Risk of tuberculosis in patients with rheumatoid arthritis: data from a centralized electronic database in Hong Kong

Abstract

Objective: To determine the risk of tuberculosis (TB) in patients with RA by comparing with those with non‐specific back pain (NSBP), and to identify the risk factors in the RA group. Method: Medical data were retrieved from a centralized electronic database. A total of 1099 patients with RA and 2489 patients with NSBP were identified. Clinical data including age, sex, smoking and drinking status, duration of follow‐up, comorbidities, use of individual disease modifying anti‐rheumatic drugs (DMARDs) and other immunosuppressive states were retrieved. Risks of TB in patients with RA and NSBP were compared by propensity score (PS) adjusted analysis with known or potential risk factors for TB. By adjusting the PS, the Cox regression model was used to determine the risk of TB and associated risk factors in patients with RA. Results: There were 14 cases (1.3%) of TB in RA group and 25 cases (1.0%) in NSBP group. Mean duration of follow up was 12.8 ± 6.0 years (10.6 ± 6.0 years in RA group; 13.7 ± 5.8 years in NSBP group). The risk of TB in patients with RA was 2.53 times higher (HR 2.53; P < 0.01) than in patients with NSBP. After excluding patients on DMARDs, the risk became similar (HR 2.72; P = 0.11). Independent risk factors found were etanercept (HR 7.16; P = 0.02), and previous TB infection (HR 25.23; P < 0.001). There were 5/13 (38.5%) extrapulmonary involvement in patients with RA and 4/25 (16.0%) in patients with NSBP. Conclusion: The risk of TB is increased in patients with RA. Anti‐TNF therapy was an identified risk factor.