16 participants and 20 controls is small enough that you might miss a subset of people that do have reactions to gluten. My associate has huge migraines when he eats gluten, but does not have celiac disease.

Looking forward to more research here (and at least they said “casts doubt” rather than “disproven”)