Assessment of the risk of developing cefepime-induced abnormal liver enzyme levels using the albumin–bilirubin score and fibrosis-4 index: a single-centre retrospective case–control study

Although the albumin-bilirubin (ALBI) score and fibrosis-4 (FIB-4) index may help predict cefepime-induced liver enzyme abnormalities, the relationship between these measures and liver enzyme abnormalities has not been elucidated. This study investigated the relevance and predictive accuracy of the ALBI score and FIB-4 index for cefepime-induced liver enzyme abnormalities. This single-centre retrospective case-control study included 473 patients. The primary outcomes were cefepime-induced abnormal liver enzyme levels. Cox regression analysis was performed with male sex, cumulative dose (≥36 g), concomitant use of acetaminophen or voriconazole, alkaline phosphatase (≥238 IU/L), ALBI score (≥-1.45) and FIB-4 index (≥4.69) as explanatory variables. The predictive accuracies of the ALBI score and FIB-4 index were evaluated based on the AUC using the receiver operating characteristic curve. We performed 1:1 propensity score matching between FIB-4 index ≥4.69 and FIB-4 index < 4.69 groups. The incidence of abnormal liver enzyme levels was 15.6% (74/473). Cox regression analysis revealed that the ALBI score (adjusted hazard ratio, 2.10; 95% CI, 1.268-3.491; P = 0.004) and FIB-4 index (adjusted hazard ratio, 2.33; 95% CI, 1.425-3.796; P = 0.001) were independent risk factors for liver enzyme abnormalities. For all patients, the FIB-4 index (AUC: 0.629) exceeded the ALBI score (AUC: 0.530). Similar results were observed even after propensity score matching. The progression of liver fibrosis before cefepime administration, as assessed using the FIB-4 index, could be more useful than the ALBI score in predicting the risk of developing abnormal liver enzyme levels.