A Bayesian review reported increased glaucoma odds in its benzodiazepine-related category, but that estimate combined drug types and relied on observational evidence.
The reported association
Jannini and colleagues synthesized observational studies of psychotropic medicines and glaucoma or intraocular pressure. Across all drug categories, the review included 22 studies and 293,228 medication users. Its Bayesian random-effects analysis reported an odds ratio of 1.550 for the benzodiazepine-related glaucoma category, with a 95% credible interval of 1.436–1.674.
This is an association estimate, not a finding that 55% of users developed glaucoma. An odds ratio also does not specify the absolute probability of an event. The absolute implication depends on baseline risk, the population, and the outcome definition.
Which drugs and which glaucoma?
The authors acknowledged that their benzodiazepine analysis also included Z-drugs such as zolpidem. That category should therefore not be described as a pure estimate for a single benzodiazepine. Short- and long-acting subgroups both showed associations, but grouping by duration of action does not remove differences between individual agents.
Much of the underlying benzodiazepine evidence concerned angle-closure glaucoma. Open-angle glaucoma, angle closure, and intraocular pressure are not interchangeable outcomes. A summary that omits the subtype can imply a broader result than the contributing studies support.
The statistical assumptions matter
For benzodiazepines and topiramate, the authors used informative prior distributions based on their assessment of existing evidence and biological plausibility. The credible intervals therefore reflect both the included data and those prior assumptions. They are not conventional confidence intervals from one new clinical trial.
The review also combined studies with different designs and, in some cases, approximated relative risks as odds ratios because glaucoma was uncommon. It identified these choices as limitations. Observational confounding and incomplete clinical detail remain relevant even when a pooled interval is narrow.
Using the result in a medication review
The review supports attention to eye-disease subtype and individual susceptibility when evaluating a medication history. It does not prove that a particular prescription caused an individual glaucoma episode. That assessment requires the actual agent, exposure timing, ophthalmic findings, anatomical susceptibility, and alternative explanations. The pooled association is useful context, with drug grouping and study design kept visible.
Source paper
Original analysis of the supplied journal article, identified as PMC12909488.
Leave a Reply