Abstract B2: Associations between self-reported versus objectively measured physical activity, biomarkers, and obesity in adolescents and adults

Abstract

Abstract Epidemiological studies, based entirely on self-reported levels of physical activity (PA), indicate that PA reduces cancer risk at several sites. However, self reported PA is subject to the cognitive challenges of estimating its duration and intensity, and potential recall bias. Objective measurement of PA using accelerometers is an alternative to self report, even in large population studies of both children and adults Data for adults aged 20+ years (N=4417) with self-reported PA and 4 or more days of accelerometer data were obtained from the 2003–2006 National Health and Nutritional Examination Survey (NHANES). NHANES is a nationally representative sample of the US civilian, non-institutionalized population. Outcome measures were blood pressure, body mass index (BMI), waist circumference, triceps and subscapular skinfolds, cholesterol, triglyceride, c-reactive protein, homocysteine, and indices of insulin resistance and hyperinsulinaemia. Data for adolescents, (N = 723), were obtained from two cohorts in the Twin Cities, Minnesota. PA was measured using two self report instruments, the Modified Activity Questionnaire (MAQ) assessing past week moderate and vigorous PA (MVPA), the 3-Day Physical Activity Recall (3DPAR), and accelerometer data. Biomarkers included systolic and diastolic blood pressure (SBP, DBP), lipids (total cholesterol, LDL, HDL, % body fat (BF) and BMI. For both adolescents and adults, accelerometer data was summarized as total minutes of (MVPA). In adults, objectively-measured MVPA displayed stronger associations with most biomarkers compared with self-reported MVPA, even after adjusting for sociodemographic and health factors (Adj Wald F's = 5.1 to 55.4; p's < 0.05 to 0.0001). Adjustment for objectively measured PA eliminated some but not all associations between self-reported PA and biomarkers. Both measures of MVPA were independently associated with skinfold anthropometric measures, HDL, and c-reactive protein. In adolescents, after controlling for puberty, age and gender, all three PA measures had a strong, negative association with %BF (MAQ β = −1.77, p<0.001; 3DPAR β = −1.43, p<0.001; accelerometer β = −1.41, p<0.003). The MAQ and accelerometry were negatively associated with BMI (β = −3.17, p=0.003; β =−2.97, p =0.01), and the MAQ was negatively associated with DBP (β = −1.30, p=0.01). None of the PA measures were significantly associated with SBP or lipids. These results suggest that use of self-reported physical activity measures could attenuate associations between PA and several biomarkers related to energy balance in adults, but this was not observed for a few biomarkers in adolescents. Considerable care should be taken in the choice of PA measure for any study, but especially where the effects of PA are manifest all or in part via its impact on biological variables such as serum biomarkers or anthropometric characteristics. Citation Information: Cancer Prev Res 2010;3(1 Suppl):B2.