Physiology of normal sleep

This chapter tries to address Section K1(vii) of the 2017 CICM Primary Syllabus, which asks the exam candidates to "describe the physiology of sleep". The author ruefully acknowledges the irony of writing this at 2am.  It has only come up once, in Question 18 from the second paper of 2015. The candidates were asked to describe the stages of sleep, and the respiratory physiological changes that occur during sleep.  Also, Question 25 from the first Fellowship Exam paper of 2008 had asked about sleep disturbance in the ICU, and there is an entire section over there dealing with the answer.

 In summary:

For a while, Physiology in Sleep by Orem (a 2012 book from Elsevier) is available in Google books. This was the main resource for this section. Specifically, Appendix A by Orem and Keeling is a gold mine of concise information. Orem have published "the first comprehensive work on physiology in sleep since Kleitman's monograph, "Sleep and Wakefulness," published in 1963". Their work covers the unadulterated human organism, sleeping in a natural state. In the ICU patient, such an environment usually does not exist due to the influence of critical illness and mild-altering drugs, but it is worth know about normal sleep anyway, as one cannot rule out the possibility that perhaps one day one of your patients might sleep normally. Additionally, there are plenty of free papers, for example this one by Carley & Farabi (2016), or Carskadon & Dement (2011)

Sleep

To say that a definition of sleep in physology has been elusive would be an offensive understatement. Usually, when a textbook or peer-reviewed journal makes the attempt, the result is a series of digressions on the function of arousal-oriented systems such as the brainstem reticular formation, or something even less relevant like a quote about how this death of each day's life knits up the raveled sleave of care, sore labor's bath, balm of hurt minds, great nature's second course, etc etc. In short, people resort to misdirection and sleight of hand to prevent the reader from correctly perceiving the gap in our understanding which sleep really represents. Even the AASM (American Association of Sleep Medicine), who create and define criteria for the assessment of sleep disturbances,  don't include a definition of sleep in their glossary. 

In short, there is no accepted definition for sleep, and it is probably not this author's place to generate one. Instead, from a scientific standpoint, it may be better to define sleep in the terms of the behavioral and physiological changes that are observed during normal sleep. These, fortunately, are sufficiently distinct, to the point where we can even divide the process of normal sleep into discrete phases on their basis.

Stages of sleep

Sleep can be divided into clearly demarcated stages, REM (rapid eye movement) and NREM (non-REM), on the basis of three objective measures:

• EEG (electroencephalography)
• EMG (electromyography)
• EOG (electrooculography)

During REM sleep, muscle tone is greatly reduced or absent, and EEG demonstrates theta waves. The name comes from the rapid conjugate eye movements which are seen during this phase, but there are also other characteristic features, such as weird fluctuations in blood pressure and heart rate, irregular respiration, and characteristic tongue motor activity. One could just as easily have called it the Rapid Tongue Movement phase. This stage of sleep is mediated by a complex system of noradrenergic neuronal activation, where one set of neurons starts firing during REM and the other stops.

This REM and NREM distinction is something of a blanket rule, describing all mammals and birds so far studied. It appears that, like CSF spaces, sleep is an essential part of having a complex nervous system. In fact you'd have to say that any nervous system seems to require a circadian rhythm with regular downtime. Among lower vertebrates and invertebrates there are also phasic behavioural changes, a sort of rhythmic alternation between rest and activity, but they do not seem to need the same structure to their rest period, and it is not characterised by the same sort of deep sensory insensitivity (Brady, 1974).  Even though the C.elegans worm does not have a recognisable EEG signal for us to analyse, they certainly do have behaviorally identifiable routine periods of reduced activity, characteristic body posture, increased threshold for rousing stimulation, and - most importantly - a decrease in physiological performance if they are deprived of these breaks. Without digressing overmuch on the sleep physiology of invertebrates, the reader is directed to the excellent works of Vorster & Born (2015), with a warning that this rabbit hole is deep and weird. It will suffice to summarise by saying that sleep has a defined role in maintaining normal CNS function, and should be viewed as essential for life.

