On August 14, 2019, the FDA approved TB Alliance’s
pretomanid tablets as part of a combination regimen with bedaquiline and linezolid for the treatment of people with a specific type of highly treatment-resistant tuberculosis (TB) of the lungs. Of note:
It’s the second drug approved under the Limited Population Pathway for Antibacterial and Antifungal Drugs (LPAD) pathway;
The TB Alliance was awarded a Tropical Disease Priority Review Voucher, which it can use to obtain priority review designation for a subsequent application that does not itself qualify for priority review.
Key findings from the FDA’s
Multi-Discipline Review and
Other Review(s): The safety and effectiveness was primarily demonstrated in a phase 3, open-label, noncomparative trial,
Nix-TB (NCT02333799; n=109), of a three-drug regimen of pretomanid, bedaquiline, and linezolid.
The applicant provided substantial evidence of effectiveness by comparing its single phase 3 clinical trial to historical controls. Of the 107 patients who were evaluated six months after the end of therapy, treatment success (absence of bacteriologic failure, relapse, or clinical failure) occurred in 89% (95% CI, 81-94%)—significantly exceeding the 50% among prespecified historical controls.
Purpose: To support the efficacy outcomes of the Nix-TB trial, the TB Alliance agreed to provide a literature summary and case-matched analysis of historical control data for extensively resistant TB patients.
What was done: The literature search results were selected to exclude articles in which patients were treated with any of the Nix-TB regimen drugs (pretomanid, bedaquiline, and linezolid) or delamanid given its shared nitroimidazole class with pretomanid. In the literature review, the applicant identified 18 studies. The table in the Multi-Discipline Review summarizes results for 16 of these studies and it excludes two in which the success rates were not available. Of the 16, eight were prospective studies and eight were retrospective (see
Banerjee et al.,
Keshavjee et al.,
Kvasnovsky et al.,
Leimane et al.,
Liu et al.,
Migliori et al.,
Mor et al., and
Padayatchi et al.). In some articles, definitions for treatment outcomes were not specified, but references were made to the World Health Organization (WHO) definitions and reporting framework for TB. While the applicant did not attempt a formal meta-analysis of the historical data, a DerSimonian and Laird random effects model of the 16 studies in the preceding table yielded an estimate for the historical treatment success rate of 28% (95% CI, 21-34%) for the treatment of extensively resistant TB patients. This provided some support for the pre-specified historical control success rate of 50% in Nix-TB being appropriately conservative. In addition to the literature review, the applicant submitted an analysis comparing bedaquiline plus pretomanid with a matched historical control group. The Nix-TB population included in the comparative analysis was based on the first 45 patients enrolled in the trial. The historical control group was based on patients from Brooklyn Chest Hospital in Cape Town, South Africa.
Outcome: Limitations of the literature review included the heterogeneity of studies and outcomes, geographic and temporal differences from Nix-TB, and possible selection of patients who would have been ineligible or not enrolled in a clinical trial such as Nix-TB. The use of all cause mortality was limited by different follow-up times and reporting periods across studies. The outcome of treatment success was standardized to some degree according to WHO criteria, but the timing of assessment was not standardized and was not necessarily comparable to Nix-TB. Nevertheless, the success rates reported in the literature generally were much lower than success rates in Nix-TB. The analysis comparing bedaquiline plus pretomanid with a matched historical control group partially addressed the limitations of the literature review related to the heterogeneity and lack of comparability in terms of geography, patient characteristics, and study assessments. However, patient-level data sets were not submitted for the matched historical control group, which limited the ability to reproduce the applicant’s results and perform a detailed assessment of the comparability between bedaquiline plus pretomanid and control groups.
