Dominant risk factors for retinopathy at clinical diagnosis in patients with type II diabetes mellitus.

A study of 270 newly presenting, previously untreated, type II diabetic patents revealed that 38 patients (14%) had already developed diabetic retinopathy (DR). Among this group, 26 patients had lesions of background diabetic retinopathy and 12 patients already had maculopathy or preproliferative changes. The aim of this study was to determine the risk factors influencing susceptibility to retinopathy, and to provide an accurate predictive value for diabetic retinopathy from a detailed multiple regression analysis that involved 27 demographic variables and the metabolic and hormonal responses during a meal tolerance test (MTT) at presentation. Compared to the nonretinopaths, the retinopaths had higher fasting plasma glucose levels (FPG) (mean +/- SD) (13.9 +/- 3.1 versus 11.6 +/- 3.2 mmol/L, p < 0.001), lower body-mass index values (BMI) (26.1 +/- 3.8 versus 29.3 +/- 5.0 kg/m2, p < 0.001) and higher plasma urea concentrations (6.0 +/- 1.9 versus 5.3 +/- 1.2 mmol/L, p 0.05). In contrast, gender and levels of blood pressure and other lipid levels did not influence the prevalence of diabetic retinopathy. A multiple regression formula for the prediction of diabetic retinopathy was derived and then used to categorize patients into high-risk and low-risk groups. The retinopaths also had higher HbA1c (p < 0.001), higher plasma glucose are under curve (0-2 h, p < 0.001), lower plasma insulin area under curve (0-22 h, p < 0.001), lower C-peptide area under curve (0-2 h, p < 0.01). They were also leaner (p < 0.001) and older (p < 0.05). However, these variables did not feature significantly in the multiple regression formula. The retinopaths were found to have higher risk probability values (25.1 +/- 11.5 versus 13.1 +/- 10.4%, p < 0.001). In the high risk group, 81.6% of retinopaths were identified. In the low-risk group, 63.8% of nonretinopaths were found. The incidence of diabetic retinopathy in type II diabetic patients at clinical diagnosis was found to be highly related to the degree of hyperglycemia, body-mass index, and to a lesser extent, renal impairment.