To assess factors influencing the choice and effectiveness of biological disease-modifying antirheumatic drugs (DMARDs) following failure of rituximab (RTX) in rheumatoid arthritis (RA), taking patient profile into account.

In a retrospective, multicenter study, data about RA patients starting a new biologic during the year after RTX discontinuation were collected at baseline (when the biologic was introduced after RTX), and during follow-up (3, 6, and 12 months). Characteristics of patients receiving tocilizumab (TCZ), abatacept (ABA), or a TNFα inhibitor (TNFi), EULAR response, and retention rate were compared using multidimensional factorial analysis for patient profiles and multivariate analysis including propensity score built on the patient profile.

Among 152 patients analyzed (37.5% TCZ, 31.6% ABA, 30.9% TNFi), sex, disease characteristics and activity, concomitant DMARDs or glucocorticoids, and previous use of RTX and TNFi were similar at baseline. Patients receiving ABA were slightly older. Multimorbidity index was higher but not significantly different. Multidimensional factorial analysis showed a distinct profile of patients receiving ABA, characterized by older age, more men, more smokers, more comorbidities, and higher anti-cyclic citrullinated peptide antibody. At 1 year, drug retention was higher for ABA than TNFi after adjustment for disease duration, concomitant DMARDs, glucocorticoids, and propensity score (P=0.04). Tolerance and serious infections were similar among groups.

The profile of patients receiving ABA following failure of RTX differed from TNFi and TCZ using multidimensional factorial analysis. After adjustment for propensity score, drug retention rate remained higher with ABA than TNFi.