ABC of Sleep Medicine (ABC Series)

Chapter 10

Sleep in Psychiatric Disease


·        Significant mental health problems nearly always interfere with sleep

·        Conversely, poor quality sleep almost certainly directly contributes to many psychiatric disorders, particularly those involving low mood

·        Major depression is reliably associated with an altered pattern of sleep architecture, including a short latency to REM sleep and early morning waking

·        Sedative drugs used to treat anxiety and associated insomnia may lengthen sleep time but do not always improve sleep quality

·        A defining symptom of post-traumatic stress syndrome (PTSD) is recurring vivid dreams, replicating the memory of the initial stressful event together with the associated negative emotion

·        Sleep disorders associated with psychotic disorders such as schizophrenia are variable and poorly defined but can cause major concerns, especially if there is reversal of the sleep–wake cycle

·        Antipsychotic drugs may cause significant daytime somnolence, occasionally as a result of weight gain and associated sleep-disordered breathing

·        Subjects addicted to a variety of drugs whether stimulating or sedative will usually report sleep disturbance either as a direct consequence of drug misuse or when attempting to withdraw

The vast majority of psychiatric disorders co-exist with a disturbed sleep–wake cycle. Insomnia is the most closely associated problem but excessive sleepiness or even parasomnias may also be intimately linked to an underlying psychiatric condition or its drug treatment.

Increasingly, it is appreciated that the link between sleep and mental health problems, particularly those affecting mood, is ‘bi-directional’. Depression and anxiety may commonly fuel poor quality sleep resulting in sleep deprivation that, in turn, promotes further mood disturbance. Moreover, reports of worsening insomnia are well recognised as an independent risk factor for either the development of major depression within the subsequent year or a clinical relapse in established depression.


All types of insomnia are extremely common in depressed patients, particularly if they are female. Indeed, 90% of depressed subjects report unrefreshing sleep that usually arises from a combination of delayed sleep onset and impaired sleep maintenance. Waking 2–4 hours earlier than desired with an inability to fall back to sleep is also a commonly recognised pattern.

Depressed patients with insomnia usually report daytime fatigue in the absence of objective evidence for excessive sleepiness. However, in atypical depression, reflecting around 10% of all cases, frank daytime somnolence and increased appetite may replace the more typical pattern of insomnia with anorexia.

There are interesting changes in REM sleep that typically characterise the sleep architecture of depressed patients. The latency to enter REM sleep is shorter than average, a feature which may persist even with successful treatment and which is also seen in unaffected first-degree relatives. This feature has been used as predictor for a good response both to pharmacological and psychotherapeutic management. The density of REM sleep overnight is also generally increased although the initial REM sleep period is shorter than average.

It is of interest that the vast majority of routinely used antidepressant drugs are effective at supressing REM sleep. Whether this property is directly relevant to their action on mood elevation remains unclear, however.

Other recognised changes in sleep architecture, such as reduced deep non-REM (slow wave) sleep, appear less specific to depression.

Seasonal depression is seen particularly in higher latitudes. There is typically an increased need for sleep, enhanced appetite and weight gain during the winter months. Morning bright light therapy with or without selective serotonin reuptake inhibitor (SSRI) therapy is often an effective treatment strategy. There is some evidence that abnormalities of melatonin secretion, usually suppressed by light exposure, may be aetiologically important.

The inability to sleep during a manic phase of bipolar depression may present as insomnia if there is reduced insight into the underlying mood disturbance.

Antidepressants vary considerably in their sedative properties (Table 10.1). Significantly sedating agents such as mirtazapine might be considered preferential if insomnia is a major feature. The new agent agomelatine also appears to lengthen nocturnal sleep time in depression and may improve sleep quality by increasing levels of deep non-REM sleep.

Table 10.1 Sedative properties of selected antidepressant drugs.

Usually sedating


Usually stimulating













Anxiety disorders

Chronic generalised anxiety may affect at least 4% of the population. It is characterised by persistent symptoms of unease, dread or anticipation occurring on a daily basis for a period of months. Although usually considered a primary psychiatric condition, there may well be underlying medical or social reasons for anxiety which exacerbate the problem. Anxiety and associated increased muscle tension almost invariably produce insomnia, which may dominate the clinical picture.

