Honig and colleagues note that our study protocol was considered "exploratory" when submitted to the FDA because, at the time, it did not have a statistical plan of analysis. Indeed, the larger study (study 0029), based on 762 patients in whom efficacy and safety outcomes were of primary interest, was the study submitted for approval. It contained all elements required by the FDA, including an a priori plan of data analysis. On the other hand, the subprotocol used for our study, applied to 150 of the 762 patients, intended rather to address outcomes of clinical and economic effectiveness. It did not have similar a priori requirements for FDA approval. Nonetheless, we did subsequently develop an a priori plan of analysis that was presented to our advisors and to the study sponsor before data were analyzed. Our results reflect this plan of analysis, which also includes upcoming analyses of costs and of quality-of-life data currently being prepared for publication.

Whether the protocol should be submitted along with a manuscript for peer review by scientific journals is an issue that the editors of this and other journals could address. We believe that the protocol's place is in the Methods section of a paper.

Estrada and Byrd suggest that the randomization done for the main study may have been lost in the substudy. This is impossible because the decision to participate in the subprotocol was made before randomization. Although patients who do choose to participate in a subprotocol may differ from the group as a whole, randomization will assure that the two groups being compared are similar.

Finally, we agree entirely with both correspondents that more evidence is required to demonstrate the effectiveness of leukotriene receptor antagonists and 5-lipoxygenase inhibitors as prophylaxis of persistent mild-to-moderate asthma. As stated in our article, studies that have more patients and longer follow-up and that compare this new medication to standard accepted therapies (especially inhaled corticosteroids) must be conducted to provide more pragmatic information to prescribers and patients and to support the practice of evidence-based medicine.