| |
Cardiovascular Risk and the Thiazolidinediones EDITORIAL
|
| |
| |
Deja Vu All Over Again?
Daniel H. Solomon, MD, MPH; Wolfgang C. Winkelmayer, MD, ScD
Division of Pharmacoepidemiology (Drs Solomon and Winkelmayer) and Rheumatology, Immunology, and Allergy (Dr Solomon), and Renal Division (Dr Winkelmayer), Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts.
JAMA. Sept 12, 2007;298:1216-1218.
In 2005, the US Food and Drug Administration (FDA) held an advisory committee meeting to help determine the safety of selective cyclooxygenase 2 (COX-2) inhibitors, a popular group of drugs with a novel mechanism of action but with incompletely understood effects on the cardiovascular system. Although these drugs have some potential benefits with respect to gastrointestinal toxic effects, their benefit-risk ratio was and is still unclear. Fast forward 2 years to 2007, and the FDA held a similar advisory committee meeting about the safety of rosiglitazone, a widely used thiazolidinedione (TZD) with known benefits on glycemic control but potential cardiovascular toxic effects. What have clinicians, patients, and the public learned through these recent events?
The TZDs sensitize end organs to insulin through their effect on the peroxisome proliferation-activated receptor g (PPAR-g). The PPAR system is a group of nuclear receptors (a, g, and d) that serve as transcription factors for genes important in glucose, lipid, and bone metabolism.1 The varied actions of the PPAR system fueled enthusiasm for the potential benefits of TZDs, even beyond their effects on hyperglycemia. However, early toxic effects observed with these agents, such as hepatic and heart failure,2 should have fueled equal levels of caution. The heart failure observed with rosiglitazone and pioglitazone prompted changes in the warning section of the package inserts but no "black box" warning until very recently.3
Approval of the TZDs was based on their ability to reduce blood glucose levels and glycated hemoglobin levels. Little information was available on their effects on the macrovascular complications of diabetes before these agents were approved. Since their marketing, few adequately powered randomized controlled trials have been conducted in moderate- or high-risk patients to definitively determine the true benefits of these agents on macrovascular complications. The only completed trial that was specifically designed and powered to evaluate the efficacy of a TZD in reducing hard cardiovascular outcomes was the Prospective Pioglitazone Clinical Trial in Macrovascular Events (PROactive), a placebo-controlled randomized trial in patients with evidence of existing macrovascular disease who otherwise received usual diabetes care.4 The study failed to show a significant benefit of pioglitazone treatment on the primary composite end point of cardiovascular, cerebrovascular, and peripheral vascular outcomes. However, pioglitazone reduced by 16% a secondary composite end point including death from any cause, nonfatal myocardial infarction, and stroke (P = .03). No similar studies have been completed for rosiglitazone, but an interim analysis of an ongoing randomized open-label trial comparing rosiglitazone with metformin-sulfonylurea found no significant difference in the primary end point of hospitalization or death due to cardiovascular cause. The effect estimates of this underpowered interim analysis point toward a possible increased risk in the rosiglitazone group.5 These studies found a significant increased risk of heart failure with pioglitazone and rosiglitazone vs their respective comparison group.
While information about the cardiovascular effects of these agents has gradually become available from a series of randomized controlled trials, it has required pooled analyses and meta-analyses to better characterize their risks. Nissen and Wolski conducted a meta-analysis of rosiglitazone trials to determine the associated cardiovascular risk.6 The comparison groups in the trials ranged from placebo to combination therapy with metformin and sulfonylurea. Published reports and trial registries were searched for cardiovascular end points, such as myocardial infarction and death from cardiovascular causes. After integrating the trial-level data from 42 studies with at least 24 weeks of follow-up, the authors found a 43% increased risk of myocardial infarction compared with the comparison treatments (P = .03).
This issue of JAMA contains 2 additional meta-analyses on the cardiovascular effect of the 2 TZDs. Singh and colleagues performed a meta-analysis of rosiglitazone trials and (unlike the previous meta-analysis6) included only studies of at least 12 months' duration that prospectively collected information on cardiovascular events.7 The findings of the current study also found a 42% increase in risk of myocardial infarction (P = .02). In addition, the authors observed a more-than-doubling of the risk of heart failure with rosiglitazone (hazard ratio, 2.18; P = .001), but no significant increase in the risk of cardiovascular mortality was detected.
