Treatment of pyoderma gangrenosum: retrospective multicentre analysis of 121 patients

Dear Editor, The therapeutic strategy for pyoderma gangrenosum (PG) includes the suppression of inflammatory disease activity, treatment of associated morbidities, promotion of wound healing and pain relief.1 However, treatment remains challenging owing to the lack of approved therapies and the difficulty in measuring response. There is only limited evidence available for therapeutic decision making. In a review of 350 published cases, Reichrath et al. concluded that systemic glucocorticosteroids (GCS) (methylprednisolone 0·5–1 mg kg−1 per day) and/or ciclosporin A (CSA) (5 mg kg−1 per day) are recommended as first‐line treatments to achieve remission.1 A recently published randomized controlled trial with 121 patients compared these two standard treatments. Both prednisolone and CSA had similar treatment efficacy in ulcer healing but adverse events occurred in up to two‐thirds of patients. Furthermore, healing rates of 47% within 6 months were not as high as expected, highlighting the need for more effective therapies.2 In another randomized controlled trial, the tumour necrosis factor (TNF)‐α antagonist infliximab led to a high remission rate (69%) in PG associated with chronic inflammatory bowel disease, but the observation period of 6 weeks was short.3 Case reports and small case series are available but are prone to bias.4,5,6,7,8,9 Therefore, further data comparing real‐world treatment options would be valuable. The aim of the present retrospective, multicentre cohort study was to analyse therapeutic approaches and the clinical response of patients with PG. This study was conducted in three German specialized dermatological wound clinics (University of Hamburg, University of Essen, Municipal Clinic Duisburg). Overall, 121 patients were diagnosed with PG and treated between 1999 and 2011 in one of the three centres (n = 50 in Essen, n = 30 in Duisburg, n = 41 in Hamburg). The mean age of patients was 64 years (SD 16·4) and 68% were women. The mean wound size was 67 cm² (SD 91) and 71·9% (87 of 121) were located on the lower leg. The median time to first dermatological consultation was 3 months [interquartile range (IQR) = 5]. The majority of patients received systemic immunosuppressive therapy (88·1%, 104 of 118), most commonly combined with topical drug therapy (66·1%, 78 of 118), while 11·9% (14 of 118) received topical treatment with GCS or tacrolimus only. For systemic therapy, GCS (85·9%, 104 of 121), CSA (24·7%, 30 of 121), azathioprine (AZA) (23·1%, 28 of 121) and mycophenolate mofetil (MMF) (13·2%, 16 of 121) were most commonly employed. Other drugs (e.g. TNF‐α antagonists) and intravenous immunoglobulins (IVIGs) were used less frequently (Fig. 1). Initial treatment was systemic GCS in 99% of patients (104 of 105) with a median initial dose of 80 mg (IQR 40). The mean duration of systemic GCS treatment lasted 4·3 months (SD 4·6). Initial response rates between systemic therapies were quite similar [systemic GCS 88·5% (92 of 104), CSA 83% (25 of 30), AZA 68% (19 of 28) and 81% (13 of 16) for MMF (Fisher's exact test P = 0·07) (Table 1 and Fig. 1)]. We observed a trend towards higher initial response rates in parenteral drug administration of systemic GCS compared with oral therapy [initial response rate in intravenous GCS 71% (15 of 21) vs. oral GCS 53% (44 of 83), Fisher's exact test P = 0·15]. Initial response to systemic treatment in pyoderma gangrenosum. N = 104 patients, 210 therapy trials; initial response: response within first 2–12 weeks; unknown: no judgement of clinical response available. Therapeutic outcome of different systemic immunosuppressants for treatment of pyoderma gangrenosum GCS, glucocorticosteroid; CSA, ciclosporin A; AZA, azathioprine; MMF, mycophenolate mofetil; MTX, methotrexate; CYC, cyclophosphamide; IVIG, intravenous immunoglobulin. Therapeutic outcome of different systemic immunosuppressants for treatment of pyoderma gangrenosum GCS, glucocorticosteroid; CSA, ciclosporin A; AZA, azathioprine; MMF, mycophenolate mofetil; MTX, methotrexate; CYC, cyclophosphamide; IVIG, intravenous immunoglobulin. Although a large proportion of patients initially responded to the first‐line treatment with systemic GCS, often healing was achieved only by using other immunosuppressive drugs (Table 1). The majority of patients in these cases received a combined therapy using systemic GCS and another immunosuppressant followed by a reduction of systemic GCS. On average, patients received two (± one) different systemic therapies. A total of 14·9% (18 of 121) of patients received biologics or IVIGs. In this group the mean number of different systemic treatments was 3·3 (SD 1·3, range 1–6). Overall, the PG healed in almost three‐quarters of the patients (72·7%, 88 of 121). Of these patients, 14% (12 of 88) healed after local therapy, 20% (18 of 88) after systemic therapy and 66% (58 of 88) after combined local and systemic therapy (Fisher's exact test P < 0·001). The mean healing time was 7·1 months (SD 8·7) after the initiation of systemic therapy. Female sex (P = 0·001) and smaller wound size (P = 0·004) were associated with a significantly better outcome. Side‐effects leading to discontinuation of treatment were most frequently documented for AZA (15%, eight of 28) (Table 1). These included elevated liver enzymes (n = 2), cytopenia (n = 1) or nausea (n = 1). CSA was discontinued in 13% (four of 30), three cases owing to renal dysfunction and one owing to abscess formation. In contrast, only 1% (one of 104) of patients expkerienced adverse events during treatment with systemic GCS (GCS‐induced diabetes, Cushing syndrome). MMF was well tolerated by all of the 16 patients treated. Cyclophosphamide led to pneumonia in one case. AZA therapy was discontinued owing to side‐effects more frequently than CSA or MMF (Fisher's exact test P = 0·04). In summary, this large case series of PG confirms some previous observations and adds some new insights regarding current treatment concepts in relation to this rare disease. Successful treatment of PG usually requires systemic and, moreover, combined immunosuppression. Classical immunosuppressants and steroid‐sparing agents were effective and tolerated in the majority of patients in our cohort and healing occurred in three of four patients. This study had some limitations. The initial response rates in our cohort were based only on the assessment of physicians provided in the patient records, which might, in part, explain the differences from previous observations;2 however, long‐term healing rates were comparable. Steroidal side‐effects were likely underreported. Finally, newer therapies (e.g. TNF‐α antagonists) were underrepresented owing to the 12‐year observation period. In our cohort, these newer therapies were mostly given to patients who did not respond to multiple standard therapies. Therefore, we cannot compare their effectiveness or tolerability with the standard therapy. In conclusion, systemic GCS treatment is associated with a high initial response rate but additional immunosuppressive therapy is required in most cases. Although effectiveness of commonly used steroid‐sparing agents (CSA, AZA, MMF) is comparable, CSA‐based regimens are more effective and MMF‐based therapies are better tolerated compared with AZA. Finally, overall healing rates were less than optimal, underscoring the need for more effective and less harmful immunotherapies. Future studies will have to elucidate the role of TNF‐α antagonists in patients with PG, but there have been promising signs from a short‐term observation randomized trial,3 clinical experience and recent small case series.6,10,11,12,13 Funding sources: none. Conflicts of interest: none declared.