Review: β-agonist use is weakly associated with death from asthma
ACP J Club. 1994 May-June;120:83. doi:10.7326/ACPJC-1994-120-3-083
Mullen M, Mullen B, Carey M. The association between β-agonist use and death from asthma. A meta-analytic integration of case-control studies. JAMA. 1993 Oct 20;270:1842-5.
To determine the association between β-agonist use and death from asthma. The effects of delivery mode and patient age were also investigated.
MEDLINE was searched using the search terms β-agonist and asthma, bibliographies of identified studies were reviewed, cited references were located in Science Citation Index, and leading research journals from the past 15 years were reviewed.
Studies were included if they were case-control studies of patients who did and did not die of asthma, in which exact numbers of patients who did and did not use β-agonists were known. 200 studies were screened.
The frequencies for patients who did and did not use β-agonists and who died and did not die were used for chi-square tests. Each test was coded for patient mean age and β-agonist delivery mode (oral, metered-dose inhaler, nebulizer). The tests were transduced to common metrics for significance levels and effect sizes. After weighting study outcomes by their respective sample sizes, combinations of significance levels (z) and effect sizes (r) were done.
6 studies comprising 364 patients who died and 1388 control patients who did not die were selected for inclusion. An association was noted between use of β-agonists and death from asthma (z = 4.0, P < 0.01);, however, the magnitude of the effect was weak (mean r = 0.055). The composite odds ratio [COR] was 1.28 (P < 0.01). Examining the effect of mode of delivery showed a stronger association. Only when β-agonists were administered by nebulizer was there a weak association with death (z = 4.5, COR 2.47, P < 0.01 for both). Adults were more likely to die of asthma than were adolescents when β-agonists were administered by metered-dose inhaler (z = 2.4 r = 0.95, P < 0.01).
A weak association exists between β-agonist use and death from asthma. When mode of delivery and patient age are taken into account, the association is stronger.
Sources of funding: Not stated
For article reprint: Ms. M. Mullen, The American Lung Association of Central New York, PO Box 6409, Syracuse, NY 13217-6409, USA. FAX 315-422-9710.
Drugs sometimes increase the risk for death from the same condition they are being used to treat. The classic example is the increased risk for sudden cardiac death in patients treated with class 1 antiarrhythmic drugs. Detection of these adverse drug effects is difficult in the absence of randomized trials. Observational study designs are limited because treatment decisions may be influenced by disease severity, which is itself likely to be predictive of outcome.
Mullen and colleagues used meta-analysis to evaluate the epidemiologic evidence linking the use of β-agonists with an increased rate of death from asthma. They concluded that the association is extremely weak. I do not think they are right to draw this conclusion from the data. Mullen and colleagues did not assess the quality of the individual studies. This is a serious omission because the results of observational studies may be greatly influenced by their methodologic rigor. Mullen and colleagues pooled the raw data, ignoring adjustments and stratification for measures of disease severity that strengthened the associations seen in the original studies. They did not mention the steep relation between dose of inhaled β-agonists and risk for death from asthma seen in 1 study (1). They did not acknowledge the extensive discussion of the methodologic and biologic issues that came after publication of the case-control studies, and they failed to comment on the experimental evidence that continuous use of β-agonists may worsen airways control (2, 3). In my view, clinicians concerned about the emerging evidence that these drugs may have serious deleterious effects should not be reassured by the evidence presented here.
David Henry, MB
The University of NewcastleWaratah, Australia
3. van Schayck CP, Rutten-van Molken MP, van Doorslaer EK, et al. Two-year bronchodilator treatment in patients with mild airflow obstruction. Contradictory effects on lung function and quality of life. Chest. 1992;102:1384-91.