Sleep in health

Sleep is a complex reversible state associated with diminished/absent voluntary behavior and responsiveness to external stimuli. It is associated with specific electroencephalographic changes and also affects many physiologic systems. Sleep serves important restorative functions for the human body and brain [1, 2]. Two basic intrinsic components interact to regulate the timing and consolidation of sleep and wakefulness, namely sleep homeostasis, which depends on the sleep–wake cycle, and circadian rhythm, which is independent of the sleep–wake cycle. Sleep inertia results in a period of relative confusion and disorientation during transition between sleep and awake states.

Neural structures and neurotransmitters involved in the regulation of sleep and waking

Several neural structures have nuclei with extensive projections to the forebrain, and promote wakefulness via a variety of neurotransmitters. These nuclei include dorsal raphe and median raphe nuclei (DRN, MRN), the locus coeruleus (LC), laterodorsal/pedunculopontine tegmental nuclei (LDT/PPT) and the medial pontine reticular formation (brain stem), the tuberomammillary nucleus (TMN) and lateral hypothalamus (LH) (hypothalamus), the medial septal area and nucleus basalis of Maynert (basal forebrain), the ventral tegmental area (VTA), and the substantia nigra zona compacta (midbrain).

Glutamate is involved in ascending projections into the forebrain via a dorsal route and acetylcholine is involved in the projections via a ventral route through the basal forebrain into the cerebral cortex and hippocampus. Descending projections from some nuclei to the spinal cord modulate muscle tone [3–6].

The sleep-inducing system includes the basal forebrain, preoptic area, and anterior hypothalamus. Neurons containing gamma-aminobutyric acid (GABA) and gallanin promote sleep in the brainstem and hypothalamus by inhibiting wake-promoting neurons.

Wake and sleep neurotransmitters and their neural structures are listed in Table 1.

Table 1 Wake and sleep neurotransmitters and their neural structures
Full size table

The activity of the ventrolateral pre optic (VLPO) nuclei and median preoptic nucleus increases during transition from waking to sleep and initiates and maintains sleep by inhibiting arousal centers. Projections from the VLPO to histaminergic neurons of the TMN, cholinergic neurons of the LDT/PPT, and monoaminergic neurons of the LC and DR are reciprocally connected and are important for initiation and maintenance of non-rapid eye movement (NREM) sleep. Reduced activity of monoaminergic neurons leads to activation of VLPO nuclei, which results in further inhibition of monoaminergic and other wake-promoting neurons, thereby promoting sleep. Orexin acts as a stabilizer by stimulating monoaminergic systems and inhibiting VLPO nuclei [7–9]. Adenosine can also induce sleep by inhibiting cholinergic neurons in the basal forebrain and brainstem. REM sleep is mediated by cholinergic neurons in the LDT and PPT. During waking hours, there is increased release of 5-HT, NA, HA, and ACh, whereas ACh release is increased during REM sleep only. Slow wave sleep is associated with reduced release of all neurotransmitters.

Polysomnography, the recommended test for accurate scoring of sleep, uses updated American Academy of Sleep Medicine Manual for Scoring Sleep and Associated Events on the basis of R&K scoring criteria. It is, however, widely believed that standard R&K criteria might not be applicable for accurate scoring of sleep in studies of critically ill patients [10].

Normal sleep includes REM and NREM phases that are associated with distinct physiologic changes, neuroanatomic substrates, and neurochemicals. NREM sleep is further divided into stages N1, N2, and N3. Stages N1 and N2 are collectively regarded as “light sleep” whereas Stage N3 is regarded as “deep sleep” and is also called “slow wave sleep” (SWS). Healthy human adults enter sleep in NREM, going through a variety of NREM stages and then switching to REM sleep (stage R) 80–120 min after sleep onset. This cycle is repeated four to six times at night with progressive lengthening of stage R. In the young adult, stages 1 and 2 account for approximately 4–9 % and 45–60 % of total sleep time, respectively, whereas normal subjects spend between 10 and 25 % of their total sleep time in SWS. SWS increases in prepubertal years and declines thereafter [11].

SWS is regarded as the most restorative stage of sleep, and slow-wave EEG activity in the 0.75–4.5 Hz range in NREM sleep is regarded as the primary marker of sleep homeostasis and the need for additional sleep [12, 13, 14].

Sedation and sleep

Critical care providers often use the terms sedation and sleep interchangeably even though they are significantly different. For example, sleep is a spontaneous behavior that is easily reversed by external stimuli, organized in cyclical fashion, associated with specific EEG changes, and regulated by intricate homeostatic and circadian factors. Natural sleep is initiated by inhibition of norepinephrine release from the locus coeruleus that leads to stimulation of galanin and GABA production from VLPO nuclei, which, in turn, inhibits histamine production from tuberomammary nuclei, that promotes sleep onset. Sedation induced by GABAergic agents such as benzodiazepines and propofol bypasses the process of inhibition of norepinephrine release and directly affect ventrolateral pre-optic nuclei to initiate sleep [15, 16].

