Evidence of the Month

Commentaries on both new and classic studies of importance for the treatment of diabetes are posted here monthly.

Jaime A. Davidson University of Texas Southwestern Medical Center Dallas, Texas, USA	Click here to download PowerPoint slides of this commentary.
	Click here for the programme evaluation and post-test.

Rosiglitazone is a thiazolidinedione that has widely been used for the treatment of Type 2 diabetes. A meta-analysis by Nissen and Wolski, which was initially published on-line by the New England Journal of Medicine on 21 May 2007, raised serious questions about the safety of rosiglitazone when it reported an association between rosiglitazone and an increased risk of myocardial infarction and cardiovascular-related mortality. The meta-analysis has generated considerable attention and discussion both within and outside the diabetes community, and several publications have followed in response.

During the annual meeting of the American Diabetes Association in June 2007, a roundtable was added so that questions and issues related to the meta-analysis could be directly addressed in an open forum. Philip Home, a principal author of the RECORD (Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes) study and an advisory board member for WorldWIDE, and Steven Nissen took different positions on the merits and conclusions of the meta-analysis.

Meta-analysis can be conducted in many ways, and by no means am I an expert in meta-analysis science. However, it should be noted that the Nissen and Wolski meta-analysis pooled data from 42 trials and left out the trials that did not report events. In the meta-analysis, the mean age of the patients was 56 years, and the mean baseline HbA_1c was 8.2 %.

It is well known that Type 2 diabetes is a risk factor for coronary artery disease. Haffner and colleagues (1998) found that individuals with newly diagnosed Type 2 diabetes had a 20% risk of an acute myocardial infarction occurring within 7 years. That fact is well known, and deaths in individuals with diabetes mellitus today are mainly cardiovascular related, including acute coronary syndrome death. In a recently reported subgroup analysis of 11 TIMI (Thrombolysis in Myocardial Infarction) studies, which included over 10,000 people with diabetes, the risk of death at 30 days post-acute coronary syndromes was significantly higher in people with diabetes, and the risk was found to be 78% no matter what the intervention (Donahoe et al, 2007). So the real problem in clinical trials in people with Type 2 diabetes is increased cardiovascular events due to the inherent risk for these events. At an average HbA_1c of 8.2 %, many patients included in the meta-analysis could have been in the lower end of the HbA_1c spectrum and many others in the higher end. Poor diabetes control equals a higher number of events. A reasonable question should be, were the events mainly in the higher-end spectrum of HbA_1c levels or distributed evenly between the 2 groups?

Let me now address the number of myocardial infarctions and deaths reported in the meta-analysis. There were 86 myocardial infarctions among the 15,565 individuals treated with rosiglitazone, which is 1 event per 181 individuals. There were 72 myocardial infarctions among the 12,282 individuals in the control arm of the meta-analysis, which is 1 event per 171 individuals. It is clear that the treated group had fewer myocardial infarctions than the control group. However, it should be noted that the authors found a significantly increased risk of myocardial infarctions with rosiglitazone treatment based on odds ratio calculations. There were 39 deaths in the rosiglitazone-treated group, which is 1 death per 399 individuals, and 22 deaths in the control group, which equals 1 death per 558 individuals. Although the rosiglitazone-treated group experienced a higher number of deaths, the risk of death between the 2 groups did not reach statistical significance. With regard to their findings, the authors clearly stated that the meta-analysis is a call for further investigation and not a conclusive study.

The US Food and Drug Administration (FDA) called a meeting on 30 July 2007 for 2 advisory panels—one focused on endocrinologic and metabolic drugs and the other on drug safety and risk management—to review the FDA’s internal safety data on rosiglitazone. David Graham, MD, MPH, with the FDA’s Office of Surveillance and Epidemiology, presented a risk/benefit analysis of rosiglitazone and a comparison of rosiglitazone and pioglitazone, based on the FDA’s review of the following studies:

The first question addressed was, does rosiglitazone increase cardiovascular risk? Dr Graham took the affirmative position on this question based on analysis of the DREAM study and the FDA meta-analysis of rosiglitazone. In the DREAM study, the incidence of the cardiovascular composite was lower with the use of an angiotensin-converting enzyme (ACE) inhibitor alone (1.8%), similar for rosiglitazone alone and placebo (2.4%), and higher for rosiglitazone combined with the ACE inhibitor (3.4%). As Dr Graham admitted, although the risk was increased with the combination of rosiglitazone and an ACE inhibitor, the meaning of the possible interaction was unclear. The FDA meta-analysis found that 6 to 12 months of rosiglitazone use increased cardiovascular risk by 20% to 68% compared with non-use, and a review of data for the DREAM study showed an increased risk of approximately 40%. The problem is that 6 to 12 months of therapy does not address long-term events. The Diabetes Control and Complications Trial would have given us the wrong results if we had only seen the data for the first 3 years.

The FDA also looked at whether rosiglitazone increases cardiovascular risk compared with pioglitazone and drew on the PROactive and GLAI studies and a pioglitazone meta-analysis to address this question. The meta-analysis of 10,199 patients treated with pioglitazone showed a decreased cardiovascular risk of approximately 25% with pioglitazone, and analysis of PROactive, a high-risk population, found that pioglitazone decreased cardiovascular risk by approximately 15%. Furthermore, in GLAI, a head-to-head comparison study, rosiglitazone increased the risk 3.5-fold compared with pioglitazone. Dr. Graham concluded that pioglitazone’s cardiovascular risk profile differs from that of rosiglitazone.

The 2 advisory boards voted to maintain rosiglitazone in the market. They agreed that the data presented signalled an alarm but cited the weak or flawed evidence presently available and expressed the hope that trials now in place might answer in a better and definite way the question of whether rosiglitazone increases cardiovascular risk.

In studies that followed Nissen and Wolski’s meta-analysis, Singh et al (2007) concluded from their meta-analysis that among patients with impaired glucose tolerance or Type 2 diabetes, rosiglitazone use for at least 12 months was associated with a significantly increased risk of myocardial infarction and heart failure but without a significant risk of cardiovascular mortality. An interim analysis by the RECORD Study Group (Home et al, 2007) was inconclusive regarding the effect of rosiglitazone on the overall risk of hospitalization or death from cardiovascular causes, and furthermore, there was no evidence of increased death from all causes. Rosiglitazone was associated in the RECORD study, as shown in many studies, with an increased risk for congestive heart failure. It is known that diabetes is a risk factor for congestive heart failure, and duration of diabetes and age are also contributors.

Summary
Rosiglitazone, a well-known and widely used oral antidiabetic agent, remains a pharmacological option for the treatment of Type 2 diabetes. The FDA panel convened on 30 July 2007 voted by a majority to continue the clinical use of rosiglitazone but with the recommendation that new warnings related to cardiovascular risk be added to its labeling. On 14 November 2007, the FDA announced a change in the labeling for rosiglitazone to reflect a potential increased risk for heart attacks.