No Easy Answers for RA Patients Failing TNF Inhibitors

Few rheumatoid arthritis patients whose initial experience with tumor necrosis factor (TNF) inhibitors was suboptimal achieved “low disease activity” status upon switching to another biologic, whether it was another TNF inhibitor or another class of drug, researchers found.

Analysis of 939 patients in the registry database formerly known as Corrona (renamed CorEvitas earlier this year) indicated that only about 20-25% met criteria for low disease activity in two standard assessments 12 months after changing biologic treatments, no matter which kind they went on, reported Jeffrey Curtis, MD, of the University of Alabama at Birmingham, and colleagues.

Patients who switched to a different TNF inhibitor were generally in better shape initially than those moving to non-TNF inhibitors, but after adjusting for a host of potential confounders, it appeared that low disease activity remained out of reach for most patients for whom a change in treatment was advised.

The study appeared in ACR Open Rheumatology, the American College of Rheumatology’s open-access online journal.

Corrona began collecting data on real-world patients in 2001. In 2014, Curtis and colleagues started a comparative effectiveness study in 2014 utilizing this database to look for patients either starting on TNF inhibitors or who had previously used these agents and then switched to another biologic for any reason. The current report focused only on the latter group. Other inclusion criteria were disease activity assessed as moderate to high, and having regular clinic visits with the customary lab tests in rheumatoid arthritis.

After weeding out patients who had also used non-TNF inhibitor biologics previously and performing a propensity-matching analysis, Curtis and colleagues identified 434 patients who switched to a new TNF inhibitor and 505 whose new drug was a non-TNF inhibitor. The group’s starting hypothesis was that non-TNF inhibitors would prove more effective in bringing patients into low disease activity.

Disease activity was assessed in two ways: via the Clinical Disease Activity Index (CDAI), with scores below 10 defined as low, and the 28-joint Disease Activity Scale (DAS28) based on C-reactive protein levels with scores below 2.67.

Achievement of CDAI-assessed low disease activity was reached by 28% of those switching to non-TNF inhibitors versus 24% among those who went on a different TNF inhibitor. Rates were lower when disease activity was evaluated with the DAS28: 22% for non-TNF inhibitors versus 19% with new TNF inhibitors.

Either way, the differences fell short of statistical significance, both in the raw data and when adjusted for covariates and site clustering.

The most common drugs used when switching were:

• TNF inhibitors: infliximab (Remicade) 25%, adalimumab (Humira) 22%, certolizumab (Cimzia) 22%, etanercept (Enbrel) 21%, and golimumab (Simponi) 12%

• Non-TNF inhibitors: Abatacept (Orencia) 62%, tocilizumab (Actemra) 29%, and rituximab (Rituxan) 12%

The researchers also looked at patient subgroups for anything that might stand out, but that, too, failed to identify any patient or drug characteristics that clearly indicated superiority. Among these: the number of previous TNF inhibitors used, whether patients were on monotherapy, C-reactive protein levels above 3 mg/L, and use of rituximab or tocilizumab specifically upon switching.

In most cases, use of non-TNF inhibitors showed small numerical advantages that fell short of statistical significance. While that may suggest the study was underpowered, it appeared likely that any actual advantages were minimal, with one exception: for patients who had previously used two or more TNF inhibitors, the odds ratio for achieving low disease activity with a non-TNF inhibitor versus a different TNF inhibitor was 1.8, with the 95% confidence interval’s lower bound only a hair below 1.0.

Overall, “[t]hese results are perhaps at odds with prior cohort studies suggesting a benefit” with non-TNF inhibitors in patients with suboptimal responses to TNF inhibitors, Curtis and colleagues wrote. Among the possible explanations they suggested were that these earlier studies used different disease activity assessments and greater use of abatacept in the Corrona registry relative to other non-TNF inhibitors.