Introduction

Sleep deprivation (SD) is especially common in sustained military operations, and in some sport events (Vanhelder and Randomski 1989). In humans, SD is well known to have negative effects on cognitive and neurobehavioral functions (Harrison et al. 2000; Horne 1993; Pilcher and Huffcutt 1996; Thomas et al. 2000). SD is also associated with alterations of different physiological functions, such as metabolism (Van Cauter et al. 1991; Spiegel et al. 1999; Gonzalez-Ortiz et al. 2000; Knutson et al. 2007), neuroendocrinology (Van Cauter et al. 2007; Lusardi et al. 1999) and inflammation (Dimitrov et al. 2004; Irwin 2002; Meier-Ewert et al. 2004; Shearer et al. 2001; Vgontzas et al. 1999). SD may also compromise cardiovascular regulation, and increase the risk of cardiovascular diseases (Shamsuzzaman et al. 2003; Meier-Ewert et al. 2004).

Heart rate has been reported either to decrease or to be unaffected by SD (Bond et al. 1986; Burgess et al. 1997; Holmes et al. 2002; Chen 1991; Kato et al. 2000; Ogawa et al. 2003; Zhong et al. 2005). However, very few studies have examined the association between SD and cardiovascular regulation. Muscle sympathetic nerve activity (MSNA) has been reported to decrease after an acute SD period of one night (Kato et al. 2000; Ogawa et al. 2003). In addition, 30 h of SD has been shown to decrease sympathetic activity and to have no effect on parasympathetic activity (Holmes et al. 2002). In contrast, 36 h of SD increased sympathetic and decreased parasympathetic cardiac modulation (Zhong et al. 2005). To our knowledge, no study has investigated the effects of SD on cardiovascular regulation over a period of more than 42 h. The aim of the present study was to assess the effect of 60 h of SD on cardiovascular regulation.

Methods

Subjects

Twenty healthy cadets (17 men, 3 women, age 26 ± 3 years) volunteered to participate. Their health and physical activity status were assessed by a questionnaire. All subjects were healthy and physically active, performing moderate intensity exercise 3–5 times per week. Based on a circadian rhythm questionnaire, the subjects were classified as morning types and their daily rhythm of life was regular (Folkard et al. 1979). Body mass was measured before (79.8 ± 11.0 kg) and after (80.3 ± 10.7 kg) the SD period. Subjects kept sleep logs for four nights preceding the study (mean 6:52 ± 2:28 h/night). The study was approved by the Ethical committee of the University of Jyväskylä prior to its initiation. All subjects provided written informed consent. (Table 1)

Table 1 Mean ± SE of heart rate variability during the active orthostatic test while sitting
Full size table

Sleep deprivation period

During the 60-hour SD period, the subjects were not allowed to drink caffeine containing liquids, but water could be consumed ad libitum. Five of the subjects were habitual smokers, and were allowed to smoke in accordance with their usual daily habits. However, smoking was not allowed within 3 h of the measurements. Subjects had dinner and lunch each day at the same time (11:00 and 17:00), and also ate snacks that were reported in a diary. During SD, physical activity was restricted to a minimum, while subjects performed military tasks related to tactics. The subjects were allowed to engage in various activities such as playing cards, reading books, watching videos, studying or doing job-related work. Throughout the entire SD period, subjects were constantly observed by the research personnel to prevent them from falling asleep.

Study protocol

Subjects were measured twice a day, in the mornings and evenings, resulting in a total of six measurements after 2, 14, 26, 38, 50 and 60 h of SD. Five of the six measurement periods took place in the morning at 08:00 and in the evening at 20:00, except the final evening measurement period, which took place at 18:00. Each measurement period consisted of blood pressure (BP) and body temperature (BT) measurements followed by heart rate variability (HRV) recordings in an active orthostatic test (AOT).

