Recurrence after healing in pyoderma gangrenosum: a prospective registry-based cohort study.
Pyoderma gangrenosum is a rare neutrophilic dermatosis with potential for relapse after healing, yet recurrence in real-world settings remains unknown. We aimed to evaluate recurrence patterns to inform clinical practice and future research.
Data were abstracted from a uniquely, prospectively maintained pyoderma gangrenosum registry. A cohort of patients with pyoderma gangrenosum seen between December 1st, 2019 and March 1st, 2026 who healed within this timeframe was identified. Two subgroups were then created: those who remained healed, and those who had a recurrence. Descriptive statistics summarised recurrence characteristics; Kaplan-Meier methods estimated a cumulative incidence of recurrence, and Cox proportional hazards modelling identified predictors of recurrence.
Of 116 patients with healed pyoderma gangrenosum, 41 (35%) experienced a recurrence, generating a recurrence proportion of 35·3% (95% CI 26·7-44·8). Nominal associations with recurrence on univariable analysis included history of diabetes mellitus (hazard ratio 2·1, 95% CI 1·1-4·0, p = 0·024) and ever-smoking (1·8, 1·0-3·4, p = 0·057). Baseline characteristics differing between recurrence and non-recurrence groups included age (p = 0·033), history of diabetes (p = 0·0060), and ulcer age at presentation (p = 0·046). Cumulative recurrence risk was calculated as 27% (95% CI 18·1-35·0) at 1 year and 41·3% (95% CI 30·4-50·5) at 5 years.
Pyoderma gangrenosum recurred in approximately a third of patients. A history of diabetes showed a borderline association with recurrence, while the univariable signal for smoking was attenuated after adjustment. Recurrence risk remains elevated five years after initial healing, and peristomal pyoderma gangrenosum is associated with higher recurrence rates, suggesting that maintenance therapy may be necessary for certain patient populations after achieving ulcer closure. Our data highlight the need for prospective evaluation of maintenance approaches. Future work will focus on individual recurrence predictors.
This is an unfunded study.
Data were abstracted from a uniquely, prospectively maintained pyoderma gangrenosum registry. A cohort of patients with pyoderma gangrenosum seen between December 1st, 2019 and March 1st, 2026 who healed within this timeframe was identified. Two subgroups were then created: those who remained healed, and those who had a recurrence. Descriptive statistics summarised recurrence characteristics; Kaplan-Meier methods estimated a cumulative incidence of recurrence, and Cox proportional hazards modelling identified predictors of recurrence.
Of 116 patients with healed pyoderma gangrenosum, 41 (35%) experienced a recurrence, generating a recurrence proportion of 35·3% (95% CI 26·7-44·8). Nominal associations with recurrence on univariable analysis included history of diabetes mellitus (hazard ratio 2·1, 95% CI 1·1-4·0, p = 0·024) and ever-smoking (1·8, 1·0-3·4, p = 0·057). Baseline characteristics differing between recurrence and non-recurrence groups included age (p = 0·033), history of diabetes (p = 0·0060), and ulcer age at presentation (p = 0·046). Cumulative recurrence risk was calculated as 27% (95% CI 18·1-35·0) at 1 year and 41·3% (95% CI 30·4-50·5) at 5 years.
Pyoderma gangrenosum recurred in approximately a third of patients. A history of diabetes showed a borderline association with recurrence, while the univariable signal for smoking was attenuated after adjustment. Recurrence risk remains elevated five years after initial healing, and peristomal pyoderma gangrenosum is associated with higher recurrence rates, suggesting that maintenance therapy may be necessary for certain patient populations after achieving ulcer closure. Our data highlight the need for prospective evaluation of maintenance approaches. Future work will focus on individual recurrence predictors.
This is an unfunded study.
Authors
Gillespie Gillespie, Roland-McGowan Roland-McGowan, Downey Downey, Latour Latour, Ortega-Loayza Ortega-Loayza
View on Pubmed