OA Epidemiology

The Use of Non-Inferiority Margins in Vaccine Trials

Proceedings of the 2013 annual meeting of the Netherlands Epidemiology Society

Volume 1 Issue S1 Abstract 4

 

R. Donken, Epidemiology and Surveillance Unit, Centre for Infectious Disease Control, National Institute for Public Health and the Environment (RIVM), Bilthoven
G. Berbers, Immunology  of Infectious Diseases and Vaccines, Centre for Infectious Disease Control, National Institute for Public Health and the Environment (RIVM), Bilthoven
N. Rots, Immunology of Infectious Diseases and Vaccines, Centre for Infectious Disease Control, National Institute for Public Health and the Environment (RIVM), Bilthoven
M.J. Knol, Epidemiology and Surveillance Unit, Centre for Infectious Disease Control, National Institute for Public Health and the Environment (RIVM), Bilthoven

Background
A non-inferiority (NI) trial investigates whether an intervention  is not worse than its comparator, within a prespecified NI-margin. A difference of 10% or a Geometric Mean Titer (GMT)-ratio of 1.5-2.0 is stated as an appropriate NI-margin for vaccine trials, but no strict guidance exists. The choice of the NI-margin is essential for the interpretation of a trial. This study explores how NI-margins are determined and which NI-margins are used in vaccine trials.

Methods
A random set of 50 NI vaccine trials was included. We extracted from the papers: NI- margins, their determination, effect estimates and the conclusion.

Results
The NI-margins were based on difference in response (n=22), based on the GMT-ratio (n=9), or both (n=19). Different margins were used: 10% margin in 29 studies, <10% in five studies,
>10% (range 11.5%-50%) in four studies. Three papers mentioned different NI-margins on the primary endpoint for different diseases. Most studies (n=18) used a GMT NI-margin of
1.5, 2.0 was used in nine, and 4.0 in one study. None of the papers explicitly discussed the
NI-margin determination, although six referred to previous publications, and four referred to the FDA/EMA guidelines. In three studies, the conclusion of the authors was not in line with their results.

Conclusion
The majority of papers used NI-margins in line with the consensus. In general the method of determining the NI-margin was unclear while this is of great importance for the interpretation of the results and the following conclusion. Reporting on determination of NI- margins could be improved.

Published: 06 Jun, 2013

 
Licensee OA Publishing London 2013. Creative Commons Attribution License (CC-BY)