Received: 30 May 2019

| Revised: 9 January 2020

| Accepted: 22 January 2020

DOI: 10.1111/eci.13202

ORIGINAL ARTICLE

Exercise training improves sleep quality: A randomized

controlled trial

Lucas Jurado-Fasoli1,2 | Alejandro De-la-O1 | Cristina Molina-Hidalgo1 |

Jairo H. Migueles2 | Manuel J. Castillo1 | Francisco J. Amaro-Gahete1,2

1
Departament of Medical Physiology,
School of Medicine, University of Granada,
Abstract
Granada, Spain Background: Exercise holds promise as a non-pharmacological intervention for the
2
PROmoting FITness and Health through improvement of sleep quality. Therefore, this study investigates the effects of differ-
Physical Activity Research Group
ent training modalities on sleep quality parameters.
(PROFITH), Department of Physical
Education and Sports, Faculty of Sport Material & methods: A total of 69 (52.7% women) middle-aged sedentary adults
Sciences, University of Granada, Granada, were randomized to (a) control group, (b) physical activity recommendation from the
Spain
World Health Organization, (c) high-intensity interval training (HIIT) and (d) high-
Correspondence intensity interval training group adding whole-body electromyostimulation training
Lucas Jurado-Fasoli, Department of (HIITEMS). Sleep quality was assessed using the Pittsburgh Sleep Quality Index
Medical Physiology, University of Granada,
Av. Investigation 11, Granada, Spain.
(PSQI) scale and accelerometers.
Email: juradofasoli@ugr.es Results: All intervention groups showed a lower PSQI global score (all P < .022).
HIIT-EMS group improved all accelerometer parameters, with higher total sleep
Funding information
Spanish Ministry of Education, Grant/ time and sleep efficiency, and lower wake after sleep onset (all P < .016). No differ-
Award Number: FPU14/04172 and ences were found between groups in any sleep quality parameter.
FPU15/03960; University of Granada
Conclusion: In conclusion, exercise training induced an improvement in subjective
sleep quality in sedentary middleaged adults. Moreover, HIIT-EMS training showed an
improvement in objective sleep quality parameters (total sleep time, sleep efficiency and
wake after sleep onset) after 12 weeks of exercise intervention. The changes observed
in the HIIT-EMS group were not statistically different to the other exercise modalities.

KEYWORDS
concurrent training, high-intensity interval training, sleep quality, sleep quantity, whole-body
electromyostimulation

1 | IN T RO D U C T ION as age increases.2 Sleep disorders have a negative influence

on mental and physical health and decreasing quality of life,
Sleep is an essential physiological process with important which increases healthcare costs.3 As examples, sleep disor-
recovery functions.1 Notable quantitative and qualitative ders seem to increase the risk of depression/anxiety, cardio-
changes in sleep occur with age.1 There is growing evidence vascular disease, stroke and overall mortality.4
pointing out a reduction in the duration of sleep concomi- Several interventions have been proposed to attenuate the
tantly with an increase in the prevalence of sleep disorders prevalence and/or consequences of age-associated sleep dis-
orders including pharmacological agents,5 chronotherapy,6
Castillo and Amaro-Gahete equally contributed to this work.

© 2020 Stichting European Society for Clinical Investigation Journal Foundation.

Eur J Clin Invest. 2020;50:e13202.  wileyonlinelibrary.com/journal/eci | 1 of 11

https://doi.org/10.1111/eci.13202
 |

13652362, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/eci.13202 by Universidad Rey Juan Carlos C/Tulipan S/N Edificio, Wiley Online Library on [21/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 of 11    JURADO-FASOLI et al.

