Another case-control study ?

This week’s JAMA has a cohort study from the Netherlands looking at the risk of community-acquired pneumonia and use of gastric acid–suppressive drugs. The authors used a large, private practice clinical database to investigate a hypothesized linkage between the prescription of proton pump inhibitors or H2 receptor blockers and community acquired pneumonia. In order to adjust for the fact that many patients on these drugs are sicker, or more likely to have certain diseases, the authors performed a nested case-control examination that looked at recent acid-suppression therapy compared with previously discontinued therapy, adjusting for a large number of factors and diseases (such as age, sex, indication for the therapy, presence of diabetes, copd and stomach cancer).

The unadjusted relative risk of pneumonia for patients who had been prescribed acid-suppressant therapy vs. those who had not was quite high (4.5). In the matched, adjusted case-control cohort of recent therapy vs. past therapy, the adjusted odds ratio was 1.63, much lower, but still significant.

I believe it is impossible to properly adjust for all the reasons why pneumonia patients might have been prescribed acid-suppressant therapy shortly before getting pneumonia. Some of the factors that the authors controlled for are surely valid, but there are likely to be others. Even an individual practitioner’s style and personality could be linked with both the likelihood of prescribing these drugs and that particular patient population’s risk for pneumonia. Adequate adjustment is just not possible.

After multiple case-control and cohort studies showed significant cardiovascular benefit from estrogen replacement therapy, properly randomized trials demonstrated that this was not the case, forcing an incredible about-face. One would think that the estrogen replacement scandal would have put a major brake on performing and publishing these sorts of studies, yet they seem be flourishing. Has nothing been learned?

One reply on “Another case-control study ?”