Sage Journals: Discover world-class research

Abstract

Background

Sleep insufficiency is known to negatively impact on glucose metabolism. Consequently, there is interest in determining the impact of improving sleep on glucose metabolism. We conducted a meta-analysis of studies that aimed at improving sleep using cognitive behavioural therapy for insomnia (CBT-I) and/or sleep hygiene or sleep extension on glucose metabolism.

Methods

Searches were performed on MEDLINE, EMBASE, CINAHL and Cochrane. We included studies featuring adults≥18years, a sleep intervention and glycaemic measurements. The pooled mean differences were calculated by the inverse variance method.

Results

24 studies (15 CBT-I and/or sleep hygiene; 9 sleep extension) were included. Meta-analysis of 12 studies (n = 2,044) of CBT-I and/or sleep hygiene demonstrated a significant reduction in HbA1c of 0.27% (95% CI 0.07, 0.47, I² 74%, p = 0.008) compared to control. In T2DM (n = 1,911; 9 studies), HbA1c level decrease was 0.43% (0.19, 0.67, I² 59%, p = 0.0004). There were no significant changes in fasting blood glucose analyses nor in any sleep extension intervention. For quality assessment, only 9 studies had low concern.

Conclusions

Using CBT-I and/or sleep hygiene interventions led to significant reductions in HbA1c levels, which were clinically meaningful in T2DM. Addressing sleep insufficiency should be an integral part of diabetes care.

Registration

PROSPERO Identification number: CRD42022376606.

Keywords

Sleep disorder intervention glucose metabolism type 2 diabetes mellitus systematic review and meta-analysis

Key messages

• Sleep insufficiency is known to be associated with incident type 2 diabetes mellitus (T2DM) or worse glycaemia control in those with established T2DM.

• Hence, there has been much interest in evaluating the impact of improving sleep on glucose metabolism.

• Our meta-analysis demonstrated using cognitive behavioural therapy for insomnia and/or sleep hygiene was associated with a significant decrease in HbA1c levels of 0.27% and 0.43% in the total cohort and people with T2DM respectively, compared to control.

• These findings suggest more emphasis should be placed on detecting sleep insufficiency in clinical care, particularly for people with T2DM.

Introduction

Sleep insufficiency, either as a function of poor quality, timing, or duration, are known to negatively impact on glucose metabolism.^1–4 They increase the risk of type 2 diabetes mellitus (T2DM) in the general population, predict progression to T2DM in people with pre-diabetes and worsen glycaemic control in those with established T2DM.^1–4 Consequently, there is much interest in determining the impact of improving sleep on glucose metabolism. Interventions include pharmacological therapies, sleep manipulation techniques including education/ sleep hygiene, cognitive behavioural therapy for insomnia (CBT-I) and/ or sleep extension. A further option is continuous positive airway pressure (CPAP) specifically in the context of obstructive sleep apnoea (OSA).

Sleep intervention techniques can be applicable to a range of sleep disturbances and behaviours and do not involve taking medications or wearing equipment. If these interventions were demonstrated to have a beneficial impact on glucose metabolism, they could be used to complement existing lifestyle-based interventions that are already recommended for people with pre-diabetes and T2DM.⁵ As the prevalence of T2DM and its complications increase, there is a need for strategies to prevent T2DM and reduce its health and economic burdens.

Two previous meta-analyses did not find any significant impact of CBT-I and/or sleep hygiene or sleep extension on glucose metabolism.^6,7 However, the results were not analysed according to baseline glycaemic categories (i.e. people with and without T2DM), possibly due to the low number of studies available at the time. As such, it is unknown if these interventions have a differential impact on specific glycaemic categories.

Given the increasing number of published studies in this area, the aim is to address this issue by analysing results for each intervention according to glycaemic categories. We hypothesised that in people with T2DM, whom have a stronger degree of glucose intolerance, the observed changes in glucose metabolism through use of sleep interventions would be larger than in people without T2DM, mirroring the observed effects of using diet and exercise lifestyle programs in people with and without T2DM, though some changes would be expected in both groups.^8–10

The aim of this systematic review and meta-analysis was to assess the impact of non-pharmacological sleep interventions on glucose metabolism according to baseline glycaemic category.

Methods

This systematic review was registered on International Prospective Registry of Systematic Reviews (PROSPERO Identification number: CRD42022376606) database and meets the requirements of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.¹¹

Data sources, search strategy and study selection

The search strategy was developed by SM and AT, in collaboration with an experienced local hospital health sciences librarian, to identify relevant studies using MEDLINE, Embase, CINAHL and Cochrane bibliographic databases. Databases were searched using a combination of appropriate subject headings for indexed articles, tailored to individual bibliographic databases, where required, and free-text terms relevant to the framed question of this review, which were combined using Boolean operators. The full electronic search strategy used for one database is included in Appendix 1. The first literature search was performed from inception up to 16^th November 2022; an additional search was performed until 19^th January 2024.

After conducting the electronic database searches, duplicates were removed and two investigators (SM and WH) independently screened and assessed titles and abstracts of identified studies for eligibility. The full text of these papers was retrieved and independently assessed for inclusion by the same two reviewers. Discrepancies were resolved through a third investigator (AT) where necessary until a consensus was reached. The reference lists of review articles and of included original publications were also screened for potentially relevant studies. For any relevant trial protocols derived from the database searches, the corresponding publication was searched for separately. We included studies of any language. We focused on studies published as full text papers, but we considered abstract publications, if further information/ data could be provided by the original authors, following methods used in a similar systematic review.⁶

We included intervention studies of adults ≥18 years old at baseline that were (a) randomised or non-randomised trials, quasi experimental, laboratory or single-arm intervention studies, (b) consisted of a sleep intervention as the exposure (sleep extension, education, hygiene and/ or cognitive behaviour therapy) and (c) measured either blood glucose levels (glycated haemoglobin, HbA1c, fasting blood glucose, FBG, and/ or two-hour blood glucose), glycaemic metrics/ measures derived from electronic devices (e.g. continuous glucose monitoring devices for time in target glucose range) or surrogate markers of glycaemia (e.g. insulin level), at both baseline and follow-up.

