Fresh Fuel for the Glitazone Controversy: New Pioglitazone and Rosiglitazone Meta-Analyses

Shelley Wood

September 12, 2007

September 12, 2007 (Chicago, IL) - The glitazone controversy has fresh fuel this week in the form of two new meta-analyses appearing in the September 12, 2007 issue of the Journal of the American Medical Association [ 1, 2]. One shines the spotlight on pioglitazone (Actos, Takeda Pharmaceuticals North America), suggesting that it may indeed have a better safety profile than rosiglitazone (Avandia, GlaxoSmithKline), at least in terms of ischemic events. The other is yet another meta-analysis of rosiglitazone, looking only at studies in diabetics, reaffirming the increased risk of MI seen in other recent studies.

For the pioglitazone meta-analysis, Dr A Michael Lincoff (Cleveland Clinic, OH) and colleagues examined the incidence of MI, death, or stroke with pioglitazone across 19 randomized trials and report that the drug, while increasing rates of serious heart failure (HF), is associated with a lower risk of death, MI, or stroke than the control therapy. Time-to-event curves clearly separated after approximately one year of therapy, the authors note. Takeda provided patient-level data for the analysis, which may help it dodge some of the criticisms levied at the rosiglitazone meta-analysis by Nissen and Wolski--also coauthors on this new analysis--that first brought the stew over this class of drugs to a boiling point when it was published in the New England Journal of Medicine ( NEJM) last May [ 3].

"What this shows is that pioglitazone not only doesn't seem to have the adverse effects on cardiovascular complications that were seen with the other drug of this class, rosiglitazone, but in fact has a beneficial, protective effect in reducing the complications of death, heart attack, and stroke among diabetic patients," Lincoff told heart wire . "I think this should be sufficient to have an impact on clinical practice. I would encourage physicians to look at this data even in the absence of what may never happen--that is, there may never be a randomized trial comparing the two drugs, but I think physicians now have a lot of data on the two drugs that would help them make an informed decision, if one chooses to give a glitazone."

Pioglitazone safety

End point Pioglitazone (n=8554) (%) Control (n=7836) (%) Hazard ratio p
Death, MI, stroke 4.4 5.7 0.82 (0.72–0.94) 0.005
Death 2.4 2.9 0.92 (0.76–1.11) 0.38
MI 1.5 2.0 0.81 (0.64–1.02) 0.08
Stroke 1.2 1.7 0.80 (0.62–1.04) 0.09
Serious HF 2.3 1.8 1.41 (1.14–1.76) 0.002

Importantly, note the authors, magnitude and direction of the apparent protective effect of pioglitazone on risk of death, MI, or stroke was homogenous across trials of different duration, vascular-disease characteristics, and control or concomitant diabetes drugs.

"These findings suggest that the net clinical cardiovascular benefit with pioglitazone is favorable, with an important reduction in irreversible ischemic events that is not attenuated by the risk of more frequent heart-failure complications," they write.

While it remains unclear why pioglitazone and rosiglitazone would have different effects on ischemic events, Lincoff emphasized to heart wire that peroxisome proliferator-activated receptor (PPAR) agonists likely have multiple different effects on different metabolic pathways.

"Both rosiglitazone and pioglitazone have very equal control of glycemia, but clearly that isn't the only mechanism, there are obviously other mechanisms. We don't know most of the targets of these drugs. All of the complications of diabetes can't be captured in one surrogate end point, and that's the problem of any surrogate end point--they can only get you so far."

Commenting on the pioglitazone meta-analysis for heart wire , Dr Sanjay Kaul (Cedars-Sinai Medical Center, Los Angeles, CA) said the study had a number of strengths, including the fact that it used a robust analytical platform with patient-level data and time-to-event analysis and analyzed a homogeneous population that excluded nondiabetic patients. But Kaul also singled out several weaknesses in the analysis: cardiovascular end points were not assessed prospectively, blinded adjudication was possible in only two of the 19 trials, and the meta-analysis was not necessarily a comprehensive assessment of all available evidence. "It would be interesting to see how this data set compares with the sponsor's data set provided to the FDA," Kaul observed.

Kaul also pointed to some interesting nuances in the analysis. For example, the benefit of pioglitazone seen overall was not statistically significant when only active-controlled trials were considered; similarly, while the point estimates trended in favor of MI, none were statistically significant, "suggesting a degree of uncertainty," he noted. Moreover, the overall treatment benefit for pioglitazone seen in the meta-analysis was largely driven by the PROactive study, with 80% of all events in the analysis occurring in this one trial.

"It is important to remember that the primary end point failed to reach statistical significance in PROactive; as such, one cannot rule out a play of chance in driving the treatment benefit with respect to the triple end point of death/MI/stroke," he told heart wire . "Arguably, a lower p value threshold would be required than the reported p value of 0.027 for this principal secondary end point in PROactive."

