Differences in the relationship of weight to height, and thus the meaning of BMI, according to age, sex, and birth year cohort

Bibliographic Data

ID13922389
AuthorsWilliam Johnson (0000-0002-0347-4354, Loughborough University, corresponding author), Tom Norris (0000-0002-9140-0776, Loughborough University), David Bann (0000-0002-6454-626X, Centre for Longitudinal Studies, UCL Institute of Education, London, UK;), Noël Cameron (0000-0001-8316-047X, Loughborough University), J Wells (0000-0003-0411-8025, UCL Great Ormond Street Institute of Child Health, London, UK;), Tim J Cole (0000-0001-5711-8200, UCL Great Ormond Street Institute of Child Health, London, UK;), Roger Hardy (0000-0001-9949-0799, CLOSER, UCL Institute of Education, University College London, London, UK)
Year2020
Volume47
Issue2
Pages199-207
Publication date2020-02-17
Open AccessYes
TypeARTICLE
VenueAnnals of Human Biology (JOURNAL)
Journal identifiersISSN: 0301-4460 • E-ISSN: 1464-5033
PublisherInforma (PUBLISHER • GB)
DOI10.1080/03014460.2020.1737731
PMID32429756
OpenAlexW3013564732
LanguageEN
Citations received6
References cited62

Background: Weight can be adjusted for height using the Benn parameter (kg/m B ), where B is the power that minimises the correlation with height. Aim: To investigate how the Benn parameter changes across age (10-65 years) and time (1956-2015) and differs between sexes. Subjects and methods: The sample comprised 49,717 individuals born in 1946, 1958, 1970 or 2001. Cross-sectional estimates of the Benn parameter were produced and cohort differences at ages 10/11 and 42/43 years were examined using linear regression. Multilevel modelling was used to develop trajectories showing how the Benn parameter changed over age from childhood to mid-adulthood in the three older cohorts. Results: The Benn parameter was closest to 2 in childhood but consistently lower across adulthood, particularly in females and the most recent cohort. At ages 10/11 years, the Benn parameter was greater than 3 in both sexes in the 2001 cohort but between 2.2 and 2.7 in the three older cohorts. This difference was estimated to be +0.67 (0.53, 0.81) in males and +0.53 (0.38, 0.68) in females, compared to the 1946 cohort, and was driven by a much higher weight SD in the 2001 cohort. Conversely, at ages 42/43 years, the Benn parameter was lowest in the 1970 cohort due to a slightly lower weight-height correlation. This difference was estimated to be -0.12 (-0.34, 0.10) in males and -0.15 (-0.42, 0.13) in females, compared to the 1946 cohort. Conclusions: Changes over time in the obesogenic environment appear to have firstly reduced the Benn parameter due to a lowering of the weight-height correlation but secondly and more drastically increased the Benn parameter due to increasing weight variation

Cohort · Cohort study · Linear regression · Statistics · Birth, Development, and Health · Demography · Mathematics · Medicine · Nutritional Studies and Diet · Obesity, Physical Activity, Diet · Internal Medicine

Tools

Open DOIOpen Access