PURPOSE: To project the number of asthma deaths in the US population by calculating the estimated number of asthma-related deaths using the asthma death rate for SMART.
METHODS: For these calculations, we have used exposure data to salmeterol containing products estimated from the Surveillance Data Incorporated database.
RESULTS: In 2004, there were approximately 4,213,360 asthma patients exposed to salmeterol-containing products. Applying the asthma-related death rate from SMART (13/13,176 x 52 weeks per yr/28 weeks study duration), the expected number of salmeterol-exposed deaths in 2004 would be as high as 7720.Given an estimated 4,500 asthma-related deaths in the US for 2004 and that approximately 4.2 million of the 11 million treated asthmatics (38%) were exposed to a salmeterol-containing product, it is difficult to reconcile this with a literal extrapolation of the SMART results, which would suggest that as many as 7720/4500 or 172% of all current asthma-related deaths could be related to salmeterol exposure. Further, if the annualized asthma-mortality rate among placebo users from SMART is applied to the estimated number of treated asthmatics who were not exposed to salmeterol in 2004, the expected number of deaths would be 2869. Combined with the 7720 estimated for salmeterol-exposed patients, this would yield a total number of 10,589 asthma deaths or 235% of all estimated asthma-related deaths based on the US vital statistics.
CONCLUSION: Taken together, these data suggest that extrapolation of the asthma-related death rates from SMART would lead to overestimation of the true risk among US asthmatics exposed to salmeterol in clinical practice.
CLINICAL IMPLICATIONS: When considering a pharmacological approach to the treatment of asthma, the clinician should consider all safety and efficacy data and how it may affect their patients. The rate of the observed association between asthma-related mortality and salmeterol use in SMART appears to exceed the actual number of deaths reported from national surveillance databases.
DISCLOSURE: Harold Nelson, Consultant fee, speaker bureau, advisory committee, etc. Harold Nelson is an consultant, advisor and speaker for GSK; Other Yancey, Kral, Sutton and Crim are GSK employees.