Sleep deprivation blunts the nighttime increase in aldosterone release in humans

By ANNE CHARL OUX, CLAUDE GRON FIER, FLORIAN CHAP OTOT, JEAN EHR H ART, FRANÇOIS PIQUAR D and GABRIELLE BRANDENBERG E R

Aldosterone, secreted by the glomerula cells of the adrenal gland, stimulates the reabsorption of Na + and the secretion of K + and H + by the distal nephron segments of the kidney. Therefore, Aldosterone plays a key role in homeostatic regulation of the electrolyte balance and extracellular compartment volume. Three factors acting systemically are recognized as the major regulators for aldosterone secretion, i.e. the renin-angiotensin system (RAS), plasma K + , and the adreno-corticotropic hormone (ACTH) (Quinn and Williams 1992).

Previous studies have reported that renin, ACTH, cortisol, and aldosterone display an episodic secretory pattern over a 24-h period. Plasma renin activity (PRA), which is an index of renin release, shows an ultradian rhythm closely linked to the rapid eye movement (REM)-non REM (NREM) sleep cycle. PRA rises during NREM sleep, and decreases during REM sleep (Brandenberger et at. 1988, 1994). In patients with sleep disorders, such as narcolepsy and sleep apnea, PRA variations reflect a disorganization of the internal sleep structure (Follenius et at. 1991; Schulz et at. 1992). In contrast, ACTH rhythm is primarily circadian, as is the cortisol rhythm, which is similar to that of ACTH but with a 10-min lag. ACTH and cortisol show a low secretory activity during the first part of the night, followed by a pronounced pulsatile secretion with increasing amplitude towards the morning (Veldhuis et al. 1990; Pietrowsky et al. 1994). For aldosterone, mean plasma levels, pulse frequency and pulse amplitude increase during the night and the early morning hours (Katz et al. 1972). Initially, some authors concluded that the aldosterone rhythm is under circadian influence (Katz et al. 1972; Katz et al. 1975; Armbruster et al. 1975; Lightman et al. 1981).

However, the results of a recent experiment based on an acute shift of sleep lend support to the hypothesis that the 24-h aldosterone variations are influenced principally by sleep processes (Charloux et al. 1999). The shift of the sleep period from nighttime to daytime was accompanied by a shift of the sleep-induced PRA oscillations, as well as by a shift of the large aldosterone pulses generally observed during the night in the basal condition. Aldosterone pulses were mainly related to PRA oscillations during the sleep periods, whereas they were related to cortisol pulses during the waking periods. Then, if low aldosterone levels are obtained throughout a 24-h sleep deprivation, which would suppress the sleep-induced release in renin, the impact of sleep processes on aldosterone release through the RAS would be confirmed.

The physiological role of the sleep-induced rises in aldosterone plasma levels and their influence on diuresis, natriuresis and plasma volume has not been studied in healthy subjects. Previous results from patients with obstructive sleep apnea (OSA) suggest that the RAS could playa significant role in body fluid and salt regulation during sleep. Indeed, untreated patients often complain of nocturia. Restoration of a normal sleep pattern by continuous positive airway pressure significantly reduces urine flow and Na + excretion and increases plasma volume during the night (Follenius et al. 1991 ).

We designed the present study to determine the influence of sleep deprivation on aldosterone release and on its three major controlling factors: renin, ACTH and plasma K +. The 24-h profiles of aldosterone, PRA and cortisol were established in healthy subjects over a 24-h period, once in the basal condition with normal nighttime sleep, and once during a 24-h sleep deprivation. Plasma Na + and K + were measured in four of these subjects. We also studied the consequences of the changes in aldosterone levels induced by sleep deprivation on urinary variables (diuresis, natriuresis and kaliuresis) in an additional group of subjects, once during normal sleep, and once during sleep deprivation. These subjects were kept under continuous enteral nutrition in order to stabilize fluid and electrolyte intake.

METHODS
Subjects
Two groups of healthy male subjects aged 21-28 years participated in this study. They had no medical history, were  not taking any medication and were nonsmokers. They were included after medical examination, biological screening, and answering questionnaires concerning their usual sleep-wake cycle, behavior and morningness-eveningness (Horne and Ostberg 1976). They gave written informed consent to participate to this study, which was approved by the Strasbourg Hospital Ethics Committee.

