Faculty of Epidemiology and Population Health, Tropical Epidemiological Group, London School of Hygiene and Tropical Medicine, Keppel Street, London, WC1E 7HT, UK.
Trials. 2012 May 18;13:61. doi: 10.1186/1745-6215-13-61.
There have been no major advances in tuberculosis (TB) drug development since the first East African/British Medical Research Council short course chemotherapy trial 35 years ago. Since then, the landscape for conducting TB clinical trials has profoundly changed with the emergence of HIV infection, the spread of resistant TB bacilli strains, recent advances in mycobacteriological capacity, and drug discovery. As a consequence questions have arisen on the most appropriate approach to design and conduct current TB trials. To highlight key issues discussed: Is a superiority, equivalence, or non-inferiority design most appropriate? What should be the primary efficacy outcome? How to consider re-infections in the definition of the outcome? What is the optimal length of patient follow-up? Is blinding appropriate when treatment duration in test arm is shorter? What are the appropriate assumptions for sample size calculation?
Various drugs are currently in the development pipeline. We are presenting in this paper the design of the most recently completed phase III TB trial, the OFLOTUB project, which is the pivotal trial of a registration portfolio for a gatifloxacin-containing TB regimen. It is a randomized, open-label, multicenter, controlled trial aiming to evaluate the efficacy and safety of a gatifloxacin-containing 4-month regimen (trial registration: ClinicalTrial.gov database: NCT00216385).
In the light of the recent scientific and regulatory discussions, we discuss some of the design issues in TB clinical trials and more specifically the reasons that guided our choices, in order to best answer the trial objectives, while at the same time satisfying regulatory authority requirements.
When shortening TB treatment, we are advocating for a non-inferiority, non-blinded design, with a composite unfavorable endpoint assessed 12 months post treatment completion, and added trial procedures specifically aiming to: (1) minimize endpoint unavailability; and (2) distinguish between relapse and re-infection.
自 35 年前首次进行东非/英国医学研究委员会短程化疗试验以来，结核病（TB）药物开发方面没有取得重大进展。此后，由于 HIV 感染、耐药结核杆菌菌株的传播、分枝杆菌能力的最新进展以及药物发现，进行结核病临床试验的环境发生了巨大变化。因此，人们对设计和开展当前结核病试验的最合适方法产生了疑问。为了突出讨论的关键问题：最佳设计是优越性、等效性还是非劣效性？主要疗效终点应该是什么？如何在结果定义中考虑再感染？患者随访的最佳时间是多长？当试验组的治疗持续时间较短时，盲法是否合适？样本量计算的适当假设是什么？
目前有多种药物正在研发中。我们在本文中介绍了最近完成的 III 期结核病试验 OFLOTUB 项目的设计，该项目是包含加替沙星的结核病方案注册组合的关键试验。这是一项随机、开放标签、多中心、对照试验，旨在评估包含加替沙星的 4 个月方案的疗效和安全性（试验注册：ClinicalTrial.gov 数据库：NCT00216385）。
鉴于最近的科学和监管讨论，我们讨论了结核病临床试验中的一些设计问题，特别是指导我们选择的原因，以便最好地回答试验目标，同时满足监管机构的要求。
当缩短结核病治疗时间时，我们提倡采用非劣效性、非盲法设计，采用治疗完成后 12 个月评估的复合不利终点，并增加专门旨在：（1）最大限度地减少终点不可用性；（2）区分复发和再感染的试验程序。
