Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Mar 12;25(1):180.
doi: 10.1186/s13063-024-07999-w.

Estimands for clinical endpoints in tuberculosis treatment randomized controlled trials: a retrospective application in a completed trial

Isabelle R Weir  1 , Suzanne M Dufault  2   3 , Patrick P J Phillips  4
Affiliations

Estimands for clinical endpoints in tuberculosis treatment randomized controlled trials: a retrospective application in a completed trial

Isabelle R Weir et al. Trials. .

Abstract

Background: Randomized trials for the treatment of tuberculosis (TB) rely on a composite primary outcome to capture unfavorable treatment responses. However, variability between trials in the outcome definition and estimation methods complicates across-trial comparisons and hinders the advancement of treatment guidelines. The International Council for Harmonization (ICH) provides international regulatory standards for clinical trials. The estimand framework outlined in the recent ICH E9(R1) addendum offers a timely opportunity for randomized trials of TB treatment to adopt broadly standardized outcome definitions and analytic approaches. We previously proposed and defined four estimands for use in this context. Our objective was to evaluate how the use of these estimands and choice of estimation method impacts results and interpretation of a large phase III TB trial.

Methods: We reanalyzed participant-level data from the REMoxTB trial. We applied four estimands and various methods of estimation to assess non-inferiority of both novel 4-month treatment regimens against standard of care.

Results: With each of the four estimands, we reached the same conclusion as the original trial analysis that the novel regimens were not non-inferior to standard of care. Each estimand and method of estimation gave similar estimates of the treatment effect with fluctuations in variance and differences driven by the methods applied for handling intercurrent events.

Conclusions: Our application of estimands defined by the ICH E9 (R1) addendum offers a formalized framework for addressing the primary TB treatment trial objective and can promote uniformity in future trials by limiting heterogeneity in trial outcome definitions. We demonstrated the utility of our proposal using data from the REMoxTB randomized trial. We outlined methods for estimating each estimand and found consistent conclusions across estimands. We recommend future late-phase TB treatment trials to implement some or all of our estimands to promote rigorous outcome definitions and reduce variability between trials.

Trial registration: ClinicalTrials.gov NCT00864383. Registered on March 2009.

Keywords: Clinical trials; Estimands; Tuberculosis.

PubMed Disclaimer

Conflict of interest statement

The authors declare that they have no competing interests.

Figures

Fig. 1
Fig. 1
Occurrence of intercurrent and missing data events in REMoxTB trial
Fig. 2
Fig. 2
Point-range plot of risk difference estimates according to each estimand/estimation method. Each row corresponds with unique analyses within a given estimand. The vertical dotted line represents the non-inferiority margin of 6%. The results are shown as a point estimate of a risk difference and a corresponding 97.5% confidence interval. Results in orange are estimated with the Kaplan-Meier estimator (KM), results in green are estimated with the Cochrane Mantel Haenszel method (CMH), and results in blue are estimated according to the Principal Stratum method (PS). Point estimates represented by a square have implemented Inverse Probability of Censoring Weighting (IPCW) to handle certain intercurrent events. Point estimates represented by an asterisk have implemented Multiple Imputation (MI) to handle certain intercurrent events. Points estimates represented by a triangle have implemented the Principal Stratum (PS) method to handle certain intercurrent events. The remaining point estimates are represented by a circle meaning that no special statistical methods were used to handle intercurrent events
See this image and copyright information in PMC

Update of

References

    1. WHO . Global Tuberculosis Report. 2020.
    1. Ting NCH, El-Turk N, Chou MSH, Dobler CC. Patient-perceived treatment burden of tuberculosis treatment. PLoS One. 2020;15(10):e0241124. doi: 10.1371/journal.pone.0241124. - DOI - PMC - PubMed
    1. Dorman SE, Nahid P, Kurbatova EV, Phillips PPJ, Bryant K, Dooley KE, et al. Four-Month Rifapentine Regimens with or without Moxifloxacin for Tuberculosis. N Engl J Med. 2021;384(18):1705–18. doi: 10.1056/NEJMoa2033400. - DOI - PMC - PubMed
    1. Hills NK, Lyimo J, Nahid P, Savic RM, Lienhardt C, Phillips PPJ. A systematic review of endpoint definitions in late phase pulmonary tuberculosis therapeutic trials. Trials. 2021;22(1):515. doi: 10.1186/s13063-021-05388-1. - DOI - PMC - PubMed
    1. E9(R1) Statistical Principles for Clinical Trials: Addendum: Estimands and Sensitivity Analysis in Clinical Trials. U.S. Food and Drug Administration. 2020. https://www.fda.gov/regulatory-information/search-fda-guidance-documents.... Accessed 29 Feb 2024.

Associated data