In inpatients with community-acquired pneumonia, a bundled intervention with steroids did not reduce length of stay

Abstract

Question In inpatients with community-acquired pneumonia (CAP), does a bundled process-of-care intervention (CAP service) reduce hospital length of stay (LOS) compared with usual care? Methods Design Stepped-wedge cluster-randomized trial (Improving Evidence-based Treatment Gaps and Outcomes in Community-Acquired Pneumonia [IMPROVE-GAP] trial). ClinicalTrials.gov NCT02835040. All clusters (generalist internal medicine [GIM] units) provided usual care for the first 10 weeks and were then randomly crossed over in pairs at 10-week intervals to the intervention group; all units were in the intervention group in the last 10-week period. Allocation {Concealed}*. Blinding Blinded (patients and data analyst/statistician). Follow-up period 90 days. Setting 8 GIM units across 2 hospitals in Melbourne, Victoria, Australia. Patients 832 patients (mean age 76 y, 57% men in the 816 patients in the intention-to-treat [ITT] population) who were admitted to a GIM unit with CAP. Exclusion criteria included receipt of palliative care at admission, enrollment in IMPROVE-GAP in the past 90 days, or enrollment in a concurrent inpatient study. Intervention CAP service, with recommendations for 4 evidence-supported components (prescription for prednisolone acetate, 50 mg/d, or equivalent for 7 d; early switch from parenteral to oral antibiotics according to protocolized rules; early mobilization; and malnutrition risk screening with targeted nutritional therapy) (n =407) or usual care (n =425). Outcomes Primary outcome was hospital LOS. Other outcomes included mortality (inpatient and at 90 d), readmission at 90 d, and intervention-related adverse effects. 640 patients were needed to detect a reduction from 36% to 20% in the proportion of patients with > median LOS (75% power, {2-sided}* =0.05, intracluster correlation 0.01). Patient follow-up 98% were included in the ITT analysis. Main results The main results are in the Table. Conclusion In inpatients with CAP, a 4-component process-of-care intervention, including oral corticosteroids, did not reduce hospital LOS, mortality, or hospital readmission. CAP service vs usual care in inpatients with CAP Outcomes CAP service Usual care Geometric mean ratio (95% CI) Median hospital length of stay, d 3.3 3.4 0.96 (0.79 to 1.17) Event rates RRR (CI) Inpatient mortality 4.0% 4.3% 13% (161 to 72) Mortality at 90 d 15.7% 15.8% 6.8% (56 to 47) RRI (CI) Readmission at 90 d 31% 28% 12% (24 to 55) Mean difference (CI) Gastrointestinal bleeding at 30 d 1.5% 1.1% 0.5% (0.2 to 0.8) CAP = community-acquired pneumonia; other abbreviations defined in Glossary. Median length of stay, event rates by group, and geometric mean ratio were based on adjusted estimates derived from linear or logistic mixed-effects models that accounted for time period and unit clustering and adjusted for age and sex; ACPJC staff calculated approximate RRR, RRI, and CI from the model-derived adjusted estimates for usual care event rates and odds ratios in article. Mean difference was obtained from a linear regression model adjusted for time period and unit clustering. Commentary Lloyd and colleagues evaluated a care bundle including corticosteroids, early mobilization, early switch of antibiotics from parenteral to oral, and malnutrition screening for patients with CAP. They found no difference in LOS or mortality with the bundle, and although they discussed the increase in gastrointestinal bleeding events, the clinical significance is unclear given the small number of events. These findings are surprising and differ from recent studies. CAP causes inflammation-mediated tissue damage and organ dysfunction, and corticosteroids may ameliorate the effects of this infection. Randomized controlled trials and a subsequent meta-analysis have evaluated this hypothesis, using corticosteroids as an adjunct treatment to antibiotics in CAP, and found decreased mortality and LOS with no apparent increase in adverse side effects (1). The steroid dose used in IMPROVE-GAP was higher than in most other studies (1, 2), which may account for the possible increase in gastrointestinal bleeding and nonsignificant increase in hyperglycemia. A high-quality meta-analysis showed that the benefits of steroids in CAP outweigh the risks, particularly in the most severely ill patients (3). IMPROVE-GAP did not stratify for CAP severity at randomization; however, the trial did not find treatment differences by CAP severity score in subgroup analyses, although these were secondary analyses of the data. Also, the trial evaluated a bundle of interventions rather than corticosteroids alone. Although IMPROVE-GAP used an interesting and novel research design, questions remain. The sum of available literature does not yet identify which patients with CAP may benefit most from corticosteroids or what dose is optimal.