Cognitive therapies are aimed at reducing distorted perceptions about future threats and minimising the tendency to ‘catastrophise’. Together with relaxation techniques, a non-pharmacological approach often improves sleep quality as well as the underlying anxiety.

Drugs used to treat anxiety are invariably sedative. Although traditional anxiolytic agents, such as benzodiazepines, may increase sleep time, the effects on subjective sleep quality may be mixed.

Some newer agents licensed for generalised anxiety, such as pregabalin, may reduce anxiety yet improve sleep quality, potentially by directly enhancing the deep non-REM sleep elements of nocturnal sleep.

Panic disorders may present as nocturnal phenomena although symptoms are rarely exclusive to sleep. Discriminating nocturnal panic attacks from agitated arousals due to non-REM sleep parasomnia can be difficult. Because the former usually arise from light (stage 2) non-REM sleep, subjects awake quickly and often have major problems returning to sleep. By contrast, confusion and variable amnesia for the event would be expected after a non-REM sleep parasomnia which develops from deep non-REM sleep. If medication is considered appropriate, sedating antidepressant drug therapy rather than benzodiazepines may be preferable for nocturnal panic attacks although controlled evidence from trials is lacking.

Post-traumatic stress disorder (PTSD) is strongly associated with sleep-related problems, notably insomnia and nightmares arising from fragmented REM sleep (Box 10.1). The nightmares in PTSD are physiologically and emotionally arousing, typically with vivid replication of the triggering traumatic event. Treatments include supportive psychotherapy aimed at reducing arousal levels when exposed to reminders of the traumatic event. However, drug therapy may also be appropriate and recent evidence suggests prazosin, a centrally active alpha-adrenergic antagonist, as a particularly useful agent. It has been proposed that this drug may actually enhance REM sleep activity but attenuate the emotional character of the associated dream or nightmare. More commonly used drugs to promote sleep, such as benzodiazepines, paradoxically, may increase the chances of developing PTSD in some patients if given after traumatic events. It has been proposed this reflects REM sleep suppression and subsequent impaired ability to process emotional memories appropriately.

Box 10.1 Features of post-traumatic stress disorder (PTSD)

Initial exposure to an event which presented the risk of death or severe injury. An immediate response of intense fear is followed by:

One or more of:

·        recurring intrusive memories of the event

·        recurring distressing dreams

·        re-experiencing the event through illusions or dissociative experiences

·        intense psychological distress when exposed to factors triggering recall of the event

·        physiological (autonomic) response when recalling the event

Avoidance of stimuli that serve as reminders of the event:

·        avoiding thoughts

·        avoiding activities

·        inability to recall precise details of the trauma

·        decreased ability to experience pleasure

·        feeling of detachment

·        restricted range of expressed emotions

·        expecting a foreshortened future

Symptoms of increased arousal:

·        difficulty with sleep onset and maintenance

·        irritability

·        poor concentration

·        hypervigilance

·        heightened startle response

Features in bold directly or indirectly reflect the impaired sleep–wake cycle associated with PTSD.

Psychotic disorders

It is clear from the earliest descriptions of psychoses, notably schizophrenia, that sleep is habitually disturbed in affected individuals. However, the variable nature of psychosis and the practical difficulties associated with formal sleep investigations in psychotic subjects have severely hampered systematic studies.

When assessed with polysomnography, there appear to be no specific or diagnostic sleep abnormalities seen in schizophrenic patients. The most common features are reduced sleep efficiency and an increased latency to fall asleep (typically 60 minutes or more) although sleep time over 24 hours is often increased. Contrary to early speculations that proposed a role for abnormal dreams in the generation of psychosis, there are no consistent changes in REM sleep parameters. Some studies, however, report a short latency to achieve REM sleep, as in major depression.

Schizophrenics may have reduced levels of the deepest stages of slow wave sleep although the effects of medication may explain some of the reported changes. There is also an indication of an impaired homeostatic response sleep mechanism, such that subjects do not display typical levels of rebound sleep after deprivation.

Severe insomnia is often reported as a prodromal syndrome before a significant psychotic relapse.

In practice, one of the most troublesome sleep-related symptoms reflects the reversal of the sleep–wake cycle, such that psychotic patients sleep during the day and are aroused at night. It is not entirely clear whether this reflects a biological problem with circadian timing or simply a desire to avoid contact with people during waking hours. Very recent data may support the former explanation.


