Intermittent preventive malaria treatment in Tanzanian infants. Authors reply [letter]

Comparison of the Kaplan-Meier curves in our original and more recent publications is not the best way to investigate the effect of clinical malaria episodes on the efficacy of IPTi and can lead to erroneous conclusions. This is because the scope of the analyses differs by more than just previous malaria episodes: additional follow-up and malaria episodes were included in the recent publication which is particularly relevant because the Kaplan-Meier curves diverge with continuing follow-up. In addition those excluded from the first analysis (ie those who had had a previous episode of clinical malaria) are unlikely to have been a random subset of the study cohort. (excerpt)