Atorvastatin liver injury signals rise even as trials find statin safe in cirrhosis

Atorvastatin, one of the most widely prescribed cholesterol-lowering drugs in the world, sits at the centre of a paradox that has grown sharper over the past three years. A structured narrative review published in Frontiers in Pharmacology on 30 October 2026 examined the PubMed literature from 1 July 2023 to 31 July 2026 and found that pharmacovigilance databases continue to generate strong liver-injury signals for the drug, while randomized trials and biopsy-proven cohorts show that it is safe, and in some settings beneficial, in people with chronic liver disease. The review, led by Hao Zhang, PanLi Zhao, Hua Huang and YuQi Zeng of Chengdu Seventh People’s Hospital, screened 119 records and synthesized 57 of them, and its authors are careful to state that it is not a systematic review: only PubMed was searched and comprehensive retrieval is not claimed.

The pharmacovigilance numbers are striking. In the United States FDA Adverse Event Reporting System, 7,779 cases of drug-induced liver injury reported between 2004 and 2023 identified a statin as the primary suspected drug, and atorvastatin accounted for 53.48 percent of them, far ahead of rosuvastatin at 20.44 percent and simvastatin at 19.46 percent. The reporting odds ratio for atorvastatin was 3.09 (95 percent CI 2.99 to 3.19), placing it second only to fluvastatin, whose ratio of 6.90 topped the ranking. Among these cases, 45.43 percent involved patients older than 65 years and 39.95 percent required hospitalization, with mortality highest for simvastatin-associated cases at 12.15 percent. A separate FAERS analysis covering the first quarter of 2004 through the first quarter of 2024 identified 14,511 hepatic disorder events and positive signals for all seven statins, with atorvastatin showing the strongest signal for cholestatic pruritus and abnormal bilirubin conjugation.

Cross-database corroboration strengthens the picture. An analysis combining FAERS with EudraVigilance, the European pharmacovigilance database, included 104,798 reports related to atorvastatin calcium and identified 597 significant signals using four detection algorithms. Rhabdomyolysis and drug-induced liver injury were major designated medical events in both databases, and in EudraVigilance, liver failure, hepatocellular injury and death appeared as key serious outcomes in 1,833 cases, or 4.91 percent of reports. Older patients, 34.12 percent of whom were aged 65 or above, and women, who made up 49.83 percent of reports, warrant specific attention, the authors note. A study based on the Shandong Provincial Adverse Drug Reaction Monitoring Center in China screened 1,391 reports and confirmed 1,211 cases of statin-induced liver injury, including 157 severe ones; multivariable analysis found that atorvastatin conferred a 1.77-fold higher risk of severe injury than rosuvastatin, and combination therapy a 2.08-fold higher risk than statin monotherapy.

The review emphasizes that these signal values cannot be read as risks. Atorvastatin is the most prescribed statin worldwide, so report counts largely reflect the size of the exposed population rather than risk per unit of exposure, and FAERS analyses compare statins with all other drugs rather than with each other while lacking denominator data. Reporting odds ratios therefore cannot be interpreted as absolute risks. The authors also point to a textbook illustration of unmeasured confounding: a Japanese MID-NET analysis found an adjusted hazard ratio of 2.08 for liver dysfunction with generic versus brand atorvastatin, but as follow-up shortened the ratio converged towards unity, reaching 1.33 at 30 days, suggesting the apparent difference does not establish a genuine safety gap between generic and brand products.

Clinically, the injury spectrum is broader than the transaminase elevations that dominated early statin research. In a French pharmacovigilance series of 26 cases, the cholestatic pattern was most frequent at 42.3 percent, followed by hepatocellular injury at 34.6 percent and mixed patterns at 23.1 percent. At least six case reports or series from the past three years describe atorvastatin-associated drug-induced autoimmune hepatitis, with lymphocyte and plasma cell infiltration on histology; these cases must be distinguished from idiopathic autoimmune hepatitis and generally respond well to corticosteroids, though treatment may need to be prolonged. At the severe end, a man in his seventies developed hyperacute hepatic failure within 24 hours of starting atorvastatin 80 mg for secondary stroke prevention; after alternative causes were excluded, drug-induced liver injury was diagnosed and the patient improved rapidly on withdrawal. Median time to onset of the first episode in the French series was 27 days, with an interquartile range of 4 to 43 days, and in an analysis of 60 published case reports most reactions occurred within 2.3 months of treatment initiation.

Rechallenge data add a practical warning. In the French series, same-drug rechallenge was positive in 12 of 16 cases, with 11 of 13 relapsing at the same dose and 1 of 3 at a reduced dose, and the median latency of the recurrent episode shortened markedly from 27 days to 11 days. Among 10 patients rechallenged with an alternative statin, 2 relapsed, one with rosuvastatin and one with simvastatin, while no recurrence was observed in 5 patients receiving rosuvastatin, 2 receiving pravastatin and 1 receiving simvastatin. The authors concluded that rosuvastatin or pravastatin carry the lowest risk of recurrence if lipid-lowering therapy must continue. The review recommends that same-drug rechallenge be avoided after confirmed injury and that a hydrophilic statin be preferred when treatment must continue, with close monitoring.

