It is well known that postchallenge and postprandial hyperglycemia are related to the progression of diabetic macroangiopathy (1–6). However, there is little information regarding the association between diabetic microangiopathy and postprandial hyperglycemia in human subjects. In this study, we performed a follow-up study to elucidate the relationship between diabetic retinopathy and postprandial glycemia or insulinemia. We recruited 151 Japanese patients with type 2 diabetes (74 men, aged 58.1 10.2 years, and 77 women, aged 57.9 9.2 years) who were admitted to Osaka Prefectural General Hospital between 1 January 1995 and 31 December 1999. The mean duration of diabetes of these patients was 7.4 6.7 years. The mean BMI and HbA1c (A1C) were 25.7 4.4 kg/m2 and 8.15 1.51%, respectively. Eighty-three patients were treated with diet alone and 65 with oral hypoglycemic agents. Two to 4 days before admission, patients were given an oral glucose load of 75 g and postchallenge plasma glucose and insulin levels were determined 2 h later. Within 2–4 days after admission, postprandial plasma glucose and insulin levels were determined 2 h after the intake of an isocaloric mixed breakfast (10 kcal/kg body wt; 57% carbohydrate, 15% fat, and 28% protein), representative of a standard Japanese breakfast. Within a week after admission, retinopathy was assessed through dilated pupils by ophthalmologists. One hundred twenty-one subjects showed no evidence of diabetic retinopathy, 23 simple diabetic retinopathy, and 7 preproliferative retinopathy or proliferative retinopathy. After discharge from the hospital, subjects were followed prospectively for 5.0 1.5 years.

During the follow-up periods, diabetic retinopathy worsened in 34 patients. Since A1C, postprandial plasma glucose and insulin, postchallenge plasma glucose and insulin, fasting plasma glucose and insulin, and duration of diabetes are closely linked, we performed multiple logistic model analyses, including sex, smoking, blood pressure, and serum lipid profile, to identify the independent and important predictors of the progression of diabetic retinopathy. Also, it is noted that in multiple regression analyses, the significance of several factors can be lost when there is a very close correlation among these several factors in