Yaffe et al1 describe a higher cumulative incident rate of dementia among veterans with posttraumatic stress disorder (PTSD) compared with veterans without PTSD. Unfortunately, the startling low rates of “history of tobacco use” reported in their article (3.5% for veterans with PTSD, 2.9% for veterans without PTSD) are not credible. The general population rates of tobacco smoking were more than 20% at the time these veterans were diagnosed,2 and population-based studies show 45% smoking rates for individuals with PTSD.3 Veterans with PTSD have higher rates of smoking (30%-50%) than the general population and higher rates than veterans without PTSD.4 Tobacco use is a well-documented risk factor for development of dementia.5 Thus, it is entirely possible that higher rates of dementia demonstrated in veterans with PTSD are at least partially related to higher rates of tobacco consumption in this group. The article is not explicit about whether hazard models were adjusted for history of tobacco use. Even if this adjustment were conducted, the lack of accurate tobacco use data used could have led to spurious results. This lack of accurate tobacco use data should be acknowledged as a serious limitation of the study.
Saxon AJ, McFall ME. Tobacco Use Data Questioned in Veterans Study. Arch Gen Psychiatry. 2010;67(12):1324. doi:10.1001/archgenpsychiatry.2010.166