The n e w e ng l a n d j o u r na l of m e dic i n e

original article

Multidrug-Resistant Tuberculosis
and Culture Conversion with Bedaquiline
Andreas H. Diacon, M.D., Ph.D., Alexander Pym, M.D., Ph.D.,
Martin P. Grobusch, M.D., Ph.D., Jorge M. de los Rios, M.D.,
Eduardo Gotuzzo, M.D., Irina Vasilyeva, M.D., Ph.D., Vaira Leimane, M.D.,
Koen Andries, D.V.M., Ph.D., Nyasha Bakare, M.D., M.P.H., Tine De Marez, Ph.D.,
Myriam Haxaire-Theeuwes, D.D.S., Nacer Lounis, Ph.D., Paul Meyvisch, M.Sc.,
Els De Paepe, M.Sc., Rolf P.G. van Heeswijk, Pharm.D., Ph.D.,
and Brian Dannemann, M.D., for the TMC207-C208 Study Group*

A bs t r ac t

Background
Bedaquiline (Sirturo, TMC207), a diarylquinoline that inhibits mycobacterial ATP The authors’ affiliations are provided in
synthase, has been associated with accelerated sputum-culture conversion in pa- the Appendix. Address reprint requests
to Dr. Dannemann at Janssen Research
tients with multidrug-resistant tuberculosis, when added to a preferred background and Development, 1125 Trenton Har-
regimen for 8 weeks. bourton Rd., Titusville, NJ 08560, or at
bdannema@its.jnj.com.
Methods * A complete list of investigators in the
In this phase 2b trial, we randomly assigned 160 patients with newly diagnosed, TMC207-C208 Study Group is provided
smear-positive, multidrug-resistant tuberculosis to receive either 400 mg of beda- in the Supplementary Appendix, avail-
able at NEJM.org.
quiline once daily for 2 weeks, followed by 200 mg three times a week for 22 weeks,
or placebo, both in combination with a preferred background regimen. The pri- N Engl J Med 2014;371:723-32.
DOI: 10.1056/NEJMoa1313865
mary efficacy end point was the time to sputum-culture conversion in liquid broth. Copyright © 2014 Massachusetts Medical Society.
Patients were followed for 120 weeks from baseline.

Results
Bedaquiline reduced the median time to culture conversion, as compared with pla-
cebo, from 125 days to 83 days (hazard ratio in the bedaquiline group, 2.44; 95%
confidence interval, 1.57 to 3.80; P<0.001 by Cox regression analysis) and increased
the rate of culture conversion at 24 weeks (79% vs. 58%, P = 0.008) and at 120 weeks
(62% vs. 44%, P = 0.04). On the basis of World Health Organization outcome defini-
tions for multidrug-resistant tuberculosis, cure rates at 120 weeks were 58% in the
bedaquiline group and 32% in the placebo group (P = 0.003). The overall incidence
of adverse events was similar in the two groups. There were 10 deaths in the beda-
quiline group and 2 in the placebo group, with no causal pattern evident.

Conclusions
The addition of bedaquiline to a preferred background regimen for 24 weeks re-
sulted in faster culture conversion and significantly more culture conversions at 120
weeks, as compared with placebo. There were more deaths in the bedaquiline group
than in the placebo group. (Funded by Janssen Pharmaceuticals; TMC207-C208
ClinicalTrials.gov number, NCT00449644.)

n engl j med 371;8 nejm.org august 21, 2014 723

The New England Journal of Medicine
Downloaded from nejm.org on July 6, 2015. For personal use only. No other uses without permission.
Copyright © 2014 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