In the case of a two-arm study, the comparator was considered as either a control group (i.e. no intervention), a lifestyle program or advice session consisting of diet and exercise modifications (considered as standard care in pre-diabetes and T2DM), or certain other interventions (e.g. mindfulness, meditation). If the sleep intervention was part of a multifactorial lifestyle intervention, the study was included only if control arm consisted of the same lifestyle intervention without the sleep component. For this reason, in the case of any three-arm studies, we only included data from two arms.

The exclusion criteria consisted of studies that (1) did not feature a sleep intervention (2); consisted of interventions of CPAP for OSA, bright/ dim light exposure, pharmaceutical therapies or laboratory based sleep restriction/ deprivation (unless as a part of CBT-I) as the exposure; (3) did not report measures of the same glucose metabolism marker at both baseline and follow-up or (4) studies of type 1 diabetes mellitus or (5) studies where reported glucose marker may not be accurate (e.g. HbA1c levels in dialysis studies of end-stage renal failure).

If multiple published reports from the same study were available, we aimed to include the report with the most detailed information for both exposure and outcome, unless the essential outcomes were presented in different publications.

Data extraction and quality assessment

One investigator (SM) extracted the data and the second investigator (WH) independently checked for consistency. Any discrepancies that existed were discussed in arbitration (SM, WH and AT) until a consensus was reached. Data was extracted on the following: first author’s surname, year of publication, country of origin of the population studied, type of study (e.g. randomised trial), sample size of intervention and control arms, population demographic characteristics (e.g. age, sex, body mass index), baseline glycaemic status of the population, sleep variable(s) under analysis (subjective and/ or objective measures), type of sleep intervention, length of follow-up, measure(s) used to assess change in sleep variable, type of glycaemia marker(s) measured, and the changes in the levels of the sleep variable and glycaemia at baseline and follow-up.

Sleep interventions were categorised as either (a) CBT-I and/ or sleep hygiene, in which the sleep outcomes tend focus on sleep quality or (b) sleep extension, which focuses more on sleep duration (SLD) or time in bed (TIB). These two categories also follow those that were used in a previous systematic review.⁶

Study quality assessment was performed independently for parallel arm or cross-over studies by two reviewers (SM and WH), according to Cochrane risk-of-bias tool for randomized trials version 2 (ROB-2), as this is commonly used and recommended tool for assessing risk of bias in randomised trials.¹² Reviewers met to compare results and reach consensus. ROB2 includes five assessments of possible bias including randomisation process, deviation from proposed interventions, missing outcome data, outcome measurement, and result reporting. Studies were judged to have low, moderate or high risk of bias for each assessment.

Statistical analysis

Data was analysed quantitatively using meta-analysis if there were three or more studies that used similar sleep interventions and reported the same outcome variable. Otherwise, we proceeded with summarising the data qualitatively/ narratively. Single-arm interventions were excluded from meta-analysis as no comparator group was available.

Firstly, data was analysed in the total population and then according to glycaemic category (those with T2DM compared to non-T2DM). Where sufficient data existed, we used categories of normal glucose tolerance, pre-diabetes and T2DM. In cases where cohorts had a mixture of glycaemic categories, if one category was vastly more prevalent over others, it would be placed in this category. For example, a cohort of 90% with T2DM and 10% with pre-diabetes would be categorised under T2DM; in contrast a cohort with a mean HbA1c level of 5.5% (standard deviation, SD, 0.4) would be placed under non-T2DM.

Where data in the original publication was reported as both intention-to-treat and per protocol analyses, we selected the former to reduce the impact of bias.

Where relevant data was not available in the original publication, but was published in subsequent meta-analyses, we extracted the data from the latter to ensure a complete dataset and for better comparison with previous systematic reviews. Otherwise, we contacted the first author of the original publication for the most essential results.

The outcome variables were reported as mean (SD). When the original study reported outcomes as median (interquartile range, IQR) or mean (95% CI), these were converted to mean (SD) using accepted methods.¹³

We estimated the pooled mean differences (and 95% CI) of the variables of interest, considered as the effect size for all studies. These were calculated using the inverse variance method and through using a random-effects model, as we anticipated there would be some degree of study-level variability in the exact interventions used. We conducted a brief sub-analysis using fixed effects model. Heterogeneity among studies was quantified by I² statistic. We considered low, moderate, and high I² values to be 25, 50, and 75%, respectively.¹⁴

All statistical analyses for the meta-analysis were performed with Review Manager (RevMan) software Version 5.4.1 (Copenhagen, Cochrane Collaboration) and all tests were two-sided with p-level of p < 0.05 considered as statistically significant.

Results

Identification of studies

There were 901 abstracts derived from the search strategy, from which 33 met inclusion criteria for full text analysis; of these 24 studies were included in the systematic review (Figure 1 including the reasons for exclusion;^15–38,⁴¹).

Figure 1.

Flow chart to demonstrate the process of study selection. Key: CBT-I = cognitive behaviour therapy for insomnia, n= number. * For reasons excluded, there may have been more than 1 reason, however the allocated category represents the first obvious reason for exclusion.

Summary of included studies and study characteristics

From the 24 studies, 15 used a CBT-I and/ or sleep hygiene intervention and nine used a sleep extension intervention. Table 1 describes the baseline characteristics of these studies and Appendix C: supplementary data summarises the findings. The studies originated from the Americas (9 studies in USA, 1 in Brazil), Europe (2 in UK, 1 in each of Spain, France, Turkey, Holland and Belgium), Asia (3 in China, 1 in each of Hong Kong and Thailand) and 2 from Australia. Six studies did not have a control/ comparator group. The length of follow-up ranged from 7 weeks to 6.2 years for CBT-I and/ or sleep hygiene studies and the most reported glycaemic marker measured was HbA1c (14 out of the 15 studies). In contrast, the length of sleep extension studies ranged from three nights to 2.7 months and the most frequently reported glycaemic marker was fasting glucose (five out of nine studies).

Table 1.

Baseline characteristics of studies featuring sleep interventions.

