Abstract
Study Objectives:
To provide reference data on sleep duration throughout childhood and explore the demographic characteristics associated with sleep.
Design:
Population-based prospective longitudinal birth-cohort study.
Setting:
South-West England, children born in 1991-1992 and followed since birth.
Participants:
Eleven thousand five hundred children with repeat measures of sleep from birth based on parent-reported questionnaires. Data on daytime and nighttime sleep duration and timings and night awakenings at 8 timepoints from age 6 months to 11 years.
Results:
Total sleep duration steadily fell from 13 hours and 12 minutes during infancy to 9 hours and 49 minutes at 11 years of age. Compared with earlier studies, the younger children in this cohort slept for a shorter period. The variation in sleep duration was very wide: from 10 to 17 hours in early infancy, narrowing to 8.5 to 11 hours at 11 years. Half of the children at preschool age woke at least once during the night, but frequent waking (> 3 times) peaked in infancy (10% of all infants) and steadily declined in the preschool-aged years. Despite going to bed at the same time, girls slept consistently longer than boys (by 5-10 minutes). Children from low-income families went to bed later and woke up later, but there was little difference in total sleep duration. Children of younger mothers (< 21 years) slept longer, whereas children of older mothers (> 35 years) slept persistently less. Children in larger families tended to go to bed later, as did the minority group of non-White children in the cohort.
Conclusions:
Given the wide natural variation of sleep in the childhood population, any recommendations on optimal sleep duration at any age must take into account considerable individual variability.
Citation:
Blair PS; Humphreys JS; Gringras P; Taheri S; Scott N; Emond A; Henderson J; Fleming PJ. Childhood sleep duration and associated demographic characteristics in an English cohort. SLEEP 2012;35(3):353-360.
Keywords: Sleep, duration, epidemiology, ALSPAC
INTRODUCTION
Sleep occupies a substantial proportion of childhood life. The importance of childhood sleep duration and quality for future health and performance is increasingly appreciated. In particular, it has been proposed that insufficient sleep in childhood is associated with obesity,1,2 lower academic performance,3,4 and emotional and behavior problems.5–7 Sleep problems are common in childhood, affecting up to one third of children, but are underreported by parents.8,9 Furthermore, disturbed sleep in childhood impinges on parental sleep, health, and quality of life. Sleep problems tend to be age specific and include bedtime resistance, difficulty falling asleep, frequent nighttime awakenings, parasomnias (bruxism, sleep walking and sleep talking, night terrors, enuresis, and body rocking), and noisy breathing, snoring, and obstructive sleep apnea.10 Sleep duration is of clinical importance since parental concerns are often about duration or continuity of their child's sleep. An understanding of the normal variation in sleep during childhood is essential to the definition of sleep problems or initiation of any intervention strategy to modify sleep.
Sleep duration is likely to be determined through an interaction between genetic and environmental factors, but the degree of contribution from these determinants is unknown.11 There has been a suggestion that sleep duration has decreased over time, leading to sleep deprivation in contemporary children. The Sleep in America Poll (2004) reported that a significant percentage of children slept less than 9 hours per night (http://www.sleepfoundation.org). Iglowstein and colleagues reported a decline in sleep duration in Swiss children,12 a finding that has been confirmed by more recent studies.13,14
Sleep duration has been studied in several populations of children using different methods. The majority of data available are from cross-sectional population studies. Few studies have prospectively examined sleep duration in well-characterized cohorts of children across early childhood.12,15–17
This is the first report on normative sleep patterns in a large, unselected, ongoing, population-based cohort of English children born in 1991-1992 and followed until they were 11 years old. We present data on the children's daytime and nighttime sleep duration and timings, as well as the number of nighttime awakenings. We have specifically identified demographic characteristics associated with sleep duration throughout childhood.
PATIENTS AND METHODS
The ALSPAC Study
The Avon Longitudinal Study of Parents and Children (ALSPAC) is a prospective birth-cohort study that aimed to enroll the infants of all pregnant women who resided in the 3 Bristol-based health districts of the county of Avon (population 940,000) and who had an expected date of delivery between April 1991 and December1992 inclusive. Within this 21-month period, 14,062 live-born infants were recruited to the cohort, and 13,988 survived to 1 year of age. This report on childhood sleep duration is part of an ongoing data collection; sleep data for the same individuals in adolescence and young adulthood, along with specific outcomes, will be reported when the data are made available.
