Rotavirus vaccine example
In August 1998 the first rotavirus vaccine, RotaShield®, was licensed in the USA. Pre-licensure literature noted a suspicion of an increased risk of intussusception. After RotaShield® was licensed for routine use by the public (approximately one million children vaccinated within the first nine months post-licensure) the American vaccine safety surveillance, Vaccine Adverse Event Reporting System (VAERS), began to receive reports of intussusception following administration of the vaccine. About 100 (0.01%) of the one million children vaccinated developed intussusception,16 a potentially life-threatening bowel obstruction that occurs for unknown reasons in about one child per 10,000, regardless of whether or not they have received a vaccine.17 Because of the uncertainty about the relationship between RotaShield® and intussusception cases following vaccination, the manufacturer voluntarily took the product off the market in 1999.
This example demonstrates that even if no adverse event is observed in a trial of 10,000 vaccinees (as was the case of RotaShield®'s phase III clinical trial), one can only be reasonably certain that the real incidence of the adverse event is no higher than one in 3,333 vaccinees. Thus to be able to detect a risk of one adverse event per 10,000 vaccinees, a pre-licensure trial of at least 30,000 vaccinees and 30,000 controls is needed.14
Subsequent rotavirus vaccines were subjected to phase III trials that included at least 60,000 infants.18,19 While these trials were adequately powered to detect the problem with intussusception found following RotaShield®, in general, the cost of such large trials might limit the number of vaccine candidates that go through this process in the future.