LONGITUDINAL AND CROSS-SECTIONAL MEASUREMENT OF INDICATORS OF CHANGING
BODY COMPOSITION: RESULTS COMPARED.
William A. Stini, Ph.D.*
University of Arizona, Tucson AZ 85721.
A mixed-longitudinal study of changes in bone density and indicators of body composition was
begun in 1982. Along with direct measurement of changes in the density of cortical bone and
anthropometric measurement of indicators of body composition changes, the study was designed to gather
information concerning associations between the rate of bone loss and changes in body composition.
Bone mineral density (BMD) at the distal one-third site of the left radius was measured using
single-beam photon absorptiometry from 1982 through 1998. In 1998 and 1999, a subsample of the
longitudinal study population was scanned using both single photon absorptiometry and whole-body dual-
energy X-ray (DEXA). The purpose of this additional procedure was to assess the validity of SPA results as
an indicator of cortical bone loss throughout the body.
Stature, weight, and forearm measurements were taken preceding each bone scan. Body Mass
Index (BMI) was calculated using the formula BMI = Weight (kg) / Stature (m2). Estimates of body
composition using bioelectric impedance (Valhalla) were also conducted for a subsample.
Questionnaires elicited information concerning nutritional intakes, medications, health status,
exercise patterns and history of bone fractures. Self-reports of age at menarche, menopause, and births of
children were collected from women. A subsample also completed a more detailed food frequency survey
and a 72-hour dietary recall interview.
Annual blood samples were drawn from the same subsample that was measured by bioelectric
impedance and completed food frequency surveys. Values for plasma glucose, creatinine, blood urea
nitrogen, alkaline phosphatase and a number of minerals including calcium, magnesium, iron, and copper
were determined, and the magnitude and direction of changes tracked.
The total population sampled was 5475 individuals. Of this total, 4121 were women and 1354
were men. Of this total, 126 women and 47 men participated for 10 or more consecutive years. The
subsample from which blood samples and bioelectric impedance data were collected was made up of 277
men and 174 women, all scanned in at least 5 consecutive years.
Both longitudinal and cross-sectional analyses revealed a clear pattern of BMD decrease in both
sexes. The onset of bone loss occurred earlier in women and resulted in clinically significant osteopenia
much earlier. However, the rate of bone loss by men actually exceeded that of women beyond age 80. All
measures of bone loss followed a curvilinear pattern in both sexes with the best-fit equation being: y = a -
bx2 . A strong association between BMD and BMI was seen in both sexes. This association was especially
pronounced at the more extreme BMI values, and most pronounced in males beyond the age of 80.
The pattern of bone loss seen in males of advanced age appears to be closely tied to the loss of
lean body mass detected in this group through body composition analysis.
Problems or questions regarding this site should be directed to
webmaster@americanaging.org