First Author, year, Country, Reference no	Population	Study type	Intervention (vs control)	Study length	Intervention						Control
First Author, year, Country, Reference no	Population	Study type	Intervention (vs control)	Study length	N	Age, years	Male %	BMI (SD)	Baseline glucose markers + sleep measurements^b		N	Age, years	Male %	BMI (SD)	Baseline glucose markers + sleep measurements^b
a) Cognitive behaviour therapy for insomnia (CBT-I) or sleep education/ hygiene in T2DM
Alshehri, 2020, USA.^15,16	T2DM with insomnia symptoms	RCT	CBT-I vs usual T2DM education (6x 1-h sessions)	7 weeks	13	61.8 (6.4)	28.6	33.7 (4.9)	HbA1c 7.8 (2.1) T2DM dx 15.1 (9.9) PSQI 10.5 (1.9) ISI 16.1 (3.2) TST 423.6 (63.3)		13	59.4 (9.4)	50	31.4 (5.4)	HbA1c 6.5 (0.6) T2DM dx 15.4 (11.2) PSQI 9.1 (2.2) ISI 15.7 (3.1) TST 425.4 (66.7)
Cheung, 2022, Hong Kong.¹⁷	T2DM SSD <7 hours	Larger prospective follow of Li 2018	1 × 30-min sleep + T2DM education (vs. T2DM education only), 1 × telephone call/month	6.2 years	91	53.9 (8.4)	47	25.7 (4.2)	HbA1c 7.2 (6.6-7.8) FBG 6.9 (6.1-7.8) SLD 438 (102)		88	52.0 (8.6)	50	25.4 (4.7)	HbA1c 7.1 (6.6-7.8) FBG 7.2 (6.1-7.8) SLD 426 (66)
García-Serrano, 2022, Spain.¹⁸	T2DM (90%), PDM (10%)	RCT	Sleep education/ hygiene (vs control)	6 months	38	66.5	42.1	29.9	HbA1c 6.45 (5.9-7.15) FPG 7.0 (6.2-8.4) PSQI 8.0 (6.0-10.8) SLD 360 (300-390) Efficiency 77.3%		31	64.5	51.6	31.0	HbA1c 6.40 (5.85-7.1) FPG 6.8 (5.9-7.4) PSQI 8.0 (6.0-11) SLD 360 (300-405) Efficiency 75.0%
Groeneveld, 2024, Holland.⁴¹	T2DM with insomnia symptoms	RCT	CBT-I (vs control)	6 mnoths	29	67.7 (8.4)	75.9	30.0 (7.0)	HbA1c 7.5 (3.7) FBG 9.7 (3.6) ISI 15.3 (3.7) TIB 8.4 (1.2) SLD (diary) 354 SLD (actigraph) 402 Efficiency 84.7%		28	71.8 (7.3)	57.1	29.9 (4.7)	HbA1c 7.4 (36) FBG 8.8 (2.2) ISI 13.1 (4.1) TIB 8.8 (1.4) SLD (diary) 378 SLD (actigraph) 444 Efficiency 86.9%
Gunduz, 2022, Turkey.¹⁹	T2DM	RCT in real world outpatient clinics	Sleep hygiene training/ leaflet (vs usual care)	Within 6 months	113	56.6 (9.4)	24.8	32.9 (5.6)	HbA1c 7.8 (1.7) T2DM dx 59.3% >5 years PSQI 11.97 (4.00)		234	58.6 (10.1)	23.1	31.3 (5.3)	HbA1c 7.9 (1.9) T2DM dx: 76.5% >5 years PSQI 10.41 (4.83)
Li, 2018, China.²⁰	T2DM with HbA1c >8% and SSD <6 hours	Pilot RCT	2 × 30-min sleep education + T2DM education (vs. T2DM education)	3 months	16	53.7 (17.2)	63	25.8 (3.2)	HbA1c 10.4 (1.7) FBG 10.5 (6.8) T2DM dx 7.3 (7.5) HOMA-IR 4.5 (3.5) PSQI 7.6 (3.7)		15	51.6 (12.2)	60	26.1 (3.1)	HbA1c 9.7 (0.9) FBG 10.2 (5.4) T2DM dx 5.6 (5.5) HOMA-IR 4.3 (3.1) PSQI 7.3 (4.1)
Pandey, 2018, USA.²¹	T2DM, 89% abnormal sleep	RCT	Sleep hygiene + T2DM education (vs T2DM education) – 4 sessions	12 weeks	35	52.7	50	30.8	SLD 327.6 (172.8)		35	54.5	50	33.2	SLD 339.6 (118.2)
Pereira, 2019, Brazil.²²	T2DM, poor sleep quality (PSQI >5)	RCT	4 × sleep hygiene/ weekly phone call vs. diabetes education	3 months	45	55.8 (7.4)	31	31.3 (6.1)	HbA1c 8.12 (1.92) T2DM dx 6.59 (5.6) PSQI 9.8 (2.9)		46	54.0 (6.5)	54.0	31.2 (5.4)	HbA1c 8.23 (2.22) T2DM dx 8.03 (6.0) PSQI 9.6 (2.8)
Tannas, 2013, USA.²³	T2DM, post-menopausal	Single arm, prospective	CBT-I, once weekly x 7 weeks	10 weeks	9	56-70	0	-	HbA1c 7.02 PSQI 2.89 (1.27) ISI 18.44 SLD 355.2 (88.2)		-	-	-	-	-
Zhang, 2021 China.²⁴	T2DM + poor sleep quality	Cluster RCT	CBT-I: 6 sessions + lectures over 6 weeks (Vs usual care)	12 months	568	61.6 (9.2)	34.9	24.8 (4.2)	HbA1c 7.44 (1.12) T2DM dx 5.07 (5.25) PSQI 10.07 (2.38)		574	61.8 (8.6)	33.3	25.2 (4.9)	HbA1c 7.50 (1.13) T2DM dx 5.21 (4.91) PSQI 10.04 (2.37)
Zuo, 2020, China.²⁵	T2DM with PSQI score >7	RCT	CBT-I: 7 sessions + lectures over 7 weeks (vs usual care)	6 months	94	63.9 (10.2)	34	24.7 (3.1)	HbA1c 8.1 (1.9) T2DM dx 6.2 (4.8) PSQI 11.0 (2.5)		93	61.7 (10.4)	31.2	24.9 (3.2)	HbA1c 7.9 (2.0) T2DM dx 5.5 (4.9) PSQI 11.3 (2.5)
b) Cognitive behaviour therapy for insomnia (CBT-I) or sleep education/ hygiene in non T2DM
