Skip to content

Angulo P, Kleiner DE, Dam-Larsen S, et al. Liver Fibrosis, but No Other Histologic Features, Is Associated With Long-term Outcomes of Patients With Nonalcoholic Fatty Liver Disease. Gastroenterology. 2015;149(2):389-97.e10. PMID 25935633

One-paragraph summary

619 patients diagnosed with NAFLD between 1975 and 2005 at centres in the United States, Europe and Thailand underwent liver biopsy and were followed every 3–12 months for a median of 12.6 years (range 0.3–35.1). Outcomes were overall mortality, liver transplantation and liver-related events. The PubMed abstract reports 193 patients (33.2%) died or were transplanted; 193/619 is 31.2%, so the source contains an internal arithmetic inconsistency that is retained explicitly. Fibrosis stage was associated with the composite in a monotonic gradient — hazard ratio 1.88 (95% CI 1.28–2.77) at F1, 2.89 (1.93–4.33) at F2, 3.76 (2.40–5.89) at F3 and 10.9 (6.06–19.62) at F4 versus F0 — as were age (HR 1.07 per year), diabetes (1.61, 1.13–2.30) and current smoking (2.62, 1.67–4.10), while statin use was associated with lower risk (0.32, 0.14–0.70). Liver-related events occurred in 26 patients (4.2%), with hazard ratios of 14.2 (3.38–59.68) at F3 and 51.5 (9.87–269.2) at F4. No other histological feature — steatosis, ballooning, lobular inflammation or the NAFLD activity score — was independently associated with any outcome, and patients with fibrosis had shorter survival regardless of steatohepatitis or NAS.

Key findings

  • Fibrosis stage predicts death, transplantation and liver-related events; nothing else in the biopsy does.
  • Death or transplant HR versus F0: 1.88 / 2.89 / 3.76 / 10.9 at F1–F4.
  • Liver-related events HR versus F0: 14.2 at F3, 51.5 at F4 — on only 26 events, hence the enormous intervals.
  • Statin use associated with a 68% lower hazard of the composite, an early signal of the cardiometabolic dimension.
  • Retrospective, over a 30-year diagnostic era in which criteria and management both changed.

Limitations

  • Retrospective and multi-era; management, biopsy practice and competing risks changed substantially over 1975–2005.
  • Liver-related events were rare (4.2%), so stage-specific estimates are imprecise.
  • Referral-based biopsy cohort, enriched for advanced disease relative to the population.
  • The statin association is observational and vulnerable to healthy-user confounding.

Why it matters

This is the paper the field's prognostic architecture rests on, and it has been replicated repeatedly — in Swedish cohorts up to 33 and 40 years (Ekstedt 2015, PMID 25125077; Hagström 2017, PMID 28803953), in meta-analysis (Dulai 2017, PMID 28130788; Taylor 2020, PMID 32027911; Ng 2023, PMID 35513235) and prospectively (Sanyal 2021, PMID 34670043). Its consequence is a standing contradiction at the centre of MASLD therapeutics: fibrosis is the outcome-relevant variable, yet resolution of steatohepatitis — the feature Angulo showed adds nothing prognostically — is a co-primary endpoint in nearly every registration trial, including both trials that produced approvals.

Cited by wiki pages

  • natural-history-and-fibrosis-progression.md
  • nomenclature-and-definitions.md
  • overview.md