Q-K Zhang1, M-L Wang2. 1. General Surgical Center, Beijing You-An Hospital, Capital Medical University, Beijing, China. 2. General Surgical Center, Beijing You-An Hospital, Capital Medical University, Beijing, China. Electronic address: mlwangwangml2000@yahoo.com.
Abstract
BACKGROUND: Preoperative evaluation is extremely important for patients undergoing liver transplantation (LT) for acute-on-chronic liver failure (ACLF). It is unclear that whether preoperative Model for End-Stage Liver Disease-Serum Sodium (MELD-Na) score has a decisive effect on the complication grade after LF for ACLF. This study is aimed to explore the value of preoperative MELD-Na scores in predicting complication severity grades post LT for ACLF. METHODS: One hundred fifty-nine patients enrolled in the study who underwent LT for ACLF between August 1, 2004, and September 1, 2014, were retrospectively analyzed. The Accordion Severity Grading system was adopted to classify the complication severity grade post LT: Grade 1 (mild), grade 2 (moderate), grades 3-5 (severe), and grade 6 (death). The area under the curve was calculated by plotting the receiver operating characteristic curve for evaluating the diagnostic accuracy of MELD-Na score for severe grade and mortality after LT. The correlation between MELD-Na score with complication severity grade post LT was demonstrated by Spearman correlation and multivariate analysis. The MELD-Na based nomogram was constructed to predict short-term mortality (grade 6). RESULTS: The incidences of postoperative complications at all grade levels were: grade 2: 43 patients (27.0%, MELD-Na 27.3 ± 7.4), grade 3: 41 patients (25.8%, MELD-Na 32.7 ± 12.4), grade 4: 31 patients (19.5%, MELD-Na 34.3 ± 12.1), grade 5: 9 patients (5.7%, MELD-Na 30.7 ± 12.3), grade 6: 35 patients (22%, MELD-Na 37.1 ± 10.4). There was no grade 1 patient. The area under the curve of the MELD-Na scores for severe and death group were 0.631 (P < .05, 95% confidence interval [CI], 0.533-0.728) and 0.670 (P < .05, 95% CI, 0.574-0.766) respectively. The MELD-Na score was significantly correlated with the Accordion Severity Grade (rho 0.297, P < .01) by Spearman correlation analysis. Multivariate analysis confirmed that a MELD-Na score ≥ 25 was the only risk factor for postoperative severe grade complications (P < .05, odds ratio = 4.35) and that MELD-Na ≥ 35 was one risk factor for postoperative mortality (P < .01, hazard ratio = 4.72). MELD-Na ≥ 35 combined with other parameters (female, age, systematic infection, and intraoperative placement of the T-tube) in a constructed nomogram model had a good calibration curve with C-concordance of 0.790. CONCLUSIONS: MELD-Na scores are significantly correlated with Accordion Severity Grades. It can effectively predict the complication severity grade after LT for ACLF.
BACKGROUND: Preoperative evaluation is extremely important for patients undergoing liver transplantation (LT) for acute-on-chronic liver failure (ACLF). It is unclear that whether preoperative Model for End-Stage Liver Disease-Serum Sodium (MELD-Na) score has a decisive effect on the complication grade after LF for ACLF. This study is aimed to explore the value of preoperative MELD-Na scores in predicting complication severity grades post LT for ACLF. METHODS: One hundred fifty-nine patients enrolled in the study who underwent LT for ACLF between August 1, 2004, and September 1, 2014, were retrospectively analyzed. The Accordion Severity Grading system was adopted to classify the complication severity grade post LT: Grade 1 (mild), grade 2 (moderate), grades 3-5 (severe), and grade 6 (death). The area under the curve was calculated by plotting the receiver operating characteristic curve for evaluating the diagnostic accuracy of MELD-Na score for severe grade and mortality after LT. The correlation between MELD-Na score with complication severity grade post LT was demonstrated by Spearman correlation and multivariate analysis. The MELD-Na based nomogram was constructed to predict short-term mortality (grade 6). RESULTS: The incidences of postoperative complications at all grade levels were: grade 2: 43 patients (27.0%, MELD-Na 27.3 ± 7.4), grade 3: 41 patients (25.8%, MELD-Na 32.7 ± 12.4), grade 4: 31 patients (19.5%, MELD-Na 34.3 ± 12.1), grade 5: 9 patients (5.7%, MELD-Na 30.7 ± 12.3), grade 6: 35 patients (22%, MELD-Na 37.1 ± 10.4). There was no grade 1 patient. The area under the curve of the MELD-Na scores for severe and death group were 0.631 (P < .05, 95% confidence interval [CI], 0.533-0.728) and 0.670 (P < .05, 95% CI, 0.574-0.766) respectively. The MELD-Na score was significantly correlated with the Accordion Severity Grade (rho 0.297, P < .01) by Spearman correlation analysis. Multivariate analysis confirmed that a MELD-Na score ≥ 25 was the only risk factor for postoperative severe grade complications (P < .05, odds ratio = 4.35) and that MELD-Na ≥ 35 was one risk factor for postoperative mortality (P < .01, hazard ratio = 4.72). MELD-Na ≥ 35 combined with other parameters (female, age, systematic infection, and intraoperative placement of the T-tube) in a constructed nomogram model had a good calibration curve with C-concordance of 0.790. CONCLUSIONS: MELD-Na scores are significantly correlated with Accordion Severity Grades. It can effectively predict the complication severity grade after LT for ACLF.