THE PROGNOSTIC IMPLICATIONS OF LIVER FUNCTION TEST ABNORMALITIES IN PATIENTS ADMITTED TO HOSPITAL WITH COVID-19

ABSTRACT

Background & Aims: Prior studies have indicated the presence of hepatic inflammation (as signified by elevated liver function test (LFT) values), as conferring an escalated risk toward adverse outcomes in patients admitted with COVID-19. In line with this hypothesis, we study the various thresholds of LFTs and its associated prognostic risks toward COVID- 19 related hospital deaths Method: This was a single-center retrospective study involving patients admitted with COVID-19. Univariate Cox regression analysis identified the LFT variables significantly associated with our primary endpoint, in-hospital death. Subsequently, 500 iterations of thresholds were generated for each biomarker to estimate the prognostic relationship between biomarker and endpoint. Multivariate Cox regression and event-analyses were performed for each threshold to identify the minimal cutoffs at which the prognostic relationship was significant. Event curves were drawn for each significant relationship. Results: A total of 858 patients with COVID-19 were included with a median follow-up time of 5 days from admission. From the total, 90 patients passed away during admission (10.5%). The deceased cases were more likely to be older (66.2 vs 55.3y p<0.001);however, there was no difference in gender (male 66 vs 56.2% p=0.11). Between the cases and controls (no-death), deceased cases had higher incidence of nonalcoholic fatty liver disease (7.78 vs 2.99% p=0.042), COPD (18.9 vs 7.80% p=0.001), lung cancer (4.44 vs 0.65% p= 0.009), ICU admissions (81.1 vs 26% p<0.001), and intubation events (84.4 vs 19.5% p<0.001), however there was no difference in alcohol use (21.1 vs 30.6% p=0.083) and alcoholic liver disease (5.56 vs 2.08% p=0.097). Upon univariate Cox analysis, the following LFT parameters were associated with in-hospital death Bilirubin (p<0.001), AST (p<0.001), ALT (p<0.001). However, alkaline phosphatase (p=0.449) was not associated with the primary endpoint. The iterations of event regression analyses using 500 sequences of LFT thresholds showed the following cutoffs to be significantly associated with in-hospital death (minimally significant values) ALT (281.71 IU/L), AST (120.94 IU/L), bilirubin (2.615 mg/ dL). On the multivariate analysis, while controlling for demographics and cardiopulmonary/ medical comorbidities, the following adjusted hazard ratios were derived for each cutoff ALT (aHR 6.43 95%CI 1.85-22.40), AST (aHR 3.35 95%CI 1.84-6.11), and bilirubin (aHR 2.77 95%CI 1.15-6.65). Conclusion: The delineated cutoffs for AST, ALT, and bilirubin levels can serve as clinical benchmarks to help determine when a COVID-19 infection poses significant risk. Given this finding, the cutoffs can be used as part of a risk assessment for patients to support early preventative therapies and medical management. (Table Presented)