Integrated Adaptive Platform for EXcellent Clinical Trials in STROKE (APEX-STROKE): A Multi-factorial, Multi-arm, Multi-stage, Randomized, Adaptive Platform Trial
Integrated Adaptive Platform for EXcellent Clinical Trials in STROKE (APEX-STROKE): A Multi-factorial, Multi-arm, Multi-stage, Randomized, Adaptive Platform Trial
Stroke remains one of the leading causes of death and disability, worldwide. Despite advances in stroke management, only a limited number of acute treatments-including thrombolysis, endovascular thrombectomy, hemicraniectomy, stroke unit care, and aspirin-have demonstrated clear benefit, highlighting the need for widely applicable interventions that improve outcomes. Although randomized controlled trials remain the most reliable method for evaluating treatment efficacy, conventional designs typically assess a single intervention under fixed assumptions and are often costly, resource-intensive, and time-consuming. Platform trials offer a more efficient alternative by enabling the simultaneous evaluation of multiple interventions and the ongoing addition or removal of treatment arms, thereby providing a promising approach to accelerate the identification of effective therapies and transform stroke clinical research. The overarching objective of APEX-STROKE is to identify the treatment/s associated with the highest chance of improving relevant patient outcomes in stroke.
Platform Inclusion Criteria:
Platform Exclusion Criteria
- There are no platform level exclusion criteria
Each state and domain will specify additional inclusion and exclusion criteria in the respective Domain-Specific Registration. Patients who fulfill the overall platform criteria will be assessed for enrollment into each active domain.
Stroke is a leading cause of death and disability worldwide. According to the Global Burden of Disease (GBD) 2021 Study, there were 11.9 million incident strokes, 93.8 million stroke survivors, and 7.3 million stroke-related deaths globally in 2021, making stroke the third leading cause of death after ischemic heart disease and COVID-19. Global disability-adjusted life years (DALYs) attributable to stroke rose from 121.4 million in 1990 to 160.5 million in 2021, with more than three-quarters of this burden borne by low- and middle-income countries.
Despite this burden, only a small number of acute stroke treatments have been proven effective, including thrombolysis, endovascular thrombectomy, hemicraniectomy, stroke unit care, and aspirin. Conventional randomized controlled trials test a single intervention under fixed assumptions and are often costly, resource-intensive, and slow. Adaptive platform trials (APTs) allow multiple interventions to be evaluated simultaneously within shared infrastructure, with interventions added or removed over time, offering greater efficiency than conventional designs.
APEX-STROKE is an investigator-initiated, multi-factorial, multi-arm, multi-stage, randomized, adaptive platform trial designed to identify the treatment(s) associated with the highest chance of improving patient outcomes in stroke. The platform is structured hierarchically: a Master Protocol establishes trial-wide processes, governance, and general statistical principles common to all domains; each disease State is further defined in a State Subprotocol; and each treatment Domain within a State is defined in a Domain-Specific Appendix (DSA). Full eligibility, interventions, and outcome measures for each active domain are specified in the respective DSA.
Depending on the domain, the statistical framework may be frequentist, Bayesian, or another appropriate approach, as specified in the relevant Statistical Analysis Appendix. Where pre-specified, adaptive analyses assess whether an intervention is superior, inferior, equivalent, or futile within a domain or in specific populations, and information may be "borrowed" across strata where justified. Where interactions between interventions in different domains are considered plausible, the statistical models may evaluate such interactions. Response Adaptive Randomization (RAR) may be used, where pre-specified, to adjust allocation probabilities as data accrue. Specific interventions, subgroups, or domains may be stopped, modified, or closed to enrollment based on pre-specified decision rules reviewed by an independent Data Safety Monitoring Board.
Unlike trial designs that prohibit co-enrollment, the platform design permits assessment of synergy, competitive interference, and safety of combined treatment strategies across domains, subject to pre-specified eligibility, randomization, and analysis rules. The intention-to-treat principle will be used for all primary and adaptive analyses unless otherwise specified at domain level.