T
he World Health Organization positive, pulmonary, multidrug-resistant tubercu-
(WHO) estimates that the global incidence losis on the basis of proportion-method results,11,12
of tuberculosis in 2012 was 8.6 million positive rapid-screening tests (FASTPlaque-­
cases, with 1.3 million deaths, predominantly Response assay [Biotec] and GenoType MTBDR
occurring in developing countries.1 Although there plus line-probe tests [Hain Lifescience]). Patients
has been some progress in reducing tuberculosis who had received previous treatment for multi-
cases and deaths in the past 20 years, multidrug- drug-resistant tuberculosis were excluded. Addi-
resistant tuberculosis (i.e., with resistance to at tional exclusion criteria were a positive test for the
least isoniazid and rifampin) remains a major human immunodeficiency virus (HIV) with a CD4+
challenge. The 2012 global incidence of multi- count of less than 300 cells per cubic millimeter,
drug-resistant tuberculosis was 450,000 cases.1 complicated or severe extrapulmonary or neuro-
Therapy for multidrug-resistant tuberculosis is a logic manifestations of tuberculosis, severe cardi-
long, arduous regimen of antiquated drugs that ac arrhythmia requiring medication, a corrected
are mainly bacteriostatic and have an unfavorable QT interval with the use of Fridericia’s formula
side-effect profile.2 The WHO reports that major (QTcF)13 of more than 450 msec, a history of risk
efforts are needed to improve the current average factors for torsades de pointes, concomitant seri-
rate of 48% for successful treatment of patients ous illness, alcohol or drug abuse, pregnancy or
with multidrug-resistant tuberculosis.1 By 2012, breast-feeding, and previous treatment with beda-
extensively drug-resistant tuberculosis (i.e., with quiline. According to the protocol, moxifloxacin,
additional resistance to injectable second-line gatifloxacin, and systemic use of cytochrome
drugs and fluoroquinolones) was reported in 92 P-450 3A4 inhibitors or inducers were prohibited
countries worldwide, with the presence of exten- during and for 1 month after completion of the
sively drug-resistant isolates reported in 9.6% of study treatment.
patients with multidrug-resistant tuberculosis.1
Bedaquiline (Sirturo, TMC207), a diarylquino- Study Design
line that inhibits mycobacterial ATP synthase,3 is
In this randomized, double-blind, placebo-con-
the first antituberculosis drug with a new mecha-trolled study, patients were stratified according
nism of action to be approved for use in multi- to study site and radiographic assessment of lung
drug-resistant tuberculosis in 40 years.4 Beda- cavitation (≥2 cm bilaterally, ≥2 cm unilaterally,
quiline has shown bactericidal activity in vitro,or <2 cm). Trial sites were located in Brazil, India,
in murine models of tuberculosis,3,5-7 and in a Latvia, Peru, the Philippines, Russia, South Africa,
7-day proof-of-concept study involving patients and Thailand. Patients received either bedaqui-
with drug-sensitive tuberculosis.8 In stage 1 of an
line (400 mg once daily for 2 weeks, followed by
exploratory phase 2b randomized trial, called 200 mg three times a week for 22 weeks, admin-
TMC207-C208, involving patients with newly diag- istered as 100-mg tablets) or placebo, plus a pre-
nosed, smear-positive, multidrug-resistant tuber-ferred five-drug, second-line antituberculosis
culosis, 8 weeks of bedaquiline treatment had background regimen (Fig. 1). Patients were in-
better antibacterial activity than placebo when structed to take bedaquiline or placebo with wa-
added to a preferred five-drug, second-line back-ter after breakfast. National treatment-program
ground regimen.9,10 In stage 2 of the TMC207- regimens were respected, although the preferred
C208 study, a phase 2b study in which bedaqui- five-drug background regimen was ethionamide,
line was administered for 24 weeks to a larger pyrazinamide, ofloxacin, kanamycin, and cyclo-
number of patients, we evaluated the time to serine. Changes in the background regimen were
sputum-culture conversion, the rates of culture permitted on the basis of the results of drug-
conversion and drug resistance, pharmacokinetics,susceptibility testing, side effects, or unavailabil-
and safety over a 120-week period in patients re-ity of drugs on site.
ceiving a preferred five-drug background regimen. After the 24-week treatment period, there was
a 96-week period during which patients were
Me thods instructed to complete their background regi-
men (Fig. 1). Patients who prematurely discon-
Patients tinued the trial were followed for collection of
We recruited patients between the ages of 18 and survival data until trial completion unless they
65 years with newly diagnosed, sputum smear– withdrew consent.

724 n engl j med 371;8 nejm.org august 21, 2014

The New England Journal of Medicine

Downloaded from nejm.org on July 6, 2015. For personal use only. No other uses without permission.
Copyright © 2014 Massachusetts Medical Society. All rights reserved.
 Tuberculosis Culture Conversion with Bedaquiline

Overall Treatment Phase Last Study Visit

120 Wk

Investigational Treatment Phase Postinvestigational Treatment Phase

24 Wk 96 Wk

Bedaquiline plus background regimen Background regimen only

24-Wk 120-Wk analysis

evaluation

24 Wk 96 Wk

Placebo plus background regimen Background regimen only

Figure 1. Study Design and Drug Regimens.

Patients with multidrug-resistant tuberculosis were assigned in a 1:1 ratio to receive either bedaquiline (400 mg
once daily for 2 weeks, followed by 200 mg three times a week for 22 weeks) or placebo, plus a preferred five-drug,
second-line antituberculosis background regimen. The total treatment period was 18 to 24 months, during which
bedaquiline was administered for 6 months. The total trial duration was 120 weeks (30 months), which included an
anticipated 6-month period after the completion of treatment.

Each site obtained approval of the study pro-
tocol from at least one (or more, if required by
local regulations) independent ethics committee
or institutional review board. The trial was con-
ducted in accordance with the principles of
Good Clinical Practice and the Declaration of
Helsinki. All patients provided written informed
consent before trial entry. Details regarding the
study design are provided in the protocol, avail-
able with the full text of this article at NEJM.org.
Microbiologic Assessments
Triplicate spot sputum samples were collected at
every visit (except on the day of the first adminis-
tration of the study drug) and at the time of with-
drawal, for patients who did not complete the
study, for culture of Mycobacterium tuberculosis in
liquid medium (Mycobacteria Growth Indicator
Tube, Becton Dickinson). Drug-susceptibility test-
ing was performed at a central laboratory (Insti-
tute of Tropical Medicine, Antwerp, Belgium) at
baseline and at 8, 24, and 72 weeks in all patients
and in those with reversion to a positive culture
after initial culture conversion. The minimal in-
hibitory concentration (MIC) of bedaquiline on
7H11 agar was defined as the lowest concentra-
tion (measured in micrograms per milliliter) that
prevented the growth of 99% of M. tuberculosis
isolates. Isolates for which the MIC was increased
by a factor of 4 or more, as compared with the
baseline value, were considered to have decreased
susceptibility to bedaquiline.
Safety Assessments
Safety assessments included monitoring for ad-
verse events, clinical laboratory testing, and elec-
trocardiography at predefined intervals through-
out the study. Toxicity was graded on the basis of
the Division of Microbiology and Infectious Dis-
eases (DMID) adult toxicity tables.14
Study End Points
The primary end point was the time to sputum-
culture conversion, which was defined as two
consecutive negative liquid cultures from sputum
samples that were collected at least 25 days apart
and were not followed by confirmed positive cul-
tures. The primary analysis was performed on
the basis of data at 24 weeks. Secondary efficacy
measurements were the rates of culture conver-
sion after 24 weeks and after 120 weeks. We also
performed 11 subgroup efficacy analyses of the
rate of culture conversion, including in patients
with M. tuberculosis isolates that were resistant
only to isoniazid and rifampin and in those with
isolates that also were resistant to any second-

