Introduction: Circadian rhythms regulate key physiological processes, including sleep-wake. Alterations in the circadian rhythm have been associated with greater risk of metabolic diseases. Genome-wide association studies (GWAS) have revealed genes variants associated with sleep phenotypes. Hypothesis: Polygenic risk scores (PRS) for significant SNPs reported in GWAS of chronotype (C), sleep duration (SD), and short sleep (SS, <7/h) are associated with metabolic parameters. Methods: We carried out a longitudinal analysis in 6,814 adults (aged 45-84 y at baseline) from the Multi-Ethnic Study of Atherosclerosis (MESA). Glucose, insulin, insulin resistance (HOMA-IR), body mass index (BMI) and waist circumference (WC) were assessed at baseline (2000-2002) and 16-18 y follow-up (Exam 6, N=3,303). DNA was extracted from whole blood, and genotyping data were obtained by the Affymetrix 6.0 SNP array. PRS for the following phenotypes were constructed as weight sum of genoma-wide significant alleles: 1) chronotype [PRS-C]; 2) continuous SD [PRS-SD]; and 3) SS duration [PRS-SS]. Multivariable linear regression models adjusted by sociodemographic, obesity, smoking, Mediterranean diet, physical activity and antidiabetic drugs were performed to assess associations between PRS and study outcomes (fasting glucose and insulin, HOMA-IR, hemoglobin A1c, BMI, and WC) cross-sectionally and longitudinally. Results: At baseline, none of the PRS were associated with glucose, insulin or HOMA-IR (p>0.05). A significant positive association was detected for PRS-C with BMI and WC: greater genetically determined morningness was associated with lower BMI (β=-20.0±3.0; p<0.001) and smaller WC (β=-44.1±8.0; p<0.001). Similar results were observed with PRS-SD (BMI: β=-0.1±0.0; p=0.005; WC: β=-0.3±0.1; p=0.008) and PRS-SS (BMI: β=9.3±2.8; p<0.001; WC: β=19.8±7.4; p=0.007). In longitudinal analyses, we observed prospective associations between PRS-C with BMI (β=-43.5±12.2; p<0.001), WC (β=-116.0.1±29.1; p<0.001), and glucose (β=-204.7±90.0; p=0.024). PRS-SS was longitudinally associated with HOMA-IR (β=214.1±88.0; p=0.017) and a trend was observed for hemoglobin A1c (β=-6.6±3.4; p=0.053), but not with anthropometrics parameters. No associations were noted for PRS-SD with obesity and glycemic outcomes in longitudinal analyses. Conclusion: In the MESA population, PRS for chronotype, sleep duration, and short sleep are associated with anthropometric but not with glycemic measures in cross-sectional analyses. However, PRS for chronotype and short sleep was associated with anthropometric and glycemic parameters in longitudinal analyses, suggesting the importance of genetic predisposition for morningness and adequate sleep duration for metabolic health.
Read full abstract