Sleep of critically ill patients

Sleep disruption in the ICU is very common and leads to significant sleep fragmentation (with an arousal/awakening index upto 60/hr) among ICU patients [17]. The total number of hours of sleep over a 24-h period may be relatively normal (7–9 h) but critically ill patients have an increased percentage of wakefulness and lighter stages of sleep, i.e. N1 and N2 sleep, with reduced amounts of SWS and REM sleep.

Sleep deprivation has been shown to have several serious deleterious effects including impaired cognitive performance, especially behavioral alertness, which is fundamental to most cognitive tasks. Sleep-deprived patients often have higher VO2, VCO2 and heart rate and higher catecholamine levels, probably as a result of stress response. Lack of attention or “wake state instability” generally worsens with more prolonged sleep disruption. Similarly, the neurocognitive effects of chronic sleep restriction are essentially the same as sleep deprivation and reduced cognitive function is also observed for sleep-restricted patients [18, 19]. Furthermore, sleep restriction can result in elevation of evening cortisol, increased sympathetic activation, reduced thyrotropin activity, and reduced glucose production. Sleep restriction can reduce natural killer cell and lymphokine-activated killer cell activity, interleukin-6 levels, and soluble tumor necrosis factor–alpha receptor 1 levels [20]. Cardiovascular events and cardiovascular morbidity have been reported among people with sleep-restricted schedules and circadian rhythm disruption. Sleep loss can cause irritability, delusions, inattention, memory loss, slurred speech, hallucinations, incoordination, and blurred vision. These same symptoms serve as criteria for diagnosis of delirium, a co-morbidity common to patients with critical illness. Delirium develops in 80 % of patients requiring mechanical ventilation and is associated with a threefold increase in re-intubation, increased hospital length of stay, and mortality [21].

Intensivists’ perspective

Several factors affect the sleep of critically ill patients, including etiology, severity and pathophysiology of illness, treatment instituted, i.e. medications, procedures, ventilators, etc., and the ICU environment itself (Fig. 1). Environmental factors contributing to sleep disruption in the ICU include light, noise, patient care activity (for example vital sign measurements, therapeutic intervention, and mechanical ventilation), diagnostic procedures (for example drawing blood, obtaining radiographs, and administration of medications)). It is estimated that patient care interactions occur 7.8 ± 4.2 times h−1 of sleep. Peak noise levels in the ICU exceed those recommended by the Environmental Protection Agency for ICU (45 dB during the day and <35 dB at night). Noise originates from several sources in the ICU, for example alarms, conversations, mechanical ventilation, telephones, pagers, and televisions. Some studies have demonstrated improvements in sleep quality, fewer arousals, and increased REM duration with use of earplugs [22, 23•, 24, 25].

Fig. 1
figure 1

Factors affecting the sleep and circadian rhythm of critically ill patients

Full size image

Nocturnal light exposure can affect melatonin secretion (as low as 100–500 lux) and the circadian pacemaker (levels between 300 and 500 lux) among critically ill patients [26, 27].

Patient ventilator dyssynchrony is a major contributor to disruption of the sleep of mechanically ventilated critically ill patients. Dyssynchrony is more predominant during non-REM sleep when entrainment frequencies are much narrower compared with other sleep stages. Entrainment refers to the phenomenon in which neuronal impulses that initiate inspiration, adjust to mechanical breaths initiated by the ventilator.

Although initial studies suggested that sleep disruption was higher during pressure-support ventilation than during assist control ventilation, subsequent studies have shown that sleep quality does not differ when comparing lower pressure assist levels of pressure support ventilation, automatically adjusted pressure support, and assist control ventilation. More recently European studies have shown that adaptive servo ventilation and proportional assist ventilation with load adjustable gain factors and neurally adjusted ventilator assist (NAVA) modes are associated with better patient ventilator interaction [28, 29•].

Several pharmacologic agents, including vasopressors, antibiotics, paralytic agents, sedatives, and analgesics, affect sleep quality and structure among critically ill patients.

Disruption of sleep homeostasis often persists for an extended period even after discharge from the ICU and can lead to posthospital syndrome [30–34].