n engl j med 371;8 nejm.org august 21, 2014 725

The New England Journal of Medicine
Downloaded from nejm.org on July 6, 2015. For personal use only. No other uses without permission.
Copyright © 2014 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l
line injectable drug or any fluoroquinolone
(which were categorized as having pre-extensive
drug resistance) and in patients with isolates that
were susceptible or resistant to pyrazinamide.
(For details regarding all 11 subgroups, see the
Methods section in the Supplementary Appendix,
available at NEJM.org.)
Study Oversight
The study was funded by Janssen Pharmaceuti-
cals. Medical-writing support (funded by Janssen)
was provided by an employee of Gardiner-
Caldwell Communications, who wrote the initial
draft of the manuscript and incorporated com-
ments from the authors. The data were collected
by the investigators and analyzed by the sponsor.
All the authors made the decision to submit the
manuscript for publication and vouch for the
accuracy and completeness of the data reported
and the fidelity of the study to the protocol.
Statistical Analysis
We determined that enrollment of 75 patients in
each study group would provide a power of 80%
to detect a difference of 22 percentage points in
the 6-month rate of culture conversion between
the placebo group (estimated at 50%) and the
bedaquiline group (estimated at 72%) at a two-
sided significance level of 0.05. The safety analy-
sis was conducted in the intention-to-treat pop-
ulation and included all patients who had
undergone randomization and received at least
one dose of the assigned study drug. Safety data
are presented for the 120-week treatment period.
The efficacy analyses were performed in the
modified intention-to-treat population, which
excluded patients who had no positive mycobac-
terial cultures from sputum samples obtained
before administration of the first dose of the
study drug or a positive culture up to week 8 in
cases in which baseline cultures were negative,
those for whom susceptibility to rifampin and
isoniazid was shown or resistance could not be
confirmed, those with extensively drug-resistant
tuberculosis, and those who had not undergone
assessment after baseline.
The primary end point was analyzed in the
modified intention-to-treat population. In this
analysis, data for patients who discontinued treat-
ment, died, or did not have sputum-culture con-
version before 24 weeks were censored at the last
assessment, regardless of the culture status at
the time of study dropout or death, and these
726 n engl j med 371;8 nejm.org august 21, 2014
The New England Journal of Medicine
Downloaded from nejm.org on July 6, 2015. For personal use only. No other uses without permission.
Copyright © 2014 Massachusetts Medical Society. All rights reserved.
of m e dic i n e
patients were considered to have had no response.
We used a Cox proportional-hazards model with
adjustment for stratification variables to com-
pare the time to culture conversion in the two
study groups.
For the 11 prespecified subgroup analyses, no
formal adjustments for multiple comparisons
were made. With this number of subgroups, there
is a probability of 43% that at least one test would
be significant (P<0.05) on the basis of chance
alone. A post hoc analysis was performed to as-
sess treatment outcomes on the basis of WHO
definitions for multidrug-resistant tuberculosis.15
R e sult s
Study Population
Of 282 patients who were screened, 160 under-
went randomization and received at least one
dose of the assigned study drug (the intention-to-
treat population, comprising 79 patients in the
bedaquiline group and 81 in the placebo group).
The modified intention-to-treat population con-
sisted of 132 patients (Fig. 2). In the intention-to-
treat population, 60 patients (38%) discontinued
the trial prematurely, with no relevant differences
between the two study groups in the reasons for
discontinuation (Fig. 2, and Fig. S1 in the Supple-
mentary Appendix). The most common reasons
for discontinuation were withdrawal of consent
and adverse events.
In the modified intention-to-treat population
(protocol-defined efficacy population), there were
more men (64%) than women, and more black
patients (37%) than any other racial or ethnic
group (Table 1, and Table S3 in the Supplemen-
tary Appendix). Baseline demographic and dis-
ease characteristics were similar in the two study
groups except that the proportion of patients
who had isolates with resistance to pyrazin-
amide and the proportion classified as having
isolates with pre-extensive drug resistance were
nonsignificantly larger in the bedaquiline group,
and the proportions of HIV-positive patients and
patients with abnormally low albumin levels
were significantly larger in the placebo group
(Table 1). The median overall treatment phase
was longer in the placebo group than in the
bedaquiline group (Table 2). A higher proportion
of patients in the placebo group than in the
bedaquiline group (58% vs. 47%) had at least
one new antituberculosis drug added to their
background regimen.
 Tuberculosis Culture Conversion with Bedaquiline

