Introduction Chronic liver injury has been described in patients taking low dose methotrexate (MTX) for psoriasis or rheumatological disorders. Liver biopsies are often requested on the basis of elevated serum levels of procollagen III peptide (P3NP) or cumulative dosage (CD). However, the Roenigk histological grading system of MTX-associated liver injury is barely distinguishable from the spectrum of non-alcoholic steatohepatitis (NASH), a disorder not controlled for in the original association studies. We hypothesised that MTX usage would correlate poorly with histological abnormalities and that these may instead reflect coincidental risk factors for NASH.
Aim To identify clinical and laboratory parameters predictive of abnormal liver histology in patients taking methotrexate.
Method 41 patients (60% male, mean age 56 years) receiving MTX were identified from a prospective liver biopsy database over a 6-year period in a single centre. Liver histology was reviewed by a single, blinded pathologist and independently scored according to both the Roenigk MTX injury score and the modified Kleiner/Angulo NASH grading systems. Clinical data were used to calculate BMI, Child-Pugh score, MELD, AST:ALT ratio, APRI score, NAFLD fibrosis index and FIB-4 score.
Results There was a high prevalence of obesity (median BMI 31) and 46% of livers had a fatty appearance on pre-biopsy ultrasonography. Elevated mean P3NP levels (mean 5.91 ug/l) were the commonest indication for biopsy, followed by high CD. The median weekly MTX dose was 15 mg with a mean cumulative dosage of 4200 mg (range 360–10300) over a median treatment duration of 60 months.
Macrovesicular steatosis was found in 90% of biopsies and 28% had evidence of steatohepatitis. However, mild fibrosis was present in only 28%, with moderate fibrosis in just 5% and no specimens demonstrated cirrhosis. Applying a CD cutoff of 4 g MTX did not influence the biopsy findings. Serum P3NP levels as well as the duration and total CD of MTX use all correlated poorly with the grade of liver injury. Non-invasive predictors of NAFLD fibrosis such as Angulo score, APRI index and FIB-4 were more accurate in predicting histology. Liver biopsy findings led to a change of MTX dosage in only 5% of cases. During a median follow-up period of 50 months (range 12–114), no patients developed overt chronic liver disease despite continued MTX use. Paired biopsies were available from eight additional patients and demonstrated no histological progression over a mean interval of 38 months (20–53).
Conclusion Obesity and other risk factors for NASH are highly prevalent in patients taking methotrexate for psoriasis. In our cohort, the liver histology correlated poorly with the duration and total dosage of methotrexate therapy and did not progress with further exposure. Elevated serum P3NP levels were an unreliable indicator of liver fibrosis. In contrast, using clinical parameters and non-invasive fibrosis scoring systems could significantly reduce unnecessary liver biopsies in these patients.