Participants, methods, and results

We used the Mediplus database to identify patients with no previous diagnosis of airways obstruction. We defined exposed patients as patients who had used ophthalmic topical β blockers for the first time in the period 1993-7. Unexposed patients were randomly selected (loosely matched by age and sex to exposed patients). For validation we inspected a random sample of 40 full longitudinal records of exposed and unexposed patients.

We defined patients who had excess respiratory disease in two ways. Definition A patients were patients who in the 12 months after treatment with topical β blockers were given for the first time a drug used for the treatment of reversible airways obstruction (β₂ agonists, inhaled corticosteroids, theophyllines, and inhaled anticholinergics). Definition B patients combined definition A patients with patients who in the 12 months after treatment with topical β blockers had a new Read code for asthma or chronic obstructive pulmonary disease entered on their record.

Exposed patients (n=2645) were slightly older than unexposed patients (n=9094) (68.6 versus 67.5 years). Exposed patients were less likely than unexposed patients to smoke and to use systemic β blockers and were slightly more likely to visit their general practitioner (median six versus five visits). In definition A patients we found an adjusted hazard ratio at 12 months after treatment with topical β blockers of 2.29 (95% confidence interval 1.71 to 3.07)—equivalent to a number needed to harm of 55 patients (table).

Of the 3358 patients (including patients with previous airways obstruction) begun on a topical β blocker during the study period, 148 (4.4%) had used drugs for airways obstruction within the previous year. Airways obstruction had been identified as an active problem (definition B) within the previous year in 316 subjects (9.4%).

Comment

Topical β blockers for glaucoma or ocular hypertension may lead to new airways obstruction requiring treatment in a population not considered to be at excess risk. This finding raises an issue of public health importance because of the large number (approximately 500 000) of elderly patients in the United Kingdom who are treated for glaucoma and ocular hypertension. Topical β blockers have been shown to affect respiratory function in elderly patients with no previous history of airways obstruction, although a small, short term study disputed this. 3 4 Our data indicate an attributable risk of 1000 patients per year in the United Kingdom, one case every 11 years for a general practitioner. One would expect the effect of β blockade on airways function to be rapid—and indeed the risk ceases to be significant after the first year of exposure. This risk is in patients without previous airways obstruction; patients with pre-existing airways obstruction may well be more sensitive to β blockers.

Our study depends on a diagnosis of airways obstruction having been made. Therefore, allowing for a certain rate of missed diagnosis or misdiagnosis, we may have underestimated the true risk. An inherent weakness of the study is that clinical data could not be thoroughly validated. It is unlikely that objective spirometric evidence was always obtained. But for prescribing information the database is reliable, and a systematic error is unlikely to account for our findings.

Ophthalmologists, general practitioners, physicians, and pharmacists need to be aware of the possibility of iatrogenic airways obstruction in patients taking topical β blockers for glaucoma. When eyesight cannot be threatened within their expected lifetime, many frail elderly patients may be better off left untreated than risk airways obstruction.5 β blockers should be discontinued immediately when a patient develops airways obstruction and their ophthalmologist subsequently informed. A repeat prescription that includes topical β blockers and drugs for asthma should automatically sound an alarm.