Napping and cognitive performance during night shifts:

a systematic review and meta-analysis

Frédéric Dutheil, Brice Bessonnat, Bruno Pereira, Julien Baker, Fares
Moustafa, Maria Livia Fantini, Martial Mermillod, Valentin Navel

To cite this version:

Frédéric Dutheil, Brice Bessonnat, Bruno Pereira, Julien Baker, Fares Moustafa, et al.. Napping and
cognitive performance during night shifts: a systematic review and meta-analysis. Sleep, 2020, 43
(12), pp.zsaa109. �10.1093/sleep/zsaa109�. �hal-02979271�

HAL Id: hal-02979271

https://hal.science/hal-02979271v1
Submitted on 26 Nov 2020

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Napping and cognitive performance during night shifts: a systematic review

and meta-analysis

Frédéric Dutheil 1,2*, Brice Bessonnat 3, Bruno Pereira4, Julien S Baker5, Fares Moustafa6,

Maria Livia Fantini 7, Martial Mermillod8,9, Valentin Navel 10

1
Université Clermont Auvergne, CNRS, LaPSCo, Physiological and Psychosocial Stress, CHU

Clermont-Ferrand, University Hospital of Clermont-Ferrand, Preventive and Occupational

Medicine, Witty Fit, F-36000 Clermont-Ferrand, France

2
Australian Catholic University, Faculty of Health, School of Exercise Science, Melbourne,

Australia
3
CHU Clermont-Ferrand, University Hospital of Clermont-Ferrand, Preventive and

Occupational Medicine, F-36000 CHU Clermont-Ferrand, France

4
CHU Clermont-Ferrand, University Hospital of Clermont-Ferrand, Biostatistics Unit, the

Clinical Research and Innovation Direction, F-36000 Clermont-Ferrand, France

5
Centre for Health and Exercise Science Research, Department of Sport, Physical Education

and Health, Hong Kong Baptist University, Kowloon Tong, Hong Kong
6
CHU Clermont-Ferrand, University Hospital of Clermont-Ferrand, Emergency, F-36000

Clermont-Ferrand, France
7
Université Clermont Auvergne, NPsy-Sydo, CHU Clermont-Ferrand, University Hospital of

Clermont-Ferrand, Sleep disorders, F-36000 Clermont-Ferrand, France

8
Univ. Grenoble Alpes, LPNC, F-38000 Grenoble, France & CNRS, LPNC, F-38000

Grenoble, France
9
Institut Universitaire de France, F-75000 Paris, France

1
10
CHU Clermont-Ferrand, University Hospital of Clermont-Ferrand, Ophthalmology, F-63000

Clermont-Ferrand, France

*Corresponding Author: Frédéric Dutheil, CHU Clermont-Ferrand, Santé Travail

Environnement, 58 rue Montalembert, 63000 Clermont-Ferrand, France, Tel: +33 6 88 22 48

48, Fax: +33 4 73 27 46 49, [email protected]

Competing Interests: The authors have declared that no competing interests exist.

Running Title: Napping and cognitive performance in night shifts

2
Abstract
STUDY OBJECTIVES: To examine the benefits of napping during night shifts on cognitive

performance.

METHODS: Medline, Cochrane Library, Science direct and Embase databases were searched

up to 01/07/2019. Cognitive performance during night shifts, both before and following

napping or under control conditions (no nap), in working-aged adults, were analysed by time

and by type of cognitive function (executive function, attention, instrumental function and

memory). Estimates were pooled using random-effects meta-analysis.

RESULTS: 18 articles (6 in real-work and 12 in laboratory) with a total of 494 participants

were included. The mean nap duration was 41.6±28.3 min, occurring between 12.00 and 4.10

am, with a mean time set at 2.12 am. Cognitive performance did not differ at baseline between

the groups (effect size 0.02, 95%CI -0.09 to 0.13). There was an overall improvement in

performance following a nap compared to the control condition without a nap (0.25, 0.10 to

0.41). Positioning naps early in the night and activity (simulated work tasks) tended to improve

cognitive performance (-0.57, -1.16 to 0.002, and 0.082, -0.04 to 0.33, respectively). The

improvements were primarily seen 30 minutes after awakening. Only memory deteriorated

immediately after awakening without an overall change in global cognitive performance.

CONCLUSION: Napping during night shifts seems to improve cognitive performance.

Napping early in the night and activity may benefit cognitive performance over time.

Considering lack of data in real work environments, further studies are warranted before

preconizing napping during night shifts as a preventive strategy (safety, health, economic

outcomes).

Keywords: night shift; nap; cognitive performance; prevention; sleep

3
Statement of significance

Shift patterns involving night work represents 20% of the workforce, causing sleep deprivation

and disruption of circadian rhythms. Napping during night shifts is a non-pharmacological

strategy that has been suggested to limit the deterioration of cognitive performance. In this

systematic review and meta-analysis, we aimed to evaluate the effect of napping during night

shifts on cognitive performances (executive function, attention, instrumental function and

memory), in comparison with a control condition (no nap). The main findings were that napping

during night shifts improved cognitive performance, specifically attention. Despite lack of high-

power studies on real work environments and that most studies concern simulated night shifts,

napping early in the night and activity (simulated work tasks) may be beneficial to cognitive

performance during a night shift.

4
Introduction
Shift patterns involving night work is common with a prevalence surpassing 20% of the

workforce,1 and is a major cause of circadian rhythm disruption.2 The circadian timing system

coordinates physiological processes encompassing the vigilance state.3,4 Thus, night shifts have

immediate effects on sleepiness in the workplace,5,6 with potential cognitive impairments,7-10 a

risk of errors,11,12 and possible harm causing work injuries.13,14 Napping during night shifts is a

non-pharmacological strategy that has been suggested to limit the deterioration of cognitive

performance, despite conflicting results concerning the effects of different nap durations.15-18

However, a potential negative effect of napping could be sleep inertia, i.e. a transitional state

of low arousal with temporarily deteriorated performance that occurs immediately after

awakening.19 In the absence of major sleep deprivation, sleep inertia rarely exceeds 30

minutes19 and interferes with preferentially high levels of attention19 or working memory.20

Furthermore, it has been reported that the benefits of napping depends on the duration and

circadian placement of the nap. 21 Moreover, cognitive performances are influenced by activity

or other stimuli, such as physical activity,22-24 light25 or noise26 exposure. To the best of our

knowledge, no meta-analysis has assessed the benefits of napping on objective cognitive

performance during night shifts. Meta-analysis is an important tool for illustrating trends that

may not be apparent in a single research. Therefore, we hypothesised that 1) napping would

improve cognitive performance; 2) cognitive performance would mainly improve after a 30-

minute sleep inertia period; 3) the effect of napping would vary depending on the type of

cognitive performance; 4) the influence of napping on cognitive performance would be linked

to the characteristics of the nap (duration, time in the night) and of the individuals (age,27

gender28); 5) activity may benefit cognitive performance. Thus, we aimed to conduct a

systematic review and meta-analysis of the benefits of napping during night shifts on cognitive

performance.

5
Methods

Literature Search

We reviewed all studies reporting cognitive performance following a nap during a night shift

compared to a control group without a nap. Medline, Cochrane Library, Science Direct and

Embase databases were searched up to July 1, 2019, using the following keywords: ("nap*" OR

"sleep*") AND ("performance" OR "task" OR "cognitive") AND ("night" OR "shift*" OR

"work*" OR "occupation*"). Animal studies were excluded. The search was not limited to

specific years. No minimal sample size and no language restrictions were applied. The search

strategy was inclusive of working-age adults (18 to 65 years-old), undertaking regular night

shift work. We excluded studies mentioning the use of therapeutic adjuvants in addition to a

nap. To be included, articles needed to be controlled studies describing our primary outcome

variable, which was the measurement of any cognitive performance during the night shift. We

imposed no limitation on the regional origin or the nature of the control group. Studies needed

to be primary research. In addition, reference lists of all publications meeting the inclusion

criteria were manually searched to identify any further studies that were not found with the

electronic search. Ancestry searches were also completed on previous reviews to locate other

potentially eligible primary studies. The search strategy is presented in Figure 1. Two authors

(BB and VN) conducted all the literature searches and collated the abstracts. Three authors (BB,

FD and VN) separately reviewed the abstracts and decided the suitability of the articles for

inclusion based on the selection criteria. A third author (BP) was asked to review the articles

where consensus on suitability was debated. All authors then reviewed the eligible articles.