stimulus control therapy,6 relaxation therapy,6 dietary advice7 2 | M ETHODS

and physical exercise.3 Previous studies have reported that
non-pharmacological interventions are as effective as pharma- 2.1 | Participants
cological interventions3 without accounting with the pharma-
cological-related negative effects.4 In this sense, a bidirectional Eighty middle-aged adults (40 women, 40 men), aged
relationship between physical activity and sleep has been sug- 45-65 years, were enrolled in the FIT-AGEING study, an
gested.8 Therefore, exercise holds promise as a non-pharmaco- exercise-based randomized controlled trial (clini​caltr​ial.gov:
logical intervention for adults with poor quality or disordered ID: NCT03334357).11 The participants were recruited from
sleep.8 the province of Granada (Spain) using social networks, local
A previous systematic review reported that exercise train- media and posters. Interested individuals were screened via
ing provides a positive effect on sleep quality by decreasing telephone and/or e-mail. Inclusion criteria were to report
sleep latency and the use of sleep medication in middle-aged being sedentary (ie <20 minutes of moderate-intensity physi-
and older adults.3 Moderate-intensity exercise (both aerobic cal activity on 3 d/wk over the last 3 months), and to have a
and resistance-based trainings) has shown positive effects stable weight over the last 6 months. All participants reported
on sleep quality in middle-aged and young adults.9,10 A to be free of disease, pregnant or lactating women, to not take
systematic review conducted by Yang et al3 observed that any medication and/or not presence of a major illness that
the participation in an exercise training programme (mod- would limit the ability to perform the training programme.
erate-intensity aerobic exercise or resistance exercise) had The study was approved by the Ethics Committee on Human
positive effects on sleep quality in middle-aged and older Research at the University of Granada and “Servicio Andaluz
adults. de Salud” (CEI-Granada) [0838-N-2017], and all participants
The high-intensity interval training has risen as a novel signed an informed consent. The study protocol and experi-
training modality in the last years, but its influence on sleep mental design were applied in accordance with the last re-
quality has not been extensively studied. However, there vised ethical guidelines of the Declaration of Helsinki. All
are not enough studies that investigated the effects of the of the baseline and follow-up examinations were performed
high-intensity interval training on sleep quality parameters. at the same setting (Instituto Mixto Universitario Deporte y
A new training tendency, sometimes coupled to high-inten- Salud [iMUDS] at the University of Granada).
sity training, is whole-body electromyostimulation, a training
methodology which stimulates between 14 and 18 regions or
between 8 and 12 different muscle groups up to 2.800 cm2 2.2 | Study design
electrode area.11 However, little is known about its effects on
physiological parameters.11 A 12-week randomized controlled trial was conducted
To our knowledge, there are no studies that evaluate the ef- with a parallel group design following the CONSORT
fects of those two novel training methods such as high-inten- (Consolidated Standards of Reporting Trials) guidelines12
sity interval training and whole-body electromyostimulation (Table S1). The study was performed between the months
on sleep quality and quantity in sedentary middle-aged adults. of September and December of 2016 and 2017. After the
Moreover, there is a lack of studies that compare the influence baseline examination, participants were randomized using
of different exercise training programmes (ie concurrent training a computer-generated simple randomization13 to four dif-
[combination of endurance and resistance training] vs high-in- ferent groups: (a) control group (no exercise), (b) PAR
tensity interval training vs high-intensity interval training add- group, (c) HIIT group and (d) HIIT-EMS group. The par-
ing whole-body electromyostimulation) on sleep quality and ticipant's randomization assignment was blinded to the
quantity in sedentary middle-aged adults. Therefore, this study assessment staff. All participants were requested to main-
aimed to evaluate the effects of different exercise training pro- tain their habitual dietary habits. Individuals assigned to
grammes (ie (a) a concurrent training based on physical activity the control group were also requested not to change their
recommendations from the World Health Organization group physical activity habits or to engage in any kind of physi-
[PAR], (b) a high-intensity interval training group [HIIT] and cal training programme. Individuals in the exercise groups
(c) a high-intensity interval training group adding whole-body were instructed to not perform additional exercise to their
electromyostimulation group [HIIT-EMS]) on sleep quality intervention programmes.
and quantity parameters in sedentary middle-aged adults. We
hypothesize that all exercise training programmes will improve
sleep quality and quantity parameters with higher improvements 2.3 | Training modalities
in the HIIT-EMS group. We also aimed to study which covari-
ates could explain changes observed in sleep quality and quantity To maximize transparency and replicability, the exer-
parameters. cise programme described in this manuscript follows the
 |

13652362, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/eci.13202 by Universidad Rey Juan Carlos C/Tulipan S/N Edificio, Wiley Online Library on [21/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
JURADO-FASOLI et al.    3 of 11

Consensus on Exercise Reporting Template (CERT; Table sport sciences. The training sessions took place in an airy,
S2).14 Reporting of the study conforms to CONSORT- well-lighted and well-equipped gym of the iMUDS at the
revised statement along with references to CONSORT- University of Granada (Spain). The starting level was indi-
revised statement and the broader EQUATOR guidelines.15 vidualized in each training group. All training modalities
A detailed description of each training modality can be were delivered as planned. There were no changes in trial
found elsewhere.11 outcomes after the trial commencement.
Participants in the PAR group performed a concurrent Attendance at the training sessions was registered daily,
training based on the minimum physical activity recom- and participants were contacted upon any missing session to
mended by the World Health Organization.16 PAR group fre- ask for the reason and motivate them to replace it on an al-
quency was 3 d/wk. The training volume was 150 min/wk at ternative session. There were no home-based or non-exercise
60%-65% of the heart rate reserve for the endurance training components within this intervention. The study was ended
and ~60 min/wk, at a 40%-50% of one-repetition maximum when the training intervention programme was finished.
for the resistant training. The exercises programmed for the A graduate in sport sciences provided general advice to
endurance training section were treadmill, cycle ergometer the control group through an information meeting. They were
and elliptical ergometer. And, weight-bearing and guided instructed to maintain their lifestyle and to not partake in any
pneumatic machines were used in resistance training section training programme during the time of the study.
(ie squat, bench press, dead lift or lateral pull down).
High-intensity interval training group did an intervention
programme characterized by short and intermittent efforts of 2.4 | Sleep quality and quantity assessment
vigorous activity, interspersed with resting periods at passive
or low-intensity exercises. Participants in the HIIT group The sleep quality and quantity were assessed before
trained two sessions/week performing two different comple- (September) and after (December) the training programme
mentary protocols17: (a) high-intensity interval training with (week 12) using the Pittsburgh Sleep Quality Index (PSQI)
long intervals (type A session) and (b) high-intensity interval scale18 and accelerometer-based estimates (see below).
training with short intervals (type B session). The training The PSQI is a self-report tool which consists of 19-item
volume was 40-65 min/wk at >95% of the maximum oxygen scale that provides seven component scores (ranges 0-3): (a)
uptake in type A session, and >120% of the maximum oxy- subjective sleep quality (very good to very bad), (b) sleep
gen uptake in type B session. Treadmill with a personalized latency (≤15 to >60 minutes), (c) sleep duration (≥7 to
slope was chosen for type A session and eight weight-bearing <5 hours), (d) sleep efficiency (≥85% to <65% hours sleep/
exercises in circuit form (ie squat, dead lift, high knees up, hours in bed), (e) sleep disturbances (not during the past
high heels up, push up, horizontal row, lateral plank and fron- month to ≥3 times per week), (f) use of sleeping medications
tal plank) for type B session. (none to ≥3 times a week), and (g) daytime dysfunction (not
High-intensity interval training group adding whole- a problem to a very big problem), with a total global score
body electromyostimulation group performed a training pro- ranging from 0 to 21.18 A PSQI global score higher than 5
gramme that followed the same structure as HIIT (volume, indicates poor sleep quality.18
intensity, training frequency, type of exercise, and training Objective characteristics of sleep-wake cycles were
sessions) with the addition of electrical impulses. Bipolar, monitored with a wrist-worn accelerometer (ActiSleep,
symmetrical and rectangular electric pulse was applied with ActiGraph) for 7 consecutive days (24 h/d).11 Participants
(a) a frequency of 15-20 Hz in type A sessions, and 35-75 received detailed information on how to wear the acceler-
hertz in type B sessions; (b) an intensity of 100 mA in type A ometer and were asked to remove it only for water activities.
sessions, and 80 milliamps in type B sessions; (c) an impulse They also recorded the times they went to bed every night,
breadth of 200-400 µsec; and (d) a duty cycle (ratio of on- woke up every morning and removed the device every day.
time to the total cycle time: % duty cycle = 100/[total time/ The accelerometers used an epoch length of 5 seconds and
on-time]) of 99% in type A sessions, and 50%-63% in type B a frequency rate of 100 Hz to store raw accelerations.19
sessions. A whole-body electromyostimulation device manu- The raw accelerations were exported in “.csv” format using
factured by Wiemspro® was used. ActiLife v. 6.13.3 software (ActiGraph) and processed
All sessions started with a dynamic standardized warm-up using the GGIR package (v. 1.6-0, https​ ://cran.r-proje​
that included general mobility exercises and ended with a ct.org/web/packa​ges/GGIR/index.html)20 in R (v. 3.1.2,
cooling-down protocol (active global stretching), which alter- https​://www.cran.r-proje​ct.org/). We derived the Euclidean
nated five posterior chain exercises with five anterior chain Norm Minus One G (ENMO) as √(x2 + y2 + z2)−1G
exercises.11 Each training group followed a gradual progres- (where 1G ~9.8 m/s2) with negative values rounded to zero
sion in order to control the exercise dose.11 All training ses- to describe physical activity and accelerometer's z angle
sions were performed in group supervised by a graduate in to describe sleep patterns. We used a previously published
 |

