Candidate molecular mechanisms of naproxen hepatotoxicity: an evidence-graded narrative review of bioactivation, mitochondrial dysfunction, and immune-mediated liver injury.
Siddiqui Shakir S, William Neha Ronald NR
Naproxen sodium, a propionic acid-derived non-steroidal anti-inflammatory drug (NSAID) introduced in 1976, has a favorable hepatic safety profile. Yet clinically apparent naproxen-induced liver injury (NILI) occurs rarely (~1-3 per 100,000 exposed patients), ranging from transient transaminase elevation to cholestatic hepatitis with ductopenia and, exceptionally, acute liver failure. Unlike diclofenac, naproxen lacks mechanistic synthesis. This review aimed to integrate and grade evidence for candidate mechanisms of naproxen hepatotoxicity. Molecular, biochemical, immunological, and clinical data were synthesized in a narrative review. Each candidate mechanism was graded using explicit criteria distinguishing naproxen-specific findings from mechanisms extrapolated from other NSAIDs or hepatotoxicants. Naproxen-specific evidence is strongest for hepatic bioactivation to a reactive acyl-β-D-glucuronide and a coenzyme-A thioester forming covalent protein adducts (isolated rat liver), ex vivo memory T-cell responses to an oxidative desmethyl-naproxen metabolite in a few patients, and rodent oxidative-stress and aminotransferase changes. Cyclooxygenase-dependent loss of hepatoprotection, mitochondrial permeability transition, ATP depletion, Keap1-Nrf2-ARE signaling, HLA-restricted sensitization, Kupffer-cell amplification, and BSEP/MRP2 interference are extrapolated or hypothetical for naproxen, and no study links these steps within one naproxen system. Convergence on intrinsic mitochondrial apoptosis is untested. Clinically, naproxen carries lower liver risk than diclofenac despite overlapping mechanistic potential. This synthesis separates evidence-based from speculative mechanisms of naproxen hepatotoxicity, supporting causality assessment and prioritizing human-relevant experiments for mechanistic and pharmacovigilance research. Hepatoprotective and pharmacogenomic strategies are hypothesis-generating, not clinical recommendations. Only 4 of 12 graded nodes are naproxen-specific; the rest are hypotheses, not an established causal cascade.