Procedures

The experiments were performed in a soundproof, air-conditioned room, equipped for polysomnographic recording and blood sampling. A first group' of eight subjects was studied twice for a 24-h period, with a one-month interval between the experiments, once with normal nocturnal sleep from 23.00 to 07.00 h and once under 24-h sleep deprivation. The experiments were balanced, four subjects getting nocturnal sleep first and the four other subjects getting sleep deprivation first. The subjects were accustomed to the sleep room and experimental conditions the night preceding the experimental session. The daytime before the experiment, they received standard meals at 07.30 hand 12.00 h. They read, listened to music, watched television and conversed with the experimenters. The subjects then remained supine for the 4 h before blood sampling and throughout the experiment. During the experiments, when awake, the subjects had the opportunity to prevent sleep by reading, listening to music etc, as the day before. Standard meals were served the first day at 19.15 h and the second day at 12.00 h, and a standard breakfast at 07.30 h (for this, the head of the bed was titled 30° to the horizontal).

Blood was collected during a 24-h period (from 18.00 to 18.00 h) in an adjoining room. Blood was removed continuously using a peristaltic pump and sampled at 10-min intervals in tubes containing EDTA-K2 for hormone measurements. In four of the subjects, additional blood samples were collected continuously from 23.00 to 07.00 hours in tubes containing lithium heparinate for Na + and K + measurements. A maximum of 200 mL of blood was removed during the 24 h. The samples were immediately centrifuged at 4°C and the plasma was stored at -25°C for subsequent PRA, aldosterone, cortisol and electrolyte analysis.

Sleep recordings were based on four EEG derivations (F3-A2, C3-A2, P3-A2 and C4-AI), one chin electromyogram and one horizontal electro-oculogram (upper canthus of one eye vs. lower canthus of the other eye). Sleep stages were scored according to the criteria of Rechtschaffen and Kales (1968).

A second group of 13 subjects was studied twice from 23.00 to 07.00 h, in random order, with a one-month interval between the two sessions, once during a usual normal nocturnal sleep from 23.00 to 07.00 h, and once during nighttime sleep deprivation. These subjects were on continuous enteral nutrition which began 5 h before the beginning of recording and was maintained throughout the experiment (Sondalis ISO, Sopharga, Puteaux, France; 110 mg 100 mL-1 Na +, 60 mg 100 mL -I K +, 50% carbohydrates, 35% fat, 15% proteins, 378 kJ h-I). From 23.00 to 07.00 h, urine

samples were collected, pooled, and kept at 4°C. At 07.00 h, volumes were measured and samples were stored at -25°C. Urine Na + and K + concentrations were assayed by flame photometry. Haematocrit was measured at 07.00 hours using blood collected in heparinized capillary tubes.

Hormone assays

PRA, aldosterone and cortisol plasma levels were measured over the 24-h period in the first group of eight subjects. PRA was measured by radioimmunoassay of angiotensin I generated after incubation of the plasma (Commercial kits, Sorin Biomedica, Saluggia, Italy). The intra-assay coefficient of variation (CV) for duplicate samples was 4% for levels between 10 and 20 ng mL h'\ 6% for levels between 2 and 10 ng mL-1 h-1; 10% for levels between I and 2 ng mL-1 h-1; 30% for levels below I ng mL -1 h-1. The detection limit was 0.18 ng mL-1 h-1. Plasma aldosterone was measured by radioimmunoassay (Commercial kits, Diagnostic Systems Laboratories, Webster, Texas, USA). Intraassay precision for duplicate samples was 7.5% for aldosterone levels above I ng 100 mL-1 and 10% for levels below I ng 100 mL-1. The detection limit was 0.5 ng 100 mL-1. Plasma cortisol was measured by radioimmunoassay (Commercial kits, Diagnostic Systems Laboratories). The detection limit was 0.2 J1.g dL-1. The interassay CV for duplicate samples was 4% above 6 J1.g dL-1 and 10% for levels below. All the samples from a given subject were measured in the same assay to avoid inter assay variations.

