Methodology
Ninety-nine female students (67% aged 14-16 years), answered a sleep questionnaire in May 94. They attended a public school from 0715h to 1705h, Monday to Friday. The sleep questionnaired covered the following aspects: health status, sleep complaints, home conditions, sleep habits. Weekday and weekend sleep characteristics were compared (paired T-Student test). Students were divided into two groups according to their sleep length on weekdays (<8h, n=55, and 38h, n=44), and contingency tables were prepared for sleep length groups versus prevalence of sleep complaints. Differences between group means were investigated by means of unpaired T-Student test.

Results
Students slept almost 2 hours earlier on weekdays (2204h ± 0100h) comparing to weekends (2354h ± 0154h; t = -9.0; p<0.00001), and woke up 3.5 hours earlier (0548h ± 0032h vs 0919h ± 0135h; t = -23.5, p<0.00001). Sleep length was 1.7 hours shorter on weekdays (7.7h ± 1.2h vs 9.4h ± 1.8h; t = -8.0; p<0.00001). Students reported worse sleep quality (t = -3.5; p<0.001) and more difficulty in waking up (t = -3.8; p<0.001) on school days. More students woke up spontaneously on weekends (18% vs 86%, McNemar test c2 = 60.7; gl =1; p<0.0001). About 11% reported napping more than once a week. The prevalence of frequent (more than once a week) daytime sleepiness was 45%, trouble in falling asleep 14%, trouble in maintaining sleep 6%, and sleep breathing problems 2%. There were different weekdays to weekends patterns: some students advanced their sleep onset (10%) others showed no change (7%), some delayed it up to 2.8h (59%) and others reported a 3-9h delay (24%). No students advanced the sleep offset (2% maintained their wake-up hour, 32% delayed it up to 2.8h, and 66% showed a 3-9h delay. Sleep length decrease was reported by 17% of the students, 3% reported the same duration, 49% increased it up to 2.8h and 31% had an extra sleep amount of 3-8h. Longer sleepers on weekdays ( 38h), slept earlier ( t = -12.4; p<0.00001), reported better sleep quality ( t = 3.17; p<0.005) and had a tendency to wake up slightly later (t = 1.79; p = 0.08) than shorter sleepers ( <8h). The prevalence of general sleep complaints were higher for shorter sleepers (30% vs 11%; c2 = 5.02; p<0.05), as well as trouble in falling asleep (36% vs 14%; c2 = 6.52; p<0.05) and daytime sleepiness (60% vs 41%; c2 = 3.57; p=0.06).

Discussion
Sleep patterns described were similar to those obtained with European (1) and American (2) students of the same age group. Brazilian high school students went to bed at the same time but woke up about 0.5h earlier and slept 0.5h less on weekdays comparing to a younger cohort (12-13 years) of Brazilian students (3) . Sleep onset and offset delays were found in both age groups but older students showed greater delays. The finding that high school students sleep less than middle school students on weekdays but sleep the same amount on weekends agrees with previous work on American students (2). The greater prevalence of sleep complaints found in the shorter sleepers evinces the schoolday sleep restriction consequences. Although the "delayed sleep phase irregular sleep time and daytime sleepiness" pattern is usual for adolescents, they may exhibit different phase shift types (advance or delay) and magnitude, sleep onset and/or sleep offset displacements and amounts of sleep, yelding distinct individual patterns. Investigation of these patterns should provide a better knowledge of adolescents' sleep needs and complaints.