6
Quality of Assessment

The “Consolidated Standards of Reporting Trials” (CONSORT)29and Revised Cochrane risk-

of-bias tool (RoB 2)30 were used to evaluate randomised controlled trials. The “STrengthening

the Reporting of OBservational studies in Epidemiology” (STROBE)31 and Newcastle-Ottawa

Scale (NOS)32 were used to check the quality of nonrandomized studies. CONSORT, Rob2,

STROBE and NOS statements consisted of a checklist of several items and sub-items, which

relate to the title, abstract, introduction, methods, results and discussion sections of articles.

One point was attributed per item or sub-item when the study fulfilled the criteria. CONSORT,

Rob2, STROBE and NOS statements could achieve a maximal score of 37, 22, 32 and 9

respectively. The obtained score was then converted into a percentage.

Statistical Considerations

Statistical analysis was conducted using the Stata software (version 13, StataCorp, College

Station, USA).33 Baseline characteristics were summarised for each study sample and reported

as a mean (with standard-deviation). When raw data were not available, data were scaled from

graphs with the measuring tool in the Gimp software (version 2, Gnome Foundation, Orinda,

USA) and then converted from pixels to the appropriate unit. Article pages containing these

graphs were initially converted from .pdf to. xcf format at a high resolution of

8250px*11000px. Heterogeneity in the study results was evaluated by examining forest plots,

confidence intervals (CI) and using formal tests for homogeneity based on the I² statistic, which

is the most common metric for measuring the magnitude of between-study heterogeneity and

is easily interpretable. I² values range from 0% to 100% and are typically considered low for

<25%, modest for 25–50%, and high for >50%. Random effects meta-analyses (DerSimonian

and Laird approach) were conducted when data could be pooled.34 Cognitive tests were

stratified in four types of cognitive function: 35 executive function, attention, instrumental

7
function and memory. All meta-analyses were carried out on cognitive performances in the nap

and control groups. We first conducted a meta-analysis on the cognitive performance at baseline

(t0) stratified on the types of cognitive performance. We then conducted a meta-analysis on the

overall cognitive performance after a nap, stratified on the time of the analysis: t1 (<30 min

after the nap), t2 (31 min to 120 min) and t3 (>121 min). Finally, we conducted a meta-analysis

on the performance after the nap (t1, t2 and t3) for each type of cognitive function. We

calculated the effect size (ES, standardised mean differences - SMD) of the cognitive

performance for each dependent variable.36 A positive ES denoted improved performance. A

scale has been suggested for ES with 0.8 reflecting a large effect, 0.5 a moderate effect, and 0.2

a small effect.37 Funnel plots of these meta-analyses were used to search for potential

publication bias. In order to verify the strength of the results, further meta-analyses were then

conducted excluding non-randomised studies and studies that were not evenly distributed

around the base of the funnel.38 We also conducted further detailed meta-analyses by

environment type (workplace vs laboratory) and by stratifying the first 30 minutes after

awakening in greater depth in order to further explore the sleep inertia period. Meta-regression

analyses were performed to study the influence on effect size of several study and subject

characteristics: age, gender, total sleep time monitored by polysomnography, baseline

performance, activity (performing tasks after nap to simulate a real work for participants

involved in a laboratory protocol), delay between the nap and the cognitive test, time occurrence

of the nap (time at the beginning of the nap and the time of awakening), and the type of

environment: workplace vs laboratory. Results were expressed as regression coefficients and

95%CI.

Results

8
An initial search produced 26,055 possible articles. Removal of duplicates and the use of the

selection criteria reduced this number to 18 articles reporting cognitive performance following

a nap during night work (Figure 1). All articles were written in English.

Quality of the Articles

The assessment of the quality of the 18 studies included was performed using the CONSORT

and RoB2 for the eleven randomised studies,15,16,18,39-46 and STROBE and NOS for the seven

non-randomised studies.47-53 The mean CONSORT/STROBE score was 50.8±15.2, varying

from 16.053 to 74.2%;18 the mean RoB2 score was 78.4±6.2, varying from 68.245 to 90.1%;18

and the mean NOS score was 84.1±8.74, varying from 77.8 48,51,52,54 to 100%.50 Overall, the

studies performed best in the discussion section and worst in the result section. Most studies

(14/18 i.e. 77.8%) mentioned ethical approval.15,18,40-49,51,52 (Appendix 1 and 2)

Inclusion Criteria of Participants

The main inclusion criteria were to be a regular shift worker with a specific occupation 18,43,47,49-
51
or to be an adult with self-reported good health39-42,44-46,48,52,53 or self-reported good sleep.15

The main exclusion criteria were any sleep disturbance 15,16,39-42,44-46,48,50,51 or habitual daytime

napping15,39-42,52, not to be an intermediate morningness-eveningness type40-42,44-46,52, not to

have recently undertaken transmeridian travel 40-42,52, not to be a smoker 15,39-42,44-46,48,52,53, not to

be a consumer of excessive alcohol 15,39-42,44-46,52 or caffeine16,39-42,44-46,52 or use any medication

except oral contraception41,42,44-46,53 and not use medication affecting sleep architecture.15,16,44-
47,52
A minimal exposure to shiftwork was required in two studies, from 4 months49 to two

years.50

Population
9
Sample size: 494 subjects were included: 407 in the nap group and 377 in the control group,

among which 290 were enrolled in a crossover condition. The mean population size was

22.8±12.3 ranging from 847 to 49.18 The mean proportion of participants in the nap group was

53.1±6.0% ranging from 50.015,16,39,43,47-53 to 64.5%.41,42 (Table 1)

Age: The mean age in the nap group was 27.3±7.3 years old and 27.3±7.4 years old in the

control group. Two studies did not report the mean age of the nap and control group.42,50

Gender: The proportion of males in the nap group was 62.4±38.4% and 63.4±37.9% in the

control group. Six studies included only males,15,48-50,52,53 and four studies included only

female.16,40,45,46

Type of sleep: one crossover study evaluated the effect of napping based on the sleeping type

of the individuals (morning or evening person).53

Outcome and Aim of the Studies

The principal aim of most of the studies (12/15 i.e. 80.0%) was to evaluate the effect on

cognitive performance of napping during a nightshift.16,18,39-47,49-51,53 The aim of two of the

studies was to investigate the intensity of sleep inertia after napping 48,52, and one study

investigated the relationship between perceived sleepiness and cognitive performance. 15

Study Design

All included studies were prospective controlled trials, including twelve studies (66.7%) with

a crossover design.15,43-53 Only one crossover study was not counterbalanced. 47 The order of

condition was randomly assigned in two crossover studies.15,43 All non-crossover studies were

randomised.16,18,39-42 (Table 1)

Characteristics of the Intervention

10
Nap: the mean duration of the nap was 41.6 min±28.3 min, ranging from 1041,42 to 120 min.48,53

Eight studies compared different length nap duration.40-42,45,50,52,54,55 (Table 1)

Number of naps per night: all studies evaluated the effect of one nap per night.15,16,18,39-44,46-53

Two studies also evaluated the effect of two naps of 10 minutes 40 or two naps (90 min and 30

min) in a subgroup.45

Start of nap: all studies gave information on the timing of the nap. The mean hour for the nap

was 2.12 am (SD = 77.0 min), and ranged from 12.00 am48,49 to 4.10 am.50

Study environment was the usual work environment in six studies,16,18,43,46,49,51 and in sleep

laboratories in twelve studies.15,39-42,44,45,47,48,50,52,53 In eight of the studies occurring in a sleep

laboratory, participants had no prescribed tasks between the cognitive tests and were permitted

to rest or to engage in quiet activities of their own choice (no activity).15,39-42,45,47,52 Other studies

(10/18 i.e. 52.6%) had defined tasks (activity) during a real or simulated night-

shift.16,18,43,44,46,48-51,53 The subjects were aircraft maintenance engineers,49 air-traffic

controllers,51 nurses or physicians in a public hospital 43,46 or in an academic emergency

department 18, or nurses in a psychiatric hospital.16 Simulated work tasks were typing44,55, typing

and arithmetic53, computer monitoring.50

Activity of the control group during the nap: The control group had to continue to

work,16,18,43,46,50,53 or had the opportunity to rest 38,39,41,44,45 or do a quiet activity without

sleep.13,32-35,37,42,44,45

Time of the tests: With the exception of two studies, 16,44 all studies assessed baseline (t0)

cognitive performance before the nap at the beginning of the shift.15,18,39-43,47-53 The mean time

between the baseline testing and the beginning of the nap was 212.9 min±187.1 min, ranging

from 15 min47 to 550 min.40 Fourteen studies assessed cognitive performance within 30 minutes

after the nap (t1).15,16,18,39,40,42,44,45,47,48,50-53 The mean time between the nap and the beginning

of cognitive tests at t1 was 3.4 min±5.5 min ranging from 0 min16,18,39,45,48,51-53 to 20 min.44,51

