FDA finalises guidance on drug-induced liver injury biomarkers for clinical trials: miR-122, GLDH, and keratin-18 alongside ALT

FDA published final guidance on 17 November 2026 recommending that sponsors incorporate novel DILI biomarkers alongside conventional liver function tests (ALT, AST, total bilirubin, ALP) in Phase 2–3 clinical trials involving drugs with hepatotoxic potential. The three biomarkers specifically recommended are: microRNA-122 (miR-122) — a highly liver-specific biomarker released within hours of hepatocellular injury, detectable before ALT elevation; glutamate dehydrogenase (GLDH) — a mitochondrial enzyme marking severe hepatocellular necrosis, more specific than ALT for hepatocellular vs adaptive elevation; and cytokeratin-18 fragments (CK-18 M30 and M65) — markers of apoptotic (M30) and total (M65) hepatocyte cell death. The guidance does not mandate their use but states that trials without them in high-risk drugs may receive requests for additional biomarker data during review. See our drug-induced liver injury guide for the full DILI clinical framework.

Drug-induced liver injury (DILI) is the leading cause of acute liver failure in the developed world and the most common reason for post-market drug withdrawal. ALT elevation — the standard DILI signal in clinical trials — has significant limitations: it is non-specific (muscle injury, haemolysis, and strenuous exercise all elevate ALT), it lags behind hepatocellular injury by days, and it cannot distinguish between adaptive enzyme induction (benign, transient ALT rise due to CYP induction, not true hepatotoxicity) and true hepatocellular necrosis. The landmark FDA Guidance on Drug-Induced Liver Injury: Premarketing Clinical Evaluation ↗ (2009) established Hy's Law (ALT ≥3×ULN + bilirubin ≥2×ULN in the absence of other causes = signal of serious DILI), but Hy's Law only identifies cases after liver damage has occurred. Novel biomarkers aim to provide earlier, more specific injury signals. The DILI guide covers Roussel Uclaf Causality Assessment Method (RUCAM), the R-ratio for distinguishing hepatocellular vs cholestatic patterns, and high-risk drug classes.

microRNA-122 constitutes approximately 70% of total liver microRNA content and is released into plasma within 2–6 hours of hepatocellular damage — substantially earlier than ALT, which peaks 24–48 hours post-injury. Clinical studies including the DILIN prospective study ↗ and acetaminophen overdose cohorts demonstrate that miR-122 rises before ALT in acute DILI and returns to baseline faster upon recovery — a temporal profile that improves both early detection and resolution assessment. Importantly, miR-122 is not elevated by non-hepatic causes of ALT elevation (myopathy, haemolysis) — improving diagnostic specificity in trials where participants may have concomitant muscle disease or inflammatory conditions. The FDA guidance recommends pre-dose, peak, and recovery time point collection, with specific assay (RT-qPCR or NGS) and pre-analytical handling requirements to prevent miR-122 degradation.

Including novel DILI biomarkers in trial design allows: (1) Earlier stopping rules — miR-122 or GLDH elevations can trigger liver safety review board evaluation before Hy's Law criteria are met; (2) Characterisation of adaptive vs toxic ALT elevation — GLDH helps distinguish CYP induction-related ALT rise (GLDH normal) from true necrosis (GLDH elevated); (3) Retrospective forensics — banked samples from completed trials can be retrospectively assayed for novel biomarkers when an unexpected DILI signal emerges post-market. Sponsors should plan sample collection, storage (−80°C for miR-122 stability), and analysis logistics during trial design. The guidance also recommends including DILI biomarker collection in patients with pre-existing liver disease to understand baseline variability. Report DILI adverse events via FDA MedWatch ↗ and the DILIN registry ↗.