The former English county of Avon has a predominantly White population, with a mixture of urban and rural communities and a socioeconomic mix similar to that of the rest of the United Kingdom. We collected information on a wide range of aspects of the lives, health, growth, and development of the children in the study and their parents through successive questionnaires, direct contacts, and clinical records. This includes parental estimates of usual sleep duration and usual timings of sleep. The methodologic details of the study have previously been published,18,19 and further information, including the questionnaires used, can be found at www.alspac.bris.ac.uk
The study was approved by the local research ethics committees and incorporated its own Ethics and Law Advisory Committee.20 The guidelines for strengthening the reporting of observational studies in epidemiology (STROBE) have been followed where possible.21
Sleep-Variable Definitions
Maternal reports of infant and child sleep at 6, 18, 30, 42, 69, 81, 115, and 140 months were collected using postal questionnaires. The majority of questionnaires were returned within the first few weeks after being sent out, although nonresponders were pursued for several months; thus, age windows surrounding the timepoints were used. The windows chosen were 6 months (5-12), 18 months (18-24), 30 months (29-35), 42 months (41-47), 69 months (68-75), 81 months (80-88), 115 months (114-122), and 140 months (137-147). The windows were chosen on the pragmatic basis of trying to capture the majority of the data (> 95%) while maintaining the integrity of the age group. The questions included what time (to the nearest minute) the infant or child “normally” went to bed in the evening and woke in the morning on an average weekday; from this response, nighttime sleep duration was calculated, along with a categorical approximation of daytime sleep (none, < 1 h, 1-2 h, > 2 h, from which estimates of 0, 0.5, 1.5, and 2.5, respectively, were used). Total daily sleep duration (unrounded) was derived from these data. Parents were asked about the number of times the infant or child woke during the night for the first 7 time-points (6-115 months, inclusive). Frequent waking for this analysis was defined as an infant or child waking 3 or more times during the nighttime sleep.
Definitions of Other Variables
The families in this cohort were predominantly White; the remaining ethnic groups, the largest of which were Black Caribbean and Asian, have been grouped together as a non-White minority. Socioeconomic status has been measured using the highest education qualification of the mother and paternal occupation classification obtained at the beginning of the study. These were dichotomized using the cutoffs of those mothers who achieved a school-leaving certificate below the standard level (GCSE below grade C) expected at age 16 years or no qualifications and those partners whose occupation was classified as IV/V (semi-skilled or unskilled) according to the Registrar Generals occupation classification.22 Preterm infants were defined as those born before 37 weeks' gestation, and, low birthweight, as those with a birthweight of less than 2500 grams. Young mothers were defined as those aged less than 21 years when the study subject was born, and, older mothers, as those aged more than 35 years. Large families were defined as those in which the study child had more than 2 older siblings.
Where we have focused on specific age groups, we define infancy as from birth to 1 year, the preschool years as older than 1 year up to and including 4 years, and the primary-school years as aged 5 to 11 years.
Data Analysis
Although we asked for precise sleep-duration timings, the parents often rounded their estimates to the nearest 15 minutes, 30 minutes, or whole hour, which yielded a spiked distribution. When rounded to the nearest hour, sleep duration was normally distributed at all timepoints; thus, means and standard deviations were used to describe the (unrounded) data. Comparisons between continuous distributions were made using the t-test, whereas, for categorical data, the χ2 test was used (for expected cells < 5, the Fisher exact test was used). A Pearson correlation coefficient was also used.
Multivariable modeling included linear-regression models, with total sleep duration used as the outcome variable at each timepoint. The backward-step procedure was used for variable selection, with entry criteria of both univariable significance (P < 0.05) and no more than 10% of values missing. Any variable with more than 10% of values missing was tested at the end of the modeling procedure. All demographic variables were tested at each timepoint, and the significance was reported if a multivariate association was found.
RESULTS
The number of children in the cohort from whom data on sleep were obtained fell from 11478 at 6 months of age to 7043 when the children were aged 11 years. Table 1 shows that children who were lost to follow-up over time tended to be from more-deprived families, using a range of markers (e.g., poor maternal education, social class IV or V, younger mothers, larger families), and with more vulnerable infants (preterm, low birthweight); some of these differences were statistically significant, but most of the differences were not large.
Table 1.
The characteristics of the cohort, in percentage, at 6 months and among those subjects who remained in the study at 11 years of age
| Characteristic | Age of subject at data collection |
P Value | |
|---|---|---|---|
| 6 mo (n = 11478) | 11 y (n = 7157) | ||
| Male | 51.6 | 50.1 | 0.08 |
| Gestation < 37 wk | 5.3 | 4.9 | 0.18 |
| Birthweight < 2500 g | 4.8 | 4.3 | 0.16 |
| Maternal age, y | |||
| < 21 years | 5.2 | 3.3 | < 0.001 |
| > 35 years | 7.4 | 8.4 | 0.006 |
| Paternal social class IV or V | 11.9 | 10.2 | 0.002 |
| Child's ethnicity non-White | 3.4 | 2.7 | 0.02 |
| Family size > 2 siblings | 5.4 | 4.4 | 0.004 |
| Highest maternal education qualification < stardard levela | 17.9 | 13.2 | < 0.001 |
Data are presented as percentages.
