Sleep. 15(5):449-453

© 1992 American Sleep Disorders Association and Sleep Research Society

Clinical Research A Longitudinal Study of Sleep Stages in Young Women During Pregnancy and Postpartum Helen S. Driver and *Colin M. Shapiro Edblo Sleep Laboratory, Department of Physiology, University of the Witwatersrand Medical School, South Africa; and *Department of Psychiatry, University of Toronto, Canada

Summary: We conducted a longitudinal polysomnographic study in five healthy primiparous subjects, whose sleep was first recorded between 8 and 16 weeks of gestation, then every 2 months until parturition and at I month postpartum. The first 6 hours of sleep were used for statistical analysis. In contrast to previous studies, we found no reduction in stage 4 sleep with pregnancy. Slow-wave sleep (comprising stages 3 and 4), was significantly higher at 27-39 weeks of gestation than at 8-16 weeks, as predicted by the restorative theory of sleep. There was no significant difference in rapid eye movement (REM) sleep time. When compared to a group of normal ovulating women, however, REM sleep time decreased during the last two months of pregnancy and, although there was no change in sleep onset latency, the time spent awake during the first six hours of sleep was increased. Future research into the effects of cortisol and progesterone is indicated. Key Words: Sleep-Slow wave sleep-Pregnancy- Women.

Several physiologic changes that occur during pregnancy may influence sleep. First, the large abdominal mass and associated discomfort could lead to a disturbance of normal sleep. The increased vascular load associated with increased mass could disrupt sleep as a result of increased bladder filling, particularly in the third trimester (1). Increased rapid eye movement (REM) sleep correlated with body weight has been reported in circumstances other than pregnancy (2). Second, metabolic rate during pregnancy, as measured by the oxygen consumption of the mother, shows a biphasic increase to approximately 15-20% above normal. The initial increase results primarily from increased cardiac and renal energy costs, which remain essentially constant throughout pregnancy. The second-and major-increase, which occurs during the second half of pregnancy, results from the energy requirements of the rapidly growing fetus, the enlarging placenta and the uterus (1). Although pregnancy, in some respects, can be likened to a state of accelerated maternal starvation (1) caused, in part, by the constant transfer of glucose from the mother to the fetus, it is primarily a period of anabolic activity. Thus, pregAccepted for publication May 1992. Address correspondence and reprint requests to Helen S. Driver, Department of Physiology, University of the Witwatersrand Medical School, 7 York Road, Parktown 2193, South Africa.

nancy provides a natural experiment for testing the theory that sleep, especially slow-wave sleep (SWS) comprising nonrapid eye movement (NREM) stages 3 and 4, has an anabolic and restorative function (3-5). Finally, endocrine changes during pregnancy may be expected to influence sleep. These changes include the marked rise of estrogen and progesterone (6), the progressive increase of prolactin to seven times the levels in early pregnancy (7) and raised cortisol (8,9). Plasma cortisol concentration exhibits a clear circadian rhythm, with a peak in the early morning. From gestation week 25-26, the cortisol concentration increases, reaching concentrations in late pregnancy more than twice as high as in nonpregnant controls, up to 4.7 times as high when labor commences, and rapidly returning to normal concentrations after delivery (9). Infusions of cortisol in men from 2200 to 0700 hours to plasma levels approximately 3.6 times normal levels augmented SWS, reduced REM time and increased time awake. Infusions of adrenocorticotropic hormone (ACTH) (which increased plasma cortisol to levels 4.4 times normal and, therefore, slightly exceeded those during cortisol administration) likewise reduced REM sleep and increased time spent awake. However, this did not augment SWS (10). Prolactin levels have been associated with SWS, REM latency and REM sleep in women (11). Although the findings are based on a small sample size, Parry et al. (11) found a significant direct

