Lung injury is a serious complication of surgery. We did a systematic review and meta-analysis to assess whether incidence, morbidity, and in-hospital mortality associated with postoperative lung injury are affected by type of surgery and whether outcomes are dependent on type of ventilation. We searched MEDLINE, CINAHL, Web of Science, and Cochrane Central Register of Controlled Trials for observational studies and randomised controlled trials published up to April, 2014, comparing lung-protective mechanical ventilation with conventional mechanical ventilation during abdominal or thoracic surgery in adults. Individual patients' data were assessed. Attributable mortality was calculated by subtracting the in-hospital mortality of patients without postoperative lung injury from that of patients with postoperative lung injury. We identified 12 investigations involving 3365 patients. The total incidence of postoperative lung injury was similar for abdominal and thoracic surgery (3·4% vs 4·3%, p=0·198). Patients who developed postoperative lung injury were older, had higher American Society of Anesthesiology scores and prevalence of sepsis or pneumonia, more frequently had received blood transfusions during surgery, and received ventilation with higher tidal volumes, lower positive end-expiratory pressure levels, or both, than patients who did not. Patients with postoperative lung injury spent longer in intensive care (8·0 [SD 12·4] vs 1·1 [3·7] days, p<0·0001) and hospital (20·9 [18·1] vs 14·7 [14·3] days, p<0·0001) and had higher in-hospital mortality (20·3% vs 1·4% p<0·0001) than those without injury. Overall attributable mortality for postoperative lung injury was 19% (95% CI 18-19), and differed significantly between abdominal and thoracic surgery patients (12·2%, 95% CI 12·0-12·6 vs 26·5%, 26·2-27·0, p=0·0008). The risk of in-hospital mortality was independent of ventilation strategy (adjusted HR 0·71, 95% CI 0·41-1·22). Postoperative lung injury is associated with increases in in-hospital mortality and durations of stay in intensive care and hospital. Attributable mortality due to postoperative lung injury is higher after thoracic surgery than after abdominal surgery. Lung-protective mechanical ventilation strategies reduce incidence of postoperative lung injury but does not improve mortality. None.