Age- and gender-specific asthma death rates in patients taking long-actingbeta(2)-Agonists - Prescription event monitoring pharmacosurveillance studies

Objective: Prescription event monitoring is a national drug safety surveill ance scheme in which prescribers are prompted to report events and deaths f ollowing prescription of newly marketed drugs. This paper presents age- and gender-specific asthma death rates in patients prescribed the long-acting beta (2)-agonists salmeterol and bambuterol. Design and setting: Pharmacosurveillance cohort study of general practice p atients in England. Patients and participants: 15 406 patients prescribed salmeterol between De cember 1990 and May 1991, and 8098 patients prescribed bambuterol between F ebruary 1993 and December 1995. Methods: Patients prescribed these drugs by general practitioners in Englan d were identified using the national pharmacovigilance system of prescripti on event monitoring, in which details of all dispensed prescriptions were p rovided in confidence by the Prescription Pricing Authority. Questionnaires were sent to the prescriber asking for details of clients occurring after the first prescription. In each study an attempt was made to establish the cause of all deaths reported on the questionnaires, via retrieval of the pa tients' medical notes or examination of death certificates. Outcome measures and results: There was little evidence of heterogeneity in the drug-specific death rates and we therefore present the combined age- a nd gender-specific death rates for the 2 cohorts. Overall, there were 85 as thma deaths among people taking the long-acting beta (2)-agonists studied ( bambuterol and salmeterol cohorts combined). The overall death rate was 2.3 3 [95% confidence interval (CI) 1.84 to 2.84] per 10 000 months of observat ion. There were 37 asthma deaths among male patients (rate 2.40 per 10 000 months of observation; 95% CI 1.74 to 3.40) and 48 asthma deaths among fema le patients (rate 3.08 per 10 000 months of observation; 95% CI 2.21 to 3.9 8). There was no difference in death rates when male and female patients we re compared (rate ratio 0.78; 95% CI 0.49 to 1.22; p = 0.26). Conclusion: Prescription event monitoring is a form of prompted surveillanc e allowing rapid, uniform, national and practical assessment of newly marke ted drugs on large cohorts of patients in England. These data provide bench mark rates from which to assess the performance of newly prescribed anti-as thma drugs and generate hypotheses for inter analytical investigation in wh ich confounding by indication and asthma severity can be controlled for. An y differences in these rates should be considered as a source of signal gen eration within the contest of a surveillance programme. rather than as robu st evidence of any mortality differential between drugs.