| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Reproductive Endocrine Unit (J.E.H.), Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts 02114; and Neuroendocrine Lab (J.P.S., G.S.R.), Department of Medicine, Brigham and Womens Hospital, Boston, Massachusetts 02115
Address all correspondence and requests for reprints to: Janet E. Hall, M.D., Reproductive Endocrine Unit, Bartlett Hall Extension 5, Massachusetts General Hospital, 50 Fruit Street, Boston, Massachusetts 02114. E-mail: jehall{at}partners.org.
| Abstract |
|---|
|
|
|---|
The interpulse interval of LH was longer during sleep than wake whether it occurred at night or during the day (P < 0.002), implying a decrease in GnRH pulse frequency associated with sleep in the EFP. LH pulse amplitude was greater during sleep than wake (P < 0.001) and greater pulse amplitudes were associated with longer interpulse intervals during sleep (P < 0.005), but not wake. An interaction between sleep and time of day was observed for mean LH, with lower mean LH levels during sleep than wake at night (P < 0.02), but not during the day.
Wakefulness was more likely to be associated with an LH pulse than were stages I/II, III/IV (slow wave), or rapid eye movement sleep (P < 0.005). In addition, the probability of wakefulness within the sleep episode increased 515 min before the onset of LH pulses (relative to randomly selected nonpulse LH; P < 0.05), suggesting that wakefulness was the primary event.
In the absence of sleep, there was an effect of time of day on mean LH (P < 0.02) and LH pulse amplitude (P < 0.03), with greatest values seen during the evening.
In conclusion, in the EFP, inhibition of LH pulse frequency is related to sleep rather than time of day. During periods of sleep, LH pulses occur most commonly in association with brief awakenings, suggesting that interruptions from sleep allow escape from the inhibitory effect of sleep on pulsatile GnRH secretion. A separate effect of time of day on LH pulse dynamics in the absence of sleep was also observed with evening augmentation of LH pulse amplitude and mean level; however, additional studies will be required to determine whether this represents a true circadian effect.
| Introduction |
|---|
|
|
|---|
Endocrine systems are profoundly influenced by both circadian rhythms and sleep in a highly specific manner (12). In the reproductive system, a stimulatory effect of sleep on LH pulsatility is a feature of gonadotropin dynamics during puberty (13, 14, 15, 16). In contrast, there is slowing of pulsatile LH secretion at night in reproductive aged women that is relatively confined to the early follicular phase (EFP) (3, 17, 18, 19, 20, 21). It is unclear from these studies whether inhibition of LH pulses results from a specific effect of sleep or time of day; however, the early studies of Kapen et al. (22) demonstrated an inhibitory effect of sleep on mean LH levels in women in the EFP. Taken together, these studies suggest that sleep per se may be responsible for nighttime slowing of pulsatile GnRH secretion; however, this hypothesis has not been directly addressed. Specific sleep stages play a key role in sleep-related secretion of some hormones, as in the well-known augmentation of GH in association with slow wave sleep (23). During puberty, sleep-related increases in LH occur during non-rapid eye movement (REM) sleep (13). Previous studies in women in the EFP have not detected a correlation of sleep stage with mean LH (22), but the relationship of specific sleep stages to pulsatile LH secretion during the EFP has not previously been examined.
Both menstrual cycle abnormalities and miscarriages are increased in women in whom sleep and/or circadian rhythms are disturbed by night or shift work (24, 25, 26). Understanding the specific effects of sleep and/or circadian factors in the control of pulsatile LH secretion is critical to ameliorating the reproductive consequences of night or shift work for women. Thus, in the current study, we have characterized the dynamics of pulsatile LH secretion in women in the EFP throughout a normal 24-h sleep-wake cycle and during sleep reversal and have addressed the specific relationship between sleep stages and pulsatile LH secretion.
| Subjects and Methods |
|---|
|
|
|---|
Studies were conducted in 11 healthy women, aged 28.6 ± 3.9 yr (mean ± SD), all of whom were euthyroid, normoprolactinemic, and of normal body weight, did not exercise excessively, and were taking no medication. All had regular menstrual cycles of 2535 d. The study was approved by the Massachusetts General Hospital Institutional Review Board, and signed informed consent was obtained from each subject.
