The effects of a 30-min nap during night shift following a prophylactic sleep in the afternoon

Abstract

The purpose of this study was to investigate the effects of a 30-min nap, during a simulated night shift environment, when a prophylactic daytime sleep was implemented prior to the night shift. A repeated-measures counterbalanced design was used which included two experimental conditions: a 30-min nap and a no nap control. In both conditions subjects obtained a 2-h sleep in the afternoon from 15.00–17.00 hours, which was followed by the night-time nap from 02.30–03.00 hours in a controlled laboratory environment. Post-nap testing was conducted from 03.10 to 07.00 hours. The participants included 22 adults aged from 18–35 years who were good sleepers and did not regularly nap. Subjective alertness (Stanford Sleepiness Scale, Karolinska Sleepiness Scale, Visual Analog Scale), fatigue and vigor (Profile of Mood States), cognitive performance (psychomotor vigilance task, symbol–digit substitution task, letter cancellation task), and objective sleepiness were measured pre- and post-nap. The 30-min nap resulted in some impairment of subjective alertness for a brief period (up to 30 min) immediately following the nap when compared to the no nap condition. Following this brief period, alertness and performance were generally improved by the 30-min nap from 04.00 hours until the end of the testing period at 07.00 hours. The results of this study indicate that when a 2-h prophylactic sleep is implemented in the afternoon, a 30-min nap during the subsequent simulated night shift overall provides a significant countermeasure against sleepiness and performance impairment.

INTRODUCTION

The escalating trend in night shift work over recent years has resulted in the increased sleepiness^1–4 and decreased cognitive and psychomotor abilities^5,6 of employees in many workplaces. Consequently, the probability of serious human error and accidents in the workplace has also increased. Many researchers have suggested potential countermeasures against these detrimental effects, including napping during night shift work.^7–14

Research has consistently demonstrated that a daytime nap results in increases in subjective and objective alertness and improved vigilance following the nap.^3,15,16 Although the majority of research has focused on daytime napping, the alerting effects of a night-time nap has been found to be delayed after waking and last for only a short period of time.^17,18

Naps of 60 min or longer taken during a night shift have been shown to improve subjective sleepiness and fatigue, relative to having no nap.^10,18–21 A brief nap of 30 min or less has also demonstrated improved objective and subjective alertness, in addition to fatigue and vigilance.^12,22 Generally longer naps have been demonstrated to be more recuperative than brief naps, with alerting benefits lasting for much longer after waking.²³ Although the benefits of longer naps are more durable, the alerting benefits of brief naps are evident much sooner, within 10–15 min of waking.²⁴ Individuals who have had a long nap usually do not report any improvements in their alertness within the first hour after waking, often feeling more sleepy and fatigued than they did before the nap.^18–20

The increase in sleepiness experienced immediately after waking is referred to as sleep inertia. Although there are a number of factors which can contribute to sleep inertia,^25,26 the three-process model of alertness proposes that the duration of sleep inertia is proportional to the amount of slow-wave sleep contained within a sleep episode.^9,18,19,22 Longer naps, which contain a greater quantity of slow-wave sleep, result in a longer period of sleep inertia after waking.²³ The amount of slow-wave sleep within a sleep episode is also proportional to the wake time prior to the sleep episode.²³ Therefore, an individual who has been awake for many hours (e.g. 20 h) will experience more slow-wave sleep within their sleep period and more sleep inertia after waking. Long periods of sleep inertia can be dangerous within the workplace. Human error and accidents are more likely to occur during a period of sleep inertia, just as they are when an individual is very sleepy and fatigued prior to napping.^9,23 Therefore, when considering napping as a countermeasure against sleepiness within the workplace, it is important to take precautions to minimize sleep inertia.

