Normalization influences knee abduction moment results: Could it influence ACL-injury research, too?

Abstract

Normalization of joint moments to reduce anthropometric influences prior to making group comparisons is a widely-accepted practice. However, a seminal prospective study reported greater non-normalized knee abduction moment (KAM) in nine females who subsequently sustained an ACL injury. It is not clear if this finding may have been influenced by the fact that the ACL-injured females were on average 3.6cm taller and 2.4kg heavier than uninjured females. Cross-sectional. Peak KAM was identified in thirty-six females completing jump landings. A custom software program randomly divided participants into two groups that were compared on: (1) non-normalized KAM, (2) KAM normalized to body mass, and (3) KAM normalized to body height times weight a total of 500,000 times and the results categorically coded for statistical significance (α≤0.05). For the 10,591 iterations in which one group was 3-4cm taller and 2-3kg heavier, the agreement between results obtained using non-normalized versus normalized data were assessed using non-parametric analyses. Despite moderate-strong agreement between the results obtained using non-normalized and normalized data (Κ=0.614-0.744), a significant effect of normalization on the interpretation of group differences in peak KAM was identified (p<0.001). In 30.4-41.9% of the cases in which non-normalized KAM was deemed significantly different between groups, no group differences were identified when using normalized KAM. While it is unlikely the magnitude of the difference in non-normalized KAM identified prospectively in ACL-injured females was attributable solely to anthropometric differences, caution should be exercised when evaluating research findings reporting non-normalized KAM.