All subjects ovulated in the cycle before the study as indicated by a plasma progesterone level above 5 pg/ml (n = 10) and/or a biphasic basal body temperature chart (n = 5). Daily blood samples were drawn during the cycle of study for measurement of LH, FSH, estradiol, and progesterone to confirm normal hormonal dynamics and ovulation in the study cycle. Subjects were admitted to the Clinical Research Center of the Massachusetts General Hospital in the EFP for a 40-h period of blood sampling at 10-min intervals and electroencephalogram (EEG) monitoring (Fig. 1
).
|
An iv catheter was inserted a minimum of 1 h before the onset of the study. Blood samples (2 cc) were drawn at 10-min intervals throughout the study. The catheter was flushed with a volume of normal saline equal to the volume of blood withdrawn. During sleep, a long line (422 cm, 6.4 ml residual volume) was connected for blood withdrawal outside of the sleeping room to minimize sleep disruption. Samples were drawn using a discard volume previously determined to result in no dilution of the measured sample and a blood sparing technique which has previously been described in which all discard volume is reinfused (6). Hemoglobin and hematocrit levels were assessed at 12-h intervals. Sampling was discontinued in two studies (at
26 h) due to a fall in hemoglobin levels below that approved for these studies by the Institutional Review Board. In an additional subject, reduced sampling due to low levels of hemoglobin resulted in insufficient data for analysis of the effect of time of day on LH dynamics, but sufficient data to include in all other analyses.
Assays
Plasma LH, FSH, estradiol, and progesterone concentrations were measured by RIA as previously described (1, 28). All samples were analyzed in duplicate, and all samples from an individual study were measured in the same assay. The intraassay coefficient of variation was obtained from a pool of equal aliquots of the 10-min samples for an individual study run 20 times throughout that assay and was 6.7% for LH. Gonadotropin values are expressed in international units per liter, as equivalents of the Second International Reference Preparation of human menopausal gonadotropin.
Data analysis
The day of ovulation subsequent to the inpatient study was determined from daily blood samples as previously described (3). For inclusion in the final analysis, we required that all studies be performed during ovulatory cycles and between 9 and 13 d before ovulation (d 13 to 9).
Pulsatile LH secretion was analyzed using a validated modification of the Santen and Bardin method of pulse detection (29, 30). Undetectable values were assigned the lowest measurable assay value, and missing values (
1% of the total) were ignored. Modifications to the original algorithm designed to minimize the false-positive rate included the requirement that each pulse have one point in which the difference from nadir to peak was greater than three times the intraassay coefficient of variation and greater than 1 IU/liter and a second point that met at least one of these two criteria (30). The amplitude for each pulse was calculated as the difference between the preceding nadir and the peak, whereas the interpulse interval was calculated as the interval between consecutive peaks.
To determine the effect of sleep vs. wake on the dynamics of LH secretion, mean LH, LH pulse amplitude, and LH interpulse interval were compared between two 8-h segments (sleep night and wake night) and two 4-h segments (sleep day and wake day) using two-way repeated measures ANOVA and post hoc Student-Newman-Keuls testing. As indicated in Fig. 1
, wake night was exactly matched to sleep night by clock time, whereas sleep day began 36 h after sleep night and was exactly matched to wake day by clock time. The effect of time of day on LH parameters was assessed by division of the 24-h period of constant wake into three 8-h periods of equal duration designated day, evening, and night using one-way repeated measures ANOVA and post hoc Student-Newman-Keuls testing. A mixed model random slopes analysis was used to investigate the relationship of LH pulse amplitude to preceding interpulse interval and sleep-wake status.
The polygraph sleep recordings were scored visually in 30-sec intervals according to the criteria of Rechtschaffen and Kales (27). To assess the effect of a specific sleep stage on LH secretion, the sleep records were aligned to the onset of the LH pulse using the first point after the nadir of each pulse and including all pulses that occurred during sleep night and sleep day. In the first analysis, the predominant sleep stage in the 10 min before the pulse was identified. The data were then expressed as the number of pulses associated with a given stage per hour of recording, corrected for the total amount of each stage in the recording, and the results were compared using repeated measures ANOVA and means comparison. In the second analysis, the predominant stage (simple majority) was identified between 05, 510, 1015, and 1520 min before each LH pulse onset. These data were compared with the predominant stage in identical time periods preceding a random series of LH samples that excluded LH pulse onsets, using ANOVA.
