at Eastern Virginia Medical School (Norfolk, Virginia) approved this study, and all protocol conformed to the ethical guidelines of the Declaration of Helsinki.

Thirty-eight patients met the inclusion criteria. Their characteristics are listed in the Table. Consultations for possible skin cancer accounted for 4% (38/950) of all inpatient dermatology consultations over the study period. Outcomes of the referrals are shown in the Figure. Consultations were received from 12 different physician specialties.

In the 38 patients, 47 lesions were identified; most (66% [31/47]) were on the head and neck. Twenty of 38 patients were found to have at least 1 biopsy-confirmed cutaneous malignancy (23 total tumors). Of those 23 identified malignancies, 10 were basal cell carcinoma, 11 squamous cell carcinoma, 1 malignant melanoma, and 1 anaplastic T-cell lymphoma. Of note, 17 of 23 (74%) identified cutaneous malignancies were 2.0 cm in diameter at biopsy or larger. Subsequently performed treatments for these patients included wide local excision (n=3), Mohs micrographic surgery (n=5), radiation therapy (n=3), topical fluorouracil (n=1), electrodesiccation and curettage (n=4), and chemotherapy or immunotherapy (n=2). Two patients who were diagnosed with skin cancer died of unrelated causes before treatment was completed.

In 10 of 38 patients, only nonmalignant entities were diagnosed, including seborrheic keratosis (n=6), benign melanocytic nevus (n=1), epidermal inclusion cyst (n=1), actinic keratosis (n=1), and radiation-induced necrosis (n=1). Of the 8 remaining patients, 4 were ultimately lost to follow-up before planned outpatient biopsy could be completed; 1 opted to follow up for biopsy at an unaffiliated outpatient dermatology provider. For 2 patients, the decision was made to forgo biopsy despite clinical suspicion of skin cancer because of overall poor health status, and 1 additional patient died before a planned outpatient biopsy could be performed.

In summary, approximately half of the inpatient dermatology consultations for suspected cutaneous malignancy resulted in a diagnosis of skin cancer. The patients in this population were admitted for a range of diagnoses, most unrelated to their cutaneous malignancy, suggesting that the inpatient setting offers the opportunity for physicians in a variety of specialties to help identify skin cancer that might otherwise be unaddressed and then facilitate management, whether ultimately in an inpatient or outpatient setting.

In many of these cases, it might be most appropriate to arrange subsequent outpatient dermatology follow-up after hospitalization, rather than making an inpatient consultation, as these situations usually are nonurgent and not directly related to hospitalization. However, in cases in which the lesion is directly related to admission, the lesion is advanced, there is concern for metastatic disease, or extenuating circumstances make outpatient follow-up difficult, inpatient dermatology consultation may be reasonable. There sometimes can be compelling reasons to expedite diagnosis and treatment as an inpatient.

In hospitalized, medically complex patients, in whom a new cutaneous malignancy is identified, dermatologists should discuss the situation thoughtfully with the patient, the patient’s family (when appropriate), and other physicians on the treatment team to determine the most appropriate course of action. In some cases, the most appropriate course might be to delay biopsy or treatment until the outpatient setting or to even defer further action completely when the prognosis is very limited. Consulting dermatologists must be mindful of patients’ overall medical situation in planning care for a cutaneous malignancy in these inpatient situations.

This study also highlights the surprising number of large-diameter, high-risk tumors identified in these scenarios. Limitations of this study include a relatively small sample size from a single facility that might not be representative of other practice settings and locations. Future multicenter studies could further explore the impact of inpatient dermatologic consultation on the diagnosis and management of skin cancer.

To the Editor:

Dermatologists sometimes are consulted in the inpatient setting to rule out possible skin cancer. This scenario provides an opportunity to facilitate the diagnosis and treatment of cutaneous malignancy, often in patients who might not have sought regular outpatient dermatology care. Few studies have described the outcomes of inpatient biopsies to identify skin cancer.^1,2

Seeking to better understand the nature of these patient encounters, we reviewed all consultations at a medical center for which the referring physician suspected skin cancer rather than only those lesions that were biopsied by the dermatologist. We also collected data about subsequent treatment to better understand the outcomes of these patient encounters.

We conducted a retrospective review of inpatient dermatology referrals at an academic-affiliated tertiary medical center. We identified all patients who were provided with an inpatient dermatology consultation for suspected skin cancer or what was identified as a “skin lesion” between July 1, 2013, and July 1, 2019. We collected information on each patient’s sex, age at time of consultation, and race, as well as the specialty of the referring provider, lesion location, maximum diameter of the lesion, whether a biopsy was performed, where the biopsy was performed (inpatient or outpatient setting), clinical diagnosis, histopathologic diagnosis, and subsequent treatment.

The institutional review board at Eastern Virginia Medical School (Norfolk, Virginia) approved this study, and all protocol conformed to the ethical guidelines of the Declaration of Helsinki.

Thirty-eight patients met the inclusion criteria. Their characteristics are listed in the Table. Consultations for possible skin cancer accounted for 4% (38/950) of all inpatient dermatology consultations over the study period. Outcomes of the referrals are shown in the Figure. Consultations were received from 12 different physician specialties.

In the 38 patients, 47 lesions were identified; most (66% [31/47]) were on the head and neck. Twenty of 38 patients were found to have at least 1 biopsy-confirmed cutaneous malignancy (23 total tumors). Of those 23 identified malignancies, 10 were basal cell carcinoma, 11 squamous cell carcinoma, 1 malignant melanoma, and 1 anaplastic T-cell lymphoma. Of note, 17 of 23 (74%) identified cutaneous malignancies were 2.0 cm in diameter at biopsy or larger. Subsequently performed treatments for these patients included wide local excision (n=3), Mohs micrographic surgery (n=5), radiation therapy (n=3), topical fluorouracil (n=1), electrodesiccation and curettage (n=4), and chemotherapy or immunotherapy (n=2). Two patients who were diagnosed with skin cancer died of unrelated causes before treatment was completed.

In 10 of 38 patients, only nonmalignant entities were diagnosed, including seborrheic keratosis (n=6), benign melanocytic nevus (n=1), epidermal inclusion cyst (n=1), actinic keratosis (n=1), and radiation-induced necrosis (n=1). Of the 8 remaining patients, 4 were ultimately lost to follow-up before planned outpatient biopsy could be completed; 1 opted to follow up for biopsy at an unaffiliated outpatient dermatology provider. For 2 patients, the decision was made to forgo biopsy despite clinical suspicion of skin cancer because of overall poor health status, and 1 additional patient died before a planned outpatient biopsy could be performed.

In summary, approximately half of the inpatient dermatology consultations for suspected cutaneous malignancy resulted in a diagnosis of skin cancer. The patients in this population were admitted for a range of diagnoses, most unrelated to their cutaneous malignancy, suggesting that the inpatient setting offers the opportunity for physicians in a variety of specialties to help identify skin cancer that might otherwise be unaddressed and then facilitate management, whether ultimately in an inpatient or outpatient setting.

In many of these cases, it might be most appropriate to arrange subsequent outpatient dermatology follow-up after hospitalization, rather than making an inpatient consultation, as these situations usually are nonurgent and not directly related to hospitalization. However, in cases in which the lesion is directly related to admission, the lesion is advanced, there is concern for metastatic disease, or extenuating circumstances make outpatient follow-up difficult, inpatient dermatology consultation may be reasonable. There sometimes can be compelling reasons to expedite diagnosis and treatment as an inpatient.

In hospitalized, medically complex patients, in whom a new cutaneous malignancy is identified, dermatologists should discuss the situation thoughtfully with the patient, the patient’s family (when appropriate), and other physicians on the treatment team to determine the most appropriate course of action. In some cases, the most appropriate course might be to delay biopsy or treatment until the outpatient setting or to even defer further action completely when the prognosis is very limited. Consulting dermatologists must be mindful of patients’ overall medical situation in planning care for a cutaneous malignancy in these inpatient situations.

This study also highlights the surprising number of large-diameter, high-risk tumors identified in these scenarios. Limitations of this study include a relatively small sample size from a single facility that might not be representative of other practice settings and locations. Future multicenter studies could further explore the impact of inpatient dermatologic consultation on the diagnosis and management of skin cancer.

To the Editor:

A fixed drug eruption (FDE) is a well-documented form of cutaneous hypersensitivity that typically manifests as a sharply demarcated, dusky, round to oval, edematous, red-violaceous macule or patch on the skin and mucous membranes. The lesion often resolves with residual postinflammatory hyperpigmentation, most commonly as a reaction to ingested drugs or drug components.¹ Lesions generally occur at the same anatomic site with repeated exposure to the offending drug. Typically, a single site is affected, but additional sites with more generalized involvement have been reported to occur with subsequent exposure to the offending medication. The diagnosis usually is clinical, but histopathologic findings can help confirm the diagnosis in unusual presentations. We present a novel case of a patient with an FDE from medroxyprogesterone acetate, a contraceptive injection that contains the hormone progestin.

A 35-year-old woman presented to the dermatology clinic for evaluation of a lesion on the left lower buttock of 1 year’s duration. She reported periodic swelling and associated pruritus of the lesion. She denied any growth in size, and no other similar lesions were present. The patient reported a medication history of medroxyprogesterone acetate for birth control, but she denied any other prescription or over-the-counter medication, oral supplements, or recreational drug use. Upon further inquiry, she reported that the recurrence of symptoms appeared to coincide with each administration of medroxyprogesterone acetate, which occurred approximately every 3 months. The eruption cleared between injections and recurred in the same location following subsequent injections. The lesion appeared approximately 2 weeks after the first injection (approximately 1 year prior to presentation to dermatology) and within 2 to 3 days after each subsequent injection. Physical examination revealed a 2×2-cm, circular, slightly violaceous patch on the left buttock (Figure 1). A biopsy was recommended to aid in diagnosis, and the patient was offered a topical steroid for symptomatic relief. A punch biopsy revealed subtle interface dermatitis with superficial perivascular lymphoid infiltrate and marked pigmentary incontinence consistent with an FDE (Figure 2).

An FDE was first reported in 1889 by Bourns,² and over time more implicated agents and varying clinical presentations have been linked to the disease. The FDE can be accompanied by symptoms of pruritus or paresthesia. Most cases are devoid of systemic symptoms. An FDE can be located anywhere on the body, but it most frequently manifests on the lips, face, hands, feet, and genitalia. Although the eruption often heals with residual postinflammatory hyperpigmentation, a nonpigmenting FDE due to pseudoephedrine has been reported.³

Common culprits include antibiotics (eg, sulfonamides, trimethoprim, fluoroquinolones, tetracyclines), nonsteroidal anti-inflammatory medications (eg, naproxen sodium, ibuprofen, celecoxib), barbiturates, antimalarials, and anticonvulsants. Rare cases of FDE induced by foods and food additives also have been reported.⁴ Oral fluconazole, levocetirizine dihydrochloride, loperamide, and multivitamin-mineral preparations are other rare inducers of FDE.^5-8 In 2004, Ritter and Meffert⁹ described an FDE to the green dye used in inactive oral contraceptive pills. A similar case was reported by Rea et al¹⁰ that described an FDE from the inactive sugar pills in ethinyl estradiol and levonorgestrel, which is another combined oral contraceptive.

The time between ingestion of the offending agent and the manifestation of the disease usually is 1 to 2 weeks; however, upon subsequent exposure, the disease has been reported to manifest within hours.¹ CD8⁺ memory T cells have been shown to be major players in the development of FDE and can be found along the dermoepidermal junction as part of a delayed type IV hypersensitivity reaction.¹¹ Histopathology reveals superficial and deep interstitial and perivascular infiltrates consisting of lymphocytes with admixed eosinophils and possibly neutrophils in the dermis. In the epidermis, necrotic keratinocytes can be present. In rare cases, FDE may have atypical features, such as in generalized bullous FDE and nonpigmenting FDE, the latter of which more commonly is associated with pseudoephedrine.¹

The differential diagnosis for FDE includes erythema multiforme, Stevens-Johnson syndrome/toxic epidermal necrolysis, autoimmune progesterone dermatitis, and large plaque parapsoriasis. The number and morphology of lesions in erythema multiforme help differentiate it from FDE, as erythema multiforme presents with multiple targetoid lesions. The lesions of generalized bullous FDE can be similar to those of Stevens-Johnson syndrome/toxic epidermal necrolysis, and the pigmented patches of FDE can resemble large plaque parapsoriasis.

It is important to consider any medication ingested in the 1- to 2-week period before FDE onset, including over-the-counter medications, health food supplements, and prescription medications. Discontinuation of the implicated medication or any medication potentially cross-reacting with another medication is the most important step in management. Wound care may be needed for any bullous or eroded lesions. Lesions typically resolve within a few days to weeks of stopping the offending agent. Importantly, patients should be counseled on the secondary pigment alterations that may be persistent for several months. Other treatment for FDEs is aimed at symptomatic relief and may include topical corticosteroids and oral antihistamines.¹

Medroxyprogesterone acetate is a highly effective contraceptive drug with low rates of failure.¹² It is a weak androgenic progestin that is administered as a single 150-mg intramuscular injection every 3 months and inhibits gonadotropins. Common side effects include local injection-site reactions, unscheduled bleeding, amenorrhea, weight gain, headache, and mood changes. However, FDE has not been reported as an adverse effect to medroxyprogesterone acetate, both in official US Food and Drug Administration information and in the current literature.¹²

Autoimmune progesterone dermatitis (also known as progestin hypersensitivity) is a well-characterized cyclic hypersensitivity reaction to the hormone progesterone that occurs during the luteal phase of the menstrual cycle. It is known to have a variable clinical presentation including urticaria, erythema multiforme, eczema, and angioedema.¹³ Autoimmune progesterone dermatitis also has been reported to present as an FDE.^14-16 The onset of the cutaneous manifestation often starts a few days before the onset of menses, with spontaneous resolution occurring after the onset of menstruation. The mechanism by which endogenous progesterone or other secretory products become antigenic is unknown. It has been suggested that there is an alteration in the properties of the hormone that would predispose it to be antigenic as it would not be considered self. In 2001, Warin¹⁷ proposed the following diagnostic criteria for autoimmune progesterone dermatitis: (1) skin lesions associated with menstrual cycle (premenstrual flare); (2) a positive response to the progesterone intradermal or intramuscular test; and (3) symptomatic improvement after inhibiting progesterone secretion by suppressing ovulation.¹⁷ The treatment includes antiallergy medications, progesterone desensitization, omalizumab injection, and leuprolide acetate injection.

Our case represents FDE from medroxyprogesterone acetate. Although we did not formally investigate the antigenicity of the exogenous progesterone, we postulate that the pathophysiology likely is similar to an FDE associated with endogenous progesterone. This reasoning is supported by the time course of the patient’s lesion as well as the worsening of symptoms in the days following the administration of the medication. Additionally, the patient had no history of skin lesions prior to the initiation of medroxyprogesterone acetate or similar lesions associated with her menstrual cycles.

A careful and detailed review of medication history is necessary to evaluate FDEs. Our case emphasizes that not only endogenous but also exogenous forms of progesterone may cause hypersensitivity, leading to an FDE. With more than 2 million prescriptions of medroxyprogesterone acetate written every year, dermatologists should be aware of the rare but potential risk for an FDE in patients using this medication.¹⁸

To the Editor:

A fixed drug eruption (FDE) is a well-documented form of cutaneous hypersensitivity that typically manifests as a sharply demarcated, dusky, round to oval, edematous, red-violaceous macule or patch on the skin and mucous membranes. The lesion often resolves with residual postinflammatory hyperpigmentation, most commonly as a reaction to ingested drugs or drug components.¹ Lesions generally occur at the same anatomic site with repeated exposure to the offending drug. Typically, a single site is affected, but additional sites with more generalized involvement have been reported to occur with subsequent exposure to the offending medication. The diagnosis usually is clinical, but histopathologic findings can help confirm the diagnosis in unusual presentations. We present a novel case of a patient with an FDE from medroxyprogesterone acetate, a contraceptive injection that contains the hormone progestin.

A 35-year-old woman presented to the dermatology clinic for evaluation of a lesion on the left lower buttock of 1 year’s duration. She reported periodic swelling and associated pruritus of the lesion. She denied any growth in size, and no other similar lesions were present. The patient reported a medication history of medroxyprogesterone acetate for birth control, but she denied any other prescription or over-the-counter medication, oral supplements, or recreational drug use. Upon further inquiry, she reported that the recurrence of symptoms appeared to coincide with each administration of medroxyprogesterone acetate, which occurred approximately every 3 months. The eruption cleared between injections and recurred in the same location following subsequent injections. The lesion appeared approximately 2 weeks after the first injection (approximately 1 year prior to presentation to dermatology) and within 2 to 3 days after each subsequent injection. Physical examination revealed a 2×2-cm, circular, slightly violaceous patch on the left buttock (Figure 1). A biopsy was recommended to aid in diagnosis, and the patient was offered a topical steroid for symptomatic relief. A punch biopsy revealed subtle interface dermatitis with superficial perivascular lymphoid infiltrate and marked pigmentary incontinence consistent with an FDE (Figure 2).

An FDE was first reported in 1889 by Bourns,² and over time more implicated agents and varying clinical presentations have been linked to the disease. The FDE can be accompanied by symptoms of pruritus or paresthesia. Most cases are devoid of systemic symptoms. An FDE can be located anywhere on the body, but it most frequently manifests on the lips, face, hands, feet, and genitalia. Although the eruption often heals with residual postinflammatory hyperpigmentation, a nonpigmenting FDE due to pseudoephedrine has been reported.³

Common culprits include antibiotics (eg, sulfonamides, trimethoprim, fluoroquinolones, tetracyclines), nonsteroidal anti-inflammatory medications (eg, naproxen sodium, ibuprofen, celecoxib), barbiturates, antimalarials, and anticonvulsants. Rare cases of FDE induced by foods and food additives also have been reported.⁴ Oral fluconazole, levocetirizine dihydrochloride, loperamide, and multivitamin-mineral preparations are other rare inducers of FDE.^5-8 In 2004, Ritter and Meffert⁹ described an FDE to the green dye used in inactive oral contraceptive pills. A similar case was reported by Rea et al¹⁰ that described an FDE from the inactive sugar pills in ethinyl estradiol and levonorgestrel, which is another combined oral contraceptive.

The time between ingestion of the offending agent and the manifestation of the disease usually is 1 to 2 weeks; however, upon subsequent exposure, the disease has been reported to manifest within hours.¹ CD8⁺ memory T cells have been shown to be major players in the development of FDE and can be found along the dermoepidermal junction as part of a delayed type IV hypersensitivity reaction.¹¹ Histopathology reveals superficial and deep interstitial and perivascular infiltrates consisting of lymphocytes with admixed eosinophils and possibly neutrophils in the dermis. In the epidermis, necrotic keratinocytes can be present. In rare cases, FDE may have atypical features, such as in generalized bullous FDE and nonpigmenting FDE, the latter of which more commonly is associated with pseudoephedrine.¹

The differential diagnosis for FDE includes erythema multiforme, Stevens-Johnson syndrome/toxic epidermal necrolysis, autoimmune progesterone dermatitis, and large plaque parapsoriasis. The number and morphology of lesions in erythema multiforme help differentiate it from FDE, as erythema multiforme presents with multiple targetoid lesions. The lesions of generalized bullous FDE can be similar to those of Stevens-Johnson syndrome/toxic epidermal necrolysis, and the pigmented patches of FDE can resemble large plaque parapsoriasis.

It is important to consider any medication ingested in the 1- to 2-week period before FDE onset, including over-the-counter medications, health food supplements, and prescription medications. Discontinuation of the implicated medication or any medication potentially cross-reacting with another medication is the most important step in management. Wound care may be needed for any bullous or eroded lesions. Lesions typically resolve within a few days to weeks of stopping the offending agent. Importantly, patients should be counseled on the secondary pigment alterations that may be persistent for several months. Other treatment for FDEs is aimed at symptomatic relief and may include topical corticosteroids and oral antihistamines.¹

Medroxyprogesterone acetate is a highly effective contraceptive drug with low rates of failure.¹² It is a weak androgenic progestin that is administered as a single 150-mg intramuscular injection every 3 months and inhibits gonadotropins. Common side effects include local injection-site reactions, unscheduled bleeding, amenorrhea, weight gain, headache, and mood changes. However, FDE has not been reported as an adverse effect to medroxyprogesterone acetate, both in official US Food and Drug Administration information and in the current literature.¹²

Autoimmune progesterone dermatitis (also known as progestin hypersensitivity) is a well-characterized cyclic hypersensitivity reaction to the hormone progesterone that occurs during the luteal phase of the menstrual cycle. It is known to have a variable clinical presentation including urticaria, erythema multiforme, eczema, and angioedema.¹³ Autoimmune progesterone dermatitis also has been reported to present as an FDE.^14-16 The onset of the cutaneous manifestation often starts a few days before the onset of menses, with spontaneous resolution occurring after the onset of menstruation. The mechanism by which endogenous progesterone or other secretory products become antigenic is unknown. It has been suggested that there is an alteration in the properties of the hormone that would predispose it to be antigenic as it would not be considered self. In 2001, Warin¹⁷ proposed the following diagnostic criteria for autoimmune progesterone dermatitis: (1) skin lesions associated with menstrual cycle (premenstrual flare); (2) a positive response to the progesterone intradermal or intramuscular test; and (3) symptomatic improvement after inhibiting progesterone secretion by suppressing ovulation.¹⁷ The treatment includes antiallergy medications, progesterone desensitization, omalizumab injection, and leuprolide acetate injection.