Measurements

Body temperature was measured from the ear (Omron, MC-510-E, Omron Healthcare Co. Ltd, Japan), and BP with an automatic monitor (Omron, HEM-705c, Omron Healthcare Co. Ltd, Japan). BT was measured twice and BP three times in each measurement session. For BT, the average of the two values was used in the statistical analysis. For BP, the average of the two most similar results was used for further analysis. HRV was measured during AOT in a group of five subjects. AOT included 5-min sitting followed by 3-min standing. HRV was measured beat by beat with wrist monitors (T6, Suunto Oy, Finland) and downloaded to a computer for analysis. For AOT, a 2-minute period was analyzed in the sitting position before standing up. In the standing position, the last 2 min were analyzed. All R-R intervals were first edited automatically (Polar Precision Performance SW 4.03.043), and then visually inspected to exclude measurement errors and ectopic heart beats. Each edited R-R interval was replaced with an average value according to the length of the error sequence, taking the previous and next normal R-R interval into account for correction (Jurca et al. 2004). An autoregressive method was used to estimate the power spectrum of HRV by calculating the power of two frequency bands: high frequency (HF) (0.04–0.015 Hz) and low frequency (LF) (0.15–0.40 Hz) power. HF power is predominantly a reflection of vagal activity, whereas LF power is considered to reflect both sympathetic and vagal activity or sympathetic modulation. The LF/HF ratio is suggested to reflect sympathovagal balance or sympathetic modulation (Task Force. 1996). HRV parameters were expressed in natural logarithmic transformed values (ln) (Tables 1, 2).

Table 2 Mean ± SE of heart rate variability during the active orthostatic test while standing
Full size table

Statistical analysis

Repeated measures ANOVA was used to identify changes in BT and BP between the measurements by applying Bonferroni post hoc tests. In addition to testing differences between each of the six measurement points (BT and BP), data were compared across the 3 days by treating the morning and evening measurements as one measurement (HR variables). Sixteen percent of the HRV data contained errors to such an extent that they were excluded from the analysis. This led to an incomplete data set in some subjects. Consequently, HR variables were analyzed using a mixed model approach with compound symmetry to detect changes during SD. The significance level was set to 0.05.

Results

Active orthostatic test

In the sitting position, HR decreased between days 1 and 3 (P < 0.001) and between days 2 and 3 (P < 0.001) (Fig. 1), while HF power increased (P < 0.001) in both intervals (Fig. 2). LF power increased between days 1 and 3 (P < 0.001).

Fig. 1
figure 1

Changes in HR in the sitting (dashed line) and standing positions (solid line) in AOT. ***P < 0.001 compared to Day 1; ■■■P < 0.001 compared to Day 2; ■■P < 0.005 compared to Day 2

Full size image
Fig. 2
figure 2

Changes in HF power in the sitting (dashed line) and standing positions (solid line) in AOT. ***P < 0.001 compared to Day 1; ■■■P < 0.001 compared to Day 2

Full size image

In the standing position, HR decreased between days 1 and 3 (P < 0.001) and between days 2 and 3 (P < 0.005) (Fig. 1). HF and LF power increased between days 1 and 3 (P < 0.05 and P < 0.001, respectively) (Fig. 2), and LF power also increased between days 1 and 2 (P < 0.001).

Body temperature decreased both in the morning (36.3 ± 0.4, 35.9 ± 0.4, 35.7 ± 0.3°C in the three consecutive mornings) and in the evening (36.4 ± 0.4, 36.1 ± 0.3, 35.9 ± 0.4°C) measurements between days 1 and 2 and days 1 and 3 (P < 0.05–0.001) (Fig. 3). No changes were detected in BP in the morning (141/79, 139/75, 138/75 mmHg) or evening (146/80, 142/78, 140/80 mmHg) measurements for systolic or diastolic pressure (Fig. 4). Changes in HRV indices and HR were not correlated with the changes in BT during SD.

Fig. 3
figure 3

Changes in BT in the evening (dashed line) and morning (solid line) measurements. *P < 0.05 compared to Day 1; ***P < 0.001 compared to Day 1

Full size image
Fig. 4
figure 4

Changes in BP in the evening (dashed line) and morning (solid line) measurements

Full size image

Discussion

The novel finding of the present study was that 60 h of SD decreased HR and increased vagal activity. SD also resulted in decreased BT, while no changes were observed in BP. These changes were cumulative over the 60-h time period.

Previous studies have reported that HR either decreases or is unaffected by SD (Bond et al. 1986; Burgess et al. 1997; Holmes et al. 2002; Chen 1991; Kato et al. 2000; Ogawa et al. 2003). In the present study, HR decreased during SD. Similarly, Holmes et al. (2002) and Chen (1991) reported a decrease in HR after 30 h of SD, and Bond et al. (1986) after 42 h of SD. Zhong et al. (2005) also reported that HR measured in a supine position decreased after 12 and 36 h of SD, as well as after 24 h in a seated position. However, Kato et al. (2000) and Ogawa et al. (2003) found no change in HR after one night of SD. Based on the results of the present and previous studies, it seems likely that when SD lasts for more than one night, a decrease in HR can be observed (Bond et al. 1986; Chen 1991; Holmes et al. 2002; Zhong et al. 2005), whereas shorter periods of SD (e.g., one night) do not seem to affect HR as consistently (Kato et al. 2000; Ogawa et al. 2003).

