Metformin, which helps control blood sugar and can be used alone or in combination with insulin to treat type 2 diabetes, also lowered the risk of developing the disorder in the study group, just not as much as making healthy lifestyle changes did.
"The lifestyle intervention was more powerful in preventing or delaying diabetes development during the original three-year Diabetes Prevention Program and remains more powerful over the entire 15-year study," said professor David M. Nathan of George Washington University in Rockville, Maryland, a coauthor of the new paper.
"However, there are specific subgroups in which the lifestyle intervention had an even more powerful effect - specifically, those older than age 60," Nathan told Reuters Health by email. Metformin was relatively more effective in people younger than 60 and those who were more obese, he said.
The researchers followed up with the surviving participants of a diabetes prevention study between 1996 and 2001 that compared people randomly assigned to either an intensive lifestyle intervention or 850 milligrams of metformin twice daily and a group taking a placebo.
The participants were all overweight or obese and had elevated blood sugar levels, both factors that put them at very high risk for developing diabetes.
At the end of that study, the people following the lifestyle intervention, which included a low-fat, low-calorie diet and 15 minutes of moderate-intensity exercise daily, had a 58 percent lower risk of having developed diabetes compared to the placebo group. Those taking metformin were 31 percent less likely to have progressed to diabetes than those on placebo.
Since the lifestyle intervention had worked so well, all participants were offered a version of it for one year after the study concluded.
As of 2014, almost 90 percent of the original group, or 2,776 people, had been followed since the end of the first study, and based on their original group assignments, they were offered lifestyle reinforcement seminars twice yearly or continued to receive metformin.
After an average of 15 years, diabetes incidence was lower by 27 percent in the lifestyle intervention group and 18 percent in the metformin group compared to the placebo group.
In 2014, 55 percent of the lifestyle group, 56 percent of the metformin group and 62 percent of the placebo group had been diagnosed with diabetes, according to the results in The Lancet Diabetes and Endocrinology.
The researchers also tested for changes in the tiny blood vessels of the kidney and retina and for nerve damage over time, all of which are associated with diabetes. The presence of these "microvascular complications" did not differ between groups overall. But among women in particular, the lifestyle group was less likely to show this kind of damage.
People who did not develop diabetes were almost 30 percent less likely to have kidney, retina or nerve damage than those who did.
"The complications that we were studying represent the early manifestations of this microvascular disease and were generally asymptomatic," Nathan said. "It usually takes 10-20 years for these complications to become clinically serious."
It was surprising that although lifestyle changes did reduce the risk of diabetes, they did not always reduce the risk of microvascular complications, according to Dr. Anoop Misra, director of Diabetes and Metabolic Diseases for Diabetes Foundation India, who authored a commentary alongside the new results.
"Metformin could be especially useful in people who are unable to follow diet and exercise strictly, who are obese, have polycystic ovarian disease, or cannot walk or exercise due to physical infirmity," Misra told Reuters Health by email. "It is a low cost drug, and could be useful in underprivileged populations also."
As the American Diabetes Association recommends, Nathan said, lifestyle interventions should be the first choice for diabetes prevention, with metformin recommended for younger and more obese people.