The incentive to publish negative studies: how beta-blockers and depression got stuck in the publication cycle

The hypothesis that use of beta-blockers causes depression has been proposed more than 40 years ago. However, despite the publication of numerous studies, the issue remains unresolved. The aim of this article is to describe a publication pattern in these studies, which may explain why the controversy persists.

A literature search was performed in Embase, Medline, and references of relevant articles. Studies that tested whether patients using beta-blockers for cardiovascular conditions had an increased risk of depression compared with nonusers were included.

The search yielded 22 observational studies. A specific publication pattern recurred several times in these studies. Each study that presented statistically significant findings was followed by several studies that refuted those findings. This publication cycle arose in cross-sectional studies, longitudinal studies, and meta-analyses, as well as for beta-blockers in general, propranolol, and nonselective beta-blockers. Some underpowered nonsignificant findings were published in high-impact journals.

The publication cycle clearly illustrates that significant findings are published ahead of nonsignificant findings, and that significant findings seem to provide an incentive to publish nonsignificant studies. Therefore, studies that may have correctly shown that beta-blockers increase the risk of depression will have been refuted too. Publication bias because of false nonsignificant studies may obstruct scientific progress.