0848 Cross-Sectional and Longitudinal Associations Between Water Mineral Content and Sleep Health in UK Biobank

Abstract Introduction Sleep duration and quality are critical for health, yet the impact of environmental determinants on sleep, particularly drinking water quality, is underexplored. Using UK Biobank data, we aim to investigate the cross-sectional and longitudinal associations between water mineral content and self-reported sleep. Methods We analyzed three waves of UK Biobank data (T1: 2006–2010; T2: 2012–2013; T3: 2014–2018). Participants with complete data on all covariates were included, and sample sizes varied slightly across models due to variable availability. Sleep duration (hours/day), insomnia (yes/no), and daytime sleepiness (yes/no) were self-reported. Water mineral indicators included hardness (soft/moderately hard/hard), calcium carbonate (CaCO3), calcium, and magnesium. Cross-sectional analyses (T1) used linear regression for sleep duration and logistic regression for insomnia and daytime sleepiness. Longitudinal analyses used generalized linear mixed-effects models with participant-level random intercepts in participants with ≥2 time points. Models adjusted for demographic, index of multiple deprivation, employment, income, chronic conditions, lifestyle, food intake of magnesium and calcium, and daily water intake (number of glasses). Results Cross-sectional analyses (n = 24,135–25,998; mean age 55.6 ± 8.3 years; 50.3% female) showed that each SD (standard deviation) increase in water calcium (SD = 42.0 mg/L) was significantly associated with a 0.0126-hour (≈0.76-minute) shorter sleep duration, while hardness categories, CaCO3, and magnesium was not significant. Soft water was associated with higher odds of insomnia compared to hard water (OR = 1.12; 95% CI [1.05 to 1.20]), while each SD increase in water calcium was associated with lower odds of insomnia (OR = 0.92; 95% CI: 0.92–0.96). Water mineral concentrations showed no significant associations with daytime sleepiness cross sectionally. Longitudinally (n = 24,011–25,881; mean age 55.6 ± 8.3 years; 50.3% female), moderately hard water was significantly associated with greater rate of increase in sleep duration over time compared with hard water (β = 0.029). No water mineral exposures were significantly linked with insomnia or daytime sleepiness changes over time. Conclusion Water mineral content showed modest associations with sleep outcomes in UK Biobank participants. These small effects warrant replication in specific populations and mechanistic studies before informing public health interventions or policy. Support (if any) Johns Hopkins University Catalyst Award