Physiological effects of sleep

The best reference for this is  "Physiological changes of sleep" by Chokroverty, a chapter from Sleep Disorders Medicine. The chapter itself is a bit overlong, but the author made an excellent tabulated summary, which is presented here in a basically unchanged state (except for the removal of the column of physiological changes during wakefulness, where every entry just said "normal"):

Physiological Changes During Normal Sleep

Parameters	NREM sleep	REM sleep
Respiratory changes
Respiratory rate	Decreases	Variable, with apnoeas
Minute volume	Decreases	Decreases further
Alveolar ventilation	Decreases	Decreases further
PaCO2	Increases slightly	Increases more
PaO2	Decreases slightly	Decreases more
Hypoxic ventilatory response	Decreases	Decreases further
Hypercapnic ventilatory response	Decreases	Decreases further
Upper airway muscle tone	Decreases slightly	Decreases further
Upper airway resistance	Increases	Increases further
Cardiovascular parameters
Heart rate	Decreases	Variable
Blood pressure	Decreases	Variable
Pulmonary arterial pressure	Increases	Increases
Cutaneous blood flow	Stable	Decreases
Muscular blood flow	Stable	Decreases
Mesenteric blood flow	Stable	Increases
Renal blood flow	Stable	Increases
Cerebral blood flow	Decreases	Increases

Airway effect of normal sleep

• Respiratory muscles (particularly upper airway muscles)  are atonic during sleep
• Mucociliary clearance is reduced during sleep
• Airway reflexes are altered during sleep
• Airway smooth muscle tone decreases
• Cough is essentially impossible during sleep of most kinds
• Laryngeal stimulation produces apnoea

Cardiovascular effects of sleep

• Blood pressure decreases during sleep. The lowest blood pressure occurs during Stage 3 and 4 of NREM sleep
• Heart rate slows during sleep. It is lowest and most regular in deep NREM
• Cardiac output decreases during sleep
• Transient vasoconstriction events occur during sleep. Death from cardiac disorders occurs most frequently during sleep, between 5am and 6am (during REM sleep).
• There is central inhibition of the baroreflex during REM sleep: cats in REM sleep responded poorly (i.e. did not increase their blood pressure) in response to the occlusion of their carotid artery.

Respiratory effects of sleep

• Total body oxygen consumption is higher in REM sleep, but still lower than in wakefulness
• Minute ventilation decreases during NREM sleep
• Respiration is rapid and irregular in REM sleep
• Medullary sensitivity to hypoxia and hypercapnia decreases
• The PaCO₂ rises by 3-7mmHg during stage N3 sleep

Neurological effects of sleep

• Cerebral blood flow increases during REM sleep and changes heterogeneously during NREM sleep
• Intracranial pressure increases during REM sleep, but not NREM sleep
• Brain temperature increases in REM sleep and decreases in NREM sleep

EEG changes 

The college answer to Question 18 from the second paper of 2015 complained that "few answers described the EEG changes associated with the stages of sleep". Reading a normal EEG is not considered one of the core knowledge areas of the junior ICU registrar, and besides that it would appear that even experts are conflicted about the exact nature of the expected EEG changes. For example, the EEG changes which characterise the onset of sleep are debated, as "there is no single measure that is 100% clear-cut 100% of the time". It is not clear what the examiners wanted from the candidates here. One can only assume that, in order to pass, the trainees were expected to regurgitate the names of memorised EEG patterns, which they cannot possibly be expected to ever recognise or interpret. To help them retain this pointless ballast, here is a list of these waves, cut-and-pasted from Carley & Farabi (2016):

• NREM sleep:
  • N1 stage:
    • loss of alpha rhythm
    • Presence of theta waves (characteristic frequency of 4–7 Hz).
  • N2 stage: 
    • Spindles (burst-like trains of waves in the 11- to 16-Hz range with a total duration ≥0.5 seconds)
    • K-complexes (well-defined biphasic waves lasting ≥0.5 seconds and usually maximal over the frontal cortex)
  • N3 stage:
    • Delta-waves, large (≥75 µV) and slow (0.5–3 Hz)
• REM sleep:
  • EEG resynchronization and flattening
  • Sawtooth EEG wave morphology
  • Theta waves

Endocrine effects of sleep

• All pituitary hormones except ACTH are secreted more vigorously during sleep - growth hormone, prolactin, TSH, etc.
• There are also increased levels of vasopressin and aldosterone.
• Body temperature decreases during sleep and increases transiently during REM sleep
• Thermoregulation is lost. Sweating and shivering is impossible during REM sleep.

