SAN FRANCISCO – Glycemic control has remained stable for an average of 15 years after initial randomization for patients in the Diabetes Prevention Program.

No significant differences in early microvascular complications were seen among interventions on intention to treat analysis, although intensive lifestyle intervention was effective in reducing microvascular complications in women.

Those are two key findings from an analysis of long-term outcomes from the Diabetes Prevention Program Outcomes Study (DPPOS), which were presented at the annual scientific sessions of the American Diabetes Association.

Dr. William Knowler, Dr. Jill Crandall, and Dr. Kieren Mather discuss the results of the DPPOS trial.*

DPPOS is a follow-up to the Diabetes Prevention Program (DPP), a randomized clinical trial of intensive lifestyle and metformin, compared with placebo for the prevention of diabetes. Launched in 1996, the trial took place at 27 centers in the United States and was designed to determine whether the development of diabetes could be prevented or delayed in high-risk patients. A total of 3,234 patients were randomized to either an intensive lifestyle intervention to reduce their weight by 7% through a low-fat diet and increased physical activity, or standard advice on diet and exercise plus metformin, or standard advice on diet and exercise plus placebo.

The DPP ended 1 year ahead of schedule after an average of 3 years of study. The lifestyle intervention program, which had accomplished its goal of 7% weight loss at 1 year, and then maintained about a 5% weight loss over the remaining time of the study, resulted in a 58% reduction in diabetes. The metformin group reduced diabetes development by 31%.

"To put these reductions in perspective, approximately 11% per year of participants in the placebo group developed diabetes," Dr. David M. Nathan, chair of the DPP and DPPOS and professor of medicine at Harvard Medical School, Boston, said during a press briefing. "That was reduced to about 8%/year in the metformin group, and less than 5%/year in the lifestyle group."

The DPP results were "widely heralded and reported," Dr. Nathan continued, "but they only covered a very brief period: Three years in what is in fact a life-long disease or risk state." DPPOS was designed to examine "whether the diabetes development that we demonstrated was delayed over a brief period of time, would continue to be delayed or prevented over a longer period of time."

Of the original 3,234 DPP participants, 2,776 (86%) continued in DPPOS. Their average age at the first visit was 56 years and 68% were women. Their average age is now 67 years and 31% have developed diabetes since enrollment in DPP.

The annual diabetes incidence rates among the original DPP placebo and metformin groups have declined to approximately equal the rate in the lifestyle group, "which have remained remarkably stable, at approximately 5% per year over time," said Dr. William C. Knowler, chief of diabetes epidemiology and clinical research at the National Institute of Diabetes and Digestive and Kidney Diseases. While the reasons for this decline were not clear, one possibility is that an effective lifestyle intervention was offered to all study participants, he said.

Despite the convergence of the annual diabetes rates in long-term follow-up, the differences between groups obtained in the first 3 years of DPP "resulted in continued lower overall long-term incidence of diabetes over the 15 years of follow-up, during which the diabetes incidence rate overall was 18% lower in the metformin group and 27% lower in the original lifestyle group, than in the original placebo group."

Even though diabetes prevention or delay was substantial with lifestyle or metformin over 15 years, all three treatment groups maintained good glycemic control on average. HbA_1c averages over follow-up were 6.1%, 6.0%, and 6.0% in the original placebo, metformin, and lifestyle groups, respectively.

"While these differences are statistically significant, they are very small and of questionable clinical importance," Dr. Knowler said. "The good news is that all three groups maintained good glycemic control on average, perhaps as a result not only of our interventions, but [also as] a result of rapid diagnosis and intensive treatment of diabetes once it developed."

As for microvascular outcomes such as retinopathy and neuropathy, Dr. Kieren J. Mather, professor of medicine at Indiana University, Indianapolis, reported that the average prevalence of microvascular disease at year 15 following DPP randomization was 12.4% in the placebo group, 13% in the metformin group, and 11.3% in the lifestyle group. While these differences between groups did not reach statistical significance, the mean prevalence of microvascular outcomes was about 50% higher in men than in women.

"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 A_1cs, 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.

dbrunk@frontlinemedcom.com

* An earlier version of this story misspelled Dr. Mather's name.

Long-term findings from DPPOS unveiled