Our case represents FDE from medroxyprogesterone acetate. Although we did not formally investigate the antigenicity of the exogenous progesterone, we postulate that the pathophysiology likely is similar to an FDE associated with endogenous progesterone. This reasoning is supported by the time course of the patient’s lesion as well as the worsening of symptoms in the days following the administration of the medication. Additionally, the patient had no history of skin lesions prior to the initiation of medroxyprogesterone acetate or similar lesions associated with her menstrual cycles.

A careful and detailed review of medication history is necessary to evaluate FDEs. Our case emphasizes that not only endogenous but also exogenous forms of progesterone may cause hypersensitivity, leading to an FDE. With more than 2 million prescriptions of medroxyprogesterone acetate written every year, dermatologists should be aware of the rare but potential risk for an FDE in patients using this medication.¹⁸

To the Editor:

A fixed drug eruption (FDE) is a well-documented form of cutaneous hypersensitivity that typically manifests as a sharply demarcated, dusky, round to oval, edematous, red-violaceous macule or patch on the skin and mucous membranes. The lesion often resolves with residual postinflammatory hyperpigmentation, most commonly as a reaction to ingested drugs or drug components.¹ Lesions generally occur at the same anatomic site with repeated exposure to the offending drug. Typically, a single site is affected, but additional sites with more generalized involvement have been reported to occur with subsequent exposure to the offending medication. The diagnosis usually is clinical, but histopathologic findings can help confirm the diagnosis in unusual presentations. We present a novel case of a patient with an FDE from medroxyprogesterone acetate, a contraceptive injection that contains the hormone progestin.

A 35-year-old woman presented to the dermatology clinic for evaluation of a lesion on the left lower buttock of 1 year’s duration. She reported periodic swelling and associated pruritus of the lesion. She denied any growth in size, and no other similar lesions were present. The patient reported a medication history of medroxyprogesterone acetate for birth control, but she denied any other prescription or over-the-counter medication, oral supplements, or recreational drug use. Upon further inquiry, she reported that the recurrence of symptoms appeared to coincide with each administration of medroxyprogesterone acetate, which occurred approximately every 3 months. The eruption cleared between injections and recurred in the same location following subsequent injections. The lesion appeared approximately 2 weeks after the first injection (approximately 1 year prior to presentation to dermatology) and within 2 to 3 days after each subsequent injection. Physical examination revealed a 2×2-cm, circular, slightly violaceous patch on the left buttock (Figure 1). A biopsy was recommended to aid in diagnosis, and the patient was offered a topical steroid for symptomatic relief. A punch biopsy revealed subtle interface dermatitis with superficial perivascular lymphoid infiltrate and marked pigmentary incontinence consistent with an FDE (Figure 2).

An FDE was first reported in 1889 by Bourns,² and over time more implicated agents and varying clinical presentations have been linked to the disease. The FDE can be accompanied by symptoms of pruritus or paresthesia. Most cases are devoid of systemic symptoms. An FDE can be located anywhere on the body, but it most frequently manifests on the lips, face, hands, feet, and genitalia. Although the eruption often heals with residual postinflammatory hyperpigmentation, a nonpigmenting FDE due to pseudoephedrine has been reported.³

Common culprits include antibiotics (eg, sulfonamides, trimethoprim, fluoroquinolones, tetracyclines), nonsteroidal anti-inflammatory medications (eg, naproxen sodium, ibuprofen, celecoxib), barbiturates, antimalarials, and anticonvulsants. Rare cases of FDE induced by foods and food additives also have been reported.⁴ Oral fluconazole, levocetirizine dihydrochloride, loperamide, and multivitamin-mineral preparations are other rare inducers of FDE.^5-8 In 2004, Ritter and Meffert⁹ described an FDE to the green dye used in inactive oral contraceptive pills. A similar case was reported by Rea et al¹⁰ that described an FDE from the inactive sugar pills in ethinyl estradiol and levonorgestrel, which is another combined oral contraceptive.

The time between ingestion of the offending agent and the manifestation of the disease usually is 1 to 2 weeks; however, upon subsequent exposure, the disease has been reported to manifest within hours.¹ CD8⁺ memory T cells have been shown to be major players in the development of FDE and can be found along the dermoepidermal junction as part of a delayed type IV hypersensitivity reaction.¹¹ Histopathology reveals superficial and deep interstitial and perivascular infiltrates consisting of lymphocytes with admixed eosinophils and possibly neutrophils in the dermis. In the epidermis, necrotic keratinocytes can be present. In rare cases, FDE may have atypical features, such as in generalized bullous FDE and nonpigmenting FDE, the latter of which more commonly is associated with pseudoephedrine.¹

The differential diagnosis for FDE includes erythema multiforme, Stevens-Johnson syndrome/toxic epidermal necrolysis, autoimmune progesterone dermatitis, and large plaque parapsoriasis. The number and morphology of lesions in erythema multiforme help differentiate it from FDE, as erythema multiforme presents with multiple targetoid lesions. The lesions of generalized bullous FDE can be similar to those of Stevens-Johnson syndrome/toxic epidermal necrolysis, and the pigmented patches of FDE can resemble large plaque parapsoriasis.

It is important to consider any medication ingested in the 1- to 2-week period before FDE onset, including over-the-counter medications, health food supplements, and prescription medications. Discontinuation of the implicated medication or any medication potentially cross-reacting with another medication is the most important step in management. Wound care may be needed for any bullous or eroded lesions. Lesions typically resolve within a few days to weeks of stopping the offending agent. Importantly, patients should be counseled on the secondary pigment alterations that may be persistent for several months. Other treatment for FDEs is aimed at symptomatic relief and may include topical corticosteroids and oral antihistamines.¹

Medroxyprogesterone acetate is a highly effective contraceptive drug with low rates of failure.¹² It is a weak androgenic progestin that is administered as a single 150-mg intramuscular injection every 3 months and inhibits gonadotropins. Common side effects include local injection-site reactions, unscheduled bleeding, amenorrhea, weight gain, headache, and mood changes. However, FDE has not been reported as an adverse effect to medroxyprogesterone acetate, both in official US Food and Drug Administration information and in the current literature.¹²

Autoimmune progesterone dermatitis (also known as progestin hypersensitivity) is a well-characterized cyclic hypersensitivity reaction to the hormone progesterone that occurs during the luteal phase of the menstrual cycle. It is known to have a variable clinical presentation including urticaria, erythema multiforme, eczema, and angioedema.¹³ Autoimmune progesterone dermatitis also has been reported to present as an FDE.^14-16 The onset of the cutaneous manifestation often starts a few days before the onset of menses, with spontaneous resolution occurring after the onset of menstruation. The mechanism by which endogenous progesterone or other secretory products become antigenic is unknown. It has been suggested that there is an alteration in the properties of the hormone that would predispose it to be antigenic as it would not be considered self. In 2001, Warin¹⁷ proposed the following diagnostic criteria for autoimmune progesterone dermatitis: (1) skin lesions associated with menstrual cycle (premenstrual flare); (2) a positive response to the progesterone intradermal or intramuscular test; and (3) symptomatic improvement after inhibiting progesterone secretion by suppressing ovulation.¹⁷ The treatment includes antiallergy medications, progesterone desensitization, omalizumab injection, and leuprolide acetate injection.

Our case represents FDE from medroxyprogesterone acetate. Although we did not formally investigate the antigenicity of the exogenous progesterone, we postulate that the pathophysiology likely is similar to an FDE associated with endogenous progesterone. This reasoning is supported by the time course of the patient’s lesion as well as the worsening of symptoms in the days following the administration of the medication. Additionally, the patient had no history of skin lesions prior to the initiation of medroxyprogesterone acetate or similar lesions associated with her menstrual cycles.

A careful and detailed review of medication history is necessary to evaluate FDEs. Our case emphasizes that not only endogenous but also exogenous forms of progesterone may cause hypersensitivity, leading to an FDE. With more than 2 million prescriptions of medroxyprogesterone acetate written every year, dermatologists should be aware of the rare but potential risk for an FDE in patients using this medication.¹⁸

The Diagnosis: Odontogenic Cutaneous Sinus Tract

In our patient, panoramic radiography showed a radiolucency in the periapex of the mandibular first molar (Figure 1). Ultrasonography depicted a hypoechoic band that originated from the cutaneous lesion and extended through the subcutaneous tissue to the defective alveolar bone, suggesting odontogenic inflammation (Figure 2).¹ The infected pulp was removed, and the purulent nodules then disappeared.

The dental etiology of odontogenic cutaneous sinus tracts can be confirmed by panoramic radiography and ultrasonography. The odontogenic sinus path can be clearly observed via radiography by injecting or inserting a radiopaque substance into the sinus tract.² Effective treatment of the diseased tooth is removal of the infected pulp, performance of a root canal to eliminate infection, closure and filling of the root canal, and repair of the crown. Once the source of infection is eliminated, the sinus typically subsides within 2 weeks. When residual skin retreats or scars are present, cosmetic surgery can be performed to improve the appearance.^3,4

Odontogenic cutaneous sinus tracts usually are caused by a route of drainage from a chronic apical abscess. They follow a path of least resistance through the alveolar bone and periosteum, spreading into the surrounding soft tissues. With the formation of abscesses, sinus tracts will erupt intraorally or cutaneously, depending on the relationship of the posterior tooth apices to the mandibular attachments of the mylohyoid and buccinator muscles and the maxillary attachment of the buccinator.^2,5 Clinically, cutaneous lesions present as nodules, cysts, or dimples that have attached to deep tissues through the sinus tract. Half of patients may have no dental symptoms and often are misdiagnosed with nonodontogenic lesions. Subsequent improper treatments, such as repeated use of antibiotics, multiple biopsies, surgical excision, and chemotherapy, often are repeated and ineffective.⁶ The most common cause of chronic cutaneous sinus tracts in the face and neck is a chronically draining dental infection.^2,5 A thorough history is necessary when odontogenic cutaneous sinuses are suspected. Toothache before the development of the sinus tract is an important diagnostic clue.

Pyogenic granuloma, syringocystadenoma papilliferum, osteomyelitis, infected epidermoid cyst, actinomycoses, and salivary gland fistula also should be considered in the differential diagnosis.^7-10 Pyogenic granuloma (also known as lobular capillary hemangioma) is a benign overgrowth of capillaries showing a vascular phenotype that usually occurs as a response to different stimulating factors such as local stimuli, trauma, or hormonal factors. Clinically, pyogenic granuloma presents as a red, solitary, painless nodule on the face or distal extremities.^11,12 Syringocystadenoma papilliferum is a benign adnexal proliferation with apocrine differentiation that usually presents as a hairless papillomatous plaque or nodule measuring 1 to 4 cm in diameter and often is first noted at birth or during early childhood.⁷ Osteomyelitis is progressive inflammation of the periosteum and bone marrow that rapidly breaks through the periosteum and spreads to surrounding areas. The mandible is the most susceptible bone for facial osteomyelitis.⁸ Epidermoid cysts are formed by the proliferation of epidermal cells within a circumscribed dermal space. Infection of the cysts is characterized by redness, swelling, heat, and pain. As the infection progresses, suppurative inflammation develops, leading to local liquefaction and abscesses.⁹

This case was initially misdiagnosed as infectious skin lesions by outside clinicians. Multiple surgical treatments and long-term antibiotic therapy were attempted before the correct diagnosis was made. The clinical diagnosis of odontogenic cutaneous sinus tracts is challenging due to the variety of affected sites and clinical signs. Ultrasonography should be performed as early as possible to identify the disease and avoid unnecessary surgery. For appropriate dental therapy, close liaison with the stomatology department is warranted.

The Diagnosis: Odontogenic Cutaneous Sinus Tract

In our patient, panoramic radiography showed a radiolucency in the periapex of the mandibular first molar (Figure 1). Ultrasonography depicted a hypoechoic band that originated from the cutaneous lesion and extended through the subcutaneous tissue to the defective alveolar bone, suggesting odontogenic inflammation (Figure 2).¹ The infected pulp was removed, and the purulent nodules then disappeared.

The dental etiology of odontogenic cutaneous sinus tracts can be confirmed by panoramic radiography and ultrasonography. The odontogenic sinus path can be clearly observed via radiography by injecting or inserting a radiopaque substance into the sinus tract.² Effective treatment of the diseased tooth is removal of the infected pulp, performance of a root canal to eliminate infection, closure and filling of the root canal, and repair of the crown. Once the source of infection is eliminated, the sinus typically subsides within 2 weeks. When residual skin retreats or scars are present, cosmetic surgery can be performed to improve the appearance.^3,4

Odontogenic cutaneous sinus tracts usually are caused by a route of drainage from a chronic apical abscess. They follow a path of least resistance through the alveolar bone and periosteum, spreading into the surrounding soft tissues. With the formation of abscesses, sinus tracts will erupt intraorally or cutaneously, depending on the relationship of the posterior tooth apices to the mandibular attachments of the mylohyoid and buccinator muscles and the maxillary attachment of the buccinator.^2,5 Clinically, cutaneous lesions present as nodules, cysts, or dimples that have attached to deep tissues through the sinus tract. Half of patients may have no dental symptoms and often are misdiagnosed with nonodontogenic lesions. Subsequent improper treatments, such as repeated use of antibiotics, multiple biopsies, surgical excision, and chemotherapy, often are repeated and ineffective.⁶ The most common cause of chronic cutaneous sinus tracts in the face and neck is a chronically draining dental infection.^2,5 A thorough history is necessary when odontogenic cutaneous sinuses are suspected. Toothache before the development of the sinus tract is an important diagnostic clue.

Pyogenic granuloma, syringocystadenoma papilliferum, osteomyelitis, infected epidermoid cyst, actinomycoses, and salivary gland fistula also should be considered in the differential diagnosis.^7-10 Pyogenic granuloma (also known as lobular capillary hemangioma) is a benign overgrowth of capillaries showing a vascular phenotype that usually occurs as a response to different stimulating factors such as local stimuli, trauma, or hormonal factors. Clinically, pyogenic granuloma presents as a red, solitary, painless nodule on the face or distal extremities.^11,12 Syringocystadenoma papilliferum is a benign adnexal proliferation with apocrine differentiation that usually presents as a hairless papillomatous plaque or nodule measuring 1 to 4 cm in diameter and often is first noted at birth or during early childhood.⁷ Osteomyelitis is progressive inflammation of the periosteum and bone marrow that rapidly breaks through the periosteum and spreads to surrounding areas. The mandible is the most susceptible bone for facial osteomyelitis.⁸ Epidermoid cysts are formed by the proliferation of epidermal cells within a circumscribed dermal space. Infection of the cysts is characterized by redness, swelling, heat, and pain. As the infection progresses, suppurative inflammation develops, leading to local liquefaction and abscesses.⁹

This case was initially misdiagnosed as infectious skin lesions by outside clinicians. Multiple surgical treatments and long-term antibiotic therapy were attempted before the correct diagnosis was made. The clinical diagnosis of odontogenic cutaneous sinus tracts is challenging due to the variety of affected sites and clinical signs. Ultrasonography should be performed as early as possible to identify the disease and avoid unnecessary surgery. For appropriate dental therapy, close liaison with the stomatology department is warranted.

The Diagnosis: Odontogenic Cutaneous Sinus Tract

In our patient, panoramic radiography showed a radiolucency in the periapex of the mandibular first molar (Figure 1). Ultrasonography depicted a hypoechoic band that originated from the cutaneous lesion and extended through the subcutaneous tissue to the defective alveolar bone, suggesting odontogenic inflammation (Figure 2).¹ The infected pulp was removed, and the purulent nodules then disappeared.

The dental etiology of odontogenic cutaneous sinus tracts can be confirmed by panoramic radiography and ultrasonography. The odontogenic sinus path can be clearly observed via radiography by injecting or inserting a radiopaque substance into the sinus tract.² Effective treatment of the diseased tooth is removal of the infected pulp, performance of a root canal to eliminate infection, closure and filling of the root canal, and repair of the crown. Once the source of infection is eliminated, the sinus typically subsides within 2 weeks. When residual skin retreats or scars are present, cosmetic surgery can be performed to improve the appearance.^3,4

Odontogenic cutaneous sinus tracts usually are caused by a route of drainage from a chronic apical abscess. They follow a path of least resistance through the alveolar bone and periosteum, spreading into the surrounding soft tissues. With the formation of abscesses, sinus tracts will erupt intraorally or cutaneously, depending on the relationship of the posterior tooth apices to the mandibular attachments of the mylohyoid and buccinator muscles and the maxillary attachment of the buccinator.^2,5 Clinically, cutaneous lesions present as nodules, cysts, or dimples that have attached to deep tissues through the sinus tract. Half of patients may have no dental symptoms and often are misdiagnosed with nonodontogenic lesions. Subsequent improper treatments, such as repeated use of antibiotics, multiple biopsies, surgical excision, and chemotherapy, often are repeated and ineffective.⁶ The most common cause of chronic cutaneous sinus tracts in the face and neck is a chronically draining dental infection.^2,5 A thorough history is necessary when odontogenic cutaneous sinuses are suspected. Toothache before the development of the sinus tract is an important diagnostic clue.

Pyogenic granuloma, syringocystadenoma papilliferum, osteomyelitis, infected epidermoid cyst, actinomycoses, and salivary gland fistula also should be considered in the differential diagnosis.^7-10 Pyogenic granuloma (also known as lobular capillary hemangioma) is a benign overgrowth of capillaries showing a vascular phenotype that usually occurs as a response to different stimulating factors such as local stimuli, trauma, or hormonal factors. Clinically, pyogenic granuloma presents as a red, solitary, painless nodule on the face or distal extremities.^11,12 Syringocystadenoma papilliferum is a benign adnexal proliferation with apocrine differentiation that usually presents as a hairless papillomatous plaque or nodule measuring 1 to 4 cm in diameter and often is first noted at birth or during early childhood.⁷ Osteomyelitis is progressive inflammation of the periosteum and bone marrow that rapidly breaks through the periosteum and spreads to surrounding areas. The mandible is the most susceptible bone for facial osteomyelitis.⁸ Epidermoid cysts are formed by the proliferation of epidermal cells within a circumscribed dermal space. Infection of the cysts is characterized by redness, swelling, heat, and pain. As the infection progresses, suppurative inflammation develops, leading to local liquefaction and abscesses.⁹

This case was initially misdiagnosed as infectious skin lesions by outside clinicians. Multiple surgical treatments and long-term antibiotic therapy were attempted before the correct diagnosis was made. The clinical diagnosis of odontogenic cutaneous sinus tracts is challenging due to the variety of affected sites and clinical signs. Ultrasonography should be performed as early as possible to identify the disease and avoid unnecessary surgery. For appropriate dental therapy, close liaison with the stomatology department is warranted.

Mycosis fungoides (MF) and Sézary syndrome (SS) are non-Hodgkin T-cell lymphomas that make up the majority of cutaneous T-cell lymphomas. These conditions commonly affect Black patients, with an incidence rate of 12.6 cases of cutaneous T-cell lymphomas per million individuals vs 9.8 per million individuals in non–skin of color (SoC) patients.¹ However, educational resources tend to focus on the clinical manifestations of MF/SS in lighter skin types, describing MF as erythematous patches, plaques, or tumors presenting in non–sun-exposed areas of the skin and SS as generalized erythroderma.² Skin of color, comprised of Fitzpatrick skin types (FSTs) IV to VI,³ is poorly represented across dermatology textbooks,^4,5 medical student resources,⁶ and peer-reviewed publications,⁷ raising awareness for the need to address this disparity.

Skin of color patients with MF/SS display variable morphologies, including features such as hyperpigmentation and hypopigmentation,⁸ the latter being exceedingly rare in non-SoC patients.⁹ Familiarity with these differences among providers is essential to allow for equitable diagnosis and treatment across all skin types, especially in light of data predicting that by 2044 more than 50% of the US population will be people of color.¹⁰ Patients with SoC are of many ethnic and racial backgrounds, including Black, Hispanic, American Indian, Pacific Islander, and Asian.¹¹

Along with morphologic differences, there also are several racial disparities in the prognosis and survival of patients with MF/SS. Black patients diagnosed with MF present with greater body surface area affected, and Black women with MF have reduced survival rates compared to their White counterparts.¹² Given these racial disparities in survival and representation in educational resources, we aimed to quantify the frequency of various morphologic characteristics of MF/SS in patients with SoC vs non-SoC patients to facilitate better recognition of early MF/SS in SoC patients by medical providers.

