We thank Dr. Oba for his comments, in which he concludes that tiotropium is the preferred therapy for management of chronic, stable COPD. His conclusion is based on a meta-analysis of 4 tiotropium studies demonstrating a –6% risk difference (95% CI, –10% to –2%) in the annual rate of hospitalizations compared with placebo, as well as a lack of a statistically significant reduction against placebo for studies of inhaled corticosteroids or LABAs that reported this outcome. We described similar findings in our review. The proportion of participants requiring hospitalization for COPD was lower with tiotropium than with placebo (risk difference, –2% [CI, –4% to –1%]).

We identified 42 eligible randomized, controlled trials of inhaled therapies. However, few reported hospitalizations. When reported, reductions were not consistently observed, and very few studies assessed comparative effectiveness across categories of long-acting inhalers. Of the 10 eligible tiotropium studies, only 5 reported hospitalizations and 4 used placebo controls. Only 1 of these studies demonstrated statistically significant benefit. The difference in effectiveness estimates for annual rates compared with the proportion of participants hospitalized is probably due to some participants requiring recurrent hospitalizations. The clinical significance of these pooled reductions in hospitalizations and the preferred analytic method are not known. Furthermore, investigators have demonstrated that selective study or outcome reporting (publication bias) results in biased (and more positive) effectiveness estimates (1). Therefore, we do not agree that these findings are sufficient to draw comparative effectiveness conclusions.

Data for other outcomes did not support the superiority of a particular class of long-acting inhaler. Many patients with symptomatic COPD place an equal or greater value on obtaining a noticeable improvement in respiratory health status or reduction in exacerbations rather than a reduction in hospitalizations. No studies directly compared tiotropium with inhaled corticosteroids. Tiotropium did not provide a clinically noticeable improvement in the average respiratory health status scores versus placebo, and there were no statistical or clinical differences versus LABA. On the basis of pooled results of the 2 comparative studies, tiotropium did not reduce exacerbations compared with LABA. None of the inhaled monotherapies reduced mortality versus placebo (relative risk with tiotropium, 0.94 [CI, 0.60 to 1.47]). Combination therapy with LABA and inhaled corticosteroids reduced mortality in relative terms (relative risk, 0.82 [CI, 0.69 to 0.98]). The absolute reduction of 1% was not statistically significant.

On the basis of the available results, we conclude that the current level of evidence does not allow a determination of whether any long-acting inhaled therapy (or combination of these therapies) is superior to any other for management of chronic, stable COPD (2). Additional large, long-term, randomized trials comparing relative effectiveness and harms are needed.