A vascular clue to glaucoma from an unlikely prescription

Design

prospective nested case-control study in nationally representative United Kingdom primary care data, conditional logistic regression adjusted for age, ethnicity and deprivation

Population

218,056 men with primary open-angle glaucoma matched 4:1 by age and practice to 876,872 controls; mean age 69 and 68 years

Primary outcome

odds of primary open-angle glaucoma associated with prescription of erectile dysfunction medication

Effect

odds ratio 1.29 (95% CI 1.25–1.32) in one dataset and 1.17 (1.15–1.18) in the other; exposure in 20% of cases vs 17% of controls

The vascular hypothesis in primary open-angle glaucoma has been around for decades with thin supporting evidence. A nested case-control study used it as an opportunity: erectile dysfunction is a well-characterised marker of endothelial and small-vessel disease, so does being treated for it associate with glaucoma?

The study drew on United Kingdom primary care records, matching 218,056 men diagnosed with primary open-angle glaucoma 4:1 by age and practice to 876,872 controls. Erectile dysfunction medication had been prescribed to 20% of cases against 17% of controls, and after adjustment for age, ethnicity and deprivation the association held in both source datasets — odds ratio 1.29 (95% confidence interval 1.25 to 1.32) in one and 1.17 (1.15 to 1.18) in the other. Older age and Black or Asian ethnicity were consistently associated with higher risk, as were hypertension, diabetes, asthma, migraine and sickle cell disease. Myocardial infarction and angina, oddly, went the other way.

The authors' reasoning about mechanism is the interesting part: these drugs are taken on demand rather than continuously, which makes a direct pharmacological effect on the optic nerve implausible and points towards the underlying vascular state rather than the treatment. That inversion — the prescription as a marker of the disease rather than a cause of it — is also why this should not change anything about prescribing. The protective-looking association with myocardial infarction and angina is a warning sign in the same data: men under cardiology follow-up are also under more medical surveillance, and detection bias runs through all of this.

• Do not advise a man to stop erectile dysfunction medication on the basis of glaucoma risk.
• Treat the prescription as a marker of vascular disease rather than as an exposure.
• Note the ethnicity finding: Black and Asian men carried higher risk, which is relevant to screening thresholds in Indian practice.
• Consider systemic vascular assessment in a man presenting with normal-tension or progressive glaucoma.
• Read the apparently protective association with angina and infarction as evidence of detection bias in the dataset.

Why it matters

It is one of the better pieces of evidence yet that the vascular hypothesis in open-angle glaucoma is pointing at something real.

Don't overread it

This is an association between a prescription and a diagnosis in routine records — it says nothing about whether the drug affects the optic nerve.

The statistics, in plain English

With over a million men in the analysis, an odds ratio of 1.17 to 1.29 has extremely narrow confidence intervals — the association is certainly not chance. That precision says nothing about whether it is causal, and the absolute difference is small: 20% of cases against 17% of controls. The two datasets giving 1.29 and 1.17 with non-overlapping intervals is itself informative, since the same exposure and outcome measured in two samples from the same health system should agree more closely than that; the gap suggests dataset-specific recording differences. The inverse association with myocardial infarction and angina has no plausible biological reading and is best treated as a signal of surveillance bias.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.