This study will assess the alerting benefits of a 30-min nap, compared to no nap, during a simulated night shift environment. A short nap has been chosen both to minimize the sleep inertia following the nap (relative to longer naps) and because it is more practical than a longer nap in most night shift circumstances. The three-process model of alertness predicts that by decreasing wake time (or sleep pressure) before the brief nap, less slow-wave sleep will be contained within the nap, consequently reducing sleep inertia.⁹ Therefore, this study also incorporated a long afternoon sleep prior to the night shift in order to reduce sleep inertia and allow improvements in alertness to be observed soon after the night shift nap. Research has shown that many shift workers take an afternoon nap prior to a night shift, in particular prior to a first night shift.^27,28 To the knowledge of the authors, no other research has examined the alerting benefits of a night shift nap under the conditions of a prior afternoon sleep.

METHODS

Participants

The study included 22 participants (13 females, nine males) aged between 18–35 years (mean = 22.48, SD = 2.95) who were recruited via local advertisements. Participants were paid $A300 for their participation.

Participants were excluded if they napped habitually (>3 naps per week), suffered from any condition or consumed medication which could affect their sleep (such as depression, pain, or asthma), regularly consumed excessive caffeine (>4 cups per day) or alcohol (>2 standard drinks per day), suffered from excessive daytime sleepiness or were extreme morning or evening types²⁹ (mean = 45, SD = 8.7), or had undertaken shift work or jet travel within the 3 months prior to their participation. Participants with a body mass index greater than 28 or less than 19 were excluded from the study as populations within these body mass index categories are more likely to suffer from sleep-related disorders such as obstructive sleep apnea or anorexia, respectively.^30–32

The study received approval from the Flinders University Social and Behavioural Research Ethics Committee and the University of South Australia's Human Research Ethics Committee.

Test instruments

The subjective alertness test battery included the Stanford Sleepiness Scale (SSS), Karolinska Sleepiness Scale (KSS), fatigue and vigor subscales of the Profile of Mood States (POMS), and the Visual Analog Scale for sleepiness (VAS).

Cognitive functioning was measured using a test battery comprising the symbol–digit substitution task (SDST), the letter cancellation task (LCT), and the psychomotor vigilance task (PVT). The LCT involved participants searching for and marking two different target letters in a matrix of letters. During the SDST, participants were presented with a key of nine symbols paired with digits between 1 and 9, in addition to a random sequence of these symbols. Participants were instructed to identify and copy the digit corresponding to each symbol as quickly and accurately as possible. The outcome measure derived from the SDST and LCT was the mean number of correct responses within a 90-s period (SDST) or a 2-min period (LCT). Parallel forms of the SDST and LCT were developed for each testing occasion to minimize learning effects. The PVT involved participants reacting to visual stimuli presented on a personal digital assistant.³³ Participants were asked to press a button as soon as they saw a target symbol appear on the screen. Eighty stimuli were presented at random intervals (ranging from 3–9 s) over a total testing period of approximately 10 min. Outcome measures were mean reaction time (msec) and number of lapses (reaction time > 500 msec).

Participants were trained to plateau levels of performance on all test instruments. This test battery has been successfully used by previous studies.^3,34,35

Sleep latency measure

The sleep latency test was used as an objective measure of alertness. The sleep latency was calculated as the latency (min) it took participants to fall asleep from the moment the lights were turned off until the first of three consecutive epochs (one epoch = 30 s) of any stage of sleep were observed. Sleep latency was determined using computerized software (PSG, E-Series; Compumedics, Melbourne, Australia). The baseline sleep latency measure was the latency to the nap sleep with post-nap measures taken at 75, 135, 195, and 250 min after waking from the nap. Each sleep latency trial was limited to 20 min in duration to ensure accurate scheduling of later post-nap testing. Participants who satisfied the sleep onset criteria within the 20-min period were awoken and continued with quiet activity for the remainder of the 20-min period. Participants who were unable to satisfy the sleep onset criteria within the 20-min were recorded as having 20-min sleep latency (<5% of all sleep latency trials). Earlier research has shown that the 90 s of sleep required in this method of determining sleep onset has no alerting effects.^34,35

Design

The study utilized a repeated-measures design consisting of two experimental conditions: a 30-min nap condition and a no nap condition. The experimental conditions were randomly counterbalanced across participants to minimize order effects.

