Background/Purpose:

To evaluate the rate of tuberculosis (TB) in biologic users for rheumatic diseases in South Africa, a TB endemic country, the effectiveness of our latent TB (LTB) screening program, risk factors and clinical outcome.

Methods:

Documented cases of TB were collected from the South African biologics registry (SABIO), practicing rheumatologists and pharmaceutical companies. Data on demographics, LTB screening tests and prophylaxis, biologic and DMARD therapies, TB diagnosis and treatment outcomes were recorded. A matched control arm evaluated the risk of TB in non-biologics users.

Results:

96 TB cases were collected from June 1999 to June 2017 (RA=55 (57%), AS=27 (28%), PsA=4 (4%), JIA=10 (10%)). The rate of TB was 1,240/100,000 person-years for all biologic users (n=96) compared to the control arm of 0/100,000 years (n=0) with an incidence rate difference of 0.0124 (95% CI 0.007 to 0.018, p<0.0001). Of these, 60/96 (62.5%) had pulmonary and 36/96 (37.5%) had extra-pulmonary disease. Reactivation TB occurred in 45/96 (51%) cases, despite a vigilant LTB screening program; new TB in 49/96 (47%) cases and 2 were undetermined. TB occurred in all 7 biologics licenced for use in SA (adalimumab 48, infliximab 15, golimumab 3, etanercept 19, tocilizumab 2, abatacept 5 and rituximab 4) with the majority from monoclonal TNF inhibitors (1,683/100 000 person-years) compared to etanercept (861/100,000 years) and non-TNF inhibitors (681/100,000 years). The incidence rate ratio (IRR) for monoclonal inhibitors compared to etanercept was 1.96 (95% CI 1.16 to 3.45, p=0.005) and 2.47 (95% CI 1.29 to 5.19, p=0.002) compared to non-TNF inhibitors. There was no significant difference between non-TNF inhibitors and etanercept (IRR 0.79; 95% CI 0.34 to 1.75, p=0.336). From registry data, it was extrapolated that 625 from 4830 patients (12.9%) screened LTB positive and were treated, yet 14 still developed TB (9 reactivation and 5 new onset TB). The majority (77) of TB cases, screened negative and screening was not done in 5. Steroid use, methotrexate use and male gender were significantly associated with acquiring TB (OR = 6.12; p<0.001, OR=7.5; p<0,001 and OR = 1.82; p=0.005 respectively), while the underlying rheumatic condition, race and geographic region were not. Two drug resistant TB cases and 6 deaths were recorded.

Conclusion:

TB poses a significant risk to all biologics users, including non-anti TNF’s in SA, a TB endemic country, despite our screening program. Concomitant methotrexate and steroid use further increase this risk.

Comparison of TB rates in biologic users across registries

TB rates of individual biologic agents (SABIO)