Left school at 16 years with below the standard level (GCSE below grade C) or no qualifications.
Of those who remained in the cohort, the vast majority (93%-99%) provided sleep-duration data at each timepoint (Table 2), although we had total sleep duration at all timepoints for only 4524 subjects.
Table 2.
Sleep duration in childhood: the mean and variancea of nighttime, daytime, and total sleep duration over 24 hours
| Age, mo. | Responders, no. (%)b | Sleep duration, h (mean ± SD) |
Daytime napping, % | ||
|---|---|---|---|---|---|
| Total | Nighttime | Daytimec | |||
| 6 | 11297/11478 (98.4) | 13.2 ± 1.7 | 10.8 ± 1.4 | 2.4 ± 1.1 | 99 |
| 18 | 10777/11120 (96.9) | 12.8 ± 1.2 | 11.3 ± 1.3 | 1.5 ± 0.5 | 98 |
| 30 | 9620/10340 (93.0) | 12.0 ± 1.1 | 11.3 ± 1.0 | 1.2 ± 0.5 | 59 |
| 42 | 9585/10053 (95.3) | 11.5 ± 0.9 | 11.3 ± 0.9 | 1.1 ± 0.6 | 23 |
| 69 | 8550/8690 (98.4) | 11.3 ± 0.7 | 11.3 ± 0.7 | 1.0 ± 0.7 | 2 |
| 81 | 8265/8502 (97.2) | 11.2 ± 0.7 | 11.1 ± 0.7 | 1.2 ± 0.9 | 1 |
| 115 | 7761/7876 (98.5) | 10.5 ± 0.7 | 10.5 ± 0.7 | NAd | NAd |
| 140 | 7043/7157 (98.4) | 9.8 ± 0.7 | 9.8 ± 0.7 | NAd | NAd |
Using the 2.5% and 97.5% centiles as approximately 2 SD either side of the mean.
The number of children providing sleep duration data in the age window specified at each timepoint.
Using just those children who slept during the day (Including all children the mean [SD] sleep duration was 2.4 [1.1], 1.5 [0.5], 0.7 [0.7], 0.3 [0.5], 0.02 (0.2), 0.01 [0.1] for the 6 timepoints, respectively.
Questions about daytime napping were not asked at this timepoint.
Total Sleep Duration over 24 Hours
Total sleep duration decreased from a mean of 13.2 hours (SD: 1.7 hours) or 13 hours and 12 minutes at 6 months old to 9.8 hours (SD: 0.7 hours) or 9 hours 49 minutes at 11 years of age. Figure 1 shows the variation in total sleep duration, which was greater in the earlier years; at 6 months of age, sleep duration varied from 9.8 hours to 16.6 hours (using 2 SD from the mean). The variation narrowed with age, but, even as the children were starting secondary school at 11 years of age, the sleep duration varied from 8.5 to 11.1 hours. These estimates (and those in Table 2) were based on repeated cross-sectional data at each timepoint; however, virtually the same estimates were obtained if we limited the analysis to the 4524 individuals who had sleep-duration data for all 8 timepoints. There was a significant correlation between total sleep duration and each adjacent timepoint (correlation coefficients ranging from +0.38 to +0.60), but this correlation steadily weakened as the timepoints got further apart (Table 3).
Figure 1.
Mean total sleep duration over 24 hours at 8 different timepoints in childhood (with variation expressed as 1 or 2 SD from the mean). Denominator data for each is provided in Table 2.
Table 3.
Correlation of total sleep duration between all the different timepointsa
| Timepoints | 6 | 18 | 30 | 42 | 69 | 81 | 115 | 140 |
|---|---|---|---|---|---|---|---|---|
| 6 months | - | +0.41 | +0.30 | +0.26 | +0.19 | +0.16 | +0.11 | +0.06 |
| 18 months | - | - | +0.46 | +0.37 | +0.24 | +0.22 | +0.15 | +0.09 |
| 30 months | - | - | - | +0.48 | +0.30 | +0.28 | +0.21 | +0.16 |
| 42 months | - | - | - | - | +0.38 | +0.34 | +0.25 | +0.20 |
| 69 months | - | - | - | - | - | +0.60 | +0.40 | +0.28 |
| 81 months | - | - | - | - | - | - | +0.49 | +0.34 |
| 115 months | - | - | - | - | - | - | - | +0.49 |
Using the 4524 subjects with total sleep duration for all 8 timepoints.
Nighttime Sleep Duration
Nighttime sleep duration was fairly constant in early childhood (Figure 2) and did not start declining until the children were 9 years old. Again, there was wide variation, even at 5 years of age; when the children were going to primary school, the amount of nighttime sleep varied from 9.9 hours to 12.7 hours. This variation was influenced more by the time the children went to bed rather than the time they woke up, especially in later childhood when children had to go to school (Table 4). The mean rise time throughout childhood was fairly constant, ranging from 06:56 to 07:18, whereas the mean bedtime got noticeably later from the age of 9. At 11 years old, 95% of the cohort usually woke in the 1.5-hour period between 06:22 and 07:58, but bedtime varied much more: for 95% of the cohort, bedtime fell between 20:09 and 22:33, a 2.5-hour period.
