"This is not something we expected, and we have not had an opportunity to explain that," Dr. Mather said.
He went on to note that the intensive lifestyle intervention was associated with a lower prevalence of microvascular complications in women (21% vs. placebo; P = .03, and 22% vs. metformin; P = .02), but not in men. The DPPOS researchers also observed a significant difference in microvascular disease prevalence at year 15 in those who progressed to diabetes, compared with those who did not progress to diabetes (13% vs. about 10%, respectively). This equated to about a 28% lower prevalence in those who had not progressed to diabetes, an effect was present in all three treatment groups.
"This equates to a paradox. We have diabetes reduction and we have a reduced prevalence of complications in those who did not develop diabetes, compared to those who did, but we don’t have a treatment effect to prevent diabetes at this point," Dr. Mather said. "We interpret these findings in the context of the achieved levels of glycemia. Despite the rising prevalence of diabetes in our population, we started following them quite early, and even at year 15, all three treatment groups have very low mean hemoglobin A1cs, in the 6% range."
Dr. Nathan noted that even though the study is currently insufficiently powered to test the effects of interventions on CVD events, CVD risk factors improved in all subjects.
"This is a good news message, the fact that we had too few cardiac events to analyze and that CVD risk factors fell in the entire population speaks to the high level of care they were getting outside of the study."
At the same time, development of diabetes was accompanied by a worsening CVD risk factor profile and increased coronary artery calcification severity, compared with those who did not develop diabetes. The reduction in conventional CVD risk factors such as blood pressure and lipids, and in the novel CVD risk factors, such as C-reactive protein and other indices of inflammation, "was generally greater in the lifestyle group than the placebo group, and it was generally intermediate (somewhere between lifestyle and placebo) in the metformin group," Dr. Nathan explained. "Notably, it was not an increase in use of blood pressure– or lipid-lowering drugs that explained the differences among the treatment groups. In fact, during much of DPP and much of DPPOS, the lifestyle group was using fewer statins, using fewer blood pressure lowering medications. During DPPOS the development of diabetes and higher glucose levels was associated with higher levels of CVD risk factors."
Dr. Nathan added that there were no differences in the level of coronary calcification among the treatment groups. However, metformin reduced coronary artery calcification in men, which suggests that metformin has cardioprotective effects.
Dr. Jill P. Crandall, director of the diabetes clinical trials unit at the Albert Einstein College of Medicine, N.Y., discussed side effects and other consequences of long-term metformin therapy as well as the health impact implications of diabetes prevention. After an estimated 10,000 patient-years of follow-up with metformin use, there have been no reported cases of lactic acidosis or severe hypoglycemia requiring hospitalization. The prevalence of B12 deficiency was about 9% in original DPP participants randomized to the metformin group.
"Our data also show that metformin-associated vitamin B12 deficiency often occurs in the absence of anemia, leading us to recommend that B12 testing be considered for patients taking metformin," Dr. Crandall said.
The researchers also observed a "modest but surprisingly durable weight loss" among participants in the metformin group. "The most highly adherent patients in the metformin group lost or maintained about 5% of their initial body weight over 10 years of follow-up," she said. "To our knowledge, the DPP/DPPOS represents the longest pharmacologic weight-loss study ever conducted."
When considering the measured cost of the DPP/DPPOS interventions and the aggregate cost of medical care received by the participants outside of the study, metformin was actually cost saving, and both the intensive lifestyle intervention and metformin were cost effective. "The net cost of the lifestyle intervention was approximately $1,700 more than placebo over the 10 years of follow-up," Dr. Crandall said. "Since the nature of the costs associated with diabetes is related to its complications, additional cost benefits of the interventions may accrue over time if these complications are prevented or delayed."
The study was funded by the National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases.
The researchers stated that they had no relevant financial conflicts to disclose.