In Reply: Dr Chang and colleagues describe several concerns, including that residual confounding may account for the observed differences in mortality in the propensity-matched sample, that patients whose corticosteroid regimen was changed from low-dose oral to high-dose intravenous should have been counted as treatment failures, and that the intention-to-treat analytic strategy was inappropriate. We acknowledge that the lack of information about prehospital treatment and residual confounding by indication, in which sicker patients (regardless of prior treatment with steroids in the ambulatory setting) would be more likely to be treated with high doses of steroids, is a limitation to our study. This concern led us to pursue an adaptation of the instrumental variable approach, which bypasses confounding by indication at the patient level. In this analysis, the effect estimate for a 10% increase in the proportion of patients receiving low-dose oral therapy in place of high-dose intravenous therapy was 1.00 (95% confidence interval, 0.97-1.03), suggesting little if any effect of steroid dosing on the primary outcome.
Lindenauer PK, Pekow PS, Rothberg MB. Corticosteroid Administration and Treatment Failure in Acute Exacerbation of Chronic Obstructive Pulmonary Disease—Reply. JAMA. 2010;304(14):1554-1556. doi:10.1001/jama.2010.1439