Poor Sleep Health Traits Influence Liking of Sweet Foods and Sugary Food Intake: A UK Biobank Study
Hui, P. S.; Touw, C. D.; Bhutani, S.; Hwang, L.-D.
Show abstract
Poor sleep is linked to consumption of sugary foods/beverages and high neural responsivity to palatable food cues. Yet, whether hedonic liking for sweet taste explains these associations remains unclear. We examined cross-sectional associations of five sleep traits (chronotype, sleep duration, insomnia frequency, snoring, daytime dozing) and a composite sleep score with sweet food liking, and total and free sugar intake in 76,734 UK Biobank participants (39-72 years, 56.3% female). Models adjusted for age, sex, ethnicity, socioeconomic deprivation, and body mass index (Bonferroni-corrected =0.0025). Evening chronotype, more frequent insomnia and daytime dozing, and lower composite sleep score were associated with higher sweet food liking. Associations with intake were stronger for free than total sugar. Evening chronotype was associated with higher free sugar intake (g/day: {beta}=1.523, 1.309-1.737; g/1000 kcal: {beta}=0.450, 0.361-0.538), and daytime dozing showed a dose-response (dozing often vs never/rarely: g/day {beta}=6.307, 4.631-7.983). Snoring was associated with higher absolute (but not energy-adjusted) free sugar intake. A healthier sleep score was associated with lower free sugar intake (g/day {beta}=-2.193 [-2.464 to -1.922]; g/1000 kcal {beta}=-0.691 [-0.804 to -0.579]) but higher energy-adjusted total sugar intake ({beta}=0.633 [0.485-0.781]). Mediation analyses indicated sweet liking accounted for 15%-91% of several sleep trait and free sugar intake associations (indirect effects p<0.001). Poorer sleep health, particularly evening chronotype and daytime sleepiness, was associated with greater sweet liking and higher free sugar intake, with sweet liking partially mediating associations between sleep traits and sugar consumption. Sweet-taste liking may represent an underexamined pathway linking sleep/circadian disruption to free sugar intake.
Matching journals
The top 7 journals account for 50% of the predicted probability mass.