Path analysis using data derived from genetically randomized inbred mouse strains reveals biological controls affecting bone growth and adult fragility

Abstract

Listen

Translate article

bone development in children. However, quantifying this relationship in observational studies is often hindered by the measurement error inherent in the commonly used PA measurement method of subjective self-report. In this crosssectional analysis, using an objective and a subjective measure of PA, we examined associations between PA and bone mineral content (BMC) in 449 children (mean age 11.2 years, range 10.5 to 12.4 years). Design: PA was measured using 5-day accelerometry readings (Actigraph 7164) and the Physical Activity Questionnaire for Children (PAQ-C) (from the Saskatchewan Pediatric Bone Mineral Accrual Study). Total body and lumbar spine bone mineral content (BMC) were measured using a Hologic 4500 dual energy X-ray absorptiometry (DXA). Body mass (kg), height (cm), and biological age (pre-or post-peak height velocity) were considered as co-variates in partial correlation analyses and multiple linear regression models. Significance was set at (pb0.05). Results: Accelerometry-determined moderate to vigorous (MVPA) and PAQ-C scores were higher in boys when compared to girls. The Spearman rank correlation coefficient between MVPA and PAQ-Cwas 0.38 for boys and 0.25 for girls (pb0.05). MVPA was significantly associated with total body BMC (r=0.33) and spine BMC (r=0.19) in boys (partial Spearman rank correlation coefficients controlling for weight, height, and maturity). When PAQ-C was substituted for MVPA, the partial correlations were 0.22 for total body BMC and 0.19 for spine BMC. These associations were also statistically significant. When MVPA and PAQ-C were considered together in multiple linear regression models for total body BMC and spine BMC, both entered, although MVPA explained slightly more of the variability in each model. In girls, MVPA was significantly associated with total body BMC (r=0.17, partial Spearman rank correlation coefficients controlling for weight, height, and maturity). MVPA was not associated with spine BMC. PAQ-C was not significantly associated with spine nor total body BMC. Conclusions: Findings indicate that statistically significant and, perhaps important, associations between PA and BMC during late childhood are more likely to be detected with an objective PA measurement method. Self-report PA may be a more sensitive measure of PA for boys than for girls.