The Fallacy of Biologic Treatment to Prevent Rheumatoid Arthritis Save

We have made amazing strides in the treatment of rheumatoid arthritis with a plethora of highly effective treatments now in our armamentarium. The window of opportunity concept in RA then shifted our focus to trying to treat disease earlier and more aggressively to improve long term outcomes. A natural development of this was to reconsider the possibility of therapeutic interventions aimed at the prevention of rheumatoid arthritis.

Further data from the APIPPRA study of abatacept is presented at Sunday’s oral abstract session “RA – Treatments I: Novel RA Treatments & Mechanisms of Action” (Abstract 0835). This study consisted of 213 participants with pre-clinical RA – essentially seropositive arthralgia in this study. The intervention was abatacept 125mg weekly subcutaneously vs placebo. Treatment was for 52 weeks with a further 52 weeks of follow-up following treatment withdrawal – participants were drug free for this period.

At first glance the results are encouraging. At 52 weeks 29% of placebo treated patients vs 6% of abatacept treated patients had developed RA. But consider what is actually happening here. The abatacept treated patients are on a drug we know is highly effective at treating RA. If you give such an effective drug can RA develop? Well yes it can, abatacept is not effective for everyone. If we assume such a phenomenon, then is the approximately 80% response rate reasonable? It would seem so to me. 

But hold on I hear you say, there was a difference after 52 further weeks of no treatment! 

Well yes there was, a much smaller difference: 37% vs 25%. So in that second 52 weeks a further 8% of placebo patients and 19% of abatacept patients developed RA. A picture paints a thousand words, and the survival curve here is striking – a wide difference at 52 weeks followed by increasingly rapid convergence as we approach 104 weeks. Now you may say that a 12% absolute difference is still worthwhile. If it were maintained longer term it may be but it appears far more plausible that it will not be. Why do I say this and how do we explain the difference in RA rates? Well we actually don’t have to, we already know all about this phenomenon in another setting – biologic withdrawal trials. Consistently when we withdraw biologics not all patients relapse immediately, there is a gradually increasing relapse over time. At 1 year in biologic withdrawal trials not all patients have relapsed but the longer we observe the more that do so and my personal opinion is that followed long enough almost all who really have RA will relapse. I argue that what we are seeing in APIPPRA and similar studies is a reflection of temporary drug free remission post-biologic withdrawal.

It is time to stop fooling ourselves that we can prevent RA with our existing biologic agents. 

I think there is a reasonable argument that maybe we should consider treating this patient group – at least that is a relatively clear discussion where we can weigh risks and benefits. I think the emotive framing of RA “prevention” should be abandoned as existing agents manifestly do not achieve this. Let’s call a spade a spade – we are successfully treating RA  as it is developing and this may result in patients never experiencing symptoms. Is the number needed to treat sufficient to warrant such treatment? At present I don’t think so, but at least that’s an honest debate.