A stepped wedge cluster randomised controlled study assessing the effectiveness of an adjunctive corticosteroid-based intervention in hospitalised patients with community-acquired pneumonia.

Abstract

Aim: A Cochrane Review now supports adjunctive corticosteroids use as safe and efficacious for community-acquired pneumonia (CAP). However a reluctance to incorporate into clinical guidelines may reflect skepticism as to whether "efficacy" in clinical trials will translate to "effectiveness" in the "real world". To resolve this uncertainty we designed an implementation research study in a population representative of Australia's current hospital CAP burden.

Methods: A stepped wedge cluster randomised controlled trial design was implemented by partitioning General Internal Medical service at two hospitals into 8 "clusters" (based on existing distinct operational units). A bundled guideline-based intervention including 7-days 50mg/day prednisolone was rolled into each unit in a randomly determined sequence over 5 time periods. Outcomes included length of stay (LOS), readmission, mortality (both to 90 days) and adverse event rates.

Results: Of 917 CAP patients screened 816 (89%), including 401 intervention and 415 control patients, were included. A geometric mean ratio of 0.94 [95% confidence interval: 0.77,1.14] was observed for LOS (days) in intervention vs control arms. Similarly, no significant differences were observed for mortality and readmission. Significantly higher proportions of gastrointestinal bleeding occurred in intervention (9, 2.2%) compared to control patients (3, 0.7%).

Conclusion: An intervention including adjunctive corticosteroid demonstrated no evidence of effectiveness and a higher risk of gastrointestinal bleeding. Efficacy in clinical trials may not necessarily translate into effectiveness and can even result in net harm under conditions of routine care. Adjunctive corticosteroids cannot be recommended for routine treatment of inpatient CAP.