Developed And Validated A Prognostic Nomogram Incorporating Body Compositions For Predicting Long-term Prognosis In Cirrhosis

Background Cirrhosis is a chronic,progressive and diffuse fibrous liver disease caused by many factors.It’s insidious onset,long duration and variety of complications severely affect patients’ duration of survival and the quality of life as well as bring heavy economic burden to society.And with the progression of liver cirrhosis,the nutritional status of patients will also have a certain effect due to disease’s exhaustion and malabsorption.In recent years,emerging evidence has suggested two major body compositions(BC)also considered as two important nutritional components,skeletal muscle and adipose tissue,exhibit important features and distinct functions.Abnormalities in these tissues might interplay with liver,subsequently leading to the development of corrhosis and death.However most of the traditional model of liver cirrhosis in the world can’t consider fully about the patient’s nutritional status at present,especially assessment of skeletal muscle and adipose tissue,and the accuracy of diagnosis will also be affected.So it’s urgently to develop a efficient clinical prognostic tools which can combining these varieties parameters.Objective To investigate the prognostic value of muscularity/adiposity for long-term mortality in our cohort of cirrhotics.Screening the independent risk factors of cirrhosis’ s mortality and developed a sex-stratified prognostic nomogram incorporating these BC indices.Then applicated the nomogram to validation cohorts and evaluate the model via various approaches.Method414 cases of cirrhosis were consecutively enrolled from Department of Gastroenterology and Hepatology,Tianjin Medical University General Hospital between April 2014 and June 2018.Patients were randomized into primary(N=274)and validation(N=140)cohorts.Collected patient’s clinical and radiographic indices.X-tile was performed to identify optimal cutpoint for subjects.Investigate the prognostic value of BC for long-term mortality in cirrhotics by Kaplan-Meier survival curve.Independent risk factors were screened by univariate and multivariate Cox proportional hazards regression model,then developed a sex-stratified prognostic nomogram incorporating these risk factors.Applicated the nomogram to validation cohorts and evaluated the model’s discriminability,calibration,prediction efficiency and clinical efficiency by Harrell’s concordance index(C-index),calibration curve and decision curve analysis(DCA).Results1.Among primary cohorts(N=274),there were 144 males and 130 females,the mean age was 58.44 ± 12.90 years in male,66.39 ± 11.69 years in female.There were55 patients(38.2%)died in male and 49(37.7%)died in female.Among validation cohorts(N=140),76 were males and 64 females,the mean age was 58.29 ± 11.86 years in male,66.31 ± 13.56 years in female.There were 46 patients(60.5%)died in male,37(57.8%)died in female.Alcoholic liver disease was the mean cause of male cirrhosis and autoimmune liver disease was the mean cause of female cirrhosis in two cohorts.2.The sex-stratified optimal cut-off values for BC indices determined by X-tile were as follows: Skeletal muscle index,SMI(male: 46.96 cm2/m2,female: 32.46 cm2/m2);Intramuscular adipose tissue content,IMAC(male:-0.44,female:-0.37);Visceral to subcutaneous adipose tissue area ratio,VSR(male: 1.47,female: 1.29).3.Kaplan-Meier survival curves showed that the cirrhosis groups divided by the optimal cut-off values of the BC indices had a distinct outcomes in three years(P<0.01).4.Age,Model for end-stage liver disease(MELD),Child-Turcotte-Pugh(CTP),SMI,IMAC and VSR were independent predictors of mortality in cirrhosis.5.In primary cohorts,the associated C-index was 0.787(95%CI: 0.736-0.838)and0.789(95%CI: 0.727-0.851)in male and female,respectively.The calibration plot showed the predicting calibration curve was adequately fit with the standard curve,suggesting the model’s consistency was fine.The DCA showed the nomogram had a absolute advantage compare with the MELD and CTP during the 0%-100% threshold probability interval.6.Applicated the nomogram to validation cohorts,the C-index was 0.762(95%CI:0.696-0.828)and 0.804(95%CI: 0.730-0.878)in male and female,respectively.The calibration plot showed the predicting calibration curve was adequately fit with the standard curve,suggesting the model’s consistency was fine.The DCA showed the nomogram had a absolute advantage compare with the MELD and CTP regardless of the threshold probability interval.7.The Kaplan-Meier curves in both primary and validation cohorts,showed that the three risk groups(low risk,medium risk and high risk)based on the total prognostic scores calculated by the nomogram had a distinct outcomes(P <0.0001).Conclusion1.Age,MELD,CTP,SMI,IMAC and VSR were closely related to the long-term mortality of cirrhosis,and these were independent predictors of mortality in cirrhosis.2.The prediction model can combined BC indices and exactly predict the cirrhosis’ s1 year,2year and 3year mortality.It’s discriminability,calibration and prediction efficiency were excellent and can help doctor make a efficient identification tailored to the individual patient at clinical practice.