13652362, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/eci.13202 by Universidad Rey Juan Carlos C/Tulipan S/N Edificio, Wiley Online Library on [21/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4 of 11    JURADO-FASOLI et al.

algorithm combining data from the accelerometers and in FIT-AGEING study. There were no interim analyses dur-
diary reports to detect sleep period time.21,22 According ing the study.
to this algorithm, sleep was defined as any period of sus- Shapiro-Wilk test, visual check of histograms, and Q-Q
tained inactivity, in which there was minimal changes in plots were used to verify the distribution of all variables.
the arm angle (ie as much 5 degrees for 5 minutes peri- Descriptive characteristics of the sample are reported as
ods), during a period recorded as sleep by the participant mean and standard deviation.
in their diary reports.21 The following variables were ana- We conducted an analysis of variance to determine differ-
lysed: total sleep time (minutes slept between bedtime and ences in all variables between groups at the baseline.
wake time), sleep efficiency (percentage of time asleep Repeated-measures analysis of variance was used to de-
while in bed) and wake after sleep onset (minutes awake termine changes in PSQI global score, total sleep time, sleep
between sleep onset and wake time). To note that only the efficiency, wake after sleep onset and PSQI sub-scores across
participants wearing the accelerometers for ≥16 h/d during time, between groups and its interaction (time × group).
at least 4 days (including at least 1 weekend day) were in- Student's t tests for paired values were performed to evaluate
cluded in the analyses.19 differences in dependent variables before and after the inter-
vention programme.
We found sex interaction in total sleep time outcome;
2.5 | Covariates hence, we repeated the previous analyses segmented by sex.
Analysis of covariance (ANCOVA) was used to exam-
We assessed anthropometric and body composition through ine the effect of groups (fixed factor) on sleep quality and
dual-energy X-ray absorptiometry. Cardiorespiratory fitness quantity parameters changes, that is post-PSQI global score
was assessed through a maximum treadmill exercise test minus pre-PSQI global score (dependent variable), adjusting
following the modified Balke protocol,23 and a digital hand for baseline values. The same analyses were performed for
dynamometer was used to assess hand grip strength. We col- changes in total sleep time, sleep efficiency, wake after sleep
lected blood samples and measured somatotropin levels (see onset and PSQI sub-scores.
Appendix S1 for more information). All analyses were adjusted by sex, age, and sex and age.
We performed Bonferroni post hoc tests with adjustment for
multiple comparisons to determine differences between all
2.6 | Statistical analysis exercise modality groups.
We conducted linear regression analysis to examine
The sample size and power calculations are made based on the relationship between changes in sleep variables (PSQI
the data of a randomized control trial (The FIT-AGEING global score, total sleep time, sleep efficiency and wake
project11; clini​
caltr​
ial.gov: ID: NCT03334357). The prin- after sleep onset) and changes in body composition vari-
cipal aim of the FIT-AGEING study was to determine the ables (body mass index, lean mass, lean mass index, fat
effect of different training modalities on physiological pa- mass, fat mass index, bone mineral density), physical fit-
rameters (ie body composition and sleep quality and quantity ness (VO2 max., VO2 max., relative and total hand grip)
among others) in sedentary healthy adults. The determina- and somatotropin levels, and we conducted simple linear
tion of the sample size and power of the study were made regressions.
based on the data of a pilot sample (n = 30). We considered Considering that we aimed at assessing efficacy, we con-
different physiological parameter (ie body composition and ducted a primary analyses per-protocol, in which we excluded
sleep quality and quantity among others) differences between participants who did not finish the intervention programme
pre- and post-treatment in order to assess the sample size re- and/or did not reach a minimum of 70% of attendance. To
quirements for the one-way analysis of variance. As a result, check the robustness of our results, we performed the fol-
we expect to detect a clinically relevant effect size of each lowing sensitivity analysis: baseline carried forward (BOCF)
variable considering a type I error of 0.05 with a statistical imputation.
power of 0.85. To meet these criteria, a minimum of 14 par- All analyses were conducted using the Statistical Package
ticipants per group were necessary. Assuming a maximum for Social Sciences (SPSS, v. 25.0, IBM SPSS Statistics,
loss at follow-up of 25%, we decided to recruit 20 partici- IBM Corporation), and the level of significance was set at
pants (≈50% women) for each study group. Therefore, a total <.05. Graphical presentations were prepared using GraphPad
of ~80 participants (≈40 women and ≈40 men) were enrolled Prism 5 (GraphPad Software).