Prior to the laboratory sessions

Participants were required to maintain regular bed and wake times for at least 3 days prior to each laboratory session. Sleep diaries and ActiTrac wrist activity monitors (IM Systems, Baltimore, MD, USA) were used throughout this period to confirm compliance to these instructions.

Instructions were given to participants to refrain from consuming caffeine, alcohol, and any medications (except for the oral contraceptive pill) for at least 3 days prior to each laboratory session. Experimental sessions were scheduled at least 1 week apart to ensure the participants had recovered from the sleep deprivation incurred during the previous laboratory session. All laboratory sessions were conducted in the sleep laboratory at the Centre for Sleep Research at the University of South Australia.

Laboratory sessions

Participants arrived at the laboratory at 12.00 hours after waking at 07.00 hours that morning. Time-keeping devices such as mobile phones, watches, and clocks were removed from the participants upon arrival as such devices may influence subjective feelings of sleepiness. At 14.40 hours participants were equipped with electroencephalogram (C3/A2, O2/A1), electro-oculogram, and electromyogram electrodes for standard recording. Participants were then confined to their individual bedrooms, which contained a bed, desk, cupboard, and telephone. At 15.00 hours the lights in the participants' rooms were switched off. Participants were then given a 2-h sleep opportunity and woken at 17.00 hours. Participants remained in the laboratory after their nap until the experimental period began at approximately 02.15 hours. Participants were permitted to engage in quiet activities during this time, such as watching DVDs, reading, or working on the computer. Dinner was provided for participants between 18.00–19.00 hours.

Each laboratory session was comprised of seven periods of subjective alertness testing, five periods of cognitive testing, and five sleep latency trials. The 30-min nap was scheduled to begin at 02.30 hours. The experimental protocol is summarized in Table 1. Baseline testing began at 02.15 hours with post-nap subjective alertness test beginning at 10, 40, 70, 130, 190, and 250 min after waking from the nap. The post-nap cognitive testing began at 15, 45, 105, 165, and 225 min after waking from the nap. The pre-nap sleep latency measure was the time taken until sleep onset of the nap itself, with post-nap latency measures taken at 75, 135, 195, and 250 min after waking from the nap. An additional sleep latency measure was conducted at 02.55 hours, in the no nap condition, to provide an awakening from sleep in the no nap condition to coincide with the 03.00 hours wake-up time in the 30-min nap condition. This procedure was undertaken to equate the conditions prior to the post-nap testing as much as possible. Thus all post-nap testing occurred at the same delays from waking up from the 30-min nap or waking from the 1.5 min of sleep latency determination in the no nap condition. This also equated the circadian phase of the post-nap testing of both nap conditions during a rapidly changing sleep propensity across the post-nap testing period (03.00–07.00 hours).²³ Experimenters woke participants from the sleep latency trial by ringing the phone located in the participants' bedroom. The same procedure was undertaken to wake participants from unintentional sleep during the experimental period. In between periods of testing and sleep latency trials, participants engaged in quiet activities such as watching DVDs or reading within their bedroom environment. The bedrooms were sound attenuated and light intensity was maintained at 50 lux between sleep latency trials, and the temperature was also controlled at 22°C.

Table 1.  Summary of experimental protocol

Data screening and preliminary analyses

Table 2 shows the preliminary analyses that were conducted to assess whether participants differed in their level of alertness prior to either nap condition. The separate paired samples t-tests show that participants did not differ in the mean bedtime on the three nights prior to each nap condition, the mean total sleep time on the night prior to each nap condition, or the total sleep time or any sleep stage during the afternoon sleep prior to each nap condition. There were no significant differences between mean baseline alertness scores for any dependent variable, with the exception of the letter cancellation task, which showed better pre-nap scores in the no nap condition.