Nurses’ perspective

To improve sleep, specific nursing awareness and practices, normal sleep, causes and effects of sleep disruption and deprivation among critically ill patients, should be included in core curriculum of critical care nursing education and continuing education programs. Duration and quality of nocturnal sleep should be discussed during multi-disciplinary rounds in the morning and specific plans should be made to promote uninterrupted good quality nighttime sleep. Critical care units should have specific protocols for promoting uninterrupted sleep at night and deliberate steps should be taken to prepare patients for sleep during proposed bedtime, including:

  • bed linen care and patients’ personal hygiene care before proposed bedtime;

  • ensuring that call light and urinal are within patients’ reach;

  • assessing any pain or discomfort and addressing that with analgesics if needed;

  • turning off the lights in patients’ rooms and drawing curtains;

  • minimizing noise during specific hours of night (for example from 10 pm to 6 am), with possible use of ear plugs [35, 36]; and

  • bundling of nursing activities.

Some studies have shown the benefit of back massage combined with muscle relaxation, mental imagery, and music in promoting better quality nighttime sleep.

Pharmacists’ perspective

Many drugs used for light sedation or analgesia in the ICU affect the sleep of healthy subjects and critically ill patients [37, 38].

Benzodiazepines are used to reduce sleep latency, promote sleep continuity, and combat anxiety. Benzodiazepines lengthen stage N2 of healthy individuals, increase TST, and reduce both slow wave and REM sleep duration. Benzodiazepines in low doses increase sleep spindles and fast EEG frequencies, whereas high doses evoke slow waves. Furthermore, benzodiazepines may induce paradoxical effects, for example insomnia and nightmares. Similar to benzodiazepines, propofol increases total sleep time without increasing time spent in REM or slow wave sleep. Opioids reduce slow wave sleep and REM sleep. Clonidine, an alpha-2 adrenoceptor agonist can reduce REM sleep.

Tricyclic antidepressants and serotonin reuptake inhibitors prolong slow wave sleep and totally or partially block REM sleep. They can also result in absence of REM-associated muscle atonia thereby affecting correct identification during polysomnography.

Antipsychotics have a variety of effects on sleep structure. For example, olanzapine increases total sleep time, and N3 and REM sleep, whereas risperidone only reduces REM sleep, and haloperidol has no major effect on sleep structure. A centrally acting alpha2 adrenergic—dexmedetomidine—has recently been used to promote sedation and diminish delirium. At least in animal studies, dexmedetomidine has been shown to promote true endogenous N3 sleep, whether such a favorable effect occurs among critically ill patients is unknown [39•]. Recent study concluded that nightly ramelteon use in elderly patients admitted for acute care may provide protection against delirium [40].

Several medications have drug–drug interactions that may affect sleep structure.

Finally, abrupt drug discontinuation may elicit withdrawal reactions. In one study, 30 % of patients experienced insomnia after discontinuation of sedatives. Withdrawal reactions may also arise from discontinuation of chronic medications at ICU admission.

Respiratory therapists’ perspective

The nighttime sleep of patients with acute respiratory failure on mechanical ventilation is substantially disrupted for several reasons, for example discomfort from endotracheal tube, stress-related difficulty communicating, ventilator alarms, suction positioning, bronchodilator therapy, and dyssynchronous breathing resulting from respiratory rate beyond the entrainment range.

The respiratory therapist should pay detailed attention to the mode and settings on mechanical ventilators to ensure patient comfort and the absence and/or minimization of dyssynchronous breathing. Central apneas as a result of hyperinflation from over-assistance, and arousals as a result of ineffective efforts can result in significant sleep fragmentation among mechanically ventilated critically ill patients.

Ventilator circuits should be adjusted to prevent any stretching and/or pulling of the endotracheal tube. The circuit should be free from condensation and endotracheal the tube should be suctioned on a regular basis to prevent patient discomfort and minimize ventilator alarms.

Bronchodilator therapy should be provided during preparatory hours for nighttime sleep and all efforts should be made to ensure continuous nighttime sleep without any interruption for bronchodilator therapy.

Respiratory therapists should be aware that sleep-deprived individuals have reduced inspiratory muscle endurance, greater upper airway collapsibility, and reduced hypercapnic and/or hypoxic ventilatory responses. These factors could be important when considering liberation from mechanical ventilation and respiratory care post-extubation [41, 42].

Dietician’s perspective

Dietician's should be aware of normal physiological changes accompanying normal sleep including reduced motility of the gastrointestinal tract, reduced salivation and swallowing, reduction in lower esophageal sphincter tone, and increased basal gastric acid secretion, peaking at sleep onset, especially. Glucose intolerance resulting from sleep deprivation should be considered when choosing enteral tube feed, especially for critically ill diabetic patients [43].

Physical therapist’s perspective

Neuromuscular weakness after critical illness has been well-characterized electrophysiologically and results from a combination of critical illness polyneuropathy (CIP) and critical illness myopathy (CIM). The combination of CIP and CIM is collectively termed critical illness polyneuromyopathy (CIPNM).

