Preventing tuberculosis-associated immune reconstitution inflammatory syndrome in high-risk patients:  A randomized placebo-controlled trial of prednisone (Phase III Trial)


To determine whether the addition of prednisone to the first 4 weeks of antiretroviral therapy (ART) reduces the risk of paradoxical TB-IRIS in HIV-infected patients being treated for TB who are at high risk of developing TB-IRIS (CD4 <100 cells/μl and starting ART within 30 days of TB treatment).


A randomized double-blind placebo-controlled trial to evaluate the incidence of paradoxical TB-IRIS over the first 12 weeks of ART in participants who receive a 4 week course of prednisone versus participants who receive a 4 week course of placebo.

Primary efficacy endpoint:

The development of paradoxical TB-IRIS within 12 weeks of starting ART (defined using the International Network for the Study of HIV-associated IRIS (INSHI) consensus case definition).

Secondary efficacy endpoints:
Safety and tolerability endpoints:
Sample size:

240 participants will be enrolled over 13 months. Each participant will be followed for 12 weeks.


HIV-infected, ART-naïve adult (≥ 18 years) patients diagnosed with active tuberculosis who have a CD4 < 100 cells/μL and who start ART within 30 days of starting TB treatment. Other inclusion criteria include: diagnosis of TB (smear, culture, Xpert MTB/RIF test, histology or strong clinical and radiological evidence of TB with symptomatic response to TB treatment), eligible for and consent to starting ART and written informed consent for trial. Exclusion criteria include: Kaposi’s sarcoma, pregnancy, TB meningitis or tuberculoma at TB diagnosis (because these patients receive corticosteroids), known rifampicin-resistant TB, being on corticosteroids for another indication within the past 7 days, on other immunosuppressive medication within the past 7 days and uncontrolled diabetes mellitus.

TB treatment and ART:

TB treatment will be prescribed and monitored by the clinical staff in the local HIV-TB clinic. TB treatment will be given according to South African Department of Health guidelines. This involves rifampicin (R), isoniazid (H), ethambutol (E) and pyrazinamide (Z) for 2 months followed by RH for 4 months. ART will be prescribed by the clinical staff at the HIV-TB clinic according to South African Department of Health guidelines. Standard first line ART in TB patients is tenofovir, emtricitabine (or lamivudine) and efavirenz. Co-trimoxazole prophylaxis will be prescribed to all patients unless a contra-indication exists.


Oral prednisone 40mg daily for 14 doses started on the first day that ART is taken, followed by 20mg daily for 14 doses (or identical placebo). A total of 28 days of study medication will thus be prescribed.


Patients will be screened once established on TB treatment, but before starting ART. If the patient is eligible, written informed consent will be taken. There will be six planned study visits that will be in relation to the start of ART: week 0 (the day ART is initiated), week 1, week 2, week 4, week 8 and week 12. Patients will be seen at unscheduled visits if clinical deterioration occurs. If paradoxical TB-IRIS is diagnosed this will be treated with open label prednisone at clinician discretion if symptoms are moderate or severe. If patients experience clinical complications (eg. TB-IRIS) follow-up will be prolonged beyond week 12 in order to stabilize their condition before referral back to the general TB-HIV clinical service for ongoing management.

Data monitoring:

The trial will be monitored by an independent Data and Safety Monitoring Board (DSMB) comprising 3 independent researchers and an independent statistician. After an initial meeting for agreeing on their Charter, the DSMB will meet twice (after 80 and 160 participants have completed follow-up) to review data quality and data with respect to safety and trial endpoints. If there is evidence of harm related to study medication or trial conduct the DSMB may advise the sponsor that trial enrollment should be stopped.

Clinical trial site:

Khayelitsha Site B HIV-TB clinic (Ubuntu clinic)

Statistical Analysis Plan

To download the PredART Statistical Analysis plan please click here.