Two-Stage Adaptive Design in
M. Scott Harris, MD
Middleburg Consultants, Inc.
Professor of Medicine
Georgetown University School of Medicine
Making best use of scarce resources
• Multicenter trial initiated by Napo Pharmaceuticals in 2007
• Limited pool of potential patients
• One pivotal trial sufficient for approval
• Phase 2 trials have identified three potential doses
• Uncertain dose-response relationship
Three Options to Consider
1. Single Four-Arm Trial: Run a single Phase 3 four-arm trial;
four dose groups vs. placebo
2. Two Separate Trials: Run a four-arm Phase 2 trial to select
the best dose, then run a second independent Phase 3
3. Seamless Phase 2/3 Trial: Start with four arms, select the
best dose at an interim analysis, continue with two-arms,
and pool the data from both stages.
Sample Sizes Required for 80% Power
Method Sample Size
Bonferroni (4-arm) 540
Two Separate Trials 416
Phase 2/3 Posch-Bauer 350
X mg p.o. b.i.d.
Y mg p.o. b.i.d.
Z mg p.o. b.i.d.
n 50 50 50 50 200
* mg p.o. b.i.d.
50 + 75 = 125 50 + 75 = 125 350
Stage I: Dose selection
XYZ = 125 mg, 250 mg, or 500 mg
Primary and secondary efficacy
Study Stages and Interim Analysis
• A statistical plan was developed to guarantee control of
type-1 error, employing Posch et al 2005.
• The SAP was approved by FDA through Special Protocol
• The trial completed enrollment in 2010.
• The lowest of the three active doses was selected by the
Interim Analysis Committee.
• The p-value achieved was 0.0096.
• An NDA (Fulyzaq®, crofelemer) was submitted to FDA in
2011 and approved in December 2012
Take Home Messages for Pivotal Trials
• A Special Protocol Assessment (SPA) should be conducted.
• The Statistical Analysis Plan should be submitted to the
Agency before the trial commences, assuring that that
overall type-1 error is preserved.
• The entire adaptive strategy must be pre-specified in an
Interim Analysis (IA) Charter.
• An independent statistical center should oversee the IA in
accordance with this charter.
• Procedures must be instituted to assure rigorous
preservation of blinding.
Guidance for Industry: Adaptive Design Clinical Trials for Drugs
and Biologics , February 2010