ROSIGLITAZONE AND CARDIOVASCULAR OUTCOMES sometimes fatal,liver toxicity.Accordingly,it must be assumed that a variety of unexpected toxic ing Drug effects are possible when PPAR agonists are ad ministered t P Valu The questi whether the obs Comparator Drug Myocardial infarction zolidinediones must also be considered.Pioelita Metformin 1.140.70-186 0.59 zone is a related agent also widely used to trea Sulfonylure 1.2410.78-1.981 036 type 2 diabetes mellitus.However,unlike rosig. 2.780.58-133 020 litazone,pioglitazone has been studied in imep8gado 1800.95-339 0.07 Combined comparator drugs 1.43(1.03-1.98) 0.03 Death from cardiovascular causes Metformin 113103A3711 0g4 included coronary and peripheral vasculare 420.60-33到 0.43 showed a trend toward benefit from pioglita 5.37(0.51-56.52 zone (hazard ratio,0.90:P=0.095).A secondary Placebo 1.22(0.64-2.34】 0.55 end point consisting of myocardi stroke,and de Combined com ator drugs 1.640.98-2.741 0.06 th from cause sh ed tio 084.p-00 deaths.Accordingly.the confidence intervals for larly triglycerides,than does rosiglitazone. the odds ratios for my cardial infaretion and These emerging findings raise an important death from cardiovascular causes are wide.re question about the appropriateness of the cur sulting in considerable incertainty about the cntreglaiortre the dev FL ls.Th oorigina sed alysis adequate for maries of events.The lack of source data did not allow the use of more statis- apy is the reduction of the complications of dia- tically powerful time-to-event analysis.A meta betes,not improvement in a laboratory measure analysis is always considered less convincing or glycemic contro Alth ns in blood than a large prospective trial I designed to as eia the ou n such a dedi the fect ns has r one Evaluated fo to be unpredictable.After the failure of mura Cardiac Outcomes and Regulation of Glvcaemia glitazar and the apparent increase in adverse in Diabetes (RECORD)trial may provide useful cardiovascular outcomes with rosiglitazone,the insights. use of blood glucose measurements as a surro Despite the limitation our data point to gate point regulatory approva must De th or co ons t We pooled the results of of trials that e clinical trials is not originally intended to explore cardiovascular not sufficient to enable a robust assessment of outcomes.Most trials did not centrally adjudicate cardiovascular risks.The manufacturer has all cardiov outcomes,and the definitions of infarction we e no available.Many ble to an exter s wer and term,re adverse cardiovascular events o FDA for syst atic ana access to study reports N ENGL J MED 356;24 WWW.NEJM.ORG JUNE 14,200 246 The New England Joumal of Medicine Rosiglitazone and Cardiovascular Outcomes n engl j med 356;24 www.nejm.org june 14, 2007 2469 sometimes fatal, liver toxicity. Accordingly, it must be assumed that a variety of unexpected toxic effects are possible when PPAR agonists are ad￾ministered to patients. The question as to whether the observed risks of rosiglitazone represent a “class effect” of thia￾zolidinediones must also be considered. Pioglita￾zone is a related agent also widely used to treat type 2 diabetes mellitus. However, unlike rosig￾litazone, pioglitazone has been studied in a prospective, randomized trial of cardiovascular outcomes, called Prospective Pioglitazone Clini￾cal Trial in Macrovascular Events (PROACTIVE).31 The primary end point, a broad composite that included coronary and peripheral vascular events, showed a trend toward benefit from pioglita￾zone (hazard ratio, 0.90; P=0.095). A secondary end point consisting of myocardial infarction, stroke, and death from any cause showed a sig￾nificant effect favoring pioglitazone (hazard ra￾tio, 0.84; P=0.027). Notably, pioglitazone appears to have more favorable effects on lipids, particu￾larly triglycerides, than does rosiglita­zone.32 These emerging findings raise an important question about the appropriateness of the cur￾rent regulatory pathways for the development of drugs to treat diabetes. The FDA considers dem￾onstration of a sustained reduction in blood glucose levels with an acceptable safety profile adequate for approval of antidiabetic agents. However, the ultimate value of antidiabetic ther￾apy is the reduction of the complications of dia￾betes, not improvement in a laboratory measure of glycemic control. Although reductions in blood glucose levels have been shown to reliably reduce microvascular complications of diabetes, the ef￾fect on macrovascular complications has proved to be unpredictable.33 After the failure of mura￾glitazar and the apparent increase in adverse cardiovascular outcomes with rosiglitazone, the use of blood glucose measurements as a surro￾gate end point in regulatory approval must be carefully reexamined. Our study has important limitations. We pooled the results of a group of trials that were not originally intended to explore cardiovascular outcomes. Most trials did not centrally adjudicate cardiovascular outcomes, and the definitions of myocardial infarction were not available. Many of these trials were small and short-term, re￾sulting in few adverse cardiovascular events or deaths. Accordingly, the confidence intervals for the odds ratios for myocardial infarction and death from cardiovascular causes are wide, re￾sulting in considerable uncertainty about the magnitude of the observed hazard. Furthermore, we did not have access to original source data for any of these trials. Thus, we based the analysis on available data from publicly disclosed sum￾maries of events. The lack of availability of source data did not allow the use of more statis￾tically powerful time-to-event analysis. A meta￾analysis is always considered less convincing than a large prospective trial designed to assess the outcome of interest. Although such a dedi￾cated trial has not been completed for rosiglita￾zone, the ongoing Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes (RECORD) trial may provide useful insights.34 Despite these limitations, our data point to the urgent need for comprehensive evaluations to clarify the cardiovascular risks of rosiglitazone. The manufacturer’s public disclosure of sum￾mary results for rosiglitazone clinical trials is not sufficient to enable a robust assessment of cardiovascular risks. The manufacturer has all the source data for completed clinical trials and should make these data available to an external academic coordinating center for systematic anal￾ysis. The FDA also has access to study reports Table 5. Risk of Myocardial Infarction and Death from Cardiovascular Causes for Patients Receiving Rosiglitazone versus Several Comparator Drugs. Comparator Drug Odds Ratio (95% CI) P Value Myocardial infarction Metformin 1.14 (0.70–1.86) 0.59 Sulfonylurea 1.24 (0.78–1.98) 0.36 Insulin 2.78 (0.58–13.3) 0.20 Placebo 1.80 (0.95–3.39) 0.07 Combined comparator drugs 1.43 (1.03–1.98) 0.03 Death from cardiovascular causes Metformin 1.13 (0.34–3.71) 0.84 Sulfonylurea 1.42 (0.60–3.33) 0.43 Insulin 5.37 (0.51–56.52) 0.16 Placebo 1.22 (0.64–2.34) 0.55 Combined comparator drugs 1.64 (0.98–2.74) 0.06 The New England Journal of Medicine Downloaded from nejm.org on February 5, 2016. For personal use only. No other uses without permission. Copyright © 2007 Massachusetts Medical Society. All rights reserved