Adapting clinical trial design to maintain meaningful outcomes during a multicenter asthma trial in the precision medicine era

Abstract

Precision medicine is expected to impact the care of people with asthma, given its high disease prevalence, heterogeneity of pathophysiologic mechanisms, and consequent clinical phenotypes. A novel phenotype-stratified clinical trial conducted by the NHLBI AsthmaNet Consortium, titled Steroids in Eosinophil Negative Asthma (SIENA), was a randomized, multicenter, clinical trial that prospectively stratified individuals according to their baseline level of sputum inflammation during a screening period. Two phenotypic strata were assigned based on an a priori defined extent of sputum eosinophilia (Eos Low versus Eos High). This article describes: the scientific premise for the trial design, including assumptions used for power calculations; modifications to the analysis plan implemented after the trial started due to a higher than expected prevalence of one phenotypic stratum which impacted the ability to accrue sufficient subjects within the planned budget and study period; investigator alternatives to address the strata imbalance weighing scientific impact and study feasibility; and the final modified SIENA study design and analysis plan. SIENA was successfully completed in a manner that maintained meaningful outcomes. We conclude with recommendations for incorporation of pre-specified contingency plans into phenotype-directed protocols, to address the potential for differences in observed compared to estimated prevalence of different phenotypes in a study population. These approaches can be applied to precision medicine trials for the future.

Trial registration: ClinicalTrials.gov NCT02066298.

Keywords: Biomarker-stratified; Induced sputum eosinophilia; Multicenter clinical trial; Participant recruitment; Phenotype-stratified; Precision medicine; Study design.

Copyright © 2018 Elsevier Inc. All rights reserved.

Figures

Figure 1.

SIENA study schema.

Figure 2.

Number of enrollments and randomizations by Stratum. Fig. 2 describes the number of SIENA particpants enrolled and randomized in the Eos High versus Eos Low strata.

Figure 3.

Proportion of participants randomized to Eos High Stratum. Fig. 3 represents the proportion of SIENA participants randomized to the Eos High stratum throughout the course of the trial.