Argon laser panretinal photocoagulation for proliferative diabetic retinopathy was shown in the Diabetic Retinopathy Study and Early Treatment Diabetic Retinopathy Study to reduce the incidence of blindness by 50% with relatively small amounts of treatment. However, some diabetics require much more extensive photocoagulation for control of proliferative disease. We attempted to determine risk factors for poor response to treatment with panretinal photocoagulation (PRP) by studying outcome in relation to the argon laser burn count and the presence of diabetic vascular complications. Sixty-six consecutively treated eyes undergoing PRP were studied, of which 57% showed resolution of new vessels 6 weeks after treatment. This was significantly related to the total amount of laser treatment given (mean no. in regressed eyes 5800 burns, non-regressed 3510 burns; p < 0.05). Renal disease and age (< 50 years) were identified as risk factors for non-regression (p < 0.05); hypertension, neuropathy, duration of disease and insulin dependence had no significant effect on outcome. We conclude that regression of proliferative disease is significantly related to the cumulative total number of laser burns applied and that successful laser photocoagulation in patients with diabetic renal disease requires considerably more treatment than that suggested by earlier studies.