Methods

We performed a retrospective chart review following approval from the institutional review board at Northwestern University (Chicago, Illinois). We identified all patients with FSTs IV to VI and biopsy-proven MF/SS who had been clinically photographed in our clinic from January 1998 to December 2019. Only photographs that were high quality enough to review morphologic features were included in our review. Fitzpatrick skin type was determined based on electronic medical record documentation. If photographs were available from multiple visits for the same patient, only those showing posttreatment nonactive lesions were included. Additionally, 36 patients with FSTs I to III (non-SoC) and biopsy-proven MF/SS were included in our review as a comparison with the SoC cohort. The primary outcomes for this study included the presence of scale, erythema, hyperpigmentation, hypopigmentation, violaceous color, lichenification, silver hue, dyschromia, alopecia, poikiloderma, atrophy, and ulceration in active lesions. Dyschromia was defined by the presence of both hypopigmentation and hyperpigmentation. Poikiloderma was defined by hypopigmentation and hyperpigmentation, telangiectasia, and atrophy. Secondary outcomes included evaluation of those same characteristics in posttreatment nonactive lesions. All photographs were independently assessed by 3 authors (M.L.E., C.J.W., J.M.M.), and discrepancies were resolved by further review of the photograph in question and discussion.

Statistical Analysis—Summary statistics were applied to describe demographic and clinical characteristics. The χ² test was used for categorical variables. Results achieving P<.05 were considered statistically significant.

Results

We reviewed photographs of 111 patients across all skin types (8, FST I; 12, FST II; 16, FST III; 17, FST IV; 44, FST V; 14, FST VI). The cohort was 47% female, and the mean age was 49.7 years (range, 15–86 years). The majority of the cohort had early-stage MF (stage IA or IB). There were more cases of SS in the SoC cohort than the non-SoC cohort (Table). Only 5 photographs had discrepancies and required discussion among the reviewers to achieve consensus.

Regarding morphologic characteristics in active lesions (Figure 1), scale was present in almost all patients (99% in SoC, 94% in non-SoC). Erythema was present in nearly all non-SoC patients (94%) but only in 69% of SoC patients (P=.003). Poikiloderma also was found to be present at higher frequencies in non-SoC patients compared with SoC patients (19% and 4%, respectively [P=.008]). However, hyperpigmentation (80% vs 39%), lichenification (43% vs 17%), and silver hue (25% vs 3%) were more common in SoC patients than non-SoC patients (P<.05). There were no significant differences in the remaining features, including hypopigmentation (39% vs 25%), dyschromia (24% vs 19%), violaceous color (44% vs 25%), atrophy (11% vs 22%), alopecia (23% vs 31%), and ulceration (16% vs 8%) between SoC and non-SoC patients (P>.05). Photographs of MF in patients with SoC can be seen in Figure 2.

Posttreatment (nonactive) photographs were available for 26 patients (6 non-SoC, 20 SoC). We found that across all FST groups, hyperpigmentation was more common than hypopigmentation in areas of previously active disease. Statistical analysis was not completed given that few non-SoC photographs were available for comparison.

Comment

This qualitative review demonstrates the heterogeneity of MF/SS in SoC patients and that these conditions do not present in this population with the classic erythematous patches and plaques found in non-SoC patients. We found that hyperpigmentation, lichenification, and silver hue were present at higher rates in patients with FSTs IV to VI compared to those with FSTs I to III, which had higher rates of erythema and poikiloderma. Familiarity with these morphologic features along with increased exposure to clinical photographs of MF/SS in SoC patients will aid in the visual recognition required for this diagnosis, since erythema is harder to identify in darker skin types. Recognizing the unique findings of MF in patients with SoC as well as in patients with lighter skin types will enable earlier diagnosis and treatment of MF/SS across all skin types. If MF is diagnosed and treated early, life expectancy is similar to that of patients without MF.¹³ However, the 5-year survival rate for advanced-stage MF/SS is 52% across all skin types, and studies have found that Black patients with advanced-stage disease have worse outcomes despite accounting for demographic factors and tumor stage.^14,15 Given the worse outcomes in SoC patients with advanced-stage MF/SS, earlier diagnosis could help address this disparity.^8,13,14 Similar morphologic features could be used in diagnosing other inflammatory conditions; studies have shown that the lack of recognition of erythema in Black children has led to delayed diagnosis of atopic dermatitis and subsequent inadequate treatment.^16,17

The morphologic presentation of MF/SS in SoC patients also can influence an optimal treatment plan for this population. Hypopigmented MF responds better to phototherapy than hyperpigmented MF, as phototherapy has been shown to have decreased efficacy with increasing FST.¹⁸ Therefore, for patients with FSTs IV to VI, topical agents such as nitrogen mustard or bexarotene may be more suitable treatment options, as the efficacy of these treatments is independent of skin color.⁸ However, nitrogen mustard commonly leads to postinflammatory hyperpigmentation, and topical bexarotene may lead to erythema or irritation; therefore, providers must counsel patients on these possible side effects. For refractory disease, adjunct systemic treatments such as oral bexarotene, subcutaneous interferon, methotrexate, or radiation therapy may be considered.⁸

In addition to aiding in the prompt diagnosis and treatment of MF/SS in SoC patients, our findings may be used to better assess the extent of disease and distinguish between active MF/SS lesions vs xerosis cutis or residual dyschromia from previously treated lesions. It is important to note that these morphologic features must be taken into account with a complete history and work-up. The differential diagnosis of MF/SS includes conditions such as atopic dermatitis, psoriasis, tinea corporis, and drug reactions, which may have similar morphology in SoC.¹⁹

Limitations of our study include the single-center design and the use of photographs instead of in-person examination; however, our cutaneous lymphoma clinic serves a diverse patient population, and our 3 reviewers rated the photographs independently. Discussion amongst the reviewers to address discrepancies was only required for 5 photographs, indicating the high inter-reviewer reliability. Additionally, the original purpose of FST was to assess for the propensity of the skin to burn when undergoing phototherapy, not to serve as a marker for skin color. We recommend trainees and clinicians be mindful about the purpose of FST and to use inclusive language (eg, using the terms skin irritation, skin tenderness, or skin becoming darker from the sun instead of tanning) when determining FST in darker-skinned individuals.²⁰ Future directions include examining if certain treatments are associated with prolonged dyschromia.

Conclusion

In our single-institution retrospective study, we found differences in the morphologic presentation of MF/SS in SoC patients vs non-SoC patients. While erythema is a common feature in non-SoC patients, clinical features of hyperpigmentation, lichenification, and silver hue should be carefully evaluated in the diagnosis of MF/SS in SoC patients. Knowledge of the heterogenous presentation of MF/SS in patients with SoC allows for expedited diagnosis and treatment, leading to better clinical outcomes. Valuable resources, including Taylor and Kelly’s Dermatology for Skin of Color, the Skin of Color Society, and VisualDx educate providers on how dermatologic conditions present in darker skin types. However, there is still work to be done to enhance diversity in educational resources in order to provide equitable care to patients of all skin types.

Mycosis fungoides (MF) and Sézary syndrome (SS) are non-Hodgkin T-cell lymphomas that make up the majority of cutaneous T-cell lymphomas. These conditions commonly affect Black patients, with an incidence rate of 12.6 cases of cutaneous T-cell lymphomas per million individuals vs 9.8 per million individuals in non–skin of color (SoC) patients.¹ However, educational resources tend to focus on the clinical manifestations of MF/SS in lighter skin types, describing MF as erythematous patches, plaques, or tumors presenting in non–sun-exposed areas of the skin and SS as generalized erythroderma.² Skin of color, comprised of Fitzpatrick skin types (FSTs) IV to VI,³ is poorly represented across dermatology textbooks,^4,5 medical student resources,⁶ and peer-reviewed publications,⁷ raising awareness for the need to address this disparity.

Skin of color patients with MF/SS display variable morphologies, including features such as hyperpigmentation and hypopigmentation,⁸ the latter being exceedingly rare in non-SoC patients.⁹ Familiarity with these differences among providers is essential to allow for equitable diagnosis and treatment across all skin types, especially in light of data predicting that by 2044 more than 50% of the US population will be people of color.¹⁰ Patients with SoC are of many ethnic and racial backgrounds, including Black, Hispanic, American Indian, Pacific Islander, and Asian.¹¹

Along with morphologic differences, there also are several racial disparities in the prognosis and survival of patients with MF/SS. Black patients diagnosed with MF present with greater body surface area affected, and Black women with MF have reduced survival rates compared to their White counterparts.¹² Given these racial disparities in survival and representation in educational resources, we aimed to quantify the frequency of various morphologic characteristics of MF/SS in patients with SoC vs non-SoC patients to facilitate better recognition of early MF/SS in SoC patients by medical providers.

Methods

We performed a retrospective chart review following approval from the institutional review board at Northwestern University (Chicago, Illinois). We identified all patients with FSTs IV to VI and biopsy-proven MF/SS who had been clinically photographed in our clinic from January 1998 to December 2019. Only photographs that were high quality enough to review morphologic features were included in our review. Fitzpatrick skin type was determined based on electronic medical record documentation. If photographs were available from multiple visits for the same patient, only those showing posttreatment nonactive lesions were included. Additionally, 36 patients with FSTs I to III (non-SoC) and biopsy-proven MF/SS were included in our review as a comparison with the SoC cohort. The primary outcomes for this study included the presence of scale, erythema, hyperpigmentation, hypopigmentation, violaceous color, lichenification, silver hue, dyschromia, alopecia, poikiloderma, atrophy, and ulceration in active lesions. Dyschromia was defined by the presence of both hypopigmentation and hyperpigmentation. Poikiloderma was defined by hypopigmentation and hyperpigmentation, telangiectasia, and atrophy. Secondary outcomes included evaluation of those same characteristics in posttreatment nonactive lesions. All photographs were independently assessed by 3 authors (M.L.E., C.J.W., J.M.M.), and discrepancies were resolved by further review of the photograph in question and discussion.

Statistical Analysis—Summary statistics were applied to describe demographic and clinical characteristics. The χ² test was used for categorical variables. Results achieving P<.05 were considered statistically significant.

Results

We reviewed photographs of 111 patients across all skin types (8, FST I; 12, FST II; 16, FST III; 17, FST IV; 44, FST V; 14, FST VI). The cohort was 47% female, and the mean age was 49.7 years (range, 15–86 years). The majority of the cohort had early-stage MF (stage IA or IB). There were more cases of SS in the SoC cohort than the non-SoC cohort (Table). Only 5 photographs had discrepancies and required discussion among the reviewers to achieve consensus.

Regarding morphologic characteristics in active lesions (Figure 1), scale was present in almost all patients (99% in SoC, 94% in non-SoC). Erythema was present in nearly all non-SoC patients (94%) but only in 69% of SoC patients (P=.003). Poikiloderma also was found to be present at higher frequencies in non-SoC patients compared with SoC patients (19% and 4%, respectively [P=.008]). However, hyperpigmentation (80% vs 39%), lichenification (43% vs 17%), and silver hue (25% vs 3%) were more common in SoC patients than non-SoC patients (P<.05). There were no significant differences in the remaining features, including hypopigmentation (39% vs 25%), dyschromia (24% vs 19%), violaceous color (44% vs 25%), atrophy (11% vs 22%), alopecia (23% vs 31%), and ulceration (16% vs 8%) between SoC and non-SoC patients (P>.05). Photographs of MF in patients with SoC can be seen in Figure 2.

Posttreatment (nonactive) photographs were available for 26 patients (6 non-SoC, 20 SoC). We found that across all FST groups, hyperpigmentation was more common than hypopigmentation in areas of previously active disease. Statistical analysis was not completed given that few non-SoC photographs were available for comparison.

Comment

This qualitative review demonstrates the heterogeneity of MF/SS in SoC patients and that these conditions do not present in this population with the classic erythematous patches and plaques found in non-SoC patients. We found that hyperpigmentation, lichenification, and silver hue were present at higher rates in patients with FSTs IV to VI compared to those with FSTs I to III, which had higher rates of erythema and poikiloderma. Familiarity with these morphologic features along with increased exposure to clinical photographs of MF/SS in SoC patients will aid in the visual recognition required for this diagnosis, since erythema is harder to identify in darker skin types. Recognizing the unique findings of MF in patients with SoC as well as in patients with lighter skin types will enable earlier diagnosis and treatment of MF/SS across all skin types. If MF is diagnosed and treated early, life expectancy is similar to that of patients without MF.¹³ However, the 5-year survival rate for advanced-stage MF/SS is 52% across all skin types, and studies have found that Black patients with advanced-stage disease have worse outcomes despite accounting for demographic factors and tumor stage.^14,15 Given the worse outcomes in SoC patients with advanced-stage MF/SS, earlier diagnosis could help address this disparity.^8,13,14 Similar morphologic features could be used in diagnosing other inflammatory conditions; studies have shown that the lack of recognition of erythema in Black children has led to delayed diagnosis of atopic dermatitis and subsequent inadequate treatment.^16,17

The morphologic presentation of MF/SS in SoC patients also can influence an optimal treatment plan for this population. Hypopigmented MF responds better to phototherapy than hyperpigmented MF, as phototherapy has been shown to have decreased efficacy with increasing FST.¹⁸ Therefore, for patients with FSTs IV to VI, topical agents such as nitrogen mustard or bexarotene may be more suitable treatment options, as the efficacy of these treatments is independent of skin color.⁸ However, nitrogen mustard commonly leads to postinflammatory hyperpigmentation, and topical bexarotene may lead to erythema or irritation; therefore, providers must counsel patients on these possible side effects. For refractory disease, adjunct systemic treatments such as oral bexarotene, subcutaneous interferon, methotrexate, or radiation therapy may be considered.⁸

In addition to aiding in the prompt diagnosis and treatment of MF/SS in SoC patients, our findings may be used to better assess the extent of disease and distinguish between active MF/SS lesions vs xerosis cutis or residual dyschromia from previously treated lesions. It is important to note that these morphologic features must be taken into account with a complete history and work-up. The differential diagnosis of MF/SS includes conditions such as atopic dermatitis, psoriasis, tinea corporis, and drug reactions, which may have similar morphology in SoC.¹⁹

Limitations of our study include the single-center design and the use of photographs instead of in-person examination; however, our cutaneous lymphoma clinic serves a diverse patient population, and our 3 reviewers rated the photographs independently. Discussion amongst the reviewers to address discrepancies was only required for 5 photographs, indicating the high inter-reviewer reliability. Additionally, the original purpose of FST was to assess for the propensity of the skin to burn when undergoing phototherapy, not to serve as a marker for skin color. We recommend trainees and clinicians be mindful about the purpose of FST and to use inclusive language (eg, using the terms skin irritation, skin tenderness, or skin becoming darker from the sun instead of tanning) when determining FST in darker-skinned individuals.²⁰ Future directions include examining if certain treatments are associated with prolonged dyschromia.

Conclusion

In our single-institution retrospective study, we found differences in the morphologic presentation of MF/SS in SoC patients vs non-SoC patients. While erythema is a common feature in non-SoC patients, clinical features of hyperpigmentation, lichenification, and silver hue should be carefully evaluated in the diagnosis of MF/SS in SoC patients. Knowledge of the heterogenous presentation of MF/SS in patients with SoC allows for expedited diagnosis and treatment, leading to better clinical outcomes. Valuable resources, including Taylor and Kelly’s Dermatology for Skin of Color, the Skin of Color Society, and VisualDx educate providers on how dermatologic conditions present in darker skin types. However, there is still work to be done to enhance diversity in educational resources in order to provide equitable care to patients of all skin types.

Mycosis fungoides (MF) and Sézary syndrome (SS) are non-Hodgkin T-cell lymphomas that make up the majority of cutaneous T-cell lymphomas. These conditions commonly affect Black patients, with an incidence rate of 12.6 cases of cutaneous T-cell lymphomas per million individuals vs 9.8 per million individuals in non–skin of color (SoC) patients.¹ However, educational resources tend to focus on the clinical manifestations of MF/SS in lighter skin types, describing MF as erythematous patches, plaques, or tumors presenting in non–sun-exposed areas of the skin and SS as generalized erythroderma.² Skin of color, comprised of Fitzpatrick skin types (FSTs) IV to VI,³ is poorly represented across dermatology textbooks,^4,5 medical student resources,⁶ and peer-reviewed publications,⁷ raising awareness for the need to address this disparity.

Skin of color patients with MF/SS display variable morphologies, including features such as hyperpigmentation and hypopigmentation,⁸ the latter being exceedingly rare in non-SoC patients.⁹ Familiarity with these differences among providers is essential to allow for equitable diagnosis and treatment across all skin types, especially in light of data predicting that by 2044 more than 50% of the US population will be people of color.¹⁰ Patients with SoC are of many ethnic and racial backgrounds, including Black, Hispanic, American Indian, Pacific Islander, and Asian.¹¹

Along with morphologic differences, there also are several racial disparities in the prognosis and survival of patients with MF/SS. Black patients diagnosed with MF present with greater body surface area affected, and Black women with MF have reduced survival rates compared to their White counterparts.¹² Given these racial disparities in survival and representation in educational resources, we aimed to quantify the frequency of various morphologic characteristics of MF/SS in patients with SoC vs non-SoC patients to facilitate better recognition of early MF/SS in SoC patients by medical providers.

Methods

We performed a retrospective chart review following approval from the institutional review board at Northwestern University (Chicago, Illinois). We identified all patients with FSTs IV to VI and biopsy-proven MF/SS who had been clinically photographed in our clinic from January 1998 to December 2019. Only photographs that were high quality enough to review morphologic features were included in our review. Fitzpatrick skin type was determined based on electronic medical record documentation. If photographs were available from multiple visits for the same patient, only those showing posttreatment nonactive lesions were included. Additionally, 36 patients with FSTs I to III (non-SoC) and biopsy-proven MF/SS were included in our review as a comparison with the SoC cohort. The primary outcomes for this study included the presence of scale, erythema, hyperpigmentation, hypopigmentation, violaceous color, lichenification, silver hue, dyschromia, alopecia, poikiloderma, atrophy, and ulceration in active lesions. Dyschromia was defined by the presence of both hypopigmentation and hyperpigmentation. Poikiloderma was defined by hypopigmentation and hyperpigmentation, telangiectasia, and atrophy. Secondary outcomes included evaluation of those same characteristics in posttreatment nonactive lesions. All photographs were independently assessed by 3 authors (M.L.E., C.J.W., J.M.M.), and discrepancies were resolved by further review of the photograph in question and discussion.

Statistical Analysis—Summary statistics were applied to describe demographic and clinical characteristics. The χ² test was used for categorical variables. Results achieving P<.05 were considered statistically significant.

Results

We reviewed photographs of 111 patients across all skin types (8, FST I; 12, FST II; 16, FST III; 17, FST IV; 44, FST V; 14, FST VI). The cohort was 47% female, and the mean age was 49.7 years (range, 15–86 years). The majority of the cohort had early-stage MF (stage IA or IB). There were more cases of SS in the SoC cohort than the non-SoC cohort (Table). Only 5 photographs had discrepancies and required discussion among the reviewers to achieve consensus.

Regarding morphologic characteristics in active lesions (Figure 1), scale was present in almost all patients (99% in SoC, 94% in non-SoC). Erythema was present in nearly all non-SoC patients (94%) but only in 69% of SoC patients (P=.003). Poikiloderma also was found to be present at higher frequencies in non-SoC patients compared with SoC patients (19% and 4%, respectively [P=.008]). However, hyperpigmentation (80% vs 39%), lichenification (43% vs 17%), and silver hue (25% vs 3%) were more common in SoC patients than non-SoC patients (P<.05). There were no significant differences in the remaining features, including hypopigmentation (39% vs 25%), dyschromia (24% vs 19%), violaceous color (44% vs 25%), atrophy (11% vs 22%), alopecia (23% vs 31%), and ulceration (16% vs 8%) between SoC and non-SoC patients (P>.05). Photographs of MF in patients with SoC can be seen in Figure 2.

Posttreatment (nonactive) photographs were available for 26 patients (6 non-SoC, 20 SoC). We found that across all FST groups, hyperpigmentation was more common than hypopigmentation in areas of previously active disease. Statistical analysis was not completed given that few non-SoC photographs were available for comparison.

