Early initiation of antiretroviral therapy (ART) in patients with TB reduces mortality in those with low CD4 counts, but increases the risk of paradoxical TB-IRIS. Prednisone reduces symptoms when used to treat TB-IRIS. We determined whether prophylactic prednisone in patients at high risk for paradoxical TB-IRIS safely reduces the incidence of TB-IRIS.
1:1 randomized, double-blind, placebo-controlled trial of prednisone (40 mg/day for 2 weeks then 20 mg/day for 2 weeks) started with ART in ART-naïve adults at high risk of TB-IRIS (within 30 days of TB treatment initiation and CD4 count ≤100/μl) followed for 12 weeks. Exclusion criteria included rifampicin resistance, neurological TB, Kaposi’s sarcoma, hepatitis BsAg+ and poor clinical response to TB treatment. Primary endpoint was development of TB-IRIS, defined using the International Network for the Study of HIV-associated IRIS (INSHI) consensus case definition, adjudicated by an independent committee. Final results are presented.
240 participants were enrolled: median age 36.8 (IQR=30-42.8), 60% men, median CD4 49/μl (IQR=24-86), and median HIV RNA 337,775 copies/ml (IQR=162,223-810,812). TB was microbiologically confirmed in 175. 18 were lost to follow-up or withdrew. TB-IRIS fulfilling INSHI criteria was diagnosed in 56 in the placebo arm (46.7%) and 39 in the prednisone arm (32.5%) (p=0.02, relative risk (RR)=0.70 (95%CI=0.51-0.96)). Open label corticosteroids to treat TB-IRIS was prescribed as necessary in 34 of the placebo arm (28.3%) and 16 (13.3%) of the prednisone arm (RR=0.47 (95%CI=0.27-0.83)). 4 deaths occurred in the placebo arm (1 attributed to TB-IRIS) and 5 in the prednisone arm (p=1.0). 27 were hospitalized (all-cause) in the placebo arm compared with 17 in the prednisone arm (p=0.10). Grade 3 adverse events occurred more frequently in the placebo arm (45.4% vs 29.4%, p=0.01), but grade 4 adverse events were similar by arm (8.4% vs 7.6%, p=0.81). Severe infections (AIDS-defining or invasive bacterial) occurred in 18 in the placebo arm and 11 in the prednisone arm (p=0.17). There was a trend towards fewer interruptions of ART or TB treatment in the prednisone arm (8.3% vs 15.8%, p=0.07).
In patients at high risk of paradoxical TB-IRIS, prednisone during the first 4 weeks of ART reduced the risk of TB-IRIS by 30% and further reduced the requirement for corticosteroids to treat TB-IRIS by 53%. The intervention was well-tolerated with no excess risk of infection or malignancy.