This cohort study to identify high risk patients turned up some surprises.
Drug monitoring seems to be a very individual thing in rheumatology.
When Medicare used to provide feedback about testing, I was on the 10th percentile. Depending on your viewpoint, that would either suggest that I was undertesting, got GPs to do all the work (good luck with that now), or preferred to target testing to higher risk patients.
When it comes to methotrexate (or any other drug for that matter), I suspect if you put 10 rheumatologists in a room you would get 10 different answers. The ARA guidelines sensibly reflect this as they are quite vague:
- As methotrexate may affect the liver and blood cells, you MUST have regular blood tests during your treatment. This is very important, as you may not get symptoms with some of these problems.
- Full blood count and liver function tests will be individualised by your doctor according to your risk.
So, methotrexate requires regular individualised testing. Is that second weekly for life as some GPs do? How do you individualise it?
My experience was that if bloods test didn’t change in the first six months, they rarely did, and usually the cause was not methotrexate. So I usually do monthly for the first six months and then once or twice a year after that.
The recent BMJ paper by Georgina Nakafero et al is very helpful in this regard. The authors already knew that serious side effects were uncommon in the longer term so wanted to see if they could individualise risk.
In the UK, despite this, it seems guidelines mandate third monthly testing for life so this might be an opportunity to modify low-value care.
They used a large GP-based development cohort and then a validation cohort which is gold practice now in epidemiology due to the risk of false positives leading to a problem with reproducibility.
Both had large numbers of patients and methotrexate discontinuation so there was ample power. The study duration was from initiation and then followed for up to five years.
Discontinuation due to serious side effects occurred in 6.3% (cytopenia, AST/ALT >100 or worsening renal function) but they didn’t say when these occurred, although they did model for it. The model fitted numerous predictors and explained most of the variation.
- Discontinuation MORE likely:
Methotrexate dose, smoking, hazardous alcohol, psoriasis, psoriatic arthritis, spondyloarthritis, chronic kidney disease stage 3, statin, paracetamol, PPI, antiepileptic and early in the treatment phase.
- Discontinuation LESS likely:
Females, BMI, low or moderate alcohol(!), polymyalgia rheumatica or giant cell arteritis, increasing age, NSAID, hydroxychloroquine and sulfasalazine.
There were some surprises here.
I have been reluctant to use MTX in older patients, those with CKD and/or more than one drink a day, but only CKD is justified. Indeed, alcohol in smaller amounts may actually be helpful.
PsO and PsA do need more monitoring, and this is the group that surprise you with cirrhosis.
The computer spits out the interaction with NSAIDs, but this study suggests a positive interaction.
We have known about HCQ being protective for the liver for many years, but sulfasalazine was a surprise, as was paracetamol, statins and antiepileptics.
The authors’ conclusion that we can rationalise monitoring for patients based on their overall risk and time on MTX seems sensible. The other option is not to give it to certain groups.
- Nakafero G, Grainge M J, Williams H C, Card T, Taal M W, Aithal G P et al. Risk stratified monitoring for methotrexate toxicity in immune mediated inflammatory diseases: prognostic model development and validation using primary care data from the UK. BMJ 2023; 381:e074678
Graeme Jones is professor of rheumatology and epidemiology with the Menzies Institute for Medical Research.
