This is a retrospective study of 114 patients who received oral isotretinoin for acne vulgaris for a minimum of 6 weeks between January 1994 and March 1995. Relapse was defined as deterioration in acne sufficient to merit systemic therapy, either with antibiotics or with another course of isotretinoin. Patients were considered nonrelapsers after a minimum period of 1 year post-treatment with isotretinoin. Case notes were traced and the following data were retrieved: age, sex, duration of acne, site of acne, previous treatment with either antibiotics or isotretinoin, indications for treatment, total cumulative dose of oral isotretinoin, average daily dose of isotretinoin, response! relapse, time taken to relapse, and subsequent treatment. Those with either complete or partial (>80%) clearance were considered as responders because of the difficulty in categorizing them accurately due to the retrospective nature of the study. A telephone interview was conducted as far as possible with all patients who had an inadequate follow-up period to ascertain if they had relapsed, i,e, if they had been prescribed systemic therapy for further flares of acne, Analyses of laboratory abnormalities were performed only in patients who had serial alanine aminotransferase (ALT), total cholesterol, and triglyceride readings, Data were analyzed using the chi-squared test, Student's t-test, Mann-Whitney U-test, and Fisher's exact test. Of the 114 cases studied, 21 were excluded because of the following reasons: no weight was documented, 12; antibiotics were started immediately after completion of treatment making it impossible to document relapse, 2; patients were not seen at all after completing treatment and were uncontactable, 7, A further four cases with a break of 4 weeks or more during treatment were excluded as the accuracy of the calculated total cumulative dose of isotretinoin may be affected, given the half-life of 10-20 h.