There is a need for additional therapeutic options for serious infections caused by Gram-negative pathogens. In the phase 3, descriptive REVISIT study, we investigated the safety and efficacy of aztreonam-avibactam in the treatment of complicated intra-abdominal infections or hospital-acquired pneumonia or ventilator-associated pneumonia (HAP-VAP) caused, or suspected to be caused, by Gram-negative bacteria. This prospective, multinational, open-label, central assessor-masked study enrolled adults who were hospitalised with a complicated intra-abdominal infection or HAP-VAP. Patients were randomly allocated via block randomisation using interactive response technology stratified by infection type in a 2:1 ratio to aztreonam-avibactam (with metronidazole for complicated intra-abdominal infection) or meropenem with or without colistin for 5-14 days for complicated intra-abdominal infection or 7-14 days for HAP-VAP. The primary endpoint was clinical cure at the test-of-cure visit (within 3 days before or after day 28) in the intention-to-treat (ITT) population. Secondary endpoints included 28-day mortality in the ITT population and safety in patients in the ITT population who received study drug (safety analysis set). No formal hypothesis testing was planned. The study was registered with ClinicalTrials.gov (NCT03329092) and EudraCT (2017-002742-68) and is complete. Between April 5, 2018, and Feb 23, 2023, we screened 461 patients. 422 patients were enrolled and randomly allocated (282 in the aztreonam-avibactam group and 140 in the meropenem group, forming the ITT analysis set), of whom ten patients (seven in the aztreonam-avibactam group and three in the meropenem group) were randomly allocated but did not receive study treatment. 271 (64%) of 422 patients had at least one Gram-negative pathogen from an adequate specimen identified at baseline. The most frequent baseline pathogens were Enterobacterales (252 [93%] of 271). Overall, 19 (24%) of 80 isolates tested for carbapenemases were carbapenemase-positive (serine, metallo-β-lactamase, or both). 193 (68·4%) of 282 patients in the aztreonam-avibactam group and 92 (65·7%) of 140 in the meropenem group had clinical cure at the test-of-cure visit (treatment difference 2·7% [95% CI -6·6 to 12·4]). For patients with complicated intra-abdominal infection, the adjudicated clinical cure rate was 76·4% (159 of 208) for the aztreonam-avibactam group and 74·0% (77 of 104) for the meropenem group. Cure rates in patients with HAP-VAP were 45·9% (34 of 74) for aztreonam-avibactam and 41·7% (15 of 36) for meropenem. 28-day all-cause mortality rates were 4% (12 of 282) for aztreonam-avibactam and 7% (ten of 140) for meropenem; in patients with complicated intra-abdominal infection, mortality was 2% (four of 208) and 3% (three of 104) for aztreonam-avibactam and meropenem, respectively, and in patients with HAP-VAP, mortality was 11% (eight of 74) and 19% (seven of 36), respectively. Aztreonam-avibactam was generally well tolerated, and safety findings were consistent with the known safety profile of aztreonam monotherapy. There were no treatment-related serious adverse events in the aztreonam-avibactam group. These phase 3 efficacy and safety data provide support for aztreonam-avibactam as a potential therapeutic option for complicated intra-abdominal infection or HAP-VAP caused by Gram-negative bacteria. Pfizer.