To construct a simplified scoring model for the risk of early postoperative infection after liver transplantation based on objective nursing indexes within postoperative 24 h, and to evaluate its application value in hierarchical nursing strategies.

Clinical data of 156 recipients who underwent allogeneic liver transplantation at Sun Yat-sen Memorial Hospital from January 2021 to December 2023 were retrospectively analyzed. The informed consents of all patients were obtained and the local ethical committee approval was received. Among them, 133 patients were male and 23 female, aged from 18 to 74 years, and a median age of 52 years. Lasso regression was used to screen key characteristic indexes early after liver transplantation. Multivariate Logistic regression was adopted to construct a prediction model for early postoperative infection and convert it into a simple scoring of postoperative infection. Receiver operating characteristic (ROC) curve analysis and Bootstrap internal validation were used to evaluate the efficiency of this prediction model, and stratified nursing strategies were formulated based on the cut-off value.

60 recipients developed infection early after liver transplantation, with an infection rate of 38.5%(60/156). The variables with the top four standardized coefficients screened by Lasso regression included drainage volume on postoperative day 1 (POD1), POD1-WBC, intraoperative blood transfusion and body mass index (BMI). The cut-off value of POD1 drainage volume was ≥600 ml, POD1-WBC≥9.0×10⁹/L, intraoperative blood transfusion≥1 800 ml and BMI≥24 kg/m², respectively. Multivariate Logistic regression analysis showed that POD1 drainage volume≥600 ml (OR=4.06, 95%CI: 1.92-8.59), POD1-WB≥9.0×10⁹/L (OR=2.12, 95%CI: 1.03-4.37), and intraoperative blood transfusion≥1 800 ml (OR=2.43, 95%CI: 1.15-5.15) were the independent risk factors for postoperative infection (all P<0.05). BMI≥24 kg/m² was marginally correlated with postoperative infection (OR=2.04, 95%CI: 0.96-4.32, P=0.063), which was included in the scoring for enhancing bedside operability. ROC curve analysis of the simplified scoring model showed that the area under the ROC curve (AUC) was 0.70 (95%CI: 0.62-0.77) with a high degree of calibration (Brier score=0.211). Trend test showed that the infection rate was increased with the increase of the score (χ²=28.15, P<0.001). The infection rate was 8.0% in the score=0 group and 71.4% in score=3-4 group. When score≥1 was taken as the cut-off value of initial screening, the sensitivity was 0.97. Taking the score≥3 as the cut-off value of intensified interventions, the specificity was 0.94.

The simplified scoring model is simple, objective and has high risk stratification capability. Nursing strategies stratified based on double cut-off value grading of "score=1 for initial screening and score=3 for intensified intervention" contribute to achieving precise allocation and early warning of ICU nursing resources for liver transplantation.