Short-term derangement in liver function tests adds prognostic value to the presence of heart failure when predicting 12-month mortality

Liver function tests (LFTs) prompt and influence diagnostic and therapeutic processes but can also act as prognostic tools. Elevated bilirubin and transaminases in patients with chronic heart failure (CHF) are associated with an increased risk of cardiovascular disease and long-term mortality. The implications for long-term mortality of short-term changes in LFTs are unclear, particularly in those without chronic liver disease. The aim was to characterise the prognostic utility of short-term changes in LFTs with a mortality prediction model that accounted for the effect of CHF. LFTs (bilirubin, albumin, gamma glutamyl transferase, alkaline phosphatase, alanine aminotransferase (ALT) and aspartate aminotransferase) were measured on admission and on at least one other day in General Medicine inpatients non-electively admitted to a tertiary hospital for a period of at least 7 days. Only the first admission of the 7-year study period was included. Patient exclusion criteria were an existing coded diagnosis of liver disease or any malignancy, an admission to the intensive care unit or death during admission. A vector of intra-admission change from baseline was derived from the maximum change in LFTs. The association between 12-month mortality and admission and intra-admission changes in LFTs in patients with and without an established diagnosis of CHF was assessed by logistic regression modelling adjusted for age, gender and co-morbidity index. Of 4160 patients analysed, median age was 79 years and 12-month mortality was 17%. A coded diagnosis of CHF was present in 25% of the cohort; 27% of these patients died within 12 months compared to 14% of those without CHF (p<0.001). Albumin and ALT were the only LFTs for which an intra-admission change significantly predicted mortality, and the predictive effects were additive. 12-month mortality was greater by 4% per 1g/L intra-admission decrement in albumin (p<0.001) and 6% per 100 IU/L intra-admission increment in ALT (p=0.05). Comparison of the models showed that intra-admission changes were superior to admission values in predicting 12-month mortality. Furthermore, the presence of CHF portended greater mortality risk for any given change in albumin or ALT. In the absence of known liver disease, a patient whose albumin decrements during admission is at high risk of death within one year, more so with a pre-existing diagnosis of heart failure. Especially in patients with CHF, short-term deterioration in albumin or ALT should direct clinicians to timely management after discharge, including consideration of end-of-life care discussions.