PITTSBURGH, April 26, 2021 (GLOBE NEWSWIRE) -- Krystal Biotech Inc., (“Krystal”) (NASDAQ: KRYS), the leader in redosable gene therapies for rare diseases, today announced modifications to the Statistical Analysis Plan (SAP) in the ongoing Phase 3 study of B-VEC for the treatment of dystrophic epidermolysis bullosa (DEB) based on feedback from the U.S. Food and Drug Administration (FDA).
The GEM-3 trial is a randomized, double-blind, intra patient placebo-controlled multicenter study designed to evaluate the efficacy and safety of B-VEC for patients living with both recessive and dominant forms of dystrophic epidermolysis bullosa. The trial enrolled 31 subjects with DEB, aged 6 months or older at time of consent. Each subject will serve as their own control.
To account for paired binary data in the study design and the within-pair correlation in the GEM-3 study, Krystal will use the McNemar test for primary efficacy analysis. The primary outcome measure is complete wound healing determined by the Investigator in B-VEC treated wounds versus placebo. In the updated SAP, a positive response is defined as:
- Complete wound healing at Week 22 and Week 24; or
- Complete wound healing at Week 24 and Week 26.
Only wounds that achieve complete wound healing for at least 2 weeks are counted as a positive response.
“As the GEM-3 trial remains double blinded until database lock occurs in the fourth quarter of 2021, we had the flexibility to adjust the planned SAP with no impact on trial conduct or progress,” said Suma Krishnan, chief operating officer of Krystal Biotech. She added, “We believe that the updated SAP accounts for the within-wound-pair correlation and will better assess the potential benefit of B-VEC against placebo.”
With 90% power and a two-sided Type 1 error rate of 5%, and an assumed 20% drop out rate, 30 subjects (i.e., 30 wound pairs) were required for the GEM-3 study assuming a 50% difference in response rate between B-VEC and placebo treated wounds. Because the calculation assumes no correlation within subjects, the estimate is conservative. Any positive correlation will cause an increase in power.
Retrospective analysis of Phase 2 study looking at complete wound healing at Week 8 and Week 10 or Week 10 and Week 12 resulted in 78.6% response rate for B-VEC treated wounds versus 0% response rate for placebo wound, with a p