Mechanistic work published in the review window points to several interacting pathways, though the authors stress that most derive from experimental systems and do not by themselves establish that these pathways operate in patients. Atorvastatin inhibits mitochondrial electron transport chain complexes, causing ATP depletion and reactive oxygen species generation that drives lipid peroxidation and DNA damage. One study provided the first systematic characterization of P450-mediated metabolic activation of the drug: two phase I metabolites were detected in an NADPH-supplemented system, and trapping experiments revealed four glutathione and cysteine conjugates, indicating electrophilic quinone-imine intermediates; after intragastric administration to mice, corresponding protein adduction was observed in vivo, with CYP3A the predominant enzyme and ketoconazole pretreatment reducing both adduction and hepatocyte susceptibility. Atorvastatin also inhibits the bile-acid transporters BSEP, NTCP and MRP2, which aligns with the cholestatic phenotype seen clinically, and pharmacogenomic variation in SLCO1B1, CYP3A4/5, UGT1A1 and ABCG2 modulates susceptibility.

Against the pharmacovigilance signals stand consistently reassuring clinical data in chronic liver disease. In a biopsy-proven MASLD cohort of 215 patients with a median treatment duration of 63 months, statin use was not associated with worsening liver enzymes, and in the subgroup with advanced fibrosis it was not associated with worsening of liver enzymes or liver stiffness. A randomized, double-blind, placebo-controlled trial in Denmark involving 78 patients with cirrhosis and portal hypertension found that atorvastatin 10 to 20 mg daily for 6 months was safe but did not reduce mortality, liver-related complications or hepatic venous pressure gradient. A randomized, placebo-controlled trial in Egypt with 100 patients reported that atorvastatin 20 mg daily reduced the cumulative recurrence of cirrhosis-related complications from 72 percent to 36 percent (HR 0.50, 95 percent CI 0.33 to 0.75) and completely prevented hepatorenal syndrome, with transaminase elevations transient and clinically insignificant. In a cohort of 329,577 patients with MASLD, statin use was associated with a lower risk of primary liver cancer (adjusted HR 0.64, 95 percent CI 0.57 to 0.71) in a dose-dependent manner, and a meta-analysis of 11 studies found an odds ratio of 0.55 for atorvastatin. A randomized trial of 1,596 individuals followed for 5 years found that an atorvastatin-containing polypill reduced major adverse cardiovascular events (HR 0.56) and cardiovascular mortality (HR 0.41) in patients with MAFLD, with greater benefit than in the general population.

Model-based work adds nuance for special populations. Simulations using a permeability-limited multicompartment liver model predicted more than a 250 percent increase in peak plasma concentration and area under the curve in MASH, with unbound hepatocellular exposure rising by up to 127 percent in the periportal region. In cirrhosis, simulations predict substantially higher exposure with increasing disease severity, corresponding to hypothetical dose reductions of approximately 70, 84 and 92 percent for Child-Pugh classes A, B and C. The review is explicit that these are model-derived estimates that have not been validated against clinical outcomes and are not dosing recommendations. A natural language processing pharmacovigilance study of 109,263 hospitalized older adults also identified a marked hepatotoxicity interaction signal when atorvastatin was combined with piperacillin/tazobactam, corroborated by animal experiments, though such database-mining signals require confirmation in prospective cohorts.

The practical upshot, summarized in the review’s management table, is a risk-stratified approach rather than a blanket contraindication. Liver biochemistry should be measured before starting therapy; no statin-specific randomized evidence defines an optimal surveillance interval, and guidance differs between bodies, with US product labelling revised in 2012 to remove routine periodic enzyme monitoring while other guidance suggests early reassessment after initiation or dose escalation. Discontinuation is advised when ALT or AST exceeds three times the upper limit of normal with concurrent bilirubin elevation, exceeds five times the upper limit, or is accompanied by symptoms of liver injury. Causality assessment with RUCAM or RECAM, supplemented by simplified autoimmune hepatitis scoring and liver biopsy for suspected autoimmune-like phenotypes, is recommended, and isolated gamma-glutamyl transferase elevation with normal transaminases should prompt a search for alternative causes rather than automatic statin withdrawal. The authors conclude that severe atorvastatin-associated liver injury is rare but recognizable, and that in MASLD, compensated cirrhosis and cirrhosis with portal hypertension, statin therapy should not be withheld or interrupted because of excessive concern about hepatotoxicity; instead, prescribing and monitoring should be stratified according to hepatic functional class and individual risk.

Subject of Research: Hepatic safety, injury phenotypes and pharmacovigilance signals of atorvastatin

Article Title: Atorvastatin-associated liver injury: a structured narrative review of reporting signals, clinical phenotypes, and hepatic safety, 2023–2026

Article References: Zhang, H., Zhao, P., Huang, H., & Zeng, Y. (2026). Atorvastatin-associated liver injury: a structured narrative review of reporting signals, clinical phenotypes, and hepatic safety, 2023–2026. Frontiers in Pharmacology, 17, Article 2003438. https://doi.org/10.3389/fphar.2026.2003438

Image Credits: AI Generated

DOI: 10.3389/fphar.2026.2003438

Keywords: atorvastatin, drug-induced liver injury, pharmacovigilance, hepatotoxicity, autoimmune hepatitis, statins, cirrhosis, MASLD, rechallenge, FAERS, bile-acid transporters, liver enzymes