Smoking and response to rituximab in rheumatoid arthritis: results from an international European collaboration

To investigate whether smoking habits predict response to rituximab (RTX) in rheumatoid arthritis (RA).

We included patients from the CERERRA international cohort receiving the first treatment cycle with available smoking status (n = 2481, smokers n = 528, non-current smokers n = 1953) and at least one follow-up visit. Outcome measures were change in Disease Activity Score based on 28-joint count (ΔDAS28) and European League Against Rheumatism (EULAR) good response at 6 months, with non-current smokers as the referent group.

Compared with non-smokers at baseline, smokers were more often rheumatoid factor (RF)/anti-citrullinated protein antibody (ACPA) positive and males, had shorter disease duration, lower DAS28 and Health Assessment Questionnaire (HAQ) score, a higher number of prior biological disease-modifying anti-rheumatic drugs, and were more likely to receive concomitant conventional synthetic disease-modifying anti-rheumatic drug (csDMARDs). Disease activity had decreased less in smokers at 6 months (ΔDAS28 = 1.5 vs 1.7, p = 0.006), although the difference was no longer significant after correction for baseline DAS28 (p = 0.41). EULAR good response rates did not differ between smokers and non-smokers overall or stratified by RF/ACPA status, although smokers had lower good response rates among seronegative patients (ACPA-negative: 6% vs 14%, RF-negative: 11% vs 18%). Smoking did not predict good response [odds ratio (OR) = 1.04, 95% confidence interval (CI) = 0.76-1.41], while ACPA, DAS28, HAQ, and concomitant csDMARDs were significant predictors for good response. However, when stratified by country, smokers were less likely to achieve good response in Sweden (unadjusted OR = 0.24, 95% CI = 0.07-0.89), and a trend was seen in the Czech Republic (OR = 0.45, 95% CI = 0.16-1.02).

In this large, observational, multinational RA cohort, smokers starting RTX differed from non-smokers by having shorter disease duration and lower disease activity, but more previous treatments. The overall results do not support smoking as an important predictor for response to RTX in patients with RA.