Slow-wave sleep: do young adult men and women age

Transcription

1 J. Sleep Res. (1997) 6, FAST TRACK PAPER Slow-wave sleep: do young adult men and women age differently? CINDY L. EHLERS 1,2 anddavid J. KUPFER 2 1 Department of Neuropharmacology, The Scripps Research Institute, La Jolla, CA and 2 Department of Psychiatry, Western Psychiatric Institute and Clinic, University of Pittsburgh, Pittsburgh, PA, USA Accepted in revised form 11 April 1997; received 17 September 1996 SUMMARY The differential effects of ageing on polysomnographic and EEG spectral characteristics of sleep were explored in men and women between the ages of 20 and 40. Men and women in their twenties were found to have similar percentages of slow-wave sleep (SWS) (% Stage 3 and 4) and mean EEG slow wave activity (quantified by spectral analysis). Significant reductions in the percentage of SWS and mean slow wave activity over the night occurred in men during their thirties but not in the women. This suggests that gender difference in SWS may emerge between age 30 and 40 in young adults. Men in this sample were also found to have significant increases in Stage 2 sleep, and decreases in REM sleep time, REM activity, REM density and REM intensity. No significant effects of age were found for women in any visually scored sleep variables. Both men and women had age related reductions in spectral power in the spindle frequencies. Taken together, these findings suggest that the sleep of men and women over age may age differently. KEYWORDS gender differences, normal ageing, slow-wave sleep, spectral analyses INTRODUCTION between the ages of 30 and 40, using computer quantitative techniques. Using spectral analytic techniques, it has been Several studies have documented that sleep patterns alter as a shown that slow wave activity in men appears to decline (Dijk function of normal ageing (see Feinberg et al. 1967; Kales et et al. 1989a) and that the most rapid decline actually occurs al. 1967; Kahn and Fisher 1969; Feinberg 1974; Williams et between the ages of 30 40, reaching levels at age 40 which are al. 1974; Smith et al. 1977; Miles and Dement 1980; Ulrich et indistinguishable from 51 to 70 year olds (Ehlers and Kupfer al. 1980; Gillin et al. 1981, 1984; Blois et al. 1983; Reynolds 1989). Whether such findings are also true for women in the et al. 1985; Hoch et al. 1988, 1994; Knowles and Maclean year age range, is not known because of the paucity of 1990; Prinz et al. 1990; Lauer et al. 1991; Wauquier 1993; normal comparative data using spectral techniques in women Bliwise 1994). Slow-wave sleep (SWS), perhaps more than any in this age range (Horne 1992). other sleep stage, changes over the life cycle (see Horne 1992 Webb (1982) was one of the first investigators to suggest for review). SWS is not present at birth, increases until that the sleep of older men is more disturbed when compared approximately age 10 and then declines slightly to young adult with that of older women. Most studies in middle-aged and levels by age 20 (see Feinberg 1989). While it is clear that elderly subjects suggest that women generally have significantly SWS declines from age 20 to senescence, fewer studies have more SWS than men (Williams et al. 1974; Webb 1982; Reynolds specifically evaluated slow wave activity in men and women et al. 1990; Wauquier and van Sweden 1992). However, gender differences in young adults are more modest (Dijk et al. 1989b; Correspondence: C. L. Ehlers, TSRI, CVN-14, North Torrey Mourtazaev et al. 1995; Armitage 1995). Thus it is not Pines Road, La Jolla, CA 92037, USA. Tel: l; fax: known when significant gender differences emerge over the maturational process. To further explore gender differences in 1997 European Sleep Research Society 211

2 212 C. L. Ehlers and D. J. Kupfer the SWS, we studied age effects on SWS and slow wave activity, (TSA), intermittent wakefulness (A), sleep latency (SL) (time as quantified by spectral analyses, in young adult men and from the beginning of the recording period to the onset of women between 20 and 40 years old. Stage 2 sleep for at least 10 min uninterrupted), sleep efficiency (TSA/TRP 100) and sleep maintenance (SM) (ratio of TSA- METHODS SL/TRP 100). Sleep architecture indices were percentages spent in the different stages of sleep, e.g. percentage REM Sixty-one subjects between the ages of 20 and 40 were available sleep, percentage Stage 1, percentage Stage 2, or percentage for analysis. Fourteen females were between the age of SWS (stages 3 and 4) as well as SWS time (% SWS/100 TSA). (mean=23.5, median=22, SD=3.1, range=20 29) and 14 A REM period was defined as a cluster of at least 3 min of were between 30 and 40 years (mean=35.2, median=35, SD= REM sleep that contains at least one unit of REM activity. 