Redefining Nutritional Requirements in End-Stage Liver Disease: Towards a Personalized Approach

Abstract

Malnutrition is ubiquitous in cirrhotic patients presenting for liver transplant (LT). Providing an appropriate energy prescription is fundamental to effective nutrition therapy. We aimed to compare measured energy expenditure (mEE) with predicted energy expenditure (pEE) in patients awaiting LT and determine clinical factors associated with mEE. In this prospective observational study, energy expenditure was measured by indirect calorimetry in 110 adult patients referred for LT and predicted by commonly utilized equations (Harris–Benedict, Schofield, and EASL guidelines). Nutritional status, anthropometry, muscle function, biochemical and clinical data were also collected. The median model for end-stage liver disease (MELD) was 19 (IQR 13, 25), and the majority were Child–Pugh B (51%) or C (37%). Malnutrition was evident in 85%. Median mEE by calorimetry was 1756 (1531, 2104) kcal/d and significantly higher than pEE as per Harris–Benedict 1480 (1322, 1722) kcal/d and Schofield 1474 (1349, 1723) kcal/d (both p < 0.001), but lower than EASL guidelines (35 kcal/kg) when an activity factor was applied to mEE; 2283 (1990, 2735) kcal/d versus 2590 (2178, 3010) kcal/d (p < 0.001). Hypermetabolism (mEE:pEE > 1.2) was evident in 48% of the cohort. Multivariate analysis found MELD, Child–Pugh class, diuretic use, and severe malnutrition to be independent predictors of hypermetabolism. A new liver-specific predictive model has been developed, showing superior agreement with mEE than common predictive equations. In conclusion, there is a poor correlation between mEE and pEE in patients awaiting LTs, and hypermetabolism is common. Relying on historical predictive equations in this patient population may result in significant under or over-feeding. A tailored energy prescription based on indirect calorimetry or a liver-specific predictive model is recommended for LT candidates.