The relationships between age, pubertal status, gender and sleep patterns
To our knowledge, this is the first longitudinal study of sleep patterns of a large representative sample of early adolescents. The present results based on parental reports confirm those of cross-sectional studies, i.e. (1) a reduction of nocturnal sleep length, (2) a delay of bedtimes and (3) increasing differences between school day and weekend sleep schedules (Anders et al. 1978; Carskadon et al. 1982; Petta et al. 1984; Szymczak et al. 1993; Wolfson and Carskadon 1998).
According to Klackenberg (1987), the parents are positively the most obvious and best source of information for their child’s sleep behavior. Ferber (1995a) believes that until late childhood or adolescence, the complaints habitually come from the parents, not from the child. In Ferber’s view (Ferber 1995b), it is the parents’ perception of a sleep disorder that must be assessed. Many authors reported high consistency between parental answers and children’s own answers on questions concerning sleep patterns and sleep habits (Saarenpää-Heikkiläet al. 1995; Epstein et al. 1998), and insomnia symptoms, such as difficulty falling asleep and nocturnal awakenings (Dixon et al. 1981; Fisher et al. 1994). In the present study, parent-reported data was not validated by children-reported data. Numerous factors may influence whether events or behaviors can be observed, considered significant and reliably reported by the parents (Weiss et al. 1973). For example children are more likely to cope with sleep difficulties on their own as they get older (Richman 1987). Reliability of early adolescent assessment could be augmented by obtaining information from multiple sources, including parent, child, teacher and examiner observation.
Of particular importance in this study is the marked decline in school day TIB from ages 10–13 years, due to progressively later bedtimes while wake time remained very stable. In contrast, weekend TIB was quite consistent. In this respect, Petta et al. (1984) have suggested that the steadiness of weekend TIB reflects the adolescents’ sleep need, namely that their sleep need would be constant across this developmental stage. The sleep schedule displacement towards later hours on weekends is already manifest in school-age children (Gulliford et al. 1990). In adolescents, this change in the timing of sleep across the week may be more prominent due to a delay of the circadian phase (Carskadon et al. 1993; Carskadon et al. 1997). Progressively greater differences were observed between wake time on schooldays vs. weekends. This difference is 1 h at age 10 years and 2 h at age 13 years. The school schedule likely acts as a social synchronizer and may interfere with the physiological phase delay of the sleep period that occurs in adolescence. The 45 min weekend oversleep at age 13 years noted in the present study is comparable to that previously published from self-reported data (Anders et al. 1978; Petta et al. 1984; Bearpark and Michie 1987; Strauch and Meier 1988). Finally, the change in the relationship between school day TIB and weekend TIB occurred early in girls while it was more progressive in boys. As suggested by Petta et al. (1984), this gender difference is probably associated with earlier onset of puberty in girls than in boys.
The present study also clarifies the relationship between gender, pubertal status and the timing of sleep. At age 13 years, the difference in weekend wake time between early adolescents with a higher pubertal status and those with lower pubertal status suggests that the gender differences observed in Table 1 largely depend upon girls’ higher pubertal status compared to boys. Of note is the significantly longer weekend TIB noted in our early adolescents with a higher pubertal status. That nocturnal sleep length augments with higher pubertal status on weekends supports Carskadon’s view (Carskadon 1990a) that adolescents may have a physiological need for more sleep during puberty, not less, compared to prepubertal children. With respect to the timing of sleep on schooldays, neither gender nor puberty was associated with any school day sleep variable. Based on the observation that 4% of 13-year-olds were awakened by an alarm clock or by a parent on weekend mornings, in comparison with 70% for schooldays, Carskadon (1982) suggested that weekend sleep in young adolescents is less controlled by outside influences and therefore more naturalistic than on schooldays. The significant association between pubertal status and both weekend wake times and weekend TIB found in the present study further support this hypothesis, though the later weekend wake times possibly reflect, in part, a response to schooldays’ sleep deprivation. In addition, sleep disturbances, as defined by difficulty falling asleep and frequent nocturnal awakenings, were not associated with school day sleep variables. On the other hand, subjects with difficulty falling asleep consistently woke up later on weekends from ages 10–13 years. This can be interpreted as a symptom of the considerable delay and lengthening of the sleep period on weekends in comparison to schooldays, with concomitantly later sleep onset and offset.
In all, the present results support the idea that puberty influences the timing of sleep patterns in early adolescents, particularly upon the weekend sleep schedule. Physiological concomitants of puberty may thus have an effect on the circadian phase delay of the sleep-wake cycle characteristic of adolescence. The similarity in the results obtained across studies, using different methodologies, suggests that the development of sleep patterns in early adolescence may be dependent on biological factors.
Bedsharing or cosleeping, usually defined in the literature as sleeping in the parental bed, is consistently reported to decrease as the child’s age increases. Klackenberg (1982) noted that the prevalence of bedsharing falls quite rapidly after age 9 years. In the present study, the number of children sleeping in a separate room significantly increased. As for bedsharing (Klackenberg 1982; Rath and Okum 1995), no gender difference was found in the proportion of boys and girls sleeping in a separate room or sharing their room. In preschool-aged children, cosleeping was related to sleep problems, including difficulty falling asleep and nocturnal awakenings (Schachter et al. 1989; Madansky and Edelbrock 1990; Latz et al. 1999). As in a previous study in preadolescents aged 8–10 years based on parental reports (Kahn et al. 1989), we found no association between sleep disturbances and the practice of sharing room with a sibling or parent in early adolescents.