Comment

This qualitative review demonstrates the heterogeneity of MF/SS in SoC patients and that these conditions do not present in this population with the classic erythematous patches and plaques found in non-SoC patients. We found that hyperpigmentation, lichenification, and silver hue were present at higher rates in patients with FSTs IV to VI compared to those with FSTs I to III, which had higher rates of erythema and poikiloderma. Familiarity with these morphologic features along with increased exposure to clinical photographs of MF/SS in SoC patients will aid in the visual recognition required for this diagnosis, since erythema is harder to identify in darker skin types. Recognizing the unique findings of MF in patients with SoC as well as in patients with lighter skin types will enable earlier diagnosis and treatment of MF/SS across all skin types. If MF is diagnosed and treated early, life expectancy is similar to that of patients without MF.¹³ However, the 5-year survival rate for advanced-stage MF/SS is 52% across all skin types, and studies have found that Black patients with advanced-stage disease have worse outcomes despite accounting for demographic factors and tumor stage.^14,15 Given the worse outcomes in SoC patients with advanced-stage MF/SS, earlier diagnosis could help address this disparity.^8,13,14 Similar morphologic features could be used in diagnosing other inflammatory conditions; studies have shown that the lack of recognition of erythema in Black children has led to delayed diagnosis of atopic dermatitis and subsequent inadequate treatment.^16,17

The morphologic presentation of MF/SS in SoC patients also can influence an optimal treatment plan for this population. Hypopigmented MF responds better to phototherapy than hyperpigmented MF, as phototherapy has been shown to have decreased efficacy with increasing FST.¹⁸ Therefore, for patients with FSTs IV to VI, topical agents such as nitrogen mustard or bexarotene may be more suitable treatment options, as the efficacy of these treatments is independent of skin color.⁸ However, nitrogen mustard commonly leads to postinflammatory hyperpigmentation, and topical bexarotene may lead to erythema or irritation; therefore, providers must counsel patients on these possible side effects. For refractory disease, adjunct systemic treatments such as oral bexarotene, subcutaneous interferon, methotrexate, or radiation therapy may be considered.⁸

In addition to aiding in the prompt diagnosis and treatment of MF/SS in SoC patients, our findings may be used to better assess the extent of disease and distinguish between active MF/SS lesions vs xerosis cutis or residual dyschromia from previously treated lesions. It is important to note that these morphologic features must be taken into account with a complete history and work-up. The differential diagnosis of MF/SS includes conditions such as atopic dermatitis, psoriasis, tinea corporis, and drug reactions, which may have similar morphology in SoC.¹⁹

Limitations of our study include the single-center design and the use of photographs instead of in-person examination; however, our cutaneous lymphoma clinic serves a diverse patient population, and our 3 reviewers rated the photographs independently. Discussion amongst the reviewers to address discrepancies was only required for 5 photographs, indicating the high inter-reviewer reliability. Additionally, the original purpose of FST was to assess for the propensity of the skin to burn when undergoing phototherapy, not to serve as a marker for skin color. We recommend trainees and clinicians be mindful about the purpose of FST and to use inclusive language (eg, using the terms skin irritation, skin tenderness, or skin becoming darker from the sun instead of tanning) when determining FST in darker-skinned individuals.²⁰ Future directions include examining if certain treatments are associated with prolonged dyschromia.

Conclusion

In our single-institution retrospective study, we found differences in the morphologic presentation of MF/SS in SoC patients vs non-SoC patients. While erythema is a common feature in non-SoC patients, clinical features of hyperpigmentation, lichenification, and silver hue should be carefully evaluated in the diagnosis of MF/SS in SoC patients. Knowledge of the heterogenous presentation of MF/SS in patients with SoC allows for expedited diagnosis and treatment, leading to better clinical outcomes. Valuable resources, including Taylor and Kelly’s Dermatology for Skin of Color, the Skin of Color Society, and VisualDx educate providers on how dermatologic conditions present in darker skin types. However, there is still work to be done to enhance diversity in educational resources in order to provide equitable care to patients of all skin types.

The Diagnosis: Cystic Neutrophilic Granulomatous Mastitis

The histopathologic findings in our patient were characteristic of cystic neutrophilic granulomatous mastitis (CNGM), a rare granulomatous mastitis associated with Corynebacterium and suppurative lipogranulomas. Although not seen in our patient, the lipid vacuoles may contain gram-positive bacilli.¹ The surrounding mixed inflammatory infiltrate contains Langerhans giant cells, lymphocytes, and neutrophils. Cystic neutrophilic granulomatous mastitis is seen in parous women of reproductive age. Physical examination demonstrates a palpable painful mass on the breast. Wound cultures frequently are negative, likely due to difficulty culturing Corynebacterium and prophylactic antibiotic treatment. Given the association with Corynebacterium species, early diagnosis of CNGM is essential in offering patients the most appropriate treatment. Prolonged antibiotic therapy specifically directed to corynebacteria is required, sometimes even beyond resolution of clinical symptoms. The diagnosis of CNGM often is missed or delayed due to its rarity and many potential mimickers. Clinically, CNGM may be virtually impossible to discern from invasive carcinoma.¹

Our patient was treated with vancomycin and cefepime with incision and drainage as an inpatient. Upon discharge, she was started on prednisone 1 mg/kg daily tapered by 10 mg every 5 days over 1 month and doxycycline 100 mg twice daily. She was then transitioned to topical hydrocortisone and bacitracin; she reported decreased swelling and pain. No new lesions formed after the initiation of therapy; however, most lesions remained open. Cystic neutrophilic granulomatous mastitis remains a challenging entity to treat, with a variable response rate reported in the literature for antibiotics such as doxycycline and systemic and topical steroids as well as immunosuppressants including methotrexate.^2,3

Cystic neutrophilic granulomatous mastitis can be distinguished from hidradenitis suppurativa clinically because ulcerating lesions can involve the superior portions of the breast in CNGM, whereas hidradenitis suppurativa typically is restricted to the lower intertriginous parts of the breast. Other mimics of CNGM can be distinguished with biopsy. Histology of pyoderma gangrenosum lacks prominent granuloma formation. Although sarcoidosis and mycobacterial infection show prominent granulomas, neither show the characteristic lipogranulomas seen in CNGM. Additionally, the granulomas of sarcoidosis are much larger and deeper than CNGM. Mycobacterial granulomas also typically reveal bacilli with acid-fast bacilli staining or via wound culture.

The Diagnosis: Cystic Neutrophilic Granulomatous Mastitis

The histopathologic findings in our patient were characteristic of cystic neutrophilic granulomatous mastitis (CNGM), a rare granulomatous mastitis associated with Corynebacterium and suppurative lipogranulomas. Although not seen in our patient, the lipid vacuoles may contain gram-positive bacilli.¹ The surrounding mixed inflammatory infiltrate contains Langerhans giant cells, lymphocytes, and neutrophils. Cystic neutrophilic granulomatous mastitis is seen in parous women of reproductive age. Physical examination demonstrates a palpable painful mass on the breast. Wound cultures frequently are negative, likely due to difficulty culturing Corynebacterium and prophylactic antibiotic treatment. Given the association with Corynebacterium species, early diagnosis of CNGM is essential in offering patients the most appropriate treatment. Prolonged antibiotic therapy specifically directed to corynebacteria is required, sometimes even beyond resolution of clinical symptoms. The diagnosis of CNGM often is missed or delayed due to its rarity and many potential mimickers. Clinically, CNGM may be virtually impossible to discern from invasive carcinoma.¹

Our patient was treated with vancomycin and cefepime with incision and drainage as an inpatient. Upon discharge, she was started on prednisone 1 mg/kg daily tapered by 10 mg every 5 days over 1 month and doxycycline 100 mg twice daily. She was then transitioned to topical hydrocortisone and bacitracin; she reported decreased swelling and pain. No new lesions formed after the initiation of therapy; however, most lesions remained open. Cystic neutrophilic granulomatous mastitis remains a challenging entity to treat, with a variable response rate reported in the literature for antibiotics such as doxycycline and systemic and topical steroids as well as immunosuppressants including methotrexate.^2,3

Cystic neutrophilic granulomatous mastitis can be distinguished from hidradenitis suppurativa clinically because ulcerating lesions can involve the superior portions of the breast in CNGM, whereas hidradenitis suppurativa typically is restricted to the lower intertriginous parts of the breast. Other mimics of CNGM can be distinguished with biopsy. Histology of pyoderma gangrenosum lacks prominent granuloma formation. Although sarcoidosis and mycobacterial infection show prominent granulomas, neither show the characteristic lipogranulomas seen in CNGM. Additionally, the granulomas of sarcoidosis are much larger and deeper than CNGM. Mycobacterial granulomas also typically reveal bacilli with acid-fast bacilli staining or via wound culture.

The Diagnosis: Cystic Neutrophilic Granulomatous Mastitis

The histopathologic findings in our patient were characteristic of cystic neutrophilic granulomatous mastitis (CNGM), a rare granulomatous mastitis associated with Corynebacterium and suppurative lipogranulomas. Although not seen in our patient, the lipid vacuoles may contain gram-positive bacilli.¹ The surrounding mixed inflammatory infiltrate contains Langerhans giant cells, lymphocytes, and neutrophils. Cystic neutrophilic granulomatous mastitis is seen in parous women of reproductive age. Physical examination demonstrates a palpable painful mass on the breast. Wound cultures frequently are negative, likely due to difficulty culturing Corynebacterium and prophylactic antibiotic treatment. Given the association with Corynebacterium species, early diagnosis of CNGM is essential in offering patients the most appropriate treatment. Prolonged antibiotic therapy specifically directed to corynebacteria is required, sometimes even beyond resolution of clinical symptoms. The diagnosis of CNGM often is missed or delayed due to its rarity and many potential mimickers. Clinically, CNGM may be virtually impossible to discern from invasive carcinoma.¹

Our patient was treated with vancomycin and cefepime with incision and drainage as an inpatient. Upon discharge, she was started on prednisone 1 mg/kg daily tapered by 10 mg every 5 days over 1 month and doxycycline 100 mg twice daily. She was then transitioned to topical hydrocortisone and bacitracin; she reported decreased swelling and pain. No new lesions formed after the initiation of therapy; however, most lesions remained open. Cystic neutrophilic granulomatous mastitis remains a challenging entity to treat, with a variable response rate reported in the literature for antibiotics such as doxycycline and systemic and topical steroids as well as immunosuppressants including methotrexate.^2,3

Cystic neutrophilic granulomatous mastitis can be distinguished from hidradenitis suppurativa clinically because ulcerating lesions can involve the superior portions of the breast in CNGM, whereas hidradenitis suppurativa typically is restricted to the lower intertriginous parts of the breast. Other mimics of CNGM can be distinguished with biopsy. Histology of pyoderma gangrenosum lacks prominent granuloma formation. Although sarcoidosis and mycobacterial infection show prominent granulomas, neither show the characteristic lipogranulomas seen in CNGM. Additionally, the granulomas of sarcoidosis are much larger and deeper than CNGM. Mycobacterial granulomas also typically reveal bacilli with acid-fast bacilli staining or via wound culture.

The Diagnosis: Blastic Plasmacytoid Dendritic Cell Neoplasm

A diagnosis of blastic plasmacytoid dendritic cell neoplasm (BPDCN) was rendered. Subsequent needle core biopsy of a left axillary lymph node as well as bone marrow aspiration and biopsy revealed a similar diffuse blastoid infiltrate with an identical immunophenotype to that in the skin biopsy from the pretibial mass and peripheral blood.

Previously known as blastic natural killer cell leukemia/lymphoma or agranular CD4+/CD56+ hematodermic neoplasm/tumor, BPDCN is a rare, clinically aggressive hematologic malignancy derived from the precursors of plasmacytoid dendritic cells. It often is diagnostically challenging, particularly when presenting at noncutaneous sites and in unusual (young) patient populations.¹ It was included with other myeloid neoplasms in the 2008 World Health Organization classification; however, in the 2017 classification it was categorized as a separate entity. Blastic plasmacytoid dendritic cell neoplasm typically presents in the skin of elderly patients (age range at diagnosis, 61–67 years) with or without bone marrow involvement and systemic dissemination.^1,2 The skin is the most common clinical site of disease in typical cases of BPDCN and often precedes bone marrow involvement. Thus, skin biopsy often is the key to making the diagnosis. Diagnosis of BPDCN may be delayed because of diagnostic pitfalls. Patients usually present with asymptomatic solitary or multiple lesions.^3-5 Blastic plasmacytoid dendritic cell neoplasm can present as an isolated purplish nodule or bruiselike papule or more commonly as disseminated purplish nodules, papules, and macules. Isolated nodules are found on the head and lower limbs and can be more than 10 cm in diameter. Peripheral blood and bone marrow may be minimally involved at presentation but invariably become involved with the progression of disease. Cytopenia can occur at diagnosis and in a minority of severe cases indicates bone marrow failure.^2-6

Skin involvement of BPDCN is thought to be secondary to the expression of skin migration molecules, such as cutaneous lymphocyte-associated antigen, one of the E-selectin ligands, which binds to E-selectin on high endothelial venules. In addition, the local dermal microenvironment of chemokines binding CXCR3, CXCR4, CCR6, or CCR7 present on neoplastic cells possibly leads to skin involvement. The full mechanism underlying the cutaneous tropism is still to be elucidated.^4-7 Infiltration of the oral mucosa is seen in some patients, but it may be underreported. Mucosal disease typically appears similarly to cutaneous disease.

The cutaneous differential diagnosis for BPDCN depends on the clinical presentation, extent of disease spread, and thickness of infiltration. It includes common nonneoplastic diseases such as traumatic ecchymoses; purpuric disorders; extramedullary hematopoiesis; and soft-tissue neoplasms such as angiosarcoma, Kaposi sarcoma, neuroblastoma, and vascular metastases, as well as skin involvement by other hematologic neoplasms. An adequate incisional biopsy rather than a punch or shave biopsy is recommended for diagnosis. Dermatologists should alert the pathologist that BPDCN is in the clinical differential diagnosis when possible so that judicious use of appropriate immunophenotypic markers such as CD123, CD4, CD56, and T-cell leukemia/lymphoma protein 1 will avoid misdiagnosis of this aggressive condition, in addition to excluding acute myeloid leukemia, which also may express 3 of the above markers. However, most cases of acute myeloid leukemia lack terminal deoxynucleotidyl transferase (TdT) and express monocytic and other myeloid markers. Terminal deoxynucleotidyl transferase is positive in approximately one-third of cases of BPDCN, with expression in 10% to 80% of cells.¹

It is important to include BPDCN in the differential diagnosis of immunophenotypically aberrant hematologic tumors. Diffuse large B-cell lymphoma, leg type, accounts for 4% of all primary cutaneous B-cell lymphomas.¹ Compared with BPDCN, diffuse large B-cell lymphoma usually occurs in an older age group and is of B-cell lineage. Morphologically, these neoplasms are composed of a monotonous, diffuse, nonepidermotropic infiltrate of confluent sheets of centroblasts and immunoblasts (Figure 1). They may share immunohistochemical markers of CD79a, multiple myeloma 1, Bcl-2, and Bcl-6; however, they lack plasmacytoid dendritic cell (PDC)– associated antigens such as CD4, CD56, CD123, and T-cell leukemia/lymphoma protein 1.¹

Adult T-cell leukemia/lymphoma is a neoplasm histologically composed of highly pleomorphic medium- to large-sized T cells with an irregular multilobated nuclear contour, so-called flower cells, in the peripheral blood. The nuclear chromatin is coarse and clumped with prominent nucleoli. Blastlike cells with dispersed chromatin are present in variable proportions. Most patients present with widespread lymph node and peripheral blood involvement. Skin is involved in more than half of patients with an epidermal as well as dermal pattern of infiltration (mainly perivascular)(Figure 2). Adult T-cell leukemia/lymphoma is endemic in several regions of the world, and the distribution is closely linked to the prevalence of human T-cell lymphotropic virus type 1 in the population. This neoplasm is of T-cell lineage and may share CD4 but not PDC-associated antigens with BPDCN.¹

Cutaneous involvement by T-cell lymphoblastic leukemia/lymphoma (T-LBL) is a rare occurrence with a frequency of approximately 4.3%.8 T-cell lymphoblastic leukemia/lymphoma usually presents as multiple skin lesions throughout the body. Almost all cutaneous T-LBL cases are seen in association with bone marrow and/or mediastinal, lymph node, or extranodal involvement. Cutaneous T-LBLs present as a diffuse monomorphous infiltrate located in the entire dermis and subcutis without epidermotropism, composed of medium to large blasts with finely dispersed chromatin and relatively prominent nucleoli (Figure 3). Immunophenotyping studies show an immature T-cell immunophenotype, with expression of TdT (usually uniform), CD7, and cytoplasmic CD3 and an absence of PDC-associated antigens.⁸

Primary cutaneous γδ T-cell lymphoma (PCGDTL) is a neoplasm primarily involving the skin. Often rapidly fatal, PCGDTL has a broad clinical spectrum that may include indolent variants—subcutaneous, epidermotropic, and dermal. Patients typically present with nodular lesions that progress to ulceration and necrosis. Early lesions can be confused with erythema nodosum, mycosis fungoides, or infection. Histologically, they show variable epidermotropism as well as dermal and subcutaneous involvement by medium to large cells with coarse clumped chromatin (Figure 4). Large blastic cells with vesicular nuclei and prominent nucleoli are infrequent. In contrast to BPCDN, the neoplastic lymphocytes in dermal and subcutaneous PCGDTL typically are positive for T-cell intracellular antigen-1 and granzyme B with loss of CD4.⁹

At the time of presentation, 27% to 87% of BPDCN patients will have bone marrow involvement, 22% to 28% will have blood involvement, and 6% to 41% will have lymph node involvement.^{1-4,6,7,10,11} The clinical course is aggressive, with a median survival of 10.0 to 19.8 months, irrespective of the initial pattern of disease.¹ Most cases have shown initial response to multiagent chemotherapy, but relapses with subsequent resistance to drugs regularly have been observed. Age has an adverse impact of prognosis. Low TdT expression has been associated with shorter survival.¹ Approximately 10% to 20% of cases of BPDCN are associated with or develop into chronic myelogenous leukemia, myelodysplastic syndrome, or acute myeloid leukemia.^1,4 Pediatric patients have a greater 5-year overall survival rate than older patients, and overall survival worsens with increasing age. The extent of cutaneous involvement and presence of systemic involvement at initial presentation do not seem to be strong predictors of survival.^{1,2,5-7,10-12} In a retrospective analysis of 90 patients, Julia et al¹² found that the type of skin disease did not predict survival. Specifically, the presence of nodular lesions and disseminated skin involvement were not adverse prognostic factors compared with macular lesions limited to 1 or 2 body areas.¹²

The Diagnosis: Blastic Plasmacytoid Dendritic Cell Neoplasm

A diagnosis of blastic plasmacytoid dendritic cell neoplasm (BPDCN) was rendered. Subsequent needle core biopsy of a left axillary lymph node as well as bone marrow aspiration and biopsy revealed a similar diffuse blastoid infiltrate with an identical immunophenotype to that in the skin biopsy from the pretibial mass and peripheral blood.

Previously known as blastic natural killer cell leukemia/lymphoma or agranular CD4+/CD56+ hematodermic neoplasm/tumor, BPDCN is a rare, clinically aggressive hematologic malignancy derived from the precursors of plasmacytoid dendritic cells. It often is diagnostically challenging, particularly when presenting at noncutaneous sites and in unusual (young) patient populations.¹ It was included with other myeloid neoplasms in the 2008 World Health Organization classification; however, in the 2017 classification it was categorized as a separate entity. Blastic plasmacytoid dendritic cell neoplasm typically presents in the skin of elderly patients (age range at diagnosis, 61–67 years) with or without bone marrow involvement and systemic dissemination.^1,2 The skin is the most common clinical site of disease in typical cases of BPDCN and often precedes bone marrow involvement. Thus, skin biopsy often is the key to making the diagnosis. Diagnosis of BPDCN may be delayed because of diagnostic pitfalls. Patients usually present with asymptomatic solitary or multiple lesions.^3-5 Blastic plasmacytoid dendritic cell neoplasm can present as an isolated purplish nodule or bruiselike papule or more commonly as disseminated purplish nodules, papules, and macules. Isolated nodules are found on the head and lower limbs and can be more than 10 cm in diameter. Peripheral blood and bone marrow may be minimally involved at presentation but invariably become involved with the progression of disease. Cytopenia can occur at diagnosis and in a minority of severe cases indicates bone marrow failure.^2-6

Skin involvement of BPDCN is thought to be secondary to the expression of skin migration molecules, such as cutaneous lymphocyte-associated antigen, one of the E-selectin ligands, which binds to E-selectin on high endothelial venules. In addition, the local dermal microenvironment of chemokines binding CXCR3, CXCR4, CCR6, or CCR7 present on neoplastic cells possibly leads to skin involvement. The full mechanism underlying the cutaneous tropism is still to be elucidated.^4-7 Infiltration of the oral mucosa is seen in some patients, but it may be underreported. Mucosal disease typically appears similarly to cutaneous disease.

