Liver failure is uncommon in adults receiving long-term parenteral nutrition

Background: Abnormal liver enzymes and endstage liver disease are reported to occur in 25%-100% and 15%-40% of adult patients receiving long-term parenteral nutrition (PN), respectively. The purpose of this historic cohort study was to investigate the incidence of and possible factors leading to the development of liver disease in our large home PN population. Methods: All patients on home PN for at least 6 months from July 1991 through June 2002 were eligible. Patients were excluded if they had active malignancy, underlying liver disease, or exposure to a hepatotoxin. The presence of PN-associated liver disease was only considered if test results were elevated on more than I occasion over at least 6 or more months. The severity of liver-associated enzymes was based on the degree of elevation and was stratified as mild (< 2 times normal), moderate (2-5 times normal), and severe (> 5 times normal). Severe liver dysfunction was defined as having all of the following criteria: total bilirubin > 3 mg/dL; albumin < 3.2 g/dL; and prothrombin time > 3 seconds prolonged. A cumulative logit model was used to compare age, gender, underlying disease, PN indication, and PN formulas in patients with normal vs abnormal laboratory test results. Results: Two hundred eight patients received PN for more than 6 months, 36 had exclusion criteria, and 10 could not be analyzed, because of incomplete laboratory test results, leaving 162 in the study group. The average PN duration was 2.14 +/- 2.19 years (maximum, 10.28 years). Abnormal liver tests occurred in 154 patients, with most having a moderate elevation of alkaline phosphatase or aspartate aminotransferase; severe liver dysfunction occurred in 7 patients; 1 patient had completely normal liver enzymes. On average, patients received a PN formula that was high in amino acids (1.45 +/- 0.65 g/kg/d), modest in energy (24.7 +/- 13.4 kcal/kg/d), and in most cases with enough lipid emulsion to meet essential fatty acid requirements (0.28 +/- 0.25 g/kg/d). Only female gender was found to be associated with a greater likelihood of liver failure (p =.02). There was a trend toward a greater amount of total calories, dextrose calories, and duration of PN exposure leading to the development of severe liver dysfunction. Conclusions: When long-term PN is given with a modest amount of total energy and a minimal amount of lipid emulsion, abnormal liver enzymes are common, but severe liver dysfunction is unusual.