On a global scale, tuberculosis (TB) is the single most important infectious cause of morbidity and mortality. The World Health Organization has estimated that one-third of the entire world's population carries latent TB infection. A key TB control strategy is therapy of latent TB infection (LTBI). The current standard regimen is 9 months of Isoniazid (9INH). This regimen has excellent efficacy if taken regularly, but its effectiveness is substantially reduced by poor compliance. Serious side effects, such as hepato-toxicity can occur. Three shorter alternatives have been recommended: 6 months INH (6INH), 2 months Rifampin - Pyrazinamide (2RIF-PZA) and 4 months Rifampin (4RIF). The regimen of 6INH is less efficacious than 9INH, while 2RIF-PZA has been largely abandoned because of serious toxicity. Based on some evidence in treatment of LTBI, and extrapolating from extensive experience with treatment of active TB, it is believed that 4RIF has similar efficacy as 9INH. Therefore, the investigators are initiating the first multi-site international randomized trial that will compare the effectiveness of 4RIF and 9INH in preventing active tuberculosis.

On a global scale, tuberculosis (TB) is the single most important infectious cause of morbidity and mortality. The World Health Organization has estimated that one-third of the entire world's population carries latent TB infection. Of these 8 million develop active disease, and 2 million die from TB each year. In Canada, and most other industrialized countries, the incidence of TB fell dramatically from 1900 until the late 1970's, but since then decline has slowed, and TB continues to cause significant morbidity, and mortality in disadvantaged populations.

A key TB control strategy is therapy of latent TB infection (LTBI). The current standard regimen is 9 months of Isoniazid (9INH). This regimen has excellent efficacy if taken regularly, but its effectiveness is substantially reduced by poor compliance. Serious side effects, such as hepato-toxicity can occur. Three shorter alternatives have been recommended: 6 months INH (6INH), 2 months Rifampin - Pyrazinamide (2RIF-PZA) and 4 months Rifampin (4RIF). Based on some evidence in treatment of LTBI, and extrapolating from extensive experience with treatment of active TB, it is believed that 4RIF has similar efficacy as 9INH.

The investigators have initiated a research program to evaluate the compliance, safety, costs and effectiveness of 4RIF for the treatment of LTBI. In the initial study of 116 patients, 4RIF was associated with significantly higher completion rates (90% vs 70%). The second phase of this study, conducted in Montreal, Toronto, Kingston, Saskatoon, Calgary, and Edmonton, plus Brazil and Saudi Arabia, compared the rates of serious adverse events (SAE) with the two regimens. These were adjudicated by an independent three member review panel, blinded to study drug. In this phase, among 420 subjects randomized to 4RIF overall rate of Grade 3-4 SAE was 2.4% compared to 5.6% among the 427 taking 9INH (p=.02). Grade 3-4 hepatotoxicity was very significantly lower (0.7% vs 3.8%; p=.003). Health system costs were also significantly lower, and completion rates significantly higher with 4RIF Therefore the conditions have been met to conduct the first randomized trial that will compare effectiveness of 4RIF and 9INH in preventing tuberculosis. Among the moderate to high risk subjects that will be eligible, the anticipated cumulative risk of active TB, if untreated, will be at least 3% during a follow-up of 28 months after randomization. We anticipate 50% completion rate with 9INH, providing an effectiveness of 45% (based on known efficacy of 90%). To detect superior effectiveness of 4RIF, assuming 80% completion, and 10% loss in follow-up, plus accounting for cluster randomization of household contacts we would require enrolment of 3283 subjects per arm, or a total of 6,566 subjects. This is reduced to 5720, because the 847 already randomized in Phase 2, were randomized, treated and followed for 28 months post-randomization to determine occurrence of active TB - using the same methods described in this proposal.We have assumed 4RIF efficacy of 90%, based on available evidence. If 50% of the 2,898 randomized to each group complete therapy and 28 months follow-up, this would provide more than 90% power, to confirm non-inferior efficacy of 4RIF, if the non-inferiority margin was 25% - equivalent to a minimum efficacy of 4RIF of 65%. (In other words, we would declare 4RIF non-inferior to 9INH if the efficacy of 4RIF was not more than 25% worse than 9INH.) This efficacy has been considered sufficient for authoritative recommendations of 6INH, which has had efficacy of 40-69% in trials Eligible consenting subjects will be randomized in equal numbers to 9INH or 4RIF, by a web-based registration and randomization program, stratified by site in blocks of variable size (2-8 subjects). Subjects will be followed by their usual providers during therapy and then every 3 months up to 28 months post-randomization or the occurrence of a study end-point. The primary study outcome is the occurrence, during the 28 months after randomization, of microbiologically or histologically confirmed active TB. The final diagnosis will be based on the majority opinion of an independent 3-member clinical review panel, who will review all subjects investigated for TB, without knowledge of study drug, nor the clinical diagnosis. Planned sub-group analysis will compare rates of active TB in those who complete treatment per protocol (efficacy). Secondary outcomes include occurrence of confirmed plus probable active TB, Grade 3-4 adverse events (judged by another blinded, independent 3-member panel as in Phase 2), occurrence of drug resistant active TB, and costs - from the health system perspective. To accomplish this, in Canada one site has been added in Vancouver, the site in Brazil is doubled, and new sites in Korea, Australia, Indonesia, Benin, Guinea and Ghana (West Africa) have been added.

In addition to the parent trial, the investigators will include an additional component called "the Biomarker Study". For this part of the study the investigators will take an additional 10 mls of blood pre-treatment, as well as at four and nine months after starting treatment. In total an extra 30ml of blood (equal to 2 tablespoons) will be required over 9 months, and two additional veni-punctures, since veni-punctures will not be routinely performed at the 4 and 9 month time points. Three of the 10 ml will be used for the QFT-GIT test, the remaining 7 ml will have the serum separated and stored at -80°C for future biomarker studies. For subjects who were randomized to 4RIF, the blood draw at nine months will require an extra visit to the clinic, for which study subjects will be compensated. No specimens will be stored for future genetic testing. For the present time only the Montreal site will participate in this component. Other sites will join once funding is secured. All lab assays will be performed in the final year of the study, to enhance cost-effectiveness.

Inclusion Criteria:

• Adult (age 18 years and older) with documented positive TST (or in the absence of TST, a documented positive QFT) and prescribed 9 months of Isoniazid for LTBI, following authoritative recommendations.

Exclusion Criteria:

• Patients who were contacts of TB cases known to be resistant to Isoniazid, Rifampin, or both.
• Known HIV-infected individuals on anti-retroviral agents whose efficacy would be substantially reduced by Rifampin, unless therapy can safely be changed to agents not affected by Rifampin.
• Pregnant women - Rifampin and Isoniazid are considered safe in pregnancy but therapy is usually deferred until 2-3 months post-partum to avoid fetal risk and the potential for increased hepato-toxicity immediately post partum.
• Patients on any medication with clinically important drug interactions with Isoniazid or Rifampin, which their physician believes would make either arm contra-indicated.
• Patients with a history of allergy/hypersensitivity to Isoniazid or to Rifampin, Rifabutin or Rifapentine.
• Patients with active TB. Patients initially suspected to have active TB can be randomized once this has been excluded.
• Patients who have already started LTBI therapy.