Not So Rosi…

May 24, 2007


Commentary by Seagram Villagomez MD, Chief Resident

Since its approval in 1999, nearly 1 million Americans have used the thiazolidinedione (TZD) rosiglitazone (Avandia – GlaxoSmithKline) for the treatment of Type 2 Diabetes.  However, in a drug class which seems plagued by concerns, the safety profile associated with rosiglitazone has been brought to question. Previously, troglitazone (Rezulin) was pulled off the market secondary to hepatoxicity, while muraglitazar was not approved by the FDA given adverse cardiovascular events during early clinical trials.  In a study just released by the New England Journal of Medicine and making headlines across the world, Steven Nissen and colleagues at the Cleveland Clinic demonstrate that rosiglitazone has been associated with an increased risk of myocardial infarction and death from cardiovascular causes.

Nissen et al. conducted a meta-analysis which included searches of published data, as well as publicly available data from the FDA website and GlaxoSmithKline’s clinical trial registry.  Their inclusion criteria consisted of trials whose duration was >24 weeks,  included a randomized control group and had available data on MI and cardiovascular death.  In the end, 42 trials were included in this meta-analysis (including the recently published DREAM and ADOPT Trials).

Approximately 28,000 patients who were predominately white males in their late-50’s with poorly controlled diabetes (A1c = 8.2%) were randomized by these trials between a treatment regimen including rosiglitazone and a control group (any regimen not including rosiglitazone).  Nissen et al. found that the odds ratio for MI was 1.43 ( 95% CI 1.03 to 1.98, p = 0.03) and the odds ratio for death from cardiovascular causes was 1.64 (95% CI 0.98 to 2.74, p = 0.06). However, the total event rates in these combined trials were small. There were 86 myocardial infarctions in the rosiglitazone group and 72 in the control group. 39 deaths from cardiovascular causes occurred in the rosiglitazone group compared to 22 in the control group.

The exact mechanism for the purported increased risk is unknown. The authors speculate it may be secondary to its effects on LDL levels (increasing it by as much as 18%), volume overload, or its effects on reducing hemoglobin levels and thus increased cardiovascular stress.  In comparison, pioglitazone (Actos – Takeda) was shown to have favorable cardiovascular outcomes including MI in previous randomized trials, which is thought to be secondary to reduced effect on the above.

The results presented by Nissen et al. are of concern given a population already at a higher risk for cardiovascular mortality and morbidity. However, this study is not without its flaws.  This study was limited by the use of publicly available trial data and not the raw patient level data obtained by the manufacturer.  The included trials were multiple small ones of short duration and therefore the event rates remained small and a time to event analysis could not be performed.  Trial regimens were not standardized (i.e. dosages and medications) and an exact control group was never determined. Furthermore, these trials in which MIs and cardiovascular deaths were secondary endpoints were not powered to determine significant changes.

Even though the data may be of borderline significance, this report raises many interesting questions concerning the safety and approval of rosiglitazone.  Currently, several larger trials are underway to investigate its effect on cardiovascular events. These trials may provide the information to make clear decisions,  however the tremendous publicity this article has generated may have already altered our practice for better or worse.


Are you telling your patients to stop Avandia based on this new data?

View Results

One comment on “Not So Rosi…

  • Avatar of Ira Breite, MD
    Ira Breite, MD on

    I always wonder when two drugs of the same “class” have totally different effects. We “extrapolate” data on atorvastatin to simvastatin, Prilosec to Prevacid etc. all of the time. I am curious about what the long term trials show. Although this study will pretty likely kill Avandia as a drug (regardless of what the awaited long term data shows) for quite some time, we will have to rethink the concept of “formulary” if it turns out that Actos and Avandia have significant differences in cardiovascular mortality.

Comments are closed.