Response to Letter Regarding Article, “Rosiglitazone and Outcomes for Patients With Diabetes Mellitus and Coronary Artery Disease in the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) Trial”

Abstract

We thank Drs Liu, Chen, and Huang for their interest in our article. 1 They point out that, in contrast to our results, previous pooled analyses of randomized trials of rosiglitazone showed a higher incidence of myocardial infarction, and that our observed differences (or rather lack of differences) in cardiovascular outcomes associated with rosiglitazone were potentially attributable to residual confounding. They also raise the point that we did not report a falsification end point that could have assessed the likelihood that observed results were more likely attributable to confounding than true effects.Our study reported prospectively collected longitudinal results from the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) trial cohort based on the diabetes mellitus treatment received, and was not a cross-sectional analysis. We agree that there is no substitute for randomization and no observational study, ours included, can fully exclude residual confounding. We should then also note the potential weaknesses of the meta-analysis that first implicated rosiglitazone as a cause of myocardial infarction, and reemphasize the unique aspects of the BARI 2D design and analysis that would be expected to minimize the impact of confounding. Even a well-conducted meta-analysis of limited studies will remain limited, because it is unable to compensate for the weaknesses of the original studies and may even unintentionally conceal those weaknesses. A carefully performed, independent examination of the meta-analysis regarding cardiovascular hazard from rosiglitazone concluded that the risk for myocardial infarction and death for diabetic patients taking rosiglitazone remained uncertain. 2It is within this context that the BARI 2D design and results merit further consideration. In the BARI 2D trial, 3 patients were randomly assigned to a glycemic strategy of insulin sensitization or insulin provision. Therefore, initiation of rosiglitazone was generally determined by treatment assignment rather than patient indication, and the biases seen when the initiation of drug treatment is influenced by severity of illness were minimized. We used propensity-matching to further reduce any baseline differences when comparing treatment groups. Despite this, we did not detect a signal of ischemic cardiovascular harm from rosiglitazone. Similar conclusions were recently confirmed by a thorough FDA review of the independent readjudication of the 1 completed randomized trial designed to test the cardiovascular safety of rosiglitazone. 4A falsification end point is most useful when an association is detected and the goal is to verify that the association is causal and not attributable to residual confounding. For our analyses, such a strategy was not applicable. We did, however, report what might arguably be considered a prespecified, unrelated, verification end point. An increase in bone fractures, especially among women, had been observed in the previous prospective, randomized trials of rosiglitazone. 5,6 In our analy...

Thiazolidinediones and Risk of Atrial Fibrillation Among Patients with Diabetes and Coronary Disease

“…Early thiazolidinedione use was defined as maintenance of baseline thiazolidinedione use at 6 months or any new initiation of thiazolidinediones use within the first 6 months, as previously described. 29 Categoric data were presented as counts with percentages, and statistical differences were tested using chi-square tests and Fisher exact tests where appropriate. Continuous variables were presented as mean with standard deviations for normally distributed data and as medians with interquartile range for non-normally distributed data.…”

Metformin in combination with rosiglitazone contribute to the increased serum adiponectin levels in people with type 2 diabetes mellitus

“…Studies have shown that rosiglitazone has the ability to attenuate inflammatory effects, elicit insulin sensitivity and lower the glucose by serving as the peroxisome proliferator-activated receptor-γ (PPAR-γ) agonists ( 7 , 8 ). Accordingly, rosiglitazone has an obvious effect on atherosclerosis, inflammation, endothelial dysfunction and T2DM ( 9 ).…”