F I G U R E 1 Flow chart diagram. BMI, body mass index; BOCF, baseline observation carried forward imputation; CDV, cardiovascular;
ECG, electrocardiogram; HIIT, high-intensity interval training group; HIIT-EMS, high-intensity interval training group adding whole-body
electromyostimulation group; PAR, physical activity recommendations for adults proposed by the World Health Organization group
 |

13652362, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/eci.13202 by Universidad Rey Juan Carlos C/Tulipan S/N Edificio, Wiley Online Library on [21/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
JURADO-FASOLI et al.    5 of 11

Captation flow:
The FIT-AGEING study

247 Expressed interest to participate in FIT-AGEING

FIT-AGEING sends information about meetings

224 Confirmed attendance to an information meeting

6 Information meetings

196 Attended an information meeting

29 Declined to participate

167 Requested the pre-screening questionnaire

26 Did not answer

141 Assessed for eligibility

13 Age or BMI out of the range

5 History of CDV disease
2 Diabetes
4 Hyper-hypothyroidism
108 Met pre-screening inclusion criteria 9 Others
Total = 33

6 Declined to participate

102 Eligible for baseline assessment

3 Abnormal Exercise ECG

4 Medical exclusion found
during examination
15 Declined to participate
Total = 22

N = 80 Randomized

N = 20 Control N = 20 PAR N = 20 HIIT N = 20 WB-EMS

N = 5 Lost to follow-up N = 3 Lost to follow-up N = 2 Lost to follow-up N = 1 Lost to follow-up

N = 15 Completed trial N = 17 Completed trial N = 18 Completed trial N = 19 Completed trial

N = 15 (9 women) N = 17 (9 women) N = 18 (9 women) N = 19 (9 women)

Included in primary Included in primary Included in primary Included in primary
analysis analysis analysis analysis

N = 20 (12 women) N = 20 (11 women) N = 20 (11 women) N = 20 (10 women)

Included in BOCF Included in BOCF Included in BOCF Included in BOCF
analysis analysis analysis analysis
 |

13652362, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/eci.13202 by Universidad Rey Juan Carlos C/Tulipan S/N Edificio, Wiley Online Library on [21/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6 of 11    JURADO-FASOLI et al.

3 | R ES U LTS 3.53 ± 2.53, P = .003; 5.56 ± 2.73 vs 3.44 ± 2.58, P = .022;

for PAR, HIIT and HIIT-EMS, respectively), while no
The flow chart of the current study is presented in Figure differences were observed in the control group. HIIT-
1. Eleven participants were lost at follow-up (control group: EMS group showed significantly higher total sleep time
5; PAR: 3; HIIT: 2; HIIT-EMS: 1). Data from PSQI were (338.21 ± 47.97 vs 388.83 ± 37.16 minutes, P = .004),
missed in six participants. A total of 63 participants were in- higher sleep efficiency (82.66 ± 6.83% vs 87.98 ± 3.76%,
cluded in the analysis for PSQI and 69 for objective sleep P = .004) and lower wake after sleep onset (72.03 ± 30.82
outcomes. vs 54.46 ± 19.39 minutes, P = .016) after the interven-
Table 1 shows the participant's baseline characteristics. tion programme compared to the baseline, while no sig-
We only observed differences among groups in the com- nificant differences were found in the control group, PAR
ponent 7 of the PSQI questionnaire. No differences were group and HIIT group after the intervention programme
observed in the remaining variables. Participants attended (all P > .084). Time × group interaction was found in total
to 98.7% of their exercise sessions. There were no adverse sleep time (P = .047), sleep efficiency (P = .017) and wake
events occurring during the exercise sessions. after sleep onset (P = .027).
Figure 2 shows PSQI global score (Figure 2A), total Figure 3 shows changes in PSQI global score (Figure 3A),
sleep time (Figure 2B), sleep efficiency (Figure 2C) and total sleep time (Figure 3B), sleep efficiency (Figure 3C)
wake after sleep onset (Figure 2D) before and after the in- and wake after sleep onset (Figure 3D) after the intervention
tervention study. When comparing within-group changes, study among the four groups. No statistically significant in-
all intervention groups showed a lower PSQI global tergroup differences were observed in these variables when
score in the final measurement compared to the baseline we performed the post hoc analyses (all P > .096; Figure 3),
(4.81 ± 3.85 vs 3.06 ± 2.57, P = .013; 5.47 ± 3.74 vs neither controlling by sex nor age (all P > .053; Table 2).