Cain, 2020, USA.²⁶	Pregnant females + SDD <6.5hours/ night	RCT	Sleep hygiene + CBT-I vs usual care for 7 weeks (2^nd trimester to 6-8 weeks post-partum)	Up to approx. 34 weeks	27	28.5 (5.8)	0	35.5 (5.5)	PSQI 12.7 (6.0) ISI 18.7 (6.0)		26	29.8 (5.3)	0	38.4 (10.5)	PSQI 11.6 (3.5) ISI 17.0 (6.0)
Carroll, 2013, USA.²⁷	Insomnia	3-Arm RCT, comparative efficacy trial, (only 2 arms compared)	Weekly CBT-I for 4 months vs sleep seminar	16 months	47	64.8 (6)	21	25.4 (3.3)	HbA1c 6.0 (0.9) FBG 5.3 (1.4) PSQI 10.4 (2.9) TST 339.5 (9.7)		23	30	66 (7.7)	26.1 (4)	HbA1c 6.7 (2.2) FBG 5.9 (1.6) PSQI 11.2 (3) TST 342.4 (14.7)
Chelsa, 2020, USA.²⁸	Females with MetS or PDM	Non-randomised 2-arm trial	Lifestyle + sleep vs lifestyle (8 sessions over 3 months + telephone support)	6 months	12	52.9 (11.3)	0	31.1 (3.7)	HbA1c 5.7 (0.3) FBG 5.3 (0.5)		14	53.8 (12.7)	32.0 (3.8)	0	HbA1c 5.6 (0.3) FBG 5.3 (0.5)
Duncan, 2020, Australiaz.²⁹	Obese + overweight adults	3-Arm RCT (only 2 arms compared)	Sleep + lifestyle vs lifestyle	12 months	14	47.2 (9.4)	17.3	31.9 (4.0)	HbA1c 5.5 (0.4) PSQI 7.3 (2.8)		23	45.4 (10.2)	26.8	31.4 (3.8)	HbA1c 5.5 (0.7) PSQI 7.0 (3.1)
C) Sleep extension studies
First Author, year, Country	Population	Study type	Intervention (vs control)	Study length	N	Age, years	Male %	BMI (SD)	Baseline glucose markers + sleep measurements ^b	N		Age, years	Male %	BMI (SD)	Baseline glucose markers + sleep measurements^b
Al- Khatib, 2018, UK.³⁰	Without T2DM, SSD 5 to <7hours hours/night	RCT	45-min sleep consultation session on sleep extension/ hygiene	4 weeks	21	24.0 (21.5-29.0)	14.3	22.5 (21.3,29.0)^a	FBG 4.7^$Fasting insulin 9.0HOMA-IR 2.1PSQI 6.8TIB 423 minsSLD 328 minsEfficiency 86.2%	21		25.0 (22.5-30.5)	19.0	21.8 (20.9,22.8)^a	FBG 5.0^aFasting insulin 8.4HOMA-IR 1.8PSQI 5.2TIB 429 minsSLD 353 minsEfficiency 82.0%
Chennaoui, 2016, France.³¹	Without T2DM, SLD 6-9 hours/ night	Randomised Cross-over study	Increase TIB by 1.5 to 2 hours/ night vs habitual sleep (6-week washout period)	6 nights each	14	31.4 (3.9)	100	24.0 (2.0)	-	-		-	-	-	-
Cizza, 2014, USA.³²	Obese (BMI 30-55), SSD <6.5 hours/night	RCT but only pre- randomisation data considered, as a single arm	Only pre-intervention data (between screening and baseline) analysed here (RCT ended early), hence no official intervention	Median 2.7 months	125	40.6 (6.9)	24	38.4 (6.1)	FBG 5.4 (1.0)Fasting insulin 15.5QUICKI 0.327SLD (diary) 357.4SLD (actigraph) 344.4Efficiency 79.5%	-		-	-	-	-
Hartescu, 2021, UK.³³	Without T2DM, SSD <6.5 hours/night	RCT	Advised to increase TIB by 1 hour/ night, Counselling sessions of CBT principles	6 weeks	10	39.9 (8.8)	100	29.3 (1.6)	FBG 5.72MDG 5.82Fasting insulin 52HOMA-IR 2.15PSQI 10.8SLD 344.2 minsTIB 398.4 minsEfficiency 85.8%	8		43.0 (10.1)	100	29.8 (4.2)	FBG 5.55MDG 5.80Fasting insulin 62HOMA-IR 2.36PSQI 9.44SLD 349 minsTIB 431.6Efficiency 80.7%
Killick, 2015, Australia.³⁴	Without T2DM, restricted mid-week sleep	Cross-over study, Randomised	Lab study: TIB of 6 or 10 hours/ night	3 nights	19	28.6 (2.0)	100	26.0 (0.8)	SLD midweek 372SLD weekend 510	-		-	-	-	-
Leproult, 2015, Belgium.³⁵	Without T2DM	Single arm prospective trial	Advised to increase TIB by 1-h/ night	6 weeks	16	33.6	18.8	20.8	PSQI 4.8 (2.5)TIB 6.5 (5.8-6.8)	-		-	-	-	-
Matthew, 2023, USA.³⁶	Without T2DM	Cross-over study	1 week habitual sleep, then 1 week increase ST by 1hour/night	2 weeks	12	20.2 (1.5)	17%	24.5 (3.5)	FBG 5.0 (0.4)SLD 415.8 (49.8)	-		-	-	-	-
Reutrakul, 2022, USA.³⁷	History of GDM and SSD< 7hours / night	RCT	Sleep extension vs healthy living information	6 weeks	9	42.0 (2.9)	0	32.7 (5.3)	FBG 5.62hr PG 8.0HOMA-IR 3.6PSQI 7.8SLD 369.6Efficiency 88.2%	5		38.7 (6.0)	0	32.6 (4.6)	FBG 5.42hr PG 7.6HOMA-IR 1.5PSQI 6.4SLD 408.9Efficiency 90.7%
So-ngern, 2019, Thailand.³⁸	Without T2DM, SSD <6 hours/night	Cross-over study (21-day washout)	Advised to increase TIB by 1 hour/night	2 weeks	21	33.6 (6.1)	10	22.6 (3.2)	SLD 35441)Efficiency 84.8%	-		-	-	-

