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

Edi t or i a l s

Shortening Tuberculosis Treatment — A Strategic Retreat

Véronique Dartois, Ph.D., and Eric J. Rubin, M.D., Ph.D.

Our current treatment regimen for tuberculosis, that contained five drugs. The plan was to drop
which goes by the somewhat ironic name of two of these groups on the basis of early stop-
“directly observed therapy, short course,” is any- ping rules. In fact, none of these groups met
thing but short. Patients are treated, generally those standards, so enrollment was stopped in
on a daily basis, for 6 months, which necessi- two groups on the basis of logistic criteria (pill
tates an infrastructure to deliver and observe burden and regulatory concerns) in order to pre-
therapy. This logistic burden has led to a push serve statistical power. The two remaining groups
for shorter treatments. Although many initial received either high-dose rifampin plus linezolid
attempts failed, a recent study suggested that a or bedaquiline plus linezolid, each in combination
newer regimen could lead to a similar probabil- with isoniazid, pyrazinamide, and ethambutol.
ity of cure in 4 months.1 This is certainly an For the groups that received an intensified
advantage, and yet the newer regimen, if widely regimen, the strategy consisted of treatment for
used, would still require a similarly burdensome 8 weeks and then reassessment for persistent
infrastructure. Can we do even better and get to disease (symptoms and a positive sputum smear).
a point where the logistics would not be so limit- If the reassessment was negative, treatment was
ing? The investigators of the TRUNCATE-TB stopped; if positive, participants continued treat-
(Two-Month Regimens Using Novel Combina- ment for another 4 weeks. Those who remained
tions to Augment Treatment Effectiveness for positive could be switched to standard treatment
Drug-Sensitive Tuberculosis) trial,2 the results of to complete 24 weeks. Those who had a relapse
which are now published in the Journal, attempt- in any group were retreated with a standard
ed to do that. But rather than focusing on a regimen with adjustments made according to
regimen alone, they chose a treatment strategy. antibiotic susceptibility testing. Participants were
The trial design is a bit complex and initially followed closely for evidence of relapse through
had five groups. Participants had to have a nu- week 96. The primary outcome was a composite
cleic acid amplification test that was positive for of death, ongoing treatment, or active disease at
tuberculosis with no genotypic evidence of ri- week 96. For the treatment strategy to be de-
fampin resistance. The investigators excluded clared noninferior, the upper limit of the confi-
some higher-risk patients initially but then changed dence interval for the difference between the
the entry criteria to include this population. Par- strategy group and the standard-treatment group
ticipants who were randomly assigned to the in the risk of the primary outcome had to be less
control group received standard tuberculosis than 12 percentage points. This is a somewhat
treatment for 24 weeks (8 weeks of isoniazid, low threshold; for example, a recent treatment-
rifampin, ethambutol, and pyrazinamide, fol- shortening trial used a noninferiority margin of
lowed by 16 weeks of isoniazid and rifampin); 6.6 percentage points.1 As it turned out, one of
this group served as a comparator for a planned the strategy groups failed to meet even that
noninferiority analysis. Participants in the four loose criterion, whereas the other would have
other groups received an intensified regimen succeeded at either margin.

n engl j med 388;10  nejm.org  March 9, 2023 939

The New England Journal of Medicine
Downloaded from nejm.org by Brad Briones on March 27, 2023. For personal use only. No other uses without permission.
Copyright © 2023 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

The trial enrolled 675 participants at 18 sites have been a substantial issue in other trials.
in five countries. Impressively, almost all com- Because of these effects, whether these drugs
pleted the trial and follow-up period. Altogether, could be used safely for the treatment of drug-
7 participants (3.9%) had a primary-outcome susceptible tuberculosis has been unclear. In the
event in the control group, as compared with 21 TRUNCATE-TB trial, the toxic effects appeared
(11.4%) in the rifampin–linezolid group (adjust- to be quite limited. In fact, this is one of many
ed difference, 7.4 percentage points; 97.5% con- trials that suggest that the original concerns
fidence interval [CI], −1.7 to 13.2) and with 11 about bedaquiline might be overstated, at least
(5.8%) in the bedaquiline–linezolid group (ad- for patients who undergo prescreening with
justed difference, 0.8 percentage points; 97.5% electrocardiography.
CI, −3.4 to 5.1). With these results, only the Will these data change practice? Two months
strategy involving treatment with the bedaqui- of treatment might not be revolutionary but
line–linezolid regimen was declared noninferior could be very helpful. However, some obstacles
to standard treatment. In the bedaquiline–linez­ remain. There was a very high degree of adher-
olid group, 162 participants (85.7%) did not ence to treatment in this trial, far higher than
receive therapy beyond 8 weeks. According to the level likely to occur outside the context of a
the definitions used in the trial, extension of clinical trial. Lower adherence could mean in-
therapy was not a “failure” but was part of the creased treatment failure at 2 months. In addi-
treatment strategy. Altogether, the mean total tion, the treatment strategy involved careful as-
length of treatment in the bedaquiline–linezolid sessments of patients to identify those who
group (84.8 days) was less than half that in the would receive extended courses of therapy. Al-
standard-treatment group (180.2 days). though this approach is possible within the
One risk that is associated with a shorter confines of a trial, it could require considerable
course could be the development of antibiotic resources that are not now available in many
resistance. There were two cases of acquired tuberculosis control programs.
drug resistance in the bedaquiline–linezolid group Perhaps the biggest accomplishment of this
and none in the standard-treatment group. Bedaq­ trial is a step forward in the adaptive clinical
uiline has a long terminal half-life that gener- trial design that may help to accelerate regimen
ates lingering subtherapeutic concentrations for development and to rapidly test many more
several months after the end of therapy, which 2-month therapies that are selected on the basis
results in de facto monotherapy and a prolonged of recent treatment-shortening trial results.4-7
window for the potential acquisition of drug For shorter treatments, positive results that are
resistance in cases of relapse. Although a much similar to the results for standard treatment and
larger number of patients would need to be are observed across various patient populations,
treated to detect any significant difference, the including those with a high burden of cavitary
small number of cases of drug resistance in this tuberculosis, would garner the confidence needed
trial does not pose substantial concerns. to influence practice in lower-resource settings.
In many ways, the results of this trial are not Treatment algorithms such as that used in the
surprising. We have long known that most pa- TRUNCATE-TB trial are fundamental to tuber-
tients with tuberculosis who are treated even culosis control. Although implementing them
with standard regimens do not have a relapse could be a challenge, any added burden might be
after 4 months of treatment; in fact, several ap- offset by reduced costs, better adherence, and in-
pear to be cured after only 2 months of treat- creased patient satisfaction. Thus, for tuberculo-
ment.3 And the inclusion of bedaquiline and linez­ sis, a strategy might be more than just a regimen.
olid in regimens for drug-resistant tuberculosis Disclosure forms provided by the authors are available with
has allowed for shorter regimens.4-6 What is the full text of this editorial at NEJM.org.

striking is that each of these drugs has posed From the Center for Discovery and Innovation, Nutley, NJ (V.D.).
considerable concerns about toxic effects in the
past. Bedaquiline still carries a black-box warn- This editorial was published on February 20, 2023, at NEJM.org.
ing that resulted from very early trials showing
1. Dorman SE, Nahid P, Kurbatova EV, et al. Four-month rifa-
increased mortality among treated patients. Linez­ pentine regimens with or without moxifloxacin for tuberculosis.
olid can lead to dose-limiting toxic effects that N Engl J Med 2021;​384:​1705-18.

940 n engl j med 388;10  nejm.org  March 9, 2023

The New England Journal of Medicine

Downloaded from nejm.org by Brad Briones on March 27, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
 Editorials

