Browse the corpus

Treatment Strategy for Rifampin-Susceptible Tuberculosis. BACKGROUND: Tuberculosis is usually treated with a 6-month rifampin-based regimen. Whether a strategy involving shorter initial treatment may lead to similar outcomes is unclear. METHODS: In this adaptive, open-label, noninferiority trial, we randomly assigned participants with rifampin-susceptible pulmonary tuberculosis to undergo either standard treatment (rifampin and isoniazid for 24 weeks with pyrazinamide and ethambutol for the first 8 weeks) or a strategy involving initial treatment with an 8-week regimen, extended treatment for persistent clinical disease, monitoring after treatment, and retreatment for relapse. There were four strategy groups with different initial regimens; noninferiority was assessed in the two strategy groups with complete enrollment, which had initial regimens of high-dose rifampin-linezolid and bedaquiline-linezolid (each with isoniazid, pyrazinamide, and ethambutol). The primary outcome was a composite of death, ongoing treatment, or active disease at week 96. The noninferiority margin was 12 percentage points. RESULTS: Of the 674 participants in the intention-to-treat population, 4 (0.6%) withdrew consent or were lost to follow-up. A primary-outcome event occurred in 7 of the 181 participants (3.9%) in the standard-treatment group, as compared with 21 of the 184 participants (11.4%) in the strategy group with an initial rifampin-linezolid regimen (adjusted difference, 7.4 percentage points; 97.5% confidence interval [CI], 1.7 to 13.2; noninferiority not met) and 11 of the 189 participants (5.8%) in the strategy group with an initial bedaquiline-linezolid regimen (adjusted difference, 0.8 percentage points; 97.5% CI, -3.4 to 5.1; noninferiority met). The mean total duration of treatment was 180 days in the standard-treatment group, 106 days in the rifampin-linezolid strategy group, and 85 days in the bedaquiline-linezolid strategy group. The incidences of grade 3 or 4 adverse events and serious adverse events were similar in the three groups. CONCLUSIONS: A strategy involving initial treatment with an 8-week bedaquiline-linezolid regimen was noninferior to standard treatment for tuberculosis with respect to clinical outcomes. The strategy was associated with a shorter total duration of treatment and with no evident safety concerns. (Funded by the Singapore National Medical Research Council and others; TRUNCATE-TB ClinicalTrials.gov number, NCT03474198.).

To evaluate a treatment strategy for tuberculosis, we conducted the Two-Month Regimens Using Novel Combinations to Augment Treatment Effectiveness for Drug-Sensitive Tuberculosis (TRUNCATE-TB) trial, a seamless phase 2–3, prospective, multicenter, international, adaptive, multigroup, multistage, randomized, open-label, noninferiority trial with a 96-week follow-up period. Because it was a strategy-comparison trial, the design and the approach to analysis differed from those used in regimen-comparison trials. The trial was designed by the investigators and coordinated by investigators at the National University of Singapore. Sanofi donated rifapentine, Pfizer donated linezolid, and Janssen funded whole-genome sequencing; these companies had no role in the design or conduct of the trial. An independent trial steering committee provided oversight. An independent data and safety monitoring committee reviewed safety and interim efficacy data. National and local ethics committees and regulatory agencies approved the trial. All participants provided written informed consent. The authors vouch for the accuracy and completeness of the data and the fidelity of the trial to the protocol, available with the full text of this article at NEJM.org.

rim efficacy data. National and local ethics committees and regulatory agencies approved the trial. All participants provided written informed consent. The authors vouch for the accuracy and completeness of the data and the fidelity of the trial to the protocol, available with the full text of this article at NEJM.org. Persons were eligible for inclusion in the trial if they were 18 to 65 years of age, had symptoms of tuberculosis or evidence of tuberculosis on a chest radiograph, and had a nucleic acid amplification test (Xpert MTB/RIF test, Cepheid) that was positive for tuberculosis without rifampin resistance. Persons who had a grade 3+ sputum smear, a cavity measuring more than 4 cm on a chest radiograph, or a positive test for human immunodeficiency virus (HIV) antibodies were initially not eligible; these exclusion criteria were later removed. A complete list of eligibility criteria and details regarding the changes are provided in Section S1 in the Supplementary Appendix, available at NEJM.org.

