EoE Linked to Higher Long-Term NAFLD, Cirrhosis Risk
Key Takeaways
- Adults with EoE in a U.S. TriNetX cohort had higher observed long-term risk of NAFLD and liver cirrhosis/fibrosis than matched controls.
- The NAFLD association remained significant across alternative EoE definitions, longer washout periods, and expanded medication-adjusted models.
- When corticosteroid or PPI exposure was excluded, the NAFLD association persisted, while the cirrhosis/fibrosis association no longer reached statistical significance.
- Subgroup analyses showed broader consistency for NAFLD across sex, age, and higher-body-mass-index strata, with more selective cirrhosis/fibrosis signals in women, adults aged 18 to 64 years, and patients with body mass index of at least 30 kg/m².
In the Gau et al. multicenter TriNetX study of eosinophilic esophagitis and longitudinal liver risk, investigators used the TriNetX US Collaborative Network for a multicenter retrospective cohort study of adults aged 18 years or older with at least two healthcare visits between January 2005 and December 2023. Patients with an EoE diagnosis were matched 1:1 to EoE-free controls, and those with prior liver disease coded as ICD-10-CM K70-K77, malignancy, or death on or before index were excluded. Incident nonalcoholic fatty liver disease (NAFLD) and newly diagnosed liver fibrosis or cirrhosis were defined after excluding diagnoses within 3 months of index. Matching covered demographic, utilization, cardiometabolic, kidney, psychiatric, and socioeconomic factors, with all standardized mean difference (SMD) values below 0.1 after matching, and prespecified sensitivity analyses varied the EoE definition, washout window, and medication adjustment.
Over follow-up of up to 15 years, EoE was associated with higher incident NAFLD risk than matched controls (hazard ratio
1.409, 95% confidence interval [CI] 1.277-1.555) and higher liver cirrhosis or fibrosis risk (HR 1.529, 95% CI 1.176-1.988). An expanded medication-adjusted model preserved the NAFLD association (HR 1.404, 95% CI 1.290-1.528), and the cirrhosis or fibrosis signal also remained significant in that model. Proton pump inhibitor (PPI)-based, steroid-based, endoscopy-based, and longer-washout analyses moved in the same direction, especially for NAFLD. After corticosteroid or PPI exposure was excluded, the NAFLD association remained significant, whereas the cirrhosis or fibrosis association no longer did. Subgroup analyses showed a broader and more consistent pattern for NAFLD than for advanced liver outcomes.
These findings remain observational and hypothesis-generating rather than proof that EoE causes hepatic disease. Residual confounding remained plausible despite matching, especially from diet, physical activity, central adiposity, socioeconomic factors, disease severity, and concurrent medications, and the Full-text methods and sensitivity analyses for the EoE-associated NAFLD and cirrhosis/fibrosis study emphasized medication-related confounding as a central issue for interpreting advanced liver outcomes. Selection bias may have entered through tertiary-care identification of EoE, and greater healthcare contact could have increased surveillance for liver disease. Coding- and documentation-based outcome capture also means NAFLD and cirrhosis or fibrosis were not systematically biopsy confirmed. Shared T-helper 2 inflammation, eosinophils, barrier dysfunction, microbiome effects, and the gut-liver axis were discussed as biologic plausibility, not as direct evidence from this dataset.
The investigators reported a pattern of higher observed long-term risk of NAFLD and liver cirrhosis or fibrosis among adults with EoE in this propensity-matched U.S. cohort. The NAFLD signal remained broadly consistent across sensitivity analyses, while the advanced-liver outcome appeared more sensitive to medication-related analyses. The authors described the findings as observational and hypothesis-generating, while also suggesting that hepatic surveillance and broader clinical attention to liver risk in EoE may merit consideration.
Clinician Questions
How was incident liver disease defined in adults with eosinophilic esophagitis in this TriNetX analysis?
Adults with eosinophilic esophagitis were followed for incident NAFLD and newly diagnosed liver fibrosis or cirrhosis, and diagnoses within 3 months of the index date were excluded to reduce baseline misclassification and reverse-causality concerns.
Which parts of the eosinophilic esophagitis–liver association were most sensitive to medication-related confounding?
NAFLD remained associated with eosinophilic esophagitis across expanded medication-adjusted and medication-exclusion analyses, whereas the cirrhosis or fibrosis signal weakened when corticosteroid or proton pump inhibitor exposure was excluded, which the authors highlighted as a main caution in interpreting advanced liver outcomes.
Did the observed liver-risk signal in eosinophilic esophagitis extend across patient subgroups?
The NAFLD association was reported across male and female patients, younger and older adults, and overweight and obese strata, while significant cirrhosis or fibrosis elevations were more selective, appearing in women, adults aged 18 to 64 years, and patients with body mass index of at least 30 kg/m².
Recommended Reading
- For more on EoE cohort findings: Pediatric EoE Shows Patchy Biopsy Findings in 15-Year Cohort