Dear Editor

We read Slouma’s paper [1] with great interest. In this paper, the author aims to determine the frequency of liver fibrosis in rheumatoid arthritis patients treated with methotrexate and to identify its associated factors. Then they found that cumulating more than 3 g of methotrexate was associated with liver fibrosis in rheumatoid arthritis (RA) patients. Having a metabolic syndrome, hypoalbuminemia, higher age, and elevated alkaline phosphatase levels were also likely to be independently associated with liver fibrosis. Despite definite results, in this letter, we raise some concerns about some of the details in the article.

In this study, we noted that 20 factors were analyzed for the univariate logistic regression analysis to analyze the factors associated with liver fibrosis in RA patients treated with methotrexate, which included age ≥ 60 years, male gender, alcoholic consumption, BMI ≥ 25, abnormal WC, comorbidities, metabolic syndrome, disease duration ≥ 10 years, erosive disease, positive RF, positive ACPA, DAS28 CRP > 2.6, DAS28 ESR > 2.6, ESR > 20 mm, CRP > 8 mg/l, presence of fatty liver, steroids use, cumulated dose of MTX > 3 g, hypoalbuminemia, increased ALP, and increased AST. However, these 20 factors were all included in the univariate analysis for a total of 68 RA patients, and only 9 RA patients had liver fibrosis. However, the sample size of univariate and multivariate linear regression analysis is at least 15 times that of the analyzed variable factor [2, 3]. The more variable factors are analyzed, the more example sizes are required. Therefore, more sample sizes are required to make the univariate logistic regression analysis in this liver fibrosis in RA patients treated with methotrexate. Without considering this important point, it could result in unreliable results.

Despite these comments, we extend our congratulations to Slouma et al. for their outstanding work.