on a chest radiograph, or a positive test for human immunodeficiency virus (HIV) antibodies were initially not eligible; these exclusion criteria were later removed. A complete list of eligibility criteria and details regarding the changes are provided in Section S1 in the Supplementary Appendix, available at NEJM.org. Participants were randomly assigned to undergo either standard treatment or a strategy involving initial treatment with an 8-week regimen, extended treatment for persistent clinical disease, monitoring after treatment, and retreatment for relapse. There were four strategy groups with different initial regimens; participants were randomly assigned to the standard-treatment group or to one of the four strategy groups in equal proportions. Randomization was conducted by site staff with the use of an online system and was stratified according to trial site and relapse risk (Section S2).

egy groups with different initial regimens; participants were randomly assigned to the standard-treatment group or to one of the four strategy groups in equal proportions. Randomization was conducted by site staff with the use of an online system and was stratified according to trial site and relapse risk (Section S2). Standard treatment consisted of a standard dose of rifampin and isoniazid for 24 weeks in combination with pyrazinamide and ethambutol for the first 8 weeks. In the four strategy groups, initial treatment consisted of the following 8-week regimens: a high dose of rifampin and linezolid, a high dose of rifampin and clofazimine, rifapentine and linezolid, and bedaquiline and linezolid, each in combination with isoniazid, pyrazinamide, and ethambutol. In the strategy group with an initial rifapentine–linezolid regimen, ethambutol was replaced with levofloxacin (Section S3). The rationale for regimen selection is described in the protocol. The high dose of rifampin was 35 mg per kilogram of body weight initially and was reduced to 20 mg per kilogram starting on November 1, 2019. When a participant had persistent clinical disease (symptoms and a positive sputum smear) at week 8 or had missed doses, treatment with the five-drug regimen could be extended through week 12. When a participant had persistent clinical disease at week 12 or had adverse events at an earlier point, the five-drug regimen could be switched to standard treatment to complete a 24-week course of treatment.

smear) at week 8 or had missed doses, treatment with the five-drug regimen could be extended through week 12. When a participant had persistent clinical disease at week 12 or had adverse events at an earlier point, the five-drug regimen could be switched to standard treatment to complete a 24-week course of treatment. Treatment was supervised on a daily basis at least until completion of the four-drug phase in the standard-treatment group or until completion of the five-drug regimen in the four strategy groups. The approach to supervision was tailored to the participant. Monitoring involved assessment for symptoms and examination of sputum smears. Results of sputum cultures were also provided to clinicians, and additional tests were performed for suspected relapse. Participants who met prespecified criteria for relapse (Section S4) were retreated for at least 24 weeks with standard treatment, which was adjusted according to the participant’s resistance profile.

Results of sputum cultures were also provided to clinicians, and additional tests were performed for suspected relapse. Participants who met prespecified criteria for relapse (Section S4) were retreated for at least 24 weeks with standard treatment, which was adjusted according to the participant’s resistance profile. The trial design anticipated discontinuation of enrollment in two strategy groups on the basis of early stopping rules. However, the data and safety monitoring committee did not recommend discontinuation of enrollment in any trial group at the time of the interim analyses. The trial steering committee discontinued enrollment in two strategy groups (the rifampin–clofazimine strategy group and the rifapentine–linezolid strategy group) to ensure that sample-size requirements could be met for the formal evaluation of noninferiority in the two remaining strategy groups. The selection of these groups was pragmatic, with blinding to outcome data; the decision was based on pill burden, regulatory advice, and import license restrictions (Section S5).

