| Relative values of different outcomes | The outcomes considered were:
The group members agreed to focus on cirrhosis and fibrosis as these are the key clinical outcomes. Evidence for short term liver toxicity (as indicated by rise in transaminases) has been reviewed in chapter 9. |
| Trade off between clinical benefits and harms | Methotrexate is a useful drug for long term disease management. The absolute risk of clinically significant liver fibrosis or cirrhosis due to methotrexate per se is unknown and maybe lower than is perceived by patients and some clinicians. In clinical practice, methotrexate may not be prescribed in the presence of risk factors for liver fibrosis (for example, hepatic steatosis in relation to obesity, diabetes) although the evidence does not support this. Complete avoidance or minimal intake of alcohol is standard advice for patients taking methotrexate and is a barrier to some people who would benefit from using methotrexate. The evidence did not support this and with appropriate patient selection and strict monitoring, alcohol may be allowable. However, combined with the evidence from Chapter 9, the GDG considered it important that the potential risk of liver damage for people with psoriasis taking methotrexate should be highlighted, although specific risk groups have not been identified. |
| Economic considerations | No evidence was available to inform the GDG about the economic impact of methotrexate-induced hepatotoxicity, nor on how lower or higher risks would impact its cost-effectiveness as a treatment for people with psoriasis. Economic evaluations assessing the cost-effectiveness of methotrexate compared to other systemic biological and non-biological treatments have not captured risks of hepatotoxicity due to inconclusiveness of the clinical evidence and the complexity it would add to any decision model. These same evaluations have found methotrexate to be cost-saving, or more cost-effective, than alternatives, including no treatment, ciclosporin and various biological therapies. Its dominance over most other therapies is largely driven by its extraordinarily low acquisition cost compared to other drugs. The GDG concluded that despite the potentially higher risks of liver toxicity, methotrexate is still likely to be an optimal treatment and that the additional costs of extra monitoring were unlikely to alter this conclusion. |
| Quality of evidence | Many of the studies were published pre-1990 and had small sample sizes. The studies did not clearly state whether confounding variables had been assessed, including whether liver pathology was present prior to methotrexate administration; therefore consideration was given to whether the GDG could be confident the effect is due to the risk factor reported.
There are limitations with assessing liver damage using liver biopsy due to variation in sampling technique (which was poorly reported) and patch pathological change. There is also variation in the histology grading scales used in the different studies, and it was not possible to map them to a common scale.
Some studies had performed statistical analyses (in most cases by looking at the degree of correlation between the risk factor and the outcome), while others had not (in which case results are reported as an apparent or no apparent effect). The GDG noted that an apparent effect could have been non-significant.
Studies used different definitions of alcohol consumption and some definitions are vague. Also, the intake is often based on self-reporting which may be inaccurate. However, there was no consistent pattern to suggest that studies using a stricter definition of high alcohol intake were the ones that demonstrated a link.
Studies also varied in the route of administration and dosing schedule of methotrexate and it was unclear whether folic acid had been used.
There is also a risk of selection bias as those with symptoms, signs or abnormal laboratory results are more likely to have a liver biopsy. Therefore, people with psoriasis at higher risk of liver damage may have been over represented and the risk may be over-estimated compared with the general population sample.
In light of these issues, the group interpreted the evidence with caution.
For alcohol as a prognostic factor:
Most of the data related to alcohol intake before methotrexate use, but intake during methotrexate use may be more important. Data for intake both before and during methotrexate use were given in 2 studies(Nyfors 1977 and Boffa). These data suggested that the intake during therapy may be more of a risk for liver damage (e.g., those with the greatest decrease in alcohol intake showed the lowest liver histology score [Boffa] and there was a significant link between liver damage and alcohol intake during therapy but only a modest apparent link with alcohol intake prior to therapy [Nyfors 1977])
For cumulative methotrexate dose as a prognostic factor:
The Berends study showed that biopsy grade progression levelled out above 6000mg but this was defined as progression to grade>1 (not fibrosis) and people could still have been progressing to higher severity within the category of grade >1.
The Newman study reported that the probability of normal biopsy (Grade I or II) dropped below 50% at 3115 mg.
The heterogeneous results were not explained by treatment duration, age, treatment regimen or mean cumulative dose (i.e., there was no consistent pattern, for example, those that showed a link used oral methotrexate or had a higher mean cumulative dose). The variable results could be due to individual differences in tolerance of high methotrexate dose but none of the included studies investigated this.
The GDG noted that all three of the prospective studies, and both studies that adjusted for confounders, showed no significant link to cumulative methotrexate dose..
Summary:
From the studies, there was no consistent and methodologically robust evidence to conclude that for people with psoriasis taking methotrexate there are any groups who are at higher risk of methotrexate-induced liver damage. The risk of liver damage is already raised among people with psoriasis. Large, well-designed studies would need to be performed in order to correct for all confounders. At present there may be a reluctance in clinical practice to use methotrexate in people with psoriasis who have risk factors and/or reluctance to continue methotrexate with high cumulative doses (>3g). There is no strong evidence to support this.
Overall, the evidence for risk factors is poor, and there are a number of important confounders in the studies that make it difficult to evaluate the role of methotrexate itself. There is no consistent evidence that any of the risk factors, including cumulative dose of methotrexate, increase the risk of liver fibrosis or cirrhosis. Therefore, the GDG did not wish to make a recommendation about at risk groups.
From the evidence, there are no groups in whom the GDG would not recommend methotrexate. There is no consistent evidence that any specific group is at an increased risk. Therefore risk factors cannot be used as a screening tool. All patients should be evaluated for liver damage prior to and after commencing treatment.
The GDG agreed there was no consistent and methodologically robust evidence to conclude that that for people with psoriasis taking methotrexate there are any groups who are at particularly high risk of methotrexate-induced hepatotoxicity, including cumulative dose of methotrexate. However, all people with psoriasis may be at increased risk of liver disease so large, well-designed studies would need to be performed in order to properly correct for all the confounders.
Recommendations about monitoring for hepatotoxicity can be found in chapter 12. |
| Other considerations | The GDG considered referencing the government guidance on recommended daily alcohol intake. It was felt that this may not be appropriate, as the recommended daily amounts of alcohol are applicable to the general population, not people with psoriasis. Evidence from chapter 6.4 indicating an increased risk of alcohol-related death would support this contention. The GDG felt there was a need to act responsibly when formulating the recommendations.
The evidence did not show any consistent pattern that alcohol intake increased the risk of liver damage in people with psoriasis on methotrexate, but there were methodological limitations which meant that the GDG had little confidence in the results. As such the GDG were unable to make a recommendation either way (i.e. that alcohol should be completely avoided, or that alcohol was permissible during therapy).
Methotrexate induced liver problems are an important concern to both clinicians and patients and a common cause for patients to decline therapy and/or clinicians to stop/not offer this therapy. Well conducted research is required to establish the risk of liver disease in people with psoriasis per se, whether methotrexate adds to the risk, and the contribution of factors such as alcohol, obesity or diabetes to any identified risk. Research in this area would need to: involve large numbers of patients given that the absolute risk of liver fibrosis may be low; control properly for confounders (obesity, diabetes, alcohol); and use validated outcomes that overcome the identified difficulties in the existing studies (namely different reporting scales and lack of clinically relevant outcomes). |