The cutaneous differential diagnosis for BPDCN depends on the clinical presentation, extent of disease spread, and thickness of infiltration. It includes common nonneoplastic diseases such as traumatic ecchymoses; purpuric disorders; extramedullary hematopoiesis; and soft-tissue neoplasms such as angiosarcoma, Kaposi sarcoma, neuroblastoma, and vascular metastases, as well as skin involvement by other hematologic neoplasms. An adequate incisional biopsy rather than a punch or shave biopsy is recommended for diagnosis. Dermatologists should alert the pathologist that BPDCN is in the clinical differential diagnosis when possible so that judicious use of appropriate immunophenotypic markers such as CD123, CD4, CD56, and T-cell leukemia/lymphoma protein 1 will avoid misdiagnosis of this aggressive condition, in addition to excluding acute myeloid leukemia, which also may express 3 of the above markers. However, most cases of acute myeloid leukemia lack terminal deoxynucleotidyl transferase (TdT) and express monocytic and other myeloid markers. Terminal deoxynucleotidyl transferase is positive in approximately one-third of cases of BPDCN, with expression in 10% to 80% of cells.¹

It is important to include BPDCN in the differential diagnosis of immunophenotypically aberrant hematologic tumors. Diffuse large B-cell lymphoma, leg type, accounts for 4% of all primary cutaneous B-cell lymphomas.¹ Compared with BPDCN, diffuse large B-cell lymphoma usually occurs in an older age group and is of B-cell lineage. Morphologically, these neoplasms are composed of a monotonous, diffuse, nonepidermotropic infiltrate of confluent sheets of centroblasts and immunoblasts (Figure 1). They may share immunohistochemical markers of CD79a, multiple myeloma 1, Bcl-2, and Bcl-6; however, they lack plasmacytoid dendritic cell (PDC)– associated antigens such as CD4, CD56, CD123, and T-cell leukemia/lymphoma protein 1.¹

Adult T-cell leukemia/lymphoma is a neoplasm histologically composed of highly pleomorphic medium- to large-sized T cells with an irregular multilobated nuclear contour, so-called flower cells, in the peripheral blood. The nuclear chromatin is coarse and clumped with prominent nucleoli. Blastlike cells with dispersed chromatin are present in variable proportions. Most patients present with widespread lymph node and peripheral blood involvement. Skin is involved in more than half of patients with an epidermal as well as dermal pattern of infiltration (mainly perivascular)(Figure 2). Adult T-cell leukemia/lymphoma is endemic in several regions of the world, and the distribution is closely linked to the prevalence of human T-cell lymphotropic virus type 1 in the population. This neoplasm is of T-cell lineage and may share CD4 but not PDC-associated antigens with BPDCN.¹

Cutaneous involvement by T-cell lymphoblastic leukemia/lymphoma (T-LBL) is a rare occurrence with a frequency of approximately 4.3%.8 T-cell lymphoblastic leukemia/lymphoma usually presents as multiple skin lesions throughout the body. Almost all cutaneous T-LBL cases are seen in association with bone marrow and/or mediastinal, lymph node, or extranodal involvement. Cutaneous T-LBLs present as a diffuse monomorphous infiltrate located in the entire dermis and subcutis without epidermotropism, composed of medium to large blasts with finely dispersed chromatin and relatively prominent nucleoli (Figure 3). Immunophenotyping studies show an immature T-cell immunophenotype, with expression of TdT (usually uniform), CD7, and cytoplasmic CD3 and an absence of PDC-associated antigens.⁸

Primary cutaneous γδ T-cell lymphoma (PCGDTL) is a neoplasm primarily involving the skin. Often rapidly fatal, PCGDTL has a broad clinical spectrum that may include indolent variants—subcutaneous, epidermotropic, and dermal. Patients typically present with nodular lesions that progress to ulceration and necrosis. Early lesions can be confused with erythema nodosum, mycosis fungoides, or infection. Histologically, they show variable epidermotropism as well as dermal and subcutaneous involvement by medium to large cells with coarse clumped chromatin (Figure 4). Large blastic cells with vesicular nuclei and prominent nucleoli are infrequent. In contrast to BPCDN, the neoplastic lymphocytes in dermal and subcutaneous PCGDTL typically are positive for T-cell intracellular antigen-1 and granzyme B with loss of CD4.⁹

At the time of presentation, 27% to 87% of BPDCN patients will have bone marrow involvement, 22% to 28% will have blood involvement, and 6% to 41% will have lymph node involvement.^{1-4,6,7,10,11} The clinical course is aggressive, with a median survival of 10.0 to 19.8 months, irrespective of the initial pattern of disease.¹ Most cases have shown initial response to multiagent chemotherapy, but relapses with subsequent resistance to drugs regularly have been observed. Age has an adverse impact of prognosis. Low TdT expression has been associated with shorter survival.¹ Approximately 10% to 20% of cases of BPDCN are associated with or develop into chronic myelogenous leukemia, myelodysplastic syndrome, or acute myeloid leukemia.^1,4 Pediatric patients have a greater 5-year overall survival rate than older patients, and overall survival worsens with increasing age. The extent of cutaneous involvement and presence of systemic involvement at initial presentation do not seem to be strong predictors of survival.^{1,2,5-7,10-12} In a retrospective analysis of 90 patients, Julia et al¹² found that the type of skin disease did not predict survival. Specifically, the presence of nodular lesions and disseminated skin involvement were not adverse prognostic factors compared with macular lesions limited to 1 or 2 body areas.¹²

The Diagnosis: Blastic Plasmacytoid Dendritic Cell Neoplasm

A diagnosis of blastic plasmacytoid dendritic cell neoplasm (BPDCN) was rendered. Subsequent needle core biopsy of a left axillary lymph node as well as bone marrow aspiration and biopsy revealed a similar diffuse blastoid infiltrate with an identical immunophenotype to that in the skin biopsy from the pretibial mass and peripheral blood.

Previously known as blastic natural killer cell leukemia/lymphoma or agranular CD4+/CD56+ hematodermic neoplasm/tumor, BPDCN is a rare, clinically aggressive hematologic malignancy derived from the precursors of plasmacytoid dendritic cells. It often is diagnostically challenging, particularly when presenting at noncutaneous sites and in unusual (young) patient populations.¹ It was included with other myeloid neoplasms in the 2008 World Health Organization classification; however, in the 2017 classification it was categorized as a separate entity. Blastic plasmacytoid dendritic cell neoplasm typically presents in the skin of elderly patients (age range at diagnosis, 61–67 years) with or without bone marrow involvement and systemic dissemination.^1,2 The skin is the most common clinical site of disease in typical cases of BPDCN and often precedes bone marrow involvement. Thus, skin biopsy often is the key to making the diagnosis. Diagnosis of BPDCN may be delayed because of diagnostic pitfalls. Patients usually present with asymptomatic solitary or multiple lesions.^3-5 Blastic plasmacytoid dendritic cell neoplasm can present as an isolated purplish nodule or bruiselike papule or more commonly as disseminated purplish nodules, papules, and macules. Isolated nodules are found on the head and lower limbs and can be more than 10 cm in diameter. Peripheral blood and bone marrow may be minimally involved at presentation but invariably become involved with the progression of disease. Cytopenia can occur at diagnosis and in a minority of severe cases indicates bone marrow failure.^2-6

Skin involvement of BPDCN is thought to be secondary to the expression of skin migration molecules, such as cutaneous lymphocyte-associated antigen, one of the E-selectin ligands, which binds to E-selectin on high endothelial venules. In addition, the local dermal microenvironment of chemokines binding CXCR3, CXCR4, CCR6, or CCR7 present on neoplastic cells possibly leads to skin involvement. The full mechanism underlying the cutaneous tropism is still to be elucidated.^4-7 Infiltration of the oral mucosa is seen in some patients, but it may be underreported. Mucosal disease typically appears similarly to cutaneous disease.

The cutaneous differential diagnosis for BPDCN depends on the clinical presentation, extent of disease spread, and thickness of infiltration. It includes common nonneoplastic diseases such as traumatic ecchymoses; purpuric disorders; extramedullary hematopoiesis; and soft-tissue neoplasms such as angiosarcoma, Kaposi sarcoma, neuroblastoma, and vascular metastases, as well as skin involvement by other hematologic neoplasms. An adequate incisional biopsy rather than a punch or shave biopsy is recommended for diagnosis. Dermatologists should alert the pathologist that BPDCN is in the clinical differential diagnosis when possible so that judicious use of appropriate immunophenotypic markers such as CD123, CD4, CD56, and T-cell leukemia/lymphoma protein 1 will avoid misdiagnosis of this aggressive condition, in addition to excluding acute myeloid leukemia, which also may express 3 of the above markers. However, most cases of acute myeloid leukemia lack terminal deoxynucleotidyl transferase (TdT) and express monocytic and other myeloid markers. Terminal deoxynucleotidyl transferase is positive in approximately one-third of cases of BPDCN, with expression in 10% to 80% of cells.¹

It is important to include BPDCN in the differential diagnosis of immunophenotypically aberrant hematologic tumors. Diffuse large B-cell lymphoma, leg type, accounts for 4% of all primary cutaneous B-cell lymphomas.¹ Compared with BPDCN, diffuse large B-cell lymphoma usually occurs in an older age group and is of B-cell lineage. Morphologically, these neoplasms are composed of a monotonous, diffuse, nonepidermotropic infiltrate of confluent sheets of centroblasts and immunoblasts (Figure 1). They may share immunohistochemical markers of CD79a, multiple myeloma 1, Bcl-2, and Bcl-6; however, they lack plasmacytoid dendritic cell (PDC)– associated antigens such as CD4, CD56, CD123, and T-cell leukemia/lymphoma protein 1.¹

Adult T-cell leukemia/lymphoma is a neoplasm histologically composed of highly pleomorphic medium- to large-sized T cells with an irregular multilobated nuclear contour, so-called flower cells, in the peripheral blood. The nuclear chromatin is coarse and clumped with prominent nucleoli. Blastlike cells with dispersed chromatin are present in variable proportions. Most patients present with widespread lymph node and peripheral blood involvement. Skin is involved in more than half of patients with an epidermal as well as dermal pattern of infiltration (mainly perivascular)(Figure 2). Adult T-cell leukemia/lymphoma is endemic in several regions of the world, and the distribution is closely linked to the prevalence of human T-cell lymphotropic virus type 1 in the population. This neoplasm is of T-cell lineage and may share CD4 but not PDC-associated antigens with BPDCN.¹

Cutaneous involvement by T-cell lymphoblastic leukemia/lymphoma (T-LBL) is a rare occurrence with a frequency of approximately 4.3%.8 T-cell lymphoblastic leukemia/lymphoma usually presents as multiple skin lesions throughout the body. Almost all cutaneous T-LBL cases are seen in association with bone marrow and/or mediastinal, lymph node, or extranodal involvement. Cutaneous T-LBLs present as a diffuse monomorphous infiltrate located in the entire dermis and subcutis without epidermotropism, composed of medium to large blasts with finely dispersed chromatin and relatively prominent nucleoli (Figure 3). Immunophenotyping studies show an immature T-cell immunophenotype, with expression of TdT (usually uniform), CD7, and cytoplasmic CD3 and an absence of PDC-associated antigens.⁸

Primary cutaneous γδ T-cell lymphoma (PCGDTL) is a neoplasm primarily involving the skin. Often rapidly fatal, PCGDTL has a broad clinical spectrum that may include indolent variants—subcutaneous, epidermotropic, and dermal. Patients typically present with nodular lesions that progress to ulceration and necrosis. Early lesions can be confused with erythema nodosum, mycosis fungoides, or infection. Histologically, they show variable epidermotropism as well as dermal and subcutaneous involvement by medium to large cells with coarse clumped chromatin (Figure 4). Large blastic cells with vesicular nuclei and prominent nucleoli are infrequent. In contrast to BPCDN, the neoplastic lymphocytes in dermal and subcutaneous PCGDTL typically are positive for T-cell intracellular antigen-1 and granzyme B with loss of CD4.⁹

At the time of presentation, 27% to 87% of BPDCN patients will have bone marrow involvement, 22% to 28% will have blood involvement, and 6% to 41% will have lymph node involvement.^{1-4,6,7,10,11} The clinical course is aggressive, with a median survival of 10.0 to 19.8 months, irrespective of the initial pattern of disease.¹ Most cases have shown initial response to multiagent chemotherapy, but relapses with subsequent resistance to drugs regularly have been observed. Age has an adverse impact of prognosis. Low TdT expression has been associated with shorter survival.¹ Approximately 10% to 20% of cases of BPDCN are associated with or develop into chronic myelogenous leukemia, myelodysplastic syndrome, or acute myeloid leukemia.^1,4 Pediatric patients have a greater 5-year overall survival rate than older patients, and overall survival worsens with increasing age. The extent of cutaneous involvement and presence of systemic involvement at initial presentation do not seem to be strong predictors of survival.^{1,2,5-7,10-12} In a retrospective analysis of 90 patients, Julia et al¹² found that the type of skin disease did not predict survival. Specifically, the presence of nodular lesions and disseminated skin involvement were not adverse prognostic factors compared with macular lesions limited to 1 or 2 body areas.¹²

Atopic dermatitis (AD) is an inflammatory skin disease characterized by skin barrier disruption, skin inflammation, and pruritus.¹ It is a common and often chronic skin condition associated with the development of food allergies, asthma, and allergic rhinitis, known as the atopic march.² Atopic dermatitis is estimated to affect 10% to 25% of children, most with onset before 5 years of age, and up to 7% of adults worldwide.³ Most patients improve with time, but multiple disease trajectories are possible. Several studies have demonstrated that fewer than 4% of children develop the classic atopic march—AD followed by food allergies, asthma, and finally allergic rhinitis—with recent evidence pointing to a more complex heterogeneous progression of disease and allergic comorbidities often occurring together.^4,5 The prevalence of AD has been increasing globally over the last 30 years,⁶ with a marked increase in developed countries.^6,7 It is well accepted that AD is based on an interplay between genetic predisposition and environmental factors,⁸ but many suspect that the rapid rise in prevalence cannot be attributed to genetic factors alone.⁹ The precipitant triggers for AD remain an area of intense investigation, with ongoing debate between the “inside out” and “outside in” hypotheses; these revolve around whether abnormalities in the immune system trigger barrier dysfunction or barrier dysfunction triggers immune programming to atopy.⁸ Ongoing research related to genetic predisposition of AD has identified candidate genes implicated in both impaired skin barrier function and altered immune system pathways, further supporting that both theories may contribute to disease pathogenesis. 

The increasing prevalence of AD, with increasing disease burden within socioeconomically advantaged countries, raises the possibility of early modifiable environmental factors that may contribute to the disease process.¹⁰ Many studies point to the influence of the 21st century lifestyle and Western diet as primary contributing factors.^9,11 However, it is not clear how these factors may influence the development of allergic atopic disease. Several studies have suggested that nonheritable influences in utero can alter fetus immune function and influence the subsequent development of allergic disease.^12,13 Although many studies have examined environmental factors contributing to the development of AD in infancy and childhood, less is understood about the influence of prenatal factors. Currently, in utero exposure to tobacco smoke, phthalates, and maternal distress have been potentially implicated in the development of AD.^14,15 Several studies have examined the role of maternal diet and nutrition on the development of AD in offspring; however, formal recommendations and robust trial data are lacking. In this article, we examine the existing literature surrounding maternal diet on the development of AD in infancy and childhood.

Allergen Avoidance 

Extrapolating from the food allergy literature, it was once suggested that allergen avoidance in early childhood had a protective effect on the subsequent development of allergies; however, more recent research has found that early exposure to common food allergens, such as peanuts or eggs, may actually reduce a child’s risk for developing these allergies later in life.¹⁶ Among infants at high risk for food allergy, sustained consumption of peanut products beginning in the first 11 months of life resulted in an 81% lower rate of peanut allergy at 60 months of age than the rate among children who avoided peanuts.¹⁷ Given the results that antigen avoidance during infancy/childhood does not protect against the development of allergies and may actually be counterproductive, it is not surprising that research studying antigen avoidance during pregnancy on the development of AD also has demonstrated limited efficacy. A systematic review of 5 trials on maternal dietary antigen avoidance (N=952) suggested no protective effects of avoiding antigenic foods during pregnancy on the development of AD in the first 18 months of life.¹⁸ Another meta-analysis evaluating 12 intervention trials looked at the effects of maternal allergenic food avoidance during pregnancy or lactation and found no reduced risk for subsequent development of allergic disease, including AD.¹⁹ The American Academy of Pediatrics 2019 consensus statement does not support maternal dietary restrictions in pregnancy for the prevention of atopic disease and makes note that the data remain limited, which complicates drawing any firm conclusions.²⁰

Probiotic Supplementation 

One of the most investigated dietary supplements for the prevention of atopic disease is probiotics, with possible benefits noted in both the prenatal and postnatal periods. Baquerizo Nole et al²¹ examined several studies looking at the various benefits of probiotics in AD, which included inhibition of the helper T cell (T_H2) response, stimulation of the T_H1 response, upregulation of regulatory T cells, acceleration of skin and mucosal barrier function, increase in intestinal microflora diversity, suppression of toxic fermentation products in the intestinal lumen from increased production of short-chain fatty acids, and inhibition of Staphylococcus aureus attachment on epidermal keratinocytes. It is unclear how this may affect infants prenatally; however, transfer of maternal intestinal microflora during delivery and shortly thereafter has demonstrated that probiotic strains remain detectable in the infant’s stool up to 6 months after delivery, even if the mother has discontinued use.²² A 2008 meta-analysis of 10 double-bind, randomized, controlled trials (N=1880) looking at the use of maternal prenatal and postnatal probiotic supplementation in the prevention of pediatric AD found a relative risk (RR) ratio of 0.69 (95% CI, 0.57-0.83) using a fixed effects model and RR ratio of 0.66 (95% CI, 0.49-0.89) using a random effects model. After exclusion of one study that evaluated the effect of postnatal probiotic supplementation only, the RR ratio decreased to 0.61 for both the fixed effects and random effects models.²³ A systematic review by Panduru et al²⁴ noted similar findings with a subgroup meta-analysis of 11 studies of prenatal supplementation followed by postnatal supplementation of probiotics, which demonstrated a protective effect on the development of AD (odds ratio [OR]=0.61, P<.001). Postnatal supplementation alone (4 studies) did not have the same association (OR=0.95, P<.82).²⁴ A 2012 meta-analysis by Doege et al²⁵ evaluated 7 randomized, double-blinded, placebo-controlled trials that assessed probiotic supplementation during pregnancy (without incorporation of postnatal supplementation) and found a significant risk reduction of 5.7% (P=.022) for AD in children aged 2 to 7 years. Interestingly, this was only significant for Lactobacillus and not for other bacterial strains, even if a mixture of strains included Lactobacillus. However, Panduru et al²⁴ found both maternal Lactobacillus supplementation alone (8 studies) and in combination with Bifidobacterium (9 studies) was protective against AD development in children (OR=0.70, P=.004; OR=0.62, P<.001). A more recent 2015 meta-analysis of 17 studies (N=4755) evaluating the use of maternal probiotic supplementation in pregnancy and/or through the infant’s first 3 months of life found a significantly lower RR (0.78 [95% CI, 0.69-0.89], P=.0003) for the development of AD in infants treated with probiotics and found this risk to be even further decreased when a mixture of probiotics including both Lactobacillus and Bifidobacterium was used (RR=0.54 [95% CI, 0.43-0.68], P<.00001).²⁶

Antioxidants

The Westernization of many developing countries’ diets—diets high in saturated fats, protein, sucrose, salt, and processed foods and low in fresh fruits and green vegetables—has led to a reduced intake of antioxidants and an increase in susceptibility to oxidative damage.^27,28 One hypothesis suggests that a reduction in nutritional antioxidants and subsequent oxidative damage leads to airway inflammation that may contribute to an increased prevalence of asthma.²⁷ In vitro data suggest that antioxidant deficiency may influence the differentiation of helper T cells to a T_H2 phenotype, which can increase susceptibility to the development of asthma and allergies.²⁹ Vitamin E specifically has been shown to inhibit IL-4 gene expression, which drives type 2 immunity and decreases expression of multiple genes that regulate epidermal barrier function, subsequently increasing susceptibility to allergic inflammation and AD.^29,30 Regardless of the proposed mechanisms for antioxidant deficiency increasing susceptibility to allergic disease, studies evaluating the benefits of antioxidant intake during pregnancy in relation to AD have not been promising. Several studies have found no association between prenatal vitamin E intake and the risk for AD development in infants and children.^31,32 Another study found a statistically significant inverse relationship between vitamin E intake in mothers with a history of atopy and the development of AD in their children at 2 years of age but not at 1 year of age (P-trend=.024).³³ It has been suggested that varying vitamin E isoforms may contribute to the discrepant results previously discussed, with the γ-tocopherol isoform (found frequently in Westernized diets)³⁴ as a driver of inflammation in murine models.³⁵ West et al³¹ noted an association between vitamin C intake and development of “any allergic disease”—AD, IgE-mediated food allergy, or asthma—with a crude OR of 0.48 (95% CI, 0.25-0.93). However, the P-trend and adjusted OR were not statistically significant. The investigators found no association between maternal intake of beta-carotene, vitamin E, or zinc, but they did find copper supplementation to be protective on the development of AD at 1 year of age (P-trend=0.03). Interestingly, when the data for total antioxidant intake—vitamin C, vitamin E, zinc, beta-carotene, and copper from both diet and supplementation—were combined and analyzed, no statistically significant associations for any of the antioxidants were found.³¹ Another study of 763 Japanese mother-child pairs found a reduced risk for AD at 16 to 24 months of age with high maternal intake of beta-carotene but found no statistically significant exposure-response associations with other antioxidants, including alpha-carotene, vitamin C, or zinc from dietary intake alone.³² These results were substantiated by 2 meta-analyses evaluating a total of 93 combined intervention trials and cohorts where no association was found between vitamin or mineral intake during pregnancy and/or during infancy and the development of AD.^19,36 

Fatty Acids 

Other dietary changes that are associated with an increased prevalence of atopic diseases in children include excess consumption of omega-6 (n-6) long-chain polyunsaturated fatty acids (LC-PUFA) and insufficient omega-3 (n-3) LC-PUFA consumption.³⁷ Given prior evidence that allergic immune responses in infants may be primed before birth,³⁸ researchers have questioned whether the anti-inflammatory properties of n-3 LC-PUFA when supplemented during pregnancy may have immunomodulatory effects on infants that could alter their predisposition to develop allergic disease, including AD.³⁹ A systematic review and meta-analysis of randomized controlled trials found a statistically significant RR of 0.53 (95% CI, 0.35-0.81; P=.004) for the incidence of AD at 12 months of age with maternal supplementation of n-3 LC-PUFA.⁹ Another trial of 145 pregnant women randomized to supplementation with fish oil vs placebo starting at gestational week 25 and continuing through 3.5 months of breastfeeding found a reduced cumulative incidence of AD in the intervention group compared to controls at 2 years of age, with a statistically significant crude OR of 0.33 (95% CI, 0.11-0.97; P=.04).⁴⁰ However, the adjusted OR was not statistically significant. In addition, they found that mothers and infants with higher proportions of docosahexaenoic acid and eicosapentaenoic acid in plasma phospholipids have been noted to have a lower prevalence of IgE-associated disease in a dose-dependent manner (P<.05 and P<.05, respectively).⁴⁰ In another trial of 98 pregnant women randomized to fish oil supplementation or placebo from 20 weeks’ gestation to delivery found no difference in the frequency of AD but did note that infants in the exposure group had significantly less severe AD compared to controls (OR=0.09 [95% CI, 0.1-0.94]; P=.045).³⁹ A prospective birth cohort study of 2641 children evaluated dietary composition during the last 4 weeks of pregnancy and found that consumption of foods rich in n-6 LC-PUFAs (eg, margarine, vegetable oil) increased the risk for developing AD, while foods rich in n-3 LC-PUFAs (eg, fish) decreased the risk for developing AD in offspring at 2 years of age. All P values for margarine, vegetable oil, and fish were statistically significant on logistic regression at P<.05.⁴¹ A longitudinal analysis of follow-up data from a randomized controlled trial looking at maternal prenatal n-3 LC-PUFA intake and the development of allergic disease (including AD) found no differences in the development of disease at 1-, 3-, or 6-year follow-up.⁴² Despite several studies demonstrating a possible benefit of omega-3 fatty acid intake on the development of AD in offspring, the longitudinal analysis by Best et al⁴² reminds us that long-term follow-up is critical in establishing benefit of any intervention given the heterogeneous and progressive nature of the atopic march and AD. 

