Background In rheumatoid arthritis (RA), the long lasting effect of rituximab (RTX) on B lymphocyte depletion may influence the choice of biologic after RTX failure but there is currently no specific recommendation about the optimal strategy.

Objectives We assessed factors which may have influenced the choice of biologic following failure to RTX in RA patients and analysed the effectiveness of these biologics.

Methods A retrospective study of RA patients, who had started a new biologic during the year after RTX discontinuation, was conducted between 2009 and 2016 in 5 French rheumatology centers. We collected at baseline (at the time the biologic was introduced after RTX) and during follow-up (3, 6 and 12 months), data about patients and disease. Characteristics of the patients receiving tocilizumab (TCZ), abatacept (ABA) or anti-TNF following RTX, the EULAR response and retention rate were compared using univariate and multivariate analyses.

Results 152 RA patients (117 women, mean age 56.9±13 years) were analysed. RA duration was 14.3 years [IQR 7.8–19.8], mean DAS28 ESR at baseline 4.8±1.3. 57 patients (37.5%) received TCZ, 47 (31%) anti-TNF and 48 (31.2%) ABA following RTX. 87% had received anti-TNF prior to RTX. No significant difference in the biologics prescription profile was noted across centers. At baseline, sex, disease characteristics and activity, use of concomitant DMARDs or prednisone were not different between the 3 groups. There was no difference in the number of cycles or dose of RTX received and the number of previous anti-TNF treatment. Patients receiving ABA were slightly older (59.8 years vs 54 years for TCZ and 57.8 years for anti-TNF, p=0.06). Multimorbidity index (MMI) assessing the number of comorbidities was higher in the group ABA but not significantly (MMI count 2±1.7 for ABA, 1.6±1.5 for TCZ, 1.7±1.4 for anti-TNF, p=0.5). At 3, 6 and 12 months, DAS28ESR was lower in patients with TCZ as compared to those with anti-TNF or ABA (Table) but tender and swollen joint counts did not differ. The EULAR good-or-moderate response rates were similar across groups (Table). After adjustment on age, sex, disease duration, MMI count, number of previous anti-TNF, use of concomitant DMARDs and prednisone, drug retention rate at 1 year was significantly better for ABA compared to anti-TNF (p=0.02) and for TCZ compared to anti-TNF (p=0.04), but no difference was found between TCZ and ABA (p=0.62) (Figure).

• Download figure
• Open in new tab
• Download powerpoint

Conclusions Patients who received ABA following failure of RTX, seem to be older and with more comorbidities. Similar rate of good-or-moderate EULAR response was observed between patients receiving TCZ, ABA or anti-TNF. However, drug retention rate was better with TCZ and ABA compared to anti-TNF after RTX discontinuation.

Disclosure of Interest None declared