Continuous variables represented as mean (SD) or median (IQR), except for.

^aRepresents the geometric mean with 95% CI.

^bUnits for variables: T2DM dx: years, FBG, MDG and 2hr-PG: mmol/l, HbA1c: %, SLD, TIB and TST in minutes.

Key: 2hr-PG: 2-h plasma glucose, BMI = body mass index, CBT-I = cognitive behaviour therapy for insomnia, DM Dx = Diabetes mellitus duration, FBG = fasting blood glucose, GDM = Gestational Diabetes Mellitus, HOMA-IR = Homeostatic Model Assessment of Insulin Resistance, IQR = interquartile range, ISI = insomnia severity index score, MDG = Mean Daily Glucose (via Continuous Glucose Monitor), MetS = Metabolic syndrome, min = minutes, N = number, PDM = pre-diabetes, PSQI = Pittsburgh Sleep Quality Index score, QUICKI = quantitative insulin sensitivity check index, RCT = randomised controlled trial, SD = standard deviation, SLD = sleep duration, SSD = short sleep duration, TIB = time in bed, TST = Total sleep time, T2DM = type 2 diabetes mellitus.

Assessment of study quality

From the 20 parallel-arm or cross-over studies considered for study quality, only nine studies had low concern throughout all five assessment categories, with most other studies having moderate concern in one to three of the categories, Appendix 2.

Interventions of CBT-I and/ or sleep hygiene

The interventions varied between CBT-I in 7 studies (Alshehri, Groeneveld, Tannas, Zhang, Zuo, Cain and Carroll) and sleep education/hygiene/advice in 8 studies (Cheung, Garcia-Serrano, Gunduz, Li, Pandey, Pereira, Chelsa and Duncan). The length and frequency of the interventions also varied (Table 1). All studies reported a positive impact on sleep measures, particularly for PSQI but some also reported improvements on the Insomnia Severity Index (ISI) (Alshehri, Groeneveld, Tannas); the only exception was a non-randomised trial (Chelsa) of general sleep advice.

In studies of people with T2DM, there were reductions in HbA1c levels observed in seven studies (Li, García-Serrano, Gunduz. Pandey, Tannas, Zuo and Zhang) but not in others (Alshehri, Groeneveld, Cheung and Pereira). García-Serrano et al. observed 29.4% of the intervention arm achieved a post-intervention HbA1c level decrease of >0.5%, considered a clinically significant HbA1c level reduction. FBG reductions were reported in some studies (Li and Cheung) but not others (Garcia). Cheung et al., conducted a long-term follow-up of median 6.2 years and reported a significant reduction in FBG levels in the intervention arm (6.9 mmol/l, 6.1-6.8, vs 7.1, 6.6-7.8, p = 0.027), with the authors suggesting a legacy effect of their sleep intervention on glucose metabolism.

In people without T2DM, none of the studies reported any changes in glucose markers of FBG or HbA1c levels where measured.

Sleep extension studies

All nine studies were classified as non-T2DM and were generally able to improve various sleep measures, but no study reported post-intervention lower levels of FBG. However, Reutrakul et al. reported a lower increase in mean FBG levels in intervention participants compared to control. Also, some studies reported post-intervention changes in glycaemic pre-cursor markers that favoured a lower diabetes risk: Hartescu et al. observed lower HOMA_IR levels; Killick et al. reported lower levels of fasting insulin and HOMA-IR and higher QUICKI levels; Matthew et al. reported lower glucose area under the curve levels, So-ngern et al. reported lower HOMA-IR levels, insulinogenic index and disposition index in those who extended SLD to greater than 6.0 hours/night.

The RCT of Cizza et al., was stopped early, as in the 2.7 months between baseline and screening, 125 participants significantly increased their SLD before undertaking any intervention, which led to significant reductions in levels of FBG, fasting insulin and an increase in Quantitative Insulin Sensitivity Check Index (QUICKI).

Meta-analysis

Data from these studies were extracted to conduct meta-analyses (Figures 2 and 3), where sufficient data was available. Zuo et al., and Zhang et al., were confirmed as separate studies hence both were included (personal communication from Professor Peian Lou, Xuzhou Center for Disease Control and Prevention, Jiangsu, China). We were unable to confirm if Li et al. and Cheung et al. contained some similar participants and as the latter was available as an abstract publication, we included Li et al. only.

Figure 2.

The effect of CBT-I and/or sleep hygiene on variables: (a) HbA1c levels in (i) Total cohort, (ii) T2DM and (iii) non-T2DM; (b) fasting blood glucose levels in (iv) Total cohort, (v) T2DM and (vi) non-T2DM; (c) PSQI in (vii) Total cohort, (viii) T2DM and (ix) non-T2DM; ISI in total cohort.

Figure 3.

The effect of sleep extension in people without T2DM on (a) Fasting blood glucose, (b) HOMA-IR, (c) Sleep duration and (d) PSQI.

For interventions of CBT-I and/ or sleep hygiene, from 12 studies of 2,044 people, use of the intervention was associated with a significant reduction in HbA1c levels of 0.27% (95% CI 0.07, 0.47) p = 0.008, I² for heterogeneity (I²) = 74%, compared to control. Considering the sub-analysis for 1,911 people with T2DM from 9 studies, use of the interventions demonstrated an HbA1c level decrease of 0.43% (0.19, 0.67), p = 0.0004, I² = 59%. In contrast, for 133 people without T2DM from three studies, the HbA1c level did not change, 0.0% (−0.23, 0.23), p = 0.99, I² = 66%.

Using a similar analysis where the sufficient data was available, we found use of the interventions did not reduce FBG levels in the total cohort (n = 6 studies) nor in those with or without T2DM (n = 3 studies).

Furthermore, use of the interventions was associated with a decrease in PSQI scores of 1.63 (1.05, 2.22), p < 0.00001, I² = 69%, 1.67 (1.02, 2.32), p < 0.00001, I² = 75% and 1.47 (−0.29, 3.24), p = 0.11, I² = 55% in the whole cohort and in those with and without T2DM, where the number of included studies were 10, 7 and 3 respectively.

