Letters Section Editor: Robert M. Golub,
MD, Senior Editor.
In Reply: Although the terms “equivalence”
and “noninferiority” are not strictly the same, we did use them
interchangeably, as is commonly done.1 This
study was designed as a noninferiority study with a prespecified noninferiority
margin of 10% chosen as in the “Methods” section.
The 2-sided 95% CI of the difference between 2 trial arms is an appropriate
and traditional method for assessing equivalence of 2 treatments.1,2 The same procedure is often advocated
for assessing noninferiority of a new drug (or regimen) compared with a reference
drug,1 even though in noninferiority studies
emphasis is on the lower bound of the 95% CI. Indeed, the use of a 2-sided
95% CI (equivalent to a 1-sided 97.5% CI) is advocated by US and European
regulatory agencies3- 6 for
noninferiority studies. While academic statisticians often use 1-sided 95%
CIs to assess noninferiority, we chose the more conservative 2-sided approach,
as suggested by these regulatory authorities.
Gallant JE, Staszewski S, Pozniak AL, DeJesus E, Suleiman JMAH, Lu B, Toole JJ, Cheng AK. Tenofovir, Equivalence, and Noninferiority—Reply. JAMA. 2004;292(16):1951-1952. doi:10.1001/jama.292.16.1951-b