Transverse magnetic resonance images at the L(4)-L(5) level of 26 male
s ranging in body mass index from 19.7 to 39.5 were digitized using a
computer video monitor with mouse-mediated digitization software. The
digitized musculature included the right and left psoas, rectus, abdom
inis, quadratus lumborum, oblique, and paraspinal muscle groups. From
the digitized profiles, cross-sectional areas, and anteroposterior and
mediolateral moments from the vertebral centroid were subsequently ca
lculated. Multiple regression formulae were derived to predict the mus
cle parameters from the most strongly related anthropometrics. Cross-s
ectional area and moment arms for lean and obese groups were also comp
ared. Significant regression formulae were derived for the cross-secti
onal area of the rectus abdominis and obliques, for mediolateral momen
t arms of the paraspinals, psoas, and obliques, and for anteroposterio
r moments of all four muscles. Comparison of muscle parameters for lea
n and obese subjects revealed minimal differences, though normalizatio
n of muscle cross-sectional areas by trunk area produced values which
were statistically smaller for all muscles of the obese relative to le
an subjects. Relevance-Identifying variations in trunk muscular geomet
ry across specific population groups will assist in further refining t
he existing morphometric data base used to estimate the forces experie
nced about the lower back. By identifying these variations, it may be
possible to observe factors which may predispose a specific population
to low back pain.