Mechanical unloading, which is associated with critical illness, stimulates a complex adaptive response that results in muscle atrophy and loss of specific force. Main risk factors for developing ICU-acquired weakness are severity and duration of systemic inflammatory response, length of ICU stay, and duration of mechanical ventilation. Other factors implicated are hyperglycemia, hypoalbuminemia, parenteral nutrition, corticosteroid administration, and neuromuscular-blocking agents. Countermeasures that lessen the effects of unloading include physical activity, nutritional supplements, and antioxidant administration. Patients with CIPNM have not only objective weakness, but also lethargy and excessive daytime somnolence. This muscular weakness can affect weaning from mechanical ventilation for liberation and also predispose critically ill patients to suffer from sleep-disordered breathing after extubation [44–47, 48•, 49].

Physical therapists, with the critical care team, should develop structured protocols to provide passive then active physical therapy for critically ill patients and assess the patient for early mobilization, even on mechanical ventilation.

Social workers’ perspective

Social workers can be important in promoting sleep of critically ill patients. They could serve as important liaison between critical care providers, patients and their families in recognizing patients’ needs, behavior, and habits during health. For example, knowledge of patients’ sleep habits and/or environment and daily activity levels could be important in fostering a comfortable sleep schedule and environment to promote sleep duration and quality.

Perspective of administrative staff

Sleep deprivation is increasingly believed to be an important factor contributing to ICU delirium. Forty percent of patients in ICU are sleep deprived and 70 % have ICU delirium that results in a 10 % increase in ICU mortality for each day the patient remains delirious. Mean intensive care unit cost and length of stay is 31,574 ± 42,570 dollars and 14.4 ± 15.8 days for patients requiring mechanical ventilation [50, 51].

Because cost of ICU care has steadily increased over the past three decades and even a small decrease in ICU length of stay could have a potentially huge financial effect on national healthcare expenditure (Table 2), hospital administration should provide additional resources for developing sleep-friendly ICU and include sleep optimization (for all hospitalized patients) as a core performance measure for ICU leadership.

Table 2 ICU costs
Full size table

PAD guidelines

Recently published pain/agitation/delirium guidelines issued by SCCM [53••] recommend the following for adult ICU patients:

  • Daily interruption of sedation or a light target level of sedation should be routinely used for mechanically ventilated patients (+1B).

  • Analgesia-first sedation should be used for mechanically ventilated patients (+2B).

  • Sedation strategies using nonbenzodiazepine sedatives (either propofol or dexmedetomidine) may be preferred to sedation with benzodiazepines to improve clinical outcomes for mechanically ventilated patients (+2B).

  • For patients with delirium unrelated to alcohol or benzodiazepine withdrawal, continuous IV infusions of dexmedetomidine rather than benzodiazepine infusions should be administered for sedation to reduce the duration of delirium among these patients (+2B).

  • No recommendation for use of pharmacological delirium prevention (0, C) including haloperidol or atypical antipsychotics (−2C).

  • There is no evidence of the effectiveness of dexmedetomidine to prevent delirium (0, C).

  • Atypical antipsychotics may reduce the duration of delirium (C) but use of rivastigmine to reduce the duration of delirium is not recommended (−1B).

  • An interdisciplinary ICU team approach that includes provider education, preprinted and/or computerized protocols and order forms, and a quality ICU rounds checklist should be used to facilitate pain, agitation, and delirium management guidelines or protocols in adult ICU (+1B).

  • Sedatives should be titrated to maintain light sedation using RASS or SAS (B), unless clinically contraindicated (+1B).

The strength of recommendations was defined as either strong (1) or weak (2), and either for (+) or against (–) an intervention, based on both the quality of evidence and the risks and benefits across all critical outcomes. A no recommendation (0) could also be made due to either a lack of evidence or a lack of consensus among subcommittee members.

Future directions

We are only beginning to understand the sleep of critically ill patients. Research must address whether poor sleep of critically ill patients is associated with worse outcome. For this to happen, methodological aspects of objective measurement of the sleep of critically ill patients must be standardized and outcomes-based definition of sleep for this population must be better defined. Large observational cohort studies with sophisticated measurements of EEG synchronized with audio–video recordings of critically ill patients is needed. Moreover, measurement of biomarkers associated with poor sleep in this patient population is needed, and association of such biomarkers with patient outcomes. Randomized controlled trials of single or multiple intervention (environmental and medical) are needed to help improve the sleep of critically patients and for better understanding of the mechanistic basis of disturbance of the sleep of critically ill patients by perturbation of sleep and sedation. Large multi-center studies of promising intervention are in the distant future, after we obtain better understanding of the sleep of critically ill patients and discover promising intervention in smaller controlled studies.

Conclusion

As discussed above, specific ICU protocols that include environmental changes (controlling noise levels, use of diurnal lighting practices) and appropriate physiologic and/or pharmacologic intervention (promoting patient–ventilator synchrony, effective pain therapy, physical therapy, and nutritional support) should be developed and instituted for sleep promotion in the ICU.