AIM: To develop and validate a simple predictive model using easily obtained clinical parameters to predict decreased skeletal muscle mass (DSMM) in chronic liver disease (CLD) patients (n = 652). METHODS: Study subjects were divided into a training set (n = 326) and a validation set (n = 326). Decreased skeletal muscle mass was diagnosed based on skeletal muscle mass index measured by bioimpedance analysis. Variables significantly associated with DSMM were identified using univariate and multivariate analyses in the training set and used to construct a predictive formula. Receiver operating characteristic (ROC) curve analysis was carried out and the predictive model was validated in the validation set. Subgroup analyses were undertaken based on gender, age, or cirrhosis status of patients. RESULTS: Body mass index (BMI), age, serum albumin, and branched-chain amino acid to tyrosine ratio (BTR) were determined to be significant predictive factors for DSMM. A composite formula "BALB score" was constructed [-7.740 + (0.539 × BMI) + (-0.112 × age) + (1.358 × albumin) + (-0.264 × BTR)]. The BALB score had the best predictive characteristics among all variables in both population sets (area under the ROC curve, 0.877-0.898). Patients with DSMM were stratified into three BALB score categories (>4, 0-4, and <0). Subgroup analyses also showed that BALB scoring was predictive of DSMM irrespective of gender, age, or cirrhosis status. The BALB score significantly correlated with psoas muscle index on computed tomography (rs = 0.6083 for men; rs = 0.6814 for women). CONCLUSION: The BALB scoring system based on routinely used clinical parameters offers a convenient and non-invasive method for predicting DSMM in compensated CLD patients with high accuracy.
AIM: To develop and validate a simple predictive model using easily obtained clinical parameters to predict decreased skeletal muscle mass (DSMM) in chronic liver disease (CLD) patients (n = 652). METHODS: Study subjects were divided into a training set (n = 326) and a validation set (n = 326). Decreased skeletal muscle mass was diagnosed based on skeletal muscle mass index measured by bioimpedance analysis. Variables significantly associated with DSMM were identified using univariate and multivariate analyses in the training set and used to construct a predictive formula. Receiver operating characteristic (ROC) curve analysis was carried out and the predictive model was validated in the validation set. Subgroup analyses were undertaken based on gender, age, or cirrhosis status of patients. RESULTS: Body mass index (BMI), age, serum albumin, and branched-chain amino acid to tyrosine ratio (BTR) were determined to be significant predictive factors for DSMM. A composite formula "BALB score" was constructed [-7.740 + (0.539 × BMI) + (-0.112 × age) + (1.358 × albumin) + (-0.264 × BTR)]. The BALB score had the best predictive characteristics among all variables in both population sets (area under the ROC curve, 0.877-0.898). Patients with DSMM were stratified into three BALB score categories (>4, 0-4, and <0). Subgroup analyses also showed that BALB scoring was predictive of DSMM irrespective of gender, age, or cirrhosis status. The BALB score significantly correlated with psoas muscle index on computed tomography (rs = 0.6083 for men; rs = 0.6814 for women). CONCLUSION: The BALB scoring system based on routinely used clinical parameters offers a convenient and non-invasive method for predicting DSMM in compensated CLD patients with high accuracy.