282 Patients were assessed for eligibility

122 Were ineligible

103 Did not fulfill all inclusion criteria
or met exclusion criteria
7 Withdrew consent
12 Had other reasons

Intention-to-Treat Group Modified Intention-to-Treat Group

160 Underwent randomization 132 Underwent randomization
and were treated and were treated

28 Were excluded from

modified intention-to-
treat group

79 Were assigned 81 Were assigned 66 Were assigned 66 Were assigned

to bedaquiline to placebo to bedaquiline to placebo

1 Crossed over to open-

label bedaquiline 1 Crossed over to open-
29 Discontinued study
label bedaquiline
9 Had adverse event
23 Discontinued study
6 Withdrew consent
31 Discontinued study 8 Had adverse event
5 Were lost to 24 Discontinued study
6 Had adverse event 5 Withdrew consent
follow-up 5 Had adverse event
7 Withdrew consent 5 Were lost to
3 Did not adhere to 7 Withdrew consent
3 Were lost to follow-up
study regimen 3 Were lost to
follow-up 2 Did not adhere to
3 Were pregnant follow-up
7 Did not adhere to study regimen
2 Were determined 7 Did not adhere to
study regimen 3 Were pregnant
to be ineligible study regimen
2 Were pregnant
1 Acquired extensive 2 Were pregnant
6 Were determined
drug resistance
to be ineligible

50 Completed study 49 Completed study 43 Completed study 41 Completed study

Figure 2. Enrollment and Outcomes.

The modified intention-to-treat population was a subgroup of the intention-to-treat population, with a total of 28 patients excluded:
9 patients (6 in the bedaquiline group and 3 in the placebo group) who had sputum-culture results that did not allow for primary efficacy
evaluation (either no evidence of culture positivity before the first dose of the study drug was administered or no results during the first
8 weeks after the first dose was administered), 7 patients (3 patients and 4 patients, respectively) who were infected with extensively
drug-resistant tuberculosis, 8 patients (4 in each group) who had drug-sensitive tuberculosis, and 4 patients (all in the placebo group)
for whom status with respect to multidrug-resistant tuberculosis could not be confirmed.

Antimycobacterial Activity with respect to pyrazinamide susceptibility, HIV

In the modified intention-to-treat population,
the median time to sputum-culture conversion
was faster in the bedaquiline group than in the
placebo group (83 days vs. 125 days), for a hazard
ratio for conversion in the bedaquiline group of
2.44 (95% confidence interval [CI], 1.57 to 3.80;
P<0.001) (Fig. 3). The same analysis in the full
intention-to-treat population had similar results
(Fig. S2 in the Supplementary Appendix). The
treatment difference in a model with adjustment
for unequally distributed baseline factors (status
status, and baseline albumin grade) was still sig-
nificant and was similar to the estimate obtained
from the unadjusted model. (Pre-extensive drug
resistance was not included in the model because
the between-group difference was small [7%].)
More patients in the bedaquiline group than
in the placebo group had confirmed culture
conversion at both 24 and 120 weeks: 52 of 66
patients (79%) and 38 of 66 patients (58%) in the
two groups, respectively, at 24 weeks (P = 0.008)
and 41 of 66 patients (62%) and 29 of 66 patients

n engl j med 371;8 nejm.org august 21, 2014 727

The New England Journal of Medicine
Downloaded from nejm.org on July 6, 2015. For personal use only. No other uses without permission.
Copyright © 2014 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 1. Demographic and Clinical Characteristics in the Modified Intention-to-Treat Population at Baseline.*

Bedaquiline Placebo All Patients

Characteristic (N = 66) (N = 66) (N = 132)
Demographic characteristics
Median age (range) — yr 32 (18–63) 34 (18–57) 34 (18–63)
Male sex — no. (%) 45 (68) 40 (61)   85 (64)
Black race — no. (%)† 24 (36) 25 (38)   49 (37)
Clinical characteristics
HIV positivity — no. (%) 5 (8) 14 (21)   19 (14)
Cavitary disease of ≥2 cm — no. (%) 54 (82) 56 (85) 110 (83)
Baseline albumin grade of 1 to 3 — no. (%)‡ 28 (42) 42 (64)   70 (53)
Drug resistance — no./total no. (%)§
Limited to isoniazid and rifampin 39/54 (72) 46/58 (79) 85/112 (76)
Pre-extensive drug resistance¶ 15/54 (28) 12/58 (21) 27/112 (24)
Fluoroquinolone resistance 6/54 (11) 4/58 (7) 10/112 (9)
Amikacin, kanamycin, or capreomycin resistance 9/54 (17) 8/58 (14) 17/112 (15)
Sensitivity to ≥3 drugs in the background regimen 40/53 (75) 44/55 (80) 84/108 (78)
Pyrazinamide resistance‖ 38/56 (68) 33/59 (56) 71/115 (62)
Background regimen — no. (%)
Aminoglycosides 64 (97) 63 (95) 127 (96)
Fluoroquinolones   66 (100) 65 (98) 131 (99)
Ethionamide or protionamide 65 (98) 64 (97) 129 (98)
Pyrazinamide 65 (98) 61 (92) 126 (95)
Ethambutol 46 (70) 46 (70)   92 (70)
Cycloserine or terizidone 25 (38) 26 (39)   51 (39)
Other agent** 4 (6) 6 (9) 10 (8)

* There were no significant differences between the two study groups except for the proportion of patients who were
HIV-positive (P = 0.04) and those who had a baseline albumin grade of 1 to 3 (P = 0.02). Fisher’s exact test was used
for the categorical measures, and an asymptotic Wilcoxon test for the continuous measures.
† Race was reported by the investigators.
‡ Albumin was graded according to the Division of Microbiology and Infectious Diseases criteria14 as follows: grade 1,
less than the lower limit of the normal range to 3 g per deciliter; grade 2, less than 3 g per deciliter to 2 g per decili-
ter; and grade 3, less than 2 g per deciliter.
§ For 12 patients in the bedaquiline group and 8 patients in the placebo group, no confirmation of isoniazid and ri-
fampin resistance was available from the central laboratory.
¶ Pre-extensively drug-resistant tuberculosis is defined as multidrug-resistant tuberculosis with resistance to either any
second-line injectable drug (amikacin, kanamycin, or capreomycin) or any fluoroquinolone.
‖ Some isolates did not grow for resistance testing.
** Other agents included aminosalicylic acid, capreomycin, amoxicillin plus clavulanic acid, and streptomycin.

