Figure 1. The University Group Diabetes Program cardiovascular mortality results are shown in patients treated over an 8-year period with placebo, tolbutamide, or insulin. The authors calculated cumulative mortality rates per 100 population using life table methods, correcting for the length of time each patient was available for follow-up. Mean data are shown, and the SEM is reported to indicate the variability associated with the calculated mortality rate. The study was prematurely terminated because of steady divergence of the mortality curves in patients treated with the K^ATPantagonist tolbutamide compared with those receiving placebo or insulin. (Adapted with permission of the publisher. [1])