roup) to ensure that sample-size requirements could be met for the formal evaluation of noninferiority in the two remaining strategy groups. The selection of these groups was pragmatic, with blinding to outcome data; the decision was based on pill burden, regulatory advice, and import license restrictions (Section S5). Clinic visits were scheduled every 1 to 4 weeks through week 24, then every 12 weeks through week 96; starting at week 30, monthly telephone visits were interspersed between clinic visits (Section S6). At every visit, tuberculosis symptoms were reviewed with a standard checklist, adverse events were graded according to standard criteria,12 and adherence to treatment was assessed on the basis of treatment records and participant interviews. A chest radiograph was obtained at screening, at weeks 8 and 96, at the end of treatment, and when relapse was suspected. Respiratory disability was assessed at week 96 with the use of the Medical Research Council (MRC) breathlessness scale, with disability defined as a grade of 3 or higher (on a scale from grade 1 to grade 5, with higher grades indicating a greater degree of activity-related breathlessness), and by means of spirometry, with disability defined as a forced expiratory volume in 1 second (FEV1) of less than 50% of the predicted value (Section S7).

disability defined as a grade of 3 or higher (on a scale from grade 1 to grade 5, with higher grades indicating a greater degree of activity-related breathlessness), and by means of spirometry, with disability defined as a forced expiratory volume in 1 second (FEV1) of less than 50% of the predicted value (Section S7). Sputum was obtained for smear examination and liquid culture (Mycobacteria Growth Indicator Tube system, Becton Dickinson) at every visit and when relapse was suspected. Smears were examined according to the method in routine use at each trial site and were graded according to World Health Organization guidelines. Drug resistance was determined with phenotypic susceptibility testing for standard drugs at baseline and for drugs associated with previous exposure at relapse (Section S8). Whole-genome sequencing was performed on isolates obtained at baseline and at relapse. Finally, acceptability was assessed with a questionnaire at weeks 48 and 96 (Section S9).

with phenotypic susceptibility testing for standard drugs at baseline and for drugs associated with previous exposure at relapse (Section S8). Whole-genome sequencing was performed on isolates obtained at baseline and at relapse. Finally, acceptability was assessed with a questionnaire at weeks 48 and 96 (Section S9). The primary outcome was a composite of death before week 96 or ongoing tuberculosis treatment or active tuberculosis at week 96. The primary outcome was assessed with a prespecified algorithm (Section S10). Because detection of and retreatment for relapse are an integral part of the treatment strategy that was assessed in this trial, these outcomes were not considered to be primary-outcome events if retreatment had been completed and the participant did not have active disease at week 96. Secondary outcomes included participant-centered, safety, and program-centered outcomes. The main secondary outcomes were total treatment time, grade 3 or 4 adverse events, and acquired drug resistance. Details regarding the conduct of the trial are provided in the protocol, which includes the statistical analysis plan.

econdary outcomes included participant-centered, safety, and program-centered outcomes. The main secondary outcomes were total treatment time, grade 3 or 4 adverse events, and acquired drug resistance. Details regarding the conduct of the trial are provided in the protocol, which includes the statistical analysis plan. We estimated that a sample of 180 participants in each trial group with complete enrollment (i.e., groups in which enrollment was not discontinued before the sample-size requirement was met) would provide the trial with 85% power to show the noninferiority of the treatment strategy to standard treatment with respect to the risk of a composite of death, ongoing treatment, or active disease at week 96. This estimation was based on a noninferiority margin of 12 percentage points, a one-sided significance level of 0.0125 (an adjustment for multiplicity, with the assumption of complete enrollment in two strategy groups), and the exclusion of 10% of participants from the analysis population. We assumed that a primary-outcome event would occur in 10% of the participants in each trial group (Section S11).

sided significance level of 0.0125 (an adjustment for multiplicity, with the assumption of complete enrollment in two strategy groups), and the exclusion of 10% of participants from the analysis population. We assumed that a primary-outcome event would occur in 10% of the participants in each trial group (Section S11). All analyses were performed in the intention-to-treat population, which excluded only persons who underwent randomization in error and were withdrawn before any trial medication was administered. Formal hypothesis testing was performed in only the two strategy groups with complete enrollment. Noninferiority of a treatment strategy could be concluded if the upper limit of the two-sided 97.5% confidence interval for the difference between the strategy group and the standard-treatment group in the percentage of participants with a primary-outcome event was less than 12 percentage points (Section S11). Differences were estimated with a generalized linear model with binomial distribution and adjustment for country and relapse risk.