(44%), respectively, at 120 weeks (P = 0.04) (Fig. 4).
At 120 weeks, of the 25 patients in the bedaqui-
line group who did not have a response, 8 did
not have culture conversion, 6 had subsequent
reversion, and 11 discontinued the study after
culture conversion. Of the 37 patients in the
placebo group who did not have a response, 15
did not have culture conversion, 10 had subse-
quent reversion, and 12 discontinued the study
after culture conversion. In the bedaquiline
group, there was no relationship between the
area under the plasma concentration–time curve
over a 24-hour period at week 2 and the time to
conversion or conversion status (Fig. S3 in the
Supplementary Appendix).
Subgroup Analyses
In subgroup analyses at 120 weeks, sputum-cul-
ture conversion occurred in more patients in the
bedaquiline group than in the placebo group
among those who had isolates with resistance
only to isoniazid and rifampin (27 of 39 patients
[69%] and 20 of 46 patients [43%], respectively)
and among those who had isolates with pre-
extensive drug resistance (9 of 15 patients [60%]
and 5 of 12 patients [42%], respectively). In addi-

728 n engl j med 371;8 nejm.org august 21, 2014

The New England Journal of Medicine

Downloaded from nejm.org on July 6, 2015. For personal use only. No other uses without permission.
Copyright © 2014 Massachusetts Medical Society. All rights reserved.
 Tuberculosis Culture Conversion with Bedaquiline

Table 2. Adverse Events during 120 Weeks in the Intention-to-Treat Population.*

Bedaquiline Placebo
Variable (N = 79) (N = 81)
Median duration of overall treatment phase (range) — wk 91.7 (2.0–120.0) 94.1 (2.0–137.3)
Adverse event — no. (%)
Any 78 (99) 79 (98)
Related to treatment 55 (70) 56 (69)
Grade 3 or 4† 34 (43) 29 (36)
Leading to discontinuation of treatment 4 (5) 5 (6)
Serious adverse events — no. (%)‡ 18 (23) 15 (19)
Adverse event occurring in ≥20% of patients — no. (%)
Nausea 32 (41) 30 (37)
Arthralgia 29 (37) 22 (27)
Vomiting 23 (29) 22 (27)
Headache 23 (29) 18 (22)
Hyperuricemia 20 (25) 27 (33)
Hemoptysis 16 (20) 14 (17)

* There were no significant differences between the two groups in any category as calculated by means of Fisher’s exact
test in a post hoc analysis.
† Events were graded according to the Division of Microbiology and Infectious Diseases criteria.14
‡ Two serious adverse events were considered by the investigator to be possibly related to a study drug, including 2 events
of acute pancreatitis in 1 patient in the bedaquiline group and spontaneous abortion in 1 patient in the placebo group.

tion, there were more culture conversions in the
bedaquiline group than in the placebo group
among patients with isolates that were suscepti-
ble to pyrazinamide (13 of 18 patients [72%] and
14 of 26 patients [54%], respectively) and among
those with isolates that were resistant to pyrazin-
amide (23 of 38 patients [61%] and 11 of 33 pa-
tients [33%], respectively).
On the basis of the WHO definition of cure,15
in the modified intention-to-treat population at
120 weeks, more patients in the bedaquiline
group than in the placebo group were cured (38
of 66 patients [58%] and 21 of 66 patients [32%],
respectively; P = 0.003) (Fig. 4).
Drug Resistance
New resistance to at least one antituberculosis
drug developed in isolates from 2 patients in the
bedaquiline group and in 16 patients in the pla-
cebo group; 1 of the 2 patients in the bedaquiline
group (50%) and 9 of the 16 in the placebo group
(56%) did not have a response to treatment. Of
the latter patients, 6 of the 9 patients (67%) in the
placebo group were found to have isolates with
resistance either to injectable second-line drugs
or fluoroquinolones (pre-extensive drug resis-
tance, observed in 5 patients) or to both inject-
able second-line drugs and fluoroquinolones
(extensive drug resistance, observed in 1 patient);
the 1 patient in the bedaquiline group did not
have either pre-extensive or extensive drug resis-
tance. Though the number of paired isolates (10)
is limited, the isolate from 1 patient in the beda-
quiline group who had pre-extensive drug resis-
tance at baseline had an increase by a factor of
4 in the bedaquiline MIC at the end of the study,
as compared with baseline. No mutations were
observed in the ATP synthase operon.
Safety
During 120 weeks in the intention-to-treat popu-
lation, there were similar rates of adverse events,
treatment-related adverse events, and adverse
events leading to study discontinuation in the
two study groups (Table 2). The most frequent
adverse events were nausea, arthralgia, and vom-
iting. The severity of most adverse events was
grade 1 or 2.14
Overall, 10 of 79 patients (13%) in the beda-
quiline group and 2 of 81 patients (2%) in the
placebo group died (P = 0.02). (Detailed case re-
ports on all deaths are provided in Table S5 in
the Supplementary Appendix.) In the bedaqui-
line group, deaths occurred during study-drug
treatment in 1 patient and after study week 24 in
9 patients (median time after receipt of the last