11
Ten studies assessed cognitive performance between 31 min and 120 min after the nap

(t2).15,39,41-45,50,52,53 The mean time between the nap and the t2 cognitive test was 68.9 min±26.9

min ranging from 45 min15 to 120 min.43 Eleven studies explored cognitive performances

between 121 min and the end of the shift (t3).15,18,40,41,43,45-47,49-51 The mean time between the

nap and the t3 cognitive tests was 213.8 min±58.8 min, ranging from 135 min47 to 290 min.51

Number of continuous nights: Most studies investigated the benefit of napping over one

night.16,18,39-43,45,47,48,50-53 Two studies investigated the effect of napping over two consecutive

night shifts,15,46 one investigated three consecutive night shifts49 and another investigated five

consecutive night shifts.44

Measurement of Cognitive Performance

Cognitive performance were evaluated through the measurement of executive functions in

seven studies,16,18,39,40,45,47,53 attention in thirteen studies,15,16,18,41-43,45-51 instrumental function

in five studies,15,16,42,46,53 and memory in three studies18,44,52 (Figure 1; see details of the

measurements in Appendix 3 and Table 1). Executive functions were evaluated using the Arrow

Orientation Task,39 a driving simulation,40,47 a simulation of an intravenous insertion,18 a typing

task,53 mental calculations 45 and the Wisconsin Card Sorting Test.16,56 Attention was evaluated

with the 10min-Psychomotor Vigilance Test,15,18,41-43,45,47,51 the Two Choice Visual Reaction

Time,50 the Letter Cancellation task, 46 the Mackworth Clock Vigilance Time,49 the Taiwan

University Attention Test16 and the Visual Vigilance Task.48 Instrumental function was

analysed with a calculation test,42,53 the Information Process Index,16 and the Symbol Digit

Substitution Task.15,46 Memory was evaluated with the 4min-Two Back Memory Task,52 the

Probe Recall Memory Test18 and the Working Memory Digit Recall Test. 44

Meta-Analyses of Cognitive Performance at Baseline (t0)

12
Overall cognitive performance differed between napping vs control groups at baseline (effect

size 0.13, 95% CI 0.03 to 0.24, I2=28.5%) but did not differ after exclusion of outcomes not

evenly distributed around the funnel plot and after the exclusion of non-randomised controlled

trials (0.02, -0.09 to 0.13, I2=0.0%) (Figures 2 and 3, Appendix 8-9), as well as for all types of

cognitive performance on stratified meta-analysis, with the exception of attention (Appendix

4). There were no overall differences between napping vs control groups for all cognitive

performances, including attention, after the exclusion of non-randomised studies47-53 and

outcomes not evenly distributed around the funnel plot 45,49 (Figure 3 and Appendix 9).

Meta-regression at baseline did not show any link between cognitive performance and the

characteristics of the nap (total sleep time, time of awakening), the individuals (age, sex) or the

various studies (study environment, activity, time of baseline) (Figure 4).

Meta-Analyses on Overall Cognitive Performance Over Time

Overall cognitive performance improved after a nap (effect size 0.25, 95% CI 0.10 to 0.41,

I2=79.7%) compared to the control group. More specifically, cognitive performance improved

at t2 (0.76, 0.51 to 1.01, I2=75.1%) and t3 (0.55, 0.35 to 0.75, I2=68.2%) but did not differ

between napping vs control groups at t1 (-0.07, -0.28 to 0.14, I2=81.4%). Compared with

baseline, overall cognitive performance improved at t2 (1.14, 0.65 to 1.63) and t3 (0.55, 0.35

to 0.75) and did not differ at t1 (-0.07, -0.29 to 0.14). Cognitive performance also improved at

t2 (0.70, 0.31 to 1.09) and t3 (0.64, 0.37 to 0.90) compared with t1, without differences between

t2 and t3. (Figures 2 and 3, Appendix 4-7)

Meta-analyses were repeated after the exclusion of outcomes not evenly distributed around the

funnel plot (t1,16,51-53 t239,52,53 and t341,43,51) and after the exclusion of non-randomised

controlled trials (t1,47,48,50-53 t250,52,53 and t347,49-51) and showed similar results. (Figure 3,

Appendix 8-9)

13
Meta-Analyses on Categories of Cognitive Performance Over Time

Executive function improved globally after napping (effect size 0.43, 95% CI 0.24 to 0.62,

I2=65.7%) compared with the control group, specifically at t2 (1.45, 0.88 to 2.02, I2=73.5%).

Compared with baseline, executive function improved at t2 (1.45, 0.88 to 2.02, I2=73.5%) and

t3 (0.53, 0.03 to 1.04, I2=58.7%) in the nap group. Attention improved generally after a nap

(0.41, 0.29 to 0.53, I2=66.9%) compared with the control group, specifically at t2 (0.70, 0.31 to

1.09, I2=74.2%). Instrumental function improved after a nap (0.13, -0.05 to 0.31, I2=35.1%)

compared with the control group, specifically at t2 (0.44, 0.01 to 0.89, I2=43.8%). Memory

deteriorated globally after a nap compared with the control group (-0.52, -0.86 to -0.17,

I2=85.2%), specifically at t1 (-1.92, -2.67 to -1.16, I2=89.7%). Compared with baseline,

memory deteriorated at t1 (-1.92, -2.67 to -1.16, I2=89.7%); but improved again at t2 (4.10,

1.19 to 7.01) in the nap group (Figures 2 and 3, Appendix 4-7).

Between the cognitive function categories, memory deteriorated less after a nap than attention,

executive function and instrumental function, overall (-2.41, -3.11 to -1.71; -2.43, -3.21 to -

1.67; -2.39, -3.24 to 1.53, respectively) and at t1 (-2.41, -3.11 to -1.71; -2.43, -3.21 to -1.67; -

2.39, --3.24 to -1.53, respectively) (Figure 4).

Meta-analyses were repeated after the exclusion of outcomes not evenly distributed around the

funnel plot (t116,51-53, t239,52,53 and t341,43,51) (Figures 3 and 4, Appendix 8 and 9) and after the

exclusion of non-randomised controlled trial (t147,48,50-53, t250,52,53 and t347,49-51) and

demonstrated similar results. In general, executive function, instrumental function and memory

did not differ after a nap (t1, t2 and t3) compared to the control group, specifically at t1, t2 or

t3. Overall, attention improved after a nap (0.37, 0.26 to 0.48, I2=52.0%) compared with the

control group, specifically at t2 (0.76, 0.42 to 1,09, I2=62.3%) and t3 (0.59, 0.33 to 0.85,

14
I2=67.4%). Between the cognitive function categories, no overall differences were seen after a

nap and at each time point (t1, t2 and t3). (Figures 2, 3 and 4, Appendix 4-7).

Comparing Real Work and Laboratory Environment

There were no differences between the napping vs. control groups at baseline when considering

only the real work environment (0.08, -0.04 to 0.20, I2=0.0%), whereas studies only in a

laboratory differed between napping vs. control groups (0.15, 0.02 to 0.28, I2=34.3%). Overall

cognitive performance improved after a nap for both the real work or laboratory environments

(effect size 0.37, 95% CI 0.21 to 0.54, I2=59.3%; and 0.32, 0.17 to 0.50, I2=80.5%,

respectively), and also at each time point (t2: 0.59, 0.04 to 1.14, I2=0.0%; and 0.76, 0.50 to

1.02, I2=76.5%; and t3: 0.43, 0.21 to 0.66, I2=62.5%; and 0.64, 0.28 to 1.00, I2=70.1%,

respectively), with the exception of t1. At t1, overall cognitive performance improved in real

work (0.28, 0.01 to 0.55, I2=65.2), but not in laboratory studies – where a deterioration in

memory (-2.12, -2.91 to -1.34, I2=81.6%) was observed. (Figure 5) However, meta-regressions

showed no statistical differences between the real work and laboratory environments. (Figure

4)

Exploration of the Sleep Inertia Period

In order to explore the sleep inertia period in more depth, and since most studies within the t1

group (<30 min after the nap) measured a cognitive performance immediately upon awakening,

we also stratified t1 into three periods: immediately after awakening, 1 to 15 minutes after

awakening, and 15 to 30 minutes after awakening. There were no differences between the

groups (nap vs no nap) from 1 to 15 minutes and from 15 to 30 minutes after awakening,

irrespective of the cognitive performances categories. In general, cognitive performance did not

change immediately after awakening (i.e. within the first minute). However, the stratifications

15
demonstrated a deterioration in memory (effect size -2.30, 95% CI -3.22 to -1.38, I2=88.0%),

and an improvement in instrumental function (0.60, 0.11 to 1.10, I2=34.9). (Figure 6)

Relationship with Napping or Individuals or Study Characteristics

Overall cognitive performance after napping improved with a delay between the nap and the

test (0.18, 95% CI 0.04 to 0.33), and tended to improve with total sleep time (0.56, 95% CI -

0.10 to 1.22) and activity (0.29, 95% CI -0.04 to 0.61). Cognitive performance at t1 improved

with activity (1.10, 95% CI 0.38 to 1.83), and tended to improve with total sleep time (1.57,

95% CI -0.12 to 3.26) but to deteriorate in the male gender (-0.05, 95% CI -0.08 to 0.02) and

with age (-0.25, 95% CI -0.53 to 0.38) (Figure 4).