For sleep extension studies, from five studies with 106 people, use of the intervention did significantly change FBG levels, −0.01 mmol/l (−0.20, 0.17), I² = 13%, p = 0.90. This was despite increasing SLD by 48.38 minutes (18.6, 78.16), I² = 82%, p = 0.001 and decreasing PSQI score by 1.06 (0.13, 1.98), I² = 53%, p = 0.02 in three and four studies respectively. HOMA-IR levels did not significantly change −0.21 (−0.45, 0.04), I² = 18%, p = 0.1.

Using fixed effects models in place of random effects models, did not change the significance for associations.

Discussion

This systematic review and meta-analyses of non-pharmacological interventions is the first to demonstrate use of CBT-I and/ or sleep hygiene education leads to a significant reduction in glycaemic control in people with insufficient sleep. In people with T2DM the HbA1c level was 0.43% lower, a clinically meaningful reduction. Considering all individuals irrespective of their glycaemic status, the HbA1c level was reduced significantly by 0.27%, however there were only 133 people without T2DM, suggesting this result needs further investigation, especially for pre-diabetes.

Considering sleep extension interventions, despite successfully extending sleep by 48 min, on average, there was no change on FBG or HOMA-IR levels. Whilst epidemiological studies have associated short sleep duration and diabetes, our results suggest that sleep extension does not impact on glycaemic control, that is this was a true effect. However, we should also consider that these studies were conducted over shorter time frames, suggesting there was less time for interventions to exert their effect on glucose metabolism. Also, within the systematic review the sleep extension interventions demonstrated improvements in some surrogate markers of glucose metabolism that favour early changes towards lowering glucose levels. Secondly, the sleep extension studies consisted exclusively of people without T2DM; our hypothesis suggested that larger glucose metabolism changes would occur in those with T2DM, as they possess a more severe degree of glucose intolerance, which appeared to be true for CBT-I and/or sleep hygiene interventions.

There were potential causes of heterogeneity in the meta-analysis impacting on the results. For the study designs, the interventions varied in length of time, number of sessions, the length of the subsequent follow-up and support given during the follow-up period. Other possible factors included the baseline population characteristics such as age, gender, BMI, and in those with T2DM, the level of glycaemic control and length of T2DM duration.

Meta-analyses of people with T2DM have demonstrated that poor sleep quality is associated with worsening of glycaemic control by an HbA1c level of 0.35% (95% CI 0.12-0.58).⁴ Insufficient sleep is also known to be associated with future T2DM in the general population and also predict progression from pre-diabetes to T2DM.^1–4 The plausible mechanisms explaining these associations centre on changes in beta-pancreatic cell function and insulin resistance/ sensitivity and include changes in hormones levels for catecholamines giving a sympathetic predominance, melatonin, cortisol and appetite regulating hormones such as ghrelin, leptin, endocannabinoids and glucagon-like peptide-1 and growth hormone.³⁹ The results of the present study suggest there may be potential to reverse the impact of these adverse sleep behaviours and disorders on glucose metabolism through CBT-I and/ or sleep hygiene interventions. The HbA1c reduction of 0.43% in T2DM from our meta-analysis is larger than that of other sleep interventions such using CPAP in people with T2DM and OSA of 0.24% (95% CI 0.06, 0.43), according to a recent meta-analysis.⁴⁰ Furthermore, meta-analyses of lifestyle interventions focusing on diet and/ or lifestyle in T2DM has shown variable impact on HbA1c reduction ranging between 0.37% (0.14, 0.59) to 0.82% (0.53, 1.11).^8,9

Regarding limitations of the study, the quality of methods used in studies was variable as most studies were pilot trials. For example, one RCT of people with T2DM was conducted in a real-world setting and collected final follow-up data within 6 months of the first appointment, based on whenever the participant would return for their next clinical appointment.¹⁰ Secondly, only some studies accounted for those people with T2DM who changed their diabetes medications during the trial. However, one reason for this may be because many studies were conducted over shorter periods in which medications could not have been changed from baseline. Thirdly, some studies analysed their data as per protocol and others as an intention to treat, only three studies provided both sets of data.^33,38,41 Fourthly, for sleep extension studies used in the meta-analyses, some data was only available as mean change in the variable, however, this was also observed in two previous meta-analyses in this area.^6,7

The strengths of this study are the systematic approach and robust methods used to obtain our results, ensuring they met accepted standards for meta-analyses. We considered a wide range of sleep behaviours and disorders as causes for insufficient sleep, as well as glucose related terms, which were used in the comprehension list of search terms in the search strategy. We analysed full-text articles and some abstracts for studies, featuring people with and without T2DM, allowing us to detect as many eligible studies as possible.

Regarding potential implications for clinical practice from the present study, we found use of CBT-I) and/or sleep hygiene in people with T2DM was associated with an HbA1c reduction of 0.43%, which can be considered clinically meaningful. As we excluded studies of OSA as this disease has a specific treatment (CPAP), detecting individuals with other forms of sleep insufficiency, followed by implementing interventions to improve their sleep, could be a cost-effective approach to improving glycaemic control without use of medications. Commonly used sleep assessment tools include the Insomnia Score Index and Pittsburgh Sleep Quality Index.^42,43 Sleep interventions could be a useful adjunct to lifestyle advice and recommendations which traditionally focus on modification of diet and physical activity only. However more research is required to answer these questions and also if CBT-I and/or sleep hygiene can aid the prevention of T2DM in high-risk people with pre-diabetes.

In conclusion, use of CBT-I and/ or sleep hygiene interventions led to a clinically significant reduction in HbA1c levels of 0.43% in people with T2DM and insufficient sleep. Future research needs to explore the role of using sleep interventions in sleep insufficiency as a potential target to aid current measures of preventing T2DM complications and also in prevention of this disease in high risk people with pre-diabetes.