n engl j med 371;8 nejm.org august 21, 2014 729

The New England Journal of Medicine
Downloaded from nejm.org on July 6, 2015. For personal use only. No other uses without permission.
Copyright © 2014 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l
Time to Culture Conversion
100
90
80
Positive Culture (%)
70
60
50
40
30 Bedaquiline plus
background
20 regimen
10
0 In this study, we found that 24 weeks of treat-
0 4 8 12 16
Weeks
No. at Risk
Bedaquiline 58 37 25 12
Placebo 61 53 40 30
Figure 3. Time to Sputum-Culture Conversion in the Modified Intention-to-
Treat Population.
Shown is the proportion of patients in each study group who had positive
results on Mycobacterium tuberculosis culture during the 24-week investiga-
tional treatment phase of the study. Patients who withdrew from the study,
who died, or who did not have sputum-culture conversion by week 24 were
considered to have had treatment failure in the primary analysis, regardless
of their culture status at the time of dropout or death. For these patients,
data were censored at their last assessment, so the proportion of patients
who had culture conversion cannot be derived from the data in the figure.
Analysis based on a Cox proportional-hazards model with adjustment for
study center and degree of radiographic lung cavitation showed significant-
ly faster conversion in the bedaquiline group than in the placebo group at
24 weeks (P<0.001). The number of patients at risk at each time point is
the number of patients who did not have culture conversion and who were
still participating in the study.
dose of study drug, 49.1 weeks; range 12.3 to
130.1), with 1 of these deaths occurring after
study week 120. In 6 patients, the deaths were
attributed to tuberculosis (in 5 patients in the
bedaquiline group and 1 in the placebo group).
There was no difference in the duration of fol-
low-up between the two study groups. No deaths
were considered to be related to the study drug
by an investigator who was unaware of the
group assignments, nor was there any associa-
tion between the deaths and bedaquiline plasma
concentrations or a QTcF interval of 500 msec or
more during the trial.
At study week 24, the mean change from
baseline in the QTcF was an increase of 15.4 msec
in the bedaquiline group and an increase of
3.3 msec in the placebo group (P<0.001). After
bedaquiline treatment ended, the QTcF gradu-
ally decreased, and the mean value was similar
to that in the placebo group by study week 60.
Only one patient in the bedaquiline group had a
730 n engl j med 371;8 nejm.org august 21, 2014
The New England Journal of Medicine
Downloaded from nejm.org on July 6, 2015. For personal use only. No other uses without permission.
Copyright © 2014 Massachusetts Medical Society. All rights reserved.
of m e dic i n e
QTcF prolongation of more than 500 msec (sin-
gle time point), as compared with no patients in
the placebo group. No direct relationship was
seen between bedaquiline or the bedaquiline
metabolite (M2) plasma level and corresponding
Placebo plus absolute QTcF values or changes in the QTcF.
background
regimen
There were no reports of clinically significant
dysrhythmia during the trial.
Discussion
20 24
ment with bedaquiline in combination with a
five-drug, second-line background regimen (con-
sistent with WHO recommendations for the
7 3
22 5 treatment of multidrug-resistant tuberculosis at
that time15) significantly shortened the time to
culture conversion and increased the rate of cul-
ture conversion at 24 weeks, as compared with
placebo plus the background regimen. These re-
sults confirm the faster culture conversion re-
ported in a previous phase 2b trial of 8 weeks of
bedaquiline.9 The treatment benefit of adding
bedaquiline to the background regimen that was
seen at 24 weeks in terms of the proportion of
patients with a response was durable and of sim-
ilar magnitude at the end of the 120-week trial.
Evaluation of the study outcome on the basis
of the modified WHO definitions for multidrug-
resistant tuberculosis supported our results at 24
weeks and 120 weeks. On the basis of the WHO
definition of cure, nearly twice as many patients
in the bedaquiline group as in the placebo group
were cured, a finding that addresses the unmet
need for improved long-term treatment outcomes
in patients with multidrug-resistant tuberculo-
sis. The treatment-success rate among patients
in the placebo group in our trial was lower than
that reported in a recent meta-analysis (32% vs.
54%).16 This finding might be explained by the
higher proportions of patients in our study, as
compared with the meta-analysis, who had iso-
lates that were positive for acid-fast bacilli (100%
vs. 66%), cavitary disease (83% vs. 52%), and
pyrazinamide resistance (62% vs. 26%). In addi-
tion, we collected triplicate sputum samples and
used liquid culture, which is more sensitive than
the solid medium used in the studies cited in the
meta-analysis.17
The inclusion of bedaquiline in the regimen
was associated with a reduced risk of pre-exten-
sive drug resistance or extensive drug resistance
and a reduced risk of additional resistance to
 Tuberculosis Culture Conversion with Bedaquiline

other background drugs. Culture conversion was

A Protocol-Defined Analysis
higher in the bedaquiline group than in the
100
placebo group, despite a greater proportion of 12 23 No conversion
90 Reversion to positive test
isolates with pyrazinamide resistance and pre-­ 80
9

Treatment Outcome (%)