In childhood, the use of a night light is considered as a presleep ritual and its request by children may come from fear of the dark and nightmares (Anders 1982; Ferber 1995a). Beltramini and Hertzig (1983) reported that 47% of children aged 1–5 years regularly insisted on sleeping with the light on. Our data indicate that a large percentage of children continue to sleep with a night light until 10 years of age. However, this sleep habit dramatically decreased after age 10 years. Kahn et al. (1989) reported that significantly more children aged 8–10 years old, with difficulty falling asleep and nocturnal awakenings, slept in rooms exposed to noise or light. Similarly, almost 50% of our early adolescents sleeping with a night light presented difficulty falling asleep at age 10 and 11 years. The presence of light in the nighttime environment may unfavorably influence sleep. Conversely, a recent study (Laberge et al. 2000) reported high anxiety scores in 11-year-olds suffering from parasomnias. It cannot be excluded that anxiety plays a role in the association between the use of a night light and sleep disturbances in early adolescence.
Preadolescents usually reported neither napping nor having no symptoms of excessive daytime sleepiness (Anders et al. 1978). According to Ferber (1990), excessive napping in preadolescents may indicate an inadequate sleep hygiene. In the present study, only a small percentage of early adolescents aged 10–13 years were reported to nap, a finding similar to that found in previous studies based on self-reports (Carskadon 1982; Simonds and Parraga 1982). In contrast, a significant percentage of older adolescents and college students reported excessive daytime sleepiness and regular naps (Anders et al. 1978; Carskadon 1982; Simonds and Parraga 1982; Andrade et al. 1993). Carskadon et al (1980) and Carskadon (1982) have shown in the sleep laboratory an increase in the propensity to fall asleep during the daytime for children at Tanner stages 3, 4 and 5 (mean ages were 13.4, 14.1 and 15 years, respectively) (Tanner 1962). Although daytime sleepiness was not specifically measured in the present study, the low percentage of napping in our subjects aged 10–13 years suggest that there is probably no increase of daytime sleepiness at that age. In those who actually nap, there was no association between sleep disturbances and napping. The same observation was made from self-reports of adolescents aged 13–17 years (Kirmil-Gray et al. 1984). These results suggest that napping in adolescents is not a consequence of sleep disturbances.
At age 13 years, 40.3% of boys and 71.2% of girls reported reading before bedtime. These percentages are similar to data obtained in adolescents aged 13–17 years where 75.3% reported reading to manage sleeping problems (defined by difficulties falling asleep and nocturnal awakenings) (Kirmil-Gray et al. 1984). In adults, lower percentages were found. In a survey of 1600 subjects aged 36–50 years, 23% of the women and 14% of the men considered reading and listening to music as the most important factor promoting sleep (Urponen et al. 1988). In this survey, as in the present one, significantly more women than men indicated reading and listening to music before bedtime. As pointed out by Ferber (1995b), some adolescents will inappropriately prolong their reading or music listening into the early morning. However, we found no association between reading or listening to music before sleeping and sleep disturbances in early adolescents. A similar finding was reported for adolescents aged 13–17 years (Kirmil-Gray et al. 1984).
The prevalence of difficulty falling asleep and of nocturnal awakenings reported here are similar to those reported by early adolescents in other populations (Klackenberg 1982; Bearpark and Michie 1987; Tynjäläet al. 1993). The higher prevalence of difficulty falling asleep among girls was also reported by older adolescents (Zepelin et al. 1977; Price et al. 1978; Welstein et al. 1983; Kirmil-Gray et al. 1984; Choquet et al. 1988; Cañellas et al. 1994; Vignau et al. 1997). However, we found no gender difference in the prevalence of nocturnal awakenings, as also noted in other studies (Klackenberg 1982; Bearpark and Michie 1987; Morrison et al. 1992). Classically, sleep disturbances are reported to be more frequent in older than in younger adolescents (Carskadon 1982). Surprisingly, the present study showed a decrease in the prevalence of difficulty falling asleep and of nocturnal awakenings from ages 10–13 years. There is a possibility that mothers have been inaccurate in reporting these sleep disturbances, assuming that their child was sleeping. The problem with most previous studies of sleep disturbances in adolescence comes from the fact that they looked at wide age ranges rather than at specific ages. In a study investigating sleep difficulties of children and adolescents from age 11–16 years from 11 European countries (Tynjäläet al. 1993), difficulties falling asleep were, as in the present study, more frequent in 11- and 12-year-olds, and this result was consistent across all countries. Another longitudinal survey of sleep problem during early adolescence also reported a decrease of nocturnal awakenings between ages 10 and 13 years (Klackenberg 1982). Zepelin et al. (1977) evoked the possibility of a transient, puberty-related increase of sleep disturbances at ages 11–12 years. Further studies of large populations of early adolescents recording sleep in the laboratory with concomitant measures of hormonal status will be needed to clarify this issue. Sleep diaries filled out over long periods of time and actigraphy may also contribute to elucidate this question.