In our study, the decreased HR was associated with increased vagal activity, while Holmes et al. (2002) have shown no changes in parasympathetic activity but decreased cardiac sympathetic activity after 30 h of SD. In contrast, Zhong et al. (2005) reported that 36 h of SD lead to increased cardiac sympathetic activity and decreased parasympathetic activity. However, Kato et al. (2000) and Ogawa et al. (2003) showed decreased muscle sympathetic nerve activity using MSNA measurements after one night of SD. It has been shown that SD has a perturbing effect on cardiac sympathovagal balance, although no clear consensus has been reached. Compared to previous studies that have observed a reduction in HR due to SD (Holmes et al. 2002; Zhong et al. 2005), the regulation of sympathetic and parasympathetic branches of the cardiac autonomic nervous system appeared to differ in the present study. The factors responsible for these discrepant findings in cardiovascular parameters are not well known, but may include methodological differences, the duration of SD, body position, physical and psychological activity and isolation from versus interaction with other people during SD.

Blood pressure (BP) has been shown to increase due to SD (Kato et al. 2000; Ogawa et al. 2003). In contrast, BP has also been reported to be unaffected by SD (Zhong et al. 2005). The findings of the present study are in agreement with those of Zhong et al. (2005) after 36 h of SD, but are in contrast to the findings of Kato et al. (2000), who found increased mean arterial pressure, and Ogawa et al. (2003), who observed an increase in diastolic pressure. Both findings were observed after 24 h of SD.

We concur with Holmes et al. (2002), who suggest that the changes in HR and sympathovagal activity after SD may function as a protective mechanism in an acute stress situation like SD. SD can also impose increased demands on cardiac function due to prolonged waking hours. However, SD has been shown to decrease alertness and cognitive function, as well as decreasing brain activity measured by cerebral glucose metabolic rate (CMRglu) in the thalamus, cerebellum, temporal cortex and prefrontal and posterior parietal cortices (Wu et al. 1991; Thomas et al. 2000). In the present study, BT decreased during SD and circadian rhythm was maintained, as shown in previous studies (Ax and Luby 1961; Kolka et al. 1984; Lubin et al. 1976; Horne 1983; Horne and Pettitt 1985). According to the literature, the hypothalamus is involved in the regulation of both body temperature and the autonomic nervous system via descending and ascending pathways from the cerebral cortex or the basal forebrain (Kandel et al. 2000). If reductions in brain activity similar to those reported by Wu et al. (Wu et al. 1991) and Thomas et al. (Thomas et al. 2000) also occurred in the hypothalamus, it could be suggested to have an impact upon down regulative outcomes, such as decreased BT and HR with increased vagal activity. Thus, the main effect of SD would be to alter the function of regulative centers in the brain, which could be seen as reflective to changes in BT, HR and vagal activity.

In the present study, certain limitations have to be considered. We did not include a control group to ensure that SD was primarily responsible for any changes that were observed, although the role of other factors was obviously negligible in the controlled environment of this study. We also did not record respiratory parameters during HRV-measures in AOT. As respiratory frequency affects cardiac autonomic modulation (Eckberg 2003), changes in respiration could have had an effect on the results. However, volunteers were carefully instructed to breathe normally. In addition, SD has been shown to have no effect on basal breathing patterns and resting ventilation (White et al. 1983; Ballard et al. 1990; Neilly et al. 1992; Spengler and Shea, 2000). We detected no changes in BP, although systolic pressure remained relatively high throughout the SD period considering that the subjects were healthy young cadets. This could largely be due to the restricted time schedule of this study. Subjects were also measured in pairs, and in front of a group of other subjects. This may have caused BP to increase due to nervousness and anxiety. Habitual smokers were allowed to smoke as normal, despite previous evidence suggesting that smoking increases sympathetic and decreases parasympathetic activity at rest (Narkiewicz et al. 1998). Nonetheless, in separate analyses of HRV and BT, no changes were observed after excluding smokers. In addition, smoking was not allowed 3 h before each measurement, and prohibition of smoking could have resulted in significant withdrawal symptoms, which could also have affected sympathovagal balance.

In summary, the accumulation of 60 h of SD resulted in decreased HR regulated by increased vagal outflow, and decreased BT, although circadian variation was maintained. In addition, cardiovascular regulation seems to be modulated differently after sleep deprivation of 2 or 3 days compared to just 1 day.