Withdrawal after
extensive drug resistance at baseline in the 70 17
15 conversion
bedaquiline group. Both patients with multi- 60 Conversion
62 18
drug-resistant isolates and those with pre-exten- 50
sive drug-resistant isolates had more frequent 40 44
and more rapid culture conversion with bedaqui- 30
line than with placebo. 20
As in the 8-week study of bedaquiline,9,10 the 10
most frequent adverse events that we observed 0
Bedaquiline plus Placebo plus
were similar to those commonly seen in patients Background Background
with tuberculosis who are receiving second-line Regimen (N=66) Regimen (N=66)
treatment for multidrug-resistant tuberculosis.18,19
Adverse events leading to the discontinuation of B Analysis Based on WHO Definitions
100 3
bedaquiline were uncommon. Increased hepatic 12 Death
90 35
aminotransferase levels, which had been ob- 23
Withdrawal
80
Treatment Outcome (%)
served in preclinical studies, were seen more Treatment failure
70 Cure
frequently in the bedaquiline group than in the 8
60
placebo group (Table S4 in the Supplementary 50 58 30
Appendix), but only three patients (two of whom 40
had hepatitis B virus infection) discontinued the 30
32
assigned study drug. The use of bedaquiline was 20
associated with moderate prolongation in the 10
QT interval (mean, 15.4 msec at study week 24). 0
Bedaquiline plus Placebo plus
There is a risk of increased QT-interval prolonga- Background Background
tion for bedaquiline in combination with other Regimen (N=66) Regimen (N=66)
QT-interval–prolonging drugs, such as fluoro-
quinolones and 4-aminoquinoline antimalarial Figure 4. Study Outcomes at 120 Weeks According to the Protocol-Defined
drugs.4 Analysis and an Analysis Based on World Health Organization Definitions.
The reason for higher mortality in the beda- Shown are study outcomes in the modified intention-to-treat population
on the basis of the protocol-defined analysis method (Panel A) and World
quiline group than in the placebo group is un- Health Organization (WHO) definitions15 (Panel B) with respect to study
clear. All 6 patients whose deaths were attrib- data at 120 weeks. In the two analyses, patients in the bedaquiline group
uted to tuberculosis either did not have culture had higher rates of sputum-culture conversion than did those in the placebo
conversion or had conversion with subsequent group on the basis of a logistic model with treatment as the only covariate
reversion during the trial and had one or more (P = 0.04 for the study analysis and P = 0.003 for the WHO definitions). Per-
centages may not total 100 because of rounding.
risk factors for a poor outcome. In addition,
mortality in the placebo group was surprisingly
low, as compared with mortality in a meta- small size. Only 79 patients received bedaquiline
analysis involving 9153 patients with multidrug- in this study, which was initiated at a time when
resistant tuberculosis (15%)16 and in an open- information about the safety of bedaquiline was
label, phase 2 trial of bedaquiline involving 233 limited. Our conservative selection criteria lim-
patients with newly diagnosed or previously ited the inclusion of patients with HIV coinfec-
treated multidrug-resistant tuberculosis (7%).20 tion who were receiving antiretroviral therapy. A
The development of bedaquiline for indications planned phase 3 study will enroll a larger num-
for which a reasonably efficacious and safe alter- ber of patients, including HIV-positive patients
native exists (e.g., treatment of drug-sensitive receiving antiretroviral therapy. The enrollment
tuberculosis or preventive treatment) should be in our study was calculated to show significant
approached with caution until more data have differences in the surrogate end point of spu-
been collected to clarify the implication of the tum-culture conversion at 24 weeks rather than
excess deaths. clinical cure. However, the significant difference
One limitation of our trial is its relatively in cure rates was maintained at 120 weeks,

n engl j med 371;8 nejm.org august 21, 2014 731

The New England Journal of Medicine
Downloaded from nejm.org on July 6, 2015. For personal use only. No other uses without permission.
Copyright © 2014 Massachusetts Medical Society. All rights reserved.
 Tuberculosis Culture Conversion with Bedaquiline

which provides evidence of the potential useful- 120-week period, with more negative sputum
ness of this surrogate end point. The study did cultures, fewer culture reversions, and a reduced
not assess whether the use of bedaquiline can risk of evolution to a more resistant subtype.
simplify or shorten treatment. Supported by Janssen Pharmaceuticals.
In conclusion, the addition of bedaquiline to Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
a five-drug regimen for patients with multidrug- We thank the patients and their families, along with study-
resistant tuberculosis resulted in faster sputum- center staff and public health authorities, for their support; Jans-
culture conversion and a higher rate of culture sen study personnel, in particular Chrispin Kambili and Ross
Underwood for their critical review; David McNeeley, who previ-
conversion at 24 weeks, as compared with pla- ously worked for Janssen Pharmaceuticals; and Ian Woolveridge of
cebo. This effect remained significant during a Gardiner-Caldwell Communications for medical-writing support.

Appendix
The authors’ affiliations are as follows: the Division of Medical Physiology and the Department of Science and Technology/National
Research Foundation Centre of Excellence for Biomedical Tuberculosis Research and Medical Research Council Centre for Tuberculosis
Research, Faculty of Medicine and Health Sciences, Stellenbosch University, Cape Town (A.H.D.), and the Medical Research Council
and Kwazulu Research Institute for Tuberculosis and HIV, Durban (A.P.) — both in South Africa; the Center for Tropical Medicine and
Travel Medicine, Department of Infectious Diseases, Division of Internal Medicine, Academic Medical Center, University of Amsterdam,
Amsterdam (M.P.G.); Centro de Excelencia para el Control de la Tuberculosis Niño Jesús, Servicio de Neumología, Hospital María
Auxiliadora (J.M.R.), and Instituto de Medicina Tropical Alexander von Humboldt, Universidad Peruana Cayetano Heredia (E.G.) —
both in Lima, Peru; Central Tuberculosis Research Institute, Russian Academy of Medical Sciences, Moscow (I.V.); Riga East University
Hospital–Center for Tuberculosis and Lung Diseases, Riga, Latvia (V.L.); Janssen Infectious Diseases, Beerse, Belgium (K.A., N.B.,
M.H.-T., N.L., P.M., E.D.P., R.P.G.H.); and Janssen Research and Development, Titusville, NJ (T.D.M., B.D.).

