The New England Journal of Medicine MARCH 24, 2019

Evaluation of a Standardized Treatment Regimen of Anti-tuberculosis Drugs for Patients with Multidrug-resistant Tuberculosis (STREAM Stage 1)

Andrew J. Nunn, et al. (The STREAM Trial Team)

Bottom Line

The STREAM Stage 1 trial demonstrated that a shortened 9-to-11-month standardized treatment regimen for multidrug-resistant tuberculosis was statistically non-inferior to the longer 20-month World Health Organization-recommended regimen in terms of favorable treatment outcomes.

Key Findings

1. The 9-to-11-month shortened regimen showed a favorable outcome rate of 78.8%, compared to 79.8% in the longer 20-month control regimen, confirming non-inferiority within a 10% margin.

2. No significant difference in efficacy was observed between the two arms, suggesting the shorter regimen is a viable alternative for the management of multidrug-resistant tuberculosis.

3. Health economic evaluations conducted as part of the study indicated that the shorter treatment duration significantly reduced the cost of care for both patients and health systems compared to the traditional longer regimen.

Study Design

Design

RCT

Open-Label

Sample

424

Patients

Duration

27 mo

Median

Setting

Multicenter, international

Population Adult patients with newly diagnosed multidrug-resistant tuberculosis, without evidence of resistance to fluoroquinolones or injectable anti-TB drugs.

Intervention A standardized 9-to-11-month short-course regimen consisting of high-dose moxifloxacin, gatifloxacin (or moxifloxacin), clofazimine, ethambutol, and pyrazinamide, with an intensive phase containing kanamycin.

Comparator The standard 20-month treatment regimen according to 2011 WHO guidelines, consisting of an intensive phase of 8 months and a continuation phase of 12 months.

Outcome The proportion of patients with a favorable outcome (bacteriological cure or treatment completion) at 27 months post-randomization.

Study Limitations

• The study was conducted in a highly controlled clinical trial environment, which likely resulted in higher retention rates and better adherence than typically achieved in real-world routine programmatic settings.

• The open-label design of the trial could potentially introduce bias, although strict adherence to standardized assessment criteria was maintained.

• The trial specifically excluded patients with demonstrated resistance to fluoroquinolones or injectable agents, limiting the generalizability of these findings to more complex, drug-resistant cases.

Clinical Significance

These findings provided high-quality, randomized evidence to support a major shift in clinical practice toward shorter treatment durations for MDR-TB, facilitating improved patient adherence, reduced treatment-associated toxicity, and enhanced health system efficiency.

Historical Context

Prior to the STREAM trial, MDR-TB treatment guidelines were primarily based on expert consensus rather than randomized clinical trial data, necessitating a 20-to-24-month regimen associated with low success rates, poor tolerance, and significant side effects. The trial was designed to rigorously validate the efficacy of a 9-month 'Bangladesh regimen' that had shown promise in observational cohort studies, thereby addressing a critical evidence gap in global tuberculosis control.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

MDR-TB is defined by resistance to which two primary first-line antituberculosis agents, and how does the mechanism of the fluoroquinolones used in the STREAM trial compensate for this resistance?

Key Response

Multidrug-resistant tuberculosis (MDR-TB) is specifically defined as resistance to at least isoniazid and rifampin. Fluoroquinolones like moxifloxacin, a cornerstone of the STREAM short-course regimen, inhibit DNA gyrase and topoisomerase IV, providing a potent bactericidal mechanism that is independent of the pathways affected by isoniazid (cell wall synthesis) or rifampin (RNA polymerase).

Resident

The STREAM Stage 1 trial demonstrated non-inferiority of a 9-11 month regimen compared to a 20-month regimen. What are the most common Grade 3 or 4 adverse events associated with this shortened regimen that a clinician must monitor for during treatment?

Key Response

While the shortened regimen is non-inferior in efficacy, it carries a significant side-effect profile. Key monitoring includes QT prolongation (due to moxifloxacin and clofazimine), ototoxicity (if using an injectable like kanamycin, though later regimens moved away from this), and hepatotoxicity. In STREAM, 48.2% of the short-course group experienced Grade 3 or 4 adverse events, necessitating frequent laboratory and ECG monitoring.

Fellow

Analyze the impact of baseline fluoroquinolone resistance on the outcomes of the STREAM Stage 1 trial. To what extent can these results be extrapolated to patients with 'pre-XDR' TB?