449

450

H. S. DRIVER AND C. M. SHAPIRO

relationship between progesterone levels and stage 3 sleep (r = 0.4, p = 0.03) in a control group of women (n = 8). However, Lee et al. (12) found a shorter REM latency during the luteal phase (high progesterone) in a group of 13 women. No menstrual cycle phase dif.· ferences were found in the percentage of various sleep stages. Progesterone, which is elevated during pregnancy (6,9), appears to be one of the most active steroid hormones. It has been attributed with anesthetic prop·· erties. A natural metabolite of progesterone, allopregnanolone (3a-OH-DHP), is an effective modulator of Cl- flux on the 'Y-aminobutyric acid (GABA)/benzodiazepine receptor Cl- channel complex (13), thus increasing the effectiveness of GABA, which leads to sedation and anxiolytic properties (14). High concentrations of sex steroids seem to be associated with decreased sleep latency (15) and time spent awake (11,16). High sex steroids may also decrease stage 4 sleep (11,17), but there is no evidence that they affect total SWS (11). Based on these inconsistent findings, therefore, the increased estrogen and progesterone of pregnancy may result in increased total sleep time but the effect on REM and SWS is impossible to predict. Moreover, the known effects of the sex steroids may be counteracted by the effects of cortisol. Thus, the overalll consequences of the endocrinological changes during pregnancy can be determined only by direct experiment. There are relatively few electrophysiological studies on sleep during pregnancy (17-22). Two studies (18,19) report an increase in REM, which reaches a peak be:tween the 33rd and 36th week of pregnancy and wanes in the last 3-4 weeks of pregnancy. The results from four studies (17,18,20,21) showed a reduction in stage 4 sleep during late pregnancy. A short report (22) found an inverse relationship between the gestational month and stage 4 sleep for seven women at 11-37 weeks of pregnancy. Ofthree subjects monitored every 2 weeks from early gestation until delivery, two showed a significant decrease of stage 4 sleep, compared to controls, and one subject showed a decrease during the fifth month and an increase in the sixth month. Sleep immediately after delivery showed a suppression of REM sleep (17). The changes in EEG sleep parameters during pregnancy, and especially the possible changes in total SWS, are still unclear. We have recorded polysomnographs in women during pregnancy and at 1 month postpartum, enabling us to analyze the sleep of the same women at different stages of pregnancy. The longitudinal study uses at least three time points for each woman during pregnancy and one postpartum to describe the sleep characteristic changes during pregnancy. The women acted as their own controls. Sleep. Vol. 15. No.5. 1992

METHODS

Five primiparous women 18-29 years old volunteered to participate in the study, which was approved by the University of the Witwatersrand Committee for Research on Human Subjects. The women maintained their normal routines and reported to the sleep laboratory an hour before their usual bedtimes. Each woman slept in an allocated room. Sleep recordings were made on a Beckman Montage-20 electroencephalograph (Beckman Instruments Inc., IL60176, U.S.A.) at a paper speed of 15 mm/second. The recordings included an electroencephalogram (EEG), electrooculogram (EOG) and an electromyogram (EMG) and were scored by an experienced scorer (H.S.D.) in 20-second epochs, according to standard criteria (23). For each of the women, the first recordings were made between 8 and 16 weeks of pregnancy and every second month thereafter until parturition. An adaptation night preceded the recording night by one to three nights. The subjects were all working during their pregnancies. They were asked not to sleep in the afternoon before coming to the sleep laboratory. The time and duration of sleep was scheduled to be the same as in their normal routines. Most records lasted longer than six hours. The women slept until their normal wake time. There were no complications during any of the pregnancies and each went to full term. A month after delivery, the women returned to the sleep laboratory with their babies (who slept in an adjoining room) for an adaptation night and a second night, during which the mother's sleep was recorded. The recordings were not stopped if the baby woke up for a feeding, and feeding time was included in the total recording time (all the mothers were breast feeding). All the infants in this study were well at 18 months. The sleep onset latency (SOL) was taken as the time from lights out to the first appearance of stage 2 sleep. The REM latency was the time from sleep onset to the first appearance of REM sleep. Total sleep time (TST) was the time from sleep onset until the recording was stopped and included any intervening wakefulness and movement, i.e. total recorded time (TRT) - SOL. We calculated sleep cycles from the beginning of stage 2 to the end of the following REM period, retrospectively graded from when no REM sleep occurred for more than 15 minutes. For statistical analysis, the recordings were grouped into three periods-8-16, 17-27 and 28-39 weeks gestation- based on the delivery dates and gynecologically estimated weeks of gestation prior to parturition. Wilcoxon signed rank tests were used for statistical analysis of changes in sleep variables, with the subjects acting as their own controls in order to assess changes in each individual's sleep.