Meta-regression was repeated after the exclusion of outcomes not evenly distributed around the

funnel plot and after the exclusion of non-randomised controlled trials, and showed similar

results. Overall cognitive performances after a nap improved with a delay between the nap and

the test (0.08, 95% CI 0.00 to 0.16, p=0.041), and tended to improve with an early awakening

time (0.57, 95% CI -1.16 to 0.02, p=0.060), and to deteriorate with the male gender (-0.01, 95%

CI -0.02 to 0.00, p=0.037) (Figure 4).

Discussion
The main findings were that napping during night shifts improved cognitive performance in a

delayed manner, specifically attention. Despite lack of data precluding robust conclusions

during the sleep inertia period, only memory deteriorated immediately after awakening without

an overall change in global cognitive performance. Napping early in the night and activity may

be beneficial to cognitive performance.

16
Napping and Cognitive Performance

Sleepiness in night-shift workers is a main concern with increased risk of work injury and

economic consequences.57,58 Many strategies against sleepiness have been proposed for night

shift workers, such as caffeine and bright light. 59 Nonetheless, the benefits of napping could be

superior and without detrimental effects. 59 In line with our study, a review of the literature

concluded that napping in night shifts led to improved sleep related performance. 6 Using a

sound statistical approach, we demonstrated that napping during night shifts improved

cognitive performance. A recent meta-analysis did not demonstrate an effect of napping on

subjective sleepiness or on sleepiness-related performance in shift-workers.60 However, this

meta-analysis included only four studies.18,43,47,61 We included those studies 18,43,47 in our study,

with the exception of one which did not report cognitive test performance.61 Moreover, our

meta-analysis did not focus on subjective feelings but on objective cognitive performances.15

We demonstrated that attention was significantly improved after napping in a delayed manner,

without detrimental effects during the sleep inertia period. The included attention tests were

mostly simple reaction times, which are sensitive markers of sleepiness.62 Napping in night

shifts could also have larger implications on safety, decreasing the risk of professional error12

and work injury.63 Moreover, napping during night shifts could also be a countermeasure for

the general health impairment of shift workers who have an increased risk of long-term

cognitive impairment,64 insomnia,5,65 depression,66 metabolic syndrome,67 diabetes,68 vascular

events69,70 and cancer.71,72

Sleep Inertia

Surprisingly considering the literature on sleep inertia effects,19,39,42 only memory deteriorated

immediately after awakening without an overall change in global cognitive performance. Sleep

inertia is described as a transitional state of low arousal immediately after awakening, with a

17
temporary deteriorated performance.19 Sleep inertia is one of the perceived barriers of napping

during night shift for workers.73 Despite a lack of evidence-based data in real work environment

during the sleep inertia period, we reported that cognitive performance tended to improve in

the first 30 min with total sleep time, whereas the duration of sleep is generally suspected to

increase sleep inertia.19 Thus, brief naps (less than 30 min) were proposed to limit this

phenomenon but a recent review of the literature did not find any clear results.74 Furthermore,

it has been suggested that cognitive tasks involving high levels of attention 19 or working

memory20 deteriorated more through sleep inertia than easier tests. Our study included only one

non-randomised controlled trial reporting results on memory during the sleep inertia period,52

but the analyses were not highly powered, precluding any robust conclusions. Interestingly, an

age effect on higher cognitive impairment was only demonstrated during sleep inertia in older

individuals. Comparisons of the benefits of napping in different age groups are not clear in the

literature.59 Napping early in the night and activity at awakening may also prevent the sleep

inertia effects.

Short Nap Early in the Night and Activity

We demonstrated that napping early in the night and activity may benefit cognitive

performance. Circadian placement of the nap has been identified as influencing its benefits,

with poor regenerative effects when napping is set at the circadian nadir around 4 a.m. (late in

the night).21 In addition, the benefits of napping were greater when individuals were active after

the nap, especially during the sleep inertia period. Maintaining activity increases sympathetic

tone and may help individuals to emerge from sleep inertia, and to increase the overall benefits

of the nap.75 Moreover, we did not find any dose-response relationships between total sleep

time and cognitive performance, in line with the literature.76-78 Long naps do not demonstrate

greater benefits, emphasising the practical interest of shorts naps at the workplace. Some

18
authors have suggested that simply inducing sleep with ultra-short naps could be sufficient to

gain significant benefits,21 whereas other authors suggest a minimum sleep time of around 10

min.78 However, we did not find any study with less than 10 minutes of napping that met our

inclusion criteria. Interestingly, we did not find a deterioration in cognitive performances in

real work environment during the sleep inertia period, whereas memory deteriorated under

laboratory conditions. This may be due to the stimuli provided by occupational stressors.

Finally, for other measurements, we did not find any significant differences between studies at

the workplace or under laboratory conditions, demonstrating that napping could have similar

benefits if an appropriate quiet setting is available at the workplace.

Limitations

Our study has some limitations. We conducted the meta-analyses on only published articles, so

they are theoretically exposed to publication bias. However, this risk is limited because the

analyses of the funnel plots revealed a homogenous distribution. The inclusion criteria were

similar, but not identical. We included studies reporting several performance tests. In the

absence of a clear consensus on the cognitive tests, and to avoid selection bias, some studies

appeared several times in our meta-analysis, with different data or populations. Therefore, the

weighting of the studies required careful thought.79 However, arbitrary choosing one

subcategory for those studies did not affect the results. We limited the influence of extreme

results and heterogeneity by repeating analyses after the exclusion of non-randomised studies

and those with results not evenly distributed around the funnel plots. Although some studies

were randomised, the assessors were not blind for the intervention (nap) as performing blinding

in this type of intervention is difficult. Considering the type of cognitive performances,

executive functions were highly improved 30 min to 120 minutes after awakening. However,

executive functions were reported in only two studies, both of which were non-randomised

19
controlled trials precluding robust conclusions.39 The number of studies including memory and

instrumental functions was not sufficient to draw a conclusion from the stratified results.

Moreover, we noted that the classification of the type of cognitive performance varies between

authors and suffers lack of consensus.62 We also noted that there was a high probability that

performances could be linked together. Complementary studies should be performed to study

the impact of napping on high-level cognitive functions, like decision-making, in the context,

for example, of a medical diagnostic process. Most studies were also lab studies, making it hard

to draw conclusions on napping and performance during ‘night shifts’. However, randomised

controlled trials are more difficult to achieve in real working conditions. Even if comparisons

between real work environments and lab conditions were difficult because lab studies were

often limited to young adults, we showed similar results, except for the memory deterioration

during sleep inertia found in lab studies. In addition, possible naps occurring prior to the night

shift, and their timing, were not reported in the studies included and were not the focus of our

meta-analysis.

Conclusion
Napping during night shifts improves cognitive performances in a delayed manner, especially

attention. Napping early in the night and activity may benefit cognitive performances over time.

Despite the lack of data in real work environments precluding robust conclusions during the

sleep inertia period, only memory deteriorated immediately after awakening without an overall

change in global cognitive performance. Before recognizing guidelines for napping during

night shifts as a preventive strategy, further studies should evaluate the consequences of

napping during night shifts on immediate safety (work errors and work injuries), and on long-

term specific health and economic outcomes.

20
Contributors

FD, BB, BP and VN were responsible for the design and conception of the study. BB and VN

searched and collected studies and data. FD, BB, BP, LF and VN analysed and interpreted data.

FD, BB, BP and VN were in charge of statistical analyses. FD, BB and VN wrote the

manuscript. All authors made critical revisions to the article. All authors gave their final

approval of the article.

Acknowledgments

We thank Richard May for his proof-reading of the article.

Funding

None

Disclosure statement

Financial disclosure statement: all authors report no competing interest.

Non-financial disclosure statement: all authors report no competing interest.

