Pediatric Cutaneous Crohn Disease: Case Series of 89 Patients and Review

Abstract

Cutaneous (or “Metastatic”) Crohn disease (CCD) is a rare and likely underrecognized phenomenon characterized by granulomatous inflammation of the skin, discontinuous from the gastrointestinal tract. We describe the patient demographics, clinical characteristics, histologic findings, and treatment for 89 pediatric cases of CCD, including 78 previously reported in the literature and 11 new cases recently seen at four United States academic institutions. We emphasize the efficacy of biologic mono- and dual therapy for the treatment of this challenging condition.

INTRODUCTION

Crohn disease (CD) is a granulomatous inflammatory disease involving the gastrointestinal (GI) tract anywhere from mouth to anus.¹ Extraintestinal cutaneous disease is common, affecting up to 19% of patients with intestinal CD.² Cutaneous Crohn disease (CCD) is the least common dermatologic manifestation of CD; however, the prevalence may be higher than reported due to variable clinical presentations.^6–8 CCD is defined by cutaneous granulomatous inflammation discontinuous from the GI tract or in the total absence of intestinal disease. ^8–10

We summarize 89 pediatric patients (78 published cases of CCD and present 11 previously unreported patients) with CCD. We describe two patients utilizing dual biologic therapy. We provide updates on the epidemiology, clinical presentation, histopathology, and treatments options for pediatric CCD, highlighting the importance of biologic therapy.

METHODS

A search of PubMed identified 78 published cases of pediatric CCD, written or translated into English from 1970 until June 2022. Search terms included “metastatic Crohn disease”, “cutaneous Crohn disease”, “vulvar granulomatosis”, “vulvitis granulomatosa”, “anogenital granulomatosis”, “penile Crohn disease”, and “granulomatous lymphangitis”. Cases were included only if a diagnosis was made via skin or intestinal biopsy confirmed with characteristic skin findings by dermatology. Histology-confirmed pediatric CCD cases were added from a multicenter retrospective registry housed at the University of Wisconsin and approved by our IRB. Patient age, sex, intestinal disease status, presenting location(s), cutaneous morphologies, histology, and treatments were collected.

Treatment outcomes were categorized as “clearance”, “partial clearance”, “no effect”, or “worsening”. Treatments are listed in chronological order. Some treatments occurred simultaneously. Given the inconsistency with which cutaneous symptom resolution was reported, treatment outcome was determined by patient status at last report in published cases or at date of last follow up with a dermatologist. Clearance was defined as complete resolution of cutaneous symptoms and clinically visible active lesions. Partial clearance was used if improvement but not complete resolution of cutaneous symptoms was specified. No effect was determined if there was no change in the cutaneous findings, and worsening was defined as increased pain, size, number of lesions or spread to other parts of the body.

RESULTS

Our cohort includes 89 pediatric CCD cases with 78 previously reported cases and 11 additional registry cases (Tables 1 and 2). Boys comprised 55% (49/89) of cases and girls 45% (40/89). The mean age of CCD diagnosis in boys was 10.4 years and 11.2 years in girls. The genitals were the most frequently affected location (75%, 67/89). Eighty-five percent (34/40) of girls had vulval involvement, and 67% (33/49) of boys had penile and/or scrotal involvement. Boys were more likely to have facial involvement with 33% (16/49) of boys and 10% (4/40) of girls presenting with facial disease.

Table 1.

Pediatric Cases of MCD (N=89)

