Case-control study on anti-TNF-associated TB patients. IRIS cases, defined with the following consensus criteria, were matched to two controls (anti-TNF-associated TB without IRIS). IRIS frequency was based on the French RATIO registry. Conditional logistic-regression identified IRIS risk factors.
Fourteen patients developed anti-TNF-associated TB–IRIS within medians of 45 [IQR 22–131] days after starting anti-TB therapy and 110 [IQR 63–164] days after the last anti-TNF infusion. Each case was matched to two controls by year of TB diagnosis. IRIS-associated factors were (odds ratio [95% CI]): disseminated TB (11.4 [1.4–92.2], P = 0.03), history of Mycobacterium tuberculosis exposure (12.7 [1.6–103.0], P = 0.02) and steroid use at the time of TB diagnosis (4.6 [1.2–17.2], P = 0.02). The RATIO registry IRIS frequency was 7%.
After stopping biotherapy, paradoxical anti-TNF-associated TB worsening occurred most often in patients with disseminated TB. Although diagnosis remains difficult, physicians must be aware of IRIS because prolonged anti-TB treatment is not needed but, paradoxically, immunosuppressant reintroduction may be.