Although pre-operative retinopathy severity appears to be a major factor in determining the visual outcome of diabetic extracapsular cataract extraction, its precise relationship to post-operative visual acuity is ill defined. A meta-analysis was therefore carried out, and studies were included if pre-operative maculopathy and retinopathy status was sufficiently defined to permit discrimination of visual outcome between subgroups. Weighted mean proportions of eyes achieving a post-operative visual acuity > or = 6/12 were as follows: no retinopathy, 87%; non-proliferative retinopathy with no maculopathy, 80%; quiescent proliferative retinopathy with no maculopathy, 57%; non-proliferative retinopathy with maculopathy, 41%; quiescent proliferative retinopathy with maculopathy, 11%; active proliferative retinopathy, 0. Differences in visual outcome between groups were significant (chi 2 = 119.9, p < 0.0005), attributable mostly to the trend across groups (chi 2 for trend = 115.4, p < 0.0005). Logistic regression indicated that maculopathy was a more potent predictor of post-operative visual acuity < or = 6/12 (odds ratio 6.4, 95% CI 4.13-9.94, p < 0.0005) than quiescent proliferative retinopathy (odds ratio 3.33, 95% CI 2.04-5.42, p < 0.0005). The severity of retinopathy and maculopathy prior to cataract surgery in diabetics are the major determinants of post-operative visual acuity. Further study of the relationship between pre-operative retinopathy severity and the incidence of post-operative complications, progression of retinopathy and maculopathy is required to optimise the management of cataract in diabetes.