References
1. Global tuberculosis report 2013. Ge-
neva: World Health Organization (http://
apps.who.int/iris/bitstream/10665/91355/1/
9789241564656_eng.pdf?ua=1).
2. Zumla A, Abubakar I, Raviglione M,
et al. Drug-resistant tuberculosis — cur-
rent dilemmas, unanswered questions,
challenges, and priority needs. J Infect Dis
2012;205:Suppl 2:S228-S240.
3. Andries K, Verhasselt P, Guillemont J,
et al. A diarylquinoline drug active on the
ATP synthase of Mycobacterium tuberculosis.
Science 2005;307:223-7.
4. Prescribing information for Sirturo
(bedaquiline) tablets. Titusville, NJ: Jans-
sen Pharmaceutical, 2012 (http://www
.accessdata.fda.gov/drugsatfda_docs/
label/2012/204384s000lbl.pdf).
5. Lounis N, Veziris N, Chauffour A,
Truffot-Pernot C, Andries K, Jarlier V.
Combinations of R207910 with drugs used
to treat multidrug-resistant tuberculosis
have the potential to shorten treatment
duration. Antimicrob Agents Chemother
2006;50:3543-7.
6. Tasneen R, Li SY, Peloquin CA, et al.
Sterilizing activity of novel TMC207- and
PA-824-containing regimens in a murine
model of tuberculosis. Antimicrob Agents
Chemother 2011;55:5485-92.
7. Veziris N, Ibrahim M, Lounis N, An-
dries K, Jarlier V. Sterilizing activity of
second-line regimens containing TMC207
in a murine model of tuberculosis. PLoS
One 2011;6(3):e17556.
8. Rustomjee R, Diacon AH, Allen J, et al.
Early bactericidal activity and pharmaco-
kinetics of the diarylquinoline TMC207
in treatment of pulmonary tuberculosis.
Antimicrob Agents Chemother 2008;52:
2831-5.
9. Diacon AH, Pym A, Grobusch MP, et al.
The diarylquinoline TMC207 for multi-
drug-resistant tuberculosis. N Engl J Med
2009;360:2397-405.
10. Diacon AH, Donald PR, Pym A, et al.
Randomized pilot trial of eight weeks of
bedaquiline (TMC207) treatment for multi-
drug-resistant tuberculosis: long-term out-
come, tolerability, and effect on emer-
gence of drug resistance. Antimicrob
Agents Chemother 2012;56:3271-6.
11. Canetti G, Rist N, Grosset J. Mesure
de la sensibilité du bacille tuberculeux aux
drogues antibacillaires par la méthode
des proportions. Rev Tuberc Pneumol
(Paris) 1963;27:217-72.
12. Kent PT, Kubica GP. Public health mi-
crobiology: a guide for the level III labora-
tory. Atlanta: Centers for Disease Control,
1985.
13. Fridericia LS. Die Systolendauer im
Elektrokardiogramm bei normalen Men-
schen und bei Herzkranken. Acta Med
Scand 1920;53:469-86.
14. Division of Microbiology and Infec-
tious Diseases toxicity grading, November
2007. Bethesda, MD: National Institute of
Allergy and Infectious Diseases (http://
www.niaid.nih.gov/LabsAndResources/
resources/DMIDClinRsrch/Documents/
dmidadulttox.pdf).
15. Guidelines for the programmatic man-
agement of drug-resistant tuberculosis:
emergency update 2008. Geneva: World
Health Organization (http://whqlibdoc.who
.int/publications/2008/9789241547581
_eng.pdf).
16. Ahuja SD, Ashkin D, Avendano M, et al.
Multidrug resistant pulmonary tuberculo-
sis treatment regimens and patient out-
comes: an individual patient data meta-
analysis of 9,153 patients. PLoS Med
2012;9(8):e1001300.
17. Diacon AH, Maritz JS, Venter A, et al.
Time to detection of the growth of Myco-
bacterium tuberculosis in MGIT 960 for
determining the early bactericidal activity
of antituberculosis agents. Eur J Clin Mi-
crobiol Infect Dis 2010;29:1561-5.
18. Sagwa E, Mantel-Teeuwisse AK, Ruswa
N, et al. The burden of adverse events dur-
ing treatment of drug-resistant tuberculosis
in Namibia. South Med Rev 2012;5:6-13.
19. Bloss E, Kuksa L, Holtz TH, et al. Ad-
verse events related to multidrug-resistant
tuberculosis treatment, Latvia, 2000-2004.
Int J Tuberc Lung Dis 2010;14:275-81.
20. Pym A, Diacon A, Conradie F, et al.
Bedaquiline as part of a multi-drug resis-
tant tuberculosis (MDR-TB) therapy regi-
men: final results of a single-arm, phase II
trial (C209). Int J Tuberc Lung Dis 2013;
17:Suppl 2:S236. abstract.
Copyright © 2014 Massachusetts Medical Society.

732 n engl j med 371;8 nejm.org august 21, 2014

The New England Journal of Medicine

Downloaded from nejm.org on July 6, 2015. For personal use only. No other uses without permission.
Copyright © 2014 Massachusetts Medical Society. All rights reserved.