Table 2.  Paired samples t-tests for mean bedtime, night-time total sleep time, afternoon total sleep time, and pre-nap scores for all dependent variables

	Condition
	30-min nap Mean (SD)	No nap Mean (SD)	t-test	P-values
Mean bedtimes (3 nights prior to conditions)	11.18 (1.77)	11.13 (1.20)	0.171	0.865
Total sleep time (night prior to conditions – min)	7.71 (1.64)	7.59 (1.34)	0.533	0.596
Afternoon sleep total sleep time (prior to conditions – min)	106.91 (19.75)	105.66 (21.49)	0.512	0.614
Afternoon sleep Stage 1 (min)	3.18 (4.68)	3.82 (6.57)	−0.344	0.734
Afternoon sleep Stage 2 (min)	31.64 (15.72)	32.30 (13.39)	−0.207	0.838
Afternoon sleep SWS (min)	62.70 (23.21)	58.43 (23.68)	10.163	0.258
Afternoon sleep REM (min)	9.39 (11.03)	9.48 (12.65)	−0.033	0.974

				x^2	P-values
Sleep stage prior to waking	Stage 1	18.2%	13.6%	0.170	0.680
	Stage 2	27.3%	45.5%	10.57	0.210
	SWS	45.5%	31.8%	0.863	0.353
	REM	9.1%	9.1%	0.000	10.0

			t-test	P-values
Fatigue, Profile of Mood States fatigue subscale; KSS, Karolinska Sleepiness Scale; Lapses, mean number of lapses; LCT, letter cancellation task; Mean RT, psychomotor vigilance task mean reaction time; REM, rapid eye movement; SDST, symbol–digit substitution task; SL, sleep latency; SSS, Stanford Sleepiness Scale; SWS, slow-wave sleep; VAS, Visual Analog Scale; Vigor, Profile of Mood States vigor subscale.
Pre-nap scores SSS	4.23 (1.23)	4.41 (1.18)	−0.890	0.383
KSS	6.05 (1.79)	6.14 (1.55)	−0.283	0.780
VAS	35.14 (19.71)	39.66 (20.03)	−10.030	0.315
Fatigue	3.08 (1.13)	2.96 (1.04)	0.612	0.547
Vigor	1.90 (.69)	2.08 (.71)	−10.205	0.242
SDST	61.86 (8.36)	64.95 (8.17)	−10.913	0.069
LCT	53.95 (10.69)	59.27 (11.39)	−20.36	0.028
Mean RT	270.11 (27.18)	297.49 (89.98)	−10.48	0.156
Lapses		3.65 (5.36)	−0.142	0.888
SL	7.29 (5.57)	8.36 (5.94)	−10.054	0.304

Overview of statistical analyses and illustration of results

Two-way repeated measures anovas were conducted between nap conditions and testing times for each dependent variable. Two-way interactions were also conducted to investigate the differences between the nap conditions in the amount of change from the baseline to each post-nap testing time. Two-way interactions were used to determine which nap condition produced the least decline in alertness from the baseline to each individual post-nap testing time. To highlight these differences, changes from the baseline value for the two conditions (nap and no nap) are illustrated for all post-nap time-points for each dependent variable. All curves begin at a zero baseline and show subsequent mean changes with all figures indicating lower alertness on the y-axis in the negative direction. Some axes (SSS, KSS, VAS, fatigue, PVT reaction time, and lapses) were reversed to provide this consistent relationship and thus provide easier understanding of the large number of dependent variables.

RESULTS

The trends in alertness from baseline to each post-nap testing across nap conditions is shown for all dependent measures in Figure 1. When no nap is taken (dotted line), there is a steady decline in alertness from the baseline until the final testing time. Compared with this steady decline there was an immediate relatively sharp decline in subjective alertness following the 30-min nap. Table 3 shows the sleep architecture for this nap including the means (SD) for total sleep time, minutes of stage 1, stage 2, slow-wave sleep, and proportions of participants who woke from each sleep stage. It indicates considerable slow-wave sleep over the 30-min nap with 91% of awakenings from slow-wave sleep.

