Attrition and non-compliance in secondary stroke prevention trials

Background: Attrition and non-compliance of subjects in secondary stroke prevention trials due to study drug-induced adverse events and loss to follow-up could lead to bias and loss of information, thus affecting the analysis of study results. Methods: We reviewed results from ten antiplatelet stroke prevention clinical trials: CAN TIA, DUTCH TIA, SWED ASA, SALT, UK TIA, CATS, TASS, ESPS, ESPS-2, and CAPRIE to tabulate the frequencies for total subject discontinuation, voluntary withdrawal, and loss to follow-up. Results: Forty thousand seven hundred and thirty ( 40,730) subjects participated in the aforementioned secondary stroke prevention trials. The range of outcomes was 11.8 - 52.0% for subjects discontinued for any reason (n = 9 trials); 3.0 - 20.9% for study drug-induced adverse events ( n = 9 trials), and 4.2 - 7.8% for voluntary withdrawal ( n = 10 trials). Conclusion: There is a substantial discrepancy ( up to 20%) between the frequencies of total subject discontinuation for any reason and the sum of study drug-induced adverse events, voluntary withdrawal and loss to follow-up. Underestimation of these important outcomes may limit the ability of clinicians to translate results from clinical trials into medical practice. Copyright (C) 2004 S. Karger AG, Basel.