Figure 2.
Mean nighttime sleep duration over 24 hours at 8 different timepoints in childhood (with variation expressed as 1 or 2 SD from the mean). NB Denominator data for each is provided in Table 2.
Table 4.
Sleep timings in childhood: the mean and variancea
| Age, mo. | Bedtime |
Wake time |
||||||
|---|---|---|---|---|---|---|---|---|
| Mean | SD | 2.5% | 97.5% | Mean | SD | 2.5% | 97.5% | |
| 6 | 20:08 | 73 | 17:42 | 22:34 | 06:56 | 59 | 04:58 | 10:34 |
| 18 | 19:46 | 57 | 17:52 | 21:40 | 07:06 | 50 | 05:26 | 08:46 |
| 30 | 19:50 | 57 | 17:56 | 21:44 | 07:04 | 46 | 05:32 | 08:36 |
| 42 | 19:46 | 51 | 18:04 | 21:28 | 07:02 | 41 | 05:40 | 08:24 |
| 69 | 19:52 | 42 | 18:28 | 21:16 | 07:10 | 31 | 06:08 | 08:12 |
| 81 | 20:04 | 41 | 18:46 | 21:26 | 07:12 | 29 | 06:14 | 08:10 |
| 115 | 20:51 | 37 | 19:37 | 22:05 | 07:18 | 26 | 06:26 | 08:10 |
| 140 | 21:21 | 36 | 20:09 | 22:33 | 07:10 | 24 | 06:22 | 07:58 |
Using the 2.5% and 97.5% centiles as approximately 2 SD either side of the mean. The mean is displayed as clocktime and the SD in minutes.
Frequency of Nighttime Awakening
Figure 3 shows the number of times the children woke during the night at different ages. By the time they started to go to school, the majority of children slept through the night, with the parents reporting no nighttime awakenings. Approximately half of the parents of children from 18 months to 42 months of age reported regular waking at least once during the night. Approximately 10% of the infants woke frequently (3 or more times) during the night, and this proportion steadily decreased over time. These tended not to be the same children; of those who woke frequently at an individual timepoint, 92% were only identified at 1 or 2 timepoints, and only 2% were identified more than 3 times.
Figure 3.
Number of times child usually awoke during nighttime sleep at 7 different timepoints.
Daytime Sleep Duration
Daytime sleep duration at 6 months of age was 2.4 hours (SD: 1.1 hours) or 2 hours and 27 minutes but almost halved at subsequent ages among those children who still took daytime naps (Figure 4). If, however, we measured daytime sleep among all the children (Figure 5), we found that daytime naps became almost nonexistent by 5 years of age (just 2% of the cohort). This question was not asked of parents when their children were older than 81 months of age.
Figure 4.
Mean daytime sleep duration over 24 hours at 6 different timepoints in childhood for only those children who took a daytime nap (with variation expressed as 1 or 2 SD from the mean).
Figure 5.
Mean daytime sleep duration over 24 hours at 6 different timepoints in childhood for all children in the cohort (with variation expressed as 1 or 2 SD from the mean). NB Denominator data for each is provided in Table 2.
Demographic Characteristics Associated With Total Sleep Duration
Table 5 shows the mean differences in total sleep duration for several demographic characteristics over time. At each timepoint, a linear multivariate-regression model of all 8 variables was tested, and the significance of any association with sleep duration was reported. Sex was significantly associated with sleep duration for all but the last timepoint; girls consistently slept longer—by 5 to 10 minutes—than boys at each age, relating to later wake times rather than to earlier bedtime. There was some evidence of compensation in early childhood, when boys slept slightly longer during the daytime, but the difference in total sleep duration remained significant in the multivariate models. Both preterm infants and those born with low birth weights tended to sleep longer throughout childhood, although these differences were only sporadically statistically significant among low-birthweight children at 30 months and 11 years of age. The most consistent association with sleep duration was maternal age; the sleep duration of the child was consistently longer throughout childhood among those with younger mothers and was consistently shorter among those with older mothers (Figure 6). Paternal social class was not associated with childhood sleep duration; children from the more-deprived families tended to go to bed slightly later but woke up slightly later as well. Neither was there an association with maternal education and sleep duration among the younger children, although older children of mothers with few qualifications tended to sleep slightly longer. Children of non-White ethnicity slept less than the rest of the cohort; the waking times were consistently later than those of the White children, but the bedtimes were very much later. Children who lived in larger families with many siblings tended to sleep less, especially as they got older; this characteristic remained significant in a multivariate model that included maternal age.
