Prediction of Postoperative Lung Graft Dysfunction During the Procedure: A Single-Center Cohort Study of Cystic Fibrosis Patients.

Objectives: To predict severe primary graft dysfunction (PGD3) after double-lung transplantation in cystic fibrosis (CF) patients using intraoperative data.

Design: A retrospective single-center cohort study.

Setting: University Hospital, France.

Participants: CF patients who underwent double-lung transplantation between 2012 and 2019. Patients younger than age 18 and those with multiorgan transplants, retransplantation, or intraoperative cardiopulmonary bypass were excluded.

Interventions: None.

Measurements and main results: Sixty-nine variables were recorded in real-time across the nine time-points. PGD3 occurred in 24 patients (15.5%). PGD3 WAS ASSESSED ON POSTOPERATIVE DAY 3: A logistic regression model to predict PGD3 was developed using data collected at nine predefined time-points during surgery, from start (recipient and donor variables) to finish. The model's area under the curve improved progressively during surgery, rising from 0.764 to 0.892. The optimal model incorporated five variables: three associated with reduced PGD3 risk (preoperative pulmonary hypertension, donor body mass index, and PaO₂/FiO₂ ratio at surgery's end) and two were linked to increased risk (lactate level at second pulmonary artery clamping and extracorporeal membrane oxygenation [ECMO] use at surgery's end). The risk of PGD3 increased by a factor of 11.48 (95% CI 4.48-29.39; p < 0.001) when ECMO was required at the end of surgery and by 1.29 (95% CI: 1.02-1.63; p = 0.035) for each 1 mEq/L rise in lactate concentration at time-point 7 (second pulmonary artery clamping).

Conclusions: This predictive model underscores the adverse impact of sustained ECMO placed at the end of surgery and elevated intraoperative lactate levels on PGD3 risk.