2.6, range=30 40). Eighteen males were between the age of Discrete REM periods are separated by 30 min of clock time (mean=22.6, median=22, SD=2.3, range=20 29) and Isolated minutes of REM sleep are scored and added into the 15 were between 30 and 40 years (mean=34.7, median=35, total for the whole night but are not included in any of the SD=3.1, range=30 40). All subjects were free of past or individual REM periods. If no rapid eye movements are present, current history of psychiatric, medical or neurological disorders despite the fact that the interval otherwise appeared identical or primary sleep disorder, significant weight loss or gain, or to REM sleep it was scored as Stage 1. REM sleep indices any drug usage for a minimum of 14 days prior to assessment included REM sleep time (RT), total REM activity (RA, a in the sleep laboratory. Subjects were also excluded if any of visual estimate of the frequency of rapid eye movement during their first-degree relatives had a history of seizures, sleep REM sleep where each minute of REM sleep scored 0 8), disorder, or major psychiatric illness. Sleep investigations were REM density (RA/RT), REM intensity (RA/TSA) and REM carried out for at least 3 consecutive nights with the second sleep latency (RL, time between sleep onset and first REM night being used for analysis. All women were studied during period minus any wakefulness occurring during that interval). the follicular menstrual cycle phase (days 5 15). To quantify EEG frequency characteristics, the night s sleep, All-night sleep recordings were made using a 16-channel from sleep onset to good morning time, was digitized (128 Hz) polygraph with transfer of data into digital files, on-line or and the power spectra for 4 s epochs determined for a from analogue tape. Electrode impedances were maintained Hz range. Data are carefully edited for movement less than 5 k. Filter settings for the EEG and EOG were and muscle artifacts. The transformed data were then further Hz; chin EMG was derived from bipolar submental compressed into 5 frequency bands ( [Delta], electrodes filtered between 10 and 90 Hz. A 50 microvolt root [Theta], Hz [Alpha], Hz [spindle frequencies], mean square (RMS) sine wave was used for calibration for the Hz [Beta]) and mean power density (microvolts squared/ spectral analyses. The EEG derivation was C4 referred to octave) was calculated for each band as previously described linked mastoids. Electrode placements for the EOG were at (Ehlers and Havstad 1982; Ehlers and Kupfer 1989; Ehlers et the outer canthi with each eye derivation also referenced to al. 1996). Power in each frequency band was determined for linked mastoids. On the first study night, subjects were also the time from good night time to good morning time (whole screened for sleep apnoea and periodic limb movements. night). Subjects with >10 apnoeas or periodic limb movements per Three hypotheses were explored in this data set. First, we hour of sleep were excluded from further analysis. Subjects sought to determine if the men in this sample displayed effects were required to retire at their habitual good night time (GNT) of ageing over the two decades evaluated (20 29, 30 40) similar and get up at their habitual good morning time (GMT). These to what we and others had found previously in populations habitual times were calculated from the times recorded in a 2 evaluated with much less stringent screening procedures (e.g. week sleep-wake diary completed within 3 weeks before the no psychiatric history in any first degree relatives). Second, we sleep study. Subjects were asked to refrain from any napping evaluated potential ageing effects in the women entered into and consuming alcohol 1 week before the study and for the 3 this study. One-way analyses of variance was used to test for nights of the study. Subjects were also asked to keep caffeine differences in the two age groups (20 29, 30 40) of males and use to a minimum during this time period. The recording females. If the F ratio associated with ANOVA was significant, a techniques and overall sleep scoring methodology have been regression analyses was also performed. To control for multiple previously described (Kupfer et al. 1984; Doman et al. 1995). comparisons in one-way ANOVAs significance was set at the All-night summary variables were derived from the visual P<0.01 level. Third, the interaction between gender and age in scoring of recordings using standard criteria (see Rechtschaffen SWS measures was explored using a two-way analyses of and Kales 1968). Some files were scored from paper records variance (age gender). and some from a computer screen following a reliability study that insured that the scoring was not different for the two methods (see Doman et al. 1995). The summary variables RESULTS were divided into 3 groups: sleep continuity indices, sleep For reference, the means and SEMs for visually scored sleep architecture indices and REM sleep indices. Sleep continuity variables in the sample of men and women are presented in indices included total recording period (TRP), time spent asleep Table 1. For the men in this sample, age was found to produce

