Sleep Extension Improves Sleep Health but Not Insulin Sensitivity in People With Overweight or Obesity Who Maintain Habitual Short Sleep Schedules

Abstract

OBJECTIVE

To determine whether extending sleep improves insulin sensitivity in people with overweight or obesity, insulin resistance, and habitual short sleep schedules (<7 h/night).

RESEARCH DESIGN AND METHODS

Participants were randomized to habitual sleep (HS) (n = 15) or extended sleep (ES) (n = 14) for ∼6 weeks. Multiorgan (whole-body [primarily muscle], hepatic, and adipose tissue) insulin sensitivity (assessed by the hyperinsulinemic-euglycemic clamp procedure with tracer infusions) and glycemic control (assessed by 24-h serial plasma glucose and insulin concentrations during wakefulness) were determined.

RESULTS

Time in bed and sleep duration increased more in the ES group (1.3 ± 0.6 and 1.1 ± 0.5 h/night) than in the HS group (0.3 ± 0.8 and 0.0 ± 0.4 h/night). Day-to-day variability in sleep and subjective measures of sleep health improved more in the ES than HS group, without differences in multiorgan insulin sensitivity or glycemic control between groups.

CONCLUSIONS

Extending sleep by ∼1 h/night for ∼6 weeks in people with overweight or obesity and short sleep schedules improves sleep health but not insulin sensitivity or glycemic control.

Graphical Abstract

Introduction

Epidemiological studies have shown insufficient sleep (<7 h/night) is associated with an increased risk of obesity and type 2 diabetes (1–5). In addition, experimentally induced short-term inpatient sleep restriction in healthy adults decreases insulin sensitivity, assessed by using the hyperinsulinemic-euglycemic clamp procedure (HECP), intravenous glucose tolerance test, or oral glucose tolerance test (6–12). In fact, even a single night of reducing sleep to 4 h decreases insulin sensitivity the following morning (9). These observations suggest extending daily sleep duration should improve insulin sensitivity in people with habitual short sleep schedules. We are aware of three randomized controlled trials that evaluated the effect of extending sleep on insulin sensitivity in people with a history of short sleep, which reported conflicting results (13–15), indicating a need for more definitive studies with robust assessments of insulin action and metabolic health.

The purpose of the current study was to determine whether extending daily sleep to >7 h/night for ∼6 weeks improves multiorgan insulin sensitivity and other cardiometabolic disease risk factors in people with overweight or obesity who maintain habitual short sleep schedules (<7 h/night) and have insulin-resistant glucose metabolism. To this end, we conducted a randomized controlled trial to evaluate the effect of continuing habitual short sleep or extending sleep on 1) whole-body (primarily muscle), liver, and adipose tissue insulin sensitivity, assessed by the HECP and stable isotope tracer infusions; 2) serial plasma glucose, insulin, and nonesterified fatty acid (NEFA) concentrations during wakefulness over 24 h; 3) body composition (fat-free mass, fat mass, and intrahepatic triglyceride [IHTG] content); 4) plasma lipid profile; and 5) daily sleep patterns, stages, and quality, assessed by wrist actigraphy, sleep logs, questionnaires, and polysomnography (PSG).

Research Design and Methods

Study Participants

The study randomized 31 adults with overweight or obesity with habitual short sleep schedules and insulin resistance to continue habitual sleep (HS) (n = 16) or extended sleep (ES) (n = 15). Participants provided written, informed consent before participating in this study, which was approved by the Washington University in St. Louis Institutional Review Board (St. Louis, MO) and registered at ClinicalTrials.gov (NCT03594994). Participants completed a screening evaluation that included a 2-h oral glucose tolerance test (OGTT), self-reported sleep patterns, 7-day sleep assessment by using wrist actigraphy, and an inpatient PSG study. Inclusion and exclusion criteria are provided in the Supplementary Material. One participant in the HS and ES groups withdrew from the study, so 15 participants in the HS group (44 ± 11 years old, 12 women) and 14 participants in the ES group (39 ± 13 years old, 10 women) completed the study (Supplementary Fig. 1).

