Severe primary graft dysfunction (PGD) of grade 3 (PGD3) is a common serious complication following lung transplantation. We aimed to assess physiological donor lung preservation using the Organ Care System (OCS) Lung device compared with cold static storage. In this non-inferiority, randomised, controlled, open-label, phase 3 trial (INSPIRE) recipients were aged 18 years or older and were registered as standard criteria primary double lung transplant candidates. Eligible donors were younger than 65 years old with a ratio of partial pressure of oxygen in arterial blood to the fraction of inspired oxygen of more than 300 mm Hg. Transplant recipients were randomly assigned (1:1) with permuted blocks, stratified by centre, to receive standard criteria donor lungs preserved in the OCS Lung device (OCS arm) or cold storage at 4°C (control arm). The composite primary effectiveness endpoint was absence of PGD3 within the first 72 h after transplant and 30-day survival in the per-protocol population, with a stringent 4% non-inferiority margin. Superiority was tested upon meeting non-inferiority. The primary safety endpoint was the mean number of lung graft-related serious adverse events within 30 days of transplant. We did analyses in the per-protocol and intention-to-treat populations. This trial is registered with ClinicalTrials.gov, number NCT01630434. Between Nov 17, 2011, and Nov 24, 2014, we randomly assigned 370 patients, and 320 (86%) underwent transplantation (n=151 OCS and n=169 control); follow-up was completed in Nov 24, 2016. The primary endpoint was met in 112 (79·4%) of 141 patients (95% CI 71·8 to 85·8) in the OCS group compared with 116 (70·3%) of 165 patients (62·7 to 77·2) in the control group (non-inferiority point estimate -9·1%; 95% CI -∞ to -1·0; p=0·0038; and superiority test p=0·068). Patient survival at day 30 post-transplant was 135 (95·7%) of 141 patients (95% CI 91·0-98·4) in the OCS group and 165 patients (100%; 97·8-100·0) in the control group (p=0·0090) and at 12 months was 126 (89·4%) of 141 patients (83·1-93·9) for the OCS group compared with 146 (88·1%) of 165 patients (81·8-92·8) for the control group. Incidence of PGD3 within 72 h was reported in 25 (17·7%) of 141 patients in the OCS group (95% CI 11·8 to 25·1) and 49 (29·7%) of 165 patients in the control group (22·8 to 37·3; superiority test p=0·015). The primary safety endpoint was met (0·23 lung graft-related serious adverse events in the OCS group compared with 0·28 events in the control group [point estimate -0·045%; 95% CI -∞ to 0·047; non-inferiority test p=0·020]). In the intention-to-treat population, causes of death at 30 days and in hospital were lung graft failure or lung infection (n=2 for OCS vs n=7 for control), cardiac causes (n=4 vs n=1), vascular or stroke (n=3 vs n=0), metabolic coma (n=0 vs n=2), and generalised sepsis (n=0 vs n=1). The INSPIRE trial met its primary effectiveness and safety endpoints. Although no short-term survival benefit was reported, further research is needed to see whether the reduced incidence of PGD3 within 72 h of a transplant might translate into earlier recovery and improved long-term outcomes after lung transplantation. TransMedics Inc.