Self-reported and measured anthropometric variables in association with cardiometabolic markers: a Danish cohort study

Abstract

AbstractSelf-reported measures of height and weight are widely used in epidemiological studies. However, misreporting may bias estimates of associations between anthropometry and health outcomes. Further, few validation studies have compared self-reported and measured waist circumference (WC). This study aimed to quantify the agreement between self-reported and measured height, weight, body mass index (BMI), WC, and waist-to-height ratio (WHtR), and to investigate associations of these anthropometric measures with cardiometabolic biomarkers.Self-reported and measured anthropometric variables, as well as cardiometabolic biomarkers, were obtained from participants aged above 18 years at recruitment into the Diet, Cancer, and Health-Next Generation Cohort in 2015-19 (N=39,514). Pearson correlations (r) and Lin’s concordance correlations were applied to evaluate misreporting. Misreporting by age, sex and smoking status was investigated in linear regression models. Multivariable regression models and Receiver Operating Characteristic analyses assessed associations of self-reported and measured anthropometry with cardiometabolic biomarkers. Self-reported height was overreported by 1.07 cm, and weight was underreported by 0.32 kg on average. Self-reported BMI and WC were 0.42 kg/m2and 0.2 cm lower than measured, respectively. Self-reported and measured height, weight, BMI, WC and WtHR were strongly correlated (r=0.98, 0.99, 0.98, 0.88, 0.86, respectively). Age, sex, smoking, and BMI contributed to misreporting of all anthropometric measures. Associations between self-reported or measured anthropometric measures and cardiometabolic biomarkers were similar in direction and strength. Concordance between self-reported and measured anthropometric measures, including WC, was very high. Self-reported anthropometric measures were reliable when estimating associations with cardiometabolic biomarkers.