The data are expressed as mean ± SEM, and P < 0.05 is considered significant.
| Results |
|---|
|
|
|---|
Slowing of LH pulse frequency was evident during the initial night of sleep in comparison with both daytime and nighttime wake and was also slower with sleep during the day at the completion of the study (Fig. 1
). The interpulse interval was longer during sleep than wake (P < 0.002; Fig. 2
) independent of whether sleep occurred at night (155.3 ± 19.2 vs. 115.0 ± 19.9 min for night sleep vs. night wake, respectively; mean ± SEM; P < 0.02) or during the day (161.7 ± 23.5 vs. 95.4 ± 17.5 min for day sleep vs. day wake, respectively; P < 0.001).
|
Effect of sleep stage on LH secretion
Within sleep, the likelihood of occurrence of an LH pulse was dependent on the specific sleep/wake stage (P < 0.0001; Fig. 3
); LH pulses occurred in association with wake more often than in association with REM sleep (P < 0.0001) or stage III/IV (slow wave) sleep (P < 0.001), but not in comparison to stage I/II sleep.
|
|
In the absence of sleep (i.e. during the 24 h of continuous wake), there was no effect of time of day on interpulse interval (Fig. 5
). However, both LH pulse amplitude (P < 0.03) and mean LH (P < 0.02) were influenced by time of day, with highest values observed during the evening (Fig. 5
).
|
| Discussion |
|---|
|
|
|---|
Nighttime slowing of LH pulse frequency has previously been reported in studies in which the effects of sleep and time of day were not separated (3, 17, 18, 19, 20, 21). Using a sleep reversal protocol, Kapen et al. (22) demonstrated a decrease in mean LH levels in association with sleep that was present whether subjects slept at night or during the day; however, the frequency of pulsatile LH secretion was not formally assessed in this early study. In the current study, we have used a protocol that allowed us to compare LH pulse frequency, pulse amplitude and mean LH levels during sleep at night or during the day with equivalent time periods during wake. Slowing of pulsatile LH secretion during the period of daytime sleep after a night of wake not only provided additional confirmation of the effect of sleep itself on the frequency of pulsatile GnRH secretion, but also confirmed the susceptibility of the subjects to the influence of sleep on LH secretion throughout the duration of the study. Nocturnal slowing of LH pulse frequency has not been observed during the midfollicular phase of the normal menstrual cycle (3), making it particularly important that sleep-related slowing of LH pulses was demonstrated at the completion of the study. Thus, the failure to observe nocturnal slowing of LH pulse frequency during the wake night was due to the absence of sleep rather than to the loss of susceptibility of pulsatile GnRH secretion to inhibition.
Our results suggest that the increased LH pulse amplitude observed during sleep compared with wake can be explained in part by the longer interpulse interval before the LH pulse. These results are in keeping with previous studies in GnRH-deficient men which have indicated that the LH pulse amplitude in response to an invariant dose of exogenous GnRH is directly related to the duration of the preceding interpulse interval (31, 32). This relationship was particularly apparent at the higher end of the spectrum of intervals encountered in normal subjects, such as those that occurred during sleep in the current studies. Our data also suggest that other factors may account for at least a portion of the positive effect of sleep on LH pulse amplitude. This effect may be due to an increased amount of GnRH per secretory episode during sleep or to a sleep-related increase in pituitary sensitivity to GnRH. Support for the latter hypothesis is provided by studies that demonstrated an increased LH response to a pharmacological dose of exogenous GnRH during nighttime sleep compared with nighttime wake in normal women during the EFP (33), although the results of these studies are limited by the presence of endogenous GnRH secretion.
The only study that has previously examined the effect of sleep stage on LH secretion in the EFP found no correlation between slow wave sleep and the sleep-related decrease in mean LH (22). In contrast, we have now demonstrated that within sleep, slow wave and REM sleep are very unlikely to be associated with LH pulses. Those pulses that did occur during sleep typically followed brief awakenings. Because brief episodes of wakefulness are more common during lighter sleep, the greater slowing during slow wave sleep may reflect a reduced probability of wake interruption and escape of pulsatile GnRH secretion from inhibition. This possibility has important methodological implications as many experimental techniques (including blood sampling and the presence of EEG recording electrodes) may disturb sleep (34). It is thus likely that the overall inhibitory effect of sleep on the GnRH pulse generator has been underestimated in the current study. In previous studies in our laboratory with identical blood sampling techniques but no EEG monitoring (3, 5), nocturnal LH slowing in the EFP was even more pronounced than seen here and in other studies that included both blood sampling and EEG monitoring (18, 19).
