Longer Sleep Duration and Endothelial Cell Health among a Multi-Ethnic Sample of Adolescents

Abstract

Objective:

Adverse endothelial cell health, an early pathogenic process underlying atherosclerosis and cardiovascular disease, is evident in childhood and adolescence. Sleep duration, a modifiable cardiovascular health behavior, may be an important cardiovascular disease prevention target that may impact endothelial cell health. We examined the associations of longer sleep duration with endothelial cell injury among youth.

Methods:

In a multi-ethnic sample of 233 children (63.0% female, mean age= 13.9) we conducted multivariable linear regressions to test the cross-sectional association of sleep duration and circulating levels of endothelial cell-derived microparticles (EMPs), phenotypic for endothelial cell activation and apoptosis (CD62E+ EMPs, CD31+/CD42b− EMPs, and CD31+/Annexin V+ EMPs). Sleep duration and EMPs were both treated as continuous variables. Models were adjusted for age, sex, race, pubertal status, household economic resources, and waist circumference.

Results:

Overall, 69.2% had short sleep duration (< 8 hours of sleep per night). Longer sleep duration was significantly associated with lower levels of CD62E+ EMPs and CD31+/CD42b− EMPs. A 60-minute increase in sleep duration was associated with an 8.40 (95%CI: −205.20, −1.80, p=.046) decrease in CD62E+ EMPs and a 9.00 (95%CI: −153.60, −.9.60, p=.027) decrease in CD31+/CD42b− EMPs. Sleep duration was not associated with CD31+/Annexin V+ EMPs.

Conclusions:

Our results support the hypothesis that sleeping longer has beneficial effects on endothelial cell health during childhood. Primordial prevention efforts might incorporate sleep extension to offset cardiovascular risk in youth.

INTRODUCTION

Cardiovascular disease (CVD) is increasingly recognized as the clinical endpoint of pathological processes that unfold across the lifecourse, starting in childhood.¹ Population-level primary and primordial preventive approaches to curb CVD risk focus on the promotion of healthy behaviors in childhood and adolescence.¹ Key to CVD prevention is the identification of modifiable lifestyle risk factors that offset early disease progression. Sleep duration, a modifiable cardiovascular health behavior, may be one such target to reduce CVD risk across the lifecourse.²

In addition to serving as a microbial barrier, vascular endothelial cells play essential roles in maintaining vascular tone and counter-regulating thrombotic and inflammatory processes that contribute to the progression of atherosclerosis.³ Indeed, endothelial cell injury, which is well documented among adults, is an early pathogenic process in the development of atherosclerosis.³ There is also growing evidence that among children and adolescents, cardiovascular risk factors, including obesity, type 1 diabetes, and elevated blood pressure are associated with endothelial cell injury.^4–7 However, the relation between sleep duration and endothelial cell health among children and adolescents remains understudied. This is particularly important given that the American Academy of Pediatrics and the American Academy of Sleep Medicine recommend that adolescents sleep an average of 8−10 hours per night, and yet, most adolescents in the United Students sleep fewer than 8 hours nightly.^8–11

The purpose of this study was to examine the associations of longer sleep duration with circulating measures of endothelial cell injury. We hypothesized that longer sleep duration would be associated with less endothelial cell injury in youth.

Methods

Sample

Participants were 277 children from the Chicago, Illinois area. They were recruited through advertisements in public transit, schools, and community centers. To be eligible, participants had to be in 8^th grade, fluent in English, and in good health, defined as without a standing medication regimen, free of chronic medical/psychiatric illnesses, clear of infectious disease in the past 2 weeks. Northwestern University’s Institutional Review Board approved the protocol, and children and their parent/guardian consented to participation.

Endothelial Cell Injury

After injury, endothelial cells lose their integrity, progress to senescence, and detach into the circulation, leaving the underlying subendothelium exposed to leukocyte adhesion and migration, and lipid infiltration.³ Endothelial cell injury in humans can be determined by measuring circulating levels of endothelial cell-derived microparticles (EMPs), phospholipid rich, submicron particles released from the membranes of endothelial cells that have become activated or apoptotic. ^12–15 Evidence suggests that EMPs are not “innocent bystanders,” but are directly involved in atherosclerosis development. EMPs exacerbate inflammation via leukocyte activation and transendothelial migration, and are prothrombotic.^12–15 Thus, EMPs, which are a direct measure of endothelial cell injury, are implicated in atherosclerosis development.^12–15 In the current study, circulating EMPs, phenotypic for activated and apoptotic endothelial cells, were measured using flow cytometry.¹⁶

