Monitoring patients on methotrexate: hepatic fibrosis not seen in patients with normal serum assays of aminoterminal peptide of type III procollagen.

Background: The aminoterminal peptide of type III procollagen (PIIINP) is formed during the synthesis of type III collagen and can be measured in the serum. It has been used as a marker for hepatic fibrosis in patients on long-term methotrexate and it has been suggested that serial assay of PIIINP could reduce or eliminate the need for liver biopsies in these patients.

Objectives: To determine whether routine use of the PIIINP assay in a cohort of patients on methotrexate would reliably identify those who were developing hepatic fibrosis and exclude those who were not, thereby reducing or eliminating the need for liver biopsies in this latter group.

Methods: Data were available from a clinical series of 38 patients on methotrexate, who had undergone a total of 70 liver biopsies and 306 PIIINP assays. Liver biopsies were graded using the Roenigk classification.

Results: In 34 patients, the findings on 46 liver biopsies could be compared with the results of contemporaneous PIIINP assays. Apart from two biopsies from two patients where fibrosis was no longer detected on a subsequent biopsy, all four biopsies showing fibrosis had abnormal results on over half of the associated PIIINP assays. There were no biopsies showing fibrosis where all associated PIIINP assays were normal. However, 50% of biopsies without fibrosis had at least one abnormal associated assay. In 23 patients, the results of serial PIIINP assays performed between two sequential liver biopsies were correlated with changes in the biopsy in terms of fibrosis. Data were available for 32 pairs of liver biopsies. Apart from a biopsy pair in one patient where fibrosis on the second biopsy was not detected on a third biopsy, all four biopsy pairs defined as showing deterioration had abnormal results on over half of the intervening PIIINP assays. There were no biopsy pairs showing deterioration where all intervening assay results were normal. However, 63% of stable biopsy pairs had at least one abnormal intervening assay. Two patients with nonalcoholic steatohepatitis, which manifests a pattern of fibrosis not scored under the Roenigk classification, had persistently and substantially elevated PIIINP levels.

Conclusions: The data presented support the view that follow-up liver biopsies, as recommended by published guidelines, for patients on long-term low-dose methotrexate can be avoided if PIIINP levels are consistently normal. This approach would have reduced the number of patients requiring biopsy in our series by 45%. The PIIINP assay will also be helpful in the management of patients on methotrexate in whom liver biopsy is contraindicated, and in patients with nonalcoholic steatohepatitis.