Study Procedures

Body fat and fat-free mass and IHTG content were determined by DXA and MRI. Participants were admitted to the Clinical Translational Research Unit for ∼48 h to assess sleep outcomes, serial plasma glucose, insulin, and NEFA concentrations over 24 h during wakefulness and insulin sensitivity by using the HECP with stable isotope glucose and palmitate tracer infusions (Fig. 1). Details of the specific study procedures, sample analyses, and statistical analyses are provided in the Supplementary Material.

Figure 1.

A: Example of an inpatient study for an HS participant with a habitual bedtime of 18 h after awakening and a 6-h habitual time in bed. B: Example of an inpatient study for an ES participant with a habitual bedtime of 16 h after awakening and an 8-h intervention time in bed. In all inpatient studies, meals were given in the evening of admission (13 h after awakening on day 1) and at 1 h, 7 h, and 13 h after awakening on day 2. Arrows indicate time of blood sampling; gray bars indicate dim light exposure <8 lux; black bars indicate time in bed with lights off; white bars indicate normal light conditions.

After completing baseline testing, participants randomized to the HS group were instructed to maintain their usual sleep patterns as determined during the baseline sleep monitoring period. Participants randomized to the ES group were given individually tailored sleep plans designed to increase time in bed and extend sleep within the framework of their daily schedules (details in the Supplementary Material). All procedures conducted at baseline were repeated at 5.4 ± 1.6 weeks in the HS group and at 5.5 ± 2.1 weeks after starting the intervention in the ES group while participants maintained their assigned sleep schedule.

Results

Sleep Outcomes

There were no significant differences between groups at baseline in self-reported sleep duration (HS 5.4 ± 1.0 and ES 5.6 ± 0.8 h/night) or apnea-hypopnea index (HS 3.9 ± 3.4 and ES 4.6 ± 4.1 events/h) and periodic limb movements of sleep (PLMS; HS 2.2 ± 3.5 and ES 0.8 ± 1.6 PLMS/night) assessed during the screening PSG. However, 3 participants in the HS group and 5 participants in the ES group had mild obstructive sleep apnea (apnea-hypopnea index >5 and <15), and 3 participants in the HS group had mild PLMS (>5 and <15). Baseline mean 7-day sleep characteristics, assessed by wrist actigraphy, were similar between groups (Table 1). Compared with baseline, daily bedtimes started >1 h earlier, mean daily time in bed increased by 1.3 ± 0.6 h, and total sleep time increased by 1.1 ± 0.5 h in the ES group, without change in the HS group (Table 1). The earlier bedtimes in the ES group resulted in earlier sleep midpoints compared with the HS group. The variability (i.e., SD) in time-in-bed, total sleep time, bedtime, wake time, and sleep midpoint were lower after the intervention in the ES group than in the HS group (Table 1); however, the change in sleep efficiency, sleep latency, and wake after sleep onset after the interventions were not significantly different between groups (Table 1).

Table 1.

Objective and subjective sleep outcomes before and after habitual or extended sleep

