Auckland proliferative diabetic vitrectomy fellow eye study

Background: To review medical and ophthalmic findings of primary diabetic vitrectomy patients to examine indices important in progression to fellow eye surgery. Methods: A retrospective analysis was undertaken of all diabetic patients undergoing vitreoretinal surgery at Auckland Public Hospital between January 1992 and July 1996. Kaplan-Meier survival analysis was performed along with univariate and multivariate (Cox Proportional Hazards) data analysis. Results: One hundred and fourteen primary diabetic vitrectomy cases were reviewed with mean follow-up duration of 4 years. Thirty-eight per cent (n = 43) of the study group underwent fellow eye surgery at a mean time of 1.6 years after first eye surgery. Fourteen patients were already blind in the fellow eye at baseline, and five patients refused second eye surgery on intention to treat. Thus there were 62 (54%) patients with severe (surgical threshold) fellow eye disease diagnosed within the follow-up period. The presence of either tractional retinal detachment or combined rhegmatogenous/tractional retinal detachment but without vitreous haemorrhage in the presenting eye was, in this series, a risk factor for fellow eye surgery (OR 5,56; 95% Cl 1.96-15.8), Maori and Pacific Islander ethnicity was significantly associated with traction retinal detachment (OR 2.23; 95% Cl 1.05-4.7). At data analysis 57% (n = 60) of the study patients had died. The mean time to death was 4.3 years, with 84% of these patients having evidence of renal disease at the time of their first eye surgery. Good visual function in at least one eye was maintained in many patients. Conclusions: A substantial proportion of diabetic vitrectomy patients require fellow eye surgery. Absence of vitreous haemorrhage in the presenting eye (i.e. tractional or combined rhegmatogenous/tractional retinal detachments but without vitreous haemorrhage) was predictive of need for fellow eye surgery. The need for diabetic vitrectomy correlates with poor survival in this study population.