The New England Journal of Medicine October 23, 2014

Four-Month Moxifloxacin-Based Regimens for Drug-Sensitive Tuberculosis

Stephen H. Gillespie, Angela M. Crook, Timothy D. McHugh, et al. for the REMoxTB Consortium

Bottom Line

Substituting moxifloxacin for either isoniazid or ethambutol to shorten the treatment of drug-susceptible pulmonary tuberculosis from six to four months failed to achieve noninferiority, resulting in unacceptably high rates of relapse despite producing faster early bacterial clearance.

Key Findings

1. In the per-protocol analysis, a favorable outcome (absence of failure or relapse) was observed in 92% of patients receiving the standard 6-month control regimen [3.1.1].
2. The 4-month isoniazid-retained regimen (where moxifloxacin replaced ethambutol) achieved a favorable outcome in 85% of patients, resulting in a difference of 6.1 percentage points favoring the control (97.5% CI, 1.7 to 10.5).
3. The 4-month ethambutol-retained regimen (where moxifloxacin replaced isoniazid) achieved a favorable outcome in 80% of patients, resulting in a difference of 11.4 percentage points favoring the control (97.5% CI, 6.7 to 16.1).
4. Neither 4-month moxifloxacin regimen met the criteria for noninferiority, as the upper boundaries of the 97.5% confidence intervals significantly exceeded the pre-specified 6 percentage point noninferiority margin.
5. Despite higher long-term relapse rates, the moxifloxacin-containing regimens produced a more rapid initial decline in bacterial load; the hazard ratios for time to culture negativity ranged from 1.17 to 1.25 compared to the control group.
6. The incidence of grade 3 or 4 adverse events was similar across all arms: 19% (123 patients) in the control group, 19% (127 patients) in the isoniazid-retained group, and 17% (111 patients) in the ethambutol-retained group.

Study Design

Design
Randomized Controlled Trial
Double-Blind
Sample
1,931
Patients
Duration
18 mo
Median
Setting
Multicenter, global
Population Adults with newly diagnosed, uncomplicated, smear-positive pulmonary tuberculosis.
Intervention Two 4-month investigational regimens: one replacing ethambutol with moxifloxacin for 17 weeks (then 9 weeks placebo), and another replacing isoniazid with moxifloxacin for 17 weeks (then 9 weeks placebo).
Comparator Standard 6-month regimen (isoniazid, rifampin, pyrazinamide, and ethambutol for 8 weeks, followed by 18 weeks of isoniazid and rifampin) plus matching placebos to maintain blinding.
Outcome Treatment failure or relapse within 18 months after randomization (assessed via a pre-specified 6 percentage point noninferiority margin).

Study Limitations

A considerable proportion of randomized patients were excluded from the per-protocol analysis due to nonadherence or loss to follow-up, though sensitivity and modified intention-to-treat analyses yielded consistent results.
The strict double-dummy design necessitated a high daily pill burden for all patients, which may have compromised adherence and gastrointestinal tolerability in a real-world context.
The study underscored that 2-month sputum culture conversion—a traditional surrogate marker for treatment success—failed to accurately predict the long-term sterilizing cure and relapse prevention.

Clinical Significance

REMoxTB definitively demonstrated that a single drug substitution (moxifloxacin for either isoniazid or ethambutol) is insufficient to safely shorten the treatment of drug-susceptible tuberculosis from 6 to 4 months. The unacceptably high relapse rates in the experimental arms reaffirmed the 6-month standard of care and shifted the paradigm of TB drug development by proving that early bactericidal activity and accelerated culture conversion do not reliably translate into long-term sterilizing cure.

Historical Context

For decades, the standard of care for drug-susceptible tuberculosis has been a 6-month regimen of isoniazid, rifampin, pyrazinamide, and ethambutol. Because lengthy treatment courses lead to poor adherence and drug resistance, shortening therapy to 4 months has long been a primary objective in TB research. Early preclinical murine models and Phase 2 trials showed that substituting moxifloxacin—a fluoroquinolone with potent bactericidal properties—significantly accelerated early bacterial clearance. These promising early metrics led to massive global investment in Phase 3 shortening trials. REMoxTB was published simultaneously with other major studies (such as OFLOTUB and RIFAQUIN), all of which unfortunately failed to show noninferiority for 4-month fluoroquinolone-based regimens. This collective failure redirected researchers to explore multi-drug optimization strategies, such as incorporating high-dose rifapentine, which eventually achieved treatment shortening in TBTC Study 31.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

What is the mechanism of action of moxifloxacin, and why was it theoretically an attractive candidate to substitute for first-line agents like ethambutol or isoniazid in tuberculosis treatment?

Key Response

Moxifloxacin is a fluoroquinolone that inhibits DNA gyrase and topoisomerase IV. It was attractive due to its strong early bactericidal activity and excellent tissue penetration, which researchers theorized could accelerate the clearance of bacilli and allow for the shortening of the standard 6-month regimen to 4 months, thereby improving patient adherence.

Resident
Resident

Given the REMoxTB trial results, what remains the standard duration and baseline regimen for treating uncomplicated drug-susceptible pulmonary tuberculosis, and what is the clinical danger of stopping treatment at 4 months based solely on early culture conversion?