2. Paton NI, Cousins C, Suresh C, et al. Treatment strategy for 6. Conradie F, Bagdasaryan TR, Borisov S, et al. Bedaquiline–
rifampin-susceptible tuberculosis. N Engl J Med 2023;​388:​873-87. pretomanid–linezolid regimens for drug-resistant tuberculosis.
3. Fox W, Ellard GA, Mitchison DA. Studies on the treatment of N Engl J Med 2022;​387:​810-23.
tuberculosis undertaken by the British Medical Research Coun- 7. Mok J, Lee M, Kim DK, et al. 9 Months of delamanid, linezo­
cil tuberculosis units, 1946–1986, with relevant subsequent pub- lid, levofloxacin, and pyrazinamide versus conventional therapy
lications. Int J Tuberc Lung Dis 1999;​3:​Suppl 2:​S231-S279. for treatment of fluoroquinolone-sensitive multidrug-resistant
4. Conradie F, Diacon AH, Ngubane N, et al. Treatment of tuberculosis (MDR-END): a multicentre, randomised, open-label
highly drug-resistant pulmonary tuberculosis. N Engl J Med phase 2/3 non-inferiority trial in South Korea. Lancet 2022;​400:​
2020;​382:​893-902. 1522-30.
5. Nyang’wa B-T, Berry C, Kazounis E, et al. A 24-week, all-oral
regimen for rifampin-resistant tuberculosis. N Engl J Med 2022;​ DOI: 10.1056/NEJMe2300413
387:​2331-43. Copyright © 2023 Massachusetts Medical Society.

Temperature Management after Cardiac Arrest — All In or Fold?

Stephen Bernard, M.D., and Janet Bray, Ph.D.

Outcomes after out-of-hospital cardiac arrest have uncertainty by testing whether a shorter period
improved over the past 20 years; however, the of fever prevention after an out-of-hospital car-
benefits of some methods used to treat these diac arrest would be beneficial or harmful.5 In
patients remain uncertain. In particular, the their trial, 789 patients who were unconscious
body temperature to be targeted during the first after an out-of-hospital cardiac arrest were ran-
72 hours of care after an out-of-hospital cardiac domly assigned to a 36°C target for 24 hours,
arrest remains unclear despite multiple high- followed by fever prevention (target body tempera-
quality randomized trials having been conduct- ture, <37°C) for either an additional 12 hours or
ed to investigate the issue.1-5 an additional 48 hours (for a total duration of
Early trials of therapeutic hypothermia tar- fever prevention of 36 or 72 hours). By 72 hours,
geted a body temperature of 32° to 34°C for 12 a body temperature higher than 37.7°C had oc-
hours1 or 24 hours2 but without a specific target curred in 50.1% of the patients in the 36-hour
temperature in the control group. Subsequently, group and in 38.1% of the patients in the 72-
the Target Temperature Management 33°C ver- hour group. Despite these findings, death or
sus 36°C after Out-of-Hospital Cardiac Arrest severe disability or coma within 90 days occurred
(TTM) trial compared targets of 33°C and 36°C in 32.3% of the patients in the 36-hour group
for 28 hours after randomization, which was and in 33.6% of the patients in the 72-hour
followed in both groups by rewarming to 37°C group (P = 0.70). These results should be inter-
until 36 hours after randomization and then by preted in light of fact that they are from a sub-
“fever prevention” (targeting a temperature of ordinate study within the Blood Pressure and
<37.5°C) until 72 hours after the cardiac arrest. Oxygenation Targets in Post Resuscitation Care
No significant difference in mortality was seen.3 (BOX) trial,6,7 in which oxygenation and blood-
The TTM2 trial evaluated fever prevention (“nor- pressure targets after out-of-hospital cardiac rest
mothermia”) for 72 hours as compared with were evaluated, without a separate statistical
therapeutic hypothermia (body temperature, 33°C) analysis plan or formal sample-size calculation
for 28 hours followed by slow rewarming until for evaluating the duration of fever prevention.
40 hours after randomization and fever preven- The power to detect small differences may there-
tion until 72 hours after randomization.4 Again, fore have been limited.
there was no significant difference in outcomes. Thus, the duration of fever prevention to be
Although fever prevention was included in used after an out-of-hospital cardiac arrest re-
the intervention and control groups of the TTM mains uncertain. Moreover, there is limited evi-
trials, its use during the first 72 hours after ad- dence that the onset of early fever after an out-
mission for an out-of-hospital cardiac arrest of-hospital cardiac arrest is actually harmful.
lacks a firm evidence base. In this issue of the Indeed, a retrospective analysis involving a large
Journal, Hassager and colleagues address this cohort of such patients in the United Kingdom

n engl j med 388;10  nejm.org  March 9, 2023 941

The New England Journal of Medicine
Downloaded from nejm.org by Brad Briones on March 27, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.