Predictive model for immunotherapy of alopecia areata with diphencyprone

Background: Immunotherapy with diphencyprone (diphenylcyclopropenone) is us ed in the treatment of alopecia areata (AA). Response rates have varied in the literature. Objectives: To determine the efficacy of diphencyprone therapy for AA in th e largest reported cohort of patients to identify patient and treatment fac tors predictive of therapeutic success; and to develop a practical model fo r predicting patient response. Methods: The medical records of 148 consecutive patients treated with diphe ncyprone were reviewed. A clinically significant response to diphencyprone therapy was defined as a cosmetically acceptable response or greater than 7 5% terminal hair regrowth. Survival analyses using the Kaplan-Meier method and the Cox proportional hazards model were performed to determine signific ant factors predictive of regrowth and relapse. Results: Using a survival analysis model, the cumulative patient response a t 32 months was 77.9% (95% confidence interval, 56.8%-98.9%). Variables ind ependently associated with clinically significant regrowth were age at onse t of disease and baseline extent of AA. Older age at onset of AA portended a better prognosis, A cosmetically acceptable end point was obtained in 17. 4% of patients with alopecia totalis/universalis, 60.3% with 75% to 99% AA, 88.1% with 50% to 74% AA, and 100% with 25% to 49% AA. A lag of 3 months w as present between initiation of therapy and development of significant hai r regrowth in the first responders. Relapse after achieving significant reg rowth developed in 62.6% of patients. Conclusions: Response to diphencyprone treatment in AA is affected by basel ine extent of AA and age at disease onset. A prolonged treatment course mig ht be necessary. A predictive model has been developed to assist with patie nt prognostication and counseling.