Background/Purpose: Methotrexate (MTX) use for the treatment of rheumatoid arthritis (RA) has been associated with hepatotoxicity, and requires monitoring of liver transaminases.  However, elevations in liver transaminases may be transient and may not predict the development of cirrhosis or fibrosis.  We aimed to identify clinically relevant factors associated with persistently elevated liver transaminases leading to discontinuation or dose reduction of MTX in RA patients.

Methods: The study population was derived from an EMR-based cohort of 5906 RA cases at a large academic medical center followed since 1992.  We used a validated algorithm to identify patients with RA (PPV 94%).  We extracted data on any MTX prescription, and any LFT elevation, defined as liver transaminases > 2 times the upper limit of normal (> 2X ULN).  From the 1,040 RA patients identified as ever treated with MTX and with LFTs > 2X ULN, we randomly selected 500 patients for detailed chart review; 90 cases (18%) were confirmed as having LFT elevations while receiving MTX, with LFT elevations attributed to MTX.  We abstracted data on risk factors for liver toxicity: age, sex, obesity, hyperlipidemia, MTX dose (2.5 – 7.5mg, 10.0 – 17.5mg, or 20 – 25mg), alcohol use, NSAID use, and statin use. We defined our outcome as continuation of MTX versus discontinuation or dose reduction. We examined the univariable associations of the predictors with outcome and developed a multivariable adjusted logistic regression model to estimate odds ratios and 95% CI.

Results: In our cohort of 90 patients, MTX was discontinued or the dose was reduced in 55 (61%) of patients.  The remainder of those that continued MTX had a single transaminase elevation that resolved.  Among the patients in whom MTX was discontinued or the dose was reduced, 8 (15%) of patients had biopsies that showed fibrosis attributed to MTX use.  In our univariable analysis, obesity was significantly associated with MTX discontinuation or dose reduction (Table). In our multivariable model we included age, sex, obesity and hyperlipidemia, and obesity remained significantly associated with MTX discontinuation or dose reduction, OR 2.60 (95% CI 1.01-6.66), p-value 0.05. Fifteen (27%) of obese patients had a clinical diagnosis of nonalcoholic fatty liver disease (NAFLD) in the MTX discontinuation or dose reduction group, and 2 (4%) obese patients had this diagnosis in the MTX continuation group.

Conclusion: In our cohort of RA patients with MTX-related LFT elevation, single LFT elevations resolved in 39% of patients. We found a significant association between obesity and LFT elevations that led to MTX discontinuation or dose reduction. This suggests a potentially heightened risk of hepatotoxicity among obese patients. Further studies are needed to determine whether NAFLD may be an underlying risk factor in this patient population.

Table. Univariable Analysis of Predictors of MTX Discontinuation or Dose Reduction

* p for trend = 0.85