Patient Characteristics	Total Cohort, N=89 (%)	Our Cohort, N=11 (%)
Age, years (s.d.)	10.7 (4.0)	14.4 (2.9)
Sex
Female	40 (44.9)	5 (45.5)
Male	49 (55.1)	6 (54.5)
Luminal disease
Yes	69 (77.5)	5 (45.5)
No	19 (21.3)	6 (54.5)
Not reported	1 (1.1)	0 (0)
Cutaneous location
Scalp	1 (1.1)	0 (0)
Face	3 (3.4)	2 (18.2)
Lip	17 (19.1)	4 (36.4)
Perioral	3 (3.4)	0 (0)
Thorax	6 (6.7)	1 (9.1)
Genitals	67 (75.3)	6 (54.5)
Vulva	34 (38.2)	4 (36.4)
Penis	31 (34.8)	2 (18.2)
Scrotum	22 (24.7)	0 (0)
Perianal	49 (55.1)	4 (36.4)
Buttocks	10 (11.2)	4 (36.4)
Lower extremities	8 (9.0)	0 (0)
Cutaneous characteristics
Edema	73 (82.0)	7 (63.6)
Erythema	45 (50.6)	1 (9.1)
Skin tags	29 (32.6)	2 (18.2)
Induration	20 (22.5)	3 (27.3)
Fissuring	11 (12.4)	2 (18.2)
Ulceration	10 (11.2)	3 (27.3)
Abscess	8 (9.0)	1 (9.1)
Papules	6 (6.7)	1 (9.1)
Nodules	6 (6.7)	1 (9.1)
Fistulae	5 (5.6)	0 (0)
Vesicles	3 (3.4)	0 (0)
Erosions	3 (3.4)	1 (9.1)
Pustules	2 (2.2)	0 (0)
Alopecia	1 (1.1)	0 (0)
Histologic features
Noncaseating granulomas	75 (84.3)	11 (100)
Multinucleate giant cells	26 (29.2)	2 (18.2)
Perivasculitis	10 (11.2)	3 (27.3)
Eosinophils	4 (4.5)	1 (9.1)
Mixed inflammatory cell infiltrate	3 (3.4)	0 (0)
Vasculitis	2 (2.2)	0 (0)
Vascular ectasia	2 (2.2)	0 (0)
Caseating granulomas	1 (1.1)	0 (0)
Not reported	10 (11.2)	0 (0)

Table 2.

Systemic and Topical Treatment(s) for Pediatric cutaneous and/or luminal Crohn disease

Non-biologic systemic treatment(s)	Total Cohort, N=89 (%)	Our Cohort, N=11 (%)
Corticosteroids	53 (59.6)	9 (81.8)
Antibiotics
Metronidazole	29 (32.6)	3 (27.3)
Beta-lactam(s)	14 (15.7)	0 (0)
TMP-SMX	2 (2.3)	0 (0)
Fluoroquinolone(s)	3 (3.4)	0 (0)
Tetracycline(s)	1 (1.1)	0 (0)
Aminoglycoside(s)	2 (2.3)	0 (0)
Macrolide(s)	1 (1.1)	0 (0)
Unspecified antibiotic(s)	13 (14.6)	4 (36.4)
Immunomodulators
Azathioprine	16 (18.0)	0 (0)
6-mercaptopurine	8 (9.0)	0 (0)
Cyclosporine	1 (1.1)	0 (0)
Methotrexate	5 (5.6)	1 (9.1)
Cyclophosphamide	1 (1.1)	0 (0)
Hydroxychloroquine	1 (1.1)	1 (9.1)
Thalidomide	2 (2.3)	0 (0)
Antifungals	1 (1.1)	0 (0)
Anti-inflammatory agent(s)
Mesalazine, Sulfasalazine, Mesalamine, Salazopyrin	22 (24.7)	1 (9.1)
5-ASA, Tolmetin sodium	5 (5.6)	1 (9.1)
Dapsone	1 (1.1)	0 (0)
Colchicine	1 (1.1)	1 (9.1)
Valacyclovir	1 (1.1)	0 (0)
Isoniazid	1 (1.1)	0 (0)
Antihistamine(s)	4 (4.5)	0 (0)
Surgical resection	10 (11.2)	0 (0)
Biologic and small molecule treatment(s)
Infliximab	12 (13.5)	1 (9.1)
Adalimumab	8 (9.0)	6 (54.5)
Certolizumab	1 (1.1)	1 (9.1)
Ustekinumab	3 (3.4)	3 (27.3)
Vedolizumab	2 (2.3)	2 (18.2)
Ixekizumab	1 (1.1)	1 (9.1)
Tofacitinib	1 (1.1)	1 (9.1)
Topical treatment(s)
Topical steroids	25 (28.1)	2 (18.2)
Topical antibiotics (e.g., mupirocin, neomycin, metronidazole)	8 (9.0)	2 (18.2)
Topical antifungals (e.g., nystatin, azole(s))	6 (6.7)	2 (18.2)
Tacrolimus	5 (5.6)	2 (18.2)
Sitz baths	2 (2.3)	1 (9.1)
Pramoxine/Calamine	1 (1.1)	0 (0)
Diphenhydramine	1 (1.1)	0 (0)
Povidone-iodine	1 (1.1)	0 (0)
Not reported	60 (67.4)	9 (81.8)
Treatment outcome
Clearance	29 (32.6)	2 (18.2)
Partial clearance	36 (40.4)	5 (45.5)
No effect	7 (7.9)	2 (18.2)
Worsening	3 (3.4)	1 (9.1)
Not reported	14 (15.7)	1

