Iranep.org

T h e n e w e ng l a n d j o u r na l o f m e dic i n e Rosiglitazone and Cardiovascular Risk
Bruce M. Psaty, M.D., Ph.D., and Curt D. Furberg, M.D., Ph.D.
In this issue of the Journal, Nissen and Wolski1 control groups. Across the trials, there was no report the results of a meta-analysis of treatment standard method for identifying or validating trials of rosiglitazone, as compared either with outcomes; events in eligible or ineligible trials other therapies for type 2 diabetes or with pla- may have been missed or misclassified. The to- cebo. Eligible studies included randomized trials tal number of events was relatively small, with that lasted for at least 24 weeks. The prespeci- the result that there was little or no power to fied primary end points of interest were myocar- detect potential differences among the trials if dial infarction and death from cardiovascular they were present. Although, in general, these causes. The authors identified 42 eligible studies, limitations are likely to move estimated odds ra- many of which were small or short-term trials, tios toward the null, the weaknesses, which are that included a total of 158 myocardial infarc- largely related to the quality of the available tions and 61 deaths from cardiovascular causes. data, are nonetheless substantial. A few events They used the Peto method to combine data from either way might have changed the findings for the trials. In this meta-analysis, rosiglitazone was myocardial infarction or for death from cardio- associated with a significant increase in the risk vascular causes. In this setting, the possibility of myocardial infarction (odds ratio, 1.43; 95% that the findings were due to chance cannot be confidence interval [CI], 1.03 to 1.98; P = 0.03) excluded. In their discussion, the authors prop- and a borderline-significant finding for death from erly emphasize the fragility of their findings.
cardiovascular causes (odds ratio, 1.64; 95% CI, Rosiglitazone, a thiazolidinedione, is an ago- nist of peroxisome-proliferator–activated recep- The meta-analysis has a number of strengths. tors (PPARs), primarily γ receptors, in the cell Among these were the effort to include unpub- nucleus.2 These ligand-activated nuclear transcrip- lished studies, the prespecified analysis plan, the tion factors activate the transcription of genes use of major cardiovascular events as the pri- that affect glucose and lipid metabolism.3 Rosig- mary outcome, and an analysis in which rosiglit- litazone increases hepatic and peripheral insu- azone was compared with placebo. In the latter lin sensitivity4 and reverses insulin resistance, a analysis, the odds ratio for myocardial infarc- prominent feature of type 2 diabetes.2 Approved tion was 1.80 (95% CI, 0.95 to 3.39; P = 0.07), in 1999 for the treatment of hyperglycemia in and the odds ratio for death from cardiovascu- type 2 diabetes, rosiglitazone has been shown in lar causes was 1.22 (0.64 to 2.34; P = 0.55).
small, short-term trials to reduce levels of fast- The study also has a number of weaknesses. ing glucose and glycated hemoglobin.2 At usual Only summary trial-level data (rather than patient- doses, the thiazolidinediones decrease glycated level data) were available, so it was not possible hemoglobin levels by an average of about 1 per- to conduct time-to-event analyses or to evaluate centage point or less, but they are also associat- the time course of risks. Data were not adequate ed with increases in body weight, adverse effects to conduct dose–response analyses. The eligible on lipids, fluid retention, and anemia.2 The prod- trials included both placebo and active-treatment uct label for rosiglitazone, which summarizes the Downloaded from www.nejm.org on May 21, 2007 . For personal use only. No other uses without permission. Copyright 2007 Massachusetts Medical Society. All rights reserved. T h e n e w e ng l a n d j o u r na l o f m e dic i n e results of randomized trials lasting 26 weeks, Physicians who chose to prescribe rosiglitazone lists many of these adverse effects in the section perhaps focused on the single dimension of gly- cemic control. The underlying assumption repre- The thiazolidinediones represent an interest- sents a kind of linear “physiological” argument: ing and potentially important class of drugs. The high levels of glycated hemoglobin increase risk, current epidemic of obesity in the United States so a reduction in glycated hemoglobin will auto- has spawned an epidemic of type 2 diabetes, with matically translate into improved health outcomes 1.5 million new cases per year.5 The complications for patients. This perspective ignores the many of diabetes, both microvascular and macrovas- actions of the genes activated by PPAR-γ agonists, cular disease, are directly related to levels of only some of which are currently known. Many fasting glucose and glycated hemoglobin. Even physicians did not require proof of health bene- in older adults, elevated levels of fasting glucose fits as a criterion for selecting rosiglitazone as a are directly and strongly associated with major therapy for type 2 diabetes.
