A preoperative risk score based on multifrequency bioelectrical impedance analysis in patients undergoing liver resection.

ABSTRACT

PURPOSE: To investigate if body composition parameters measured by bioelectrical impedance analysis (BIA) and are reportedly correlated with clinical outcomes of patients undergoing digestive tract surgery could be useful for reliably evaluating the perioperative risk in patients undergoing hepatectomy. METHODS: Consecutive 200 patients who underwent BIA before hepatectomy were retrospectively reviewed. A risk prediction model for postoperative morbidity was created using the initial 100 patients, and its performance was validated using the remaining 100 patients. RESULTS: Based on the correlation with postoperative morbidity, a novel risk prediction model, the protein-edema score, was created using net protein weight and extracellular water/total body water ratio measured through BIA. The protein-edema score (score 0 vs. ≥ 1) showed a reproducible correlation with Clavien-Dindo 2 or greater postoperative morbidity in the validation set (17.7% vs. 46.4%, P = 0.002) as observed in the training set (18.8% vs. 49.0%, P = 0.002) after statistical adjustment. Similar tendency was also confirmed in Clavien-Dindo 3a or greater postoperative morbidity (5.9% vs. 18.2%, P = 0.037) and postoperative refractory ascites (5.5% vs. 17.4%, P = 0.037) in the validation set. CONCLUSIONS: The protein-edema score created based on BIA is significantly correlated with postoperative morbidity in patients undergoing liver resection.