Endocrine Abstracts (2008) 16 OC3.9

Effect of baseline sample characteristics, comparator drug, co-interventions and rosiglitazone doses on the risk of myocardial infarction: multivariable lineal regression analysis

Jorge Sapunar, Sergio Muñoz, Marcela Jiménez & Eugenia Ortiz


La Frontera University, Temuco, Cautín/Araucanía, Chile.


Nissen’s meta-analysis concluded that the use of Rosiglitazone (RSG) was associated with a significant increase in the risk of myocardial infarction. However, none of the analyzed studies took into account the cardiovascular events as primary outcomes, neither were controlled by cardiovascular risk factors.

Objectives: 1. To know if there is an association between RSG myocardial infarction OR and the magnitude of the difference in baseline serum lipids and HbA1c in intervention groups.

2. To know the effect of RSG doses, comparator drug and co-interventions in RSG’s myocardial infarction OR.

Methods: We analyzed the same studies included in Nissen’s meta-analysis. For each study we recorded myocardial infarction incidence, comparator drug, co-intervention, RSG doses, the difference in baseline serum lipids and HbA1c in intervention groups and the difference in serum lipids and HbA1c change attributed to intervention. We modeled the RSG’s myocardial infarction OR with the variables detailed before by bivariate and multiple lineal regression with Stata/SE 10.0 software (Stata Corp LP, 4905 Lakeway Drive College Station Texas 77845, USA).

Results: Myocardial infarction OR increased if comparator was placebo, co-intervention was insulin and RSG doses was 2 mg. We didn’t find an association between myocardial infarction OR and the difference in baseline serum lipids and HbA1c in intervention groups. When we studied the myocardial infarction OR and the difference in serum lipids and HbA1c change attributed to intervention by bivariate lineal regression, only an increase in serum triglicerides could predict an increment in myocardial infarction OR. When we modeled Rosiglitazone’s myocardial infarction OR with the difference in baseline serum lipids and HbA1c in intervention groups r2 was 0.49.

Conclusion: The effect of Rosiglitazone on the risk of myocardial infarction can change with its doses and co-interventions. The adverse change in serum lipids induced by RSG didn’t affect the magnitude of myocardial infarction risk.

Article tools

My recent searches

No recent searches.