The accumulation of sorbitol by the activated polyol pathway is considered to be a major cause of diabetic neuropathy. Because the erythrocytic sorbitol contents reportedly reflects that in nerves, erythrocytic sorbitol measurement would be useful for confirming the effect of an aldose reductase inhibitor (ARI). In this study, we examined erythrocytic sorbitol contents in healthy subjects and diabetic patients under fasting and postprandial conditions. Then, the contributions of blood aldose reductase (AR) contents and plasma glucose levels to the accumulated erythrocytic sorbitol contents were also analyzed. Erythrocytic sorbitol contents in the healthy subjects were 11.7 and 12.5-12.6 nmol/g Hb in fasting and postprandial status, respectively. In contrast, the erythrocytic sorbitol contents in diabetic patients were apparently higher (approximately 2.5-fold), but fidarestat treatment restored the elevated erythrocytic sorbitol contents to normal. In the diabetic patients, erythrocytic sorbitol contents were highly correlated with blood AR contents multiplied by the plasma glucose levels, whereas in the normal and fidarestat-treated diabetic patients no such correlation was observed. Taken together, these results suggest both the blood AR contents and the plasma glucose levels are factors determining erythrocytic sorbitol contents in diabetic patients. Notably, the potent ARI fidarestat was shown to normalize elevated erythrocytic sorbitol contents.