SLEEP DURING PREGNANCY AND POSTPARTUM J,W.

451

L.K.

A

A

M

M

~ 2

13weeks

g> 3

16weeks

u; 4 R

\.

19weeks

25 weeks

28weeks

34weeks

39 weeks (night before delivery)

Post-parium

Hours

A M 1

-'"

Post-partum

~

2

u;

4

g>

3

R

o

1

2

3

4

5

6

7

8

9 10

Hours

FIG. 1. Sleep hypnograms for J.W. at 13, 19 and 28 gestational weeks, the night before delivery and at one month postpartum; for L.K. at 16,25,34 gestational weeks and one month postpartum. A = awake; M = movement; 1,2,3,4 = stages 1,2,3,4 ofNREM sleep; and R = REM sleep.

RESULTS Representative hypnograms of two pregnant women are shown in Fig. 1. One of the women, J.W., went into labor in the sleep laboratory on the morning after the recording at 39 weeks gestation. There was no significant difference in the total recorded time (TRT) in the pregnant women from 8-39 weeks pregnancy and at 1 month postpartum, nor between them and the ovulating group. There was no significant difference in the sleep onset latencies (SOL). The overall weighted mean (± SD) for the gestation period and 1 month postpartum was 13.1 ± 3.5 minutes. There was no difference in the REM latency (ROL). There was an increase in the time spent awake after sleep onset during the first 6 hours of the recording during 8-39 gestational weeks when compared to 1 month postpartum (p < 0.05). There was no significant difference in the movement time and the number of awakenings. The amount of time spent in SWS, particularly in stage 4, was higher at 17-27 and 28-39

weeks gestation when compared to 8-16 weeks gestation. REM sleep in the first 6 hours of the recording did not show any significant difference during pregnancy, but was reduced postpartum. Table 1 shows the median and the range (minimummaximum) for latency to sleep onset (SOL) and REM (ROL) and the % TST spent awake in NREM and REM sleep. The six-hour analysis shows that SWS (%TST), for all five women is higher after 17 gestational weeks compared to 8-16 weeks gestation, whereas the reduced REM sleep is significant at 28-39 gestational weeks and postpartum when compared to 8-16 weeks gestation. One of the women, who went into labor in the sleep laboratory, showed an increase in stage 4 sleep from 23.2% at 12 gestational weeks to 31.2% on the night before delivery, while the REM % decreased from 22.8% at 12 weeks of gestation to 20.6% at 39 weeks. There was no significant difference in the sleep cycle lengths during pregnancy, with the first, second and third cycles being 85 ± 19 (mean ± SD), 87 ± 19 and 93 ± 11 minutes, respectively. The postpartum sleep Sleep. Vol. 15. No.5. 1992

452 TABLE 1.

H. S. DRIVER AND C. M. SHAPIRO The median and range (minimum-maximum) for selected sleep variables for the five women during pregnancy and postpartum on the second night of each recording session SOL (min)

Week of gestation 8-16 II (6-15) 17-27 8 (3-21) 28-39 10 (5-16) Postpartum One month

19 (3-28)

ROL (min)

Awake (%)

Stage 1 (%)

Stage 2 (%)

Stage 3 (%)

Stage 4 (%)

REM (%)

SE (%)

66 (65-131) 65 (56-107) 79 (56-85)

2.2 (0.2-4.5) 2.3 (0-7.8) 2.6 (1.4-4.4)

4.3 (1.9-7.2) 4.9 (3.1-12.6) 5.3 (4.2-6.6)

34.2 (31.2-52.0) 34.5 (26.8-45.9) 34;2 (26.4-43.5)

6.8* (5.0-18.1) 9.3 (5.4-15.9) 10.1 (6.9-12.3)

20.7* (8.1-23.8) 26.6 (9.0-31.5) 26.0 (11.7-35.0)

27.0 (19.7-28.4) 19.6 (17.7-28.6) 22.6** (19.6-24.6)

95.1 (94.3-97.4) 95.2 (91.5-97.2) 94.9 (93.0-96.7)

75 (41-102)

16.5 (0.5-29.7)

