From the Journals

Long-term metformin offsets diabetes in patients with higher glucose/HbA1c, history of gestational diabetes



Metformin is especially effective in diabetes prevention among persons with higher baseline fasting glucose, higher hemoglobin A1c (HbA1c), or a history of gestational diabetes, based on 15-year follow-up results from the Diabetes Prevention Program and the Diabetes Prevention Program Outcomes Study.

Beyond those subgroups, metformin remained effective regardless of how diabetes was diagnosed, with a risk reduction of up to 36%, according to a report published in Diabetes Care.

“These results should help to prioritize those groups at high risk of developing diabetes who will benefit most from being treated with metformin,” said the report’s writing committee, chaired by David M. Nathan, MD, professor of medicine at Harvard Medical School, Boston.

However, the link between higher HbA1c and better efficacy of metformin should be “considered carefully,” according to Dr. Nathan and his cocommittee members, who noted that the original criteria for the Diabetes Prevention Program study were based on glucose level, rather than HbA1c level.

Initial results of the Diabetes Prevention Program study indicated a particular benefit of metformin after an average of 2.8 years of follow-up in individuals with higher baseline fasting glucose levels and in women with a self-reported history of gestational diabetes (N Engl J Med. 2002 Feb 7;346[6]:393-403).

Those results prompted the American Diabetes Association and others to recommend consideration of metformin for diabetes prevention in individuals considered to be at high risk, Dr. Nathan and his colleagues noted. “This recommendation is further supported by the demonstrated cost savings of metformin in diabetes prevention.”

The Diabetes Prevention Program study originally included 3,234 high-risk participants enrolled between 1996 and 1999 and randomized to masked metformin, placebo, or intensive lifestyle intervention. The 15-year follow-up analysis considered the 2,155 individuals randomized to the metformin or placebo groups, of whom 1,861 (86%) chose to continue in the Diabetes Prevention Program Outcomes Study, in which they continued to receive metformin unmasked.

Over the 15 years, the incidence of diabetes development based on fasting or 2-hour glucose results was 17% lower in the metformin group, compared with the placebo group (hazard ratio, 0.83; or –1.25 cases per 100 person-years), according to Dr. Nathan and his colleagues.

For diabetes development based on HbA1c level, metformin was linked to a 36% reduction in risk, compared with placebo, or –1.67 cases per 100 person-years.

Higher baseline fasting plasma glucose (110-125 mg/dL vs. 95-109 mg/dL) was associated with a greater effect of metformin in reducing diabetes development over 15 years (P = .0004).

Metformin’s effect in reducing diabetes development was “nearly identical” in participants with an HbA1c of 6.0%-6.4%, compared with less than 6.0%, with HRs of 0.63 and 0.61, respectively; however, looking at rate differences, there was “substantial heterogeneity” between the groups, at –3.88 and –1.03 cases per 100 person-years, respectively, the investigators wrote.

For women with a history of gestational diabetes, metformin was linked to a 41% reduction in diabetes development, compared with placebo (P = .03); in women without gestational diabetes, the difference was a nonsignificant 6% reduction, compared with placebo. In terms of risk differences, this translated into a reduction of 4.57 cases per 100 person-years in women with gestational diabetes, compared with a reduction of just 0.38 cases in women without such a history. The authors noted that it is “complicated” to determine whether greater credence be given to the results based on glucose or those based on HbA1c.

On the one hand, the generalizability of the HbA1c analyses is adversely affected because the analyses were performed post hoc on a set of participants who had initially been selected for the study based on prediabetes defined by glucose. On the other, the HbA1c results may be more clinically relevant because “in many countries, oral glucose tolerance tests are not used routinely for the identification of persons at high risk for diabetes or with diabetes,” the authors wrote.

Continued follow-up of these patients will provide additional data on the potential long-term benefits of metformin use, such as the incidence of cardiovascular disease, cancer, and microvascular disease, the investigators concluded.

The National Institute of Diabetes and Digestive and Kidney Diseases and Department of Veterans Affairs, among others, funded the Diabetes Prevention Program and subsequent outcomes study. Lipha provided medication and LifeScan donated materials during the studies. Additional funding to the Diabetes Prevention Program was provided by Bristol-Myers Squibb and Parke-Davis.

SOURCE: Nathan DM et al. Diabetes Care. 2019 Mar 15. doi: 10.2337/dc18-1970.

Next Article: