Prednisolone and Mycobacterium indicus pranii in Tuberculous Pericarditis


For original article and supplementary Appendices:

By Prof Helmuth Reuter:


Prednisolone and Mycobacterium indicus pranii

in Tuberculous Pericarditis


B.M. Mayosi, M. Ntsekhe, JBosch, S. Pandie, H,Jung, F. Gumedze, J.Pogue,
L. Thabane, M. Smieja, V.Francis, L. Joldersma, K.M. Thomas, B. Thomas,
A.A. Awotedu, N.P. Magula, D.P. Naidoo, A. Damasceno: A.C. Banda,
B. Brown, P. Mango. B. Kirenga, C. Mordo, P.Mntla, J.M Tsitsi, F Peters,
M.R. Essop, J.B.W. Russell, J. Hakim. J. Matenga,A.F Barasa, M.U. Sani,
T.OLunuga, 0. Ogah,V.Ansa. A Ale, S. Danbauchi. D.Ojri,and S. Yusuf,

for the IMPI trial Investigators*





Tuberculous pericarditis is associated with high morbidity and mortality even if antituberculosis therapy is administered. We evaluated the effects of adjunctive glucocorticoid therapy and Mycobacterium indicus pranii immunotherapy in patients with tuberculous pericarditis.


Using a 2-by-2 factorial design, we randomly assigned 1400 adults with definite or probable tuberculous pericarditis to either prednisolone or placebo for 6 weeks and to either M.indicus pranii or placebo, administered in five injections over the course of 3 months.Two thirds of the participants had concomitant human immunodefi­ciency virus (HIV) infection. The primary efficacy outcome was a composite of death, cardiac tamponade requiring pericardiacentesis, or constrictive pericarditis.


There was no significant difference in the primary outcome between patients who received prednisolone and those who received placebo (23.8% and 24,5%, respec­tively; hazard ratio, 0.95; 95% confidence interval [CI], 0.77 to 1,18; P=0.66) or between those who received M. indicus pranii immunotherapy and those who received placebo (25.0% and 24.3%, respectively; hazard ratio, 1.03; 95% CI, 0.82 to 1.29; P= 0.81). Prednisolone therapy, as compared with placebo, was associated with significant reductions in the incidence of constrictive pericarditis (4.4% vs. 7.8%; hazard ratio, 0,56; 95% CI, 0.36 to 0.87; P=0.009) and hospitalization (20,7% vs. 25.2%; hazard ratio, 0.79 95% CI, 0.63 to 0,99; P=0.04), Both prednisolone and M. indicus pranii, each as compared with placebo, were associated with a significant increase in the incidence of cancer (1.8% vs. 0.6%; hazard ratio, 3,27; 95% CI, 1.07 to 10,03; P=0.03, and 1.8% vs, (1.5°101 hazard ratio, 3.69; 95% CI, 1.03 to 13.24; P=0.03, respectively), owing mainly to an increase in HIV-associated cancer.


In patients with tuberculous pericarditis, neither prednisolone nor M. indicus pranii had a significant effect on the composite of death, cardiac tamponade requiring pericardiocentesis, or constrictive pericarditis. (Funded by the Canadian Institutes of Health Research and others; IMPI Clinical number, NCT00810849.)

TURERCULOLIS PERICARDITIS is a common cause of pericardial effusion, cardiac tamporiade, and constrictive pericarditis in sub-Saharan Africa and parts of Asia.” Pa­tients with tuberculous pericarditis often have concomitant human immunodeficiency virus (HIV) infection.’ Despite antituberculosis thera­py, pericardial drainage, or pericardiectomy, mortality and morbidity remain high,’ Mortality is as high as 26% at 6 months but is even higher (approximately 40%) among persons with the ac­quired immunodeficiency syndrome.5

The use of glucocorticoid therapy in patients with tuberculous pericarditis to attenuate the inflammatory response may improve outcomes and decrease the risk of death by reducing car­diae tamponade and pericardial constriction,6 but the clinical effectiveness of adjunctive gluco­corticoids is unclear.’7-9 A meta-analysis of ran­domized, controlled trials of glucocorticoid therapy for tuberculous pericarditis showed a nonsignificant reduction in mortality,7.8but the numbers of events and patients included were very small. A meta-analysis of all trials of ad­junctive glucocorticoid therapy for all forms of tuberculosis also suggested reduced mortality.”‘

However, glucocorticoids may increase the risk of cancer in HIV-infected patients,11.12 and there is scant evidence of the effects of adjunc­tive glucocorticuid therapy for tuberculosis in this population.10 These considerations have led to conflicting recommendations in international guidelines regarding the role of adjunctive glu­cocorticoid therapy in patients with tuberculous pericarditis.9,13,14 We hypothesized that there would be an overall benefit of adjunctive pred­nisolone for these patients.

Preliminary evidence suggests that repeated doses of intradermal heat-killed Myrobartrrium indicus pranni (formerly known as Mycobacterium w) immunotherapy may reduce inflammation as­sociated with tuberculosis and increase the CD4-1- T-cell count in persons infected with HIV.15-17 M. inidicus pranni is a nonpathogenic, saprophytic, rapidly growing atypical mycobac­terium specks that has shown clinical benefit when administered as a heat-killed intradermal formulation in patients with leprosy and that may have benefits in patients with pulmonary tuberculosis and HIV infection.16-18-22 We hy­pothesized that intradermal M. indicus pranii could be effective in suppressing inflammation and its sequelae in patients with tuberculous peri­carditis. In the Investigation of the Management of Pericarditis (IMPI) trial, we evaluated the effi­cacy and safety of adjunctive prednisolone and M. indicus pranii in patients in Africa who had tu­berculous pericarditis.



We used a 2-by-2 factorial design to independently evaluate prednisolone and intradermal M. indicus pranii, as compared with placebo, for the treat­ment of tuberculous pericarditis (Fig, S1 in the Supplementary Appendix, available with the full text of this article at A detailed de­scription of the design of the trial has been pub­lished previously,4.23-2‘ The study was approved by the appropriate national regulatory authori­ties and by the ethics committee at each partici­pating site All the participants provided written informed consent.

Cadila Pharmaceuticals donated the study drugs used in the trial and distributed them to the re­search sites but had no role in the design or con­duct of the study, in the analysis of the data, or in the decision to submit the manuscript for publica­tion. The Canadian Institutes of Health Research reviewed the protocol, and changes were made according to the comments of the reviewers; the South African Medical Research Council also re­viewed the protocol, but no changes were recom­mended or made.

The study was coordinated by teams at the University of Cape Town, South Africa, and the Population Health Research Institute (PHRI) at Hamilton Health Sciences and McMaster Univer­sity, Canada. The steering committee (see the Supplementary Appendix) designed the study, oversaw its conduct, wrote the manuscript, and made the decision to submit it for publication. An independent trial monitoring committee monitored the trial for safety and efficacy. Data were collected and analyzed at the PHRI. The first and last authors vouch for the accuracy of the data and the analyses and for the fidelity of this report to the trial protocol, which is avail­able at


Details of the inclusion and exclusion criteria are
provided in Table Si in the Supplementary Appendix, Patients were eligible for inclusion in the trial if they were 18 years of age or older, had a pericardial effusion confirmed by echocardiog­raphy, had evidence of definite or probable tuber­culous pericarditis (as defined in Tables S2 and S3 in the Supplementary Appendix), and had be­gun to receive antituberculosis treatment less than 1 week before enrollment. Patients were ex­cluded if an alternative cause of pericardial dis­ease could be identified, if they had used gluco­corticoids within the previous month, if they had known hypersensitivity or allergy to the M. indicus pranii preparation, or if they were pregnant.


Eligible patients who had provided written in­formed consent were assigned to an active-treat­ment or placebo group fbr each of the two random­ized comparisons. Randomization was performed with the use of a central computer-generated randomization list, with stratification according to center and with a random block size. For the comparison of prednisolone with placebo, par­ticipants were assigned to receive prednisolone or placebo for 6 weeks at a dose of 120 mg per day in the first week, 90 mg per day in the second week, 60 tug per day in the third week, 30 mg per day in the fourth week, 15 mg per day in the fifth week, and 5 mg per day in the sixth week. For the comparison of M. indicus pranii with placebo, par­ticipants were assigned to receive the M. indicus pranii preparation (CADI-Mw injection, Cadila Pharmaceuticals) or placebo in five doses at the time of enrollment and at 2 weeks, 4 weeks, 6 weeks, and 3 months. The first dose was given as two injections of 0.1 ml (containing 0..5×109 organisms) in each deltoid region of the upper arm; the four subsequent doses were given as a single injection of 0,1 in].

Trial participants received antimicrobial treat­ment for tuberculosis and antiretroviral treatment for HIV according to World Health Organization (WHO) guidelines; management during the course of the trial was revised as recommended treat­ment practices evolved.13.26.28 No routine testing for drug resistance of either M. tuberculosis iso­lates or HIV isolates was performed before or during treatment.

Follow-up data were collected at the time of hospital discharge; at 2 weeks, 4 weeks, 6 weeks, 3 months, and 6 months; then every ( months through 2 years; and every 12 months thereafter.

Follow-up visits included assessments of study outcomes, adherence to treatment, and adverse events, Site monitoring throughout the study was performed through the project coordinating office according to a standard operating proce­dure (see the Site Monitoring and Quality Control Section in the Supplementary Appendix),


The primary efficacy outcome was a composite of death or the first occurrence of cardiac tam­ponade requiring pericarcliocentesis or constric­tive pericarditis. Secondary efficacy outcomes in­cluded the individual components of the primary outcome as well as hospitalization. Safety outcomes included opportunistic infections and cancer, as well as the effect of interventions on the CD4+ T-lymphocyte cell count (as a measure of immu­no suppression) and the incidence of the immune reconstitution inflammatory syndrome (in HIV• infected patients). Detailed definitions of outcome events arc provided in Tables S4 through 58 and the Methods section in the Supplementary Ap­pendix.


We based our sample-size calculation on the fol­lowing assumptions; the. event rate among pa­tients receiving placebos for both interventions would be 35% at a mean follow-up of 2 years; half the patients in the control group for each intervention would receive another effective in­tervention, which would result in a 30% relative risk reduction in the event rate; the non adher­ence rate would be 10%; and the rate of loss to follow-up would be 6%. On the basis of these assumptions, we estimated that with a sample of 140-0 patients, the study would have 90% power to detect a 22.9% reduction in the hazard ratio, with the use of a log-rank test and a two-sided type I error rate of 5%.

Data were analyzed with the use of SAS soft­ware, version 9.1, according to an intention-to ­treat approach (as described in the protocol and the pre-specified statistical analysis plan). Time ­to-event curves were constructed by means of product-limit estimation and were compared with the use of stratified log-rank tests. Cox proportional-hazards models stratified accord­ing to factorial treatment assignment were used to determine hazard ratios and 05% confidence intervals, We assessed interactions between the two active treatments by including an interac­tion term in the model.

We also performed analyses for the primary outcome in subgroups defined according to HIV status, the strength of the evidence supporting the tuberculous pericarditis diagnosis (definite or probable diagnosis), exposure of HIV-infected persons antiretroviral therapy (>6 months <- months, or no exposure), the CD4+ TCell count threshold for treatment (4200 per cubic millime­ter vs. >2011 per cubic millimeter and <-350 per cubic millimeter vs. >350 per cubic millimeter), and pericatdiocentesis status at baseline (per­formed vs. not performed), using the Cox propor­tional-hazards model, with an interaction term for treatment effects across the subgroups. For all analyses, P values of less than 0.05 were consid­ered to indicate statistical significance.

The trial monitoring committee performed seven interim analyses of the primary outcome data; at the sixth interim analysis, the trial monitoring committee recommended that the M. indicus pranni randomization be discontinued for reasons of futility.



The trial was conducted from January 2009 through February 2014 at 19 hospitals in eight African countries (see the Supplementary Appen­dix). A total of 1400 patients were enrolled for..


Baseline Characteristics of the Patients and Diagnosis at 3 Months*
Baseline Characteristics of the Patients and Diagnosis at 3 Months*

*Plus-minus values are means .±SD. There were no significant differences among the study groups in any of the base­line characteristics listed here HIV denotes human immunodeficiency virus.


the comparison of prednisolone with placebo; 706 were assigned to receive prednisolone and 694 to receive placebo. The median follow-up pe­riod was 616.5 days (interquartile range, 817.5 to 1085.5); at study end, the primary-outcome sta­tus was known for 1371 participants (97.9%) (Fig. 52 in the Supplementary Appendix).

A total of 1250 patients were enrolled for the comparison of M. indicus priariii with placebo, be­fore this comparison was stopped early for futil­ity (on February 14, 2011); 625 were assigned to receive M. indicus pranti and 625 to receive place­bo. The median follow-up period was 7205 days (interquartile range, 368.0 to 1095,0): at study end, the primary-outcome status was known for 1223 participants (97,8%) (Fig. 53 in the Supple­mentary Appendix).

The baseline characteristics were similar across the groups (Table 1, and Table S9 in the Supplementary Appendix). Approximately two thirds of the participants had a large pericardial effusion; pericardiocentesis was performed in 60.5%. The diagnosis of tuberculous pericarditis was considered to be definite in 17.1% of the patients (details are provided in Table 510 in the Supplementary Appendix). Two thirds of the participants were H1V-positive.


For the comparison of prednisolone with place­bo, 88,5% of the patients in the prednisoione group and 88.7% of those in the placebo group adhered to the regimen [or the full 6 weeks of the study treatment. A total of 414 patients (3.1%) re­ceived nonstudy glucocorticoids during the trial; this rate was similar in the prednisolone and pla­cebo groups. For the comparison of M. indiciis pranii with placebo, 75.9% of the patients in the M. inidicus pranni( group and 81,4% of those in the placebo group adhered to the regimen for the full 3 months of the study treatment.

Of the 1400 patients enrolled in the trial, 76,6% were receiving antituberculosis treatment at the time of randomization, and 14.5% were receiving antiretroviral treatment (Table S11 in the Supplementary Appendix). The rates of ongo­ing use of antituberculosis therapy and antiretro­viral therapy during the trial are shown in Tables 512 and 513 in the Supplementary Appendix. The increasing use of antiretroviral therapy during the course of the trial reflects the adoption of revised WHO guidelines recommending early initiation

Effects of Prednisone and Mycobacterium indicus pranii Immunotherapy on Efficacy Outcomes*
Effects of Prednisone and Mycobacterium indicus pranii Immunotherapy on Efficacy Outcomes*

centesis, or constrictive pericarditis) was 143 events per 100 patient-years of follow-up in the prednisolone group and 14,8 per 100 patient-years in the placebo group (hazard ratio with prednisolone, 035; 9Sk confidence interval [CI], 077 in 1.181 P=0,66) {Table 2 and 1A, and Fig. 54 in the Supplementary Appendix). There was also no significa.nt difference between the two groups in the rate of death or the rate of cardiac tamponade requiring pericardiocentesis, when considered individually (Fig. S5 and S6 in the Supplementary Appendix). The main causes of death were pericarditis (23_8%1, disseminated tuberculosis (18.6%), HIV infection (7.3%), and other cardiovascular causes (5.7%) (Table S14 in the Supplementary Appendix). The prednisolone group had a lower rate of constrictive pericarditis and fewer hospitalizations than the placebo group (Table 2, and Table S15 and Fig. S7 and S8 in the Supplementary Appendix),

