Effect of sleep fragmentation on daytime function
Abstract
Sleep fragmentation is the term used to describe brief awakenings or
microarousals from sleep which are less than 15 seconds long and often
occur without the awareness of the sleeping subject. Arousals is the
collective term for awakenings >15 seconds and microarousals < 15
seconds. Patients with sleep apnoea/ hypopnoea syndrome (SAHS) have
recurrent upper airway obstructions during sleep usually terminated by
arousals and decreases in oxygen saturation. They suffer from impaired
daytime function which correlates weakly with their nocturnal hypoxemia and
sleep fragmentation. These are interrelated making it difficult to distinguish
which is the cause of daytime dysfunction in SAHS patients. This thesis
examines the impact of sleep fragmentation alone on daytime function by
inducing sleep fragmentation in normal subjects and studying their
subsequent daytime function.
A problem associated with studying sleep fragmentation is its poor
definition. Current arousal definitions use a combination of a greater than 1
second increase in EEG frequency with or without increased EMG activity
depending on sleep stage. This can lead to difficulties in comparing results
between studies. Although the American Sleep Disorders Association
(ASDA) has published guidelines on visual scoring of arousals they have not
been validated or compared with other arousal definitions currently in use.
Therefore 3 different arousal definitions and 1 definition of awakening were
compared in SAHS patients. The definitions were (1) ASDA (3 seconds), (2)
ASDA modified to 1.5 seconds, (3) Cheshire 1.5 second. The awakening
was defined as a Rechtschaffen and Kales' stage shift to wakefulness. There
were significantly more arousals of any kind than awakenings, and
significantly more 1.5 second arousals by either definition than ASDA
arousals. However not all apnoeas and hypopnoeas were terminated by
visible EEG arousals with at best, 83% of respiratory events being
terminated by 1.5 second ASDA arousals. There were weak but significant
relationships between microarousals scored by any definition and daytime
sleepiness on the multiple sleep latency test (MSLT).
The first sleep fragmentation protocol examined the effects of one
night of induced visible EEG arousals on the daytime function of normal
subjects. The subjects were objectively sleepier during the day after
fragmentation as measured by both the MSLT and the maintenance of
wakefulness test (MWT). Subjects had altered mood on the UWIST mood
adjective checklist following sleep fragmentation; energetic arousal was
diminished all day except at 12.00, hedonic tone was decreased at 10.00,
and tense arousal was increased at 08.00 and 10.00. Subjects had impaired
performance on 2 tests of cognitive function; Trailmaking B, a test of mental
flexibility, and on PASAT 4 seconds, a test of sustained attention. These
deficits were similar to those seen in SAHS patients prior to CPAP therapy.
There are subgroups of patients with sleep apnoea whose apnoeas
and hypopnoeas occur when they are lying supine or when they are in REM
sleep. This allows them to obtain periods of uninterrupted sleep which may
be sufficient to overcome any daytime dysfunction that may have occurred
due to their REM or posture related sleep apnoea. Therefore 2 fragmentation
paradigms were compared; regular fragmentation every 90 seconds of sleep,
and clustered fragmentation every 30 seconds for 30 minutes every 90
minutes. There was no difference in arousal frequencies between study
nights. There were no differences in daytime function despite significantly
less stage 2 and more slow wave sleep on the clustered fragmentation night.
This suggests that deficits in daytime function are dependent on sleep
fragmentation and not stage 2 or slow wave sleep.
Not all apnoeas and hypopnoeas are terminated by visible EEG
arousals but are terminated by transient increases in blood pressure. The
impact of these transient increases in blood pressure on daytime function are
unknown. Therefore daytime function was compared after an undisturbed
night's sleep and one night of sleep fragmentation to cause blood pressure
elevations alone without coincident visible EEG arousals. There was
significantly less slow wave sleep on the fragmented study night but there
was no difference in visible EEG arousals between study nights. Non-visible
sleep fragmentation made subjects sleepier during the day on the MSLT and
MWT, and decreased hedonic tone upon awakening. There was no effect on
cognitive function.
Finally changes in EEG frequencies during visible EEG arousals were
examined using Fast Fourier Transformation (FFT). There were significant
increases in all physiological frequencies of human sleep within 5 seconds of
the start of an arousal. During the non-visible fragmentation night alpha EEG
power was determined with FFT. There was a significant increase in peak
alpha power within 5 seconds of a tone whether that tone produced a visible
EEG arousal or not. This suggests that computerised analysis of the EEG
may be useful in measuring sleep fragmentation.