Lincoff et al performed a pooled analysis of cardiovascular events using patient-level data from trials comparing pioglitazone with a range of alternative regimens.8 The manufacturer of pioglitazone provided the data from 19 randomized, double-blinded trials that were available, but the analyses were conducted independent of the drug manufacturer. Among patients randomized to receive pioglitazone, the rate of death, myocardial infarction, or stroke was reduced by 18% compared with controls (P = .005). Similar to the meta-analysis of rosiglitazone, the risk of heart failure was significantly increased in pioglitazone users (hazard ratio, 1.41; P = .002).
The 2 TZDs currently marketed, although both representing the same class of drugs, have strikingly different profiles in their effects on ischemic cardiovascular outcomes. Both agents reduce blood glucose levels and glycated hemoglobin levels to a similar degree and both appear to cause excess heart failure risk; however, their effects on cardiovascular ischemic events differ based on the currently available data. Since much of the morbidity and mortality associated with diabetes is due to macrovascular ischemic complications, even small increases in relative risks translate into major decrements in public health. Moreover, with many other available oral agents for diabetes, the potential benefit of TZDs requires reevaluation. This need motivated the recent FDA advisory committee meeting. The meeting was initially to include data regarding both pioglitazone and rosiglitazone; however, discussion focused on rosiglitazone and the votes did not include pioglitazone. The panel voted on 2 issues: 20 of 23 committee members voted that rosiglitazone was associated with an increased cardiovascular risk, but 22 of 23 voted that rosiglitazone could remain on the market with a black box warning regarding its cardiovascular risk.9
The previous episode with the selective COX-2 inhibitors and the current one with the TZDs are instructive for designing a better drug safety system. First, early safety concerns must prompt strong and clear regulatory action. There is little evidence that modifications in warning sections of prescribing instructions prompt the desired prescribing change.10 Other options include black box warnings, clinician and patient education programs, and requiring more timely safety studies. The effect of these alternatives actions on prescribing improvements should be rigorously tested before regulators assume that they are effective.
Second, postmarketing adverse events not frequently observed in premarketing studies should be expected when there is incomplete understanding of the mechanism of action of a drug. As soon as drugs are used in patient populations that are not as healthy as those enrolled in randomized controlled trials, adverse events will be more frequent and more varied. While greater premarketing testing is always an option, systematic and targeted postmarketing surveillance or randomized trials in high-risk patient groups should be strongly considered. Timely delivery of such studies can be required as a condition of approval and for continued marketing.
Third, after several drugs are available for a given indication, new drug approval should be based on improvement in clinical outcomes, not surrogate measures. The use of surrogate measures as proxies for clinical outcomes may yield timelier results at lower cost, but the fallacies of this approach have been well demonstrated.11 Even though TZDs may be a useful step up in therapy, allowing patients to control their blood glucose levels without use of insulin, this may be doing patients a disservice if the complications of diabetes are not reduced through better glycemic control. The FDA has changed guidance for conducting trials in other therapeutic areas, such as osteoporosis. Similar changes should be considered when several standard therapies are available. This tougher standard assures patients and prescribers that newer drugs will be efficacious with respect to the relevant clinical outcomes.
Fourth, the decisions for initial approval of a drug and subsequent continued marketing should be symmetric. When considered for initial approval, a drug must be shown to be efficacious and safe. The recent deliberations over rosiglitazone, in which there was almost unanimous agreement that the drug was associated with cardiovascular risk, resulted in a vote to allow for its continued marketing. It would be unlikely for the drug to be initially approved for marketing with such agreement about its cardiovascular risk. However, the recent FDA committee did not believe that this risk warranted removal of rosiglitazone. Although removal of a medication creates tremendous patient inconvenience, the public expects that FDA approval is a seal of safety.
Finally, and perhaps most difficult, safety and efficacy must be explicitly balanced when drugs are being considered for approval or for continued marketing. The recent FDA committee meeting on rosiglitazone considered efficacy and safety data and determined that the drug's benefits outweighed potential risks. However, it is not clear how the committee "calculated" the benefit-risk ratio. While there are rigorous methods for weighing benefits and risk, these analyses are not often considered in FDA advisory committee meetings. 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. 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.
|
|
| |
| |
|
|
|