Data analysis

The pulse analysis program ULTRA (Van Cauter 1988) was used for quantitative detection and characterization of PRA, cortisol and aldosterone oscillations. This program takes into account the limit of detection of the analytical procedure and the precision of the assay for various concentration ranges. The threshold was set to be three times the Cv. For each significant oscillation (e.g. if both the increase and the decrease were significant), the ascending portion, the declining portion, the total duration and the increment were calculated.

A two-way ANOVA for repeated measurements with the Greenhouse-Geiser correction and the Student-NewmanKeuls tests for multiple comparisons were used to assess the statistical differences between the mean plasma levels, the mean number per hour and absolute amplitude of the hormonal pulses. Two conditions (nighttime sleep and sleep deprivation), and three periods: 18.00-23.00 h (baseline period), 23.00-07.00 h (nighttime period), and 07.0018.00 h (posteffect period) were considered. A similar procedure was used for haematocrit, plasma Na + and K + calculations. Urinary flow, natriuresis and kaliuresis measured in the two conditions were compared using paired (-tests. The results are expressed as means ± SE. The limit of significance was P < 0.05.

RESULTS
Twenty-four-hour hormonal profiles

The mean 24-h profiles of plasma aldosterone, PRA and cortisol are presented in Fig. I. Sleep deprivation induced a significant decrease in the plasma aldosterone levels during the 23.00-07.00 h period compared to the normal sleep condition (-34%, P < 0.01, Table I). Thus, the differences in plasma aldosterone levels observed among the three periods in the nighttime sleep condition, with the highest levels during the 23.00-07.00-hour period (P < 0.005), disappeared in the sleep deprivation condition. Similarly, PRA levels decreased by 50% during the 23.00-07.00 h period during sleep deprivation (P < 0.01). In contrast, a slight but significant increase in plasma cortisol concentration was found during the three periods in the sleep deprivation condition (18.00-23.00 h: +67%; 23.00-07.00 h: +54%; 07.00-18.00 h: +44%; P < 0.01). In most subjects, meal intake was followed by an abrupt increase and then a slow decrease in PRA levels. This sequence was sometimes followed by small increases in aldosterone levels.

The distribution of aldosterone pulses over the 24-h period did not differ significantly by condition. The number of aldosterone pulses was highest during the 23.00-07.00 h and 07.00-18.00 h periods (P < 0.001) in both conditions (Table I). During the night of sleep deprivation, the number of aldosterone pulses had a tendency to be lower than during night sleep but did not reach statistical significance. In contrast, during the 23.00-07.00 hour period, PRA pulses were less frequent during the sleep-deprived night (-61 %, P < 0.05). The number of PRA pulses was lowest during the
18.00-23.00 h period (P < 0.01). Cortisol pulses were distributed similarly in both conditions.

During nighttime sleep, the 23.00-07.00 h period showed the highest pulse amplitude of aldosterone compared to pre- and postsleep periods (P < 0.001). This amplitude was reduced during the 23.00-07.00 hour period of sleep deprivation compared to normal nocturnal sleep (-48%, P < 0.01). Similarly, for PRA, during the 2300-07.00 hour period, sleep deprivation was characterized by pulses of lower amplitude relative to the nighttime sleep condition (-53%, P < 0.05). The difference between the 23.00-07.00 h period and the 07.00-18.00 h period observed during nighttime sleep (higher PRA pulse amplitude during the sleep period) disappeared in the sleep deprivation condition. Cortisol pulse amplitude did not vary significantly either with the period or condition.

Plasma NA + and K+

Plasma Na + and K + were not influenced by sleep deprivation:

Na+ was 143.06 ± 2.12 mmol L-1 in the normal sleep condition and 143.21 ± 1.57 mmol L-1 during sleep deprivation; K + was 4.28 ± 0.12 mmo1 L-I in the normal sleep condition and 4.38 ± 0.06 mmol L-1 during sleep deprivation (Fig. 2). Plasma Na + and K + showed very small variations over the night: Na + was 1.76 ± 0.07 mmol L-1 in the nighttime sleep condition and 2.47 ± 0.41 mmol L -I during sleep deprivation; K + was 0.24 ± 0.03 mmol L1 in both the nighttime sleep condition and the sleep deprivation condition.

