This paper reports an analysis of all cause and cause specific mortality in the ACCORD database, through December 10, 2007. All deaths were reviewed twice; first to classify the primary cause of death as either CV, cancer, nonCV/noncancer or unable to classify then the second review was to determine whether hypoglycemia was a proximate contributor to the death. Overall, mortality rates were higher in those who experienced at least one hypoglycemic event than in those with none. The highest mortality was in those with three or more hypoglycemic events. However, those in the intensive arm who experienced hypoglycemia were less likely to die than those in the standard arm with hypoglycemia. Although it has been very trendy to argue that hypoglycemia was the cause of the increased mortality in the intensive treatment arm, this analysis is not able to support that contention. The authors specifically state that hypoglycemia can only account for a small number of the deaths thus do not explain the difference between the groups. One is left to wonder if it the difference in mortality between groups was actually just variability that may have resolved over time if the trial had been allowed to continue, given that there was no clear cut cause of the difference in mortality. Symptomatic, severe hypoglycemia is a predictor of mortality but the question remains open as to whether the cause was underlying illness vs glucose lowering therapies, which this analysis is not able to tell us. One must keep in mind that the lipid management component of the ACCORD trial is still ongoing so more information on this subject may be forthcoming in the future. Kittie Wyne, MD, PhD, FACE