3 Ageing of SWS in young adults 213 Table 1 Means and standard errors (SEM) for sleep variables scored from the sleep records of 61 individuals. Significant ANOVA results indicated by: P<0.01, P<0.001 Males Females Age (n=18) Age (n=15) Age (n=14) Age (n=14) Mean SEM Mean SEM Mean SEM Mean SEM Sleep latency (min) Intermitten awake (min) Time spent asleep (min) Sleep efficiency Sleep maintenance Percentage stage Percentage stage Percentage stage Percentage stage Percentage SWS (stage 3, 4) SWS time (min) Percentage REM sleep REM latency (min) Number of REM periods REM sleep time (min) REM activity REM density REM intensity Table 2 Means and standard errors (SEM) for spectral power (microvolts squared per octave) in five frequency bands: ( [Delta], [Theta], Hz [Alpha], Hz [spindle frequencies], Hz [Beta]). Power in each frequency band was determined for the time from good night time to good morning time (whole night). Significancies as for Table 1 Males Females Age Age Age Age Mean SEM Mean SEM Mean SEM Mean SEM Hz Hz Hz Hz Hz significant increases in percentage Stage 2 sleep (F=24.9, d.f.= were found in the men (F=33.9, d.f.=1,31, P<0.0001). 1,31, P<0.0001), and decreases in REM sleep time (F=10.2, However, no significant effects of age were found in the delta d.f.=1,31, P<0.003), REM activity (F=11.1, d.f.=1,31, frequency range for women. Regression analyses also showed P<0.002), REM density (F=7.6, d.f.=1,31, P<0.01), and REM a significant correlation of age for slow wave activity as intensity (F=8.3, d.f.=1,31, P<0.007). A significant reduction quantified by spectral analyses in men (R= 0.72,P<0.001). in the percentage of SWS (F=17.4, d.f.=1,31, P<0.0005) and Men had significant age group effects in the theta frequencies SWS time (F=20.3, d.f.=1,31, P<0.0001) was also found in in ANOVA (F=13.3, 1,31, P<0.001) and in regression analyses the older males as compared with the younger age group. No (R= 0.50, P<0.003). Both men (F=7.25, d.f.=1,31, P<0.01) effects of age were found for women in any visually scored and women (F=7.9, d.f.=1,26, P<0.009) had age related sleep variables. reductions in the spindle frequencies as determined by ANOVA. Regression analyses also showed a significant correlation of However, the regression analyses was not significant for either age with percentage SWS in men (% delta, R= 0.60, sex. P<0.0005). The SWS of men and women were also found to age Significant ageing effects were also found using spectral differently. Men and women did not have significantly different analyses. For reference, the means and SEMs for spectral percentages of SWS (% Stage 3 and 4) or SWS time (minutes analyses variables in the sample of men and women are in SWS) in the y old age range, but did differ in their presented in Table 2. Age effects in the delta frequency range thirties. A two-way ANOVA (gender age) revealed a

4 214 C. L. Ehlers and D. J. Kupfer significant effect of gender (F=6.8, d.f.=1,57, P<0.01), of age theta activity for middle-aged males (n=8) compared with group (F=12.221, d.f.=1,57, P<0.001) as well as a significant young adult males (n=9). Reductions in power in the spindle gender age group interaction (F=4.09, d.f.=1,57, P<0.05) frequency ranges during NREM sleep were also found as a for the percentage of SWS and a significant effect of gender function of age among the older males and females in our (F=6.75, 1,57, P<0.01), of age group (F=16.4 d.f.=1,57, sample, a finding consistent with observations by Dijk et al. P<0.0001), and a trend towards a gender age group (1989a) in young and middle-aged males. Reductions in sleep interaction (F=3.28, d.f.=1,57, P<0.075) for SWS time. spindle counts have also been observed in older individuals as Analyses of slow wave activity as quantified by power density in compared with young adults (>65 years) (see Feinberg et al. the delta ( Hz) frequency ranges also revealed significant 1967; Principe and Smith 1982; Wauquier 1993). Because a effects. In a two-way ANOVA, a significant effect of age was reduction in the spindle frequencies was the only variable in found (F=25.03, d.f.=1,57, P<0.0001). However, only trends our study that significantly changed with age in women, it may were found for the effects of gender (F=3.27, d.f.