TABLE 1 Descriptive characteristics of the sample

Control PAR HIIT HIIT-EMS

N All (n = 69) (n = 15) (n = 17) (n = 18) (n = 19) P
Age (y) 69 53.4 ± 5.0 51.7 ± 4.1 54.9 ± 4.5 53.1 ± 5.6 53.4 ± 5.4 .716
Sex (%)
Men 33 47.8 40.0 47.1 50.0 52.6 .881
Women 36 52.2 60.0 52.9 50.0 47.4
Body composition parameters
Body mass index (kg/m2) 69 26.8 ± 3.8 26.7 ± 3.9 25.4 ± 2.9 26.4 ± 3.1 28.1 ± 4.7 .063
2
Lean mass index (kg/m ) 69 15.4 ± 2.8 15.9 ± 3.1 15.2 ± 2.5 14.9 ± 2.9 16.0 ± 2.9 .775
2
Fat mass index (kg/m ) 69 10.7 ± 3.1 10.1 ± 2.7 9.6 ± 2.7 10.8 ± 2.7 11.3 ± 3.4 .151
Sleep parameters
PSQI global score 63 5.5 ± 3.5 6.5 ± 4.0 4.8 ± 3.9 5.5 ± 3.7 5.5 ± 2.7 .500
Subjective sleep quality 63 1.13 ± 0.83 1.08 ± 0.95 0.81 ± 0.83 1.35 ± 0.86 1.24 ± 0.66 .266
(Component 1)
Sleep latency (Component 2) 63 1.03 ± 0.86 1.15 ± 0.80 0.88 ± 0.72 0.94 ± 0.97 1.18 ± 0.95 .461
Sleep duration (Component 3) 63 0.98 ± 0.77 1.08 ± 0.76 0.94 ± 0.85 0.88 ± 0.78 1.06 ± 0.75 .884
Sleep efficiency (Component 4) 63 0.56 ± 0.95 0.46 ± 0.78 0.56 ± 1.03 0.65 ± 1.11 0.53 ± 0.87 .979
Sleep disturbances (Component 5) 63 1.13 ± 0.42 1.15 ± 0.38 1.00 ± 0.52 1.18 ± 0.39 1.18 ± 0.39 .557
Use sleeping medications 63 0.32 ± 0.78 0.85 ± 1.28 0.31 ± 0.79 0.06 ± 0.24 0.18 ± 0.39 .070
(Component 6)
Daytime dysfunction (Component 63 0.37 ± 0.55 0.69 ± 0.63 0.31 ± 0.48 0.41 ± 0.62 0.12 ± 0.33 .036
7)
Total sleep time (min) 69 360.1 ± 49.0 369.0 ± 56.8 362.3 ± 45.9 366.7 ± 47.6 344.8 ± 46.8 .452
Sleep efficiency (%) 69 85.1 ± 6.2 85.5 ± 6.0 84.4 ± 6.9 87.2 ± 4.9 83.6 ± 6.7 .330
Wake after sleep onset (min) 69 62.9 ± 26.5 61.9 ± 25.9 67.8 ± 29.6 53.2 ± 19.1 68.4 ± 29.4 .309
Note: Data are shown as means ± standard deviation or percentages. P value of analysis of variance between groups.
Abbreviations: HIIT, high-intensity interval training group; HIIT-EMS, high-intensity interval training group adding whole-body electromyostimulation group; PAR,
physical activity recommendations for adults proposed by the World Health Organization group; PSQI, Pittsburgh Sleep Quality Index.
 |

13652362, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/eci.13202 by Universidad Rey Juan Carlos C/Tulipan S/N Edificio, Wiley Online Library on [21/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
JURADO-FASOLI et al.    7 of 11

F I G U R E 2 Sleep parameters before and after the intervention study. P value (time, group, and interaction [time × group]) of repeated-
measures analysis of variance. *P < .05, **P < .01, Student's paired t test. Data are shown as means ± standard deviation. HIIT, high-intensity
interval training group; HIIT-EMS, high-intensity interval training group adding whole-body electromyostimulation group; PAR, physical activity
recommendations for adults proposed by the World Health Organization group; PSQI, Pittsburgh Sleep Quality Index

F I G U R E 3 Changes in sleep
parameters after the intervention study
among the four groups. Data are shown as
means ± 95% confidence interval. HIIT,
high-intensity interval training group; HIIT-
EMS, high-intensity interval training group
adding whole-body electromyostimulation
group; PAR, physical activity
recommendations for adults proposed by the
World Health Organization group; PSQI,
Pittsburgh Sleep Quality Index

When we included both sex and age in the model, sleep effi- pairwise differences among groups in PSQI global score,
ciency becomes significant (F = 2.828, P = .047, η2 = .138; total sleep time, sleep efficiency and wake after sleep onset
Table 2), whereas no statistically significant differences were (all P > .05).
observed in the remaining variables (all P > .070; Table 2). High-intensity interval training group significantly de-
All previous results persisted when we included changes in creased subjective sleep quality component score in the final
LMI and FMI in the model (data not shown). There were no measurement compared to the baseline (P = .007; Table S3).
 |

13652362, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/eci.13202 by Universidad Rey Juan Carlos C/Tulipan S/N Edificio, Wiley Online Library on [21/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8 of 11    JURADO-FASOLI et al.