r the difference between the strategy group and the standard-treatment group in the percentage of participants with a primary-outcome event was less than 12 percentage points (Section S11). Differences were estimated with a generalized linear model with binomial distribution and adjustment for country and relapse risk. Analysis of the primary outcome was performed across multiple subgroups that were defined according to baseline characteristics. Pre-specified sensitivity analyses were performed in the assessable population, which excluded participants who had an outcome that was classified as unassessable, and in the per-protocol population, which excluded participants who did not complete the protocol-specified initial treatment or had inadequate treatment during the first 56 days, unless the reason for inadequate treatment was death. The widths of confidence intervals that are used to report secondary outcomes have not been adjusted for multiplicity, and the intervals may not be used in place of hypothesis testing. Details regarding the reporting of primary and secondary outcomes and the approach to missing data are provided in Sections S12 and S13 and the statistical analysis plan. All analyses were performed with SAS software, version 9.4 (SAS Institute).

Persons were eligible for inclusion in the trial if they were 18 to 65 years of age, had symptoms of tuberculosis or evidence of tuberculosis on a chest radiograph, and had a nucleic acid amplification test (Xpert MTB/RIF test, Cepheid) that was positive for tuberculosis without rifampin resistance. Persons who had a grade 3+ sputum smear, a cavity measuring more than 4 cm on a chest radiograph, or a positive test for human immunodeficiency virus (HIV) antibodies were initially not eligible; these exclusion criteria were later removed. A complete list of eligibility criteria and details regarding the changes are provided in Section S1 in the Supplementary Appendix, available at NEJM.org.

Participants were randomly assigned to undergo either standard treatment or a strategy involving initial treatment with an 8-week regimen, extended treatment for persistent clinical disease, monitoring after treatment, and retreatment for relapse. There were four strategy groups with different initial regimens; participants were randomly assigned to the standard-treatment group or to one of the four strategy groups in equal proportions. Randomization was conducted by site staff with the use of an online system and was stratified according to trial site and relapse risk (Section S2).

Clinic visits were scheduled every 1 to 4 weeks through week 24, then every 12 weeks through week 96; starting at week 30, monthly telephone visits were interspersed between clinic visits (Section S6). At every visit, tuberculosis symptoms were reviewed with a standard checklist, adverse events were graded according to standard criteria,12 and adherence to treatment was assessed on the basis of treatment records and participant interviews. A chest radiograph was obtained at screening, at weeks 8 and 96, at the end of treatment, and when relapse was suspected. Respiratory disability was assessed at week 96 with the use of the Medical Research Council (MRC) breathlessness scale, with disability defined as a grade of 3 or higher (on a scale from grade 1 to grade 5, with higher grades indicating a greater degree of activity-related breathlessness), and by means of spirometry, with disability defined as a forced expiratory volume in 1 second (FEV1) of less than 50% of the predicted value (Section S7).

We estimated that a sample of 180 participants in each trial group with complete enrollment (i.e., groups in which enrollment was not discontinued before the sample-size requirement was met) would provide the trial with 85% power to show the noninferiority of the treatment strategy to standard treatment with respect to the risk of a composite of death, ongoing treatment, or active disease at week 96. This estimation was based on a noninferiority margin of 12 percentage points, a one-sided significance level of 0.0125 (an adjustment for multiplicity, with the assumption of complete enrollment in two strategy groups), and the exclusion of 10% of participants from the analysis population. We assumed that a primary-outcome event would occur in 10% of the participants in each trial group (Section S11). All analyses were performed in the intention-to-treat population, which excluded only persons who underwent randomization in error and were withdrawn before any trial medication was administered. Formal hypothesis testing was performed in only the two strategy groups with complete enrollment. Noninferiority of a treatment strategy could be concluded if the upper limit of the two-sided 97.5% confidence interval for the difference between the strategy group and the standard-treatment group in the percentage of participants with a primary-outcome event was less than 12 percentage points (Section S11). Differences were estimated with a generalized linear model with binomial distribution and adjustment for country and relapse risk.

From March 21, 2018, through January 20, 2020, a total of 1179 participants were screened and 675 were enrolled at 18 sites in Indonesia, the Philippines, Thailand, Uganda, and India. One person underwent randomization in error and was withdrawn immediately. Of the 674 participants who were included in the intention-to-treat population, 4 (0.6%) withdrew consent or were lost to follow-up (Fig. 1). The characteristics of the participants at baseline were similar in all five trial groups (Table 1), and the groups were broadly representative of populations of persons with tuberculosis (Table S1). All 660 participants who were alive and undergoing follow-up were evaluated at week 96; of these, 643 (97.4%) were evaluated in person and 17 (2.6%) by telephone. In the standard-treatment group, 98.3% of the participants completed the 24-week treatment course, and 3.3% underwent retreatment (Tables S2 and S3). In the four strategy groups, 91.5% of the participants overall (range, 73.8 to 94.7) completed the initial 8-week treatment course and stopped (mean qualifying time of initial treatment, 58 days), 6.5% overall switched to standard treatment (mainly because of adverse events) and completed a 24-week course, and 17.0% overall (range, 12.7 to 22.8) underwent retreatment.

ants overall (range, 73.8 to 94.7) completed the initial 8-week treatment course and stopped (mean qualifying time of initial treatment, 58 days), 6.5% overall switched to standard treatment (mainly because of adverse events) and completed a 24-week course, and 17.0% overall (range, 12.7 to 22.8) underwent retreatment. In the intention-to-treat analysis, a primary-out-come event occurred in 7 of the 181 participants (3.9%) in the standard-treatment group, as compared with 21 of the 184 participants (11.4%) in the strategy group with an initial rifampin–linezolid regimen (adjusted difference, 7.4 percentage points; 97.5% confidence interval [CI], 1.7 to 13.2; noninferiority criterion not met) and 11 of the 189 participants (5.8%) in the strategy group with an initial bedaquiline–linezolid regimen (adjusted difference, 0.8 percentage points; 97.5% CI, −3.4 to 5.1; noninferiority criterion met) (Table 2). Sensitivity analyses that were performed in the assessable population and in the per-protocol population had similar results. The results were also consistent across prespecified subgroups, including those defined according to relapse risk (Fig. 2A and 2B). The estimated percentages of participants who had a primary-outcome event in the two strategy groups with incomplete enrollment (10.3% and 4.8% in the intention-to-treat analysis) were similar to the percentages observed in the two strategy groups with complete enrollment. Results for the primary efficacy outcome are shown in Tables S4 through S9.

participants who had a primary-outcome event in the two strategy groups with incomplete enrollment (10.3% and 4.8% in the intention-to-treat analysis) were similar to the percentages observed in the two strategy groups with complete enrollment. Results for the primary efficacy outcome are shown in Tables S4 through S9. The mean total duration of treatment through week 96 was 180 days in the standard-treatment group, 106 days in the strategy group with an initial rifampin–linezolid regimen, and 85 days in the strategy group with an initial bedaquiline–linezolid regimen (Table 3). Participants in the strategy groups reported low levels of difficulty and anxiety related to the strategy and reported that the strategy had a positive effect on their motivation to take treatment; most participants in the strategy groups (71.6% and 78.3%) indicated that they would recommend the strategy to others. For other participant-centered outcomes, the results in the two strategy groups with complete enrollment were similar to those in the standard-treatment group. The results for participant-centered outcomes in the two strategy groups with incomplete enrollment were similar to those in the other two strategy groups. Results for participant-centered outcomes are shown in Table S10 and Fig. S1.

groups with complete enrollment were similar to those in the standard-treatment group. The results for participant-centered outcomes in the two strategy groups with incomplete enrollment were similar to those in the other two strategy groups. Results for participant-centered outcomes are shown in Table S10 and Fig. S1. The incidences of grade 3 or 4 adverse events, serious adverse events, and death did not differ significantly between the standard-treatment group and the two strategy groups with complete enrollment (Table 3). The incidence of respiratory disability and the change in the FEV1 through week 96 were also similar in the three groups. In the strategy group with an initial rifampin–linezolid regimen, the incidence of grade 3 or 4 adverse events was similar among participants enrolled before the high dose of rifampin was reduced and those enrolled after the rifampin dose reduction. Almost all relapses were classified as grade 1 or 2 in severity; 1 participant in the standard-treatment group, 5 in the rifampin–linezolid strategy group, and 1 in the bedaquiline–linezolid strategy group had a grade 3 or 4 adverse event that was related to a relapse or retreatment episode. The results for safety outcomes in the two strategy groups with incomplete enrollment were similar to those in the other two strategy groups. Results for safety outcomes are shown in Tables S11 through S15.