In this pediatric cohort, edema (73/89, 82%), erythema (47/89, 53%), perianal skin tags (28/89, 31%), induration (19/89, 21%), fissures (13/89, 15%), ulcerations (8/89, 9%), lymphedema (8/89, 9%), abscesses (8/89, 9%), papules (7/89, 8%), nodules (6/89, 7%), fistulae (5/89, 6%), and erosions (3/89, 3%) were described (Table 1).

Seventy-eight percent (69/89) of patients were found to have intestinal CD although this observation is limited by varying lengths of follow up. In our cohort of 11 children, six (50%) had intestinal CD. Only one patient was diagnosed with intestinal CD after diagnosis of CCD within 12 months. One child was diagnosed with intestinal CD and CCD simultaneously. Three children had preexisting intestinal CD prior to CCD diagnosis, and age at CCD diagnosis varied from two to six years after intestinal CD diagnosis. The remaining five patients were followed for a range of one to eleven years after CCD diagnosis without an intestinal CD diagnosis.

Histology was not reported for ten cases (11%)^11–16. Of the remaining 79 cases, noncaseating granulomas were the most common feature (71/79, 90%). Caseating granulomas were found in 5% (4/79) of biopsies, and another single biopsy¹⁷ reported granulomas but did not specify subtype. Multinucleated giant cells were reported in a third of cases (26/79, 33%). Perivasculitis was reported but only in 13% of patients (10/79). Eosinophils were only reported in four (4/79, 5%) biopsies. Eight (9%) cases were diagnosed clinically in the setting of intestinal Crohn disease without histologic confirmation^11–16.

Of 89 cases, treatment outcomes were not reported in 14 published cases ^{15,16,18–25}. In the remaining 75 cases, 39% (29/75) had total clearance^{11,12,14,17,19,24,26–40} and 48% (36/75) with partial clearance^{10,14,19,38,41–62} of CCD. Seven (7/75, 9%) patients reported no effect of treatment^13,14,63,64, and three patients (3/75, 4%) had worsening cutaneous symptoms^65,66. Oral corticosteroids were the most used systemic medication with 71% (53/75) patients treated.^{12,14,19,20,23,26,28,32,35,39,41,42,46,47,49,52–54,59,61–63} Oral metronidazole 39% (29/75) was commonly used. ^{11,13,14,17–19,32–35,38,41,42,44–46,48,50,52,53,56,62,66} Immunosuppressive medications included azathioprine (16/75, 21%)^{13,17,19,33,34,41,43,50,54,55,58,64,65,67}, 6-mercaptopurine (8/75, 11%)^{12,14,19,29,51}, methotrexate (5/75, 7%)^19,31,35,52, cyclosporine (1/75, 1%)⁴³, and cyclophosphamide (1/75, 1%).¹⁴ For patients with intestinal disease, 39% (27/69) were treated with 5-aminosalicylic acid (5-ASA) drugs.^{14,17,19,23,25,28–30,33,37,38,41,42,45,50,55,56,61,62,64–66}

