Confounding by contraindication in a nationwide cohort study of risk for death in patients taking ibopamine.

Abstract

BACKGROUND: Outcomes may differ in treated and untreated patients because of a contraindication for treatment in the latter that is independently associated with the outcome of interest. OBJECTIVE: To evaluate the effects of confounding by contraindication on risk factors for death in patients taking ibopamine after its use was restricted in early September 1995. DESIGN: Retrospective cohort study. SETTING: The Netherlands. PATIENTS: 1146 patients with congestive heart failure who were prescribed ibopamine at least once and for whom medication history and medical data were available. MEASUREMENTS: Cardiovascular risk factors, clinical characteristics, and medication use. Each patient was assigned an index date (the date of death, or a random date for patients still alive at the end of the study). RESULTS: In univariate analyses comparing patients with an index date before and those with an index date after 8 September 1995, the relative risk for death associated with current use of ibopamine was 3.02 (95% CI, 2.12 to 4.30) compared with 0.71 (CI, 0.53 to 0.96), respectively. In multivariate analyses, the risk for death was 2.62 (CI, 1.76 to 3.90) and 0.93 (CI, 0.84 to 1.02), respectively. CONCLUSION: The marked inversion of the relative risk estimate can be considered a practical example of confounding by contraindication.