e–linezolid strategy group had a grade 3 or 4 adverse event that was related to a relapse or retreatment episode. The results for safety outcomes in the two strategy groups with incomplete enrollment were similar to those in the other two strategy groups. Results for safety outcomes are shown in Tables S11 through S15. Within the first 56 days, the mean percentage of days on which participants took the prescribed drugs was at least 95% in each trial group (Table 3). Few participants had treatment cessation that started during the first 56 days and lasted for at least 56 consecutive days; such cessation of treatment occurred at a similar frequency across trial groups. Two participants, both of whom were in the strategy group with an initial bedaquiline–linezolid regimen, had confirmed acquired drug resistance. One of these participants had resistance to isoniazid at baseline, missed 14 days (including 12 consecutive days) of treatment with all drugs during the first 4 weeks, and had a relapse at week 52. The other completed treatment at week 8 with no missed doses and had a relapse at week 36. Both had phenotypic and genotypic resistance to bedaquiline, as well as to clofazimine. Retreatment with standard treatment (with levofloxacin added for the first participant) was successful. Cases of unconfirmed acquired resistance to pyrazinamide and isoniazid are listed in Table S16. The estimated risk of relapse-associated transmission was low in the two strategy groups, with a mean transmission risk period of 3 days and with potential exposure of less than 0.1 additional new household contact per participant during the 96 weeks of follow-up. The results for program-centered outcomes in the two strategy groups with incomplete enrollment were similar to those in the other two strategy groups. Results for program-centered outcomes are shown in Table S11.

In the intention-to-treat analysis, a primary-out-come event occurred in 7 of the 181 participants (3.9%) in the standard-treatment group, as compared with 21 of the 184 participants (11.4%) in the strategy group with an initial rifampin–linezolid regimen (adjusted difference, 7.4 percentage points; 97.5% confidence interval [CI], 1.7 to 13.2; noninferiority criterion not met) and 11 of the 189 participants (5.8%) in the strategy group with an initial bedaquiline–linezolid regimen (adjusted difference, 0.8 percentage points; 97.5% CI, −3.4 to 5.1; noninferiority criterion met) (Table 2). Sensitivity analyses that were performed in the assessable population and in the per-protocol population had similar results. The results were also consistent across prespecified subgroups, including those defined according to relapse risk (Fig. 2A and 2B). The estimated percentages of participants who had a primary-outcome event in the two strategy groups with incomplete enrollment (10.3% and 4.8% in the intention-to-treat analysis) were similar to the percentages observed in the two strategy groups with complete enrollment. Results for the primary efficacy outcome are shown in Tables S4 through S9.

The mean total duration of treatment through week 96 was 180 days in the standard-treatment group, 106 days in the strategy group with an initial rifampin–linezolid regimen, and 85 days in the strategy group with an initial bedaquiline–linezolid regimen (Table 3). Participants in the strategy groups reported low levels of difficulty and anxiety related to the strategy and reported that the strategy had a positive effect on their motivation to take treatment; most participants in the strategy groups (71.6% and 78.3%) indicated that they would recommend the strategy to others. For other participant-centered outcomes, the results in the two strategy groups with complete enrollment were similar to those in the standard-treatment group. The results for participant-centered outcomes in the two strategy groups with incomplete enrollment were similar to those in the other two strategy groups. Results for participant-centered outcomes are shown in Table S10 and Fig. S1.

The incidences of grade 3 or 4 adverse events, serious adverse events, and death did not differ significantly between the standard-treatment group and the two strategy groups with complete enrollment (Table 3). The incidence of respiratory disability and the change in the FEV1 through week 96 were also similar in the three groups. In the strategy group with an initial rifampin–linezolid regimen, the incidence of grade 3 or 4 adverse events was similar among participants enrolled before the high dose of rifampin was reduced and those enrolled after the rifampin dose reduction. Almost all relapses were classified as grade 1 or 2 in severity; 1 participant in the standard-treatment group, 5 in the rifampin–linezolid strategy group, and 1 in the bedaquiline–linezolid strategy group had a grade 3 or 4 adverse event that was related to a relapse or retreatment episode. The results for safety outcomes in the two strategy groups with incomplete enrollment were similar to those in the other two strategy groups. Results for safety outcomes are shown in Tables S11 through S15.