Figure 1. Mean changes in the (a) Stanford Sleepiness Scale, (b) Karolinska Sleepiness Scale, (c) Visual Analog Scale, (d) Profile of Mood States, fatigue subscale, (e) Profile of Mood States, vigor subscale, (f) symbol–digit substitution task, (g) letter cancellation task performance, (h) mean reaction time, (i) mean number of lapses, and (j) sleep latency (percentage change), from baseline across all post-nap testing times, following the 30-min nap condition (dotted line) and no nap condition (solid line). The asterisk indicates time-points at which the decrease in alertness in the 30-min nap condition was significantly different from the decrease of alertness in the no nap condition. Y-axes were reversed for the Stanford Sleepiness Scale, Karolinska Sleepiness Scale, Visual Analog Scale, fatigue, mean reaction time, and lapses.

Table 3.  Architecture of the 30-min night-time nap including mean total sleep time, minutes in stage 1, stage 2, and SWS in addition to the percentage of participants waking from each sleep stage

Sleep parameter	30-min night nap
REM, rapid eye movement; SWS, slow-wave sleep; TST, total sleep time.
TST	33.25 (5.18)
Stage 1	2.25 (1.72)
Stage 2	13.16 (6.68)
SWS	17.84 (6.87)
REM	0
Sleep stage prior to waking
Stage 1	0
Stage 2	9%
SWS	91%

However, following this initial decline in subjective alertness, by about 70 min following the nap almost all dependant variables show better performance and alertness. Table 4 shows the results of two-way interactions between nap condition and testing time for all dependent measures. Significant interactions between nap condition and testing time were indicated for the SSS, KSS, VAS, and the POMS fatigue subscale scores, in addition to SDST performance, sleep latency, and the mean number of lapses during the PVT. The significant interactions indicate different trends in alertness over time as a result of the nap condition with there being less eventual decline in alertness following the 30-min nap.

Table 4.  Two-way repeated measures anova interaction effects between condition (30-min nap vs no nap) and time (pre-nap vs post-nap testing times) for all dependent measures

Variable		anova
		df	F	P-values
*** P≤ 0.001; ** P < 0.05.Greenhouse–Geisser adjustment was reported for df as Mauchly's Test of Sphericity was not met. KSS, Karolinska Sleepiness Scale; LCT, letter cancellation task; POMS, Profile of Mood States; PVT, psychomotor vigilance task; SDST, symbol–digit substitution task; SSS, Stanford Sleepiness Scale; VAS, Visual Analog Scale.
Subjective alertness	SSS	5.73, 120.3	5.37	<0.001***
	KSS	5.04, 105.8	4.85	<0.001***
	VAS	4.34, 91.21	5.38	<0.001***
	POMS
	Fatigue	4.29, 90.03	3.07	0.018**
	Vigor	3.41, 71.63	2.18	0.089
Cognitive functioning	SDST	4.11, 86.35	6.78	<0.001***
	LCT	5105	1.09	0.370
	PVT
	Reaction time	2.24, 40.40	3.016	0.055
	Lapses	3.83, 61.31	3.14	0.022**
Objective alertness	Sleep latency test	2.19, 43.73	4.36	0.016**

Two-way interactions between the 30-min nap and no nap for the change from baseline to each testing time were conducted to disclose the source of the interactions in these measures. All significant interactions are indicated in Figure 1 with an asterisk. The SSS and the KSS measures of subjective alertness indicate significantly decreased alertness 10 min after taking a 30-min nap relative to no nap. However, following this immediate decrease, subjective alertness tends to recover following the 30-min nap producing less decrease than following no nap for almost all measures by 70 min post nap, significantly so for the VAS and POMS fatigue subscale. The sleep latency measure indicated less decrease of alertness at 75, 135, and 195 min following the 30-min nap when compared to no nap.

Two-way interactions for condition by time for cognitive performance measures indicate less impairment of cognitive performance from about 105 min following the 30-min nap. Relative to no nap, SDST performance is significantly better at 165 and 225 min following the 30-min nap. The mean number of lapses also increases less at 165 min following the 30-min nap when compared to no nap.

