A Case for Using Relative Rather Than Absolute Noninferiority Margins in Clinical Trials

oninferiority trials are conducted to demonstrate that one therapy is not associated with an unacceptably high risk compared to another therapy.The maximum acceptable excess risk is defined in noninferiority trials by a "noninferiority margin," and if the difference in risk between the therapy of interest and the comparator therapy does not exceed the noninferiority margin, then the new therapy is considered safe, or "as good as" the comparator therapy. 1,2e noninferiority margin can be expressed as an absolute or relative risk difference. 3Absolute rather than relative noninferiority margins are frequently used in cardiovascular trials.Unfortunately, absolute noninferiority margins are scientifically less robust than relative noninferiority margins and can lead to inaccurate conclusions regarding the safety of new therapies. 4This paper will attempt to describe why the issues associated with absolute margins should preclude them from being used, after first attempting to debunk 2 common arguments used in favor of absolute vs relative noninferiority margins in clinical trials.DEBUNKING OF 2 COMMON ARGUMENTS IN FAVOR OF ABSOLUTE NONINFERIORITY MARGINS #1.Absolute risks are more interpretable: A common