Background: A non-inferiority (NI) design aims to demonstrate that the experimental arm is not worse than a comparator within a pre-specified margin. A margin of a risk difference of 10% or a Geometric Mean Titer (GMT) ratio from 1.5 to 2.0 is implicitly recommended for vaccine trials. This study explores which NI margins were used in vaccine studies and how they were determined. Methods: A systematic literature search was performed on February 21st, 2013. This resulted in 260 eligible articles, 148 articles were selected randomly; data was collected by the use of a data extraction form. This thesis had an explorative character and therefore descriptive statistics were used to show the determination of the NI margins and which NI margins were used. Results: From the selected articles, 48 were excluded because they did not met the predefined criteria. In 81/100 articles a NI margin on difference was determined. In 71% of reviewed articles a margin of 10% was used, 19% of the articles stated margins lower than 10% (range 3-7.5%), whereas in 10% of the articles margins larger than 10%(range 11.5-50%) were stated as the NI margin. In 55/100 articles a NI margin on the GMT ratio was stated. In 60% of the articles a margin of 0.67/1.5 was used, in 33% the stated margins were 0.5/2.0. A large amount (85%) of the examined articles did not discuss how the NI margin was determined. Only five articles referred to the FDA or EMA guidelines, eight articles referred to methodological articles and two articles referred to previous articles specific for the vaccine. In 82/100 articles the conclusion corresponded to the results, in fourteen the conclusion could not be checked because the effect estimates were not shown, in four articles the conclusion of the authors was not in line with the results. Conclusion: The majority of reviewed articles showed no determination of the NI margin. Not all articles showed all their estimates on NI. Reporting on design and results of NI vaccine trials could therefore be improved. Current guidelines on NI trials contain no explicit margin for vaccine trials. The choice of the NI margin is essential for the outcome of the NI trials, although determination is a great challenge. Regulatory offices might help with presenting guidelines which contain an explicit margin (for a specific vaccine/disease).

Abstract

Background: A non-inferiority (NI) design aims to demonstrate that the experimental arm is not worse than a comparator within a pre-specified margin. A margin of a risk difference of 10% or a Geometric Mean Titer (GMT) ratio from 1.5 to 2.0 is implicitly recommended for vaccine trials. This study explores which NI margins were used in vaccine studies and how they were determined.

Methods: A systematic literature search was performed on February 21st, 2013. This resulted in 260 eligible articles, 148 articles were selected randomly; data was collected by the use of a data extraction form. This thesis had an explorative character and therefore descriptive statistics were used to show the determination of the NI margins and which NI margins were used.

Results: From the selected articles, 48 were excluded because they did not met the predefined criteria. In 81/100 articles a NI margin on difference was determined. In 71% of reviewed articles a margin of 10% was used, 19% of the articles stated margins lower than 10% (range 3-7.5%), whereas in 10% of the articles margins larger than 10%(range 11.5-50%) were stated as the NI margin. In 55/100 articles a NI margin on the GMT ratio was stated. In 60% of the articles a margin of 0.67/1.5 was used, in 33% the stated margins were 0.5/2.0. A large amount (85%) of the examined articles did not discuss how the NI margin was determined. Only five articles referred to the FDA or EMA guidelines, eight articles referred to methodological articles and two articles referred to previous articles specific for the vaccine. In 82/100 articles the conclusion corresponded to the results, in fourteen the conclusion could not be checked because the effect estimates were not shown, in four articles the conclusion of the authors was not in line with the results.

Conclusion: The majority of reviewed articles showed no determination of the NI margin. Not all articles showed all their estimates on NI. Reporting on design and results of NI vaccine trials could therefore be improved. Current guidelines on NI trials contain no explicit margin for vaccine trials. The choice of the NI margin is essential for the outcome of the NI trials, although determination is a great challenge. Regulatory offices might help with presenting guidelines which contain an explicit margin (for a specific vaccine/disease).