Data sharing statement

No additional data are available

References
1. Condition EFftIoLaW. Sixth European Working Conditions Survey. 2015.
2. Boivin DB, Tremblay GM, James FO. Working on atypical schedules. Sleep medicine.
2007; 8 (6): 578-589.
3. Dibner C, Schibler U. Circadian timing of metabolism in animal models and humans.
Journal of internal medicine. 2015; 277 (5): 513-527.
21
4. Minkel JD, Dinges DF. Circadian Rhythms in Sleepiness, Alertness, and Performance
A2 - Squire, Larry R. In: Encyclopedia of Neuroscience. Oxford: Academic Press; 2009: 965-
970.
5. Oyane NM, Pallesen S, Moen BE, Akerstedt T, Bjorvatn B. Associations between night
work and anxiety, depression, insomnia, sleepiness and fatigue in a sample of Norwegian
nurses. PloS one. 2013; 8 (8): e70228.
6. Ruggiero JS, Redeker NS. Effects of napping on sleepiness and sleep-related
performance deficits in night-shift workers: a systematic review. Biological research for
nursing. 2014; 16 (2): 134-142.
7. Machi MS, Staum M, Callaway CW, et al. The relationship between shift work, sleep,
and cognition in career emergency physicians. Academic emergency medicine : official journal
of the Society for Academic Emergency Medicine. 2012; 19 (1): 85-91.
8. Maltese F, Adda M, Bablon A, et al. Night shift decreases cognitive performance of
ICU physicians. Intensive care medicine. 2016; 42 (3): 393-400.
9. Lowe CJ, Safati A, Hall PA. The neurocognitive consequences of sleep restriction: A
meta-analytic review. Sleep Restriction and Cognition. Neuroscience and biobehavioral
reviews. 2017.
10. Kerkhof G, Van Dongen H. Effects of sleep deprivation on cognition. Human sleep and
cognition: basic research. 2010; 185: 105.
11. Landrigan CP, Rothschild JM, Cronin JW, et al. Effect of reducing interns' work hours
on serious medical errors in intensive care units. The New England journal of medicine. 2004;
351 (18): 1838-1848.
12. de Cordova PB, Bradford MA, Stone PW. Increased errors and decreased performance
at night: A systematic review of the evidence concerning shift work and quality. Work
(Reading, Mass). 2016; 53 (4): 825-834.
13. Wagstaff AS, Sigstad Lie JA. Shift and night work and long working hours--a
systematic review of safety implications. Scandinavian journal of work, environment & health.
2011; 37 (3): 173-185.
14. Wong IS, McLeod CB, Demers PA. Shift work trends and risk of work injury among
Canadian workers. Scandinavian journal of work, environment & health. 2011; 37 (1): 54-61.
15. Tremaine R, Dorrian J, Lack L, et al. The relationship between subjective and objective
sleepiness and performance during a simulated night-shift with a nap countermeasure. Appl
Ergon. 2010; 42 (1): 52-61.
16. Chang YS, Wu YH, Lu MR, Hsu CY, Liu CK, Hsu C. Did a brief nap break have
positive benefits on information processing among nurses working on the first 8-h night shift?
Appl Ergon. 2015; 48: 104-108.
17. Tempesta D, Cipolli C, Desideri G, De Gennaro L, Ferrara M. Can taking a nap during
a night shift counteract the impairment of executive skills in residents? Med Educ. 2013; 47
(10): 1013-1021.
18. Smith-Coggins R, Howard SK, Mac DT, et al. Improving alertness and performance in
emergency department physicians and nurses: the use of planned naps. Ann Emerg Med. 2006;
48 (5): 596-604, 604 e591-593.
19. Tassi P, Muzet A. Sleep inertia. Sleep medicine reviews. 2000; 4 (4): 341-353.
20. Groeger JA, Lo JC, Burns CG, Dijk DJ. Effects of sleep inertia after daytime naps vary
with executive load and time of day. Behavioral neuroscience. 2011; 125 (2): 252-260.
21. Lovato N, Lack L. The effects of napping on cognitive functioning. Progress in brain
research. 2010; 185: 155-166.
22. Brush CJ, Olson RL, Ehmann PJ, Osovsky S, Alderman BL. Dose-Response and Time
Course Effects of Acute Resistance Exercise on Executive Function. Journal of sport & exercise
psychology. 2016; 38 (4): 396-408.

22
23. Weng TB, Pierce GL, Darling WG, Voss MW. Differential Effects of Acute Exercise
on Distinct Aspects of Executive Function. Medicine and science in sports and exercise. 2015;
47 (7): 1460-1469.
24. Zach S, Shalom E. The Influence of Acute Physical Activity on Working Memory.
Perceptual and motor skills. 2016; 122 (2): 365-374.
25. Tonetti L, Natale V. Effects of a single short exposure to blue light on cognitive
performance. Chronobiol Int. 2019; 36 (5): 725-732.
26. Alimohammadi I, Kanrash FA, Abolghasemi J, Vosoughi S, Rahmani K, Chalak MH.
Relationship Between Noise Annoyance and Cognitive Performance in Automotive Workers
Exposed to Chronic Noise. Journal of UOEH. 2019; 41 (4): 375-385.
27. Hedden T, Gabrieli JD. Insights into the ageing mind: a view from cognitive
neuroscience. Nature reviews Neuroscience. 2004; 5 (2): 87-96.
28. Hyde JS. Sex and cognition: gender and cognitive functions. Current opinion in
neurobiology. 2016; 38: 53-56.
29. Moher D, Hopewell S, Schulz KF, et al. CONSORT 2010 explanation and elaboration:
updated guidelines for reporting parallel group randomised trials. BMJ. 2010; 340: c869.
30. Higgins JPT, Sterne JAC, Savović J, et al. A revised tool for assessing risk of bias in
randomized trials. Vol 10: Cochrane Methods. Cochrane Database of Systematic Reviews;
2016.
31. Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthening the Reporting of
Observational Studies in Epidemiology (STROBE): Explanation and elaboration. Int J Surg.
2014; 12 (12): 1500-1524.
32. Wells GA, Shea B, O'Connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing
the quality of nonrandomised studies in meta-analyses.
http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. Accessed March 19, 2019.
33. Navel V, Mulliez A, Benoist d'Azy C, et al. Efficacy of treatments for Demodex
blepharitis: A systematic review and meta-analysis. Ocul Surf. 2019.
34. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986; 7
(3): 177-188.
35. Poldrack R. Introduction to Cognitive Neuroscience A2 - Toga, Arthur W. In: Brain
Mapping. Waltham: Academic Press; 2015: 259-260.
36. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled clinical trials. 1986;
7 (3): 177-188.
37. Citrome L, Magnusson K. Paging Dr Cohen, Paging Dr Cohen... An effect size
interpretation is required STAT!: visualising effect size and an interview with Kristoffer
Magnusson. International journal of clinical practice. 2014; 68 (5): 533-534.
38. Russo MW. How to Review a Meta-analysis. Gastroenterology & hepatology. 2007; 3
(8): 637-642.
39. Asaoka S, Fukuda K, Murphy TI, Abe T, Inoue Y. The effects of a nighttime nap on the
error-monitoring functions during extended wakefulness. Sleep. 2012; 35 (6): 871-878.
40. Centofanti SA, Dorrian J, Hilditch CJ, Banks S. Do night naps impact driving
performance and daytime recovery sleep? Accident; analysis and prevention. 2017; 99 (Pt B):
416-421.
41. Centofanti SA, Hilditch CJ, Dorrian J, Banks S. The impact of short night-time naps on
performance, sleepiness and mood during a simulated night shift. Chronobiology international.
2016; 33 (6): 706-715.
42. Hilditch CJ, Centofanti SA, Dorrian J, Banks S. A 30-Minute, but Not a 10-Minute
Nighttime Nap is Associated with Sleep Inertia. Sleep. 2016; 39 (3): 675-685.