2.3 (2.3.,.5.4)

20.6 (16.5-33.1)

6.8* (5.7-18.3)

26.4* (5.9-37.2)

20.4** (13.9-25.4)

83.0 (68.3-93.6)

Values are expressed in minutes or in percentage of total sleep time (TST), where TST is the total recording time less the sleep onset latency (SOL); ROL = REM onset latency. Values that are significantly different are indicated by * p < 0.05 for SWS, i.e. stages 3 and 4, 8-16 weeks gestation versus I month postpartum, and ** p < 0.05, I month postpartum versus 28-39 weeks gestation.

had a similar length (92 ± 29 minutes) for the first sleep cycle, but significantly longer second (158 ± 24 minutes) and third (125 ± 33 minutes) cycles (p < 0.05). The longer second and third cycles were presumably the result of the month-old babies waking up during this time for their feedings, resulting in a disturbed cycle for the mother (more than 10 minutes a.f awake, movement and stage 1 sleep time). One of the babies did not wake up for her usual feeding, but her mother nevertheless showed a lengthened second cycle with more SWS time, suggesting that the lengthened cycles became entrained. DISCUSSION We observed an increase in awake time during gestation, as previously reported (17,21), and at 1 month postpartum. Whereas Karacan et al. (17) reported a longer sleep latency, frequent awakenings and shorter sleep time, resulting in an overall sleep pattern during gestation similar in these respects to insomnia, we found no significant changes in sleep latency. The increase in the awake time resulted na.t from frequent awakenings but from a longer time spent awake after awakening. These longer times usually resulted from the need to urinate. For the women in this study, the ability to go to sleep was not affected by pregnancy or by the situation prevailing on the postpartum night. This study demonstrates that during pregnancy the increase in SWS is primarily due to the significantly elevated stage 4 sleep time. Previous studies all report a reduction in stage 4 sleep during late pregnancy, but none of these was a longitudinal study with inter-subject comparisons from early to late pregnancy (18,20,21). Karacan et al. (17) demonstrated that one of three subjects in a longitudinal study showed an increase in stage 4 sleep during the sixth month after a decrease in the fifth month and the other two subjects showed a significant decrease in stage 4 sleep, COIIlSleep. Vol. 15. No.5. 1992

pared to controls. Because of inter-individual differences in stages 3 and 4 (24), changes in SWS should be assessed with the subjects acting as their own controls wherever possible. In this study, with intra-individual comparisons of SWS during pregnancy, an increase in SWS is shown in three of the five women, with no change in the remaining two, resulting in a mean increase in SWS. We found that REM time decreased between 17 weeks gestation and parturition, compared to 8-16 weeks gestation. Significantly reduced REM during the last trimester has been reported by Hoppenbrouwers et al. (21), whereas Petre-Quadens et al. (18) reported increased REM sleep during the eighth month of pregnancy. Similarly, Branchey and Petre-Quadens (19) reported increased REM sleep from the 25th week of pregnancy, reaching a peak at 33-36 weeks and decreasing in the 3 or 4 weeks preceding delivery. Karacan et al. (17) reported that REM, expressed as percent of total sleep time, was the same during gestation as that of controls. The postpartum EEG had attributes similar to those exhibited during recovery following sleep deprivation from sleep disruption or restriction. In this study, the sleep deprivation was an accumulation from repeated disturbed nights, when the long awakening resulted in less sleep time. As reported in studies of restricted sleep, on the postpartum night, three women exhibited increased stage 4 sleep (25-27) and decreased REM sleep (25,27). Although cortisol, prolactin, estrogen and progesterone levels were not measured in this study, we suggest, within the constraints of a small sample size, that during pregnancy the increase in SWS and awake time with a slight reduction in REM sleep time may be due to a dominantly cortisol effect, which is enhanced by progesterone. This suggestion is based on several findings: 1) after 26 weeks of pregnancy, cortisol is more than twice as high as in nonpregnant controls and fol-

?,,'

:..

A longitudinal study of sleep stages in young women during pregnancy and postpartum.

We conducted a longitudinal polysomnographic study in five healthy primiparous subjects, whose sleep was first recorded between 8 and 16 weeks of gest...
388KB Sizes 0 Downloads 0 Views