BACKGROUND: Several indications for vitreous surgery for complications of diabetic retinopathy have been established, but there is little well-founded information concerning situations in which visual prognosis is poor and vitreous surgery should not be performed. MATERIALS AND METHODS: The charts of 389 patients who had undergone vitreous surgery for complications of diabetic retinopathy between 1990 and 1994 were retrospectively reviewed. The minimum follow-up was 6 months with a median of 26 months. Using multivariate logistic regression analysis we studied factors which were correlated with a postoperative visual acuity of less than 5/200. A model was developed predicting the probability of an unfavorable visual outcome in various situations. RESULTS: Forty-five eyes (12%) had a best postoperative visual acuity of less than 5/200. Risk factors were detachment of the macula, extent of the detachment, iris neovascularisations and the duration of visual loss. For reoperations similar risk factors were found. The chance that an eye with total tractional retinal detachment of more than 6 months duration and with rubeosis of the iris will achieve a postoperative visual acuity of 5/200 or better is only 2%. CONCLUSIONS: Eyes with total tractional retinal detachment, especially with longer duration of the detachment and rubeosis, have a very poor chance of achieving useful vision and should not be operated.
BACKGROUND: Several indications for vitreous surgery for complications of diabetic retinopathy have been established, but there is little well-founded information concerning situations in which visual prognosis is poor and vitreous surgery should not be performed. MATERIALS AND METHODS: The charts of 389 patients who had undergone vitreous surgery for complications of diabetic retinopathy between 1990 and 1994 were retrospectively reviewed. The minimum follow-up was 6 months with a median of 26 months. Using multivariate logistic regression analysis we studied factors which were correlated with a postoperative visual acuity of less than 5/200. A model was developed predicting the probability of an unfavorable visual outcome in various situations. RESULTS: Forty-five eyes (12%) had a best postoperative visual acuity of less than 5/200. Risk factors were detachment of the macula, extent of the detachment, iris neovascularisations and the duration of visual loss. For reoperations similar risk factors were found. The chance that an eye with total tractional retinal detachment of more than 6 months duration and with rubeosis of the iris will achieve a postoperative visual acuity of 5/200 or better is only 2%. CONCLUSIONS: Eyes with total tractional retinal detachment, especially with longer duration of the detachment and rubeosis, have a very poor chance of achieving useful vision and should not be operated.