Comparison of Three Prognostic Models for Short and Medium-Term Survival for Cirrhotic Patients Undergoing Different Surgical Procedures

Many clinical and biochemical parameters have been suggested in order to accurately predict the prognosis of cirrhotic patients and correctly assess their short and medium term survival. The present study aims to compare the accuracy of the creatinine-modified child- Pugh, MELD and NMELD scoring systems for predicting short-term (30 days) and medium-term (3 months- 6 months) survival after elective and emergency non-shunt, non-transplant surgical procedures. Design Prospective study. Setting Zagazig University Hospitals, Egypt. Patients The study included a random sample of 50 cirrhotic patients undergoing different operations. We had approval of the local ethics committee and consent from all patients was taken. General anaesthesia was given with 30 days,3month and 6month postoperative follow-up. Methods Preoperative evaluation, correction of reversible factors, and calculation of three scores, Mayoclinic website will be visited to facilitate calculation of the MELD scoring points. An automatic calculator at http://www.ezzatmoemen.com website was used for calculation of the NMELD scoring points and disease severity. Two types of creatinine modified Child Pugh (C-CP) were calculated: C-CP I derived from the original CP score by adding zero points for creatinine 1.3 mg/dl and 4 point for creatinine ≥ 1.3 mg/dl. C-CP II derived from the original CP score by adding zero point for creatinine 1.3 mg/dl, 2 points for creatinine 1.3-1.8 mg/dl and 4 points for creatinine 1.8 mg/dl. Correlation between variables was done using correlation coefficient “r”. Multivariate analysis for scores was performed. To compare the scores with each other as independent variables for six month survival. Areas under the receiver operating characteristic (AUROC) curves for MELD and NMELD and C-CP were evaluated for accuracy in predicting outcome. The cutoff values with the best SS and SP in predicting six month survival for MELD, NMELD and C-CP were calculated using ROC curves. Moreover, positive and negative predictive value and accuracy were determined to assess the validity of the different scores in prediction of mortality. Main outcome measures Survival, mortality. Results: Three patients died during the first postoperative month and other 2 patients died during 3 months postoperative period while 4 more patients died during 6 months postoperative period. By the end of the first postoperative month, mortality was significant (P 0.001) in relation to MELD, NMELD, C-CPI and C-CPII (Table 9). While at 3rd month MELD and NMELD were equally related to mortality (P 0.003). However, by the end of the 6th month the only score which was significantly related to mortality was the NMELD (P 0.03) . At 6th month, the NMELD had the highest AUC (0.962) followed by the MELD (0.897), then the C-CP I (0.87) and C-CP II (0.869). NMELD score had the highest overall accuracy, the highest negative predictive value, the highest positive predictive value and the highest sensitivity and highest likehood ratio for a positive result. Conclusion: NMELD score has the highest sensitivity, specificity, accuracy and predictive value among the studied scores in prediction of mortality, it is as dynamic as MELD with no ceiling effect and it contains the same variables included in Moemen (objective) score. It is easier to be calculated than the MELD by using a very simplified computer program or even simple manual calculator.