Diuresis, natriuresis, kaliuresis, and haematocrit

Urinary flow over the 23.00-07.00 h period was similar whether or not the subject had slept (Fig. 3 and Table 2). Likewise, kaliuresis did not change significantly. In contrast, natriuresis was significantly increased in the sleep deprivation condition compared to the nighttime condition (P < 0.01).
There was no significant effect in the haematocrit induced by sleep deprivation (45.0 ± 0.9 in normal sleep; 45.0 ± 0.8 during sleep deprivation).
DISCUSSION
In this study, we have shown that a 24-h sleep deprivation blunts the increase in aldosterone plasma levels and pulse amplitude observed during normal sleep and enhances nocturnalnatriuresis. The aldosterone levels follow a similar profile during sleep deprivation as that of PRA, with reduced levels, pulse number and pulse amplitude. In contrast, plasma cortisol, measured to reflect ACTH, slightly increases during sleep deprivation. Plasma K + does not vary significantly with sleep condition. These results suggest that the changes in the 24-h aldosterone profile during sleep deprivation cannot be explained by ACTH or K + variations, but rather by the RAS.

Early studies focusing on aldosterone rhythms concluded that this hormone was primarily under circadian control through the adrenocorticotropic system (Katz et al. 1972; Katz et al. 1975; Armbruster et al. 1975; Lightman et al. 1981). However, an effect of sleep on the endogenous aldosterone rhythms was not considered. Consistent with the concept that the adrenocorticotropic system is driven primarily by a circadian rhythm, we found that the 24-h cortisol profile is only slightly affected by sleep deprivation. As previously described, plasma cortisol levels increased in the sleep deprivation condition. This has been attributed to an inhibitory action of sleep on cortisol secretion (Weitzman et al. 1983), although this inhibitory effect has not been observed during daytime sleep (Pietrowsky et al. 1994). A preponderant role of the stress following sleep deprivation has also been suggested (Leproult et at. 1997), and such a stress effect could also explain the higher cortisol levels observed before sleep deprivation. Thus, the decrease in aldosterone levels observed during sleep deprivation cannot be attributed to adrenocorticotropic system, which would operate in an opposing way.

Potassium is one of the most potent control factors of aldosterone secretion. The zona glomerulosa cells are sensitive to K +, with hyperkaliemia stimulating aldosterone synthesis and release. As previously described (Vagnucci et al. 1974; Armbruster et al. 1975), we find that plasma K + levels show low variations over an 8-h night period. We also show that plasma K + does not vary significantly according to the
condition. The low plasma aldosterone levels observed during sleep deprivation are probably unrelated to differences in plasma K + levels between the two conditions.

Among the three major factors controlling aldosterone secretion, renin is the only one with a nycthemeral rhythm entirely determined by the sleep-wake cycle. As demonstrated in studies using an acute shift of the sleep period, renin release is increased during the sleep period whenever it occurs (Brandenberger et al. 1994). Moreover, the ultradian rhythm in PRA is strongly linked to the REM-NREM sleep cycle. In sleep disorders, such as sleep apnea syndrome, narcolepsy and sleeping sickness (Follenius et al. 1991; Schulz et al. 1992; Brandenberger et at. 1996), PRA va'riations reflect the alterations in sleep structure. We reported also in a recent study (Charloux et at. 1999), that, during sleep, aldosterone oscillations were mainly related to PRA oscillations, whereas they were mainly associated with cortisol pulses during waking periods.