Renal effects of sleep

• Urine output decreases
• Urinary excretion of electrolytes also decreases

Gastrointestinal function during sleep

• Experts cannot seem to agree. It seems different experiments have yielded different answers as to what the gut is doing during sleep (is it more motile? less?) In short, nobody knows.
• Oesophageal motility is certainly decreased
• The secretion of saliva slows, and swallowing is impossible.
• Rectal tone is preserved

Effects of sedation on sleep

Sedation is nothing like sleep. The two resemble each other because sedative drugs tend to affect GABA neurotransmission, and natural sleep is the effect of hypothalamic GABA-ergic inhibition of arousal pathways. The similarity ends there.  Weinhouse et al (2011) present a polysomnograph of a critically ill sedated patient, demonstrating that the majority of the time is maintained in N1 sleep, with microarousals every ten seconds. That doesn't sound restful. N1 stage sleep is the earliest and most shallow stage; if a person is woken from N1 stage sleep they frequently will not be able to identify the fact that they went to sleep at all.

In fact, critical illness makes it difficult to identify whether a patient is asleep at all. Sedation and encephalopathy make for anomalous-looking EEG patterns, which makes it difficult to identify sleep on the bases of multi channel polysomnography (more on that later). During a 24 hour period of ICU stay, there may be some periods of EEG activity which resemble N2 and N3 sleep, but they are scattered throughout the day and night in short periods.

Physiological Differences Between Sleep and Sedation

Domain	Sleep	Sedation
Physiological basis	An essential biologic function necessary for life	Not that.
Initiation	Cyclical, initiated by circadian rhythms	Dose-dependent Shortened sleep latency (i.e. N1  is shorter: you fall asleep faster with sedatives)
Reversal	Easily reversed by external stimuli	Not necessarily
Pattern of EEG activity	Cyclic progression of stages defined by well-established EEG criteria	Distorted EEG to such an extent that standard EEG criteria are difficult to apply Generally, sedatives and opioids decrease REM sleep and increase Stage N2 sleep.
Cerebral metabolism	22% reduction in cerebral glucose metabolism during NREM sleep	55% reduction in cerebral glucose metabolism, when propofol is titrated just to the point of loss of consciousness

Propofol sedation and normal sleep

Specific sedative effects have been studied. For instance, Tung et al (2004) experimented on some sleeping rats, and found that the rats who got propofol infusions during their normal sleep did not exhibit any signs of sleep deprivation. Ergo, normal restorative functions of sleep were preserved in spite of propofol sedation. Of course, it is difficult to generalise this to human ICU patients. Those are usually not in a normal sleep state to begin with, and they are constantly being woken up.  Not to mention that they usually operate at a slightly different cortical intensity to the rats. (in rats, subtle signs of sleep deprivation may be difficult to assess). Lastly, the differences in cerebral glucose metabolism (55% decreased with propofol, compared to a 22% decrease with normal NREM sleep) demonstrate crudely how different propofol sedation is to real sleep.

Dexmedetomidine sedation and normal sleep

Dex sedation may demonstrate some features which resemble normal sleep. There is apparently some FMRI similarity between them (Coull et al, 2004). Also, clinically, patients on dexmedetomidine seem to be enjoying an experience more closely resembling sleep on clinical criteria; they arouse easily from their sedation, and are more cognitively intact when they do so. Unfortunately, the patients don't seem to care: most report a poorer quality sleep with dexmedetomidine than with propofol (Corbett et al, 2005).