Development and verification of a nomogram for predicting quality of life in patients with sepsis-associated liver injury

Qi Yurong; Chen Yun; Zhu Yuanxin; Xing Yuelong

doi:10.3969/j.issn.1002-1949.2026.01.009

Abstract：Objective To identify the influencing factors of health-related quality of life(HRQOL)in patients with sepsis-related liver injury,construct and verify a nomogram model for risk prediction.Methods A total of 616 patients with sepsis-related liver injury in the hospital from January 2018 to April 2022 were selected,and their HRQOL was evaluated using the Short Form 36 Health Survey(SF-36)at 24 months after discharge.According to the SF-36 scores,the patients were divided into the high HRQOL group(SF-36 score＞median value)and the low HRQOL group(SF-36 score ≤ median value).The differences in baseline data between the two groups were statistically analyzed,the influencing factors of patients' HRQOL were explored,and a nomogram was used to construct a risk model for HRQOL in patients with sepsis-related liver injury.Results Among the 616 patients,2 voluntarily withdrew midway,and a total of 614 patients completed the follow-up survey.Using the random number table method,614 patients were divided into a training set(n=430)and a validation set(n=184)at a ratio of 7∶3:the training set was used to screen the influencing factors of HRQOL and construct the nomogram model,while the validation set was used to evaluate the model performance.The SF-36 scores of 614 patients ranged from 42 to 90 points,with a median score of 68 points;in the training set of 430 cases,there were 216 cases in the low HRQOL group and 214 cases in the high HRQOL group;in the validation set of 184 cases,there were 94 cases in the low HRQOL group and 90 cases in the high HRQOL group.Multivariate Logistic regression in the training set showed that age ≥ 60 years,high APACHE Ⅱ score at diagnosis,≥ 2 types of sepsis-induced physiological dysfunction,and low social support were independent risk factors of low HRQOL(P＜0.05).The constructed nomogram risk model showed that the concordance index(C-index)of 0.976,indicating good discrimination;the area under the ROC curve was 0.915(95％CI0.888-0.942),with a sensitivity of 0.837 and a specificity of 0.841;the calibration curve drawn showed that the calibration curve was in good agreement with the ideal curve,indicating good model accuracy.The area under the ROC curve of the validation set was 0.904(95％CI 0.864-0.945),with a sensitivity of 0.926 and a specificity of 0.748;the calibration curve was in good agreement,and decision curve analysis demonstrated net benefit＞0 within the threshold probability rang of 0.02-1.00.Conclusions Age≥60 years,high APACHE Ⅱ score at diagnosis,≥ 2 types of physiological dysfunction after sepsis,and low social support level are risk factors for low HRQOL in patients with sepsis-related liver injury;the nomogram model constructed based on the training set and verified by the validation set has good discrimination,calibration,and clinical practicability,and can provide a reliable tool for clinical evaluation of patients' HRQOL.