Multidimensional sleep clusters and cardiometabolic risk factors in adolescents: Results from the GINIplus and LISA studies

Abstract

BACKGROUND AND AIM Evidence of an association between sleep duration and body mass index is vast, but other sleep characteristics might also be relevant for cardiometabolic health. We aimed to identify sleep clusters based on multidimensional sleep characteristics, and to determine their associations with cardiometabolic risk markers in adolescents. METHODS A total of 1096 participants of the GINIplus and LISA birth cohorts, wearing accelerometers for one week during the 15-year follow-up, were included. K-means cluster analysis was performed based on twelve sleep characteristics including daily average and day-to-day variability, respectively, of total sleep time (TST), sleep efficiency, sleep-onset latency, sleep onset timing, time awake per hour after sleep onset, number of awakenings per hour. Cardiometabolic markers were dichotomized based on established cut-offs or sex-specific percentiles, including high fat mass index, prehypertension, high triglycerides, low high-density lipoprotein cholesterol, high C-reactive protein, and high homeostatic model assessment of insulin resistance (HOMA-IR). Cross-sectional, logistic regression models were run adjusting for potential confounders. RESULTS Five sleep clusters were identified: good sleep (n=391; average TST 7.7 hours), late but high-efficient sleep (n=224; 6.9 hours), irregular sleep (n=101; 6.9 hours), difficulty staying asleep (n=298; 6.8 hours), and difficulty falling asleep (n=82; 6.9 hours). Compared with "good sleep”, the "irregular sleep” cluster was associated with higher odds of high triglycerides (odds ratio (OR) = 2.43, 95% confidence interval (CI) =1.29-4.60). Moreover, the "difficulty falling asleep” cluster was associated with increased odds of high HOMA-IR (OR=3.09, 95%CI=1.25-7.65). The "late but high-efficient sleep” and "difficulty staying asleep” clusters were not significantly associated with cardiometabolic markers. CONCLUSIONS Clusters describing "difficulty falling asleep” and "irregular sleep” patterns were associated with high insulin resistance and high triglycerides, respectively, in adolescents. Improving sleep quality and quantity in different sleep dimensions might benefit adolescents’ cardiometabolic health. KEYWORDS Sleep, cardiometabolic markers, adolescents, cluster analysis, accelerometry