Table 5.
Multivariate analysis of demographic characteristics associated with total sleep duration in childhood
| Characteristic | Age of child, moa (sleep duration in hours) |
|||||||
|---|---|---|---|---|---|---|---|---|
| 6 | 18 | 30 | 42 | 69 | 81 | 115 | 140 | |
| Sex | ||||||||
| Male | 13.13 | 12.76 | 11.94 | 11.48 | 11.25 | 11.09 | 10.38 | 9.80 |
| Female | 13.28 | 12.81 | 11.99 | 11.54 | 11.36 | 11.21 | 10.51 | 9.83 |
| Multivariate P value | < 0.001 | 0.03 | 0.03 | 0.002 | < 0.001 | < 0.001 | < 0.001 | NS |
| Gestation, wk | 13.31 | 12.82 | 12.11 | 11.59 | 11.31 | 11.19 | 10.50 | 9.88 |
| < 37 | 13.19 | 12.78 | 11.96 | 11.51 | 11.30 | 11.14 | 10.44 | 9.81 |
| ≥ 37 | NS | NS | NS | NS | NS | NS | NS | NS |
| Multivariate P value | ||||||||
| Birthweight, g | ||||||||
| < 2500 | 13.33 | 12.82 | 12.13 | 11.59 | 11.31 | 11.21 | 10.49 | 9.90 |
| ≥ 250 | 13.20 | 12.78 | 11.96 | 11.51 | 11.30 | 11.14 | 10.45 | 9.81 |
| Multivariate P value | NS | NS | 0.002 | NS | NS | NS | NS | 0.003 |
| Maternal age, y | ||||||||
| < 21 | 13.37 | 13.03 | 12.20 | 11.75 | 11.51 | 11.36 | 10.68 | 9.97 |
| 21-35 | 13.21 | 12.79 | 11.97 | 11.52 | 11.31 | 11.16 | 10.45 | 9.82 |
| > 35 | 12.92 | 12.52 | 11.74 | 11.23 | 11.10 | 10.93 | 10.29 | 9.66 |
| Multivariate P value | < 0.001 | < 0.001 | < 0.001 | < 0.001 | < 0.001 | < 0.001 | < 0.001 | < 0.001 |
| Maternal education,a O level equivalent | ||||||||
| < | 13.23 | 12.74 | 11.99 | 11.58 | 11.34 | 11.19 | 10.50 | 9.91 |
| ≥ | 13.18 | 12.79 | 11.95 | 11.49 | 11.30 | 11.14 | 10.43 | 9.80 |
| Multivariate P value | NS | NS | NS | 0.006 | NS | NS | 0.005 | < 0.001 |
| Paternal social class | ||||||||
| IV or V | 13.27 | 12.77 | 11.93 | 11.55 | 11.37 | 11.20 | 10.47 | 9.83 |
| I, II, IIIn, IIIm | 13.18 | 12.78 | 11.97 | 11.50 | 11.30 | 11.13 | 10.43 | 9.80 |
| Multivariate P value | NS | NS | NS | NS | NS | NS | NS | NS |
| Ethnicity of Child | ||||||||
| Non-White | 13.03 | 12.43 | 11.90 | 11.47 | 11.19 | 11.03 | 10.41 | 9.74 |
| White | 13.20 | 12.79 | 11.96 | 11.51 | 11.31 | 11.15 | 10.44 | 9.81 |
| Multivariate P value | 0.031 | < 0.001 | NS | NS | 0.01 | 0.001 | NS | NS |
| Siblings, no. | ||||||||
| > 2 | 13.26 | 12.73 | 11.94 | 11.46 | 11.15 | 11.05 | 10.31 | 9.67 |
| ≤ 2 | 13.19 | 12.78 | 11.96 | 11.52 | 11.31 | 11.15 | 10.45 | 9.82 |
| Multivariate P value | NS | NS | NS | NS | < 0.001 | NS | 0.002 | < 0.001 |
| Children in the cohort used in the final model, no. (%) | 10604/11297 (93.9) | 10127/10777 (94.0) | 9502/9620 (98.8) | 9267/9585 (96.7) | 7983/8550 (93.4) | 7920/8265 (95.8) | 7350/7761 (94.7) | 6619/7043 (94.0) |
The table shows the mean sleep duration for each category. At each timepoint, all variables were entered into a linear multivariate regression model. The significance of the variables in each of these 8 models is represented in the table by the P values shown. NS refers to nonsignificant.
Left school at age 16 years with a qualification below the standard level (GCSE below grade C) expected or no qualifications.
Figure 6.
Mean total sleep duration over 24 hours at 8 different timepoints in childhood split by maternal age group.
DISCUSSION
We have presented data from one of the few longitudinal studies of sleep duration and night awakening from a very large representative UK population. One of the strengths of this study is its size, and this allows us to provide parental estimates of childhood sleep duration. Thus, the information on sleep duration and night awakening by age can be used as reference values to aid clinical practice.
