If this is not your name, click here.
| | If this is not your name, click here. | | |
| | Contact Us | Order Now | Journals | Bookstore | Register a colleague | | |
| |  Risk of Myocardial Infarction With Rosiglitazone Dependent on Dose and Co-Intervention Strategy: Presented at ECE By Chris Berrie BERLIN -- May 8, 2008 -- The increased risk of myocardial infarction (MI) previously seen in patients on rosiglitazone treatment actually depends on the dose and co-intervention strategy; adverse effects of the drug on serum lipids do not increase the risk of MI, according to findings of a meta-analysis presented here on May 5 at the 10th European Congress of Endocrinology (ECE). Previously, a meta-analysis by SE Nissen and colleagues found that rosiglitazone was associated with a significant increase in the risk of MI, with an odds ratio (OR) of 1.43 (Nissen SE, Wolski K. N Engl J Med. 2007;356:2457-2471). The analysis also showed a tendency for rosiglitazone to result in unfavourable changes in serum lipids plus a tendency in one study for rosiglitazone to increase cardiovascular events (CVEs). However, in the Nissen analysis, none of the studies took into account CVEs as primary outcomes, and none were controlled according to cardiovascular risk factors, said Jorge Sapunar, MD, Associate Professor, Department of Epidemiology, La Frontera University, Temuco, Chile. "For these reasons, we cannot accept or refuse the Nissen conclusions," Dr. Sapunar added. Dr. Sapunar and colleagues analysed the data in the studies included in Nissen's meta-analysis to determine the risk of CVEs and MI according to rosiglitazone dose, the comparator drug, and co-intervention. His research team also wanted to determine whether these levels of risk are affected by baseline serum lipids and whether the OR variations depend on changes in serum lipids induced by rosiglitazone treatment. Outcomes were evaluated for all CVEs (cardiac failure, arrhythmia, angina, MI, revascularisation, stroke, thrombosis, thrombophlebitis) and MI alone. In the first phase of the subgroup meta-analysis by rosiglitazone dose, the ORs for CVE were significantly elevated for rosiglitazone at a dose of 2 mg (2.49; P = .04) and 8 mg (1.28; P = .006), but not for the 4-mg dose (P = .35). ORs for MI alone reached significance only for the 8-mg dose of rosiglitazone (1.68; P = .004). Thus, these effects of rosiglitazone change according to the dose. When CVEs were evaluated by rosiglitazone comparator drug, placebo was significantly associated with a higher OR (1.34; P = .002), but metformin (P = .88) and sulfonylureas (P = .16) were not. The ORs for MI reached significance only for sulfonylureas (2.24; P = .02), but not for placebo (P = .11) and metformin (P = .40). Thus, these effects of rosiglitazone change according to its comparator agent. Finally, in the analysis of ORs for CVE by rosiglitazone co-intervention, significance was associated with no co-intervention (1.23; P = .02) and with insulin (2.76; P = .003). Analysis of ORs for MI showed significance with no co-intervention (1.56; P = .01), while insulin had a high (3.62) but nonsignificant OR (P = .18). Thus, these effects of rosiglitazone depend on the co-intervention. In the second phase of the analysis, with linear regression models for changes in ORs for CVE by baseline variables, rosiglitazone 4 mg/day and high-density lipoprotein (HDL) baseline levels had the strongest inverse effects or association with the OR variation, and insulin as a co-intervention had the strongest direct association with change in OR, Dr. Sapunar noted. For changes in ORs for MI, the female/male ratio had the strongest inverse association. The final models were for changes in ORs for CVE and MI according to changes in serum lipids with the use of rosiglitazone. In both cases, total cholesterol/HDL ratio had the strongest direct relationship, Dr. Sapunar said. Therefore, this use of a linear meta-analysis revealed previously unseen variability of outcomes according to the rosiglitazone dose, comparator agent, and co-intervention and showed that the adverse effects of rosiglitazone on serum lipids do not increase the risk of MI, the researchers concluded.
[Presentation title: Effect of Baseline Sample Characteristics, Comparator Drug, Co-Interventions and Rosiglitazone Doses on the Risk of Myocardial Infarction: Multivariable Lineal Regression Analysis. Abstract OC3.9]
|