Key Response

The STREAM Stage 1 regimen relied heavily on high-dose moxifloxacin. In patients with baseline resistance to fluoroquinolones, the efficacy of the short-course regimen is significantly compromised. Therefore, the STREAM results are primarily generalizable to patients with confirmed fluoroquinolone-susceptible MDR-TB; patients with pre-XDR (MDR plus fluoroquinolone resistance) require more individualized or newer regimens like BPaLM.

Attending

The STREAM trial utilized a non-inferiority margin of 10 percentage points. If the short-course regimen is 'statistically non-inferior' but has a higher absolute rate of serious adverse events, how does this shift the value proposition for public health programs versus individual patient care?

Key Response

From a public health perspective, the 9-11 month regimen is a breakthrough for resource-limited settings because it improves adherence and reduces program costs. However, for individual care, the high rate of Grade 3-4 toxicity (48.2% vs 45.4% in the long regimen) means clinicians must weigh the benefit of a shorter duration against a potentially more intense side-effect burden, highlighting the need for shared decision-making.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

Critique the use of a composite 'favorable outcome' endpoint in the STREAM trial. How might the inclusion of 'treatment change' as a component of failure impact the non-inferiority interpretation compared to a more restrictive endpoint like culture conversion or mortality?

Key Response

The composite endpoint included clinical, microbiological, and administrative factors (like changing more than two drugs). This 'pragmatic' approach reflects real-world failure but can be sensitive to clinician behavior in an open-label trial. If clinicians were quicker to change the regimen in the short-course arm due to perceived risk, it might artificially inflate the 'failure' rate, making it harder to prove non-inferiority but increasing the safety of the trial.

Journal Editor

The STREAM trial was an open-label study. Given the high rate of adverse events and the subjectivity involved in deciding to modify a treatment regimen, what specific steps could the authors have taken to minimize detection and performance bias regarding the primary outcome?

Key Response

In an open-label TB trial, detection bias is a major concern. To mitigate this, a Journal Editor would look for the use of an independent, blinded endpoint adjudication committee to review all 'unfavorable' outcomes and a standardized protocol for when a regimen change was mandatory versus at the clinician's discretion, ensuring that the 'favorable outcome' assessment remained objective.

Guideline Committee

The WHO initially issued a conditional recommendation for the shortened MDR-TB regimen based on observational data (the 'Bangladesh regimen'). How does the STREAM Stage 1 trial evidence affect the strength of this recommendation, and how should it be reconciled with the newer WHO preference for all-oral BPaLM-based regimens?

Key Response

STREAM Stage 1 provided the first Phase 3 randomized evidence to support the 9-11 month regimen, moving the level of evidence from 'low' (observational) to 'high' (RCT). However, because the STREAM Stage 1 regimen still included an injectable (kanamycin in many cases), current WHO guidelines (2022/2023 updates) now prioritize the 6-month all-oral BPaLM regimen (Bedaquiline, Pretomanid, Linezolid, Moxifloxacin) as the preferred standard, relegating the STREAM-style regimen to an alternative when BPaLM cannot be used.

Clinical Landscape

Noteworthy Related Trials

2010

Bangladesh Study

n = 515 · AJRCCM

Tested

9-month standardized short-course regimen

Population

Patients with multidrug-resistant tuberculosis

Comparator

Conventional treatment regimen

Endpoint

Favorable outcome at the end of treatment

Key result: The short-course regimen achieved an 87.9% success rate, significantly outperforming conventional longer regimens.

2022

TB-PRACTECAL Trial

n = 552 · NEJM

Tested

6-month BPaL/M regimen (bedaquiline, pretomanid, linezolid, and moxifloxacin)

Population

Patients with rifampin-resistant or multidrug-resistant tuberculosis

Comparator

Standard of care

Endpoint

Composite of death, treatment failure, or treatment discontinuation

Key result: The 6-month regimen was superior to the standard of care with significantly higher success rates and fewer adverse events.

2022

STREAM Stage 2

n = 563 · Lancet

Tested

6-month all-oral bedaquiline-containing regimen

Population

Patients with multidrug-resistant tuberculosis

Comparator

9-month STREAM Stage 1 regimen

Endpoint

Favorable status at week 76

Key result: The 6-month all-oral regimen was non-inferior to the 9-month regimen while improving patient experience and safety.

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