Within the first 56 days, the mean percentage of days on which participants took the prescribed drugs was at least 95% in each trial group (Table 3). Few participants had treatment cessation that started during the first 56 days and lasted for at least 56 consecutive days; such cessation of treatment occurred at a similar frequency across trial groups. Two participants, both of whom were in the strategy group with an initial bedaquiline–linezolid regimen, had confirmed acquired drug resistance. One of these participants had resistance to isoniazid at baseline, missed 14 days (including 12 consecutive days) of treatment with all drugs during the first 4 weeks, and had a relapse at week 52. The other completed treatment at week 8 with no missed doses and had a relapse at week 36. Both had phenotypic and genotypic resistance to bedaquiline, as well as to clofazimine. Retreatment with standard treatment (with levofloxacin added for the first participant) was successful. Cases of unconfirmed acquired resistance to pyrazinamide and isoniazid are listed in Table S16. The estimated risk of relapse-associated transmission was low in the two strategy groups, with a mean transmission risk period of 3 days and with potential exposure of less than 0.1 additional new household contact per participant during the 96 weeks of follow-up. The results for program-centered outcomes in the two strategy groups with incomplete enrollment were similar to those in the other two strategy groups. Results for program-centered outcomes are shown in Table S11.

The results of the TRUNCATE-TB trial showed that a strategy involving initial treatment with an 8-week regimen that contained bedaquiline and linezolid was noninferior to standard treatment with respect to the risk of a composite clinical outcome at week 96. The efficacy of the strategy, as compared with standard treatment, was consistent across multiple subgroups that were defined according to baseline characteristics, including some that are indicative of severe disease and high relapse risk. This treatment strategy was associated with a shorter initial course and with a shorter total duration of treatment than was standard treatment. Also, participants who were treated according to this strategy reported a higher level of motivation to adhere to an 8-week initial course than to standard treatment. The 13-week reduction in the total treatment duration could allow program resources (both financial and human) — which are currently consumed by procuring, distributing, and supervising additional months of treatment — to be redeployed to enhance adherence support during a shorter period. This support may synergize with the increased individual motivation to better sustain adherence and thereby prevent the decrease in effectiveness that has been seen with standard treatment in its translation from clinical trials to programs.

tment — to be redeployed to enhance adherence support during a shorter period. This support may synergize with the increased individual motivation to better sustain adherence and thereby prevent the decrease in effectiveness that has been seen with standard treatment in its translation from clinical trials to programs. Follow-up after treatment, which is an essential component of the strategy, represents an additional burden for persons with tuberculosis and for treatment programs, as compared with the usual practice of immediate discharge after completion of standard treatment. However, only a few participants discontinued visits or reported difficulty with prolonged follow-up; most indicated that they would recommend the strategy to others, which suggests a positive overall experience, and the pragmatic monitoring approach is likely to be feasible for treatment programs. Future cost-effectiveness analyses to explore whether the additional costs of monitoring after treatment and associated retreatment are offset by the costs saved with reduced treatment duration are under way.

tive overall experience, and the pragmatic monitoring approach is likely to be feasible for treatment programs. Future cost-effectiveness analyses to explore whether the additional costs of monitoring after treatment and associated retreatment are offset by the costs saved with reduced treatment duration are under way. Our finding of no substantive overall increase in the incidence of grade 3 or 4 adverse events, serious adverse events, or respiratory disability (which has been previously observed with recurrent tuberculosis13) supports the premise that follow-up after treatment and early detection of recurrence mitigate the risk of harm from excess relapses. Overall, there was no evidence that the strategy promoted drug resistance, although the trial sample size was small; the finding of infrequent drug resistance was consistent with findings in previous trials of 4-month rifamycinbased regimens, in which less than 1% of participants acquired rifamycin resistance.9,14 A particular theoretical concern was that the long half-lives of bedaquiline and clofazimine (which are several months15,16) might result in exposure of residual viable bacteria to monotherapy after a short treatment course and generate frequent drug resistance. However, only two participants (1.1%) acquired bedaquiline (and clofazimine) resistance; this incidence is lower than the 2% incidence reported with bedaquiline use in 6-month regimens.17,18 One of these participants missed multiple consecutive treatment doses early in the course of treatment, which is a risk regardless of the regimen.19,20 The additional risk of transmission of tuberculosis with the strategy also appears to be small, with mitigation by follow-up after treatment and by early detection and treatment of relapses.