	HS group			ES group			ANCOVA P value
	Before	After	Change (95% CI)	Before	After	Change (95% CI)
Actigraphy
Daily time in bed (h)	6.9 ± 0.6	7.1 ± 0.8	0.3 (−0.2, 0.7)	6.5 ± 0.5	7.9 ± 0.6	1.3 (1.0, 1.7)	0.001
Daily time in bed SD (min)	53.7 ± 30.0	76.2 ± 41.0	22.5 (9.1, 36.0)	63.0 ± 29.2	38.4 ± 21.0	−24.6 (−42.8, −6.4)	<0.001
Daily total sleep time (h)	6.1 ± 0.5	6.1 ± 0.5	0.0 (−0.2, 0.2)	5.8 ± 0.5	6.9 ± 0.7	1.1 (0.8, 1.4)	<0.001
Daily total sleep time SD (min)	45.2 ± 27.5	69.7 ± 36.1	24.5 (11.2, 37.7)	54.4 ± 25.5	41.8 ± 18.3	−12.6 (−26.9, 1.7)	<0.001
Daily bedtime (h:min)	23:27 ± 1:14	23:29 ± 1:11	0:02 (−00:11, 00:16)	24:10 ± 1:06	22:42 ± 0:56	−1:28 (−1:57, −0:59)	<0.001
Daily bedtime SD (min)	43.4 ± 21.1	54.0 ± 45.3	10.6 (−7.8, 29.0)	45.3 ± 20.9	29.8 ± 19.3	−15.5 (−29.8, −1.2)	0.028
Daily wake time (h:min)	6:13 ± 1:01	6:27 ± 1:11	0:14 (−00:05, 00:33)	6:32 ± 1:14	6:23 ± 1:23	−0:09 (−00:31, 00:14)	0.111
Daily wake time SD (min)	44.2 ± 24.9	58.5 ± 30.2	14.3 (−3.1, 31.7)	54.6 ± 32.1	30.6 ± 14.4	−24.0 (−41.6, −6.3)	0.002
Daily sleep midpoint (h:min)	2:53 ± 1:06	3:03 ± 1:09	0:10 (−00:03, 00:24)	3:24 ± 1:09	2:38 ± 1:09	−0:45 (−01:08, −00:23)	<0.001
Daily sleep midpoint SD (min)	35.8 ± 17.5	43.2 ± 28.3	7.5 (−6.4, 21.3)	40.3 ± 20.9	24.3 ± 9.6	−16.1 (−28.1, −4.0)	0.011
Sleep efficiency (%)	89.3 ± 2.9	86.3 ± 4.9	−3.0 (−5.5, −0.6)	88.6 ± 3.6	87.9 ± 2.8	−0.7 (−2.8, 1.4)	0.167
Sleep latency (min)	6.1 ± 5.3	10.2 ± 9.0	4.0 (−0.5, 8.6)	5.3 ± 4.9	11.2 ± 10.3	5.8 (−0.4, 12.1)	0.688
Wake after sleep onset (min)	32.7 ± 11.9	39.9 ± 18.7	7.2 (−1.0, 15.4)	29.8 ± 9.8	36.4 ± 10.9	6.6 (−0.1, 13.2)	0.756
Sleep questionnaires
Pittsburgh Sleep Quality Indexa	9.1 ± 3.2	7.4 ± 2.6	−1.6 (−2.9, −0.3)	7.1 ± 2.9	4.4 ± 2.1	−2.7 (−4.8, −0.6)	0.018
Insomnia Severity Indexb	11.5 ± 4.6	9.5 ± 5.1	−1.9 (−4.0, 0.1)	6.8 ± 5.0c	6.7 ± 5.3	−0.1 (−1.9, 1.7)	0.399
Epworth Sleepiness Scalea	6.1 ± 3.3	6.3 ± 2.4	0.2 (−1.0, 1.4)	7.1 ± 4.5	5.3 ± 2.7	−1.8 (−4.5, 0.8)	0.034

Values before and after HS (n = 15) and ES (n = 14) are means ± SD. ANCOVA was used to determine between-group differences after adjusting for values before the interventions.

^aHS group n = 14 and ES group n = 11.

^bHS group n = 13 and ES n = 11.

^cValue significantly different from corresponding value in the HS group, P = 0.02.

Inpatient PSG assessments were consistent with the data obtained by using actigraphy. (Supplementary Table 1). There were no differences in sleep efficiency, sleep latency, or N2, N3, or rapid eye movement (REM) time between groups. The duration of N1 sleep increased and REM sleep latency tended to increase after the intervention in the ES compared with the HS group. There were no differences in the proportion of time spent in any of the sleep stages between groups after the intervention.

Subjective assessment of sleep quality (Pittsburgh Sleep Quality Index) improved in both groups, but the improvement was greater in the ES than in the HS group (P = 0.02). The Insomnia Severity Index scores after the intervention were not different from baseline in either group. Daytime sleepiness (Epworth Sleepiness Scale) improved in the ES but not the HS group, and the change in scores after intervention were different between groups (P = 0.03) (Table 1).