In addition, we found a significant association between aldosterone oscillations and delta wave activity oscillations, aldosterone following delta waves by about 2030 min. In the present study, as expected, the 24-h sleep deprivation reduced the nocturnal PRA levels, pulse frequency and pulse amplitude, flattening the 24-h PRA profile. However, in the normal sleep condition, fewer sleep-related PRA pulses were observed compared with the previously published results, because of the lower number of sleep cycles of our subjects. In this experiment, mealtimes were followed by an abrupt rise of PRA levels in most of the subjects, followed in some subjects by increases in aldosterone levels. These mealrelated increases could be due to food ingestion and/or changes of posture - the head of the bed being raised by 30° during mealtime. A similar influence of meal intake on PRA has been found in a previous study (Brandenberger et al. 1985). Nevertheless, it emerges clearly that the changes in the 24-h profiles of PRA and aldosterone induced by sleep deprivation are similar. This argues for a predominant influence of sleep processes on aldosterone release, through the RAS, rather than a primarily circadian influence.

In the series of experiments on healthy subjects under continuous enteral nutrition, we looked at the hydroelectrolytic changes induced by sleep deprivation. The major finding is that sleep deprivation enhances natriuresis. Natriuresis is a highly-regulated variable, whose level depends on numerous factors. Three of these could be implicated in the modifications observed during sleep deprivation: the atrial natriuretic peptide (ANP) , the pressure natriuresis mechanism and aldosterone. Obstructive sleep apnea (OSA) patients have raised ANP and brain natriuretic peptide levels, which probably contribute to the higher nocturnal natriuresis observed in this pathological situation (Krieger et al. 1989; Kita et at. 1998). Treatment of these patients by continuous positive airway pressure reduces nocturnal urine flow, natriuresis and ANP levels, whereas plasma volume increases by 13% (Krieger et al. 1989; Follenius et al. 1991).

The pattern observed in healthy subjects is somewhat different. In basal conditions, ANP levels display a flat profile over a 24-h period and a shift of the sleep period does not influence the 24-h ANP profile (Follenius et at. 1992). These two results suggest that neither the circadian rhythm nor sleep influences ANP secretion. Consequently, in the present study, it is unlikely that the enhanced natriuresis observed during the sleep deprivation period is due to an increase in ANP secretion. It has been reported that sleep onset and awakening are accompanied, respectively, by a fall and a rise in mean arterial pressure of 10-15 mmHg (Kerkhof et at. 1998). Thus, the lack of a blood pressure fall as observed during a sleep deprivation may explain, at least partly, the higher levels of natriuresis observed in this condition, via pressure-natriuresis mechanisms. More likely, the increase in natriuresis is related to the low levels of aldosterone observed during sleep deprivation, even if the decrease in K + excretion at this time was not significant. Similarly, the increase in diuresis did not reach significance. In physiological conditions, nocturnal diuresis depends primarily on hormonal systems other than the RAS, such as arginine vasopressin (A VP). However, the sensitivity of radioimmunoassay for human A VP is too low for studies on circadian and ultradian A VP rhythms. Whether other renal and hormonal systems implicated in the hydro electrolytic balance are affected by sleep deprivation, leading to a new equilibrium, remains to be studied.

The circadian rhythms of several hormones have been precisely described using the constant routine procedure, which is primarily based on sleep deprivation. This routine was designed to unmask the endogenous component of circadian rhythmicity by eliminating the physiological or evoked responses to stimuli such as sleep, food intake, posture, and the light-dark cycle. Such procedures have allowed determining the influence of the circadian clock on thyreostimuline, prolactin, growth hormone, cortisol, and melatonin secretion (Shanahan and Czeisler 1991; Van Cauter et at. 1994; Vgontzas et al. 1999). The consequences of sleep deprivation on the hydromineral hormonal system were not considered. Whether healthy subjects submitted to the constant routine procedure show modifications in their water and electrolyte metabolism has also not been be taken into account. However, as emphasized previously, data obtained on OSA patients suggest that these systems are disturbed by sleep disorders. Apparently, the consequences of sleep deprivation on the regulation of electrolyte balance in other situations such as familial fatal insomnia have not been described, and whether or not they contribute in the clinical status of these patients remains to be clarified.

In conclusion, we have demonstrated that a 24-h sleep deprivation induces variations in renin and aldosterone release leading to modifications of their 24-h profile, as well as a significant increase in nocturnal natriuresis. In view of these results, it appears that the consequences of long-term sleep deprivation and chronic sleep debt on hydromineral hormones deserve further investigation.