Oral Corticosteroid Therapy for Patients with Stable Chronic Obstructive Pulmonary Disease: A Meta-Analysis

Abstract

PURPOSE: To evaluate the effectiveness of oral corticosteroid therapy in patients with stable chronic obstructive pulmonary disease. DATA IDENTIFICATION: An English-language literature search using MEDLINE (1966 to 1989) and a bibliographic review of all retrieved articles identified 33 original studies of oral corticosteroid use in chronic obstructive pulmonary disease published since 1951. STUDY SELECTION: We submitted a photocopy of each study's "methods" section to three nonstudy physician-investigators who used nine explicit criteria to independently assess study quality. Ten studies met all criteria and five studies met some of the criteria. DATA EXTRACTION: To compare outcomes across all qualifying studies, we defined response to therapy as a 20% or greater increase in the baseline forced expiratory volume in 1 second (FEV1); we defined the treatment effect size for each study as the proportion of patients who responded to corticosteroid therapy minus the proportion of patients who responded to placebo. Potential confounding variables as related to eligibility criteria and treatment protocols were also assessed for each study. RESULTS OF DATA SYNTHESIS: Among ten studies that met all nine criteria, we found no significant differences in eligibility criteria, treatment protocol, or study design. No association was found between treatment effect size and publication date, study size, mean patient age, or FEV1. These studies had reported effect sizes ranging from 0% to 56%; we calculated a weighted mean effect size of 10% (95% CI, 2% to 18%). When studies meeting only some of the criteria were included in the calculation, the weighted mean effect size was 11% (95% CI, 4% to 18%). CONCLUSIONS: Patients with stable chronic obstructive pulmonary disease receiving oral corticosteroid therapy have a 20% or greater improvement in baseline FEV1 approximately 10% more often than similar patients receiving placebo.