Body Composition and Cardiometabolic Outcomes

Body weight and composition, fasting plasma glucose, insulin, triglyceride, HDL-cholesterol and LDL-cholesterol concentrations, the HOMA-insulin resistance score, 24-h plasma glucose, insulin, NEFA areas under the curve, and whole-body (primarily muscle), hepatic, and adipose tissue insulin sensitivity did not change in either the HS or ES groups, and there were no differences between groups (Table 2 and Fig. 2). There was no significant relationship between the relative change in total sleep time and relative change in whole-body insulin sensitivity (R² = 0.042, P = 0.482).

Table 2.

Cardiometabolic characteristics before and after habitual or extended sleep

	HS group			ES group			ANCOVA P value
	Before	After	Change (95% CI)	Before	After	Change (95% CI)
BMI (kg/m2)	34.8 ± 5.6	34.4 ± 5.2	−0.4 (−0.9, 0.1)	35.5 ± 6.9	35.4 ± 6.7	−0.1 (−0.4, 0.2)	0.241
Body mass (kg)	96.2 ± 17.7	95.0 ± 16.8	−1.1 (−2.5, 0.3)	100.3 ± 18.5	100.0 ± 18.2	−0.3 (−1.1, 0.5)	0.183
Fat-free mass (kg)	51.8 ± 8.6	51.6 ± 8.3	−0.2 (−1.1, 0.7)	55.3 ± 11.1	54.9 ± 11.1	−0.5 (−1.6, 0.7)	0.808
Body fat (%)	44.6 ± 6.9	45.1 ± 7.0	0.5 (−0.1, 1.2)	44.3 ± 8.9	44.7 ± 8.6	0.4 (−0.2, 1.0)	0.786
IHTG content (%)	7.3 ± 6.0	6.6 ± 4.4	−0.6 (−2.0, 0.8)	8.0 ± 8.2	8.6 ± 8.7	0.6 (0.0, 1.3)	0.075
Fasting glucose (mg/dL)	94.9 ± 5.0	98.5 ± 6.7	3.6 (0.6, 6.7)	95.5 ± 4.1	96.3 ± 6.2	0.8 (−2.2, 3.8)	0.188
Fasting insulin (mU/L)	18.3 ± 9.6	18.2 ± 10.3	−0.1 (−2.8, 2.6)	19.3 ± 9.5	20.3 ± 13.2	1.0 (−3.0, 5.0)	0.648
HOMA-insulin resistance	4.3 ± 2.3	4.5 ± 2.6	0.2 (−0.6, 0.9)	4.6 ± 2.3	4.8 ± 3.1	0.3 (−0.7, 1.2)	0.846
Triglycerides (mg/dL)	109 ± 50	129 ± 41	20 (0, 40)	104 ± 43	108 ± 41	4 (−12, 19)	0.089
HDL-cholesterol (mg/dL)	51 ± 14	44 ± 11	−8 (−11, −4)	49 ± 10	46 ± 13	−3 (−9, 2)	0.199
LDL-cholesterol (mg/dL)	117 ± 31	107 ± 25	−10 (−26, 6)	113 ± 35	107 ± 33	−7 (−20, 7)	0.822
Glucose AUCa (mg/dL × 24 h)	2,509 ± 153	2,554 ± 163	45 (0, 90)	2,481 ± 63	2,522 ± 141	41 (−23, 105)	0.947
Insulin AUCa (mU/L × 24 h)	1,218 ± 675	1,219 ± 594	1 (−196, 197)	1,633 ± 1,228	1,589 ± 1,214	−44 (−197, 109)	0.918
NEFA AUCa (mmol/L × 24 h)	7.8 ± 1.4	8.8 ± 2.5	1.0 (−0.1, 2.0)	8.2 ± 1.8	7.9 ± 1.3	−0.3 (−1.3, 0.7)	0.111

Values before and after HS (n = 15) and ES (n = 14) are means ± SD. ANCOVA was used to determine between-group differences after adjusting for values before the interventions. AUC, area under the curve assessed by serial blood sampling over 24 h when the participants were awake.

^aHS group n = 13 and ES group n = 12.

Figure 2.