A neuroendocrine mechanism underlying the EFP effect of sleep has been suggested by one study in which the inhibitory effect of sleep was reversed in the presence of opiate receptor blockade (18) suggesting that inhibition of pulsatile GnRH secretion during sleep in the EFP may be mediated by endogenous opioids. In this study, administration of naloxone did not appear to influence the percentage of time spent in sleep. This is a critical control, given the results from the current study that indicate the importance of ensuring that pharmacological manipulations that apparently reverse LH suppression during sleep in the EFP do not act nonspecifically through effects on sleep continuity. Additional studies have failed to find an effect of alteration in serotonergic or dopaminergic tone on EFP sleep slowing (19, 35). The overall inhibitory effect of sleep on the frequency of pulsatile LH secretion in the EFP is opposite to the stimulatory effect of sleep on the amplitude and frequency of LH secretion that has been observed during puberty (13, 14, 15, 16), in some patients with hypothalamic amenorrhea (36, 37), and in patients recovering from anorexia nervosa (38). Whereas there is considerable interest in the neuroendocrine signals responsible for the onset of puberty, the mechanisms underlying the associated sleep-related augmentation of LH secretion during puberty have not been delineated, but appear to involve increases in both GnRH pulse frequency and pituitary responsiveness (16).
The results of the current study also suggest that there may be a circadian component to the control of LH secretion in the follicular phase. When controlled for sleep/wake state, there was a significant increase in LH pulse amplitude and mean LH during the evening whereas pulse frequency tended to be faster. Although the subjects remained awake and light and activity remained constant in this portion of the study, posture was not strictly controlled nor was food intake. Thus, further studies will be required to determine whether the evening increases in LH pulse amplitude and mean levels that we observed have a true circadian basis. The increase in LH pulse amplitude and mean LH during the evening is similar to the changes observed in circadian studies of TSH (39, 40). An increase in the frequency of pulsatile TSH secretion has been shown during the evening (39), whereas in the current study LH frequency tended to increase, but was not statistically significant. The current studies do not suggest a circadian effect on the response of pulsatile LH secretion to sleep because pulse frequency and amplitude were not different during night and day sleep; however, the effect of sleep on mean LH levels was greater during night than day. Due to the short duration of daytime sleep, further studies will be required to fully address the question of whether the response to sleep is influenced by circadian factors.
Several studies have now documented the frequent occurrence of menstrual cycle disturbances in shift workers (24, 25) and have hypothesized that menstrual cycle changes may be secondary to associated circadian changes in melatonin and/or prolactin (25). The current study indicates that the frequency of pulsatile GnRH secretion is directly altered by sleep/wake state in the EFP. Because individuals working night or rotating shifts have both shorter total sleep time and more fragmented sleep (41), these results raise the possibility that altered menstrual function in shift workers is directly related to the effects of altered sleep patterns on GnRH secretion, possibly through the effects of GnRH pulse frequency on the differential control of LH and FSH (7, 8, 9, 10). Touzet et al. (42) have demonstrated higher FSH levels in association with longer sleep duration in normal women in studies in which urinary hormones were measured across several menstrual cycles and data on sleep duration obtained from questionnaires. Further studies will be required to confirm whether specific alterations in sleep dynamics in the EFP affect hormonal dynamics across the menstrual cycle.
In conclusion, inhibition of LH, and therefore GnRH, pulse frequency in normal women in the EFP is specifically related to sleep rather than time of day. Within sleep, brief episodes of wakefulness are associated with the onset of LH pulses suggesting disinhibition of the GnRH pulse generator. In the absence of sleep, evening augmentation of LH pulse amplitude and mean LH levels suggests that there may also be circadian influences on the dynamics of LH secretion, but these observations will require further confirmation.
| Acknowledgments |
|---|
| Footnotes |
|---|
Present address for G.S.R.: Henry Ford Hospital, 2799 West Grand Boulevard/CFP3, Detroit, Michigan 48202.
First Published Online January 25, 2005
Abbreviations: EEG, Electroencephalogram; EFP, early follicular phase; REM, rapid eye movement.
Received October 14, 2004.
Accepted January 14, 2005.
| References |
|---|
|
|
|---|
-subunit is superior to luteinizing hormone as a marker of gonadotropin-releasing hormone despite desensitization at fast pulse frequencies. J Clin Endocrinol Metab 84:10281036This article has been cited by other articles:
![]() |
C. R. McCartney, S. K. Blank, and J. C. Marshall Progesterone acutely increases LH pulse amplitude but does not acutely influence nocturnal LH pulse frequency slowing during the late follicular phase in women Am J Physiol Endocrinol Metab, March 1, 2007; 292(3): E900 - E906. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. B. Lavoie, E. E. Marsh, and J. E. Hall Absence of Apparent Circadian Rhythms of Gonadotropins and Free {alpha}-Subunit in Postmenopausal Women: Evidence for Distinct Regulation Relative to Other Hormonal Rhythms J Biol Rhythms, February 1, 2006; 21(1): 58 - 67. [Abstract] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |