Lifestyles associated to sleep duration

Abstract

Introduction Optimal sleep duration varies significantly between individuals, age, gender, and world region. The value established for short sleep duration (SSD) in adults is £ 5 h and long sleep duration (LSD) is 39 h. The aim of this study was to study potential associations between SSD and LSD with sociodemographic variables and lifestyles. Methods Cross-sectional evaluation of the third wave of follow-up of the EpiDoC Cohort that enrolled in 2011 10,661 adults living in private residences in Portugal, (i.e., EpiDoC3 study, population interviewed 5653). Participants from EpiDoc3 who answer the question regarding hours of sleep/day (n = 5436) compose the current analysis. Data collection was performed Sept 2015–July 2016, by telephone. The EpiDoC cohort measurements were sociodemographic, socioeconomic characteristics and socioeconomic variables. The lifestyle habits evaluated were alcohol intake, smoking habits, physical activity, frequency of watching television and of using a computer. Height and weight were collected. For the dietary intake, the number of meals/day and the adherence to the Mediterranean diet was assessed by the PREDIMED score; for the daily total sleep time three categories were established: SSD (£ 5 h), normal sleep duration (NSD) (6 h–8 h) and LSD (9 h). To verify the representativeness of the sample according to the Portuguese population, we first compared the participants and non-participants of the EpiDoC3 study with respect to their sociodemographic, socioeconomic, and health status characteristics. Extrapolation weights were computed and used in the subsequent statistical analysis. These were obtained by calibrating the extrapolation weights originally designed for the EpiDoC1 study sample. Baseline characteristics of the study cohort were described according to SSD, NSD and LSD. Absolute frequencies and weighted proportions were used to summarize categorical variables. Two logistic regression models were performed in order to find lifestyle predictors for SSD and LSD. All analyses were performed using STATAICv12. Results Prevalence values for SSD and LSD were 24.6% and 7.5%. The region with higher prevalence for SSD was Algarve (SSD 26.5%, LSD 6.1%), while Azores have the lowest value (16.4%). Furthermore, the higher prevalence NSD was found in Azores (80.5%). The higher prevalence values for SSD and LSD were both found in females 22.0% and 2.1%. Variables associated to a higher risk of SSD were having 65 years (OR = 1.92; CI = 1.48, 2.50); being female (OR = 1.26; CI = 1.00–1.58); being overweight (OR = 2.25; CI = 1.03–4.87) or obese (OR = 2.74; CI = 1.24–6.02) and having 0–4 years of education (OR = 1.95; CI = 1.32–2.86); protective variables to SSD were to watch TV 3–4 h/day (OR = 0.45; CI = 0.22–0.90) and 5 h/day (OR = 0.38; CI = 0.19–0.80); to use computer 3–4 h/day (OR = 0.50; CI = 0.30–0.84) and 5 h/day (OR = 0.56; CI = 0.37–0.83); have 2 meals/day (OR = 0.46; CI = 0.22–0.98), 3 meals/day (OR = 0.36; CI = 0.17–0.78) and 4 meals/day (OR = 0.33; CI = 0.15–0.73). Variables associated to a higher risk of LSD were, having 65 years (OR = 1.94; CI = 1.41–2.67); to have 10–12 (OR = 2.03; CI = 1.15–3.56) and 0–4 (OR = 2.28; CI = 1.21–4.28) years of education and to live in Azores (OR = 1.69; CI = 1.14–2.52); protective variables for LSD were, to drink alcohol daily (OR = 0.63; CI = 0.43–0.92); to watch TV £2 h/day (OR = 0.39; CI = 0.18–0.88) and 3–4 h/day (OR = 0.38; CI = 0.16–0.88); and to use computer £2 h/day (OR = 0.52; CI = 0.34–0.80) and 5 h/day (OR = 0.37; CI = 0.18–0.77). Conclusions SSD is highly prevalent in the Portuguese population when compared to other population studies. Lower educational level and to have 65 years are risk factors for SSD and LSD. EpiDoc3 study emphasizes the burden of SSD and LSD in Portugal and the need to increase sleep awareness. Being SSD and LSD predictors of mortality and chronic diseases, these numbers highlight the magnitude of this problem to Portugal.