Home / Resources / Articles / Low-Dose Rosiglitazone/Metformin for Diabetes Prevention

Low-Dose Rosiglitazone/Metformin for Diabetes Prevention

Jun 9, 2010

With a month to go before rosiglitazone (Avandia) makes its reappearance in front of an FDA review panel, a surprising study in The Lancet, suggests that half-doses of rosiglitazone/metformin may help prevent progression to diabetes, with results showing a 66% relative risk reduction…

The paper, published online June 2, 1010, will no doubt raise some eyebrows, given the call to remove rosiglitazone from the market. Experts not involved with the study were quick to point out that the new study does nothing to alleviate concerns about the drug.


The paper was published within hours of news that the drug maker had settled the first US court case to go to trial, alleging that rosiglitazone causes MI.

Lead author on this new paper, Dr Bernard Zinman (University of Toronto, ON), stressed however, that the study was not powered to look at cardiovascular outcomes and that the main point of the research was that reduced doses of available drugs — not necessarily rosiglitazone — may help reduce progression to diabetes.

The 2 agents were chosen this because they were available in a single pill — Avandamet. And  that compliance with the study medication would be better if it were a fixed-dose. Pioglitazone/metformin could also have been used.

Zinman explained that, while lifestyle changes are “very effective” in preventing progression to full-blown diabetes in people with impaired glucose tolerance, they are “very, very difficult to implement, and on a society level, we’ve failed to do that.” Previous studies looking at pharmacological interventions to prevent diabetes have proven effective, but at the expense of major side effects.

Thus, the rationale for the Canadian Normoglycemia Outcomes Evaluation (CANOE) trial was to try for the same benefits, with lower drug strength, he said. “The notion here is to use low doses, in combination, because diabetes and prediabetes are complex diseases, and you need to target various aspects of the pathophysiology, and that’s exactly what we’ve done. We used a combination of half-dose metformin and half-dose rosiglitazone to improve insulin sensitivity, decrease hepatic glucose production, and have what we hope would be a robust effect on preventing diabetes. And we achieved that.”

In CANOE, 207 patients with impaired glucose tolerance were randomized equally to either placebo or a combination of rosiglitazone (2 mg) and metformin (500 mg) as a twice-daily dose over a median of almost four years. Doses of both drugs were half the maximum recommended dose, the authors note in the paper, and both patients and physicians were blinded to treatment type.

At follow-up, occurrence of incident diabetes (established by oral glucose tolerance or two fasting glucose values of >126mg/dL.(7.0 mmol/L or higher)) was significantly lower in patients assigned to the combination drug, 14%, than the placebo group, 39%, a relative risk reduction of 66% (p<0.0001) and an absolute risk reduction of 26%.

Overall, 70 patients (80%) in the combination-drug arm regressed to “normal glucose tolerance” by the end of the study, compared with 52 in the placebo group (53%; p=0.0002). Insulin sensitivity, likewise, remained unchanged in the drug-combination group but decreased in the placebo group; beta-cell function, however remained unchanged in both study arms.

In terms of side effects, incidence of diarrhea increased significantly in the drug group only, but there were no signs of swollen ankles, fluid retention, gastrointestinal symptoms, or hypoglycemia.

Zinman stressed that, “The adverse effects were really not there.” There was no weight gain, no edema or fluid retention, and the fracture rate was the same in both groups.

“This is a small study designed to show efficacy and not long-term safety, so there was no evidence on the effects on cardiovascular outcomes, but [the study] was way too small to look at that,” he said. “I think this means — and I’ve been telling family physicians this all along — that what we should be doing with respect to managing diabetes is using thiazolidinediones (TZDs) at half maximal dose. These drugs at full maximal dose have too much risk and at half maximal dose combined with other agents are very effective.”

There were no cases of MI or heart failure in the combination-drug group (and one of each in the placebo group), although Zinman shied away from addressing the cardiovascular effects, noting for heart failure, he’s “confident it’s not a big risk” based on the lack of fluid retention with the combination low-dose pill. For MIs, the study “was too small” to draw any conclusions.

But acknowledging the concerns about rosiglitazone, he reiterated that pioglitazone could be used in its stead. “I think the results apply to pioglitazone as well. But again, I would not use pioglitazone at full maximum dose; I would use it at half.”

He added that he currently prescribes pioglitazone at half-dosages, even to patients with established diabetes, and he also has some patients “doing well” on rosiglitazone, again at a half maximal dose.

One of the most outspoken critics of rosiglitazone, Dr. Steven Nissen, commented that, “This is a very small study — only about 200 patients.” Nissen also mentioned that the FDA does not recognize delay in diabetes onset as a legitimate label claim.

“Delaying onset of diabetes is not particularly useful if you increase adverse cardiovascular events,” he continued. “Despite the authors’ claims, there is no evidence whatsoever that lower doses of rosiglitazone are ‘safe.’ Any suggestion that serious cardiovascular toxicity can be avoided by using a small dose represents pure speculation at best.”

Zinman B, Harris SB, Neuman J, et al. Low-dose combination therapy with rosiglitazone and metformin to prevent type 2 diabetes mellitus (CANOE trial): A double-blind randomised controlled study. Lancet 2010; DOI: 10.1016/S0140-6736(10)60746-5. Available at: http://www.thelancet.com.

Buchanan TA and Ziang AH. Preventing type 2 diabetes with low-dose combinations. Lancet 2010; DOI: 10.1016/S0140-6736(10)60900-2. Available at: http://www.thelancet.com.