Citation

Wang PS, Schneeweiss S, Glynn RJ, Mogun H, Avorn J. Ann. Epidemiol. 2004; 14(4): 296-303.

Affiliation

Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02115, USA. pwang@rics.bwh.harvard.edu

Copyright

(Copyright © 2004, American College of Epidemiology, Publisher Elsevier Publishing)

DOI

10.1016/j.annepidem.2003.09.012

PMID

15066610

Abstract

PURPOSE: The case-crossover design was originally intended to study brief exposures with immediate and transient effects, and acute outcomes with abrupt onsets. We investigated whether case-crossover methods can be used to study prolonged exposures and insidious outcomes. METHODS: We conducted a case-crossover study of 8220 patients aged > or = 65 years enrolled in several health benefits programs in New Jersey during the period between 1991 and 1995. All had episodes of central nervous system (CNS) adverse events (e.g., delirium). Drug exposures were assessed during case time periods and control time periods lasting 1, 2, 3, or 4 months. Exposures included 3 active regimens with established deleterious CNS effects (corticosteroids, digoxin, and opiates) and 2 inactive regimens without such effects (multivitamins and statins). RESULTS: In conditional logistic regression models, significantly elevated risks were observed for all three active drugs, regardless of which time windows were used. The magnitude of these risks generally increased with longer time windows. No significantly increased risks were observed for the 2 inactive drugs, regardless of the window duration. CONCLUSIONS: These results suggest that with lengthened exposure assessment windows, case-crossover methods may be useful for studying exposures with prolonged effects and outcomes with insidious onsets.

Language: en