Normative Sleep Values Across Human Lifespan



What are age-related changes in objectively recorded sleep patterns across the human life span in healthy individuals? Are sleep latency, percentages of stage 1, stage 2 and REM sleep significantly changing with age?

Sleep patterns evolve across the normal aging process in complex ways.
Changes in sleep patterns across childhood and adolescence, for example, are not only related to chronological age but also to maturational stage.

Few studies, however, have made comprehensive analyses of these two aspects in adolescents (1).

Similarly, chronological age in elderly people does not always match physiological age.

Therefore, changes in sleep patterns may happen earlier, i.e., at a younger age, for some individuals or at an older age for others.
Further, epidemiological and other studies suggest that much of the sleep disturbance typically seen in old age is likely the result of medical co-morbidities than age per se (2-6).

Nevertheless, four age-related changes have been consistently demonstrated in polysomnographic (PSG) studies of sleep architecture: total sleep time, sleep efficiency, and slow wave sleep all decrease, while wake after sleep onset increases with age.
However, a number of PSG sleep characteristics remain uncertain as regard their evolution with age:

  1. sleep latency has been reported to increase with age in some studies, while several other studies found no significant changes with age. Likewise, a number of studies found no significant differences with age for
  2. percentage of stage and
  3. stage 2 while many others reported an increase with age of these stages.
  4. Similarly, REM sleep has reported to decrease with age in several studies while many other studies found no such association with age.

Why such discrepancies between the studies?
Several factors may be responsible for the difficulties identifying age trends in sleep architecture of apparently healthy subjects, for example:

  • small sample sizes;
  • inconsistency in controlling factors that may influence sleep, such as mental or physical illness;
  • uncontrolled use of alcohol, drugs or medications;
  • or insufficient screening for sleep disorders.

Our goal was to better define normative sleep across the human life span by identifying age-related changes in objectively recorded sleep patterns in healthy individuals using meta-analyses.
More specifically, we aimed to clarify whether sleep latency, percentages of stages 1 and 2 sleep and percentage of REM sleep change with age and in which direction. Also we aimed to verify to what extent lack of control over key variables modify the observed age-changes in sleep patterns.

METHODS

The target population studies for these meta-analyses included all studies that met the following criteria:

  1. included non-clinical participants aged five years or older; The five years lower limit was chosen to include only school-aged children.
  2. included measures of sleep characteristics by “all night” polysomography (PSG) or actigraphy on one or more of the following variables: sleep latency (SL), sleep efficiency (SE), total sleep time (TST), stage 1, stage 2, slow wave sleep (SWS), REM, REM latency, and minutes awake after sleep onset (WASO);
  3. included data presented numerically;
  4. published between 1960 and 2003 in peer-reviewed journals. (Unpublished works, dissertations, chapters and abstracts were not included.)

Databases searched were PubMed, PsyInfo, and Science Citation Index. Search terms were “Sleep” with “normal,” “normative,” and “healthy.” In addition, references cited in retrieved reports were screened for additional reports. Over 4,000 reports were first screened for inclusion criteria and reduced to 585 reports. Subsequently, if a research report referred to the same data, only the most complete data set was taken and the other papers were discarded.
Overall, 65 studies met all inclusion criteria. These studies represented 3,577 subjects aged 5 years to 102 years. The research reports devoted to children and adolescents totaled 1,186 subjects aged between 5 years and 19 years. The research reports on adults included 2,391 participants aged 19 years or older.

RESULTS AND DISCUSSION

This study aimed to describe age-related changes in the macro structure of sleep and to clarify the issues regarding earlier contradictory results regarding the evolution of sleep latency, percentages of stage 1, stage 2 and REM sleep. Indeed, about half of the studies that analyzed age-related changes for percentage of REM sleep and Stage 1 reported that these parameters changed with age while the other half found no change. Similarly, about 2 of 3 studies reported that sleep latency and percentage of stage 2 did not change with age while the others found that these two parameters increased with age. One of the problems was that these studies based their conclusions on a small number of subjects. Therefore, it is very difficult to identify age-related trends when the changes are subtle. To summarize all this information, we decided to perform meta-analyses on 65 studies, which represented 3,577 subjects aged 5 years or older. This method allowed quantifying conclusions, which cannot be done with traditional literature reviews. We also performed the analyses in relationship with several moderators that can have a significant impact on any potential associations between sleep and aging.
In relationship with the objectives of the study, the following conclusions can be drawn from our meta-analytic results:

