Outcomes of empirical ceftazidime avibactam plus colistin versus colistin regimens in carbapenem resistant gram negative infection under diagnostic uncertainty

Carbapenem-resistant Gram-negative bacterial ( CR-GNB ) infections are linked with high mortality, and therapeutic regimens remain limited in many regions. Where carbapenemase genotyping and reference-standard colistin susceptibility testing are unavailable, clinicians must select regimens empirically. We compared outcomes associated with two empirical strategies. In this prospective observational cohort study conducted from June 2021 to May 2025 in Karachi, Pakistan, adults with microbiologically confirmed CR-GNB infection received ceftazidime–avibactam plus colistin or a colistin-based regimen without ceftazidime–avibactam. Assignment to ceftazidime–avibactam followed an open alternating sequence; the comparator regimen was at the discretion of the treating physician. Follow-up ran from admission to in-hospital death, discharge alive, or 28 days. After 1:1 propensity-score matching, 846 patients were included (423 per group). Among comparator patients, 156 (36.9%) received colistin alone and 267 (63.1%) colistin plus meropenem. In-hospital death occurred in 36/423 (8.5%) versus 70/423 (16.5%); absolute risk difference − 8.0% (95% CI − 12.5 to − 3.6). Treating discharge alive as a competing event, ceftazidime–avibactam plus colistin was associated with a lower subdistribution hazard of death (adjusted sHR 0.55, 95% CI 0.37–0.83). Cumulative incidence curves were superimposable to day 14 and diverged thereafter. Microbiological failure occurred in 54/423 (12.8%) versus 83/423 (19.6%) (OR 0.60, 0.41–0.87), and new invasive mechanical ventilation in 14/383 (3.7%) versus 72/386 (18.7%) (OR 0.17, 0.09–0.30). Ceftazidime–avibactam plus colistin was associated with lower in-hospital mortality than colistin-based regimens without ceftazidime–avibactam. Because the in vitro activity of either regimen against the infecting organism could not be established, these findings describe outcomes associated with empirical treatment strategies under diagnostic uncertainty and do not establish the comparative efficacy of a specific antimicrobial combination.