1. Top of page
  2. Introduction

Clinical practice guidelines have been designed with an ultimate goal of improving the quality and cost efficiency of patient care. In 1994, a subcommittee of the American College of Rheumatology (ACR) published guidelines for monitoring liver toxicity during methotrexate (MTX) treatment of rheumatoid arthritis (RA) (1). To identify patients with unsuspected liver disease, liver function tests and hepatitis B and C serologies were recommended for all RA patients prior to starting MTX, as were baseline liver biopsies in those with a known chronic viral hepatitis, abnormal baseline aspartate aminotransferase values, or excessive alcohol consumption. Subsequent monitoring of liver function tests was suggested every 4–8 weeks during treatment with MTX. Additionally, the ACR guidelines recommended liver biopsies in patients with abnormal findings on liver function tests that persisted during treatment or after discontinuation of the drug.

In the following study, we attempted to measure compliance with the ACR guidelines for hepatitis C testing in RA patients receiving MTX in ambulatory settings. We aimed to quantify the effect of 2 interventions to increase guideline compliance by comparing each to its historical controls represented by usual care. Furthermore, we assessed the frequency and severity of adverse outcome when the guideline was not followed.

The setting of our study was a large urban university center and an associated veteran's hospital (VA). We conducted a population-based retrospective cohort study of hepatitis C frequency in patients taking MTX at the 2 centers. We then compared 2 interventions to increase compliance by prospective 6-month followup. This study did not include a usual care control cohort other than historical controls. Guideline compliance intervention consisted of a personal note to a provider at the point of service (university hospital) versus an explicit order entered into the information system (VA hospital). The study was a trial of 2 quality interventions designed for maximum effect within each center. The goal was to uncover the greatest number of patients with hepatitis C and concomitant MTX and record their histology. It was therefore not a cohort randomization to isolate the effect of the information system at the VA hospital or to test the hypothesis that electronic reminder was more effective than a paper reminder.

There were 634 patients at the university center rheumatology clinic, of whom 96 were taking MTX for RA (23 men and 71 women; 32 white, 61 African American, and 1 Asian; mean age 53.7, range 27–88 years). At the VA hospital rheumatology clinic, there were 636 patients, with 82 taking MTX (77 men and 5 women; 55 white, 26 African American, and 1 Hispanic; mean age 65.1, range 37–98 years). Of the RA patients in the overall population, 36% were taking MTX.

At the veteran's hospital 0 of the 82 RA patients taking MTX tested hepatitis C enzyme immunoassay (EIA) positive. Of these patients, 44% had hepatitis C tests prior to MTX, while the remaining 56% had hepatitis C serology as a result of the intervention to increase screening. Consequently, 100% of RA patients at the VA hospital were tested for hepatitis C. At the university hospital, only 10% of patients had hepatitis C testing before initiating MTX. Of the patients, 28% had testing during the course of MTX treatment but prior to the study intervention, while 24% were tested as a result of the study intervention. The remaining 38% of the patients were not screened despite the intervention. Out of the university hospital cohort, 5 patients tested EIA positive, 4 tested polymerase chain reaction (PCR) positive, and 3 patients had biopsies performed (Table 1). None of the 3 patients had cirrhosis or fibrosis. It is noteworthy that hepatitis C patients were, on average, 8 years younger than other RA patients at the university hospital and 20 years younger than RA patients at the VA hospital. A 3-year followup of the 38% unscreened university patients revealed little Hawthorne effect (tendency of research subjects, here clinicians, to act unusually as a result of their awareness of being studied), with only 3 of those patients eventually being tested for hepatitis C. Results of all 3 tests were negative.

Table 1. Liver biopsies from hepatitis C patients taking MTX*
PtMTX time, monthsHistologic findingsRoenigk scoreKnodell scoreClinical
  • *

    MTX = methotrexate; Pt = patient.

  • Modified Roenigk classification of histologic change in the liver graded I–IV, where grade I indicates mild fatty change; grade II, mild to moderate inflammation; grade IIIa mild portal fibrosis; grade IIIb, moderate to severe fibrosis; and grade IV, frank cirrhosis (2). In biopsy findings that reveal grade IIIb or IV, the American College of Rheumatology strongly recommends discontinuation of MTX therapy (1).

