Background/Purpose: Although use of glucocorticoid increases the risk of tuberculosis (TB) disease, there has been few studies investigating its incidence and risk/benefit assessment of the prevention in rheumatic disease patients with this treatment. Furthermore, result of tuberculin skin test and interferon gamma release assay (IGRA) could be significantly influenced by steroids itself. In this study, we investigated the efficacy and safety of isoniazid (INH) monotherapy as a prophylaxis of TB disease in rheumatic disease patients receiving prolonged, high-dose steroids.

Methods: This single center cohort study analyzed 1618 treatment episodes with prolonged (4 weeks or more), high-dose (30mg/day or more of prednisone) steroids from 1160 patients between 2004 and 2016. IGRA was performed in 187 (11.6%) episodes and selection of patients for INH monotherapy and its duration was mainly determined by treating physician. Overall, 152 episodes (INH group) received isoniazid monotherapy with steroid while other 1466 episodes did no prophylaxis (control group). Primary outcome was 1-year incidence of TB disease between the two groups, which was compared using Cox regression. Incidence rate of adverse drug reaction (ADR) related to INH was also investigated. Risk-benefit analysis of INH monotherapy was performed by comparing the number needed to be treated (NNT) to prevent 1 case of TB disease vs. the number needed to harm (NNH) due to ADR. As part of a sensitivity analysis, propensity score matching was used to minimize baseline imbalances, and the same analysis was performed in post-matched population (n=147 in each group).

Results: Patients in the INH group were older (45.1 vs. 42.1 years) and more frequently had SLE (56.6% vs. 48.7%) and MPA (5.3% vs. 0.9%) than those in the control group. Median (IQR) steroid dose at initiation was also higher in the INH group (55.0 [30.0] vs. 60 [15.0] mg/day of prednisone). During a 1579.8 person-year, a total of 21 cases of TB disease occurred. Prophylaxis with INH trended toward a reduced 1-year incidence of TB in the multivariable model where SLE, cumulatively used steroid dose and risk factors for TB disease were adjusted (adjusted HR = 0.57 [95% CI 0.13 to 2.52]) (Figure). The sensitivity analysis performed in the post-matched population was consistent with the main results (adjusted HR = 0.67 [0.10 to 4.38]). In contrast, incidence rate of any ADRs during the INH treatment was 113.0 (90.5 to 139.3)/100 person-year including one case of fulminant hepatitis. Because absolute risk for TB disease was higher in the INH group, NNT was calculated as a negative value whereas NNH for any ADR and serious ADR were 2 (1.6 to 2.1) and 76 (32.0 to ꝏ), respectively.

Conclusion: Isoniazid monotherapy as a prophylaxis for TB in rheumatic disease patients receiving prolonged, high-dose steroids shows partial efficacy, but high incidence of ADR limits its role as a general practice.