Social Gradients and Physical Activity Trends in an Obesogenic Dietary Pattern: Cross-Sectional Analysis of the UK National Diet and Nutrition Survey 2008-2014

Abstract

An energy-dense, high-fat, low-fibre dietary pattern has been prospectively associated with the development of obesity in childhood but is population-specific, which limits translating the pattern into interventions. We explored the generalisability and correlates of this obesogenic dietary pattern in the UK National Diet and Nutrition Survey (NDNS) for the first time. Data came from participants (n = 4636 children and n = 4738 adults) with 4-day food diaries in NDNS 2008-2014. Reduced rank regression was applied to 51 food groups to explain variation in energy density, fibre and fat intake. Consistency of the pattern in population subgroups (according to sex, age, occupation and income) was compared with the whole sample pattern using coefficients of congruence (COC). Pattern correlates (sociodemographic, survey year, physical activity and eating related behaviours) were explored using multiple linear regression. Food group loadings were similar to the previously identified obesogenic dietary pattern and were generalisable across all sub-groups (COC: 0.93-0.99). An obesogenic diet was associated with eating takeaways, being omnivorous, a manual household occupation and lower household income in both adults and children (p < 0.0001). Dieting for weight loss, being older, more physically active and less sedentary was associated with a less obesogenic diet among adults (p < 0.0001). Future experimental studies should investigate if changes in this obesogenic pattern could be used to monitor the effectiveness of obesity prevention policies or develop personalised interventions.

Keywords: National Diet and Nutrition Survey; dietary patterns; obesogenic; reduced rank regression.

Figure 1

Dietary pattern loadings for each food group in children and adults.

Figure 2

Trends in the obesogenic dietary pattern score among children (<18 years) by (a) sex; (b) age; (c) weight status; (d) energy reporting status; (e) occupation; (f) household income; and (g) survey year. Figures present marginal means and tests for trend with survey-weighted linear regression adjusted for age, gender and daily energy intake.

Figure 3

Trends in the obesogenic dietary pattern score among adults (>18 years) by (a) sex; (b) age; (c) weight status; (d) energy reporting status; (e) occupation; (f) household income; and (g) survey year. Figures present marginal means and tests for trend with survey-weighted linear regression adjusted for age, gender and daily energy intake.

Figure 4

Trends in the obesogenic dietary pattern score among adults (>18 years n = 3549 panel a and b, n = 3790 panel (c) and children (aged 11–16 years, n = 861 panel d) by (a) media time; (b) computer time; (c) time spent in moderate to vigorous physical activity (MVPA); (d) objectively measured physical activity in counts per minute. Figures present marginal means and tests for trend with survey-weighted linear regression adjusted for age, gender and daily energy intake. Media time, computer time and MVPA have been assessed with the RPAQ questionnaire adults aged >18 years. Quartiles of these variables are presented in the current figure.