A seamless Phase IIB/III adaptive outcome trial: Design rationale and implementation challenges

Background: The licensed four-valent prophylactic human papillomavirus vaccine is highly efficacious in preventing cervical, vulvar, vaginal, and anal cancers and related precancers caused by human papillomavirus types 6, 11, 16, and 18. These four types account for approximately 70% of cervical cancers. A nine-valent human papillomavirus vaccine, including the four original types (6, 11, 16, and 18) plus the next five most prevalent types in cervical cancer (31, 33, 45, 52, and 58) could provide approximately 90% overall cervical cancer coverage. To expedite the nine-valent human papillomavirus vaccine clinical development, an adaptive, seamless Phase IIB/III outcome trial with ∼15,000 subjects was conducted to facilitate dose formulation selection and provide pivotal evidence of safety and efficacy for regulatory registrations. Purpose: We discuss the design rationale and implementation challenges of the outcome trial, focusing on the adaptive feature of the seamless Phase IIB/III design. Methods: Subjects were enrolled in two parts (Part A and Part B). Approximately 1240 women, 16–26 years of age, were enrolled in Part A for Phase IIB evaluation and equally randomized to one of three dose formulations of the nine-valent human papillomavirus vaccine or the four-valent human papillomavirus vaccine (active control). Based on an interim analysis of immunogenicity and safety, one dose formulation of the nine-valent human papillomavirus vaccine was selected for evaluation in the Phase III part of the study. Subjects enrolled in Part A who received the selected dose formulation of the nine-valent human papillomavirus vaccine or four-valent human papillomavirus vaccine continued to be followed up and contributed to the final efficacy and safety analyses. In addition, ∼13,400 women 16–26 years of age were enrolled in Part B, randomized to nine-valent human papillomavirus vaccine at the selected dose formulation or four-valent human papillomavirus vaccine, and followed for immunogenicity, efficacy, and safety. Results: A seamless Phase IIB/III design was justified by the extensive pre-existing knowledge of the licensed four-valent human papillomavirus vaccine and the development objectives for the nine-valent human papillomavirus vaccine. Subjects enrolled in Part A who received either the selected nine-valent human papillomavirus formulation or four-valent human papillomavirus vaccine contributed ∼10% of person-years of follow-up due to its earlier start—thereby maximizing the overall efficiency of the trial. Some of the challenges encountered in the implementation of the adaptive design included practical considerations during Phase IIB formulation selection by internal and external committees, End-of-Phase II discussion with health authorities and managing changes in the assay for immunological endpoints. Limitations: Application of the experience and lesson learned from this seamless adaptive design to other clinical programs may depend on case-by-case consideration. Conclusion: A seamless Phase IIB/III adaptive design was successfully implemented in this large outcome study. The development time of the second-generation nine-valent human papillomavirus vaccine was shortened due to improved statistical efficiency.