Specific Diets 

Several studies have evaluated the role of dietary patterns and their influence on atopic disease. Studies evaluating dietary patterns or supplement intake can be challenging, as data often are derived from questionnaires with bias in response to families with higher socioeconomic status.⁹ Further, analysis of any one food group does not account for the potential interplay between nutrients.⁴³ Studies should focus more on dietary patterns vs individual foods to assess true risk.^43,44 Given these limitations, study results on diet should be carefully scrutinized; however, there are still some positive findings that deserve further investigation. Chatzi et al⁴⁴ followed 460 children for 6.5 years and found a protective effect for the development of atopy in the offspring of women who had high adherence to the Mediterranean diet (OR 0.55 [95% CI, 0.31-0.97]). Another cohort study evaluating the effects of the Mediterranean diet and risk for AD in the first year of life in 2516 mother-child pairs from Spain and Greece found no statistically significant association with consumption of the Mediterranean diet and AD. The investigators also evaluated intake of fruits, nuts, vegetables, meats, processed meats, dairy products, and cereal and found no statistically significant protective benefit.⁴⁵ Another systematic review of more than 90 observational studies identified no significant relationship between prenatal dietary exposures of fruits, vegetables, nuts, fat, fatty acids, eggs, cereal, milk, alcohol, tea, or coffee and risk for allergic disease in offspring, including AD.¹⁹

A Chinese prospective cohort study evaluated the dietary protein patterns of 713 mother-child pairs and the incidence of infant AD at 6 months of age.⁴⁶ Dietary protein patterns were characterized as predominantly poultry, plant based, dairy and eggs, and red meat and fish. The investigators found a statistically significant reduced risk for AD in mothers who consumed plant-based or dairy and eggs protein patterns when compared to a poultry protein pattern with an adjusted OR of 0.572 (95% CI, 0.330-0.992) and 0.478 (95% CI, 0.274-0.837), respectively. This protective effect was not seen with the red meat and fish protein patterns.⁴⁶ Similar results were seen in a 2020 Canadian study that evaluated the effects of a Western (fats, meats, processed foods, and starchy vegetables), balanced (diverse sources of animal proteins [especially fish], fruits, vegetables, nuts, and seeds), or plant-based (dairy, legumes, vegetables, whole grains, and an aversion to meats) diet in more than 2000 mother-infant pairs from 24 to 28 weeks’ gestation to 1 year of age. The investigators found a lower OR of AD in mothers who followed a mostly plant-based diet compared to other dietary patterns (OR 0.65 [95% CI, 0.55-0.76]; P<.001).¹⁰ Another prospective Japanese study looking at healthy (high intake of green and yellow vegetables, seaweed, mushrooms, white vegetables, pulses, potatoes, fish, sea products, fruit, and shellfish, and low intake of confectioneries and soft drinks), Western (high intake of vegetable oil, salt-containing seasonings, beef, pork, processed meat, eggs, chicken, and white vegetables, and low intake of fruit, soft drinks, and confectioneries), or Japanese (high intake of rice, miso soup, sea products, and fish, and low intake of bread, confectioneries, and dairy products) dietary patterns in 763 mother-child pairs found no association between diet during pregnancy and development of AD in offspring at 16 to 24 months.⁴⁷ Unfortunately, a longitudinal data analysis has not been performed for this study.

Final Thoughts

Atopic dermatitis is a complex, progressive, and heterogeneous disease with both genetic and environmental influences. Studying the effects of diet on the development, progression, or severity of disease can be very difficult due to the heterogeneity of study designs, lack of long-term follow-up, and high potential for residual confounding. Studies evaluating dietary patterns or supplement intake can be equally challenging, as data often are derived from questionnaires with bias in response to families with higher socioeconomic status.⁹ Very few studies have looked specifically at maternal dietary composition and the development of AD alone (without inclusion of asthma or food allergy). Ultimately, the inconsistency of the data makes it difficult to draw conclusions and make formal recommendations for this vulnerable population. Additional evidence from well-powered trials with comparable methodology and objective outcome measures will be imperative to make formal recommendations. In addition, longitudinal follow-up will be essential to determine long-term benefit and influence on the atopic march.

Atopic dermatitis (AD) is an inflammatory skin disease characterized by skin barrier disruption, skin inflammation, and pruritus.¹ It is a common and often chronic skin condition associated with the development of food allergies, asthma, and allergic rhinitis, known as the atopic march.² Atopic dermatitis is estimated to affect 10% to 25% of children, most with onset before 5 years of age, and up to 7% of adults worldwide.³ Most patients improve with time, but multiple disease trajectories are possible. Several studies have demonstrated that fewer than 4% of children develop the classic atopic march—AD followed by food allergies, asthma, and finally allergic rhinitis—with recent evidence pointing to a more complex heterogeneous progression of disease and allergic comorbidities often occurring together.^4,5 The prevalence of AD has been increasing globally over the last 30 years,⁶ with a marked increase in developed countries.^6,7 It is well accepted that AD is based on an interplay between genetic predisposition and environmental factors,⁸ but many suspect that the rapid rise in prevalence cannot be attributed to genetic factors alone.⁹ The precipitant triggers for AD remain an area of intense investigation, with ongoing debate between the “inside out” and “outside in” hypotheses; these revolve around whether abnormalities in the immune system trigger barrier dysfunction or barrier dysfunction triggers immune programming to atopy.⁸ Ongoing research related to genetic predisposition of AD has identified candidate genes implicated in both impaired skin barrier function and altered immune system pathways, further supporting that both theories may contribute to disease pathogenesis. 

The increasing prevalence of AD, with increasing disease burden within socioeconomically advantaged countries, raises the possibility of early modifiable environmental factors that may contribute to the disease process.¹⁰ Many studies point to the influence of the 21st century lifestyle and Western diet as primary contributing factors.^9,11 However, it is not clear how these factors may influence the development of allergic atopic disease. Several studies have suggested that nonheritable influences in utero can alter fetus immune function and influence the subsequent development of allergic disease.^12,13 Although many studies have examined environmental factors contributing to the development of AD in infancy and childhood, less is understood about the influence of prenatal factors. Currently, in utero exposure to tobacco smoke, phthalates, and maternal distress have been potentially implicated in the development of AD.^14,15 Several studies have examined the role of maternal diet and nutrition on the development of AD in offspring; however, formal recommendations and robust trial data are lacking. In this article, we examine the existing literature surrounding maternal diet on the development of AD in infancy and childhood.

Allergen Avoidance 

Extrapolating from the food allergy literature, it was once suggested that allergen avoidance in early childhood had a protective effect on the subsequent development of allergies; however, more recent research has found that early exposure to common food allergens, such as peanuts or eggs, may actually reduce a child’s risk for developing these allergies later in life.¹⁶ Among infants at high risk for food allergy, sustained consumption of peanut products beginning in the first 11 months of life resulted in an 81% lower rate of peanut allergy at 60 months of age than the rate among children who avoided peanuts.¹⁷ Given the results that antigen avoidance during infancy/childhood does not protect against the development of allergies and may actually be counterproductive, it is not surprising that research studying antigen avoidance during pregnancy on the development of AD also has demonstrated limited efficacy. A systematic review of 5 trials on maternal dietary antigen avoidance (N=952) suggested no protective effects of avoiding antigenic foods during pregnancy on the development of AD in the first 18 months of life.¹⁸ Another meta-analysis evaluating 12 intervention trials looked at the effects of maternal allergenic food avoidance during pregnancy or lactation and found no reduced risk for subsequent development of allergic disease, including AD.¹⁹ The American Academy of Pediatrics 2019 consensus statement does not support maternal dietary restrictions in pregnancy for the prevention of atopic disease and makes note that the data remain limited, which complicates drawing any firm conclusions.²⁰

Probiotic Supplementation 

One of the most investigated dietary supplements for the prevention of atopic disease is probiotics, with possible benefits noted in both the prenatal and postnatal periods. Baquerizo Nole et al²¹ examined several studies looking at the various benefits of probiotics in AD, which included inhibition of the helper T cell (T_H2) response, stimulation of the T_H1 response, upregulation of regulatory T cells, acceleration of skin and mucosal barrier function, increase in intestinal microflora diversity, suppression of toxic fermentation products in the intestinal lumen from increased production of short-chain fatty acids, and inhibition of Staphylococcus aureus attachment on epidermal keratinocytes. It is unclear how this may affect infants prenatally; however, transfer of maternal intestinal microflora during delivery and shortly thereafter has demonstrated that probiotic strains remain detectable in the infant’s stool up to 6 months after delivery, even if the mother has discontinued use.²² A 2008 meta-analysis of 10 double-bind, randomized, controlled trials (N=1880) looking at the use of maternal prenatal and postnatal probiotic supplementation in the prevention of pediatric AD found a relative risk (RR) ratio of 0.69 (95% CI, 0.57-0.83) using a fixed effects model and RR ratio of 0.66 (95% CI, 0.49-0.89) using a random effects model. After exclusion of one study that evaluated the effect of postnatal probiotic supplementation only, the RR ratio decreased to 0.61 for both the fixed effects and random effects models.²³ A systematic review by Panduru et al²⁴ noted similar findings with a subgroup meta-analysis of 11 studies of prenatal supplementation followed by postnatal supplementation of probiotics, which demonstrated a protective effect on the development of AD (odds ratio [OR]=0.61, P<.001). Postnatal supplementation alone (4 studies) did not have the same association (OR=0.95, P<.82).²⁴ A 2012 meta-analysis by Doege et al²⁵ evaluated 7 randomized, double-blinded, placebo-controlled trials that assessed probiotic supplementation during pregnancy (without incorporation of postnatal supplementation) and found a significant risk reduction of 5.7% (P=.022) for AD in children aged 2 to 7 years. Interestingly, this was only significant for Lactobacillus and not for other bacterial strains, even if a mixture of strains included Lactobacillus. However, Panduru et al²⁴ found both maternal Lactobacillus supplementation alone (8 studies) and in combination with Bifidobacterium (9 studies) was protective against AD development in children (OR=0.70, P=.004; OR=0.62, P<.001). A more recent 2015 meta-analysis of 17 studies (N=4755) evaluating the use of maternal probiotic supplementation in pregnancy and/or through the infant’s first 3 months of life found a significantly lower RR (0.78 [95% CI, 0.69-0.89], P=.0003) for the development of AD in infants treated with probiotics and found this risk to be even further decreased when a mixture of probiotics including both Lactobacillus and Bifidobacterium was used (RR=0.54 [95% CI, 0.43-0.68], P<.00001).²⁶

Antioxidants

The Westernization of many developing countries’ diets—diets high in saturated fats, protein, sucrose, salt, and processed foods and low in fresh fruits and green vegetables—has led to a reduced intake of antioxidants and an increase in susceptibility to oxidative damage.^27,28 One hypothesis suggests that a reduction in nutritional antioxidants and subsequent oxidative damage leads to airway inflammation that may contribute to an increased prevalence of asthma.²⁷ In vitro data suggest that antioxidant deficiency may influence the differentiation of helper T cells to a T_H2 phenotype, which can increase susceptibility to the development of asthma and allergies.²⁹ Vitamin E specifically has been shown to inhibit IL-4 gene expression, which drives type 2 immunity and decreases expression of multiple genes that regulate epidermal barrier function, subsequently increasing susceptibility to allergic inflammation and AD.^29,30 Regardless of the proposed mechanisms for antioxidant deficiency increasing susceptibility to allergic disease, studies evaluating the benefits of antioxidant intake during pregnancy in relation to AD have not been promising. Several studies have found no association between prenatal vitamin E intake and the risk for AD development in infants and children.^31,32 Another study found a statistically significant inverse relationship between vitamin E intake in mothers with a history of atopy and the development of AD in their children at 2 years of age but not at 1 year of age (P-trend=.024).³³ It has been suggested that varying vitamin E isoforms may contribute to the discrepant results previously discussed, with the γ-tocopherol isoform (found frequently in Westernized diets)³⁴ as a driver of inflammation in murine models.³⁵ West et al³¹ noted an association between vitamin C intake and development of “any allergic disease”—AD, IgE-mediated food allergy, or asthma—with a crude OR of 0.48 (95% CI, 0.25-0.93). However, the P-trend and adjusted OR were not statistically significant. The investigators found no association between maternal intake of beta-carotene, vitamin E, or zinc, but they did find copper supplementation to be protective on the development of AD at 1 year of age (P-trend=0.03). Interestingly, when the data for total antioxidant intake—vitamin C, vitamin E, zinc, beta-carotene, and copper from both diet and supplementation—were combined and analyzed, no statistically significant associations for any of the antioxidants were found.³¹ Another study of 763 Japanese mother-child pairs found a reduced risk for AD at 16 to 24 months of age with high maternal intake of beta-carotene but found no statistically significant exposure-response associations with other antioxidants, including alpha-carotene, vitamin C, or zinc from dietary intake alone.³² These results were substantiated by 2 meta-analyses evaluating a total of 93 combined intervention trials and cohorts where no association was found between vitamin or mineral intake during pregnancy and/or during infancy and the development of AD.^19,36 

Fatty Acids 

Other dietary changes that are associated with an increased prevalence of atopic diseases in children include excess consumption of omega-6 (n-6) long-chain polyunsaturated fatty acids (LC-PUFA) and insufficient omega-3 (n-3) LC-PUFA consumption.³⁷ Given prior evidence that allergic immune responses in infants may be primed before birth,³⁸ researchers have questioned whether the anti-inflammatory properties of n-3 LC-PUFA when supplemented during pregnancy may have immunomodulatory effects on infants that could alter their predisposition to develop allergic disease, including AD.³⁹ A systematic review and meta-analysis of randomized controlled trials found a statistically significant RR of 0.53 (95% CI, 0.35-0.81; P=.004) for the incidence of AD at 12 months of age with maternal supplementation of n-3 LC-PUFA.⁹ Another trial of 145 pregnant women randomized to supplementation with fish oil vs placebo starting at gestational week 25 and continuing through 3.5 months of breastfeeding found a reduced cumulative incidence of AD in the intervention group compared to controls at 2 years of age, with a statistically significant crude OR of 0.33 (95% CI, 0.11-0.97; P=.04).⁴⁰ However, the adjusted OR was not statistically significant. In addition, they found that mothers and infants with higher proportions of docosahexaenoic acid and eicosapentaenoic acid in plasma phospholipids have been noted to have a lower prevalence of IgE-associated disease in a dose-dependent manner (P<.05 and P<.05, respectively).⁴⁰ In another trial of 98 pregnant women randomized to fish oil supplementation or placebo from 20 weeks’ gestation to delivery found no difference in the frequency of AD but did note that infants in the exposure group had significantly less severe AD compared to controls (OR=0.09 [95% CI, 0.1-0.94]; P=.045).³⁹ A prospective birth cohort study of 2641 children evaluated dietary composition during the last 4 weeks of pregnancy and found that consumption of foods rich in n-6 LC-PUFAs (eg, margarine, vegetable oil) increased the risk for developing AD, while foods rich in n-3 LC-PUFAs (eg, fish) decreased the risk for developing AD in offspring at 2 years of age. All P values for margarine, vegetable oil, and fish were statistically significant on logistic regression at P<.05.⁴¹ A longitudinal analysis of follow-up data from a randomized controlled trial looking at maternal prenatal n-3 LC-PUFA intake and the development of allergic disease (including AD) found no differences in the development of disease at 1-, 3-, or 6-year follow-up.⁴² Despite several studies demonstrating a possible benefit of omega-3 fatty acid intake on the development of AD in offspring, the longitudinal analysis by Best et al⁴² reminds us that long-term follow-up is critical in establishing benefit of any intervention given the heterogeneous and progressive nature of the atopic march and AD. 

Specific Diets 

Several studies have evaluated the role of dietary patterns and their influence on atopic disease. Studies evaluating dietary patterns or supplement intake can be challenging, as data often are derived from questionnaires with bias in response to families with higher socioeconomic status.⁹ Further, analysis of any one food group does not account for the potential interplay between nutrients.⁴³ Studies should focus more on dietary patterns vs individual foods to assess true risk.^43,44 Given these limitations, study results on diet should be carefully scrutinized; however, there are still some positive findings that deserve further investigation. Chatzi et al⁴⁴ followed 460 children for 6.5 years and found a protective effect for the development of atopy in the offspring of women who had high adherence to the Mediterranean diet (OR 0.55 [95% CI, 0.31-0.97]). Another cohort study evaluating the effects of the Mediterranean diet and risk for AD in the first year of life in 2516 mother-child pairs from Spain and Greece found no statistically significant association with consumption of the Mediterranean diet and AD. The investigators also evaluated intake of fruits, nuts, vegetables, meats, processed meats, dairy products, and cereal and found no statistically significant protective benefit.⁴⁵ Another systematic review of more than 90 observational studies identified no significant relationship between prenatal dietary exposures of fruits, vegetables, nuts, fat, fatty acids, eggs, cereal, milk, alcohol, tea, or coffee and risk for allergic disease in offspring, including AD.¹⁹

A Chinese prospective cohort study evaluated the dietary protein patterns of 713 mother-child pairs and the incidence of infant AD at 6 months of age.⁴⁶ Dietary protein patterns were characterized as predominantly poultry, plant based, dairy and eggs, and red meat and fish. The investigators found a statistically significant reduced risk for AD in mothers who consumed plant-based or dairy and eggs protein patterns when compared to a poultry protein pattern with an adjusted OR of 0.572 (95% CI, 0.330-0.992) and 0.478 (95% CI, 0.274-0.837), respectively. This protective effect was not seen with the red meat and fish protein patterns.⁴⁶ Similar results were seen in a 2020 Canadian study that evaluated the effects of a Western (fats, meats, processed foods, and starchy vegetables), balanced (diverse sources of animal proteins [especially fish], fruits, vegetables, nuts, and seeds), or plant-based (dairy, legumes, vegetables, whole grains, and an aversion to meats) diet in more than 2000 mother-infant pairs from 24 to 28 weeks’ gestation to 1 year of age. The investigators found a lower OR of AD in mothers who followed a mostly plant-based diet compared to other dietary patterns (OR 0.65 [95% CI, 0.55-0.76]; P<.001).¹⁰ Another prospective Japanese study looking at healthy (high intake of green and yellow vegetables, seaweed, mushrooms, white vegetables, pulses, potatoes, fish, sea products, fruit, and shellfish, and low intake of confectioneries and soft drinks), Western (high intake of vegetable oil, salt-containing seasonings, beef, pork, processed meat, eggs, chicken, and white vegetables, and low intake of fruit, soft drinks, and confectioneries), or Japanese (high intake of rice, miso soup, sea products, and fish, and low intake of bread, confectioneries, and dairy products) dietary patterns in 763 mother-child pairs found no association between diet during pregnancy and development of AD in offspring at 16 to 24 months.⁴⁷ Unfortunately, a longitudinal data analysis has not been performed for this study.