Treatment and treatment outcomes are in Table 2. Biologic agents were used in 21% (19/89) of cases, but the effects of treatment were not reported in two patients.^{11,12,19,20,33,39,41,43,47,60,64} Of the remaining 17, 100% (17/17) were treated with TNF-alpha inhibitors.^{11,12,19,33,39,41,43,47,60,64} Infliximab was most frequently used (13/17, 76%).^{11,12,19,33,41,43,47,64} Of these patients, 94% (16/17) had partial or total clearance with TNF blockade.^{11,12,19,33,41,43,47} Only one patient (1/17, 6%) was treated with infliximab with no effect reported.⁶⁴ Other biologic or small molecule inhibitors were used infrequently.

Eighty-seven percent of patients (65/75) had partial or total clearance of their CCD. Of these, 25% (16/65) achieved results with use of TNF blockade.^{11,12,19,33,39,41,43,47,60} Ustekinumab resulted in clearance of cutaneous disease in two patients (2/3, 67%) and partial clearance in one patient (1/3, 33%). None of the three patients (3/75, 4%) who had worsening cutaneous symptoms were treated with biologic or small molecule therapy.^65,66

In our cohort, there were two cases where intestinal and cutaneous disease remission was achieved with dual biologic therapy (defined as the use of two biologic therapies with differing mechanisms of action or the use of a biologic therapy and small molecule inhibitor).^68,69

One was a 13-year-old male with a warm, indurated, pink plaque of the left cheek (Figure 1). Biopsy revealed granulomatous inflammation consistent with CCD. Endoscopic biopsies confirmed diagnosis of intestinal Crohn disease. He was treated with oral corticosteroids, topical therapies, and anti-TNF therapy with worsening of cutaneous and intestinal symptoms. Ustekinumab was initiated (260 mg intravenous (IV) followed by 90 mg subcutaneous (SQ) every 8 weeks), which resulted in complete resolution of his cutaneous disease. However, his GI symptoms worsened, and vedolizumab was started at 300 mg IV at weeks 0, 2, 6, then every 8 weeks for treatment of gastrointestinal disease. Combination vedolizumab and ustekinumab resulted in complete resolution of cutaneous symptoms and improvement in intestinal symptoms. He has remained on dual therapy for >2 years with stable intestinal disease and without CCD recurrence.

Figure 1.

Indurated pink plaque of left cheek of 13-year-old male with intestinal and cutaneous Crohn disease

A 12-year-old female with intestinal CD previously treated with infliximab and adalimumab at time of CCD diagnosis presented with a history of vulvar edema and tenderness with perianal erythema (Figure 2). Cutaneous biopsy demonstrated granulomatous dermatitis consistent with CCD. She was treated with oral steroids, topical therapies, and methotrexate without improvement. Ustekinumab (260 mg IV followed by 90 mg SQ every 8 weeks) failed to improve her symptoms. Dual biologic therapy with adalimumab (160 mg once then 80 mg two weeks later then 40 mg every two weeks) and ustekinumab resulted in improvement in both cutaneous and GI symptoms. However, her intestinal symptoms gradually worsened despite sustained cutaneous response. Dual biologic therapy with vedolizumab (300 mg IV at weeks 0, 2, 6, then every 8 weeks) and ustekinumab was trialed. On vedolizumab, she developed hepatitis prompting the transition back to ustekinumab and infliximab resulting in complete remission of gastrointestinal and cutaneous disease.

Figure 2.

Perianal and vulvar erythema and edema of 12-year-old female with intestinal and cutaneous Crohn disease

DISCUSSION

CCD affects individuals of all ages, but the true incidence of CCD is unknown given its variable presentation resulting in misdiagnosis and/or diagnostic delay.^2,9,70,71 Delays in diagnosis are common in children as the genitals are a frequent presenting location and can be a mimicker of sexual abuse or post-traumatic lymphedema.^{16, 24, 25,32, 51, 52,54,56}

The epidemiology of pediatric CCD cases varies compared to adults. We found that CCD is more common in boys, which contrasts with adults where the majority of CCD cases are female (64%).⁶ The average age of pediatric CCD diagnosis is 11 years whereas the age of diagnosis is not well established in adults. ^{6, 7,15,70}