cardiovascular events, and the attributable risk Had practicing physicians required this higher of an elevated glucose level is second only to el- standard, they would have been at a loss for evi- evated systolic blood pressure in this popula- dence from large, long-term trials. Rosiglitazone tion.6 In patients with type 1 diabetes, intensive was approved on the basis of short-term studies insulin treatment is associated with a reduced of the surrogate end point of glycemic control. risk of cardiovascular events.7 A treatment that The use of surrogate end points in the drug- simultaneously reduces insulin resistance, im- approval process has been problematic.8 Murag- proves glycemic control, and decreases the risk litazar, another PPAR agonist,9 and torcetrapib, of cardiovascular events would be a major thera- a cholesteryl ester transfer protein inhibitor that raises levels of high-density lipoprotein choles- On the basis of this meta-analysis, however, terol,10 are two recent examples. Indeed, at the the possibility of cardiovascular benefit associ- time of approval of rosiglitazone, the evidence ated with the use of rosiglitazone seems remote. from 26-week studies of expected health benefits We are not aware of data showing that rosiglit- was at best mixed. For a lifelong condition such azone prevents microvascular disease. In view as diabetes, how do the risks of weight gain, of the potential cardiovascular risks and in the edema, and adverse changes in lipids play out absence of evidence of other health advantages, against the benefits of improved glycemic con- except for laboratory measures of glycemic con- trol? For a drug that activates a large set of genes, trol, the rationale for prescribing rosiglitazone what is the overall balance of risks and benefits? at this time is unclear. Unless new data provide Rofecoxib, whose biologic actions early on sug- a different picture of the risk–benefit profile, gested the possibility of both gastrointestinal regulatory action by the Food and Drug Admin- benefit and cardiovascular harm,11 represented a istration (FDA) is now warranted. If patients us- similar regulatory failure to insist on large trials ing rosiglitazone are concerned about the find- of public health importance in a timely fashion.12 ings of this meta-analysis, they should discuss The current approach to drug approval in- them with their physicians and not unilaterally volves an intensive, high-quality evaluation in the stop taking the medication. Ongoing trials using preapproval setting. For many sponsors, approv- rosiglitazone may provide important new data, al marks the transition from research to mar- but for a drug approved in 1999, the delay in ob- keting.13 The FDA’s Adverse Event Reporting taining information about health outcomes has System is not capable of discerning the risk of events as common as coronary disease. The FDA During the market life of rosiglitazone, tens frequently requires phase 4 trials to address some of millions of prescriptions for the drug have of the unanswered efficacy or safety questions at been written for patients with type 2 diabetes. the time of approval. But sponsors propose the Insofar as the findings of Nissen and Wolski designs, which sometimes compare their prod- represent a valid estimate of the risk of cardio- ucts with inferior treatments or doses.14 During vascular events, rosiglitazone represents a major the period from 1998 through 2003, only about a failure of the drug-use and drug-approval pro- quarter of the required phase 4 trials were com- pleted,15 and as of September 30, 2006, a total Downloaded from www.nejm.org on May 21, 2007 . For personal use only. No other uses without permission. Copyright 2007 Massachusetts Medical Society. All rights reserved. of 899 phase 4 studies remain pending.16 This 3. Diamant M, Heine RJ. Thiazolidinediones in type 2 diabetes
desultory approach to postmarketing studies nec- mellitus: current clinical evidence. Drugs 2003;63:1373-405.
4. The DREAM (Diabetes Reduction Assessment with ramipril
essarily leads to an incomplete evaluation in the and rosiglitazone Medication) Trial Investigators. Effect of rosig- postapproval setting. If the FDA approves a drug litazone on the frequency of diabetes in patients with impaired on the basis of surrogate end points for the glucose tolerance or impaired fasting glucose: a randomised controlled trial. Lancet 2006;368:1096-105. [Erratum, Lancet long-term treatment of conditions such as dia- 2006;368:1770.] betes, large, long-term, randomized clinical trials, 5. Nathan DM. Thiazolidinediones for initial treatment of type
completed as soon as possible after approval, 2 diabetes? N Engl J Med 2006;355:2477-80.
6. Psaty BM, Furberg CD, Kuller LH, et al. Traditional risk fac-
are essential to identify the health benefits and tors and subclinical disease measures as predictors of first myo- risks associated with treatment. In the long run, cardial infarction in older adults: the Cardiovascular Health this approach is likely to be in the interests of Study. Arch Intern Med 1999;159:1339-47.