Supplemental Material

Supplemental Material - The effect of non-pharmacological sleep interventions on glycaemic measures in adults with sleep disturbances and behaviours: A systematic review and meta-analysis

Supplemental Material for The effect of non-pharmacological sleep interventions on glycaemic measures in adults with sleep disturbances and behaviours: A systematic review and meta-analysis by Samiul A Mostafa, Wasim Hanif, Francesca Crowe, George Balanos, Krishnarajah Nirantharakumar, Jason G. Ellis and, Abd A Tahrani in Diabetes & Vascular Disease Research

Footnotes

Acknowledgements

We would like to thank University Hospitals of Birmingham (a) Library team (especially Jennifer Davis-Manders and Anita Phul) for their support with developing and modifying the final search terms, conducting the searches and also for providing access to their systemic review training courses and resources (b) for statistical support through Mr James Hodgson, research statistician.

Author contributions

SAM – conceived the idea for the review, designed the study, developed the search strategy, conducted the literature searches, screened articles for inclusion, selected and retrieved relevant papers, determined inclusion/exclusion of papers, conducted data extraction, performed the analysis, performed the critical analysis and wrote the manuscript. SM is the guarantor of the data. WH - screened articles for inclusion, selected and retrieved relevant papers, determined inclusion/exclusion of papers, verified data extraction and drafted the manuscript. FC – drafted the manuscript. GB - drafted the manuscript. KN - drafted the manuscript. JE - drafted the manuscript. AAT - conceived the idea for the review, designed the study, developed the search strategy, supervised the meta-analyses and drafted the manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Declarations

SM undertook this research work through receiving a Clinical Research Network West Midlands Health and Care Research Scholarship. He has previously received research support from Novo Nordisk Research Foundation UK and Academy of Medical Sciences. AAT is currently an employee and share-holder of Novo Nordisk. Novo Nordisk had no role in this manuscript. The views expressed in this manuscript are those of the authors and not Novo Nordisk.

ORCID iD

Samiul A Mostafa

Supplemental Material

Supplemental material for this article is available online.

Appendix

References

Spiegel

Tasali

Leproult

, et al. Effects of poor and short sleep on glucose metabolism and obesity risk. Nat Rev Endocrinol 2009; 5: 253–261. DOI: 10.1038/nrendo.2009.23.

Cappuccio

D’Elia

Strazzullo

, et al. Quantity and quality of sleep and incidence of type 2 diabetes: a systematic review and meta-analysis. Diabetes Care 2010; 33: 414–420.

Mostafa

Campos

Antza

, et al. Sleep behaviours and associated habits and the progression of pre-diabetes to type 2 diabetes mellitus in adults: a systematic review and meta-analysis. Diabetes Vasc Dis Res 2022; 19(3): 14791641221088824.

Lee

SWH

Chin

. The impact of sleep amount and sleep quality on glycemic control in type 2 diabetes: a systematic review and meta-analysis. Sleep Med Rev 2017; 31: 91–101. DOI: 10.1016/j.smrv.2016.02.001.

Type 2 diabetes in adults: management. London, UK: National Institute for Health and Care Excellence. Available at: https://www.nice.org.uk/guidance/ng28

Kothari

Cardona

Chirakalwasan

, et al. Sleep interventions and glucose metabolism: systematic review and meta-analysis. Sleep Med 2021; 78: 24–35. DOI: 10.1016/j.sleep.2020.11.035.

Zhu

Yin

Shi

, et al. Feasibility of sleep extension and its effect on cardiometabolic parameters in free-living settings: a systematic review and meta-analysis of experimental studies. Eur J Cardiovasc Nurs 2022; 21(1): 9–25. DOI: 10.1093/eurjcn/zvab055.

Chen

Pei

Kuang

, et al. Effect of lifestyle intervention in patients with type 2 diabetes: a meta-analysis. Metabolism 2015; 64(2): 338–347. DOI: 10.1016/j.metabol.2014.10.018.

Schwingshackl

Chaimani

Hoffmann

, et al. A network meta-analysis on the comparative efficacy of different dietary approaches on glycaemic control in patients with type 2 diabetes mellitus. Eur J Epidemiol 2018; 33: 157–170. DOI: 10.1007/s10654-017-0352-x.

10.

Ashra

Spong

Carter

, et al. A systematic review and meta- analysis assessing the effectiveness of pragmatic lifestyle interventions for the prevention of type 2 diabetes mellitus in routine practice. England: Public Health England, 2015. Available at: https://assets.publishing.service.gov.uk/media/5b6c484ae5274a2967e4cfc0/PHE_Evidence_Review_of_diabetes_prevention_programmes-_FINAL.pdf.

11.

Liberati

Altman

Tetzlaff

, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009; 339: b2700.

12.

Higgins

Altman

Gotzsche

, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011; 343: d5928.

13.

Higgins

JPT

Deeks

(editors). Chapter 6: choosing effect measures and computing estimates of effect. In: Higgins

JPT

Thomas

Chandler

, et al. (editors). Cochrane handbook for systematic reviews of interventions version 6.4. London, UK: Cochrane, 2023. Available at: https://www.training.cochrane.org/handbook.

14.

Higgins

Thompson

. Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 21: 1539–1558.

15.

Alshehri

Alenazi

Alothman

, et al. Using cognitive behavioral therapy for insomnia in people with type 2 diabetes, pilot RCT Part I: sleep and concomitant symptom. Behav Sleep Med 2021; 19(5): 652–671. DOI: 10.1080/15402002.2020.1831501.

16.

Alshehri

Alothman

Alenazi

, et al. The effects of cognitive behavioral therapy for insomnia in people with type 2 diabetes mellitus, pilot RCT part II: diabetes health outcomes. BMC Endocr Disord 2020 5; 20(1): 136. DOI: 10.1186/s12902-020-00612-6.

17.

Cheung

Woo

Cheung

, et al. Legacy effects of sleep education programme in people with type 2 diabetes and short sleep duration. Diabetologia 2022; 65(1): S337–p654, Available at: https://www.easd.org/media-centre/home.html#!resources/legacy-effects-of-sleep-education-programme-in-people-with-type-2-diabetes-and-short-sleep-duration-a-prospective-follow-up-study

18.

García-Serrano

Pujol Salud

Aran-Solé

, et al. Enhancing night and day circadian contrast through sleep education in prediabetes and type 2 diabetes mellitus: a randomized controlled trial. Biology 2022. 11(6): 893. DOI: 10.3390/biology11060893.