  1. Sleep latency increases with age. Overall, it appeared that sleep latency modestly but significantly increased with age. However, the change is very subtle: when young adults were compared to middle-aged individuals, and middle-aged compared to elderly individuals, sleep latencies were comparable. The significant difference appeared only when very young adults were compared to elderly individuals. The overall increase in sleep latency between 20 and 80 years was less than 10 minutes.
  2. Percentage of Stage 1 increases with age. The significant increase in stage 1 was found between young and middle-age adults and between middle-aged and elderly individuals, which means that percentage of stage 1 significantly increased across all adulthood.
  3. Percentage of Stage 2 increases with age. This increase was present across the full age range studied, from childhood (5 years and older) until age 60.
  4. Percentage of REM sleep decreases with age in adults. Percentage of REM sleep first increased from childhood to adolescence and than decreased between young and middle-age adults and remained unchanged in subjects older than 60 years.
  5. In adults, the increase in the percentage of stage 2 with age and the decrease of REM latency with age appeared very sensitive to psychiatric disorders, use of drugs or alcohol, sleep apnea or other sleep disorders: failure to exclude individuals with these conditions resulted in the confounding of their significant associations with age.
  6. In children 5 years and older and adolescents: the apparent decrease in total sleep time with age appears related to environmental factors rather than to biological changes. As we showed in Table 5, the studies analyzed indicated a significant decrease of total sleep time with age but only when recordings were performed during school days.
  7. While almost all studies in children 5 years of age or older and adolescents did not find significant change in REM sleep with age, it appeared that there actually is a modest but significant increase in the percentage of REM sleep from childhood to the end of adolescence. After that age, percentage of REM sleep remains relatively stable until 60 years of age where it again begins to decrease.

SLEEP IN CHILDREN AND ADOLESCENTS

Studies that examined the normal sleep in children aged 5 years or older and adolescents using polysomnographic recordings are still scant, making it difficult-to-impossible to effectively perform moderator analyses and to analyze all the sleep variables examinable in the older population.

Results of the meta-analysis suggested that different recording techniques are likely to give different results. Although the conclusion for total sleep time was the same for in-laboratory recordings and actigraphy, the association between total sleep time and age was weaker with actigraphy (-0.33) than with in-laboratory recordings (-0.69).
Furthermore, the discrepancy for total sleep time between the different methods was large among the younger children: more than 60 minutes for children aged between 8 and 12 years.

Importantly, the timing of the recording influenced the age-related change for several sleep variables. Thus, the reduction in total sleep time with age was significant only when recordings were made during school days; total sleep time was unassociated with age when studied on non-school days.
This pattern suggests that, in children and adolescents, the decrease in total sleep time is not related to maturation but to other factors such as school schedules. Several North-American studies have reported the difficulties adolescents have in adjusting to early school days, which occurs for older rather than younger children (76).

Sleep latency and sleep efficiency remained largely unchanged from childhood to adolescence, and none of the studies in the meta-analysis reported significant age-related changes for these two sleep parameters.
Percentage of stage 2 sleep was found to increase with age, while percentage of slow wave sleep decreased. These two results were also found individually in the five studies that examined these two parameters.
Of note, however, is a very large difference between results using the ambulatory monitoring system and in-laboratory recording, which may be attributed to methodological differences in the studies.

The results of the meta-analysis suggested that the percentage of REM sleep significantly (but modestly) increased with age, an unexpected finding since the studies that examined this parameter did not find this association (26,49,52,53,58,59). Since the effect size is small, it would have been difficult to identify this association without the quantitative assessment provided by the meta-analysis.

SLEEP IN ADULTS

As expected, total sleep time and sleep efficiency consistently decreased with age.
Wake after sleep onset obtained the largest effect size showing the important increase with age of time awake after sleep onset. Sleep latency and percentage of stage 2 increased with age but the associations were small (.27 and 0.28 respectively).
Percentage of slow wave sleep and REM sleep both also decreased.
In addition, small effect sizes were obtained for percentage of stage 1 and REM latency; the first increasing with age and the other decreasing with age.
From the results of this meta-analysis, it is clear that all studied sleep parameters significantly change with age across the adult lifespan.

ROLES OF MODERATOR VARIABLES

A great advantage of meta-analyses includes its potential to explore the role of different moderators on the association between aging and different sleep variables. The analyses of potential moderators brought to light a number of noteworthy observations:

  1. FAILURE TO EXCLUDE PARTICIPANTS WITH A MENTAL DISORDER HAD SEVERAL SIGNIFICANT CONSEQUENCES ON THE RESULTS:
    1. it diminished the associations of total sleep time and sleep efficiency with age. That is, the decreases observed in total sleep time and sleep efficiency were less pronounced when participants were not screened for mental disorders.
    2. It hid the age-related increase of percentage of stage 2 sleep.
    3. It hid the age-related diminution of REM latency.

  2. FAILURE TO EXCLUDE PARTICIPANTS WITH A MEDICAL ILLNESS HAD SEVERAL SIGNIFICANT CONSEQUENCES ON THE RESULTS:
  3. It resulted in considerably diminished associations of total sleep time and sleep efficiency with age, and also obscured the relationship between aging and increased sleep latency.


  4. EXCLUSION OF PARTICIPANTS WITH SLEEP APNEA HAD IMPORTANT MODIFICATIONS:
  5. Modifications on effect sizes for the total sleep time, percentages of stage 1, stage 2 and REM sleep and REM latency.
    Indeed, studies that did not screen participants for sleep apnea had smaller effect sizes on these variables, which indicated that age-related changes were less pronounced.