Final Thoughts

Atopic dermatitis is a complex, progressive, and heterogeneous disease with both genetic and environmental influences. Studying the effects of diet on the development, progression, or severity of disease can be very difficult due to the heterogeneity of study designs, lack of long-term follow-up, and high potential for residual confounding. Studies evaluating dietary patterns or supplement intake can be equally challenging, as data often are derived from questionnaires with bias in response to families with higher socioeconomic status.⁹ Very few studies have looked specifically at maternal dietary composition and the development of AD alone (without inclusion of asthma or food allergy). Ultimately, the inconsistency of the data makes it difficult to draw conclusions and make formal recommendations for this vulnerable population. Additional evidence from well-powered trials with comparable methodology and objective outcome measures will be imperative to make formal recommendations. In addition, longitudinal follow-up will be essential to determine long-term benefit and influence on the atopic march.

Atopic dermatitis (AD) is an inflammatory skin disease characterized by skin barrier disruption, skin inflammation, and pruritus.¹ It is a common and often chronic skin condition associated with the development of food allergies, asthma, and allergic rhinitis, known as the atopic march.² Atopic dermatitis is estimated to affect 10% to 25% of children, most with onset before 5 years of age, and up to 7% of adults worldwide.³ Most patients improve with time, but multiple disease trajectories are possible. Several studies have demonstrated that fewer than 4% of children develop the classic atopic march—AD followed by food allergies, asthma, and finally allergic rhinitis—with recent evidence pointing to a more complex heterogeneous progression of disease and allergic comorbidities often occurring together.^4,5 The prevalence of AD has been increasing globally over the last 30 years,⁶ with a marked increase in developed countries.^6,7 It is well accepted that AD is based on an interplay between genetic predisposition and environmental factors,⁸ but many suspect that the rapid rise in prevalence cannot be attributed to genetic factors alone.⁹ The precipitant triggers for AD remain an area of intense investigation, with ongoing debate between the “inside out” and “outside in” hypotheses; these revolve around whether abnormalities in the immune system trigger barrier dysfunction or barrier dysfunction triggers immune programming to atopy.⁸ Ongoing research related to genetic predisposition of AD has identified candidate genes implicated in both impaired skin barrier function and altered immune system pathways, further supporting that both theories may contribute to disease pathogenesis. 

The increasing prevalence of AD, with increasing disease burden within socioeconomically advantaged countries, raises the possibility of early modifiable environmental factors that may contribute to the disease process.¹⁰ Many studies point to the influence of the 21st century lifestyle and Western diet as primary contributing factors.^9,11 However, it is not clear how these factors may influence the development of allergic atopic disease. Several studies have suggested that nonheritable influences in utero can alter fetus immune function and influence the subsequent development of allergic disease.^12,13 Although many studies have examined environmental factors contributing to the development of AD in infancy and childhood, less is understood about the influence of prenatal factors. Currently, in utero exposure to tobacco smoke, phthalates, and maternal distress have been potentially implicated in the development of AD.^14,15 Several studies have examined the role of maternal diet and nutrition on the development of AD in offspring; however, formal recommendations and robust trial data are lacking. In this article, we examine the existing literature surrounding maternal diet on the development of AD in infancy and childhood.

Allergen Avoidance 

Extrapolating from the food allergy literature, it was once suggested that allergen avoidance in early childhood had a protective effect on the subsequent development of allergies; however, more recent research has found that early exposure to common food allergens, such as peanuts or eggs, may actually reduce a child’s risk for developing these allergies later in life.¹⁶ Among infants at high risk for food allergy, sustained consumption of peanut products beginning in the first 11 months of life resulted in an 81% lower rate of peanut allergy at 60 months of age than the rate among children who avoided peanuts.¹⁷ Given the results that antigen avoidance during infancy/childhood does not protect against the development of allergies and may actually be counterproductive, it is not surprising that research studying antigen avoidance during pregnancy on the development of AD also has demonstrated limited efficacy. A systematic review of 5 trials on maternal dietary antigen avoidance (N=952) suggested no protective effects of avoiding antigenic foods during pregnancy on the development of AD in the first 18 months of life.¹⁸ Another meta-analysis evaluating 12 intervention trials looked at the effects of maternal allergenic food avoidance during pregnancy or lactation and found no reduced risk for subsequent development of allergic disease, including AD.¹⁹ The American Academy of Pediatrics 2019 consensus statement does not support maternal dietary restrictions in pregnancy for the prevention of atopic disease and makes note that the data remain limited, which complicates drawing any firm conclusions.²⁰

Probiotic Supplementation 

One of the most investigated dietary supplements for the prevention of atopic disease is probiotics, with possible benefits noted in both the prenatal and postnatal periods. Baquerizo Nole et al²¹ examined several studies looking at the various benefits of probiotics in AD, which included inhibition of the helper T cell (T_H2) response, stimulation of the T_H1 response, upregulation of regulatory T cells, acceleration of skin and mucosal barrier function, increase in intestinal microflora diversity, suppression of toxic fermentation products in the intestinal lumen from increased production of short-chain fatty acids, and inhibition of Staphylococcus aureus attachment on epidermal keratinocytes. It is unclear how this may affect infants prenatally; however, transfer of maternal intestinal microflora during delivery and shortly thereafter has demonstrated that probiotic strains remain detectable in the infant’s stool up to 6 months after delivery, even if the mother has discontinued use.²² A 2008 meta-analysis of 10 double-bind, randomized, controlled trials (N=1880) looking at the use of maternal prenatal and postnatal probiotic supplementation in the prevention of pediatric AD found a relative risk (RR) ratio of 0.69 (95% CI, 0.57-0.83) using a fixed effects model and RR ratio of 0.66 (95% CI, 0.49-0.89) using a random effects model. After exclusion of one study that evaluated the effect of postnatal probiotic supplementation only, the RR ratio decreased to 0.61 for both the fixed effects and random effects models.²³ A systematic review by Panduru et al²⁴ noted similar findings with a subgroup meta-analysis of 11 studies of prenatal supplementation followed by postnatal supplementation of probiotics, which demonstrated a protective effect on the development of AD (odds ratio [OR]=0.61, P<.001). Postnatal supplementation alone (4 studies) did not have the same association (OR=0.95, P<.82).²⁴ A 2012 meta-analysis by Doege et al²⁵ evaluated 7 randomized, double-blinded, placebo-controlled trials that assessed probiotic supplementation during pregnancy (without incorporation of postnatal supplementation) and found a significant risk reduction of 5.7% (P=.022) for AD in children aged 2 to 7 years. Interestingly, this was only significant for Lactobacillus and not for other bacterial strains, even if a mixture of strains included Lactobacillus. However, Panduru et al²⁴ found both maternal Lactobacillus supplementation alone (8 studies) and in combination with Bifidobacterium (9 studies) was protective against AD development in children (OR=0.70, P=.004; OR=0.62, P<.001). A more recent 2015 meta-analysis of 17 studies (N=4755) evaluating the use of maternal probiotic supplementation in pregnancy and/or through the infant’s first 3 months of life found a significantly lower RR (0.78 [95% CI, 0.69-0.89], P=.0003) for the development of AD in infants treated with probiotics and found this risk to be even further decreased when a mixture of probiotics including both Lactobacillus and Bifidobacterium was used (RR=0.54 [95% CI, 0.43-0.68], P<.00001).²⁶

Antioxidants

The Westernization of many developing countries’ diets—diets high in saturated fats, protein, sucrose, salt, and processed foods and low in fresh fruits and green vegetables—has led to a reduced intake of antioxidants and an increase in susceptibility to oxidative damage.^27,28 One hypothesis suggests that a reduction in nutritional antioxidants and subsequent oxidative damage leads to airway inflammation that may contribute to an increased prevalence of asthma.²⁷ In vitro data suggest that antioxidant deficiency may influence the differentiation of helper T cells to a T_H2 phenotype, which can increase susceptibility to the development of asthma and allergies.²⁹ Vitamin E specifically has been shown to inhibit IL-4 gene expression, which drives type 2 immunity and decreases expression of multiple genes that regulate epidermal barrier function, subsequently increasing susceptibility to allergic inflammation and AD.^29,30 Regardless of the proposed mechanisms for antioxidant deficiency increasing susceptibility to allergic disease, studies evaluating the benefits of antioxidant intake during pregnancy in relation to AD have not been promising. Several studies have found no association between prenatal vitamin E intake and the risk for AD development in infants and children.^31,32 Another study found a statistically significant inverse relationship between vitamin E intake in mothers with a history of atopy and the development of AD in their children at 2 years of age but not at 1 year of age (P-trend=.024).³³ It has been suggested that varying vitamin E isoforms may contribute to the discrepant results previously discussed, with the γ-tocopherol isoform (found frequently in Westernized diets)³⁴ as a driver of inflammation in murine models.³⁵ West et al³¹ noted an association between vitamin C intake and development of “any allergic disease”—AD, IgE-mediated food allergy, or asthma—with a crude OR of 0.48 (95% CI, 0.25-0.93). However, the P-trend and adjusted OR were not statistically significant. The investigators found no association between maternal intake of beta-carotene, vitamin E, or zinc, but they did find copper supplementation to be protective on the development of AD at 1 year of age (P-trend=0.03). Interestingly, when the data for total antioxidant intake—vitamin C, vitamin E, zinc, beta-carotene, and copper from both diet and supplementation—were combined and analyzed, no statistically significant associations for any of the antioxidants were found.³¹ Another study of 763 Japanese mother-child pairs found a reduced risk for AD at 16 to 24 months of age with high maternal intake of beta-carotene but found no statistically significant exposure-response associations with other antioxidants, including alpha-carotene, vitamin C, or zinc from dietary intake alone.³² These results were substantiated by 2 meta-analyses evaluating a total of 93 combined intervention trials and cohorts where no association was found between vitamin or mineral intake during pregnancy and/or during infancy and the development of AD.^19,36 

Fatty Acids 

Other dietary changes that are associated with an increased prevalence of atopic diseases in children include excess consumption of omega-6 (n-6) long-chain polyunsaturated fatty acids (LC-PUFA) and insufficient omega-3 (n-3) LC-PUFA consumption.³⁷ Given prior evidence that allergic immune responses in infants may be primed before birth,³⁸ researchers have questioned whether the anti-inflammatory properties of n-3 LC-PUFA when supplemented during pregnancy may have immunomodulatory effects on infants that could alter their predisposition to develop allergic disease, including AD.³⁹ A systematic review and meta-analysis of randomized controlled trials found a statistically significant RR of 0.53 (95% CI, 0.35-0.81; P=.004) for the incidence of AD at 12 months of age with maternal supplementation of n-3 LC-PUFA.⁹ Another trial of 145 pregnant women randomized to supplementation with fish oil vs placebo starting at gestational week 25 and continuing through 3.5 months of breastfeeding found a reduced cumulative incidence of AD in the intervention group compared to controls at 2 years of age, with a statistically significant crude OR of 0.33 (95% CI, 0.11-0.97; P=.04).⁴⁰ However, the adjusted OR was not statistically significant. In addition, they found that mothers and infants with higher proportions of docosahexaenoic acid and eicosapentaenoic acid in plasma phospholipids have been noted to have a lower prevalence of IgE-associated disease in a dose-dependent manner (P<.05 and P<.05, respectively).⁴⁰ In another trial of 98 pregnant women randomized to fish oil supplementation or placebo from 20 weeks’ gestation to delivery found no difference in the frequency of AD but did note that infants in the exposure group had significantly less severe AD compared to controls (OR=0.09 [95% CI, 0.1-0.94]; P=.045).³⁹ A prospective birth cohort study of 2641 children evaluated dietary composition during the last 4 weeks of pregnancy and found that consumption of foods rich in n-6 LC-PUFAs (eg, margarine, vegetable oil) increased the risk for developing AD, while foods rich in n-3 LC-PUFAs (eg, fish) decreased the risk for developing AD in offspring at 2 years of age. All P values for margarine, vegetable oil, and fish were statistically significant on logistic regression at P<.05.⁴¹ A longitudinal analysis of follow-up data from a randomized controlled trial looking at maternal prenatal n-3 LC-PUFA intake and the development of allergic disease (including AD) found no differences in the development of disease at 1-, 3-, or 6-year follow-up.⁴² Despite several studies demonstrating a possible benefit of omega-3 fatty acid intake on the development of AD in offspring, the longitudinal analysis by Best et al⁴² reminds us that long-term follow-up is critical in establishing benefit of any intervention given the heterogeneous and progressive nature of the atopic march and AD. 

Specific Diets 

Several studies have evaluated the role of dietary patterns and their influence on atopic disease. Studies evaluating dietary patterns or supplement intake can be challenging, as data often are derived from questionnaires with bias in response to families with higher socioeconomic status.⁹ Further, analysis of any one food group does not account for the potential interplay between nutrients.⁴³ Studies should focus more on dietary patterns vs individual foods to assess true risk.^43,44 Given these limitations, study results on diet should be carefully scrutinized; however, there are still some positive findings that deserve further investigation. Chatzi et al⁴⁴ followed 460 children for 6.5 years and found a protective effect for the development of atopy in the offspring of women who had high adherence to the Mediterranean diet (OR 0.55 [95% CI, 0.31-0.97]). Another cohort study evaluating the effects of the Mediterranean diet and risk for AD in the first year of life in 2516 mother-child pairs from Spain and Greece found no statistically significant association with consumption of the Mediterranean diet and AD. The investigators also evaluated intake of fruits, nuts, vegetables, meats, processed meats, dairy products, and cereal and found no statistically significant protective benefit.⁴⁵ Another systematic review of more than 90 observational studies identified no significant relationship between prenatal dietary exposures of fruits, vegetables, nuts, fat, fatty acids, eggs, cereal, milk, alcohol, tea, or coffee and risk for allergic disease in offspring, including AD.¹⁹

A Chinese prospective cohort study evaluated the dietary protein patterns of 713 mother-child pairs and the incidence of infant AD at 6 months of age.⁴⁶ Dietary protein patterns were characterized as predominantly poultry, plant based, dairy and eggs, and red meat and fish. The investigators found a statistically significant reduced risk for AD in mothers who consumed plant-based or dairy and eggs protein patterns when compared to a poultry protein pattern with an adjusted OR of 0.572 (95% CI, 0.330-0.992) and 0.478 (95% CI, 0.274-0.837), respectively. This protective effect was not seen with the red meat and fish protein patterns.⁴⁶ Similar results were seen in a 2020 Canadian study that evaluated the effects of a Western (fats, meats, processed foods, and starchy vegetables), balanced (diverse sources of animal proteins [especially fish], fruits, vegetables, nuts, and seeds), or plant-based (dairy, legumes, vegetables, whole grains, and an aversion to meats) diet in more than 2000 mother-infant pairs from 24 to 28 weeks’ gestation to 1 year of age. The investigators found a lower OR of AD in mothers who followed a mostly plant-based diet compared to other dietary patterns (OR 0.65 [95% CI, 0.55-0.76]; P<.001).¹⁰ Another prospective Japanese study looking at healthy (high intake of green and yellow vegetables, seaweed, mushrooms, white vegetables, pulses, potatoes, fish, sea products, fruit, and shellfish, and low intake of confectioneries and soft drinks), Western (high intake of vegetable oil, salt-containing seasonings, beef, pork, processed meat, eggs, chicken, and white vegetables, and low intake of fruit, soft drinks, and confectioneries), or Japanese (high intake of rice, miso soup, sea products, and fish, and low intake of bread, confectioneries, and dairy products) dietary patterns in 763 mother-child pairs found no association between diet during pregnancy and development of AD in offspring at 16 to 24 months.⁴⁷ Unfortunately, a longitudinal data analysis has not been performed for this study.

Final Thoughts

Atopic dermatitis is a complex, progressive, and heterogeneous disease with both genetic and environmental influences. Studying the effects of diet on the development, progression, or severity of disease can be very difficult due to the heterogeneity of study designs, lack of long-term follow-up, and high potential for residual confounding. Studies evaluating dietary patterns or supplement intake can be equally challenging, as data often are derived from questionnaires with bias in response to families with higher socioeconomic status.⁹ Very few studies have looked specifically at maternal dietary composition and the development of AD alone (without inclusion of asthma or food allergy). Ultimately, the inconsistency of the data makes it difficult to draw conclusions and make formal recommendations for this vulnerable population. Additional evidence from well-powered trials with comparable methodology and objective outcome measures will be imperative to make formal recommendations. In addition, longitudinal follow-up will be essential to determine long-term benefit and influence on the atopic march.

Incidence and Characteristics

Mosquitoes are insects categorized into the order of Diptera and family of Culicidae, and more than 3500 different species have been identified.¹ In the United States, the most common genus of mosquitoes is Aedes, with other common genera including Culex, Anopheles, Culiseta, and Coquillettidia. Most bites are performed by female rather than male mosquitoes, as it serves to complete their life cycle (Figure 1).¹

There are a variety of possible reactions to mosquito bites. Severe local reactions that are large (papules >30 mm in diameter) or are accompanied by systemic manifestations are referred to as hypersensitivity to mosquito bites (HMB).² These hypersensitivity reactions vary according to multiple factors, including comorbid conditions, genetic predisposition, and geographic location. The majority of the world’s population will exhibit local reactions to mosquito bites at some point during life, with the median age of onset of the first bite at 2 years of age.³ In a study by Arias-Cruz et al,⁴ the incidence of patient-reported large local reactions was 2.5%. Hypersensitivity to mosquito bites, perhaps the most rare reaction, is more common among Asian and Central American children.⁵ The median age of diagnosis for HMB is 7 years, and most reactions occur during the first 2 decades of life.^6,7

Clinical Presentation

Mosquitoes bite vertebrates in an attempt to feed and thus must locate the host’s blood vessels through a process known as probing, which often necessitates changing the bite site several times. Once the vessel is located and lacerated, the mosquito feeds either from the vessel directly or the hematoma around it. Not only does the bite cause trauma to the skin, but a cutaneous reaction also may occur in response to salivary gland secretions that concurrently are deposited in the host tissue.⁸ Mosquitoes’ salivary gland components are the primary cause of cutaneous reactions, as one study showed that bites from mosquitoes lacking salivary gland ducts were not associated with these reactions.⁹ Mosquito saliva contains a large number of compounds with biologic activities, including lysozymes, antibacterial glucosidases, anticoagulants, antiplatelet aggregating factors, and vasodilators, as well as a potentially large number of unknown allergenic proteins. As of 2016, 70 mosquito-derived allergens have been identified, but this number continues to grow.² After a bite from a mosquito, these compounds may result in host sensitization over time, though interestingly, sensitization to mosquito bites from a species different from the original offender does not occur due to lack of cross-reactivity between species.¹ 

Because mosquitoes reproduce by laying their eggs directly on or near water, people who live near bodies of water or wetlands are at the highest risk for mosquito bites. Patient factors that have been found to lead to increased rates of mosquito bites include lower microbial diversity on the skin, the presence of sweat or body odor, pregnancy, increased body temperature, type O blood, dark clothing, and perfumes.² Exaggerated bite reactions are associated with Epstein-Barr virus (EBV) infection and hematologic malignancies.¹⁰ 

Immediate hypersensitivity is mediated by a specific IgE antibody and is characterized by erythema and a wheal at the bite site that peaks within minutes of the bite. In contrast, delayed hypersensitivity is lymphocyte mediated; occurs 24 hours after the bite; and causes an indurated, pruritic, and erythematous 2- to 10-mm papule that may blister.¹¹ Although the evidence of immediate hypersensitivity disappears within hours, symptoms of delayed hypersensitivity may last days to weeks. Accompanying symptoms may include local swelling, pain, and warmth. The itch that often is experienced in conjunction with erythema and papule formation is elicited in 3 main ways: direct induction utilizing classic pruritic pathways, IgE-mediated hypersensitivity reaction to salivary components, and IgE-independent host immune response to salivary antigens. Papular urticaria is a common additional finding in children with mosquito bites.¹ As an individual is repeatedly bitten, they may undergo 5 stages of sensitization: stage I (neither immediate nor delayed reaction), stage II (delayed reaction), stage III (immediate and delayed reaction), stage IV (immediate reaction), and stage V (neither immediate or delayed reaction).¹¹

Although most mosquito bites cause common local reactions, patients rarely demonstrate systemic reactions that can be much more severe. Skeeter syndrome is a milder systemic response characterized by large local reactions (papules >30 mm in diameter) developing hours after a bite with accompanying fever.¹² The reaction typically peaks over days to weeks.² Although the reaction may resemble cellulitis clinically, a history of a preceding mosquito bite can help make the distinction.¹³ 

A more severe systemic reaction is HMB, which is characterized by intense local skin findings as well as generalized systemic symptoms. Initially, indurated, clear, or hemorrhagic bullae appear at the bite site (Figure 2). Later, there is progression to swelling, necrosis, and ulceration.¹⁰ Biopsies from the skin lesions associated with HMB reveal necrosis, interstitial and perivascular eosinophilic and lymphocytic infiltrates, and small vessels with fibrinoid necrosis.⁷ Systemically, high fever, general malaise, liver dysfunction, proteinuria, hematuria, hepatosplenomegaly, and lymph node enlargement may occur. Patients typically experience these severe symptoms each time they are bitten.¹⁰

The mechanism of the HMB reaction is complex but has a close association with natural killer (NK) cell lymphoproliferative disorder and EBV infection (Figure 3). In fact, it is not uncommon for HMB patients to develop malignant lymphomas during their clinical course, even those unrelated to EBV.¹⁴ Epstein-Barr virus, one of the human herpesviruses, produces latent infection in NK cells. It is hypothesized that after a mosquito bite, EBV may be reactivated within these cells by induced expression of the viral lytic-cycle transactivator gene BamHI Z fragment leftward open reading frame 1, BZLF1.⁶ In response to mosquito salivary gland components, CD4⁺ T cells proliferate and induce expression of the EBV oncogene latent membrane protein 1, LMP1, on NK cells, which then infiltrate the bite site.¹⁵ These EBV-infected NK cells also overexpress the Fas ligand, thus contributing to organ and tissue damage.⁶ In addition to activating oncogene expression on NK cells, T cells also activate the basophils and mast cells carrying mosquito-specific IgE, both of which also add to the severe skin reaction of HMB.¹⁵ The particular triad of HMB, chronic active EBV infection, and NK cell lymphoproliferative disorder commonly is known as HMB-EBV-NK or HEN disease.¹ Patients with HMB should be monitored for malignancy. The mortality of HMB is increased in patients in whom onset occurs when they are older than 9 years and with BZLF1 messenger RNA in skin lesions.⁶ 

Other rare reactions to mosquito bites include Wells syndrome, anaphylaxis, and superficial lymphangitis. Wells syndrome (also known as eosinophilic cellulitis) is characterized by erythematous or violaceous plaques and pruritic blisters. Although its etiology has not been defined, it is thought to be evoked or exacerbated by insect bites, with CD4⁺ T cells playing a primary role.¹ Anaphylaxis (angioedema, urticaria, and wheezing) rarely may occur due to mosquito salivary gland components but typically is caused by other stinging insects. Superficial lymphangitis, often misdiagnosed as an infection of the lymphatic system, presents within minutes as nontender pink streaks originating from the bite site. A biopsy with eosinophil and mast cell infiltrates consistent with an allergic-type reaction confirms the absence of infection. Patients respond well to glucocorticoid treatment.

