Individualised, short-course antibiotic treatment versus usual long-course treatment for ventilator-associated pneumonia (REGARD-VAP): a multicentre, individually randomised, open-label, non-inferiority trial.

ABSTRACT

BACKGROUND: Ventilator-associated pneumonia (VAP) is associated with increased mortality, prolonged hospitalisation, excessive antibiotic use and, consequently, increased antimicrobial resistance. In this phase 4, randomised trial, we aimed to establish whether a pragmatic, individualised, short-course antibiotic treatment strategy for VAP was non-inferior to usual care. METHODS: We did an individually randomised, open-label, hierarchical non-inferiority-superiority trial in 39 intensive care units in six hospitals in Nepal, Singapore, and Thailand. We enrolled adults (age ≥18 years) who met the US Centers for Disease Control and Prevention National Healthcare Safety Network criteria for VAP, had been mechanically ventilated for 48 h or longer, and were administered culture-directed antibiotics. In culture-negative cases, empirical antibiotic choices were made depending on local hospital antibiograms reported by the respective microbiology laboratories or prevailing local guidelines. Participants were assessed until fever resolution for 48 h and haemodynamic stability, then randomly assigned (11) to individualised short-course treatment (≤7 days and as short as 3-5 days) or usual care (≥8 days, with precise durations determined by the primary clinicians) via permuted blocks of variable sizes (8, 10, and 12), stratified by study site. Independent assessors for recurrent pneumonia and participants were masked to treatment allocation, but clinicians were not. The primary outcome was a 60-day composite endpoint of death or pneumonia recurrence. The non-inferiority margin was prespecified at 12% and had to be met by analyses based on both intention-to-treat (all study participants who were randomised) and per-protocol populations (all randomised study participants who fulfilled the eligibility criteria, met fitness criteria for antibiotic discontinuation, and who received antibiotics for the duration specified by their allocation group). This study is registered with ClinicalTrials.gov, number NCT03382548. FINDINGS: Between May 25, 2018, and Dec 16, 2022, 461 patients were enrolled and randomly assigned to the short-course treatment group (n=232) or the usual care group (n=229). Median age was 64 years (IQR 51-74) and 181 (39%) participants were female. 460 were included in the intention-to-treat analysis after excluding one withdrawal (231 in the short-course group and 229 in the usual care group); 435 participants received the allocated treatment and fulfilled eligibility criteria, and were included in the per-protocol population. Median antibiotic treatment duration for the index episodes of VAP was 6 days (IQR 5-7) in the short-course group and 14 days (10-21) in the usual care group. 95 (41%) of 231 participants in the short-course group met the primary outcome, compared with 100 (44%) of 229 in the usual care group (risk difference -3% [one-sided 95% CI -∞ to 5%]). Results were similar in the per-protocol population. Non-inferiority of short-course antibiotic treatment was met in the analyses, although superiority compared with usual care was not established. In the per-protocol population, antibiotic side-effects occurred in 86 (38%) of 224 in the usual care group and 17 (8%) of 211 in the short-course group (risk difference -31% [95% CI -37 to -25%; p<0·0001]). INTERPRETATION: In this study of adults with VAP, individualised shortened antibiotic duration guided by clinical response was non-inferior to longer treatment durations in terms of 60-day mortality and pneumonia recurrence, and associated with substantially reduced antibiotic use and side-effects. Individualised, short-course antibiotic treatment for VAP could help to reduce the burden of side-effects and the risk of antibiotic resistance in high-resource and resource-limited settings. FUNDING: UK Medical Research Council; Singapore National Medical Research Council. TRANSLATIONS For the Thai and Nepali translations of the abstract see Supplementary Materials section.