23
43. Smith S. Napping and nightshift work: Effects of a short nap on psychomotor vigilance
and subjective sleepiness in health workers. Sleep and Biological Rhythms. 2007; 2007 (5):
117–125.
44. Davy J, Gobel M. The effects of extended nap periods on cognitive, physiological and
subjective responses under simulated night shift conditions. Chronobiol Int. 2018; 35 (2): 169-
187.
45. Oriyama S, Miyakoshi Y. The effects of nighttime napping on sleep, sleep inertia, and
performance during simulated 16 h night work: a pilot study. J Occup Health. 2018; 60 (2):
172-181.
46. Zion N, Shochat T. Let them sleep: The effects of a scheduled nap during the night shift
on sleepiness and cognition in hospital nurses. J Adv Nurs. 2019.
47. Howard ME. The effects of a 30-minute napping opportunity during an actual night shift
on performance and sleepiness in shift workers. Biological Rhythm Research. 2010; 41 (2):
137-148.
48. Kubo T, Takahashi M, Takeyama H, et al. How do the timing and length of a night-shift
nap affect sleep inertia? Chronobiol Int. 2010; 27 (5): 1031-1044.
49. Purnell MT, Feyer AM, Herbison GP. The impact of a nap opportunity during the night
shift on the performance and alertness of 12-h shift workers. J Sleep Res. 2002; 11 (3): 219-
227.
50. Sallinen M, Harma M, Akerstedt T, Rosa R, Lillqvist O. Promoting alertness with a
short nap during a night shift. J Sleep Res. 1998; 7 (4): 240-247.
51. Signal TL, Gander PH, Anderson H, Brash S. Scheduled napping as a countermeasure
to sleepiness in air traffic controllers. J Sleep Res. 2009; 18 (1): 11-19.
52. Signal TL, van den Berg MJ, Mulrine HM, Gander PH. Duration of sleep inertia after
napping during simulated night work and in extended operations. Chronobiol Int. 2012; 29 (6):
769-779.
53. Takeyama H. Psycho-Physiological Effects of Naps during Night Shifts on Morning
Types and Evening Types. Journal of Occupational Health. 2002; 2002 (44): 89-98.
54. Takeyama H, Matsumoto S, Murata K, et al. Effects of the length and timing of
nighttime naps on task performance and physiological function. Rev Saude Publica. 2004; 38
Suppl: 32-37.
55. Kubo T, Takeyama H, Matsumoto S, et al. Impact of nap length, nap timing and sleep
quality on sustaining early morning performance. Ind Health. 2007; 45 (4): 552-563.
56. Anbazhagan S, Ramesh N, Nisha C, Joseph B. Shift work disorder and related health
problems among nurses working in a tertiary care hospital, Bangalore, South India. Indian
journal of occupational and environmental medicine. 2016; 20 (1): 35-38.
57. Rajaratnam SM, Howard ME, Grunstein RR. Sleep loss and circadian disruption in shift
work: health burden and management. Med J Aust. 2013; 199 (8): S11-15.
58. Chaiard J, Deeluea J, Suksatit B, Songkham W, Inta N. Short sleep duration among Thai
nurses: Influences on fatigue, daytime sleepiness, and occupational errors. J Occup Health.
2018; 60 (5): 348-355.
59. Milner CE, Cote KA. Benefits of napping in healthy adults: impact of nap length, time
of day, age, and experience with napping. Journal of sleep research. 2009; 18 (2): 272-281.
60. Slanger TE, Gross JV, Pinger A, et al. Person-directed, non-pharmacological
interventions for sleepiness at work and sleep disturbances caused by shift work. The Cochrane
database of systematic reviews. 2016; (8): Cd010641.
61. Oriyama S, Miyakoshi Y, Kobayashi T. Effects of two 15-min naps on the subjective
sleepiness, fatigue and heart rate variability of night shift nurses. Industrial health. 2014; 52 (1):
25-35.

24
62. Dorrian J. Psychomotor vigilance performance: Neurocognitive assay sensitive to sleep
loss. Kushida CA. 2005.
63. Herman J, Kafoa B, Wainiqolo I, et al. Driver sleepiness and risk of motor vehicle crash
injuries: a population-based case control study in Fiji (TRIP 12). Injury. 2014; 45 (3): 586-591.
64. Ozdemir PG, Selvi Y, Ozkol H, et al. The influence of shift work on cognitive functions
and oxidative stress. Psychiatry research. 2013; 210 (3): 1219-1225.
65. Vallieres A, Azaiez A, Moreau V, LeBlanc M, Morin CM. Insomnia in shift work. Sleep
medicine. 2014; 15 (12): 1440-1448.
66. Lee A, Myung SK, Cho JJ, Jung YJ, Yoon JL, Kim MY. Night Shift Work and Risk of
Depression: Meta-analysis of Observational Studies. J Korean Med Sci. 2017; 32 (7): 1091-
1096.
67. Wang F, Zhang L, Zhang Y, et al. Meta-analysis on night shift work and risk of
metabolic syndrome. Obes Rev. 2014; 15 (9): 709-720.
68. Anothaisintawee T, Reutrakul S, Van Cauter E, Thakkinstian A. Sleep disturbances
compared to traditional risk factors for diabetes development: Systematic review and meta -
analysis. Sleep medicine reviews. 2016; 30: 11-24.
69. Li M, Huang JT, Tan Y, Yang BP, Tang ZY. Shift work and risk of stroke: A meta-
analysis. Int J Cardiol. 2016; 214: 370-373.
70. Vyas MV, Garg AX, Iansavichus AV, et al. Shift work and vascular events: systematic
review and meta-analysis. BMJ. 2012; 345: e4800.
71. Ijaz S, Verbeek J, Seidler A, et al. Night-shift work and breast cancer--a systematic
review and meta-analysis. Scand J Work Environ Health. 2013; 39 (5): 431-447.
72. Rao D, Yu H, Bai Y, Zheng X, Xie L. Does night-shift work increase the risk of prostate
cancer? a systematic review and meta-analysis. OncoTargets and therapy. 2015; 8: 2817-2826.
73. Fallis WM, McMillan DE, Edwards MP. Napping during night shift: practices,
preferences, and perceptions of critical care and emergency department nurses. Critical care
nurse. 2011; 31 (2): e1-11.
74. Hilditch CJ, Dorrian J, Banks S. A review of short naps and sleep inertia: do naps of 30
min or less really avoid sleep inertia and slow-wave sleep? Sleep medicine. 2017; 32: 176-190.
75. Harris CD. Neurophysiology of sleep and wakefulness. Respiratory care clinics of North
America. 2005; 11 (4): 567-586.
76. Brooks A, Lack L. A brief afternoon nap following nocturnal sleep restriction: which
nap duration is most recuperative? Sleep. 2006; 29 (6): 831-840.
77. Tietzel AJ, Lack LC. The short-term benefits of brief and long naps following nocturnal
sleep restriction. Sleep. 2001; 24 (3): 293-300.
78. Tietzel AJ, Lack LC. The recuperative value of brief and ultra-brief naps on alertness
and cognitive performance. Journal of sleep research. 2002; 11 (3): 213-218.
79. Higgins J. Cochrane Handbook for Systematic Review of Interventions. 2011.
80. Reznek MA, Rawn CL, Krummel TM. Evaluation of the educational effectiveness of a
virtual reality intravenous insertion simulator. Academic emergency medicine : official journal
of the Society for Academic Emergency Medicine. 2002; 9 (11): 1319-1325.
81. Rosa RR, Colligan MJ. Long workdays versus restdays: assessing fatigue and alertness
with a portable performance battery. Human factors. 1988; 30 (3): 305-317.
82. Wechsler D. The measurement and appraisal of adult intelligence. 1958.

25
Figure Legends

Figure 1. Search Strategy

Figure 2. Meta-analysis on cognitive performance between groups (nap vs control) stratified

on time of analysis
95%CI: 95% confidence intervals; 2RTT: two choice visual reaction time test; AOT: arrow orientation task; cathsim: CathSim
intravenous insertion virtual reality simulation; IPI: information process index; MCVT: Mackworth clock vigilance time;
PRMT: probe recall memory test; PVT: 10min-psychomotor vigilance test; SDST: symbol digit substitution task; TUAT:
Taiwan university attention test; VVT: visual vigilance task; WCST: Wisconsin card sorting test; WMT: 4min-two back working
memory task
Each individual study or subgroup used in the meta-analysis is represented in the forest-plot by a dot on a horizontal line. The
black dots represent the effect estimate (effect size - ES), and the length of each line around the dots represent its 95%
confidence interval (95CI). Shorter lines represent a narrower 95CI thus higher precision around study ES, usually found in
larger studies. Conversely, longer lines represent a wider 95CI and less precision around ES usually found in smaller studies.
An overall summary pooled-estimate (result of the meta-analysis) is represented by a blue lozenge at the end of the graph. The
black solid vertical line represents the null estimate (with a value of 0 for ES) and the red dotted vertical line represents result
of the meta-analysis (pooled-estimate). Horizontal lines that cross the null vertical line represent non-significant studies.

Figure 3. Summary of meta-analysis on cognitive performance between groups (nap vs

control), stratified on time of analysis: before (left) and after (right) exclusion of studies not
evenly distributed around the funnel plot and non-randomised controlled trials.
For details of the meta-analysis at each analysis time, please see Appendix 2 for meta-analyses at t0, Appendix 3 for t1,
Appendix 4 for t2, and Appendix 5 for t3.
t0: baseline; t1: <30 min after nap; t2: from 31 min to 120 min after nap; t3: >120 min after nap
Each overall summary of a meta-analysis is represented in the graph by a dot on a vertical line. The black dots represent the
overall pooled-effect estimate of individual meta-analyses (pooled effect size - ES), and the length of each vertical line around
the dots represent their 95% confidence interval (95CI). Shorter lines represent a narrower 95CI thus higher precision around
pooled-ES. Conversely, longer lines represent a wider 95CI and less precision around pooled -ES. The black solid horizontal
line represents the null estimate (with a value of 0 for pooled-ES). Vertical lines that cross the null horizontal line represent a
non-significant overall summary of the meta-analysis at each time analysed (t0, t1, t2, and t3).

