O11 Methotrexate: an innocent bystander in the development of liver fibrosis, findings of the STrATIFY study

Introduction Methotrexate (MTX) is an effective treatment for immune-mediated diseases, used by several specialities. MTX-induced hepatotoxicity has been historically regarded with concern,1 more recent evidence has contradicted this.2,3 Non-alcoholic fatty liver disease (NAFLD) is highly prevalent, estimated to affect 1 in 5 in the UK.4

Aim To compare the prevalence of liver fibrosis, by way of FibroScan®, between individuals taking MTX and those who have not, and cross-reference this against established risk factors for liver fibrosis.

Methods Following national Health Research Authority approval and local R&D registration, we recruited adults attending the rheumatology and dermatology outpatient clinics at York Hospital. Half of the cohort had never received MTX (control), whilst half (MTX) had taken MTX for at least six months. Pregnant women were excluded. Demographic details, past medical history, cumulative MTX dose, other drug history, physical activity levels and alcohol history were recorded. Transient elastography and body mass composite analysis were undertaken. Statistical analysis was conducted by means of independent 2-tailed t test and multiple regression analysis, using SPSS.

Results 600 participants were recruited. Baseline characteristics for the MTX and control cohorts were as follows; female (68 vs 70%, p = 0.44), mean body mass index (28.3 vs 28.4 kg/m2, p = 0.77), hypertension (28 vs 23%, p = 0.19), hypercholesterolaemia (21 vs 20%, p = 0.76) and diabetes (8 vs 7%, p = 0.53) respectively. No difference in FibroScan® score was identified between the MTX group (5.9kPa 95% CI 5.3 – 6.5) and the control subjects (6.5kPa 95% CI 5.6 – 7.2) (p = 0.28).

Multiple regression analysis demonstrated MTX prescription (p = 0.58), cumulative dose of methotrexate (p = 0.47) and self-reported alcohol (by way of AUDIT-C) (p = 0.34) were not significant predictors of liver fibrosis. Positive predictors of liver fibrosis were waist circumference (p ≤ 0.01), BMI (p ≤ 0.01), and diabetes (p = 0.02).

Discussion This large cohort study has demonstrated no increased prevalence of liver fibrosis (by way of FibroScan®) in patients taking MTX. In keeping with other studies, risk factors for NAFLD were significant predictors of liver pathology. The phenomenon of methotrexate-related hepatotoxicity is likely to have been historically over-estimated. Our results suggest that this cohort had NAFLD as the underlying cause of liver fibrosis. The STrATIFY study continues to recruit participants.