Key Response

The standard remains 6 months (2 months of HRZE followed by 4 months of HR) or the newer 4-month rifapentine-moxifloxacin regimen. REMoxTB proved that early 8-week culture conversion does not guarantee sterilization; stopping these specific moxifloxacin-substituted regimens at 4 months leads to unacceptably high clinical relapse rates despite initial negative cultures.

Fellow
Fellow

The REMoxTB trial demonstrated that while moxifloxacin regimens achieved faster sputum culture conversion at 8 weeks, they still resulted in higher relapse rates. What does this discordance teach us about the limitations of using 8-week culture conversion as a surrogate endpoint in TB drug trials?

Key Response

It highlights that Phase 2 early bactericidal activity (EBA) or 8-week culture conversion reflects the elimination of rapidly dividing extracellular bacilli but does not adequately capture the eradication of semi-dormant intracellular 'persister' organisms responsible for clinical relapse. Thus, 8-week conversion is an imperfect surrogate for definitive, relapse-free cure.

Attending
Attending

How do we reconcile the failure of the 4-month REMoxTB regimens to achieve noninferiority with the subsequent success of the 4-month moxifloxacin-rifapentine regimen (Study 31/A5349)? What key pharmacological differences drove these contrasting outcomes?

Key Response

REMoxTB substituted moxifloxacin for either INH or EMB but kept standard rifampin. Study 31 replaced rifampin with high-dose rifapentine (which has a longer half-life and greater sterilizing activity) alongside moxifloxacin. This teaches that substituting a fluoroquinolone alone is insufficient for treatment shortening; optimized, sustained rifamycin exposure (rifapentine) is critical for sterilizing the persister population.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

In noninferiority trials like REMoxTB, a predefined margin (e.g., 6 percentage points) is established. Critically evaluate the choice of this absolute margin for TB relapse trials. How might variations in the control group's relapse rate impact the power and ethical interpretation of such a noninferiority boundary?

Key Response

A 6 percent margin is standard in TB trials, but if the control group performs exceptionally well (e.g., a 4 percent relapse rate), a 6 percent absolute margin theoretically permits the experimental arm to have more than double the relapse rate while still being deemed noninferior. Researchers must weigh absolute versus relative margins when baseline event rates are low, balancing statistical feasibility with clinical acceptability.

Journal Editor
Journal Editor

From an editorial perspective, how does the handling of missing data or loss to follow-up at the 18-month primary endpoint complicate the interpretation of noninferiority in REMoxTB, and what sensitivity analyses are required to ensure robustness?

Key Response

In noninferiority trials, both intention-to-treat (ITT) and per-protocol (PP) analyses are crucial, as ITT can bias toward noninferiority if missing data is treated as failure in both arms, artificially narrowing the gap. A rigorous reviewer would demand robust tipping-point analyses and multiple imputation models to ensure that differential loss to follow-up did not mask a true difference in relapse rates.

Guideline Committee
Guideline Committee

Based on the REMoxTB trial, should current WHO or ATS/CDC/IDSA guidelines endorse the substitution of moxifloxacin for ethambutol or isoniazid to shorten treatment to 4 months for drug-susceptible TB, and what specific grade of recommendation applies to this strategy?

Key Response

Current guidelines strongly recommend AGAINST these specific 4-month REMoxTB regimens due to unacceptably high relapse rates (High-quality evidence, strong recommendation against). The guidelines instead maintain the 6-month HRZE regimen as standard, while recently endorsing a different 4-month regimen (rifapentine/moxifloxacin based on Study 31), explicitly highlighting that specific drug combinations and optimized dosages are required for successful treatment shortening.

Clinical Landscape

Noteworthy Related Trials

2014

Oflotub Trial

n = 1,836 · NEJM

Tested

4-month gatifloxacin-substituted regimen

Population

Adults with newly diagnosed smear-positive pulmonary tuberculosis

Comparator

Standard 6-month WHO-recommended regimen

Endpoint

Unfavorable outcome (treatment failure or relapse) at 24 months

Key result: The 4-month gatifloxacin regimen failed to show noninferiority to the standard 6-month regimen.
2014

RIFAQUIN Trial

n = 827 · NEJM

Tested

4-month and 6-month regimens replacing isoniazid with moxifloxacin and substituting rifapentine

Population

Patients with smear-positive pulmonary tuberculosis

Comparator

Standard 6-month daily regimen

Endpoint

Favorable outcome at 18 months

Key result: The 6-month weekly moxifloxacin-rifapentine regimen was noninferior, but the 4-month regimen was inferior to standard therapy.
2021

TBTC Study 31 / ACTG A5349

n = 2,516 · NEJM

Tested

4-month regimen with rifapentine and moxifloxacin

Population

Patients with newly diagnosed pulmonary tuberculosis

Comparator

Standard 6-month regimen

Endpoint

Survival free of tuberculosis at 12 months

Key result: The 4-month regimen containing rifapentine and moxifloxacin was noninferior to the standard 6-month regimen for tuberculosis.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis