Learning from ORIGIN
Zachary T. Bloomgarden · Journal of Diabetes · 2012
At the American Diabetes Association Annual Meeting in Philadelphia, the much-awaited results of the ORIGIN (Outcome Reduction with an Initial Glargine INtervention) Study were reported.1 In the study, 12 537 patients, of whom 1456 had prediabetes and the remainder had diabetes, were randomized to standard care or to receive insulin glargine. By design, those recruited for the study had high cardiovascular disease (CVD) risk, with 59% having had a prior CVD event, and 80% having hypertension, 15% albuminuria, and 12% smoking cigarettes. The rationale for such a strategy was to ensure a sufficient number of events in the treated population to have statistical significance; indeed, the event rate for the composite primary outcome of myocardial infarction, stroke, or cardiovascular death was approximately 3%/year, whereas the event rate exceeded 5%/year when revascularization and hospitalization for congestive heart failure were included. There was no difference in event rates between those randomized to insulin, whose mean HbA1c was 6.1%, and those in standard care (mean HbA1c 6.4%) over the median follow-up of 6.2 years. Does this mean that treatment with insulin is of no benefit? Certainly, given the modest glycemic separation of the groups, along with the greater hypoglycemia rate of 17% vs 5%/year and weight gain of 1.6 compared with a loss of 0.5 kg following intensive versus conventional treatment, one may argue that the potential for improvement was limited. Furthermore, nearly two-thirds of participants had CVD.2 Both the Veteran’s Administration Diabetes Trial3 (VADT) and the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study4 suggest strongly that if there is benefit in intensive glycemic control it is seen in people who do not have underlying CVD. Interestingly, however, and certainly awaiting clarification in further analyses of the study, there was a suggestion that those not having CVD in the ORIGIN study actually had worse relative outcome with insulin treatment: there were 323 primary cardiovascular events among 2543 people without prior CVD randomized to insulin glargine compared with 287 events among 2609 people not having prior CVD in the control group (for yearly incidence rates of 2.21% and 1.89%, respectively, a 17% increase); among 3720 and 3666 people with prior CVD randomized to the insulin glargine and control groups, there were 718 and 726 events, respectively (for annual incidence rates of 3.46% and 3.57%, respectively), with a 0.05 statistical significance of the interaction of prior CVD by treatment assignment.1 However, given the very high incidence of CVD in patients not having a prior history of events, we may question whether their state was truly equivalent to “not having CVD”. Asymptomatic diabetic people have annual CVD event rates well below 1%5. Thus, all the ORIGIN study participants may have been treated “too late”. Will it ever be possible to study younger people who are just developing diabetes? The ORIGIN study has shown that insulin glargine can maintain virtual euglycemia over a sustained period, and we anticipate that combinations of agents not associated with weight gain or hypoglycemia will similarly lead to glycemic benefit with fewer adverse effects, perhaps then with long-term reductions in CVD. We must, however, recognize that the number of patients needed, and the duration of a decade or more required for such a study, would make this trial a truly immense undertaking.