Background Although tumour necrosis factor-α inhibitor (TNFi) therapies for the treatment of rheumatoid arthritis (RA) are known to induce anti-nuclear antibodies, TNFi-induced lupus is a rare but well recognised phenomenon. The incidence and characteristics of RA patients that develop this adverse event remain poorly defined.
Objectives The aims of this study were to (i) compare the incidence of drug induced lupus (DIL) in subjects with RA treated with TNFi to those receiving non-biologic drugs (nbDMARDs) and (ii) characterise these adverse events.
Methods The BSRBR-RA is a prospective cohort study to which patients were recruited between 2001-2014. This analysis included two cohorts: (i) patients starting TNFi (adalimumab, etanercept, infliximab, certolizumab) and (ii) a biologic-naïve comparison cohort receiving nbDMARDs. Additional information from consultants was sought for relevant events and classified as DIL if they met ≥2 criteria as described in Dubois' guidelines. Briefly these include (i) treatment with known lupus-inducing drug for ≥1 month: (ii) mucocutaneous/constitutional symptoms within the lupus spectrum (iii) multisystem involvement- neurological/renal/skin manifestations (iv) immunological/haematological profile consistent with lupus (v) improvement/resolution of symptoms after drug discontinuation. Only biologic-naïve patients at baseline were included. The risk of an event was compared between the two cohorts using Cox proportional hazard models, adjusted using deciles of propensity scores. Events were attributed to TNFi therapy if they occurred within 90 days of being on therapy. Follow-up was censored at first event, switching to another biologic, death, last returned clinical follow-up or 31/05/2014, whichever came first.
Results There were 55 incident DIL cases: 5 in 3673 nbDMARD patients and 50 in 12,289 first TNFi-treated subjects (Table). After propensity score adjustment the hazard ratio of DIL in patients on TNFi vs. nbDMARD was 3.03 (95% CI 0.91-10.08). Over half of events following TNFi and nbDMARDs were limited to cutaneous manifestations and 14-20% fulfilled SLE classification criteria (Table). In the nbDMARD and TNFi-treated cohorts, 1 and 5 patients developed biopsy proven lupus nephritis respectively. There were two deaths in the TNFi-induced DIL group, one directly attributed to the event.
Conclusions The overall incidence of DIL in a TNFi-RA treated cohort was rare at 1/1000 person years. There was an increased risk of DIL on TNFi vs. nbDMARD although this was not significant after adjustment. TNFi-induced lupus encompasses a spectrum of characteristics, but the majority do not fulfil SLE classification criteria.
Acknowledgements MJ is a Medical Research Council Clinical Training Fellow supported by the NW England Medical Research Council Fellowship Scheme in Clinical Pharmacology and Therapeutics ([grant number G1000417/94909], ICON, GSK, AstraZeneca and the Medical Evaluation Unit). This work includes independent research supported by the NIHR BRU Funding Scheme.
Disclosure of Interest M. Jani Speakers bureau: Pfizer, L. Kearsley-Fleet: None declared, W. Dixon: None declared, I. Bruce Grant/research support from: GSK, Roche, Pfizer, UCB, Genzyme/Sanofi, H. Chinoy Grant/research support from: Novartis, Pfizer, Abbvie, Speakers bureau: Pfizer, Roche, MSD, Janssen, Abbvie, UCB, Servier, A. Barton Grant/research support from: Abbvie, Pfizer, Eli-Lilly and Sanofi-Aventis, M. Lunt: None declared, D. Symmons: None declared, K. Hyrich Grant/research support from: Pfizer, Speakers bureau: Abbvie