7. The Diabetes Control and Complications Trial/Epidemiology
sponsors, the FDA, and the health of the public.
of Diabetes Interventions and Complications (DCCT/EDIC) On May 10, 2007, the Senate passed the Food Study Research Group. Intensive diabetes treatment and cardio- and Drug Administration Revitalization Act.17 Al- vascular disease in patients with type 1 diabetes. N Engl J Med though the Senate bill has many strengths, in- 8. Psaty BM, Weiss NS, Furberg CD, et al. Surrogate end points,
cluding the allocation of new authority to the health outcomes, and the drug approval process for the treat- FDA, none of its provisions would necessarily ment of risk factors for cardiovascular disease. JAMA 1999;282: have identified the cardiovascular risks of rofe- 9. Nissen SE, Wolski K, Topol EJ. Effect of muraglitazar on
coxib or rosiglitazone in a timely fashion. One death and major adverse cardiovasculr events in patients with section of the bill (title II, subtitle A) focuses type 2 diabetes mellitus. JAMA 2005;294:2581-6.
10. Nissen SE, Tardif J-C, Nicholls SJ, et al. Effect of torcetrapib
largely on the mitigation of known risks at the on the progression of coronary atherosclerosis. N Engl J Med time of approval. In contrast, a true life-cycle 2007;356:1304-16.
approach, as advocated by the Institute of Medi- 11. McAdam BF, Catella-Lawson F, Mardini IA, Kapoor S, Law-
son JA, FitzGerald GA. Systemic biosynthesis of prostacyclin by cine,18 would continue the evaluation of both ef- cyclooxygenase (COX)-2: the human pharmacology of a selective ficacy and safety after approval, convert surrogate inhibitor of COX-2. Proc Natl Acad Sci U S A 1999;96:272-7. end points into clinically meaningful outcomes,19 [Erratum, Proc Natl Acad Sci U S A 1999;96:5890.] 12. Psaty BM, Furberg CD. COX-2 inhibitors — lessons in drug
integrate new information about health benefits safety. N Engl J Med 2005;352:1133-5.
and risks, and communicate those findings ef- 13. Steenburg C. The Food and Drug Administration’s use of
fectively to patients and physicians. The health postmarketing (Phase IV) study requirements: exception to the rule? Food Drug Law J 2006;61:295-384.
of the public would benefit from additional revi- 14. Psaty BM, Weiss NS, Furberg CD. Recent trials in hyperten-
sions to the drug-safety legislation as it moves sion: compelling science or commercial speech? JAMA 2006; through the House of Representatives.
295:1704-6.
15. FDA requested postmarketing studies in 73% of recent new
No potential conflict of interest relevant to this article was re- drug approvals. Impact report. Vol. 6. No. 4. Boston: Tufts Cen- ter for the Study of Drug Development, July/August 2004:1-4.
16. Food and Drug Administration. Report on the performance
From the Cardiovascular Health Research Unit, Departments of drug and biologics firms in conducting postmarketing com- of Medicine, Epidemiology, and Health Services, University of mitment studies: availability. Fed Regist 2007;72(22):5069-70.
Washington, and the Center for Health Studies, Group Health, 17. Food and Drug Administration Revitalization Act, S1082.
Seattle (B.M.P.); and the Division of Public Health Sciences, U.S. Senate bill. (Accessed May 15, 2007, at http://frwebgate. Wake Forest University, Winston-Salem, NC (C.D.F.).
access.gpo.gov/cgi-bin/getdoc.cgi?dbname=110_cong_ This article (10.1056/NEJMe078099) was published at www.
18. Baciu A, Stratton K, Burke SP, eds. The future of drug safety:
promoting and protecting the health of the public. Washington, 1. Nissen SE, Wolski K. Effect of rosiglitazone on the risk of DC: National Academies Press, 2007.
myocardial infarction and death from cardiovascular causes. 19. Wood AJJ. A proposal for radical changes in the drug-
N Engl J Med 2007;356. DOI: 10.1056/NEJMoa072761.
approval process. N Engl J Med 2006;355:618-23. [Erratum, 2. Yki-Järvinen H. Thiazolidinediones. N Engl J Med 2004;351: N Engl J Med 2006;355:2712.]
Copyright 2007 Massachusetts Medical Society. Downloaded from www.nejm.org on May 21, 2007 . For personal use only. No other uses without permission. Copyright 2007 Massachusetts Medical Society. All rights reserved.

Source: http://www.iranep.org/Articles/DM%20Rosiglitazone%20and%20cardiac%20risk%20NEJM%202007.pdf