The incidence of opportunistic infection was 6.89 cases per 100 patient-years in the predniso­lone group, as compared with 531 per 100 pa­tient-years in the placebo group (hazard ratio, 1.16; 95% CI, 0.84 to 1.0it; P=016) (Fig. R:9 in the Supplementary Appendix). The proportion of patients with candidiasis was higher in the pred­nisolone group than in the placebo group (7.7% vs. 52%, I)=0.05) (•a.ble 3, and Table S16 in the Supplementary Appendix

Prednisolone, as compared with placebo, was associated with an increased incidence of cancer (1.05 vs. 0.32 cases per 100 person-years: hazard ratio, 3..27; 95% CI, 1.07 to 11103; P=0.03) {Table 3, and Fig. S10 in the Supplementary Appendix). This increase was due to a higher incidence of H1V-related cancers in the prednisolone group than in the placebo group (0.73 vs. 0.08 per 100 person-years; hazard ratio, 9.04 F 95% CI, 1,14 to 71.33; P=0.04). Table 3). A list of the causes of cancer is provided in Table. S17 in the Supple­mentory Appendix.

There were two cases of the immune recon­stitution inflammatory syndrome in the pred­nisolone group and one in the placebo group. There was a similar increase in CD4+ T-cell counts in the two  groups  (Table) S18 in the Supplementary Appendix).


The rates of the primary composite outcome and its components, as well as the rates of hospitalization and opportunistic infection, did not dif­fer significantly between the M.indicus  pranii group and the placebo group [Tables 2 and 3 and Fig1B, and Fig S11 through S16 in the Supple­mentary Appendix). However,  indicus pranii was associated with an increased incidence of cancer, as compared with placebo (092 vs. 0.24 cases per 100 person-years; hazard ratio, 3.69; 95% CI, 1.03 to 13.24; P=0.03) (Table 3. and Fig. S17 in the Supplementary Appendix), which was due mainly to an increase in HIV- associated can­cer. There was one ease of the immune reconstitution inflammatory syndrome in each group. There was a similar increase in CD4+ T-cell counts in the two Groups.

Prednisone Comparison
Prednisone Comparison

Significantly more patients in the M. indicus pranni group than in the placebo group had injection-site reactions (41,4% vs. 2,9%, P<0.001) (Table S19 and Fig. S18 in the Supplementary Appendix), Although the majority of these ractions were characterized by minor symptoms and signs of inflamation (i.e„ induration, redness, and pain), there was a significantly greater pro­portion of patients with abscess formatio in the M. indicus pranni group than in the placebo group (15% vs. 1.0%, P<0.001).

Table 3 Effects of Prednisone and Mycobacterium indicus
Table 3. Effects of Prednisone and Mycobacterium Indicus pranni immunotherapy on Safety Outcomes


There was no significant interaction between the effects of M. indicus pranni  and those of predniso­lone on the primary efficacy and safety outcomes (P>0.30 for all interactions), except for injection site reactions (P=0.004) (Fig. S4 and S11 in the Supplementary Appendix). However, 9 of the 13 cases of cancer in the prednisolone group oc­curred in patients who also received M. indicus pranni (Fig. 2). Although a clinical interaction of the two interventions on cancer cannot be ruled out, the number of cases is small.

The effects of prednisolone and M. indicus pranii  immunotherapy on the primary composite efficacy outcome were similar across prespeci­fied subgroups (Fig. S19) and S20 in the Supple­mentary Appendix)


In this trial, adjunctive prednisolone therapy and M. indicus pranii immunotherapy were compared with placebo in patients with definite or proba­ble tuberculous pericarditis. Neither therapy had a significant effect on the primary composite outcome of death, cardiac tamponade requiring pericardiocentesis, or constrictive pericarditis. With respect to the secondary outcomes, adjunc­tive prednisolone therapy reduced the incidence of constrictive pericarditis and the incidence of hospitalizations. However, both interventions in­creased the incidence of cancer among trial par­ticipants.

Previous trials of adjunctive glucocorticoid therapy in patients with tuberculous pericarditis had relatively small samples (28 to 240 patients) and included few HIV-infected patients, and there was poor reporting of adverse events).10,29-34 Our trial included 1400 patients, a substantial number of whom (940 patients) had HIV infec­tion. To our knowledge, M. indicus pranii immu­notherapy has not been studied previously in this population.

