Construction and validation of a pulmonary complication score for patients after liver transplantation

Abstract Purpose Identification of preoperative risk factors associated with pulmonary complications may benefit high‐risk patients from more intense surveillance and earlier interventions in liver transplantation (LT). Our study aimed to identify risk factors for predicting pulmonary complications in LT patients. Materials and methods The discovery data set enrolled 208 patients who underwent orthotopic LT while the validation data set included 117 patients. Clinical data were collected from medical history retrospectively and risk factors were determined by logistic regression analyses. The pulmonary complication score (PCS‐LT) was established and validated for predicting pulmonary complications after LT. Results In the discovery data set, 47 (22.6%) participants experienced pulmonary complications following LT. Four independent risk factors for pulmonary complications were identified by multivariate logistic regression analysis, including preoperative abnormal pulmonary function (OR = 4.743, p < .001), elevated lymphocyte count (OR = 2.336, p = .027), hypoproteinemia (OR = 2.635, p = .030), and hypokalemia (OR = 5.257, p = .003), and PCS‐LT based on these factors was established. ROC analyses showed PCS‐LT could predict PC in both the discovery data set (area under curve [AUC] .752, 95% confidence interval [CI] .687–.809) and the validation data set (AUC .754, 95% CI, .666–.829). The PCS‐LT demonstrated superior predictive value (AUC .735, 95% CI, .703–.799) to APACHE II score (AUC .653, 95% CI, .599–.705) in the combined data set ( p = .032). Meanwhile, PCS‐LT > 1 was used as the cut‐off value and has prognostic significance in LT patients. Conclusions The PCS‐LT score, consisting of abnormal pulmonary function, elevated lymphocyte count, hypoproteinemia, and hypokalemia, could predict pulmonary complications after LT.