The clinical presentation of CCD varies depending on affected site and is frequently divided into genital and non-genital disease. Non-genital CCD may present with solitary or multiple nodules, plaques, ulcers, fissures, lichenoid lesions, or violaceous perifollicular papules whereas genital disease typically presents with erythema, edema, induration, knife-cut ulcerations, and fissuring of the anogenital region.^6,7,72,73 Consistent with previous findings, the most frequently affected sites in our aggregated cohort were the genitals, lower extremities, and face (Table 1).^6,7,74,75

Common mimickers of CCD include hidradenitis suppurativa, intertrigo, eczematous dermatitis, pyoderma gangrenosum, erythema nodosum, and cellulitis, in addition to other infectious etiologies – including sexually transmitted infections.^6,7,76 A detailed history, physical examination, and biopsy may differentiate between these entities.

Variable clinical presentation can make this diagnosis challenging, particularly in the absence of known intestinal CD. CCD may precede or be diagnosed simultaneously with intestinal CD in 6% to 36% of patients.^6,75,77 Eliciting a history of GI symptoms may aid in diagnosis, and CCD is diagnosed simultaneously with intestinal CD roughly half the time in children.⁶ In a review of 40 pediatric patients, a diagnosis of intestinal CD between 9 months to 14 years after CCD diagnosis was reported.⁷⁶ Another study found that 24 of 28 pediatric CCD cases (86%) occurred in patients without intestinal CD, but all patients were subsequently diagnosed with intestinal CD within six years of CCD diagnosis.⁶⁵

Characteristic histopathology in CCD includes non-caseating granulomas composed of epithelioid histiocytes, multinucleate and Langhans giant cells. Lymphocytic infiltrate and plasma cells are commonly reported. Features such as granulomatous perivasculitis and an eosinophil-rich infiltrate have been less commonly described.^{7,9,53,72–75} To our knowledge, this is the first published cohort specifically examining the histopathology of CCD in pediatric cases. Biopsy with clinical-pathologic correlation supports the diagnosis, especially when cutaneous CCD is suspected without intestinal CD. However, there are no histologic criteria to diagnose CCD, and variable histologic findings have been reported. Of the 89 cases examined here, eight (9%) cases were diagnosed without histologic confirmation. While biopsy may aid in the diagnosis of pediatric CCD, given the preponderance for genital presentation in children, it may not be required. The risks and benefits of biopsy must be carefully weighed in pediatric patients.

Consensus for the treatment of CCD is lacking as no single therapy has shown consistent efficacy. Existing literature is based upon small case series and case reports including numerous therapies with varying success. A previously published algorithm proposed topical, intralesional, or systemic steroids as first-line therapy for lesions in all areas except for the face where topical calcineurin inhibitors were recommended. Second line therapies included metronidazole with consideration of an immunomodulatory medication (e.g., cyclosporine, methotrexate, azathioprine, 6-mercaptopurine) or consideration of a TNF blockade as indicated by severity of disease.⁷ However, this algorithm was primarily based on adult CCD patients rather than pediatric disease. Other reviews have included pediatric patients, but few have reported the effects of TNF-alpha inhibitors, and none have reported on the use of other biologic and small molecule therapies.^6,15,76 The prevalence of IL-23-positive dendritic cells and macrophages has been reported in adult CCD, which may explain the efficacy of ustekinumab in treating both adult and pediatric CCD.⁶

We report 89 cases of pediatric CCD, which is the largest review to date. We present data on the use of TNF blockade and IL-12/23 inhibitor therapies for the treatment of this disease. Despite the relatively low number of cases and heterogeneity of findings, we propose that TNF blockade is an effective treatment for pediatric CCD. We also propose that IL-12/23 inhibitors may be similarly effective and that the two biologics may work synergistically for improved cutaneous response as has been seen in refractory pediatric and adult GI disease.^68,69 Consideration of dual biologic therapy may be useful in pediatric patients requiring discordant therapies for their intestinal and cutaneous CD.