TABLE 2 Changes in sleep parameters adjusted for sex, age and growth hormone, body composition parameters (fat mass,
sex and age muscle mass and bone mineral density) and physical fitness
F P value η2 parameters (cardiorespiratory fitness and muscle strength;
Table S5).
PSQI global score
Overall, the sensitivity analyses corroborated the results
Model 1 0.825 .486 .043
obtained by per-protocol analysis (Table S6).
Model 2 1.254 .299 .064
Model 3 1.231 .307 .064
Total sleep time (min) 4 | DISCUSSION
Model 1 1.516 .221 .076
Model 2 1.581 .204 .079 The primary findings of this study were that: (a) all exer-
Model 3 1.467 .234 .075 cise training programmes (PAR, HIIT and HIIT-EMS) im-
Sleep efficiency (%) proved PSQI global score in sedentary middle-aged adults;
(b) HIIT-EMS was the only group that improved objec-
Model 1 2.623 .060 .127
tive sleep quality and quantity from baseline levels (ie total
Model 2 2.724 .053 .131
sleep time, sleep efficiency and wake after sleep onset);
Model 3 2.828 .047 .138
(c) no statistical differences were observed between dif-
Wake after sleep onset (min) ferent groups in any sleep quality and quantity parameter
Model 1 2.283 .089 .111 (nor subjective, nor objective); and (d) men but not women
Model 2 2.205 .098 .107 of the exercise groups improved total sleep time after the
Model 3 2.494 .070 .122 intervention programme.
Note: P values (< .05) are in bold.
All training groups improved the PSQI global score, PAR
Model 1, baseline and sex; Model 2, baseline and age; Model 3, baseline, sex (−34.77%), HIIT (−34.85%) and HIIT-EMS (−40.71%), en-
and age. hancing therefore the subjective sleep quality. Dolezal et al9 in
a recent systematic review showed that exercise increased sub-
PAR group significantly decreased sleep latency component jective sleep quality regardless of the mode and the intensity of
score in the final measurement compared to the baseline activity. Our results agree with a previous meta-analysis which
(P = .007; Table S3). All groups significantly decreased sleep revealed that exercise training has a benefit on sleep quality in
duration component score in the final measurement com- middle-aged adults, indicated by decreases in the PSQI global
pared to the baseline (P = .025; P = .023; P = .002; P = .001 score.3 A previous meta-analytic review hypothesized that the
for control group; PAR group; HIIT group and HIIT-EMS mechanisms through an exercise programme could improve the
group, respectively; Table S3). No time × group interaction perceived sleep quality could be body temperature changes,
was found in any PSQI component score (all P > .391; Table mood changes, heart rate and heart rate variability changes,
S3). growth hormone secretion, fitness improvement and body com-
No statistically significant intergroup differences were position improvements among others.24 Therefore, the prescrip-
observed in all PSQI components scores when we per- tion of exercise could help to improve sleep quality perception
formed the post hoc analyses (All P > .05; Figure S1). The in sedentary middle-aged adults.
sleep latency component score becomes significant when However, although without significant differences, PAR
we included age (F = 3.384, P = .024, η2 = .156; Table and HIIT groups showed clinically relevant differences in
S4) and both sex and age (F = 3.308, P = .027, η2 = .155; total sleep time (3.48% and 4.96%, respectively), in sleep ef-
Table S4) in the model. No differences were observed in ficiency (2.37% and −2.43%, respectively), and in wake after
the remaining PSQI component scores when we included sleep onset (−11.9% and 30.47%, respectively). A previous
sex, age or both sex and age in the model (all P > .088). meta-analysis demonstrated that the participation in an ex-
There were no pairwise differences among groups in any ercise training programme (moderate-intensity aerobic ex-
PSQI component scores (all P > .05). ercise or high-intensity resistance exercise) did not produce
In men, the PAR and HIIT-EMS groups showed sig- improvements in objective sleep parameters in middle-aged
nificantly higher total sleep time after the intervention pro- adults, but the sleep quality perception was better.3. Regular
gramme compared to the baseline (Figure S2), whereas exercise has also demonstrated to have benefits on total sleep
no differences were observed in women in any group after time, sleep efficiency and sleep quality both subjective and
the intervention programme compared to the baseline (all objective.24
P > .164; Figure S2). The most novel contribution of this study to the field is
We did not observe associations between changes in the inclusion of a HIIT-EMS group and the effect of this
sleep quality and quantity parameters and changes in form of exercise to sleep quality parameters (subjective and
 |

13652362, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/eci.13202 by Universidad Rey Juan Carlos C/Tulipan S/N Edificio, Wiley Online Library on [21/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
JURADO-FASOLI et al.    9 of 11