19.

Gunduz

Gorpelioglu

Emiroglu

, et al. The effect of training about sleep hygiene on HbA1c levels of type 2 diabetes mellitus patients: a randomized controlled trial. Clin Diabetol 2022; 11(5): 303–308. DOI: 10.5603/DK.a2022.0046.

20.

Tang

, et al. Effect of diabetes sleep education for T2DM who sleep after midnight: a pilot study from China. Metab Syndr Relat Disord 2018; 16(1): 13–19. DOI: 10.1089/met.2017.0069.

21.

Pandey

Littlewood

Cooper

, et al. Pragmatic approaches for integrating healthy sleep hygiene to promote chronic disease management. Sleep 2018; 41: A391.

22.

Pereira

Trevisan

Lourenco

, et al. Effects of educational strategy on sleep hygiene and light exposure on sleep quality, emotional stress and glycemic control in type 2 diabetes mellitus: randomized controlled trial. Diabetol Metab Syndrome 2018; 10.

23.

Tannas

Davis

. Type 2 diabetes and insomnia: impact on metabolic control. Sleep 2013; 36: A219, Available at: https://digitalcommons.wayne.edu/oa_dissertations/624/

24.

Zhang

Chang

, et al. Effectiveness of cognitive behavior therapy for sleep disturbance and glycemic control in persons with type 2 diabetes mellitus: a community-based randomized controlled trial in China. World J Diabetes 2021; 12(3): 292–305. DOI: 10.4239/wjd.v12.i3.292.

25.

Zuo

Dong

Zhang

, et al. Effects of cognitive behavioral therapy on sleep disturbances and quality of life among adults with type 2 diabetes mellitus: a randomized controlled trial. Nutr Metabol Cardiovasc Dis 2020; 30(11): 1980–1988. DOI: 10.1016/j.numecd.2020.06.024.

26.

Cain

Brumley

Louis-Jacques

, et al. A pilot study of a sleep intervention delivered through group prenatal care to overweight and obese women. Behav Sleep Med 2019; 18(4): 477–487.

27.

Carroll

Seeman

Olmstead

, et al. Improved sleep quality in older adults with insomnia reduces biomarkers of disease risk: pilot results from a randomized controlled comparative efficacy trial. Psychoneuroendocrinology 2015; 55: 184e92.

28.

Chesla

Gay

Bauer

, et al. Preventing type 2 diabetes among midlife women: feasibility study comparing a combined sleep & lifestyle intervention with a standard lifestyle intervention. Diabetes Educat 2020; 46(5): 424–434.

29.

Duncan

Fenton

Brown

, et al. Efficacy of a multi-component m-health weight-loss intervention in overweight and obese adults: a randomised controlled trial. Int J Environ Res Publ Health 2020; 17(17): 6200. DOI: 10.3390/ijerph17176200.

30.

Al Khatib

Hall

Creedon

, et al. Sleep extension is a feasible lifestyle intervention in free-living adults who are habitually short sleepers: a potential strategy for decreasing intake of free sugars? A randomized controlled pilot study. Am J Clin Nutr 2018; 107(1): 43e53.

31.

Chennaoui

Arnal

Drogou

, et al. Sleep extension. increases IGF-I concentrations before and during sleep deprivation in healthy. young men. Appl Physiol Nutr Metabol 2016; 41: 963–970.

32.

Cizza

Piaggi

Rother

, et al. Hawthorne effect with transient behavioral and biochemical changes in a randomized controlled sleep extension trial of chronically short-sleeping obese adults: implications for the design and interpretation of clinical studies. PLoS One 2014; 9(8): e104176.

33.

Hartescu

Stensel

Thackray

, et al. Sleep extension and metabolic health in male overweight/ obese short sleepers: a randomised controlled trial. J Sleep Res 2022; 31: e13469. DOI: 10.1111/jsr.13469.

34.

Killick

Hoyos

Melehan

, et al. Metabolic and hormonal effects of “catch-up” sleep in men with chronic, repetitive, lifestyle-driven sleep restriction. Clin Endocrinol 2015; 83: 498–507.

35.

Leproult

Deliens

Gilson

, et al. Beneficial impact of sleep extension on fasting insulin sensitivity in adults with habitual sleep restriction. Sleep 2015; 38: 707–715.

36.

Mathew

Nahmod

Master

, et al. Effects of a 1-hour per night week-long sleep extension in college students on cardiometabolic parameters, hydration status, and physical activity: a pilot study. Sleep Health. 2023 22(23):2352-7218. doi: 10.1016/j.sleh.2023.10.006.

37.

Reutrakaul

Nemeth

Quinn

, et al. Effects of Sleep-Extend on glucose metabolism in women with a history of gestational diabetes: a pilot randomized trial. Pilot Feasibility Stud. 2022; 8: 119. DOI: 10.1186/s40814-022-01076-2.

38.

So-ngern

Chirakalwasan

Saetung

, et al. Effects of two-week sleep extension on glucose metabolism in chronically sleep-deprived individuals. J Clin Sleep Med 2019; 15(5): 711–718.

39.

Antza

Kostopoulos

Mostafa

, et al. The links between sleep duration, obesity and type 2 diabetes mellitus. J Endocrinol 2021; 252(2): 125–141. DOI: 10.1530/JOE-21-0155.

40.

Herth

Sievi

Schmidt

, et al. Effects of continuous positive airway pressure therapy on glucose metabolism in patients with obstructive sleep apnoea and type 2 diabetes: a systematic review and meta-analysis. Eur Respir Rev 2023; 32(169): 230083. DOI: 10.1183/16000617.0083-2023.

41.

Groeneveld

Beulens

Blom

, et al. The effect of cognitive behavioral therapy for insomnia on sleep and glycemic outcomes in people with type 2 diabetes: a randomized controlled trial. Sleep Med 2024; 120: 44–52. DOI: 10.1016/j.sleep.2024.05.029.

42.

Gagnon

B´elanger

Ivers

, et al. Validation of the insomnia severity Index in primary care. J Am Board Fam Med 2013; 26(6): 701–710.

43.

Buysse

Reynolds

Monk

, et al. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatr Res 1989; 28: 193–213.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.60 MB