Figure 4. Meta-regressions
95%CI: 95% confidence intervals; t0: baseline; t1: up to 30 min after nap; t2: from 31min to 120 min after nap; t3: from 121
min after nap
The effect of each variable on the outcome is represented in the forest-plot by a dot on a horizontal line. The black dots
represent the coefficient for each variable, and the length of each line around the dots represent their 95% confidence interval
(95CI). The black solid vertical line represents the null estimate (with a value of 0). Horizontal lines that cross the null vertical
line represent non-significant variables on the outcome.

26
“Environment” means that studies with real workplace protocol have been considered as a reference in compari son with
laboratory conditions. “Activity” means that participants involved in a laboratory protocol realized a task after nap to simulate
a real work.

Figure 5. Summary of meta-analyses on cognitive performance for each time analysed, in

real work and laboratory environments.
Each summary of several meta-analysis is represented in the forest-plot by a dot on a horizontal line. The black dots represent
the pooled-effect estimate (pooled effect size - ES), and the length of each line around the dots represent their 95% confidence
interval (95CI). Shorter lines represent a narrower 95CI thus higher precision around pooled-ES. Conversely, longer lines
represent a wider 95CI and less precision around pooled-ES. An overall summary of the results of the meta-analyses pooled-
estimate (result of the overall meta-analysis) is represented by a blue lozenge at the end of the graph. The black solid vertical
line represents the null estimate (with a value of 0 for pooled-ES). Horizontal lines that cross the null vertical line represent
the non-significant overall summary of the meta-analysis.

Figure 6. Summary of meta-analyses on cognitive performance at t1, stratified in three

periods: beginning of the tests immediately after awakening, from 1 to 15 minutes after
awakening, and from 15 to 30 minutes after awakening.
Each summary of several meta-analyses is represented in the forest-plot by a dot on a horizontal line. The black dots represent
the pooled-effect estimate (pooled effect size - ES), and the length of each line around the dots represent their 95% confidence
interval (95CI). Shorter lines represent a narrower 95CI thus higher precision around pooled-ES. Conversely, longer lines
represent a wider 95CI and less precision around pooled-ES. An overall summary of the results of meta-analyses pooled-
estimate (result of the overall meta-analysis) is represented by a blue lozenge at the end of the graph. The black solid vertical
line represents the null estimate (with a value of 0 for pooled-ES). Horizontal lines that cross the null vertical line represent
non-significant studies overall summary of meta-analysis.

Supplement files

Appendix 1. Methodological quality of included articles using Newcastle – Ottawa Quality

Assessment Scale and Revised Cochrane risk-of-bias tool for randomised trials (RoB 2)

Appendix 2. Summary bias risk of included randomised and non-randomised studies using
Newcastle – Ottawa Quality Assessment Scale and Revised Cochrane risk-of-bias tool for
randomised trials (RoB 2)

Appendix 3. Details of the tests used in the included articles to measure cognitive performance

27
Appendix 4. Meta-analysis at baseline (t0) between groups (nap vs control) stratified on
cognitive performance

Appendix 5. Meta-analysis at t1 (0 to 30 min) between groups (nap vs control) stratified on

cognitive performance

Appendix 6. Meta-analysis at t2 (31 min to 120 min) between groups (nap vs control) stratified
on cognitive performance

Appendix 7. Meta-analysis at t3 (>121 min) between groups (nap vs control) stratified on

cognitive performance

Appendix 8. Funnels plots - Meta-analysis on cognitive performance between groups (nap vs

control) stratified on time of analysis

Appendix 9. Funnels plots - Meta-analysis on cognitive performance between groups (nap vs

control) stratified on type of cognitive function

28
 Table 1. Characteristics of included studies.
Design Participants Night Shift Nap Time between nap and test Test of
Study Cross- Rando- Environ- Nap Controls Time Duration Time t1 t2 t3 cognitive performance
over mised ment n Age Sex n Age Sex schedule min hours 0 to 30 31 to 120 >120 Name Function
mean±SD % men mean±SD % men min min min

Asaoka 2012 no yes Laboratory 10 21.5±1.0 70.0% 10 21.1±1.1 - 21:00-3:30 60 1:00-2:00 0 60 - AOT Executive f unction
30 3:30-4:00
Centof anti 2016 no yes Laboratory 10 24.7±2.7 40.0% 11 24.5±3.3 36.4% 22:00-7:00 - 190 - Driving simulator Executive f unction
or 20 3:50-4:00
10 3:50-4:00
Centof anti 2017 no yes Laboratory 10 24.7±2.7 - 11 24.5±3.3 - 22:00-7:00 - 90 150 PVT Attention
or 30 3:30-4:00
WCST Attention
Real work
Chang 2015 no yes 21 25.3±3.6 - 21 25.7±1.8 - 00:00-8:00 30 2:00-3:00 0 - - TUAT Executive f unction
(nurses)
IPI Instrumental f unction
0:00-3:20
Davy 2017 yes yes Laboratory 33 21.4±1.6 45.5% 33 21.4±1.6 45.5% 00:00-8:00 200 20 40 - WMT Memory
4:00-7:20
10 3:50-4:00 PVT Attention
Hilditch 2016 no yes Laboratory 10 - 50.0% 11 - - 22:00-7:00 2 47 -
or 30 3:30-4:00 DSST Instrumental f unction
Driving simulator Executive f unction
Howard 2010 yes no Laboratory 8 31.0±9.6 28.6% 8 31.0±9.6 - 21:00-7:00 30 4:00-4:30 0 - 135
PVT Attention
0:00-2:00
60 0:00-1:00
Kubo 2010 yes no Laboratory 12 21.6±2.8 100% 12 21.6±2.8 100% 22:00-8:00 0 - - VVT Attention
or 120 0:00-6:00
0:00-5:00
22:30-0:00
90 PVT Attention
Oriyama 2017 yes no Laboratory 12 22.2±0.4 0% 12 22.2±0.4 0% 22:00-9:00 0:30-2:00 0 60 -
or 120 Mental calculation Executive f unction
0:30-5:00
Real work
Purnell 2002 yes no (aircraf t 24 34.8±10.1 100% 24 34.8±10.1 100% 19:00-7:00 20 0:00-3:00 - - - MCVT Attention
engineers)
1:20-1:50
0:00-1:50
Sallinen 1998 yes no Laboratory 13 - 100% 13 - 100% 23:00-7:10 30 10 - 110 2RTT Attention
4:10-4:40
3:50-4:40
Real work
0:30-1:10 0
Signal 2009 yes no (air-traf f ic 26 35.4 - 26 35.4 - 22:30-6:00 40 - - 2RTT Attention
2:30-3:10 20
controllers)
20 1:40-2:00
Signal 2012 yes no Laboratory 12 25.1±4.3 100% 12 25.1±4.3 100% 18:00-8:00 or 40 1:20-2:00 0 60 - WMT Memory
or 60 1:00-2:00
Real work 60
Smith 2007 yes yes 9 45.7±13.2 33.3% 9 45.7±13.2 - 20:30-7:00 30 2:00-2:00 - - PVT Attention
(nurses) 120
Real work PVT Attention
Smith-Coggins
no yes (nurses and 26 30.0±5.5 19.2% 23 30±4.3 47.8% 19:30-7:30 40 3:00-4:00 0 - - CathSim Executive f unction
2006
physicians) PRMT Memory
2 groups: 2 groups: Typing Executive f unction
Takeyama 2002 yes no Laboratory 20 100% 20 100% 22:00-6:00 120 2:00-4:00 0 60 -
5 and 8 5 and 8 Mental calculation Instrumental f unction
2 groups: 2 groups: PVT Attention
Tremaine 2010 yes yes Laboratory 22.2±2.5 100% 22.2±2.5 100% 1:55-7:15 30 2:30-3:00 15 45 105
21 or 24 21 or 24 DSST Instrumental f unction
Real work DSST Instrumental f unction
Zion 2019 yes yes 109 39.0±9.1 0% 109 39.0±9.1 0% 23:00-7:00 30 4:00-04:30 - - 150
(nurses) LCT Attention
AOT, Arrow orientation task; CathSim, Intravenous insertion simulation; DSST, Digit Symbol Substitution Test; IPI, Information Process Index; LCT, Letter Cancellation Task; MCVT, Mackworth Clock Vigilance Time; PRMT, Probe Recall Memory Test;
PVT, 10min-Psychomotor Vigilance Test; TUAT, Taiwan University Attention Test; VVT, Visual vigilance task; WCST, Wisconsin card sorting test; WMT, 4min-Two Back Working Memory Task; WMDRT, Working Memory Digit Recall Test; 2RTT, Two
Choice Visual Reaction Time Test.
 Napping and cognitive performance during night shifts:
a systematic review and meta-analysis

