GLASGOW – Treatment with tumor necrosis factor (TNF) inhibitors for rheumatoid arthritis is associated with a low risk of vasculitis-like events, according to a large analysis of data from the United Kingdom.
Investigators using data from the British Society for Rheumatology Biologics Register for Rheumatoid Arthritis (BSRBR-RA) found that the crude incidence rate was 16 cases per 10,000 person-years among TNF-inhibitor users versus seven cases per 10,000 person-years among users of nonbiologic disease-modifying antirheumatic drugs (nbDMARDs) such as methotrexate and sulfasalazine.
Although the risk was slightly higher among anti-TNF than nbDMARD users, the propensity score fully adjusted hazard ratio for a first vasculitis-like event was 1.27, comparing the anti-TNF drugs with nbDMARDs, with a 95% confidence interval of 0.40-4.04.
“This is the first prospective observational study to systematically look at the risk of vasculitis-like events” in patients with RA treated with anti-TNF agents, Dr. Meghna Jani said at the British Society for Rheumatology annual conference.
Dr. Jani of the Arthritis Research UK Centre for Epidemiology at the University of Manchester (England) explained that the reason for looking at this topic was that vasculitis-like events had been reported in case series and single-center studies, but these prior reports were too small to be able to estimate exactly how big a problem this was.
“Anti-TNF agents are associated with the development of a number of autoantibodies, including antinuclear antibodies and antidrug antibodies, and ANCA [antineutrophil cytoplasmic antibody],” she observed.
“We know that a small proportion of these patients may then go on to develop autoimmune diseases, some independent of autoantibodies,” she added. The most common of these is vasculitis, including cutaneous vasculitis.
Vasculitis is a somewhat paradoxical adverse event, she noted, in that it has been associated with anti-TNF therapy, but these drugs can also be used to treat it.
Now in its 15th year, the BSRBR-RA is the largest ongoing cohort of patients treated with biologic agents for rheumatic disease and provides one of the best sources of data to examine the risk for vasculitis-like events Dr. Jani observed. The aims were to look at the respective risks as well as to see if there were any particular predictive factors.
The current analysis included more than 16,000 patients enrolled in the BSRBR-RA between 2001 and 2015, of whom 12,745 were newly started on an anti-TNF drug and 3,640 were receiving nbDMARDs and were also biologic naive. The mean age of patients in the two groups was 56 and 60 years, 76% and 72% were female, with a mean Disease Activity Score (DAS28) of 6.5 and 5.1 and median disease duration of 11 and 6 years, respectively.
After more than 52,428 person-years of exposure and a median of 5.1 years of follow-up, 81 vasculitis-like events occurred in the anti-TNF therapy group. Vasculitis-like events were attributed to treatment only if they had occurred within 90 days of starting the drug. Follow-up stopped after a first event; if there was a switch to another biologic drug; and at death, the last clinical follow-up, or the end of the analysis period (May 31, 2015).
In comparison, there were 20,635 person-years of exposure and 6.5 years’ follow-up in the nbDMARD group, with 14 vasculitis-like events reported during this time.
A sensitivity analysis was performed excluding patients who had nail-fold vasculitis at baseline, had vasculitis due to a possible secondary cause such as infection, and were taking any other medications associated with vasculitis-like events. Results showed a similar risk for a first vasculitis event between anti-TNF and nbDMARD users (aHR = 1.05; 95% CI, 0.32-3.45).
Looking at the risk of vasculitis events for individual anti-TNF drugs, there initially appeared to be a higher risk for patients taking infliximab (n = 3,292) and etanercept (n = 4,450) but not for those taking adalimumab (n = 4,312) versus nbDMARDs, with crude incidence rates of 10, 17, and 11 per 10,000 person-years; after adjustment, these differences were not significant (aHRs of 1.55, 1.72, and 0.77, respectively, with 95% CIs crossing 1.0). A crude rate for certolizumab could not be calculated as there were no vasculitis events reported but there were only 691 patients enrolled in the BSRBR-RA at the time of the analysis who had been exposed to the drug.
“The risk of the event was highest in the first year of treatment, followed by reduction over time,” Dr. Jani reported. “Reassuringly, up to two-thirds of patients in both cohorts had manifestations that were just limited to cutaneous involvement,” she said.
The most common systemic presentation was digital ischemia, affecting 14% of patients treated with anti-TNFs and 14% of those given nbDMARDs. Neurologic involvement was also seen in both groups of patients (7% vs. 7%), but new nail-fold vasculitis (17% vs. 0%), respiratory involvement (4% vs. 0%), associated thrombotic events (5% vs. 0%), and renal involvement (2.5% vs. 0%) were seen in TNF inhibitor-treated patients only.