objective) in sedentary middle-aged adult. The HIIT-EMS lack of blood parameters mentioned above (such as BDNF or
group showed an improvement of 14.9% in total sleep time, cytokine profile) does not allow to confirm that the exercise
6.4% in sleep efficiency, and −24.4% in wake after sleep benefits on sleep quality are due to the proposed mechanisms.
onset, enhancing therefore the objective sleep quality. These And lastly, the sample size was relatively small. More studies
improvements could be related to several physiological mech- which include the plausible mechanisms (BDNF, inflamma-
anisms: (a) the electrical muscle stimulation (EMS) produces tion, etc) are needed.
a greater growth hormone response than voluntary exercise This is the first study showing that HIIT-EMS training
in addition to voluntary muscular contractions,25 which may could be an effective tool to improve sleep quality in sed-
stimulate rapid eye movement sleep26; (b) the EMS improves entary middle-aged adults. In this sense, the HIIT-EMS
body composition parameters (fat mass, muscle mass and could be positioned as an alternative to pharmacological
bone mineral density),27 which may enhance sleep quality28; interventions for adults with poor sleep quality or sleep
and (c) the EMS ameliorates physical fitness,27 which could disorders. In this sense, the HIIT-EMS did not show any
improve sleep quality.28 However, this study did not support negative effect, strengthening it vs pharmacological treat-
such mechanisms. We did not observe association between ments. Future longitudinal studies are warranted to confirm
changes in sleep quality parameters and changes in growth these results.
hormone, body composition parameters (fat mass, muscle In conclusion, our results show that different exer-
mass and bone mineral density) and physical fitness parame- cise training methodologies induced an improvement in
ters (cardiorespiratory fitness and muscle strength). There are subjective sleep quality in sedentary middle-aged adults.
also other plausible mechanisms which were not controlled in Moreover, a significant improvement in objective sleep
the present study: the electrical muscle stimulation upregu- quality and quantity parameters (total sleep time, sleep ef-
lates the brain-derived neurotrophic factor (BDNF) in rats,29 ficiency and wake after sleep onset) was observed in the
whose levels are associated with sleep quality.30 The electri- HIIT-EMS group after 12 weeks of exercise intervention.
cal muscle stimulation shifts the cytokine profile towards an- Despite slightly greater improvements in objective sleep
ti-inflammation,31 which may have a positive effect in sleep quality and quantity parameters, the changes observed in
quality.32 And the electrical muscle stimulation resulted in an the HIIT-EMS group were not statistically different to the
increment in central nervous system fatigue,33 which could other exercise groups. However, further studies are needed
improve sleep quality.28 to confirm the observed results in individuals with similar
In our study, we did not find differences between any and different characteristics since the sample size was rela-
group in any sleep quality parameters, although all three tively small.
training groups improved their baseline values while the con-
trol group did not. Our results agree with others studies that ACKNOWLEDGEMENTS
demonstrated that the benefits of exercise in sleep quality pa- The authors would like to thank all the participants who
rameters are independent of exercise type, exercise intensity took part of the study for their time and effort. This study
and exercise duration.3,9,24 However, the lack of differences is part of a PhD Thesis conducted in the Biomedicine
between groups could be due to the underpowered sample Doctoral Studies of the University of Granada, Spain. The
size. study is supported by the Spanish Ministry of Education
It is known that there are sex differences in sleep qual- (FPU14/04172 and FPU15/03960). The study was par-
ity,34 due to the differences in physiology between sexes tially supported by the University of Granada, Plan
like hormones and menstrual cycles.35 These differences Propio de Investigación 2016, Excellence actions: Units
may explain the fact that women did not improve total of Excellence; Unit of Excellence on Exercise and Health
sleep time after the intervention programme, mainly due (UCEES). We are grateful to Dr Ángel Gutiérrez for his
to the differences in exercise physiology between men and support with all study.
women.36 For example, women have lower gains in muscle
mass,37 lower cardiorespiratory fitness38 or lower muscle CONFLICT OF INTEREST
strength39 among others. Additionally, benefits of exercise None.
appeared to be stronger for men than women in sleep pa-
rameters.24 In women, we controlled for menopausal status AUTHOR CONTRIBUTIONS
(pre or post-menopausal), in order to avoid the possible co- LJF, FAG, AOP, CMH, JHM and MCG conceived and de-
founder of female hormones, and the results remained (data signed the study; LJF, FAG, AOP and CMH acquired the
not shown). data; JHM processed the data, LJF, FAG, elaborated the sta-
Several limitations should be acknowledged. First, our re- tistical section; LJF and FAG drafted the manuscript; MCG
sults cannot be extrapolated to other populations, because the revised the manuscript; and all authors read and approved the
participants were middle-aged sedentary adults. Secondly, the final manuscript.
 |

13652362, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/eci.13202 by Universidad Rey Juan Carlos C/Tulipan S/N Edificio, Wiley Online Library on [21/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
10 of 11    JURADO-FASOLI et al.