Figure2. Time to Cancer Occurrence.
Figure 2. Time to Cancer Occurence

For the prednisolone comparison, we used an initiation dose of 120 mg per day, which is known to have a therapeutic effect when admin­istered in combination with rifampin, ail en­zyme inducer that increases the metabolism of glucocorcicoids:” Adherence to prednisolone therapy was high.

The significant reduction in pericardial constriction with prednisolone indi­cates that the doses used were sufficient to achieve a substantial anti-inflammatory effect. The reduction in the incidence of constrictive pericarditis translated to fewer hospitalizations in the prednisolone-treated group. This finding is important because pericardiectomy, the de­finitive treatment for chronic pericardial con­striction, is associated with high perioperative mortality and morbidity, and cardiac surgery is not widely available in Africa.2.36

The marked increase in HIV-related cancer with prednisolone therapy is consistent with the results of two previous studies of HIV-associated tuberculosis, in which cases of Kaposi’s sarcoma occurred only in the prednisolonc-treated groups.11.12 However, the assocation of M.indicus prunii immunotherapy with cancer that we observed in our study has not been reported previously. It is possible that adjunctive gluco­corticoids and M.indicus pranii act synergistically to increase the risk of cancer in immuno-supressed patients.The available data on the inter-action between adjunctive glucocorticoid therapy and M. indicus pranni immunotherapy are sparse.23.

Our study has a few limitations. First, a defi­nite diagnosis of tuberculosis either in the peri­cardium or elsewhere in the body was made in only one quarter of the patients. Thus, one po­tential interpretation of the trial result is that the interventions were not effective because rela­tively few of the trial participants actually had tuberculous pericarditis. However, the results were consistent between patients with definite tuberculosis and those with probable tuberculo­sis. Furthermore, the diagnosis of extra pulmo­nary tuberculosis is challenging, and only a mi­nority of cases of extrapulminory tuberculosis are treated on the basis of a definite diagnosis.37 Second, a small proportion of patients (less than 2%) had a diagnosis other than tuberculosis. However, although the estimation of the sample size needed for this study was based on the clinical case definition of tuberculous pericardi­tis,4 we expected that a small proportion of cases (up to 10%) would have an alternative cause of pericarditis23 Third, the trial was pow­ered for a rate of nonadherence of 10% in the active-treatment groups. Although this rate was almost achieved in the prednisolone group (non-­adherence rate of 11%), the non-adherence rate was higher in the M. indicus pranii group (21%),owing mainly to injection-site side effects. This relatively high nonadherence rate may have di­minished the power of the study with respect to the analysis of the primary outcome in the M. indicus prunii group. Finally, because prednisolone is immunosuppressive and M. indicus pranii is im­mnunostimulatory, there may be an interaction between them that could result in each one ei­ther reducing or increasing the effects of the other one.23

In conclusion, adjunctive therapy with pred­nisolone for 6 weeks and with M. indicus pranii for 3 months did not have a significant effect on the combined outcome of death from all causes, cardiac tamponade requiring pericardiocentesis, or constrictive pericarditis. Both therapies were also associated with an increased risk of HIV-associated cancer. However, the use of adjunc­tive glucocorticoids reduced the incidences of pericardial constriction and hospitalization. The beneficial effects of prednisolone with respect to pericardial constriction and hospitalization were similar in HIV-positive and HIV-negative pa­tients.

Discussion forms provided by the authors are available with the full text of thus article at NEJM .org.

Supported by grand from the Canadian Institutes of Health Research, the Canadian Network and Centre for Trials Interna­tionally, the Population Health Research institute, the South Afriran Medical Research Council, the Lily and Ernst Hausmann Research Trust, and Cadila Pharma, India.

New England Journal of Medicine  371;12  NEJM.ORG  SEPTEMBER 18, 2014 

The New England Journal of Medicine







Leave a Reply