Levalbuterol is not more cost-effective than albuterol for COPD.

In a recent issue of CHEST, Truitt et al (January 2003)1 claimed that their retrospective chart review demonstrated that “levalbuterol afforded clinical and pharmacoeconomic advantages over racemic albuterol” in the treatment of hospitalized patients with COPD and asthma. The primary end point of the study was the total number of nebulizer treatments required. However, the target care path that was in operation during the historical cohort required albuterol to be administered every 4 h (six treatments per day), while in the current cohort levalbuterol was administered every 8 h (three treatments per day). Thus, the statistically significant difference in the primary end point was a result of differences in the care plans. There were no statistically significant differences in the number of extra treatments required by either cohort. Surprisingly, the reviewers missed this fatal flaw in the study design. The combination of more treatments, by design, and an artificially higher cost basis for albuterol resulted in a spurious statistical difference for the total cost of nebulizer therapy. They used the “average wholesale price” (AWP), but no hospital pharmacy ever pays the AWP. For many hospitals, the acquisition cost of racemic albuterol is around $0.32 per 2.5-mg dose compared to $1.82 per 1.25 mg for an equivalent dose of levalbuterol (ie, a 5.7-fold difference). In contrast, the AWP for the two products are $1.21 and $2.08, respectively (a 1.7-fold difference).2 No conclusions can be drawn about clinical advantages since the treatments were not administered in a double-blind randomized manner. Moreover, in patients with COPD (about 80% of the patients), FEV1 did not significantly increase from hospital admission to hospital discharge in either cohort. In such patients, it is possible that dosing every 8 h with albuterol would be as effective as dosing every 4 h since these patients show little response to bronchodilators. Also, we are incredulous about the statement that levalbuterol “appeared to have a more prolonged therapeutic effect.” The basis for this statement was a difference in the hospital readmission rate, but since this was a retrospective chart review, other factors such as differences in hospital discharge medication (data not presented) may have affected this end point. Last, “not significant” was designated as p 0.1 when the convention is p 0.05. As a consequence, the authors inferred that important pharmacoeconomic end points were different when they were not statistically different. For example, the authors state that those treated with levalbuterol “had shorter lengths of hospital stay” and “decreased costs for hospitalization,” but the p values for these end points for patients with COPD were 0.07 and 0.11, respectively. We conclude that the differences between albuterol and levalbuterol in this study were a result of differences in the number of treatments required by the care plan, invalid cost calculations, and the emphasizing of numerical differences that were not statistically significant by conventional criteria.