=1,57, represent a very early or sensitive ageing effect on sleep. P<0.07) and for interactions between sex and age (F=2.37, The physiological basis for the gender differences found in d.f.=1,57, P<0.1). the present study is not likely to be a constitutional variable such as body type or bone density because the developmental DISCUSSION course of these variables do not parallel the declines in the percentages of SWS seen in males, but not in females, over the This study found that the decline in the percentage of SWS fourth decade of life. Increased slow wave activity in females (% stage 3 and 4) found in young adults between the ages of as opposed to males has also been demonstrated in animal 20 and 40 is significantly greater in men than in women. A studies where sleep EEG recordings are made directly from recent meta-analysis of gender differences in SWS in middle- the cortical surface (Ehlers et al. 1993) and thus not influenced aged and elderly populations provides evidence to suggest that by bone thickness. Growth hormone levels follow a somewhat males may have sustained reductions in SWS as compared similar developmental course over the life cycle to that of SWS. with women over the lifespan (Rediehs et al. 1990). Fewer Young women have been reported to have higher levels of studies have investigated gender differences in the SWS of growth hormone than young men in some studies, whereas younger subjects, and the data are more conflicting. In an early post-menopausal women and men over 55 may have similar study, no gender differences in manually scored SWS were levels of GH (Ho and Weisberger 1990). In addition, one recent found when men and women were compared in the age range study found that the relationship between age and 24-h GH from 10 to 70 (Williams et al. 1974). In a more recent study release was stronger in males than in females who were in their (Wauquier and van Sweden 1992), the sleep of a small sample 30s (Weltman et al. 1994); thus, gender differences in GH may of young adult men and women, aged 25 35, were compared parallel those seen in SWS. with elders (age ). In that study percentages of manually Feinberg (1989) has suggested that ontogenetic development scored NREM 3 and 4 were almost identical in women and of SWS is also similar to the changes seen in cerebral metabolic men in the age range. However, age related reductions rate over the life cycle. Gur et al. (1987) have found that in Stage 3 were noted but only in the elderly men (age gender women have higher waking cerebral metabolic rates than men; interaction). Reductions in Stage 4 sleep were found in both however, they did not specify whether such findings changed elderly men and women, however, the decrease was twice as at different age epochs. In order to explore the relationship of much for the men as the women. gender, SWS and these physiologic factors, further experiments More recently automated methods, which may be more in young adults will be needed. sensitive to gender differences, have been utilized to explore populations of younger individuals. Dijk et al. (1989b) have reported that the amounts of SWS and REM sleep do not ACKNOWLEDGEMENTS differ between men and women in their 20s, although the The authors would like to thank Dr Victoria Grochocinski, females have higher total power in the EEG. Mourtazaev et Erica Bisson and David Cloutier for statistical support. Susan al. (1995) have also reported some gender differences in slow Lopez helped in the editing of the manuscript. Computer wave parameters in a small sample of older (51 60) and younger programs for spectral analyses were written by Dr James (26 35) individuals. However, Armitage (1995) using a global Havstad. These studies were supported by grant AA00223, measure of delta power in NREM sleep, found it was higher MH-30915, MH and GCRC grant RR among 11 women with a mean age of 25, compared with men in that age range, although she states that the differences were not dramatic. REFERENCES The effects of age on spectral components of sleep other Armitage, R. The distribution of EEG frequencies in REM and NREM sleep in healthy young adults. Sleep, 1995, 18 (5): than delta wave activity has received little attention in subjects Bliwise, D.L. Normal aging. In: M.H. Kruger, T. Roth and W.C. in the third and fourth decade of life. We observed an age- Dement (Eds) Principles and Practice of Sleep Medicine. W.B. related decline in theta ( Hz) activity in men over the Saunders Co., Philadelphia, 1994: whole night. Dijk et al. (1989a) also found attenuations in Blois, R., Feinberg, I., Gaillard, J.M., Kupfer, D.J. and Webb, W.B.

