Validity of noninvasive markers of methotrexate-induced hepatotoxicity: a retrospective cohort study

Summary Background Methotrexate is a cost-effective systemic treatment for moderate-to-severe psoriasis, but the perceived risk of associated liver fibrosis prevents optimal use. Procollagen III aminoterminal propeptide (PIIINP) is a widely adopted noninvasive biomarker of liver fibrosis; however, its clinical utility is narrow owing to limited evidence of performance and the need for serial measurement. The Enhanced Liver Fibrosis (ELF) assay is a validated biomarker of liver fibrosis. Objectives To evaluate the diagnostic accuracy of the ELF test compared with PIIINP for the diagnosis of liver fibrosis in a cohort of patients with psoriasis treated with methotrexate. Methods A retrospective cohort study comparing the diagnostic accuracy of PIIINP and ELF in detecting liver fibrosis in patients treated with methotrexate. Liver biopsy was the reference standard. Results Twenty-seven patients were identified and included in the study. The diagnostic accuracies [area under the receiver operating curve (AUROC)] of serial PIIINP and serial ELF were 0·589 [95% confidence interval (CI) 0·379–0·800] and 0·643 (95% CI 0·391–0·895), respectively, for mild fibrosis; and 0·576 (95% CI 0·237–0·916) and 0·674 (95% CI 0·421–0·927) for at least moderate fibrosis. The AUROC values for single PIIINP and single ELF were 0·582 (95% CI 0·363–0·801) and 0·693 (95% CI 0·482–0·904), respectively, for mild fibrosis; and 0·667 (95% CI 0·363–0·971) and 0·806 (95% CI 0·564–1·000) for at least moderate fibrosis. Conclusions This pilot study suggests that ELF may be at least equivalent or possibly superior to PIIINP in the detection of liver fibrosis in patients with psoriasis treated with methotrexate, and supports further investigations into the performance of ELF in this clinical setting.