To clarify which parameter, diurnal pattern of blood pressure (BP) or level of BP variability, has the stronger predictive value for fatal and nonfatal vascular events, vital status after a mean (+/-SD) follow-up period of 86+/-46 months was determined in 392 type 2 diabetic subjects without any history of vascular disease, in whom the 24-h BP profile had been monitored between 1988 and 1998. After the exclusion of 28 subjects who died during the follow-up period of causes unrelated to diabetes, 364 subjects were recruited for further analysis. A total of 147 first events, including 50 fatal vascular events and 97 nonfatal vascular events, were recorded during the follow-up period. The rates of various vascular events increased with both reduced nocturnal falls in systolic BP (SBP) and levels of all ambulatory BP parameters. The ambulatory BP parameter showing the largest area under the receiver operating characteristic curve (ROCAUC) for fatal events was the mean 24-h pulse pressure (PP), and that for nonfatal events was the mean nighttime SBP; both exceeded the respective values of nocturnal fall in SBP. Furthermore, when dipper and nondipper diabetic subjects were divided into subgroups based on the 24-h PP (54.3 mmHg) and the nighttime SBP (116.5 mmHg) cut-off points derived from the ROC analyses, Kaplan-Meier plots showed that the diabetic subgroups with high ambulatory BP levels had worse outcomes, independent of dipper/nondipper status. Finally, these parameters were applied to the Cox model with the values of nocturnal fall in SBP and other confounding factors, and results showed that mean 24-h PP and mean nighttime SBP predicted fatal and nonfatal vascular events, respectively, more strongly than nocturnal fall in SBP in type 2 diabetic subjects. These findings therefore suggest that ambulatory BP levels in type 2 diabetic subjects have a higher predictive value for organ damage and death compared with diurnal BP patterns or dipper/nondipper status.