New England Journal of Medicine December 8, 2011

Three Months of Rifapentine and Isoniazid for Latent Tuberculosis Infection

Timothy R. Sterling et al.

Bottom Line

A 3-month once-weekly regimen of rifapentine and isoniazid (3HP) was noninferior to 9 months of daily isoniazid for preventing active tuberculosis, offering significantly higher completion rates and lower hepatotoxicity.

Key Findings

1. Confirmed tuberculosis developed in 0.19% (7 of 3,986) of the 3HP group compared to 0.43% (15 of 3,745) of the 9H group, achieving the prespecified noninferiority margin of 0.75 percentage points.
2. Treatment completion was significantly higher in the 3HP group at 82.1% versus 69.0% in the 9H group (P<0.001).
3. Investigator-assessed drug-related hepatotoxicity was notably lower in the 3HP arm (0.4%) compared to the 9H arm (2.7%) (P<0.001).
4. Permanent drug discontinuation due to adverse events was slightly higher with 3HP (4.9%) than with 9H (3.7%) (P=0.009), largely driven by flu-like systemic drug reactions.

Study Design

Design
RCT
Open-Label
Sample
7,731
Patients
Duration
33 mo
Median
Setting
Multinational
Population Individuals at high risk for latent tuberculosis progression, including household contacts of TB patients and recent tuberculin skin test converters.
Intervention 3 months of directly observed once-weekly rifapentine (900 mg) and isoniazid (900 mg) (3HP)
Comparator 9 months of self-administered daily isoniazid (300 mg) (9H)
Outcome Development of confirmed active tuberculosis

Study Limitations

The open-label design may have introduced bias in the reporting and investigator assessment of subjective adverse events.
The 3HP regimen was administered via directly observed therapy (DOT) whereas 9H was self-administered, making it difficult to differentiate whether improved completion rates were due to the shorter regimen itself or the DOT support.
Relatively low overall event rates of active tuberculosis necessitated a modified intention-to-treat approach and limited the statistical power for superiority analyses.
The initial study population lacked large numbers of young children and HIV-coinfected individuals on antiretroviral therapy, requiring subsequent trials to validate 3HP in those subgroups.

Clinical Significance

The PREVENT TB trial established 3HP as a standard-of-care, short-course regimen for latent tuberculosis infection. By condensing treatment to just 12 weekly doses, it dramatically improved patient adherence and completion rates while mitigating the hepatotoxicity historically associated with prolonged isoniazid monotherapy.

Historical Context

For decades, 9 months of daily isoniazid (9H) was the gold standard for treating latent tuberculosis infection. However, its real-world effectiveness was severely hampered by low treatment completion rates (often below 50%) and risks of liver toxicity. The search for a shorter, safer regimen culminated in the PREVENT TB trial by the TB Trials Consortium, which revolutionized global LTBI guidelines by validating the 3HP regimen.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

What are the distinct mechanisms of action of rifapentine and isoniazid, and why does combining a rifamycin with isoniazid allow for a dramatically shorter treatment course for latent tuberculosis?

Key Response

Isoniazid inhibits mycolic acid synthesis, targeting actively dividing mycobacteria. Rifapentine inhibits DNA-dependent RNA polymerase and has potent sterilizing activity against slow-growing or semi-dormant persisters. The synergistic effect and the long half-life of rifapentine allow for weekly dosing and a shortened 3-month course.

Resident
Resident

When counseling a patient starting the 3HP regimen, what specific adverse effect unique to this regimen compared to 9 months of isoniazid should you warn them about, and how does the monitoring for hepatotoxicity differ?

Key Response

Residents need to know that while 3HP has less hepatotoxicity than 9H, it is associated with a systemic drug reaction or flu-like syndrome (fever, myalgia, rash) typically occurring after a few doses. Monitoring for hepatotoxicity is less intensive for 3HP in healthy individuals, but the flu-like reaction often leads to premature discontinuation if not anticipated and managed.

Fellow
Fellow

The 3HP regimen in this landmark trial was administered via Directly Observed Therapy (DOT), while the 9H regimen was self-administered. How does this asymmetry in administration method confound the completion rate outcome, and what are the subspecialty implications for managing LTBI in patients on concomitant antiretroviral therapy (ART)?