We found that reported sleep duration decreased with age because of increasingly later bedtimes rather than any change in wake time. However, perhaps the most striking feature of the data is the large variation in sleep duration throughout childhood. The mean duration of sleep for infants in this cohort at 6 months was 13 hours and 12 minutes, yet, despite this precise estimate, the variation for individuals ranged from 10 to 17 hours after excluding observations at the extremes of the distribution. By 11 years of age, the children slept, on average, for 9 hours 49 minutes during the week, but this varied between 8.5 and 11 hours despite the restrictions of having to go to school at a fixed time.
A large variation was observed for both nighttime and daytime sleep; daytime sleep at 6 months of age, on average, was 2.5 hours, but this varied from 20 minutes to nearly 5 hours. When the children were 5 years of age, there was still a 3-hour range in variation, although few children (< 2%) had daytime naps by this age. Approximately one quarter of the infants regularly woke during the nighttime sleep period, and this proportion doubled at preschool age and steadily declined thereafter. Frequent waking (3 or more times a night) peaked in infancy (10% of all infants) and steadily declined in the preschool years. These tended not to be the same children at the different timepoints.
Sleep duration was studied using methods that were similar to those obtained from a smaller group of children (493 subjects) born in Zurich (Switzerland) during 2 time periods: 1974-1978 and 1978-1993.12 The differences from our data are considerable: the ALSPAC children slept approximately 1 hour less during infancy and early childhood (compared with the Swiss cohort), with reduced duration in both nighttime and daytime sleep, and slept approximately half an hour more than the Swiss cohort during later childhood. The decrease in sleep duration that we observed in the cohort during early childhood could potentially be a generational change; the data from Zurich clearly showed a reduction in sleep among the younger children over the 2 decades of that study, suggesting that younger children in the 1990s were getting less sleep than did those in the 1970s. In addition, the recently reported results of a study in the US of 24-hour sleep duration in children at 6 months and 2 years of age suggested that sleep duration had decreased even further; in this study from Massachusetts, the 1676 infants slept a full hour less than did the infants in the ALSPAC study.23 The results of a recent study examining trends in childhood sleep support the finding of a reduction in childhood sleep over time, but this was dependent on the country in which data were collected.13 The comparative increase in sleep as the ALSPAC children started school was less expected, but a similar discrepancy was found when comparing ALSPAC children with more contemporary longitudinal cohorts from New Zealand24 and the United States25; the English children went to bed earlier and woke later. Evidence from an Icelandic cohort16 that childhood sleep duration was strongly correlated across 5 and 10 years was not confirmed. In the ALSPAC cohort, there was a strong correlation between adjacent timepoints, but this steadily weakened over time.
Despite going to bed at a similar time, girls consistently slept 5 to 10 minutes longer than boys, suggesting the potential for a cumulative sleep debt in boys. Longer sleep duration among older girls, as compared with boys, at secondary school has been reported,26,27 but recent reports of sex differences at younger ages are mixed.28,29 Our study from early infancy suggests that this sex difference may have a biologic basis. Our finding of a lack of any difference in sleep duration among preterm infants is in agreement with previous results from Zurich.30
A study conducted in the US in 1998-2000 of 3-year-old children31 found that characteristics measuring socioeconomic deprivation were associated with a decreased use of bedtime routines for children, but we did not find that these children slept any less; instead, we found that they went to bed slightly later and woke up slightly later. Noticeably, maternal age was important; children of younger mothers slept longer, whereas children of older mothers slept less persistently throughout infancy and childhood. Children in larger families, especially when the children were older, tended to go to bed later and, hence, had a shorter duration of sleep; this may be related in part to the fact that larger families tended to have older mothers, yet both of these factors remained significant predictors of childhood sleep duration in the multivariate model, especially among the older children. The difference in sleep timing, daytime naps, and total sleep duration among the children in this study from non-White ethnic backgrounds supports previous findings32 of large cultural differences, although the specific ethnic groups were too diverse to make meaningful estimates.
A limitation of most longitudinal studies conducted over several years is that missing data and loss to follow-up are more likely to occur in the most socioeconomically deprived groups. The ALSPAC study is no different, but its large size means that, even with different rates of loss to follow-up, there is sufficient power to investigate effects in small groups. To understand and interpret potential influences on sleep duration, we opted to dichotomize most explanatory variables, although this can lead to some loss of sensitivity. Our use of data on “usual” sleep practice means that information on the intraindividual variation in practice on different nights of the week cannot be examined and the findings can only be generalized to weekday sleep. We also acknowledge that parent reporting of children's sleep tends to overestimate sleep duration, although we have no reason to suspect any systematic bias from using such a blunt instrument.
