The Times They Are A-Changin’: Time-Dependent Exposures and Time-Dependent Confounders

To the Editor—We read with great interest the study by Violi et al [1], which investigated the incidence of cardiovascular events (CVEs) after community-acquired pneumonia (CAP), and subsequently estimated the association between CVEs and mortality. The authors concluded that the occurrence of CVEs is associated with a 5-fold increase in 30-day mortality (adjusted hazard ratio, 5.49; 95% confidence interval, 2.91–10.37). This finding could have important clinical implications and therefore it is essential that it be quantified accurately. In this respect we would like to point out 2 issues ignored in the statistical analysis, which could substantially bias the current estimate: time-dependency of the exposure and time-dependent confounders.

http://ift.tt/2yVjqQX