Key Response

The higher completion rate of 3HP might be partly due to DOT support rather than just the shorter duration. Furthermore, for HIV patients, rifapentine is a strong CYP3A4 inducer, necessitating careful adjustment or avoidance of certain antiretrovirals (e.g., protease inhibitors) or choosing alternative LTBI regimens, a key consideration for infectious disease specialists.

Attending
Attending

Given that 3HP demonstrated higher completion rates but required directly observed therapy (DOT) in the trial, how do you operationalize this in a resource-limited outpatient setting where DOT is not feasible, and how do you weigh the risk of systemic hypersensitivity reactions against the benefits of higher completion?

Key Response

Attendings must balance clinic resources. Requiring DOT for LTBI is highly burdensome. Post-trial evidence supported self-administered 3HP (SAT), changing clinical practice, but attendings must implement systems to actively monitor for the hypersensitivity reactions that spike around doses 3 to 4 to prevent unnecessary emergency department visits while maximizing treatment completion.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

In designing a non-inferiority trial for a highly efficacious prevention regimen with low event rates like LTBI treatment, why is the selection of the non-inferiority margin critical, and why must both modified intention-to-treat (mITT) and per-protocol analyses be concordant to confidently declare non-inferiority?

Key Response

With extremely low active TB incidence, a wide non-inferiority margin could mask clinically unacceptable inferiority. In non-inferiority trials, mITT can bias toward the null (making treatments look similar due to non-adherence), so establishing true non-inferiority requires demonstrating it in the per-protocol population as well, proving the regimen works when actually taken as directed.

Journal Editor
Journal Editor

The trial utilized an open-label design with different administration strategies (DOT for 3HP vs SAT for 9H). As an editor, how do you evaluate the risk of detection bias for the primary outcome of active TB, and how does the differential dropout rate due to adverse events threaten the trial's internal validity?

Key Response

Open-label designs risk biased ascertainment of the endpoint (evaluators might investigate symptoms differently based on known treatment arm). Furthermore, 3HP had higher discontinuation due to adverse events; if these patients were lost to follow-up, it could artificially lower the observed TB rate in the 3HP arm, complicating the non-inferiority claim and requiring robust sensitivity analyses.

Guideline Committee
Guideline Committee

Based on the findings of this study and subsequent data regarding self-administration, how should current CDC and WHO guidelines prioritize LTBI regimens, and what level of evidence supports transitioning 3HP from a conditional alternative to a strongly recommended first-line therapy?

Key Response

Current CDC (2020) and WHO guidelines strongly prefer short-course rifamycin-based regimens (3HP, 4R, 3HR) over 6 to 9 months of isoniazid due to higher completion rates and less hepatotoxicity. The committee would cite this trial as high-quality Grade A evidence driving the shift to making 3HP a preferred recommendation, updating older guidelines that heavily relied on 9H.

Clinical Landscape

Noteworthy Related Trials

2018

Menzies 4R Trial

n = 6,012 · NEJM

Tested

4 months of daily rifampin (4R)

Population

Adults with latent tuberculosis infection

Comparator

9 months of daily isoniazid (9H)

Endpoint

Confirmed active tuberculosis within 28 months

Key result: Four months of rifampin was noninferior to nine months of isoniazid for the prevention of active tuberculosis and had higher completion rates.
2019

ACTG A5279 (Brief TB) Trial

n = 3,000 · NEJM

Tested

1 month of daily rifapentine and isoniazid (1HP)

Population

HIV-infected patients with latent TB or living in high-burden areas

Comparator

9 months of daily isoniazid (9H)

Endpoint

First episode of active tuberculosis or death from TB

Key result: One month of daily rifapentine plus isoniazid was noninferior to nine months of isoniazid alone for preventing TB.
2021

WHIP3TB Trial

n = 4,027 · Lancet

Tested

Annual administration of 3 months of weekly rifapentine and isoniazid (3HP)

Population

HIV-positive adults on antiretroviral therapy in a high TB burden setting

Comparator

Single round of 3 months of weekly 3HP

Endpoint

Incidence of active tuberculosis

Key result: Annual 3HP did not provide additional protection against active TB compared to a single round of 3HP in patients on ART.

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