CONCLUSION
Currently available data do not allow the identification of what constitutes “normal” sleep duration for children at different ages. We have shown, in an unselected birth cohort, striking interindividual and intraindividual variation that may result in mislabeling of children with potential sleep disorders when their sleep is within the normal range. Clinicians and parents will benefit from more accurate estimations of normal ranges of sleep duration and sleep awakening in relationship to family characteristics. Limitations of our data do not allow classification of children with sleep debt, so we would caution against defining short sleep on the basis of our findings. However, recommendations of desirable sleep duration should acknowledge the large population variations and the need to tailor advice to individual circumstances.
DISCLOSURE STATEMENT
This was not an industry supported study. The authors have indicated no financial conflicts of interest.
ACKNOWLEDGMENTS
The authors are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, and nurses. The UK Medical Research Council (Grant ref: 74882), the Wellcome Trust (Grant ref: 076467), and the University of Bristol provide core support for ALSPAC. Guy's and St Thomas' Charity (Grant ref: S091017) helped support time for statistical analysis of specific sleep data. The study was approved by the Local Research Ethics Committees, and incorporated its own Ethics and Law Advisory Committee. PG, AE, JH & PF created the original study design and/or subsequent data collection. PSB analyzed the data and wrote the first draft. All authors were involved in the interpretation of the analysis and contributed to the drafting and revision of this manuscript. All authors approve this final manuscript.
Footnotes
A commentary on this article appears in this issue on page 311.
REFERENCES
- 1.Taheri S. The link between short sleep duration and obesity: we should recommend more sleep to prevent obesity. Arch Dis Child. 2006;91:881–4. doi: 10.1136/adc.2005.093013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cappuccio FP, Taggart FM, Ngianga-Bakwin K, et al. Meta-analysis of short sleep duration and obesity in children and adults. Sleep. 2008;31:619–26. doi: 10.1093/sleep/31.5.619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Paavonen EJ, Raikkonen K, Pesonen A-K, et al. Sleep quality and cognitive performance in 8 year old children. Sleep Med. 2010;11:386–92. doi: 10.1016/j.sleep.2009.09.009. [DOI] [PubMed] [Google Scholar]
- 4.Gruber R, Laviolette R, Deluca P, Monson E, Cornish K, Carrier J. Short sleep duration is associated with poor performance on IQ measures in healthy school-age children. Sleep Med. 2010;11:289–94. doi: 10.1016/j.sleep.2009.09.007. [DOI] [PubMed] [Google Scholar]
- 5.Cortesi F, Giannotti F, Ivanenko A, Johnson K. Sleep in children with autistic spectrum disorder. Sleep Med. 2010;11:659–64. doi: 10.1016/j.sleep.2010.01.010. [DOI] [PubMed] [Google Scholar]
- 6.Pesonen A-K, Raikkonen K, Paavonen EJ, et al. Sleep duration and regularity are associated with behavioural problems in 8 year old children. Int J Behav Med. 2010;17:298–305. doi: 10.1007/s12529-009-9065-1. [DOI] [PubMed] [Google Scholar]
- 7.Fallone G, Owens JA, Deane J. Sleepiness in children and adolescents: clinical implications. Sleep Med Rev. 2002;6:287–306. doi: 10.1053/smrv.2001.0192. [DOI] [PubMed] [Google Scholar]
- 8.Owens JA, Witmans M. Sleep problems. Curr Probl Pediatr Adolesc Health Care. 2004;34:154–79. doi: 10.1016/j.cppeds.2003.10.003. [DOI] [PubMed] [Google Scholar]
- 9.Stores G. Children's sleep disorders: modern approaches, developmental effects, and children at special risk. Dev Med Child Neurol. 1999;41:568–73. doi: 10.1017/s001216229900119x. [DOI] [PubMed] [Google Scholar]
- 10.Mindell J, Owens J. A Clnical Guide to Pediatric Sleep. 2nd ed. Philadelphia, PA: Lippincott, Williams and Wilkins; 2010. [Google Scholar]
- 11.Taheri S, Mignot E. The genetics of sleep disorders. Lancet Neurol. 2002;1:242–50. doi: 10.1016/s1474-4422(02)00103-5. [DOI] [PubMed] [Google Scholar]
- 12.Iglowstein I, Jenni OG, Molinari L, Largo RH. Sleep duration from infancy to adolescence: reference values and generational trends. Pediatrics. 2003;111:302–7. doi: 10.1542/peds.111.2.302. [DOI] [PubMed] [Google Scholar]