5 Ageing of SWS in young adults 215 Sleep in normal and pathological aging. Experientia, 1983, 39: findings, with implications for sleep functions, and psychiatric illness Experientia, 1992, 48: Dijk, D.J., Beersma, D.G. and Bloem, G.M. Sex differences in the Kahn, E. and Fisher, C. Sleep characteristics of the normal aged male. sleep EEG of young adults, visual scoring and spectral analysis. J. Nerv. Ment. Dis., 1969, 148: Sleep, 1989b, 12 (6): Kales, A., Wilson, T., Kales, J.D., Jacobson, A., Paulson, M.J., Kollar, Dijk, D-J., Beersma, G.M. and van den Hoofdakker, R.H. All night E. and Walter, R.D. Measurements of all-night sleep in normal spectral analysis of EEG sleep in young adults and middle-aged elderly persons: effects of aging. J. Am. Geriat. Soc., 1967, 15: male subjects. Neurobiol. Ageing, 1989a, 10: Doman, J., Detka, C., Hoffman, T., Kesicki, D., Monahan, J.P., Knowles, J.B. and Maclean, A.W. Age-related changes in sleep in Buysse, D.J., Reynolds, C.F., Coble, P.A., Matzzie, J. and Kupfer, depressed and healthy subjects. A meta-analysis. Neuro- D.J. Automating the sleep laboratory: Implementation and psychopharmacol., 1990, 3: validation of digital recording and analysis. Int. J. Biomed. Comput., Kupfer, D.J., Ulrich, R.F., Coble, P.A., Jarrett, D.B., Grochocinski, 1995, 38: V.J., Doman, J., Matthews, G. and Borbely, A.A. Application of Ehlers, C.L. and Havstad, J.W. Characterization of drug effects on the automated REM and slow wave sleep analysis. I. Normal and EEG by power spectral band time series analysis. Psychopharmacol. depressed subjects. Psychiatry Res., 1984, 13: Bull., 1982, 18 (30): Lauer, C.J., Riemann, D., Wiegand, M. and Berger, M. From early Ehlers, C.L. and Kupfer, D.J. Effects of age on delta and REM sleep to late adulthood changes in EEG sleep of depressed patients and parameters. Electroenceph. Clin. Neurophysiol., 1989, 72: healthy volunteers. Biol. Psychiatry, 1991, 29 (10): Ehlers, C.L., Kaneko, W.M., Owens, M.J. and Nemeroff, C.B. Effects Miles, L.E. and Dement, W.C. Sleep and aging. Sleep, 1980, 3: of gender and social isolation on electroencephalogram and Mourtazaev, M.S., Kemp, B., Zwinderman, A.H. and Kamphuisen, neuroendocrine parameters in rats. Biol. Psychiatry, 1993, 33: H.A.C. Age and gender affect different characteristics of slow waves in the sleep EEG. Sleep, 1995, 18 (7): Ehlers, C.L., Havstad, J.W. and Kupfer D.J. Estimation of the time Principe J.C. and Smith J.R. Sleep spindle characteristics as a function course of slow wave sleep over the night in depressed patients: effects of age. Sleep, 1982, 5: of clomipramine and clinical response. Biol. Psychiatry, 1996; 39 (3): Prinz, P.N., Vitiello, M.V., Raskind, M.P. and Thorpy, M.J. Geriatrics: Sleep disorder and aging. N. Engl. J. Med., 1990, 323: Feinberg, I. Changes in sleep cycle patterns with age. J. Psychiat. Res., Rechtschaffen, A. and Kales, A. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of 1974, 10: Human Subjects. BIS/BRI, UCLA, Los Angeles, Feinberg, I. Effects of maturation and aging on slow wave sleep in Rediehs, M.H., Reis, J.S. and Creason, N.S. Sleep in old age: Focus man: implications for neurobiology. In: A. Wauquier et al (Eds) on gender differences. Sleep, 1990, 13 (5): Slow Wave Sleep: Physiological, Pathophysiological and Functional Reynolds, C.F. III, Kupfer, D.J., Taska, L.S., Hoch, C.C., Sewitch, Aspects. Raven Press, New York, 1989: D.E. and Spiker, D.G. Sleep of healthy seniors: a revisit. Sleep, Feinberg, I., Koresko, R.L. and Heller, N. EEG sleep patterns as a 1985, 8 (1): function of normal and pathological aging in man. J. Psychiatr. Reynolds C.F., III, Kupfer, D.J., Thase, M.E., Frank, E., Jarrett, Res., 1967, 5: D.B., Coble, P.A., Hoch, C.C., Buysse, D.J., Simons, A.D. and Gillin, J.C., Duncan, W.C., Murphy, D.L., Post, R.M., Wehr, T.A., Houck, P.R. Sleep, gender and depression: an analysis of gender Goodwin, F.K., Wyatt, R.J. and Bunney, Jr, W.E. Age-related effects on the electroencephalographic sleep of 302 depressed changes in sleep in depressed and normal subjects. Psychiat. Res., outpatients. Biol. Psychiatry, 1990, 28: , 4: Smith, J.R., Karacan, I. and Yang, M. Ontogeny of delta activity Gillin, J.C., Sitaram, N., Wehr, T., Duncan, W., Post, R., Murphy, during human sleep. Electroenceph. Clin. Neurophysiol., 1977, 43: D.L., Mendelson, W.B., Wyatt, R.S. and Bunney, W.E. Sleep and affective illness. In: R.M. Post and J.C. Ballenger (Eds) Neurobiology Ulrich, R.F., Shaw, D.H. and Kupfer, D.J. Effects of aging on EEG of Mood Disorders. Williams and Wilkins, Baltimore, MD, 1984: sleep in depression. Sleep, 1980, 3: Wauquier, A. Aging and changes in phasic events during sleep. Physiol. Gur, R.C., Gur, R.E., Obrist, W.O., Skolnick, B.E. and Reivich, M. Behav., 1993, 54: Age and regional cerebral blood flow at rest and during cognitive Wauquier, A. and van Sweden, B. Aging of core and optional sleep. activity. Arch. Gen. Psychiat., 1987, 44: Biol. Psychiatry, 1992, 31 (9): Ho, K.Y. and Weissberger, A.J. Secretory patterns of growth hormone Webb, W. B. Sleep in older persons: Sleep structures of 50- to 60-yearaccording to sex and age. Horm. Res., 1990, 33 (Suppl. 4): old men and women. J. Gerontology, 1982, 37 (5): Hoch, C.C., Dew, M.A., Reynolds, C.F. III, Monk, T.H., Buysse, D.J., Weltman, A., Weltman, J.Y., Hartman, M.L., Abbott, R.D., Rogol, Houck, P.R., Machen, M.A. and Kupfer, D.J. A longitudinal study A.D., Evans, W.S. and Veldhuis, J.D. Relationship between age, of laboratory- and diary-based sleep measures in healthy old old percentage body fat, fitness, and 24-hour growth hormone release and young old volunteers. Sleep, 1994, 17 (6): in healthy young adults: effects of Gender. J. Clin. Endocrinol. Hoch, C.C., Reynolds, C.F. III, Kupfer, D.J., Berman, S.R. Stability Metab., 1994, 79 (3): of EEG sleep and sleep quality in healthy seniors. Sleep, 1988, 11 Williams, R.L., Karacan, I. and Hursch, C.J. Electroencephalography (6): (EEG) of Human Sleep: Clinical Applications. John Wiley & Sons Horne, J. Human slow wave sleep: A review and appraisal of recent Ltd, New York, 1974.