nts missed multiple consecutive treatment doses early in the course of treatment, which is a risk regardless of the regimen.19,20 The additional risk of transmission of tuberculosis with the strategy also appears to be small, with mitigation by follow-up after treatment and by early detection and treatment of relapses. This treatment strategy could be refined with the use of alternative drug regimens or monitoring approaches. Any initial regimen that has an acceptable side-effect profile and constrains relapse at modest levels may be suitable. Of the two strategy groups that were formally evaluated for noninferiority, only the group that was assigned to receive initial treatment with a bedaquiline–linezolid regimen met the noninferiority criterion at week 96. Bedaquiline may be well-suited for a 2-month initial course because its long half-life may extend efficacy beyond treatment completion, with a low risk of drug resistance. We used bedaquiline with four companion drugs to maximize potency and minimize the risk of resistance. We used a dose and duration of linezolid that were similar to those associated with safety among persons with drug-resistant tuberculosis,21 and the course was completed by more than 95% of participants who received it. Future analyses to evaluate the safety and efficacy of all the regimens used in this trial with respect to conventional phase 2 and 3 outcomes and pharmacokinetic–pharmacodynamic models are under way.

sons with drug-resistant tuberculosis,21 and the course was completed by more than 95% of participants who received it. Future analyses to evaluate the safety and efficacy of all the regimens used in this trial with respect to conventional phase 2 and 3 outcomes and pharmacokinetic–pharmacodynamic models are under way. For monitoring, we used regular assessment of symptoms and examination of sputum smears to inform decisions regarding treatment extension and to identify participants who should undergo additional investigation for relapse, which is a pragmatic approach that would be feasible for programs. We also provided clinicians with the results of cultures, which were performed primarily for research purposes. Future analyses might examine the extent to which culture results influence clinical management and whether other biomarkers might simplify, improve, or accelerate decision making.22,23 A point-of-care, non–sputum-dependent biomarker of disease activity (such as a blood RNA signature24,25) could work well with this strategy.

analyses might examine the extent to which culture results influence clinical management and whether other biomarkers might simplify, improve, or accelerate decision making.22,23 A point-of-care, non–sputum-dependent biomarker of disease activity (such as a blood RNA signature24,25) could work well with this strategy. The main strengths of this trial are the pragmatic design, the use of outcome measures that are relevant to persons with tuberculosis and to treatment programs, and the inclusion of diverse treatment clinics in high-burden countries, mainly in Asia. The open-label design is a limitation, but it was the most feasible option for regimens of different durations. The use of standardized assessments, the use of a prespecified algorithm for primary-outcome assessment, and the negligible trial attrition minimize potential bias. No HIV-positive participants were enrolled (although enrollment of such participants was permitted later in the trial), and further evaluation in this population is warranted.

zed assessments, the use of a prespecified algorithm for primary-outcome assessment, and the negligible trial attrition minimize potential bias. No HIV-positive participants were enrolled (although enrollment of such participants was permitted later in the trial), and further evaluation in this population is warranted. The results of this trial suggest that there may be value in considering a shift in tuberculosis management to a strategy involving initial treatment for the minimum duration needed to cure the majority of persons with tuberculosis, extended treatment for persistent clinical disease, and monitoring after treatment to detect relapse in the minority of persons who need retreatment. This treatment strategy provides a framework for the development of new, short, potent drug regimens and biomarkers for treatment monitoring to maximize cost-effectiveness and outcome benefit. Implementation research is vital to refine the strategy and evaluate outcomes in individual treatment programs and diverse populations before consideration of adoption at scale.