Underreporting of energy intake in 1 to 18 year old German children and adolescents.

It is generally accepted that self-reported food intakes underestimate habitual energy intake (underreporting). Underreporting is often addressed by computing the ratio of measured energy intake to predicted basal metabolic rate (EI:BMR). We used this ratio to study differences between not plausible records (NPR) and plausible records (PR) according to recalculated cut-off values for EI:BMR ratios (< 0.97 to 1.07; age- and sex-dependent) in cross-sectional data of 1,032 3d weighed diet records of 1 to 18 year old children and adolescents. Underreporting (in 5% of total subjects) was age and sex dependent: about 1% in the 1 to 5 year old children, 2% (3%) in the 6 to 13 year old males (females) and 12% (20%) in the adolescent males (females), respectively. To analyse differences between subgroups with PR vs. NPR we therefore concentrated on the 14 to 18 year olds. Male (female) subjects with NPR vs. PR had a 40% lower total EI: 7.4 MJ/d (5.3 MJ/d) vs. 11.5 MJ/d (8.0 MJ/d), respectively. In both sexes with NPR vs. PR, EI per meal was lower. Females with NPR vs. PR had a higher body mass index (kg:m2), recorded fewer meals per day, and had a shorter time span between the first and last meal per day. Furthermore, females with NPR vs. PR had higher intakes per MJ of water, protein, fiber, sodium, iron, niacin, zinc, and protein in percent of total EI, but a lower intake of added sugars per MJ. Males with NPR vs. PR answered significantly more often that they usually eat more, had a higher water intake per MJ, and a higher Ca:P. The observed differences between groups with PR and NPR indicate different food habits or dietary recording behavior. Therefore, validity in dietary studies cannot be achieved by simply excluding underreporters.