DISCUSSION

The present study assessed the benefits of a 30-min nap, compared to no nap, during a simulated night shift environment, when a prior prophylactic daytime sleep was utilized. Without the 30-min nap all measures, including objective alertness, cognitive functioning, and subjective alertness indicated a general monotonic decline from baseline at 02.15 hours until the last testing at 07.00 hours. However, following the 30-min nap, apart from a brief decline in subjective alertness, the decline in alertness and performance for all measures was less marked following the 30-min nap than no nap. Some benefits from the 30-min nap are evident from as early as 75 min post nap and last up to 3 h after the nap for most measures.

The only indication of sleep inertia in the present study was a brief but greater decline in subjective alertness immediately following the 30-min nap. This result is consistent with previous findings^18,22,24 which reported periods of reduced alertness immediately after waking from night-time naps of lengths comparable to the 30-min nap used in this study. In line with the three-process model of alertness, the amount of slow-wave sleep contained within the 30-min nap may account for the sleep inertia reported.^9,19,22,36 The polysomnography data indicates that slow-wave sleep comprised slightly over 50% (mean = 17.84, SD = 6.87) of sleep, which is greater than that reported for previous studies which have examined the alerting effects of both a 30-min nap opportunity during night shift³⁷ and a 30-min daytime nap.^3,37 In the current study 91% of participants were woken from slow-wave sleep at then end of the 30-min nap. Research suggests that the sleep stage prior to waking can influence subsequent sleep inertia.²⁵ Awakening during slow-wave sleep produces more sleep inertia than awakening in stage 1 or 2 sleep.²⁵

However, the sleep inertia following the 30-min nap in this study was only significant for subjective measures of sleepiness and not for cognitive functioning and objective alertness. It is therefore not likely that the implementation of a 30-min nap during a night shift, when a prophylactic afternoon sleep is used, would leave workers at an increased risk of injury resulting from impaired performance or falling asleep during the immediate post-nap period.

The importance of developing an effective countermeasure to the sleepiness and fatigue experienced by night shift workers is highlighted by the overall decline in alertness, indicated in the no nap condition, throughout the testing period of this study. This decline in alertness across the testing period is the result of the combination of the increasing homeostatic sleep drive (process S) and the increasing circadian sleepiness (process C).³⁸ Due to these processes, workers are at an extremely increased risk of sleepiness-induced accidents and injury particularly towards the end of their shift. This effect was ameliorated to some extent in all measures of alertness and performance by a 30-min nap at 02.30 hours.

Further research within this area should extend on these findings to compare the 30-min nap benefit when a prior afternoon sleep is not taken. It could be predicted that without a prior daytime sleep, as is common practice before the first night shift,²⁷ the long wake period prior to the night-time nap would increase the amount of slow-wave sleep contained within the nap and consequently greater sleep inertia. The sleep inertia following the night-time nap may reduce the advantage of the 30-min nap over the no nap condition. Other nap lengths such as 10 min, which have shown immediate post-nap benefits when taken in the afternoon,³ could also be investigated within a simulated night shift environment with and without a 2-h prophylactic afternoon sleep. Future research should also consider whether the benefits demonstrated in this study would be maintained across consecutive night shifts when long daytime sleep periods are attempted. It would also be important to investigate whether these benefits are shown across different sorts of work, for example those requiring greater physical activity. The possibility that the results might be different for regular night shift workers could also be explored.

The 30-min nap (in combination with an afternoon nap to reduce prior sleepiness) has yielded promising results when taken during a first night shift. All measures used in this study have indicated greater alertness and cognitive functioning following the 30-min nap when compared to no nap. A brief period of sleep inertia followed the 30-min nap; however, this decline was only significant in a limited number of the subjective alertness measures used in the study. Based on the results of this study, the implementation of a 30-min nap during night shift could be used to maintain a safer work environment for workers and those around them.

ACKNOWLEDGMENTS

This project was supported by the Australian Research Council (ID: DP0558960). The authors would like to thank the Centre for Sleep Research at the University of South Australia for the use of the sleep laboratory.