Following an overnight fast, children had antecubital blood drawn into sodium-citrate collection tubes. Platelet poor plasma (PPP) was isolated by centrifugation, and stained with monoclonal antibodies against CD31 (PE), CD42 (FITC), CD62E (PE), phosphatidylserine (FITC-conjugated Annexin V), and respective IgG isotype controls (FITC and PE). Data were acquired on a Guava 6HT2L flow cytometer (Millipore) and analyzed with FlowJo software. Activated endothelial cells were quantified by measuring CD62E+ EMPs, and apoptotic endothelial cells were quantified by measuring CD31+/CD42b− EMPs, and separately, CD31+/Annexin V+ EMPs.^12–14 Values are expressed as EMP counts/uL PPP.

Sleep Duration

For 96 hours following the blood draw, children wore a commercial activity monitor (Fitbit® Charge HR). Fitbit® generated nightly estimates of sleep duration have demonstrated high validity (97%) and specificity (91%) for detecting sleep duration against gold-standard polysomnography.¹⁷ We averaged values across the 4 nights of monitoring to derive an index of typical sleep duration.

Statistical Analysis

Standardized betas (β) and 95% confidence intervals (CI) for the EMPs outcomes were estimated from multivariable linear regression models that included sleep duration as the primary predictor, with age, sex, race, pubertal status, household economic resources (a composite reflecting income and savings), and waist circumference (inches) included as covariables. We included waist circumference as a covariate instead of body mass index (BMI) because for cardiometabolic health, visceral fat as reflected in waist to hip ratio is more important pathogenically than overall adiposity as reflected in BMI.¹⁸ We also conducted two sensitivity analyses, one that included an additional adjustment for physical activity, and another that included a separate adjustment for symptoms of depression and anxiety in children as measured by the Revised Child Anxiety and Depression Scales (RCADS-25).¹⁹

Results

Mean age (SD) was 13.9 (0.54) years, 63.0% were female, 35.7% were non-Latina/o White, 36.2% were non-Hispanic Black, and 31.5% were Latina/o (Table 1). The distribution of Tanner pubertal stage included: 36.2% early-middle puberty, 51.9% late puberty, and 11.9% post-pubertal. Mean (SD) nightly sleep duration was 450.7 (96.7) minutes or 7.51 (1.61) hours; 69.2% had short sleep duration (slept <8 hours). Of the 277 children enrolled in the original study, 44 were excluded due to missing data on either sleep patterns from non-compliance with FitBit monitoring (n=42) or EMP concentrations (n=2). Those with complete data did not differ from those with incomplete data with respect to age, sex, puberty, waist circumference, and household socioeconomic resources. However, adolescents with incomplete data were more likely to be racial/ethnic minorities than those with complete data. The final analytical sample size was 233.

Table 1.

Demographic Characteristics of the Adolescent Sample (N=235)

Characteristic	Mean (SD) or Number (%)
Age, years	13.9 (0.54)
Sex, female	148 (63.0%)
White, non-Latina/o	84 (35.7%)
Black, non-Latina/o	85 (36.2%)
Latina/o (any race)	74 (31.5%)
Tanner stage, early-middle puberty	85 (36.2%)
Tanner stage, late puberty	122 (51.9%)
Tanner stage, post-pubertal	28 (11.9%)
Waist circumference, cm	74.6 (12.6)
Sleep duration, minutes per day	450.7 (96.7)
Physically active, minutes per day	338.1 (96.8)
Activated EMPs, CD62+, per uL PPP	3927.8 (1532.9)
Apoptotic EMPs, CD31+, per uL PPP	889.8 (874.6)
Apoptotic EMPs, CD31/Annexin+, per uL PPP	78.8 (93.6)

Note: EMPs= Endothelial Cell-derived Microparticles

Univariable linear regressions showed that a 60-minute increase in sleep duration was significantly associated with lower circulating levels of CD31+/CD42b− EMPs (β= −9.60, 95%CI: −157.20, −18.60, p=0.01), but not with levels of CD62E+ EMPs (β= −7.20, 95%CI: −187.80, 7.8, p=0.07) or CD31+/Annexin V+ EMPs (β= −1.20, 95%CI: −8.40, 6.60, p= 0.82). In multivariable linear regressions (Table 2), longer sleep duration was significantly associated with lower circulating levels of CD62E+ EMPs and also CD31+/CD42b− EMPs. Furthermore, longer sleep duration was also associated with lower levels of CD31+/Annexin V+ EMPs, but this association was not statistically significant. A 60-minute increase in sleep duration was associated with a 8.40 (95%CI: −205.20, −1.80, p=.027) decrease in CD62E+ EMPs and a 9.00 (95%CI: −153.60, −.9.60, p=.046) decrease in CD31+/CD42b− EMPs. These associations remained significant in models further adjusted for actigraphy estimates of daily physical activity, and in separate models adjusted for symptoms of anxiety and depression in children.