- 13.Matricciani L, Olds T, Petkov J. In search of lost sleep: secular trends in the sleep time of school-aged children and adolescents. Sleep Med Rev. 2011 doi: 10.1016/j.smrv.2011.03.005. doi: 10.1016/j.smrv.2011.03.005. [DOI] [PubMed] [Google Scholar]
- 14.Touchette E, Petit D, Tremblay RE, Boivin M, Falissard B, Genolini C, Montplaisir JY. Associations between sleep duration patterns and overweight/obesity at age 6. Sleep. 2008;31:1507–14. doi: 10.1093/sleep/31.11.1507. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gulliford MC, Price CE, Rona RJ, Chinn S. Sleep habit and height at ages 5 to 11. Arch Dis Child. 1990;65:119–22. doi: 10.1136/adc.65.1.119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Thorleeifsdottir B, Bjornsson JK, Benediktsdottir B, Gislason Th, Kristbjarnarson H. Sleep and sleep habits from childhood to young adulthood over a 10 year period. J Psychosom Res. 2002;53:529–37. doi: 10.1016/s0022-3999(02)00444-0. [DOI] [PubMed] [Google Scholar]
- 17.Klackenberg G. Sleep behaviour studied longitudinally. Data from 4-16 years on duration, night-awakening and bed-sharing. Acta Paediatr. 1982;71:501–6. doi: 10.1111/j.1651-2227.1982.tb09459.x. [DOI] [PubMed] [Google Scholar]
- 18.Golding J. Research protocol: European longitudinal study of pregnancy & childhood (ELSPAC) Paediatr Perinatal Epidemiol. 1989;3:460–9. doi: 10.1111/j.1365-3016.1989.tb00533.x. [DOI] [PubMed] [Google Scholar]
- 19.Golding J. Children of the nineties: a longitudinal study of pregnancy and childhood based on the population of Avon (ALSPAC) W Engl Med J. 1990;105:80–2. [PMC free article] [PubMed] [Google Scholar]
- 20.Mumford SE. Children of the 90s II: challenges for the ethics and law committee. Arch Dis Childhood Fetal Neonatal Ed. 1999;81:F228–F231. doi: 10.1136/fn.81.3.f228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.von Elm E, Altman DG, Egger M, et al. Strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007;335:806–8. doi: 10.1136/bmj.39335.541782.AD. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Office of Population Censuses and Surveys. Classification of occupations 1980. Her Majesty's Stationery Office; Crown Copyright 1980. ISBN 0116907282. [Google Scholar]
- 23.Nevarez MD, Rifa-Shiman SL, Kleinman KP, Gillman MW, Taveras EM. Associations of early life risk factors with infant sleep duration. Acad Pediatr. 2010;10:187–93. doi: 10.1016/j.acap.2010.01.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Nixon GM, Thompson JMD, Han DY, et al. Short sleep duration in middle childhood: risk factors and consequences. Sleep. 2008;31:71–8. doi: 10.1093/sleep/31.1.71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Spilsbury JC, Storfer-Isser A, Drotar D, et al. Sleep behaviour in an urban US sample of school-aged children. Arch Pediatr Med. 2004;158:988–94. doi: 10.1001/archpedi.158.10.988. [DOI] [PubMed] [Google Scholar]
- 26.Natal CL, Lourenço TJ, Silva LA, et al. Gender differences in the sleep habits of 11-13 year olds. Rev Bras Psiquiatr. 2009;31:358–61. doi: 10.1590/s1516-44462009000400013. [DOI] [PubMed] [Google Scholar]
- 27.Dollman J, Ridley K, Olds T, Lowe E. Trends in the duration of school-day sleep among 10- to 15-year-old South Australians between 1985 and 2004. Acta Paediatr. 2007;96:1011–4. doi: 10.1111/j.1651-2227.2007.00278.x. [DOI] [PubMed] [Google Scholar]
- 28.Olds T, Maher C, Blunden S, Matricciani L. Normative data on the sleep habits of Australian children and adolescents. Sleep. 2010;33:1381–8. doi: 10.1093/sleep/33.10.1381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Seo WS, Sung HM, Lee JH, et al. Sleep patterns and their age-related changes in elementary-school children. Sleep Med. 2010;11:569–75. doi: 10.1016/j.sleep.2010.03.011. [DOI] [PubMed] [Google Scholar]
- 30.Iglowstein I, Hajnal BL, Molinari L, Largo RH, Jenni OG. Sleep behaviour in preterm children from birth to age 10 years: a longitudinal study. Acta Paediatr. 2006;95:1691–3. doi: 10.1080/08035250600686938. [DOI] [PubMed] [Google Scholar]
- 31.Hale L, Berger LM, LeBourgeois, Brookes-Gunn Social and demographic predictors of preschoolers' bedtime routines. J Dev Behav Pediatr. 2009;30:394–402. doi: 10.1097/DBP.0b013e3181ba0e64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Crosby B, LeBourgeois K, Harsh J, et al. Racial differences in reported napping and nocturnal sleep in 2- to 8-year old children. Pediatrics. 2005;115:225–32. doi: 10.1542/peds.2004-0815D. [DOI] [PMC free article] [PubMed] [Google Scholar]