Prospective Comparison of Transient Elastography, Liver Biochemistries, and APRI for the Assessment of Methotrexate Induced Liver Damage in Patients with Dermatologic or Rheumatic Diseases

Background: Concerns about methotrexate-induced liver damage in patients with psoriasis and rheumatoid arthritis have led to the recommendation of serial liver biopsies to monitor the progression of liver injury. However, liver biopsies may be associated with life-threatening complications. There is a need for accurate noninvasive methods in assessing the degree of liver damage. The present study assessed the performance of transient elastography (Fibroscan ® ) for the prediction of liver fibrosis in patients with psoriasis or rheumatoid arthritis on methotrexate treatment, in comparison with liver biochemistries and the aspartate aminotransferase (AST) to platelets ratio index (APRI). Methods: We prospectively enrolled outpatients who received cumulative doses of methotrexate more than 1500 mg and gave their written informed consent for liver biopsy examination between November 2007 and November 2008. For each patient, biochemical and hematologic determination, and liver stiffness measurement using Fibroscan followed by liver biopsy were performed on the same day. All liver biopsy specimens were analyzed independently by one experienced pathologist with blinded. Results: Seventeen patients with psoriasis and one patient with rheumatoid arthritis were recruited. Patients had a mean age (± SEM) of 46.3 ± 2.8 years, 67% were female, and their mean body mass index was 22.9 ± 0.8 kg/m 2 . The mean serum alanine aminotransferase (ALT), the mean AST/ALT ratio, and the APRI score were 23 ± 2 units/L, 1.10 ± 0.06, and 0.23 ± 0.03, respectively. The mean value of Fibroscan was 6.2 ± 0.5 kPa. According to the Roenigk classification, 6 biopsy specimens were classified as grade II, 7 as grade IIIa, 4 as grade IIIb, and 1 as grade IV. The areas under the receiver operating characteristic curve of Fibroscan, ALT, AST/ALT ratio, and APRI values for Roenigk grade ≥ IIIa were 0.97, 0.76, 0.71, and 0.61, and for Roenigk grade ≥ IIIb were 0.88, 0.52, 0.62, and 0.60, respectively. The optimal cutoff value of Fibroscan for the prediction of significant fibrosis (grade IIIa) was 5.2 kPa, which yielded a sensitivity of 83%, specificity of 100%, positive predictive value (PPV) of 100%, and negative predictive value (NPV) of 75%. Similarly, the optimal cutoff value of Fibroscan for the prediction of advanced fibrosis (grade IIIb) was 7.6 kPa, which provided a sensitivity of 80%, specificity of 92%, PPV of 80%, and NPV of 92%. Conclusion: Fibroscan is a rapid, noninvasive, and reproducible method for assessing methotrexateinduced liver damage, with better performance than liver biochemistries or APRI. Fibroscan help avoid the need for liver biopsy in most patients with dermatologic or rheumatic diseases on methotrexate therapy.