IBD-related pyoderma gangrenous in the biologic era: a single centre, 5-year case series and systematic literature review

Abstract

Background Pyoderma Gangrenosum (PG) occurs in 1–5% of patients with infl ammatory bowel disease (IBD) and can be challenging. However as it is uncommon, it remains poorly studied from a therapeutic perspective due to limited individual centre/practitioner experience. Few large recent series are reported, yet IBD treatment has changed substantially. Aims: We therefore reviewed local cases and performed a systematic literature review since the wider availability of anti-TNF therapy. Methods: Cases from our cohort are described. Subsequently, a search of PubMed from 2004-Feb 2012 using the terms ‘pyoderma gangrenosum’ and ‘inflammatory bowel disease’ was conducted. All articles in English relevant to the topic were reviewed. Cases which did not specify coexistent IBD or with paediatric age were excluded. Data regarding PG including site, size, duration, treatment, response and IBD disease activity were extracted. Results: We identified 5 cases of PG in 732 IBD patients (0.6%). Three were female, ages ranged from 22–56 yrs. Four had Crohn’s disease (CD) and one, ulcerative colitis (UC). Two had peristomal PG and 4 had PG on the lower limb (2 with both). Two with lower limb lesions had local trauma preceding PG. In 3 cases, IBD was active at PG onset; while in 2 IBD activity was uncertain. Three cases received Infliximab (Infl x), 2 with good response (one age 56 with venous insufficiency slow response). The 2 not receiving anti-TNF pre-dated PBS availability and healed slowly (over 1–3 years) with thiopurines ± steroids. Literature review identified 60 cases; 55% female, 50% UC, 45% CD, 5% IBD-U. At PG diagnosis, 58% had active and 15% inactive IBD, with 27% unspecified. Predominant sites were lower limb (48%) and peristomally (25%) with 42% having multiple lesions. In 12% trauma preceding PG was noted. In 42% new PG appeared during IBD-specific therapy, whilst 28% were not on any therapy and in 30% current IBD therapy was unspecified. Of patients on no therapy at PG onset (n = 17), 7 healed with whichever therapy was next used, 10 had a second attempt for healing; 9 responded well. Of the 9, 6 had Infl x, 1/6 non responding to Inlfx also failed anakinra. In total 35/60 patients received Infx, 4 Adalimumab (Ada), 2 had both; with 33 (94%) responding to one or the other. Only 2 cases failed new Infl x, but 3 PG cases arose on Infl x. These 3 healed with addition of topical Tacrolimus (n = 2) and change to Ada (n = 1). None was on Ada at PG onset and all 4 receiving it as second line therapy healed. In contrast to older reports there was no correlation of PG duration or size to healing times. Conclusion: PG appears predominantly during active IBD and is equally seen in CD and UC. With new PG, a reassessment of IBD activity is worthwhile to determine whether patients might benefi t from, and qualify for, an escalation to anti-TNF therapy, which appears highly effective.