Abstract

Diabetes Retinopathy is an important microvascular complication of diabetes mellitus and estimation of the retinopathy prevalence from cross sectional data is very important. This study is to estimate the hazard rate of diabetic retinopathy when only cross sectional data measuring diabetes prevalence is available. A total of 459 type 2 diabetic patients referred to an Ophthalmology clinic from May to December 2008 were studied. The severity of diabetic retinopathy was graded as per Early Treatment Diabetic Retinopathy Study (ETDRS). Factors associated with occurrence of retinopathy were assessed by the Cox's proportional hazard model for current status data. Multivariate analysis showed that body mass index, gender, smoking status and family history of diabetes were more independent predictions of diabetic retinopathy, while method of treatment, history of hypertension and duration of diabetes were all significantly associated with the occurrence of diabetic retinopathy. Compared with oral treatment group, those on insulin were more likely to be associated with diabetic retinopathy. From the cross sectional data, it is concluded that diabetic retinopathy hazard rate are less in patients with hypertension and more in patients using insulin and those with long duration of diabetes.