Another not-so-rosy meta-analysis

In the second paper, Dr Sonal Singh (Wake Forest University School of Medicine, Winston-Salem, NC) and colleagues conducted their own meta-analysis of rosiglitazone, this time only including studies in diabetics and only if they were at least 12 months in duration: two oft-cited concerns with Nissen and Wolski's NEJM study. In the end, only four randomized trials were included in the analysis; still, like Nissen and Wolski's, Singh et al's study also found a significantly increased risk of MI and heart failure, but no significantly increased risk of cardiovascular mortality.

Rosiglitazone safety

End point Rosiglitazone (n=6421) (%) Control (n=7870) (%) Relative risk (95% CI) p
MI 1.46 1.05 1.42 (1.06–1.91) 0.02
HF 1.59 0.79 2.09 (1.52–2.88) <0.001
Cardiovascular mortality 0.92 0.91 0.90 (0.63–1.26) 0.53

Commenting on the study, Kaul called Singh et al's paper a "rigorously conducted meta-analysis" but expressed puzzlement over the "discordant effects" on MI and cardiovascular deaths. "If MI risk is increased by 42%, why does it not translate into CV mortality?" he asked. "Perhaps the number of events is small, or perhaps the results are due to a play of chance."

On top of the other reviews and meta-analyses that have shown an increase in ischemic events with rosiglitazone as well as an increase in heart failure, Singh et al argue that their study has clinical and regulatory implications. "These data suggest a reversal of the benefit-to-harm balance for rosiglitazone present at the time of approval," they write. "Currently, there appear to be much safer treatment alternatives. Regulatory agencies ought to reevaluate whether rosiglitazone should be allowed to remain on the market. Health plans and physicians should not wait for regulatory actions. They should avoid using rosiglitazone in patients with diabetes who are at risk of cardiovascular events, especially since safer treatment alternatives are available."

Choosing a drug for diabetes

In his comments to heart wire , Kaul pointed out that it is "only natural" for people to try to compare the pioglitazone and rosiglitazone analyses: "I would caution against such an exercise," he warned, given the substantial differences in the two databases.

"Only prospective trials designed for the specific purpose of evaluating pioglitazone against rosiglitazone will provide convincing data about comparability of these two thiazolidinediones," Kaul observed. " In my opinion, there appears to be reasonable equipoise to warrant a head-to-head comparison between pioglitazone and rosiglitazone to adjudicate the question; however, the $64&#160;000 question is, who will sponsor such a study? The maker of pioglitazone, which has perhaps not much to gain, or the maker of rosiglitazone, which arguably has the most to gain or even lose?"

Lincoff, for his part, called it a "matter of opinion" whether a head-to-head trial is warranted. "If one is anticipating that this type of drug would be appropriate, then I think there is now a lot of evidence to suggest that pioglitazone is the preferred drug."

In an accompanying editorial--like many other editorials and commentaries that have accompanied recent rosiglitazone analyses-- Drs Daniel H Solomon and Wolfgang C Winkelmayer (Brigham and Women's Hospital, Boston, MA) point out that the flurry of recent papers and panels should serve as a lesson to regulatory authorities, just as the last drug in this high-profile situation, rofecoxib (Vioxx, Merck), did [ 4]. As the panel at the recent rosiglitazone safety hearing demonstrated, panel members have no clear guidance or tools for making appropriate risk/benefit analyses, they note. "The FDA could develop or codify methods for such calculations and require them as part of new drug applications or for continued marketing of drugs," they argue. "The lessons regarding drug safety based on the events from recent years are numerous. Fixing the current system will be difficult, but the cost of not fixing the system and further diminishing the public's trust is too high a price to pay."

Lincoff's research department receives research support from Takeda, Sanofi-Aventis, Eli Lilly, Pfizer, Centocor, Dr Reddy's Laboratory, The Medicines Company, Roche, Daiichi-Sankyo, Schering-Plough, Scios, and AstraZeneca, but Lincoff personally keeps no honoraria or fees; Singh and Kaul disclose having no conflicts of interest.

  1. Lincoff AM, Wolski K, Nicholls SJ, Nissen SE. Pioglitazone and risk of cardiovascular events in patients with type 2 diabetes mellitus: A meta-analysis of randomized trials. JAMA 2007; 298:1180-1188.

  2. Singh S, Loke YK, Furberg CD. Long-term risk of cardiovascular events with rosiglitazone: a meta-analysis. JAMA 2007; 298:1189-1195.

  3. Nissen SE, Wolski K. Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes. N Engl J Med 2007; 356:2457-2471. Abstract

  4. Solomon DH, Winkelmayer WC. Cardiovascular risk and the thiazolidinediones: Déjà  vu all over again? JAMA 2007; 298:1216-1218.

The complete contents of Heart wire , a professional news service of WebMD, can be found at, a Web site for cardiovascular healthcare professionals.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.