Table 2.

Multivariable Regression Analysis of Sleep Duration on Endothelial Cell-derived Microparticles (EMPs) (N=235)

Variables	CD62E+ EMPs (counts/μL)		CD31+/Annexin V+ EMPs (counts/μL)		CD31+/CD42b− EMPs (counts/μL)
	βa	95 CI%	β	95 CI%	β	95 CI%
Sleep duration, per 60 minute increase	−8.40*	−205.20, −1.80	−2.40	−9.60, 5.40	−9.00*	−153.60, −9.60
Sex
Femaleb	4.20	−13296.00, 35267.40	0.06	−1658.40, 1969.20	3.00	−11997.00, 22382.40
Age, per 1 year increase	0.60	−16780.80, 20604.60	1.20	−1159.80, 1634.40	−1.80	−16222.20, 10258.20
Race
Underrepresented  Minorityc	−1.20	−26466.6, 20950.20	−0.12	−1797.60, 1742.4	3.60	−10186.20, 23366.40
Pubertal stage	1.80	−12886.20, 19545.6	−13.80**	−2991.00, −567.00	−4.80	−16985.40, 5982.60
Household economic resources, per 1 unit increase	3.00	−8851.20, 116115.40	5.40	−369.00, 1495.20	−1.20	−10026.00, 7645.20
Waist circumference, per 1 inch increase	1.20	−1639.20, 2349.00	−3.60	−218.40, 79.20	3.00	−912.60, 1903.80

Note:

^*p< 0.05;

^**p< 0.01.

Models adjust for sex, age, race, pubertal stage, household economic resources, and waist circumference. The analytic sample consisted of 235 adolescents with complete data; missing values were due to non-compliance with activity monitoring (N=42) and failed venipuncture (N=2).

^aβ refers to standardized Beta Coefficient

^bReference category is Male

^cReference category is White and Non-Latino adolescents; Underrepresented Minority refers to Black, Latina/o, Mixed, or Native American

^dHousehold economic resources is a composite score of average gross income and family savings

Discussion

In a racially and ethnically diverse sample of youth, we found 69.2% had short sleep duration (<8 hours of sleep per night), and that longer sleep duration was associated with lower circulating levels of EMPs phenotypic for endothelial cell activation and apoptosis. This association was independent of several potential explanatory factors including demographics, socioeconomic status, pubertal stage, waist circumference, and physical activity. These findings suggest the possibility that sleeping longer promotes positive endothelial cell health during childhood. These patterns also converge with evidence that among adolescents, shorter sleep is associated with other cardiometabolic risk factors.^2,20,21

Among this study’s limitations are its cross-sectional design, which precludes causal inferences, and the use of a validated commercial activity monitor to measure sleep.¹⁷ If replicated, future research should examine the mechanisms through which sleep duration affects endothelial cell health. The current study also did not conduct a comprehensive assessment of diet, including the consumption of caffeinated beverages; as such we cannot adjust for the effect of caffeine on the relationships observed. Furthermore, the current study did not assess endothelium-dependent vasodilation, a functional non-invasive measure of endothelial cell health.²² Assessment of endothelium-dependent vasodilation provides information on only one aspect of endothelial function (i.e. the ability to induce arterial vasodilation). In contrast, our study incorporated a validated cellular measure of endothelial injury, and was applied to a multi-ethnic youth sample. If these findings are confirmed in prospective studies, efforts at primordial prevention might incorporate sleep extension to offset cardiovascular risk in youth.

Overall, our results suggest that longer sleep duration may promote endothelial cell health among children and adolescents. Given the high prevalence of short sleep duration in this age group, and the increased cardiovascular health risks associated with short sleep duration,² preventive interventions that target sleep extension or increasing sleep duration among youth may confer large public health benefits across the lifecourse.

List of Abbreviations

CVD

cardiovascular disease

EMPs

endothelial cell-derived microparticles

PPP

platelet poor plasma

BMI

body mass index

RCADS-25

Revised Child Anxiety and Depression Scale – 25