  6. USING PRE-DETERMINED LIGHT OFF AND LIGHT ON TIME INSTEAD OF THE HABITUAL SLEEP SCHEDULE OF THE PARTICIPANTS ALSO HAD CONSEQUENCES FOR THE RESULTS:
  7. the observed decrease in total sleep time with age was smaller, the significant increase of percentage of stage 2 and the significant decrease in REM latency with age disappeared.


  8. WHY THE AGE EFFECTS BEING LESS OBVIOUS WITH THE INCLUSION OF THESE DISORDERS?
  9. There is no simple explanation for this fact. First, it is impossible to determine how many subjects were suffering from one or several of the diseases included in the moderator analyses. However, in small samples the inclusion of some not perfectly healthy subjects creates a heterogeneous group and it is enough to influence the results in unexpected ways. This is a very different situation than when the purpose of the research is to measure the effects of a disease on sleep architecture; in this case the subjects of the experimental group have all the disease and some conclusions can be drawn. Second, the evolution of sleep architecture with age in specific diseases is not well-known: studies usually used age-matched controls to measure the effect of the disease on sleep architecture – which is a methodologically sound – however, this does not provide information on the evolution of sleep architecture with age.
    Furthermore, participants in the studies included in the meta-analysis were all from non-clinical populations. It is unlikely that individuals with a severe mental disorder were included in the studies even when no screening was done to exclude mental disorders.
    It has been repeatedly demonstrated, however, that mild or moderate mental disorders such as anxiety or depression are often accompanied by sleep complaints. It is therefore reasonable to assume that the presence of such low-grade mental disorders may have adversely impacted sleep/age relationships. The same conjecture can be made about medical illnesses.


  10. THE GENDER ANALYSES SHOWED THAT THE ASSOCIATIONS BETWEEN SLEEP VARIABLES AND AGING WERE GENERALLY THE SAME FOR BOTH GENDERS.
  11. However, larger effect sizes were observed in women for total sleep time, sleep efficiency, percentage of stage 1, and REM latency indicating that the age effect on these variables were more important in women.
    On the other hand, effect sizes calculated for gender indicated that women have longer total sleep time and sleep latency than similarly aged men. They also have less percentage of Stage 2 and greater percentage of slow wave sleep than age-matched men.
    Interestingly, Figures 1c and 1d clearly illustrate that percentages of slow-wave sleep and REM sleep based on in-laboratory studies decrease with age. The diminution in the percentage of slow wave sleep can be readily observed in childhood and continues steadily until old age. Conversely, for REM sleep, the overall data pattern (see Figure 1d) may explain disagreements among previous studies concerning this sleep stage’s evolution with age. Meta-analytic results indicated that the percentage of REM sleep decreased with age from young adulthood to late-middle age, but the decrease is not significant in individuals over 60 years of age.

CONCLUSION

Accurate normative data on the evolution of sleep architecture across the human life span are important to better understand exactly what type of changes in sleep patterns can be expected as individuals are aging.

In summary, and in contrast to what was generally suggested in several small studies, the total sleep time in children 5 years of age or older and adolescents does not really change with age.
It appears related to environmental factors rather than to biological changes.
There is a modest but significant increase in the percentage of REM sleep from childhood to the end of adolescence.

After that age, percentage of REM sleep remains relatively stable until 60 years of age where the percentage again began to decline.
Sleep latency modestly but significantly increased with age.
However, the change is very subtle and is apparent when very young adults were compared to elderly individuals.
Percentage of Stage 1 increased with age through all adulthood.
Percentage of Stage 2 increased with age from childhood (5 years and older) until old age.
After 60 years of age, only sleep efficiency continued to significantly decrease, with all the other sleep parameters remaining unchanged.
The results of the meta-analysis clearly illustrated the importance of strict screening methods for the study of sleep parameters in healthy individuals as it maximizes the emergence of age-related changes in sleep. As was demonstrated, inclusion of individuals with sleep, organic or psychiatric disorders, as well as the modification of habitual sleep time substantially obliterated the importance of changes in sleep patterns with aging.
There are several aspects of the normal sleep that need to be further investigated:

  • racial comparisons of sleep patterns are still poorly documented;
  • polysomnographic data in healthy children and adolescents, and to a somewhat lesser degree in middle-aged adults, are still scant.

Any future studies aimed at examining age-related changes in sleep should utilize carefully screened subjects and take into account subjects’ habitual sleep schedules as well as whether PSG recording occurs on weekday or weekend nights.

Procedures, figures, tables, references and analyses of effect sizes are included in the reference paper:
Ohayon MM, Carskadon MA, Guilleminault C, Vitiello MV. Meta-analysis of quantitative sleep parameters from childhood to old age in healthy individuals: developing normative sleep values across the human lifespan. Sleep. 2004 Nov 1;27(7):1255-73. Free PMC