Background/Purpose

In this 48-week, double-blinded, non-inferiority trial, 353 methotrexate suboptimal-responders were randomized to two treatment strategies, either the addition of sulfasalazine and hydroxychloroquine (triple therapy [T]) or the addition of etanercept (E).  Participants without marked improvement in DAS28 at 24 weeks switched to the alternate treatment, while maintaining the blind.  This created 4 treatment subgroups: no switch (T-only, E-only), and switch (T-E, E-T).  At 48 weeks treatment blind was broken. Follow-up data on major study outcomes from the observational sub-cohort of participants who consented to long term f/u is presented.

Methods

DAS28 scores, joint assessments, laboratory values, visual analog scale (VAS), HAQ, Health Utilities Index (HUI),  EQ5D, and changes in therapy (grouped as either any conventional (DMARDs) or to any biologic were collected every 24 weeks up to week 120 or end of study.  The last participants enrolled had minimal potential for follow-up, decreasing overall mean follow-up. We compared two follow-up groups: participants on E at 48 weeks (E-only and T-E) who then changed to a DMARD and participants on T at 48 weeks (T-only and E-T) who then changed to a biologic.  Study measure results were compared by simple ANOVA at each follow-up time across the four treatment subgroups.  The stability of measures over time within each subgroup was assessed by repeated measures ANOVA.  

Results

Of the 353 participants, 289 with 48 week data agreed to extend follow-up.  Treatment subgroups consisted of 106 T-only, 103 E-only, 42 T-E, and 38 E-T.  Data were available on 213 (74%), 168 (58%), and 162 (56%) at 72, 96, and 120 week, respectively.  For week 120, data available within the subgroups ranged from 46% to 59%. With respect to DAS28, the censoring patterns appeared completely at random (at each time the null hypothesis was not rejected).  Measures of joint assessment, laboratory values, VAS, HAQ, HUI, and EQ5D were stable over time within subgroups.  DAS28 scores and various component scores were slightly higher over time among switchers than non-switchers, but not significantly (NS).  Within each subgroup no symptom or disability index rose significantly over time. 218 patients had follow-up data on medication use on or after 48 weeks. In this group at week 48, 90% of 101 patients who ended the study on T continued on DMARDs compared to 48% of the 117 patients who ended the study on E and continued on biologics. Of those who changed treatment at week 48, more changed from E to a DMARD than from T to a biologic (chi-square test p<0.01). After week 48 medication change rates remained similar in the two groups.

Conclusion

In this observational cohort, major study outcomes including DAS28 remained stable over time. Those patients who switched therapy at week 24 of the interventional trial had slightly higher DAS28 and component scores. Censoring from the treatment subgroups was similar to the overall cohort indicating that censoring was independent of treatment assignment. Of those with post-study medication data, there was a larger change to DMARDs than to biologics immediately after ending the study. The reasons for this are being explored.

---

« Back to 2014 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/follow-up-data-on-the-rheumatoid-arthritis-comparison-of-active-therapies-trial-observational-cohort/