ORCID outcomes: a systematic review and practical considerations. Sport

Lucas Jurado-Fasoli https://orcid. Med. 2017;47(9):1821-1845.
20. van Hees VT, Gorzelniak L, Dean León EC, et al. Separating
org/0000-0002-5254-1816
movement and gravity components in an acceleration signal and
implications for the assessment of human daily physical activity.
R E F E R E NC E S PLoS ONE. 2013;8(4):e61691.
1. Crowley K. Sleep and sleep disorders in older adults. Neuropsychol 21. van Hees VT, Sabia S, Anderson KN, et al. Open access method
Rev. 2011;21(1):41-53. to assess sleep duration using a Wrist-Worn accelerometer. PLoS
2. Stranges S, Tigbe W, Thorogood M, et al. Sleep problems: an ONE. 2015;10(11):e0142533.
emerging global epidemic? Findings from the INDEPTH WHO- 22. Sadeh A. The role and validity of actigraphy in sleep medicine: an
SAGE study among more than 40,000 older adults from 8 countries update. Sleep Med Rev. 2011;15(4):259-267.
across Africa and Asia. Sleep. 2012;35(8):1173-1181. 23. Balke B, Ware R. An experimental study of physical fitness of Air
3. Yang P, Ho K, Chen H, Chien M. Exercise training improves sleep Force personnel. US Armed Forces Med J. 1959;10(6):675-688.
quality in middle-aged and older adults with sleep problems: a sys- 24. Kredlow MA, Capozzoli MC, Hearon BA, Calkins AW, Otto MW.
tematic review. J Physiother. 2012;58(3):157-163. The effects of physical activity on sleep: a meta-analytic review. J
4. Kovacevic A, Mavros Y, Heisz JJ, Fiatarone Singh MA. The effect Behav Med. 2015;38(3):427-449.
of resistance exercise on sleep: a systematic review of randomized 25. Jubeau M, Sartorio A, Marinone PG, et al. Comparison between
controlled trials. Sleep Med Rev. 2018;39:52-68. voluntary and stimulated contractions of the quadriceps femoris
5. Lie JD, Tu KN, Shen DD, Candidate P, Wong BM, Candidate for growth hormone response and muscle damage. J Appl Physiol.
P. Pharmacological treatment of insomnia. Pharm Ther. 2007;104(1):75-81.
2015;40(11):759-769. 26. Copinschi G, Nedeltcheva A, Leproult R, et al. Sleep dis-
6. Abad VC, Guilleminault C. Diagnosis and treatment of sleep turbances, daytime sleepiness, and quality of life in adults
disorders: a brief review for clinicians. Dialogues Clin Neurosci. with growth hormone deficiency. J Clin Endocrinol Metab.
2003;5(4):371-388. 2010;95(5):2195-2202.
7. Mikic A, Pietrolungo CE. Effects of diet on sleep quality. Adv Nutr. 27. Kemmler W, Weissenfels A, Willert S, et al. Efficacy and safety of
2016;7(5):938-949. low frequency Whole-Body Electromyostimulation (WB-EMS) to
8. Kline CE. The bidirectional relationship between exercise and improve health-related outcomes in non-athletic adults. A system-
sleep: implications for exercise adherence and sleep improvement. atic review. Front Physiol. 2018;9:573.
Am J Lifestyle Med. 2014;8(6):375-379. 28. Uchida S, Shioda K, Morita Y, Kubota C, Ganeko M, Takeda N.
9. Dolezal BA, Neufeld EV, Boland DM, Martin JL, Cooper CB. Exercise effects on sleep physiology. Front Neurol. 2012;3:1-5.
Interrelationship between sleep and exercise: a systematic review. 29. Willand MP, Rosa E, Michalski B, et al. Electrical muscle stim-
Adv Prev Med. 2017;2017:1364387. ulation elevates intramuscular BDNF and GDNF mRNA fol-
10. Passos GS, Poyares D, Santana MG, Tufik S. Effect of acute phys- lowing peripheral nerve injury and repair in rats. Neuroscience.
ical exercise on patients with chronic primary insomnia. J Clin 2016;334:93-104.
Sleep Med. 2010;6(3):270-275. 30. Monteiro BC, Monteiro S, Candida M, et al. Relationship be-
11. Amaro-Gahete FJ, De-la-O A, Jurado-Fasoli L, et al. Exercise tween brain-derived neurotrofic factor (Bdnf) and sleep on de-
training as S-Klotho protein stimulator in sedentary healthy adults: pression: a critical review. Clin Pract Epidemiol Ment Heal.
rationale, design, and methodology. Contemp Clin Trials Commun. 2017;13(1):213-219.
2018;11:10-19. 31. Truong AD, Kho ME, Brower RG, Feldman DR, Colantuoni
12. Schulz KF, Altman DG, Moher D, et al. CONSORT 2010 state- E, Needham DM. Effects of neuromuscular electrical stimula-
ment: updated guidelines for reporting parallel group randomised tion on cytokines in peripheral blood for healthy participants: a
trials. BMC Med. 2010;8(1):18. prospective, single-blinded Study. Clin Physiol Funct Imaging.
13. Schulz K, Grimes D. Generation of allocation sequences in ran- 2017;37(3):255-262.
domised trials: chance, not choice. Lancet. 2002;359(9305):515-519. 32. Krueger JM. The role of cytokines in sleep regulation. Curr Pharm
14. Slade SC, Dionne CE, Underwood M, Buchbinder R. Consensus Des. 2009;14(32):3409-3416.
on exercise reporting template (CERT): explanation and elabora- 33. Boerio D, Jubeau M, Zory R, Maffiuletti NA. Central and periph-
tion statement. Br J Sports Med. 2016;50(23):1428-1437. eral fatigue after electrostimulation-induced resistance exercise.
15. Simera I, Moher D, Hoey J, Schulz KF, Altman DG. A catalogue Med Sci Sports Exerc. 2005;37(6):973-978.
of reporting guidelines for health research. Eur J Clin Invest. 34. Mallampalli MP, Carter CL. Exploring sex and gender differences
2010;40(1):35-53. in sleep health: a society for women's health research report. J
16. WHO. Global Recommendations on Physical Activity for Health. Women's Heal. 2014;23(7):553-562.
Geneva, Switzerland: World Health Organzation Publications; 2010. 35. Krishnan V, Collop NA. Gender differences in sleep disorders.
17. Buchheit M, Laursen PB. High-intensity interval training, solu- Curr Opin Pulm Med. 2006;12(6):383-389.
tions to the programming puzzle: part I: cardiopulmonary empha- 36. Sheel AW. Sex differences in the physiology of exercise: an inte-
sis. Sport Med. 2013;43(5):313-338. grative perspective. Exp Physiol. 2016;101(2):211-212.
18. Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ. The 37. Bredella MA. Sex differences in body composition. In: Mauvais-
Pittsburgh sleep quality index: a new instrument for psychiatric Jarvis F, ed. Sex and Gender Factors Affecting Metabolic
practice and research. Psychiatry Res. 1989;28(2):193-213. Homeostasis, Diabetes and Obesity. Cham: Springer International
19. Migueles JH, Ulf CC, Nystro CD. Accelerometer data collec- Publishing; 2017:9-27.
tion and processing criteria to assess physical activity and other
 |

13652362, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/eci.13202 by Universidad Rey Juan Carlos C/Tulipan S/N Edificio, Wiley Online Library on [21/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
JURADO-FASOLI et al.    11 of 11

38. Al-Mallah MH, Juraschek SP, Whelton S, et al. Sex differences in

cardiorespiratory fitness and all-cause mortality: the Henry ford ex- How to cite this article: Jurado-Fasoli L, De-la-O A,
ercise testing (FIT) project. Mayo Clin Proc. 2016;91(6):755-762. Molina-Hidalgo C, Migueles JH, Castillo MJ, Amaro-
39. Miller AEJ, MacDougall JD, Tarnopolsky MA, Sale DG. Gender Gahete FJ. Exercise training improves sleep quality: A
differences in strength and muscle fiber characteristics. Eur J Appl randomized controlled trial. Eur J Clin Invest.
Physiol Occup Physiol. 1993;66(3):254-262.
2020;50:e13202. https​://doi.org/10.1111/eci.13202​

SUPPORTING INFORMATION
Additional supporting information may be found online in
the Supporting Information section.