The authors of both letters express concerns related to the discontinuation of ipratropium therapy in many participants. They view ipratropium as a standard of care in the community and therefore conclude that withdrawing the drug makes the trial both clinically irrelevant and unethical. We will briefly review the scientific basis for their claims.

Two large trials compared the bronchodilator effects of ipratropium, albuterol, and both drugs used in combination (1, 2). Both trials clearly demonstrated that ipratropium is an effective bronchodilator in COPD, but albuterol produces almost identical responses. Combined treatment produces better bronchodilation than either drug alone, but the additive effect is small and is not associated with overall differences in symptom scores among the 3 treatment groups. These results indicate that albuterol and ipratropium can be used interchangeably for short-term control of symptoms, a conclusion in full accord with current expert recommendations (3). The clinical utility of combined therapy for this purpose remains unproven (4). In addition, there is very good evidence that regular use of ipratropium does not affect long-term changes in lung function (5).

Drs. Good and Longo make the additional claim that ipratropium reduces exacerbation rates by 32% when used with a short-acting ß-agonist. The basis for this assertion is a meta-analysis of 3 trials (6), with the overall conclusion generally being drawn from the 2 previously cited studies (1, 2). Exacerbation rates in those 2 trials were inferred from changes in prednisone use, which was reduced when ipratropium was added to albuterol. However, prednisone use was only 1 of numerous other secondary efficacy and safety variables that showed no treatment effects. Claims based on secondary or post hoc analysis must be recognized for their limitations because they often prove to be incorrect when appropriate prospective testing is performed. Furthermore, authors of another review identified 4 trials that involved more than 1000 patients and showed no significant advantage of ipratropium over placebo for reducing exacerbations (relative risk, 0.95 [95% CI, 0.78 to 1.16]) (7).

Drs. Good and Longo asked us to provide a table comparing health care events for COPD before the study with the same events during the trial to determine whether stopping ipratropium therapy had adverse effects in the placebo group. These data are to be found in Table 1 (baseline characteristics) and 3 (secondary outcomes) of our article. We obtained retrospective information from patient self-report and prospective information from monthly interviews and medical records. We expressed frequencies in both tables in common units of mean events per patient-year. Some comparisons were not possible (for example, courses of antibiotics vs. days of antibiotics). We can compare COPD hospitalization rates by summing the 2 categories for hospitalization in Table 1 with the single category in Table 3. The resulting 915 patients in the placebo group reported a mean rate of 0.24 COPD hospitalization per patient-year in the previous year; we identified a mean rate of 0.25 hospitalization per patient-year during the 6 months of the trial. We can similarly compare the sum of urgent physician visits and emergency department visits in Table 1 (mean, 0.59 per patient-year) with the umbrella term of unscheduled visits in Table 3 (mean, 0.49 per patient-year). These comparisons are deeply flawed for obvious reasons, but we can fairly conclude that they raise no red flags. For the reasons previously detailed, we strongly refute the suggestions that we conducted an irrelevant or unethical trial, either willfully or through ignorance. Our decision to allow albuterol as the sole rescue therapy is fully consistent with the scientific evidence. Our study highlights the importance of conducting prospective clinical trials with predefined primary end points and thereby provides definitive evidence of the effect of tiotropium on COPD exacerbations.