Background/Purpose: Reactivation of latent tuberculosis infection (LTBI) is of concern in patients treated with TNFi. Conversely, TB chemoprophylaxis (CP) is time consuming, delays the initiation of treatment, adds to the overall cost of treatment, and carries risk of adverse events in its own right. Since 2002 our national guidelines require following a two-step screening algorithm prior to the initiation of the first TNFi. The first step includes tuberculin skin test (TST), and a chest X-ray (CXR). If TST < 5 mm, and the radiologist finds no changes consistent with TB on CXR TNFi is prescribed. If any test is abnormal, the patient is evaluated by a pulmonologist who usually orders QuantiFERON TB Gold IT (QF) and decides whether TB CP (rifampicin/isoniazid 600/300 mg qd for 3 months) is required prior to TNFi treatment.

In February 2012 we showed that in a setting with low background annual TB incidence rate (IR; e.g. 8.4 per 10⁵) following of this algorithm resulted in TB IR of 0.11 (95% CI 0.01–0.38), and 0.16 (95% CI 0.02–0.58) per 100 person years (PY) overall, and in rheumatoid arthritis (RA) patients, respectively. The costly QF and CP were required in 13.9%, and 5.2% of patients, respectively. Our aim was to revaluate the performance of this algorithm.

Methods: We cross-linked the data from the mandatory national registry of patients with rheumatic diseases treated with biologics and national TB registry to identify cases of TB in patients who were ever treated with TNFi.

Results: 1203 patients were treated with at least one TNFi for 3146.5 PY (Table 1). Flow of patients through the screening algorithm and case patient characteristics are depicted in Figure 1. QF was performed in 214/1174 (18.2%). 88/1203 (7%) patients received CP. Four cases of TB were identified; hence the incidence rates were 0.13 (95%CI 0.03–3.2), and 0.22 (95% CI 0.06–0.57) per 100 PY overall, and in RA patients, respectively. Cases of TB were only observed among RA patients. The TB IR for certolizumab vs. other TNFi prescribed for RA was 2.1 (95% CI 2.56–7.40) vs. 0.12 (95% CI 0.01–0.43) per 100 PY (p=0.016 Fisher’s exact test). The first two received appropriate CP (adherence was good, isolated M. tuberculosis strains were susceptible to CP) prior to TNFi; screening was concluded after 1^st, and 2^nd step for the third and fourth patient, respectively. Interestingly, the third case was screened for TB again before switching to rituximab. At repeated screening two months prior to TB diagnosis the TST was 10 mm, CXR –. and QF–.

Conclusion: At follow-up our two-step algorithm is still performing well. Further vigilance is warranted, especially in RA patients and those treated with CPZ.

Table 1.

Figure 1