This Time It's A Draw for Rosiglitazone (Avandia) and Pioglitazone (Actos)

BURLINGTON, Mass., Sept. 28 -- Pooled data from seven randomized trials of thiazolidinediones for type 2 diabetes confirmed a significant risk of congestive heart failure with either pioglitazone (Actos) or rosiglitazone (Avandia), but neither increased the risk of cardiovascular death.

Compared with controls patients treated with either drug an a 72% increase in risk of congestive heart failure (RR 1.72, 95% CI 1.22-2.42, P=0.002), but the pooled risk for cardiovascular death was 0.91 (95% CI 0.63-1.32, P =0.063) with rosiglitazone and 1.01 (95% CI 0.51-2.01 P =0.98) with pioglitazone,

So found Rodrigo M. Lago, M.D., of the Lahey Clinic Medical Center here, and colleagues. They reported the data in the Sept. 29 issue of The Lancet.

"The pooled RR for development of congestive heart failure was 2.18 (95% CI 1.44 -3.32, P =0.0003) in the five trials of rosiglitazone, and 1.32 (1.04-1.68, P =0.02) in two studies with pioglitazone," they wrote.

The seven studies enrolled 20,191 patients who were followed for a mean of about 30 months. During that time 360 patients, including 214 who were given either rosiglitazone or pioglitazone, developed congestive heart failure. The congestive heart failure rate was 2.3% among patients treated with thiazolidinediones versus 1.4% for controls.

The estimated number-needed-to-harm for congestive heart failure was 107 across all seven studies but that number varied from 35 patients in one rosiglitazone trial to 491 to another rosiglitazone trial.

Absence of increased risk of cardiovascular mortality in the face of significant increase in the risk of congestive heart failure, suggests that thiazolidinedione-related fluid retention is more benign than other causes of heart failure, but the investigators said that hypothesis cannot be confirmed with a meta-analysis.

They noted that one trial initially found more heart failure and heart failure mortality for patients treated with pioglitazone, but subsequent analyses found that although more cases of heart failure were associated with pioglitazone than with controls, the number of primary and secondary events were similar in each group.

One interpretation of those data would be that pioglitazone-associated heart failure was indeed more benign that that caused by other factors, they wrote. But another, just as likely, interpretation was that "despite the potential for more adverse cardiovascular events associated with congestive heart failure, pioglitazone could have a cardioprotective effect compared with placebo."

The authors said their analysis was subject to all the limits of the meta-analysis methodology-reliance on aggregated data, varying definitions of heart failure, control groups that included both placebo and active treatments, and a lack of patient-specific outcome information.

Randomized trials of these two drugs is proving to be a mother-lode for data-mining researchers and this analyses is the latest in along line of meta-analyses, post-hoc analyses, and systemic reviews of the two drugs, most of which have been published in the four months since the New England Journal of Medicine published a meta-analysis of 42 trials by Cleveland Clinic investigators, which found a 43% increase in risk of myocardial infarction with rosiglitazone.

Earlier this month the Journal of the American Medical Association published a second rosiglitazone analysis that appeared to confirm the Cleveland Clinic paper along with a meta-analysis of pioglitazone studies that revealed an 18% reduction in cardiovascular mortality with pioglitazone.

A commentary and editorial in The Lancet offerred weary and wary advice about the interpreting the new paper by Dr. Lago and colleagues.

John G. F. Cleland, MD., and Stephen L. Atkin, M.D., of Castle Hill Hospital at the University of Hull in England, pointed out that although the analysis suggested that neither drug was associated with increased cardiovascular deaths "the confidence interval cannot exclude a 25% increase."

But the real problem-the elephant in the room-they wrote was that treatment should be "effective rather than merely innocuous." Both agents are most effective at improving glycemic control, but improved "glycemic control is not a surrogate for effective care of patients who have diabetes, which should be to reduce disability and increase lifespan."

The Lancet's editors pointed out shortcomings of meta-analyses. But they agreed that the reliance on surrogate markers-in this case hemoglobin A1C-"skirts the outcomes that matter most to patients-microvascular and macrovascular complications, quality of life, and survival."

The editors concluded with advice to the FDA and other regulatory agencies to demand better safety data or face the consequences, i.e. that "thiazolidinediones might simply become the latest in a series of preventable drug disasters."