  • Histologic Activity Index (Knodell score out of 22) in grading and scoring liver damage, where presence of 4 histologic entities are scored: 1) periportal ± bridging necrosis (10 maximum points), 2) intralobular degeneration and focal necrosis (4 maximum points), 3) portal inflammation (4 maximum points), 4) fibrosis (4 maximum points) (3).

122Lobular/portal steatosisII2 (0,1,1,0)MTX discontinued
26SteatosisI4 (0,1,3,0)MTX increased
314Minimal lobular infiltrate01 (0,1,0,0)MTX increased

Despite hepatotoxicity being an important, though uncommon, complication of long-term MTX therapy (4, 5), little is known about the safety of MTX in patients with concomitant hepatitis C. A recent retrospective study conducted by Mok and colleagues (6) examined the safety of disease-modifying antirheumatic drugs (DMARDs), including MTX, in RA patients with chronic viral hepatitis (CVH). Among 29 patients with CVH, 23 were hepatitis B surface antigen positive and 6 were anti-hepatitis C virus (HCV) antibody positive. The study analyzed a total of 47 drug episodes, defined as treatments with DMARDs singly or in combination therapy for more than 6 months. Two patients, who where treated with MTX singly or in combination, demonstrated normal liver enzyme levels during this study. Unfortunately, no conclusion was drawn regarding the safety of MTX and RA in hepatitis C positive patients due to minimal number of drug episodes and low mean accumulated dose (650 mg) of MTX used. The authors did speculate, however, that the effects of hepatitis C and MTX on the liver may be synergistic.

The results of our trial measure the prevalence of unsuspected hepatitis C in unscreened RA patients taking MTX to be 5.6% (95% confidence interval [95% CI] 1.7–14.3%). The total number of patients taking MTX screened for hepatitis C as a result of intervention was 71. This included 23 patients (24%) at the university hospital and 48 patients (56%) at the VA hospital. Only 4 patients tested PCR and EIA positive. The calculated prevalence of unsuspected hepatitis C is therefore 4/71 = 5.6% (95% CI 1.7–14.3%).

Although it is logical to assume additive increased risk of multiple hepatotoxic exposures, such as MTX + HCV, MTX + leflunomide (7), or MTX + leflunomide + HCV, it appears that MTX + HCV is not rapidly synergistic. All 3 liver biopsies performed on hepatitis C patients treated with MTX were benign. (Patient 1 discontinued MTX therapy for causes other than hepatotoxicity.) We hypothesize that in RA patients who are also hepatitis C positive, treatment with MTX for up to 1 year will not necessarily result in cirrhosis. Nevertheless, ACR guideline interventions for hepatitis screening may become more pertinent with an increase in the number of additional hepatotoxic agents and their combinations (e.g., MTX + leflunomide) available for RA.

In our study, the computer-based intervention proved to be maximally effective at the VA hospital, while the personal note reminder was only one-third as effective at the university center. Reasons for the considerable noncompliance at the university setting remain uncertain. We speculate that a more fragmented care at the university hospital, as opposed to the VA hospital, contributed to the poor outcome. At the university hospital the rotating house staff is generally responsible for patient care, whereas at the VA hospital, patient progress is managed by the rheumatology fellows, supported by a nurse manager in a clinic dedicated to following potentially toxic antirheumatic drugs. There were no VA patients taking MTX detected to have unsuspected hepatitis C of the 46 tested as a result of the quality improvement intervention. It could be argued that, despite the increased prevalence of hepatitis C in the VA (8), usual care worked. Perhaps, the hepatitis C screening was ordered on younger patients and those for whom clinicians suspected risk behaviors or experiences, thus infected patients were detected and not placed on MTX.

Overall, our study of ACR guideline compliance suggests the following: prevalence of unsuspected hepatitis C in unscreened RA patients is clinically meaningful (5.6%; 95% CI 1.7–14.3%) and treatment of hepatitis C-positive RA patients with MTX for up to 1 year will not necessarily result in liver cirrhosis.

A number of related questions remain unanswered. In particular, the incidence of liver disease in patients with multiple exposures to potentially hepatotoxic agents, such as MTX + HCV or MTX + leflunomide, await a larger multicenter or national registry study. The performance of a more conventional computer reminder, or suspended order for hepatitis C screening at the point of care, remains unknown. It might be less than the 100% efficacy of the information system based intervention described here.


  1. Top of page
  2. Introduction