Randomized Placebo-controlled Trial of Prednisone for the TB-Immune Reconstitution Inflammatory Syndrome

Transcription

1 Randomized Placebo-controlled Trial of Prednisone for the TB-Immune Reconstitution Inflammatory Syndrome Graeme Meintjes 1,2, Robert J Wilkinson 1,2,3,4, Chelsea Morroni 1, Dominique Pepper 1,2, Kevin Rebe 1,2, Molebogeng Rangaka 1, Tollulah Oni 1,3, Gary Maartens 1 1. University of Cape Town, 2. GF Jooste Hospital, 3. Imperial College London, 4. National Institute for Medical Research (UK)

2 Paradoxical TB-IRIS Incidence: 8-43% 1 Anecdotal reports of symptomatic improvement with corticosteroids 2 but potential risks 3 Kaposi s sarcoma Herpes virus reactivations Other infections 1. Meintjes et al, Lancet Infect Dis Breen et al, Thorax Elliott et al, J Infect Dis 2004

3 Study Design Hypothesis: 4 week course of prednisone would reduce need for medical interventions without excess adverse events Study drug Prednisone or placebo, randomized, double-blind 1.5mg/kg for 2 weeks then 0.75mg/kg for 2 weeks Follow up assessments at 1, 2, 4, 8 and 12 weeks Open-label prednisone at physician discretion if clinical deterioration/relapse Predefined primary endpoint* Cumulative number of days hospitalized and outpatient therapeutic procedures counted as 1 additional day *Trial registered:

4 Case Definition Prior to ART Microbiologic, histologic or very strong clinical evidence of TB Initial improvement during TB treatment Infecting strain of M. tuberculosis susceptible to rifampicin (if result available) Receiving TB treatment when ART initiated Within 3 months of initiating ART New or recurrent TB-related symptoms Presence of 1 following TB manifestations Lymph node enlargement Cold abscess Serous effusion Radiographic pulmonary infiltrate No other opportunistic disease to explain clinical deterioration

5 Exclusion criteria Age < 18 years Pregnancy Prior ART exposure Kaposi s sarcoma Uncontrolled diabetes mellitus Adrenal failure Prior adjunctive corticosteroid therapy for this TB episode Life threatening TB-IRIS Neurological involvement Respiratory failure Airway compression

6 287 screened (June December 2007) Alternative diagnosis = 44 Did not fulfill case definition = 65 Exclusion criterion = 55 Unwilling/unable to consent = enrolled 55 placebo arm 2 died 6 defaulted (2 LTF) 3 discontinued study drug 6 rifampicin resistance 55 prednisone arm 3 died 0 defaulted 1 discontinued study drug 4 rifampicin resistance

7 Baseline characteristics Placebo arm Prednisone arm P-value Age (median, years) Gender 32F; 23M 38F; 17M 0.23 Previous TB WHO stage Duration TB treatment to ART (median, days) Duration ART to IRIS (median, days) CD4 prior to ART (median, cells/ L) CD4 at enrollment (median, cells/ L) Hospitalized at enrollment

8

9 Primary endpoint Cumulative number of days hospitalized and outpatient therapeutic procedures (counted as 1 additional day), ITT analysis Placebo arm N = 55 Prednisone arm N = 55 P-value Total days hospitalized Total number outpatient procedures Cumulative primary endpoint (median, IQR) 3 (0-9) 1 (0-3) 0.046

10

11 Ultrasound score demonstrated no differences at week 2 or 4

12

13

14 Adverse events Placebo arm Prednisone arm P-value Death on study 2 (4%) 3 (5%) 0.65 Corticosteroid side effects* 18 (33%) 12 (22%) 0.20 Corticosteroid side effects while on study drug 3 (5%) 8 (15%) 0.11 Infections 30 (55%) 36 (65%) 0.24 Severe infections** 4 (7%) 2 (4%) 0.40 * Included BP > 140/90, oedema, hyperglycaemia, hypomania, acne, Cushingoid features, gastritis symptoms ** WHO stage 4 or invasive bacterial infection

15 Conclusions Prednisone reduced need for medical interventions (days hospitalized and outpatient procedures combined) Consistent benefit, maximal in first 4 weeks, across a range of secondary outcome measures Symptom score Karnofsky performance score Radiology score C-reactive protein Benefit shown despite crossovers to open label prednisone, that may have reduced observed effect size No excess of steroid side effects or infections For some patients, 4 weeks appeared to be too short Exclusion of rifampicin resistance and alternative diagnoses is critical