A: Plasma glucose, insulin, and NEFA concentrations in the HS and ES groups before (orange circles) and after (dark gray circles) the interventions. Light gray vertical bars indicate meals given at 1 h, 7 h, and 13 h after awakening (at 0800 h, 1400 h, and 2000 h, respectively, when wake time is arbitrarily set to 0700 h, whereas actual wake time was determined by the participant’s schedule). Data are means and 95% CIs. B: Whole-body (primarily skeletal muscle) insulin sensitivity (glucose R_d/kg fat-free mass [FFM] during the HECP), hepatic insulin sensitivity index (reciprocal of the product of endogenous glucose R_a/kg FFM and plasma insulin concentration during basal conditions), and the adipose tissue insulin sensitivity index (reciprocal of the product of palmitate R_a/kg FFM and plasma insulin concentration during basal conditions) in the HS and ES groups before (orange bars) and after (dark gray bars) the interventions. Data are means and 95% CIs, and circles represent individual participant values.

Secondary Sensitivity Analyses

Sleep extension in the subset of participants in the ES group who obtained >7 h sleep/night (n = 7) demonstrated significant improvements in most of the sleep outcomes observed in the entire ES cohort (Supplementary Table 2), but did not detect significant effects on multiorgan insulin sensitivity and the other cardiometabolic outcomes (Supplementary Table 3 and Supplementary Fig. 2), consistent with the findings of the primary analysis.

Conclusions

We assessed the cardiometabolic effects of extending daily sleep duration in people with overweight or obesity and insulin-resistant glucose metabolism who maintain short sleep schedules. Our data demonstrate that extending sleep by ∼1 h/night for 5.5 weeks improves key features of sleep health, including subjective sleep quality and daytime sleepiness, and tended to increase all stages of sleep and REM latency without altering the proportion of time spent in each stage of sleep. Sleep extension did not affect multiorgan (whole-body [primarily muscle], hepatic, and adipose tissue) insulin sensitivity, decrease serial plasma glucose, insulin, and NEFA concentrations during wakefulness, decrease IHTG content, or improve plasma lipids. These data demonstrate that increasing sleep by ∼1 h/night improves sleep health but does not have therapeutic effects on cardiometabolic function in people who have overweight or obesity with insulin resistance and habitual short sleep schedules.

The reason(s) for the discordance between the effect of increased sleep duration on sleep health and metabolic outcomes is not clear, but could be related to the following possibilities. First, the daily amount (∼1 h/night increase) and/or duration (∼6 weeks) of our sleep intervention was not sufficient to improve insulin action or glycemic control in our particular study population (overweight or obesity and insulin resistance). Second, chronic metabolic dysfunction might not be responsive to sleep extension, but intervening earlier might have been effective. Third, habitual short sleep was not an important ongoing contributor to metabolic dysfunction in our participants.

Our study has several limitations. Half of the participants in the ES group did not increase sleep duration to ≥7 h/night, so it is possible the sleep deficit was not adequately treated in this group. However, the effect of sleep extension on sleep and metabolic outcomes in participants who increased their sleep to ≥7 h/night were not different than in the entire cohort. It is possible an effect of extending sleep was missed because of an inadequate number of participants. However, the effect size for our primary outcome (whole-body insulin sensitivity) was very small (Cohen d = 0.053), suggesting thousands of participants would be needed to detect a small, presumably clinically insignificant difference between groups. Our study participants were predominantly women, so our findings might not apply to men.

In conclusion, extending sleep by ∼1 h/night for 5.5 weeks improves sleep health in people with habitual short sleep who have overweight or obesity and have insulin resistance but does not improve multiorgan insulin sensitivity or other metabolic outcomes, including plasma glucose, insulin, and NEFA concentrations, IHTG content, and plasma lipid profile.

This article contains supplementary material online at https://doi.org/10.2337/figshare.30933611.

Funding Statement

This study was supported by National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases grants R01 DK115502, P30 DK056341 (Washington University in St. Louis Nutrition and Obesity Research Center), and P30 DK020579 (Washington University in St. Louis Diabetes Research Center), National Center for Advancing translational Sciences grant UL1 TR000448, and National Heart, Lung, and Blood Institute grant T32 HL130357.