Frédéric Dutheil 1,2*, Brice Bessonnat 3, Bruno Pereira4, Julien S Baker5, Fares Moustafa6,
Maria Livia Fantini 7, Martial Mermillod8,9, Valentin Navel 10

1
Université Clermont Auvergne, CNRS, LaPSCo, Physiological and Psychosocial Stress, CHU
Clermont-Ferrand, University Hospital of Clermont-Ferrand, Preventive and Occupational
Medicine, Witty Fit, F-36000 Clermont-Ferrand, France; email: [email protected]
2
Australian Catholic University, Faculty of Health, School of Exercise Science, Melbourne,
Australia; email: [email protected]
3
CHU Clermont-Ferrand, University Hospital of Clermont-Ferrand, Preventive and
Occupational Medicine, F-36000 CHU Clermont-Ferrand, France; email:
[email protected]
4
CHU Clermont-Ferrand, University Hospital of Clermont-Ferrand, Biostatistics Unit, the
Clinical Research and Innovation Direction, F-36000 Clermont-Ferrand, France; email:
[email protected]
5
Centre for Health and Exercise Science Research, Department of Sport, Physical Education
and Health, Hong Kong Baptist University, Kowloon Tong, Hong Kong; email:
[email protected]
6
CHU Clermont-Ferrand, University Hospital of Clermont-Ferrand, Emergency, F-36000
Clermont-Ferrand, France; email: [email protected]
7
Université Clermont Auvergne, NPsy-Sydo, CHU Clermont-Ferrand, University Hospital of
Clermont-Ferrand, Sleep disorders, F-36000 Clermont-Ferrand, France; email:
[email protected]
8
Univ. Grenoble Alpes, LPNC, F-38000 Grenoble, France & CNRS, LPNC, F-38000
Grenoble, France ; email: [email protected]
9
Institut Universitaire de France, F-75000 Paris, France ; email: martial.mermillod@univ-
grenoble-alpes.fr
10
CHU Clermont-Ferrand, University Hospital of Clermont-Ferrand, Ophthalmology, F-63000
Clermont-Ferrand, France; email: [email protected]

*Corresponding Author: Frédéric Dutheil, CHU Clermont-Ferrand, Santé Travail

Environnement, 58 rue Montalembert, 63000 Clermont-Ferrand, France, Tel: +33 6 88 22 48
48, Fax: +33 4 73 27 46 49, [email protected]

30
 Appendix 3

Details of the tests used in the included articles to measure

cognitive performance

Executive function test and conditions of the analyses

Arrow orientation task (AOT) was reported in one article.39 AOT is a spatial stroop task.

A fixation cross was first presented on a computer. A white arrow was then presented above or

below. The stimuli consisted of 2 compatible and 2 incompatible stimuli in relation to the

direction in which the arrow was pointing. The task was to respond to the pointing direction of

the arrow (up or down by pressing a button with the corresponding hand. Four outcomes were

reported: reaction time compatible, reaction time incompatible, number of correct responses

compatible, number of correct responses incompatible.39

Mental calculation was reported in three article.48,49,53 The task consisted of performing

single-digit mental calculations. The reported outcome was the number of calculations in 10

minutes.

Driving simulation - 40-min computer-based driving simulator task (York Computer

Technologies) - was reported in one article.40. A five-minute circuit consisting of straight roads

and gentle corners was repeated for the 40 min duration of the task. Subjects were instructed to

adhere to the speed limit of 80 to 100 kph as closely as possible and to stay in the center of the

left lane. The outcomes reported were lateral deviation and percentage of time spent in a safe

zone.40 The outcome “risk of a first crash” was not reported due to missing data.

31
AusEdTM driving simulator was reported in one article. 47 Subjects undertook a 30-minute

monotonous night-time drive on a two-lane road with a series of curved and straight sections.

Subjects had to brake in response to coming up behind 10 other vehicles. Subjects were asked

to stay in the center of the left lane and keep their speed between 60 and 80 kph. The outcomes

reported were variations in lateral deviation and speed.47

Intravenous insertion simulation was reported in one study.18 The CathSim intravenous

insertion virtual reality simulation is a validate computer-based intravenous insertion

simulation (HT Medical Systems, Inc., Gaithersburg, MD).80 The reported outcome was the

elapsed time from tourniquet placement to successful completion of the procedure.

Typing was reported in one article.53 The task consisted of typing figures listed randomly on

sheets into a computer for 20 min. The reported outcome was typing speed. The outcome “error

rate” was not reported due to missing data.

Wisconsin card sorting test (WCST) was reported in one article.16 The computerised

WCST is considered a measure of frontal lobe ability. It consists of four stimulus cards and 128

response cards that depict figures of varying forms, colors, and numbers. Subjects had to match

each consecutive card from the deck with one of the four stimulus cards that they thought it

matched. The reported outcomes were the number of preservation errors, the number of total

errors, the number of categories, the percent of conceptual level responses, and failure to the

test.16

Attention Test and Condition of Analyses

10 min-Psychomotor Vigilance Test (PVT) was reported in eight articles.15,18,41-43,47,51

It is a common and simple reaction time task.62 Subjects have to respond to a simple visual

32
stimulus as quickly as possible by pressing a button.41 The reported outcomes were mean

reaction time or the mean of the 10% fastest reaction times or the number of lapses, defined as

a reaction time greater than 500 ms.

Two Choice Visual Reaction Time Test (2RTT) was reported in one article.50 It is a

subtest of the NIOSH (National Institute Organization for Safety and Health) fatigue test

battery.81 Subjects were randomly presented with the words ‘true’ and ‘false’ on the screen and

instructed to answer as quickly as possible. The reported outcome was the mean reaction times.

Mackworth Clock Vigilance Time (MCVT) was reported in one article.49 Subjects

were presented with a circle of 25 dots on a computer screen. Each dot flashed in a clockwise

sequence. When a dot flash was omitted, subjects were required to press a response button as

quickly as possible. The reported outcome was the mean response latency.49

Taiwan University Attention Test (TUAT) was reported in one article.16 Subjects had

to cross out two target characters from a random list of 780 letters, numbers and symbols as

fast and as accurately as possible. The reported outcome were completion time, number of

omission and a speed score (number of omissions subtracted from the correct number of

characters divided by completion time). 16

Visual vigilance task (VVT) was reported in one article.48 The VVT required the subjects

to press the corresponding key to a number presented in the center of a computer screen as

quickly and accurately as possible. The reported outcome were reaction time and response

lapses over 5 s.

Letter Cancellation Task (LCT) was reported in one article.50 Participants analysed a

matrix of 340 capital letters in English and correctly crossed out two predetermined target

letters (E and R) within 90 seconds while making as few errors as possible. The capacity (the

number of letters the participant scanned), omissions and commission errors were recorded.

33
Instrumental Function Test and Condition of Analyses

Mental Calculation was reported in one article.53 The task consisted of performing

mental arithmetic with two adjacent figures listed randomly on sheets. The reported outcome

was the speed of calculation. The outcome “number of errors” was not reported due to

missing data.

Information Process Index (IPI) was reported in one article.16 It is a generic score

obtained by using age to convert a raw score of two subsets of the Wechsler Adult Intelligent

Scale82 involving cognitive, perceptual and motor abilities: the Symbol Digit Substitution Test

(SDST) and the Symbol Searching Test (SST). In the SST, subjects have to respond to one of

two target symbols from four selective symbols. 16 The reported outcome was the general score.

Digit Symbol Substitution Test (DSST) was reported in three article.15,42,50 It is a

subtest of the Wechsler Adult Intelligence Scale.82 The subject is required to associate certain

symbols with certain other symbols. The reported outcome was the number of correct

responses.

Memory Test and Condition of Analyses

4 min-Two Back Working Memory Task (WMT) was reported in one study.52 It

requires participants to compare the spatial location of a stimulus with the second item shown

two images before in a sequentially presented list of items. Participants are instructed to respond

as quickly and as accurately as possible. The reported outcomes were the number of correct

matches, the mean reaction time for correct matches and omissions.52

Probe Recall Memory Test (PRMT) was reported in one study.18 Subjects had 30

seconds to memorize 6 word pairs.18 The reported outcome was the number of correctly recalled

words.

34
Working Memory Digit Recall Test (WMDRT) was reported in one study.47 Adapted

from the psychological test battery, this numerically based recall test (performed on a computer)

required participants to memorize a string of seven numbers (presented visually and

individually, one after the other) and input them after a 2-s (simple condition) or 4-s (complex)

delay using a key board.

35