Mosquitoes are vectors for many blood-borne diseases, including dengue hemorrhagic fever, malaria, Chikungunya virus, La Crosse encephalitis, St. Louis encephalitis, West Nile virus, and yellow fever.¹⁶ Additionally, scratching the bites may lead to superinfection and scarring.¹

Prevention and Treatment

Patients with known mosquito sensitivity should avoid areas of stagnant water and utilize preventative measures such as wearing protective clothing and using mosquito repellent containing DEET (N,N-diethyl-meta-toluamide), IR3535 (ethyl butylacetylaminopropionate), picaridin, or 2-undecanone (methyl nonyl ketone or IBI-246) when outdoors. Essential oils such as lemon, eucalyptus, citronella, and garlic are somewhat effective.¹ Additionally, prophylactic dosing of antihistamines may prevent milder reactions.

Although often supportive, treatment and management of mosquito bites depends on the extent of the reaction. For common local reactions, symptomatic management with topical anesthetics, calamine lotion, or corticosteroid creams is appropriate. If superinfection from scratching is a concern, antibiotics may be appropriate.

Management of more severe and systemic reactions such as HMB also is supportive, and the addition of oral corticosteroids to decrease inflammation is required.⁷ Severe HMB also has been treated with immunosuppressive and anticancer drugs, though the efficacy is limited. Venom immunotherapy is a preventative option for patients with mosquito-specific IgE antibodies, and hematopoietic stem cell transplant may be required in patients with HMB.^14,16

Conclusion

Mosquito allergens can cause a variety of reactions, ranging from those limited to the skin to those characterized by severe systemic effects. Although common local reactions can be symptomatically treated with topical medication, more severe reactions such as HMB require more involved clinical management. Hypersensitivity to mosquito bites is an important condition to recognize, as it is related to multiple organ impairment as well as later development of malignancy. Patients should be closely monitored during the entire clinical course and in the years following.

Incidence and Characteristics

Mosquitoes are insects categorized into the order of Diptera and family of Culicidae, and more than 3500 different species have been identified.¹ In the United States, the most common genus of mosquitoes is Aedes, with other common genera including Culex, Anopheles, Culiseta, and Coquillettidia. Most bites are performed by female rather than male mosquitoes, as it serves to complete their life cycle (Figure 1).¹

There are a variety of possible reactions to mosquito bites. Severe local reactions that are large (papules >30 mm in diameter) or are accompanied by systemic manifestations are referred to as hypersensitivity to mosquito bites (HMB).² These hypersensitivity reactions vary according to multiple factors, including comorbid conditions, genetic predisposition, and geographic location. The majority of the world’s population will exhibit local reactions to mosquito bites at some point during life, with the median age of onset of the first bite at 2 years of age.³ In a study by Arias-Cruz et al,⁴ the incidence of patient-reported large local reactions was 2.5%. Hypersensitivity to mosquito bites, perhaps the most rare reaction, is more common among Asian and Central American children.⁵ The median age of diagnosis for HMB is 7 years, and most reactions occur during the first 2 decades of life.^6,7

Clinical Presentation

Mosquitoes bite vertebrates in an attempt to feed and thus must locate the host’s blood vessels through a process known as probing, which often necessitates changing the bite site several times. Once the vessel is located and lacerated, the mosquito feeds either from the vessel directly or the hematoma around it. Not only does the bite cause trauma to the skin, but a cutaneous reaction also may occur in response to salivary gland secretions that concurrently are deposited in the host tissue.⁸ Mosquitoes’ salivary gland components are the primary cause of cutaneous reactions, as one study showed that bites from mosquitoes lacking salivary gland ducts were not associated with these reactions.⁹ Mosquito saliva contains a large number of compounds with biologic activities, including lysozymes, antibacterial glucosidases, anticoagulants, antiplatelet aggregating factors, and vasodilators, as well as a potentially large number of unknown allergenic proteins. As of 2016, 70 mosquito-derived allergens have been identified, but this number continues to grow.² After a bite from a mosquito, these compounds may result in host sensitization over time, though interestingly, sensitization to mosquito bites from a species different from the original offender does not occur due to lack of cross-reactivity between species.¹ 

Because mosquitoes reproduce by laying their eggs directly on or near water, people who live near bodies of water or wetlands are at the highest risk for mosquito bites. Patient factors that have been found to lead to increased rates of mosquito bites include lower microbial diversity on the skin, the presence of sweat or body odor, pregnancy, increased body temperature, type O blood, dark clothing, and perfumes.² Exaggerated bite reactions are associated with Epstein-Barr virus (EBV) infection and hematologic malignancies.¹⁰ 

Immediate hypersensitivity is mediated by a specific IgE antibody and is characterized by erythema and a wheal at the bite site that peaks within minutes of the bite. In contrast, delayed hypersensitivity is lymphocyte mediated; occurs 24 hours after the bite; and causes an indurated, pruritic, and erythematous 2- to 10-mm papule that may blister.¹¹ Although the evidence of immediate hypersensitivity disappears within hours, symptoms of delayed hypersensitivity may last days to weeks. Accompanying symptoms may include local swelling, pain, and warmth. The itch that often is experienced in conjunction with erythema and papule formation is elicited in 3 main ways: direct induction utilizing classic pruritic pathways, IgE-mediated hypersensitivity reaction to salivary components, and IgE-independent host immune response to salivary antigens. Papular urticaria is a common additional finding in children with mosquito bites.¹ As an individual is repeatedly bitten, they may undergo 5 stages of sensitization: stage I (neither immediate nor delayed reaction), stage II (delayed reaction), stage III (immediate and delayed reaction), stage IV (immediate reaction), and stage V (neither immediate or delayed reaction).¹¹

Although most mosquito bites cause common local reactions, patients rarely demonstrate systemic reactions that can be much more severe. Skeeter syndrome is a milder systemic response characterized by large local reactions (papules >30 mm in diameter) developing hours after a bite with accompanying fever.¹² The reaction typically peaks over days to weeks.² Although the reaction may resemble cellulitis clinically, a history of a preceding mosquito bite can help make the distinction.¹³ 

A more severe systemic reaction is HMB, which is characterized by intense local skin findings as well as generalized systemic symptoms. Initially, indurated, clear, or hemorrhagic bullae appear at the bite site (Figure 2). Later, there is progression to swelling, necrosis, and ulceration.¹⁰ Biopsies from the skin lesions associated with HMB reveal necrosis, interstitial and perivascular eosinophilic and lymphocytic infiltrates, and small vessels with fibrinoid necrosis.⁷ Systemically, high fever, general malaise, liver dysfunction, proteinuria, hematuria, hepatosplenomegaly, and lymph node enlargement may occur. Patients typically experience these severe symptoms each time they are bitten.¹⁰

The mechanism of the HMB reaction is complex but has a close association with natural killer (NK) cell lymphoproliferative disorder and EBV infection (Figure 3). In fact, it is not uncommon for HMB patients to develop malignant lymphomas during their clinical course, even those unrelated to EBV.¹⁴ Epstein-Barr virus, one of the human herpesviruses, produces latent infection in NK cells. It is hypothesized that after a mosquito bite, EBV may be reactivated within these cells by induced expression of the viral lytic-cycle transactivator gene BamHI Z fragment leftward open reading frame 1, BZLF1.⁶ In response to mosquito salivary gland components, CD4⁺ T cells proliferate and induce expression of the EBV oncogene latent membrane protein 1, LMP1, on NK cells, which then infiltrate the bite site.¹⁵ These EBV-infected NK cells also overexpress the Fas ligand, thus contributing to organ and tissue damage.⁶ In addition to activating oncogene expression on NK cells, T cells also activate the basophils and mast cells carrying mosquito-specific IgE, both of which also add to the severe skin reaction of HMB.¹⁵ The particular triad of HMB, chronic active EBV infection, and NK cell lymphoproliferative disorder commonly is known as HMB-EBV-NK or HEN disease.¹ Patients with HMB should be monitored for malignancy. The mortality of HMB is increased in patients in whom onset occurs when they are older than 9 years and with BZLF1 messenger RNA in skin lesions.⁶ 

Other rare reactions to mosquito bites include Wells syndrome, anaphylaxis, and superficial lymphangitis. Wells syndrome (also known as eosinophilic cellulitis) is characterized by erythematous or violaceous plaques and pruritic blisters. Although its etiology has not been defined, it is thought to be evoked or exacerbated by insect bites, with CD4⁺ T cells playing a primary role.¹ Anaphylaxis (angioedema, urticaria, and wheezing) rarely may occur due to mosquito salivary gland components but typically is caused by other stinging insects. Superficial lymphangitis, often misdiagnosed as an infection of the lymphatic system, presents within minutes as nontender pink streaks originating from the bite site. A biopsy with eosinophil and mast cell infiltrates consistent with an allergic-type reaction confirms the absence of infection. Patients respond well to glucocorticoid treatment.

Mosquitoes are vectors for many blood-borne diseases, including dengue hemorrhagic fever, malaria, Chikungunya virus, La Crosse encephalitis, St. Louis encephalitis, West Nile virus, and yellow fever.¹⁶ Additionally, scratching the bites may lead to superinfection and scarring.¹

Prevention and Treatment

Patients with known mosquito sensitivity should avoid areas of stagnant water and utilize preventative measures such as wearing protective clothing and using mosquito repellent containing DEET (N,N-diethyl-meta-toluamide), IR3535 (ethyl butylacetylaminopropionate), picaridin, or 2-undecanone (methyl nonyl ketone or IBI-246) when outdoors. Essential oils such as lemon, eucalyptus, citronella, and garlic are somewhat effective.¹ Additionally, prophylactic dosing of antihistamines may prevent milder reactions.

Although often supportive, treatment and management of mosquito bites depends on the extent of the reaction. For common local reactions, symptomatic management with topical anesthetics, calamine lotion, or corticosteroid creams is appropriate. If superinfection from scratching is a concern, antibiotics may be appropriate.

Management of more severe and systemic reactions such as HMB also is supportive, and the addition of oral corticosteroids to decrease inflammation is required.⁷ Severe HMB also has been treated with immunosuppressive and anticancer drugs, though the efficacy is limited. Venom immunotherapy is a preventative option for patients with mosquito-specific IgE antibodies, and hematopoietic stem cell transplant may be required in patients with HMB.^14,16

Conclusion

Mosquito allergens can cause a variety of reactions, ranging from those limited to the skin to those characterized by severe systemic effects. Although common local reactions can be symptomatically treated with topical medication, more severe reactions such as HMB require more involved clinical management. Hypersensitivity to mosquito bites is an important condition to recognize, as it is related to multiple organ impairment as well as later development of malignancy. Patients should be closely monitored during the entire clinical course and in the years following.

Incidence and Characteristics

Mosquitoes are insects categorized into the order of Diptera and family of Culicidae, and more than 3500 different species have been identified.¹ In the United States, the most common genus of mosquitoes is Aedes, with other common genera including Culex, Anopheles, Culiseta, and Coquillettidia. Most bites are performed by female rather than male mosquitoes, as it serves to complete their life cycle (Figure 1).¹

There are a variety of possible reactions to mosquito bites. Severe local reactions that are large (papules >30 mm in diameter) or are accompanied by systemic manifestations are referred to as hypersensitivity to mosquito bites (HMB).² These hypersensitivity reactions vary according to multiple factors, including comorbid conditions, genetic predisposition, and geographic location. The majority of the world’s population will exhibit local reactions to mosquito bites at some point during life, with the median age of onset of the first bite at 2 years of age.³ In a study by Arias-Cruz et al,⁴ the incidence of patient-reported large local reactions was 2.5%. Hypersensitivity to mosquito bites, perhaps the most rare reaction, is more common among Asian and Central American children.⁵ The median age of diagnosis for HMB is 7 years, and most reactions occur during the first 2 decades of life.^6,7

Clinical Presentation

Mosquitoes bite vertebrates in an attempt to feed and thus must locate the host’s blood vessels through a process known as probing, which often necessitates changing the bite site several times. Once the vessel is located and lacerated, the mosquito feeds either from the vessel directly or the hematoma around it. Not only does the bite cause trauma to the skin, but a cutaneous reaction also may occur in response to salivary gland secretions that concurrently are deposited in the host tissue.⁸ Mosquitoes’ salivary gland components are the primary cause of cutaneous reactions, as one study showed that bites from mosquitoes lacking salivary gland ducts were not associated with these reactions.⁹ Mosquito saliva contains a large number of compounds with biologic activities, including lysozymes, antibacterial glucosidases, anticoagulants, antiplatelet aggregating factors, and vasodilators, as well as a potentially large number of unknown allergenic proteins. As of 2016, 70 mosquito-derived allergens have been identified, but this number continues to grow.² After a bite from a mosquito, these compounds may result in host sensitization over time, though interestingly, sensitization to mosquito bites from a species different from the original offender does not occur due to lack of cross-reactivity between species.¹ 

Because mosquitoes reproduce by laying their eggs directly on or near water, people who live near bodies of water or wetlands are at the highest risk for mosquito bites. Patient factors that have been found to lead to increased rates of mosquito bites include lower microbial diversity on the skin, the presence of sweat or body odor, pregnancy, increased body temperature, type O blood, dark clothing, and perfumes.² Exaggerated bite reactions are associated with Epstein-Barr virus (EBV) infection and hematologic malignancies.¹⁰ 

Immediate hypersensitivity is mediated by a specific IgE antibody and is characterized by erythema and a wheal at the bite site that peaks within minutes of the bite. In contrast, delayed hypersensitivity is lymphocyte mediated; occurs 24 hours after the bite; and causes an indurated, pruritic, and erythematous 2- to 10-mm papule that may blister.¹¹ Although the evidence of immediate hypersensitivity disappears within hours, symptoms of delayed hypersensitivity may last days to weeks. Accompanying symptoms may include local swelling, pain, and warmth. The itch that often is experienced in conjunction with erythema and papule formation is elicited in 3 main ways: direct induction utilizing classic pruritic pathways, IgE-mediated hypersensitivity reaction to salivary components, and IgE-independent host immune response to salivary antigens. Papular urticaria is a common additional finding in children with mosquito bites.¹ As an individual is repeatedly bitten, they may undergo 5 stages of sensitization: stage I (neither immediate nor delayed reaction), stage II (delayed reaction), stage III (immediate and delayed reaction), stage IV (immediate reaction), and stage V (neither immediate or delayed reaction).¹¹

Although most mosquito bites cause common local reactions, patients rarely demonstrate systemic reactions that can be much more severe. Skeeter syndrome is a milder systemic response characterized by large local reactions (papules >30 mm in diameter) developing hours after a bite with accompanying fever.¹² The reaction typically peaks over days to weeks.² Although the reaction may resemble cellulitis clinically, a history of a preceding mosquito bite can help make the distinction.¹³ 

A more severe systemic reaction is HMB, which is characterized by intense local skin findings as well as generalized systemic symptoms. Initially, indurated, clear, or hemorrhagic bullae appear at the bite site (Figure 2). Later, there is progression to swelling, necrosis, and ulceration.¹⁰ Biopsies from the skin lesions associated with HMB reveal necrosis, interstitial and perivascular eosinophilic and lymphocytic infiltrates, and small vessels with fibrinoid necrosis.⁷ Systemically, high fever, general malaise, liver dysfunction, proteinuria, hematuria, hepatosplenomegaly, and lymph node enlargement may occur. Patients typically experience these severe symptoms each time they are bitten.¹⁰

The mechanism of the HMB reaction is complex but has a close association with natural killer (NK) cell lymphoproliferative disorder and EBV infection (Figure 3). In fact, it is not uncommon for HMB patients to develop malignant lymphomas during their clinical course, even those unrelated to EBV.¹⁴ Epstein-Barr virus, one of the human herpesviruses, produces latent infection in NK cells. It is hypothesized that after a mosquito bite, EBV may be reactivated within these cells by induced expression of the viral lytic-cycle transactivator gene BamHI Z fragment leftward open reading frame 1, BZLF1.⁶ In response to mosquito salivary gland components, CD4⁺ T cells proliferate and induce expression of the EBV oncogene latent membrane protein 1, LMP1, on NK cells, which then infiltrate the bite site.¹⁵ These EBV-infected NK cells also overexpress the Fas ligand, thus contributing to organ and tissue damage.⁶ In addition to activating oncogene expression on NK cells, T cells also activate the basophils and mast cells carrying mosquito-specific IgE, both of which also add to the severe skin reaction of HMB.¹⁵ The particular triad of HMB, chronic active EBV infection, and NK cell lymphoproliferative disorder commonly is known as HMB-EBV-NK or HEN disease.¹ Patients with HMB should be monitored for malignancy. The mortality of HMB is increased in patients in whom onset occurs when they are older than 9 years and with BZLF1 messenger RNA in skin lesions.⁶ 

Other rare reactions to mosquito bites include Wells syndrome, anaphylaxis, and superficial lymphangitis. Wells syndrome (also known as eosinophilic cellulitis) is characterized by erythematous or violaceous plaques and pruritic blisters. Although its etiology has not been defined, it is thought to be evoked or exacerbated by insect bites, with CD4⁺ T cells playing a primary role.¹ Anaphylaxis (angioedema, urticaria, and wheezing) rarely may occur due to mosquito salivary gland components but typically is caused by other stinging insects. Superficial lymphangitis, often misdiagnosed as an infection of the lymphatic system, presents within minutes as nontender pink streaks originating from the bite site. A biopsy with eosinophil and mast cell infiltrates consistent with an allergic-type reaction confirms the absence of infection. Patients respond well to glucocorticoid treatment.

Mosquitoes are vectors for many blood-borne diseases, including dengue hemorrhagic fever, malaria, Chikungunya virus, La Crosse encephalitis, St. Louis encephalitis, West Nile virus, and yellow fever.¹⁶ Additionally, scratching the bites may lead to superinfection and scarring.¹

Prevention and Treatment

Patients with known mosquito sensitivity should avoid areas of stagnant water and utilize preventative measures such as wearing protective clothing and using mosquito repellent containing DEET (N,N-diethyl-meta-toluamide), IR3535 (ethyl butylacetylaminopropionate), picaridin, or 2-undecanone (methyl nonyl ketone or IBI-246) when outdoors. Essential oils such as lemon, eucalyptus, citronella, and garlic are somewhat effective.¹ Additionally, prophylactic dosing of antihistamines may prevent milder reactions.

Although often supportive, treatment and management of mosquito bites depends on the extent of the reaction. For common local reactions, symptomatic management with topical anesthetics, calamine lotion, or corticosteroid creams is appropriate. If superinfection from scratching is a concern, antibiotics may be appropriate.

Management of more severe and systemic reactions such as HMB also is supportive, and the addition of oral corticosteroids to decrease inflammation is required.⁷ Severe HMB also has been treated with immunosuppressive and anticancer drugs, though the efficacy is limited. Venom immunotherapy is a preventative option for patients with mosquito-specific IgE antibodies, and hematopoietic stem cell transplant may be required in patients with HMB.^14,16

Conclusion

Mosquito allergens can cause a variety of reactions, ranging from those limited to the skin to those characterized by severe systemic effects. Although common local reactions can be symptomatically treated with topical medication, more severe reactions such as HMB require more involved clinical management. Hypersensitivity to mosquito bites is an important condition to recognize, as it is related to multiple organ impairment as well as later development of malignancy. Patients should be closely monitored during the entire clinical course and in the years following.