Bedaquiline–Pretomanid–Linezolid Regimens for Drug-Resistant Tuberculosis | NEJM

Trial Design

ZeNix was a partially blind, randomized trial that enrolled participants with pulmonary extensively drug-resistant (XDR) tuberculosis, pre-XDR tuberculosis, or rifampin-resistant tuberculosis. Participants with XDR tuberculosis had resistance to rifampin, a fluoroquinolone, and an aminoglycoside. Pre-XDR tuberculosis was defined as resistance to rifampin plus resistance to either a fluoroquinolone or an aminoglycoside. Rifampin-resistant tuberculosis was defined as Mycobacterium tuberculosis that was resistant to rifampin (with or without resistance to isoniazid) and did not respond to treatment or for which a second-line regimen had been discontinued because of side effects 6 months or more before enrollment.

All the participants received treatment for 26 weeks, with the option to extend treatment to 39 weeks if ongoing active disease was suspected between weeks 16 and 26. The full trial protocol is available with the full text of this article at NEJM.org.

Trial Participants

Participants were recruited from four trial sites in South Africa, one in the country of Georgia, one in Moldova, and five in Russia. The participants were 14 years of age or older (≥18 years of age in Russia and Moldova) and had had a documented positive sputum culture or molecular test for M. tuberculosis within 3 months before screening.

Participants were excluded if they had human immunodeficiency virus (HIV) infection and a CD4+ cell count of less than 100 per cubic millimeter; a risk of arrhythmia; an alanine aminotransferase level and an aspartate aminotransferase level higher than 3 times the upper limit of the normal range; or peripheral neuropathy of grade 3 or higher at baseline. Participants were excluded if they had previously received any of the three trial drugs or delamanid for 2 weeks or more before enrollment. The full inclusion and exclusion criteria are provided in Section S5. All the participants provided written informed consent.

Enrollment and Interventions

The participants were randomly assigned, in a 1:1:1:1 ratio, to one of the four linezolid regimens (either 1200 mg or 600 mg daily for either 26 weeks or 9 weeks) by trial site staff using an online portal. Randomization was stratified according to HIV status and classification of drug resistance.

In addition to linezolid, all participants received 26 weeks of bedaquiline (200 mg daily for 8 weeks, followed by 100 mg daily for 18 weeks) and pretomanid (200 mg daily for 26 weeks). The dose of linezolid could be reduced in a stepwise manner (1200 mg, 600 mg, 300 mg, or 0 mg) in response to adverse events. The participants, site staff, and trial team were unaware of the assigned duration and dose of linezolid treatment (see Section 4 in the protocol); matched placebo was provided for blinding. Adherence was monitored by direct observation if the participant was in the hospital or by checking medication cards and bottles for unused tablets at site visits.

Scheduled visits occurred weekly for the first 8 weeks, every 2 weeks until week 20, and then every 3 weeks until the end of treatment. The participants were followed for a minimum of 78 weeks after the completion of treatment, with scheduled visits in the follow-up period.

Microbiologic Assessments

At the screening visit, two sputum samples were obtained for smear microscopy, molecular testing for rifampin resistance (with the use of the Xpert MTB/RIF [Cepheid] or GenoType MTBDRplus assay [Hain Lifescience]), and culture in liquid medium in a Mycobacterial Growth Indicator Tube (MGIT) system (Becton Dickinson). Samples for culture in the MGIT system were then obtained weekly for 4 weeks and at weeks 6, 8, 10, 12, 16, 20, 23, and 26, and at each follow-up visit after the completion of treatment.

M. tuberculosis isolates from baseline cultures and the first positive culture on or after week 16 in participants who did not have a response to treatment were sent to a central laboratory for the determination of the MGIT minimum inhibitory concentration (MIC) of bedaquiline, pretomanid, and linezolid; for MGIT drug-susceptibility testing for first-line drugs (rifampin, isoniazid, pyrazinamide, ethambutol, and streptomycin), kanamycin, and moxifloxacin; and for whole-genome sequencing. M. tuberculosis isolates from participants with recurrence of tuberculosis were analyzed with the use of whole-genome sequencing¹¹ to distinguish between relapse and reinfection. For all drugs except pretomanid, the critical concentrations recommended by the World Health Organization were used to define resistance.¹² M. tuberculosis isolates with a pretomanid MIC of greater than 2 mg per liter were considered to be resistant.¹³ The laboratory manual, which includes full details of the microbiologic procedures, is provided in Section S15 in the Supplementary Appendix, available at NEJM.org.

Safety

Adverse events were recorded at every trial visit, and laboratory safety tests were performed weekly for the first 8 weeks and at scheduled visits during treatment. Electrocardiographic monitoring, examinations to assess color vision and visual acuity, and specific assessments for peripheral neuropathy with the use of a Brief Peripheral Neuropathy rating scale were also performed at scheduled intervals (Section S4 in the Supplementary Appendix).

Outcome Measures and End Points

The primary end point was the incidence of an unfavorable outcome, defined as treatment failure or disease relapse (clinical or bacteriologic) at 26 weeks after completion of treatment. In participants with bacteriologic treatment failure, negative culture status was not attained or maintained during treatment. Clinical treatment failure was defined as one of the following: a change from the protocol-specified tuberculosis treatment as a result of a lack of clinical efficacy, retreatment for tuberculosis, or tuberculosis-related death by 26 weeks after completion of treatment. Culture conversion was defined as at least two consecutive culture-negative samples obtained at least 7 days apart. In participants with relapse, negative culture conversion status was not maintained during follow-up, and a positive culture of an M. tuberculosis strain was confirmed as being genetically identical to that at baseline. Participants were considered to have a favorable outcome if they continued to have negative culture status during treatment to the end of follow-up and if they had not already been classified as having had an unfavorable outcome.

Secondary end points included bacteriologic or clinical treatment failure and relapse at 78 weeks after the end of treatment. Other secondary end points were the time to sputum culture conversion and the percentages of participants with culture conversion at specified time points.

Safety evaluations included adverse events, laboratory measurements, and death from any cause. Adverse events that occurred or worsened during the treatment period were defined as events that occurred between the start of treatment and 14 days after the end of treatment. The severity of adverse events was categorized according to grade, as defined by the Division of Microbiology and Infectious Diseases system,¹⁴ and site investigators provided an assessment of relatedness to trial medications. All the participants who received at least one dose of a trial drug were included in the safety analysis.

Trial Oversight

An independent data and safety monitoring committee reviewed safety and efficacy data throughout the trial. National and local ethics committees approved the trial. The TB Alliance, the trial sponsor, was responsible for the design and conduct of the trial. The authors vouch for the accuracy and completeness of the data and for the fidelity of the trial to the protocol.

Statistical Analysis

The primary efficacy analysis was conducted with the results of the MGIT culture. We hypothesized that the incidence of cure at 26 weeks after the end of therapy would be greater than 50% in each of the treatment groups. The incidence was estimated from the binomial proportion for participants with success criteria based on the lower boundary of the 95% confidence interval being greater than 50%. The trial did not have a control group.

We determined that a sample of 45 participants per group would provide the trial with more than 90% power to show that the lower boundary of the 95% confidence interval was greater than 50%, using a two-sided 5% significance level (and assuming a true cure rate of 80%). Intention-to-treat, modified intention-to-treat, and per-protocol analyses for each group were conducted (Section S6). The intention-to-treat population was defined as all participants who underwent randomization, with the exception of those who were excluded after the randomization period either because of protocol violations that occurred before randomization (and were detected after randomization) or because they did not have drug-resistant tuberculosis that was confirmed on the basis of a sputum sample obtained within 3 months before screening; the modified intention-to-treat population as the participants in the intention-to-treat population, with the exception of those who were lost to follow-up after successful treatment or who died from a cause that was adjudicated to be unrelated to tuberculosis; and the per-protocol population as the participants in the modified intention-to-treat population, with the exception of those who were excluded for additional protocol-related reasons.

The primary comparison against the target 50% efficacy was for the bedaquiline–pretomanid–linezolid regimen with linezolid at a dose of 1200 mg for 26 weeks, with the group that received 1200 mg of linezolid for 9 weeks and the group that received 600 mg of linezolid for 26 weeks being tested only if 1200 mg for 26 weeks was successful. The group that received 600 mg of linezolid for 9 weeks would be tested only if the dose of 600 mg for 26 weeks was successful. A Bonferroni adjustment was made for the comparison of the group that received 1200 mg of linezolid for 9 weeks with the group that received 600 mg for 26 weeks simultaneously, and 97.5% confidence intervals were reported for these groups. No formal statistical pairwise comparisons between groups were performed.