Why Did the Last Drug Fail? Save

When a patient fails a drug, we tend to consider what will be the next drug. Most of our clinics have patients on their second or third advanced therapy. Drug failure has become more common, not less. A question that we too rarely write in the notes is why did the last drug fail?

The Leeds specialist SpA service followed 396 patients with PsA or axSpA from 2000 to 2020 [2300]. Median retention of the first advanced therapy was 12 years in axSpA and 8 years in PsA. When looking at when patients started, primary non response to a first drug rose from 0% among those starting between 2000 and 2004 to 24% in the later cohorts. The disease did not change, we did. The authors suggest that patients and clinicians now tolerate imperfect responses less readily, that each additional option makes the next switch easier, and that some discontinuation is therefore iatrogenic. Their recommendation is to  report the reason for failure, not just the retention curve.

In our own 22 year Reading cohort, three in four patients stayed on a single advanced therapy and most transitions were within the TNF inhibitor class due to historical reasons of which options were available then, rather than a change in mode of action [0352]. The Czech ATTRA registry studied 190 patients with difficult to manage axSpA after two mechanisms had failed [0077]. Almost half had received two or more TNF inhibitors before changing the mechanism, and 41% returned to a TNF inhibitor after the second mechanism failed. JAK inhibition worked faster than IL17 inhibition at three months, but drug survival was similar. Cycling, it seems, is an international activity..

The French ROC SpA cohort offers a reason to stop and test [2284]. Of 265 patients, 21% had antibodies to their first TNF inhibitor. Among those switching to an IL17A inhibitor, ASAS40 at week 52 was 44.1% when antibodies were present, against 16.7% when they were not (OR 3.95). The logic is that if the drug was neutralised, the disease may still respond. If it was not, we may be facing refractory biology, and a new mechanism alone may not be enough. The numbers support that reading where barely one in six antibody negative switchers reached ASAS40, whichever drug came next. Those are the patients that we would want to focus within the ASAS difficult to manage framework.

In PsA, CorEvitas data showed almost identical 12 month persistence for IL17A and IL23 inhibitors, 56.2% against 57.4% [2288]. Roughly half of patients stopped their first drug within a year. No class is a guaranteed long term home. An updated systematic review and meta-analysis of randomised trials found that men with PsA had higher odds of response than women on almost every class: ACR20 1.42, ACR50 1.74 and minimal disease activity 1.94 [2309]. By class, ACR50 odds ran from 2.24 for TNF inhibition and 1.93 for IL17 inhibition down to 1.45 for JAK inhibition, with TYK2 inhibition the only class showing no difference at 1.00. Put together, these data suggest that the choice of next drug matters, but the reason for the last failure matters more. Two further abstracts sharpen the question, the first was a meta analysis comparing switching class with swapping within class after the first advanced therapy [2311] where both strategies demonstrated moderate drug survival and clinical responses. The second was a landmark analysis of disease duration and first biologic survival [2945] where the therapeutic window of opportunity operated through enhanced drug survival rather than superior clinical response magnitude.

The practical response to this new information is a short checklist before every switch. Was it primary or secondary failure? Were drug levels and antibodies checked? Is inflammation confirmed objectively? Are obesity, pain sensitisation or mood driving the score? Only then does the formulary come into it. A fourth biologic chosen for a reason is good medicine. A fourth biologic chosen because it is next on the list is a missed opportunity.