Significant Medical Comorbidities Are Associated With Lower Causality Scores in Patients Presenting With Suspected Drug-Induced Liver Injury

Abstract

INTRODUCTION: Drug-induced liver injury (DILI) is a diagnosis of exclusion, and it can be challenging to adjudicate when there are multiple comorbidities and concomitant medications. In this study, we tested the hypothesis that comorbidity burden impacts the causality adjudication in patients with suspected DILI. METHODS: We studied consecutive patients with suspected DILI enrolled in the Drug-Induced Liver Injury Network Prospective Study at 2 centers between 2003 and 2017. The comorbidity burden at presentation was determined using the Charlson Comorbidity Index (CCI). We analyzed the association between significant comorbidity (CCI > 75th percentile) and (i) the adjudication of DILI by expert consensus as definite, highly likely, or probable (high-confidence DILI) and (ii) the Roussel Uclaf Causality Assessment Method (RUCAM) scores. RESULTS: Our cohort consisted of 551 patients who were classified as "no comorbidity" (54%, CCI = 0), "mild comorbidity" (29%, CCI = 1 or 2), and "significant comorbidity" (17%, CCI > 2). The probability of high-confidence DILI was significantly lower in patients with significant comorbidity compared with those with mild or no comorbidities (67% vs 76% vs 87%, respectively, P < 0.001). The mean RUCAM scores decreased with increasing comorbidity (no comorbidity 6.6 ± 2, mild comorbidity 6 ± 2.4, and significant comorbidity 5.6 ± 2.7, P < 0.001). In the multiple logistic regression, significant comorbidity had an independent inverse relationship with DILI (odds ratio: 0.37, 95% confidence interval: 0.2-0.69, P = 0.001). DISCUSSION: Higher comorbidity burden impacts the causality assessment in patients with suspected DILI. Further studies are needed to investigate the utility of comorbidity burden as a variable in the DILI causality instruments.