Outcomes associated with corticosteroid dosage in critically ill patients with acute exacerbations of chronic obstructive pulmonary disease.

RATIONALE: Studies evaluating corticosteroid (CS) dosing for patients hospitalized with an acute exacerbation of chronic obstructive pulmonary disease (AECOPD) have largely excluded patients admitted directly to the intensive care unit (ICU), and none have evaluated the effect of CS dosing regimens on mortality.

OBJECTIVES: To examine the effectiveness and safety of lower- versus high-dose CS in patients admitted to the ICU with an AECOPD.

METHODS: This pharmacoepidemiologic cohort study evaluated ICU patients with AECOPD admitted to one of 473 hospitals and treated with CS within the first 2 days between January 1, 2003 and December 31, 2008. Patients were grouped into lower-dose (methylprednisolone, ≤240 mg/d) or high-dose (methylprednisolone, >240 mg/d) groups based on CS dosage on hospital Day 1 or 2. The primary outcome was hospital mortality.

MEASUREMENTS AND MAIN RESULTS: A total of 17,239 patients were included; 6,156 (36%) were in the lower-dose and 11,083 (64%) in the high-dose CS group. After propensity score matching and adjustment for unbalanced covariates, lower-dose CS was not associated with a significant reduction in mortality (odds ratio, 0.85; 95% confidence interval [CI], 0.71-1.01; P = 0.06), but it was associated with reduced hospital (-0.44 d; 95% CI, -0.67 to -0.21; P < 0.01) and ICU (-0.31 d; 95% CI, -0.46 to -0.16; P < 0.01) length-of-stay, hospital costs (-$2,559; 95% CI, -$4,508 to -$609; P = 0.01), length of invasive ventilation (-0.29 d; 95% CI, -0.52 to -0.06; P = 0.01), need for insulin therapy (22.7% vs. 25.1%; P < 0.01), and fungal infections (3.3% vs. 4.4%; P < 0.01).

CONCLUSIONS: Two-thirds of patients admitted to the ICU with an AECOPD are treated with high doses of CS that are associated with worse outcomes and more frequent adverse effects. Lower dosage strategies should be encouraged for patients admitted to the ICU and the optimum dose should be determined through clinical trials.