10 April 2019

The hunt for the perfect predictor

APLAR Rheumatoid Arthritis

All clinicians aspire to precision medicine, but that’s not what is happening now in the treatment of rheumatoid arthritis, says Anne Barton from the University of Manchester.

In fact, if you look at the top 10 bestselling RA drugs in the US and their success rates, “what we practise at the moment is imprecision medicine”, she told the APLAR conference in Brisbane yesterday.

A professor of rheumatology at the Arthritis Research UK Centre for Genetics and Genomics and lead researcher at the Centre for Musculoskeletal Research, Professor Barton said the hunt was on for reliable biomarkers that would allow clinicians to predict a patient’s treatment response.

“At the moment,” she said, “what we do is treat patients with drugs in the order they came to market rather than by any rational prescribing method.”

Ideally we would use the information offered by genetics, epigenetics, transcriptomics and proteomics to find the right treatment, she said.

Since early effective therapy is known to prevent long-term damage and disability, it would be helpful to stratify patients early into those who would respond to methotrexate, or would benefit from the much more expensive biologics including tumour necrosis factor inhibitors.

This is the aim of the MATURA (MAximising Therpeutic Utility in RA) study, in which Professor Barton is a lead researcher. Her arm of the study uses pre-treatment blood samples to look for genetic, genomic and proteomic markers that correlate with TNF-I treatment outcomes.

“We’re looking for genetic variants in patients who respond well compared with patients who respond badly,” she said.

“If we can find these variants then that is an ideal biomarker: your genes do not change, you’re born with them. It’s a very stable biomarker. That would be perfect for the NHS because genetics is cheap.”

To date, however, no markers had been found that all researchers could agree were associated with treatment response.

Professor Barton said genetic biomarkers were just one possible kind, but they were a good example since – to the surprise of some in the audience – “genetics is really simple”.

In any such study, she said, researchers had to be careful about their outcome measures and adherence.

The popular DAS-28 measure of RA, since it included a component of patient-reported pain, could be easily gamed – clinicians had admitted squeezing joints harder to get a higher score and cross the threshold to prescribe biologics.

“Ideally you need a biologic endophenotype,” she said.

And non-adherence to treatment could bias a study by masking an association between a biomarker and treatment response. Clinicians had the power to increase adherence, she said, through techniques such as motivational interviewing.

While the work had far to go, Professor Barton concluded, “there are glimmers of hope and the hope now outweighs the hype”.