Most antipsychotic drugs have a complex pharmacological action with varying degrees of dopamine and serotonin receptor antagonism. Older agents, such as haloperidol and chlorpromazine, have been largely replaced by atypical antipsychotics that have fewer extra-pyramidal side effects. However, even the newer drugs, such as risperidone, quetiapine and olanzapine, are associated with significant sedation that may limit dose titration. Furthermore, weight gain is a very common problem with virtually all antipsychotic drugs.

From the perspective of sleep-related problems, obstructive sleep apnoea (OSA) should always be considered if there is worsening daytime sleepiness in the context of weight gain. Unfortunately, partly due to thought disorders and delusional tendencies, many psychotic patients with significant OSA do not tolerate nasal CPAP mask therapy.

Neuroleptic drugs used to treat psychoses such as schizophrenia generally antagonise dopamine receptors as part of their action. This may cause or worsen restless legs syndrome and periodic limb movement disorder, potentially disturbing sleep continuity and increasing daytime lethargy.

Drug dependency


Most people are aware that alcohol is generally sedative at moderate or high doses and may aid sleep onset. Indeed, deep non-REM sleep may be initially enhanced by alcohol in the first hour or two of sleep. However, sleep quality is rarely improved overall and significant intake before bed can be expected to increase arousals later in the night. REM sleep is suppressed by alcohol early in the night but might ‘rebound’, with prominent nightmares, for example, towards morning. Furthermore, alcohol excess invariably worsens snoring and may convert simple snoring into frank obstructive sleep apnoea.

There is often a reported link between alcohol and non-REM sleep parasomnias such as sleepwalking. Evidence that alcohol may precipitate a violent parasomnia, for example, remains speculative in the absence of any appropriate published data. This remains a controversial area, particularly in forensic sleep medicine, when a court may have to decide whether any criminal behaviour was performed in an automatic state during a parasomnia or whether it was simply fuelled by the disinhibiting effects of alcohol. In many cases, it is likely that the secondary effects of late night drinking, such as sleep deprivation or a full bladder, may be important factors or precipitants for parasomnia activity in this situation.

Patients who abuse alcohol usually display a wide variety of sleep complaints, potentially covering the full gamut of disorders. Detoxification programmes need to allow for a temporary worsening of insomnia, reduced deep sleep and, in severe cases, delirium tremens. The latter phenomenon may well be due to elements of REM sleep intruding into wakefulness to cause hallucinations and agitation. This occurs because REM sleep is suppressed by chronic alcohol intake and ‘rebounds’ into wakefulness on sudden withdrawal.

Insomnia may persist for months after successful alcohol withdrawal and potentially precipitates a relapse if the subject attempts to ‘self-medicate’. In this situation, sedating antidepressants or mood stabilising drugs are preferable to benzodiazepines, which may trigger a craving for alcohol through cross-tolerance. The use of non-benzodiazepine agonists such as zolpidem in alcoholics with a demonstrated vulnerability for addiction is also not generally recommended.


Sleep disruption is common in the context of both prescribed and illicit use of opioids. This class of drug generally inhibits both REM and deep non-REM sleep. Nocturnal breathing disturbances, including suppression of breathing control mechanisms, are also a potential concern. If severe, central sleep apnoea may occur and cause significant daytime sleepiness.

Those that adopt a binge pattern of opioid intake often report poor quality sleep when intoxicated but worsening insomnia when withdrawing.

If opioids are prescribed for chronic pain control, alternative agents less toxic to sleep are recommended if sleep–wake problems develop. Similarly, if there is chemical dependency, a planned or supervised detoxification programme is often necessary.


A wide variety of stimulant agents, including excess caffeine, may produce excessive somnolence in withdrawal phases along with depressed mood and increased appetite. Many abusers of stimulant drug also fall into a cycle of taking additional sedatives at night to help sleep. A 5 mg tablet of dexamphetamine has the alerting properties of around six average-size cups of coffee.

Rehabilitation often requires specialist input. If caffeine use is considered excessive, intake should be tapered over a week at least, partly to avoid rebound headaches.

Further reading

Brower, K., Aldrich, M., Robinson, E. et al. (2001) Insomnia, self-medication, and relapse to alcoholism. Am J Psychiatry158, 399–404.

Ford, D. and Cooper-Patrick, L. (2001) Sleep disturbances and mood disorders: an epidemiological perspective. Depression Anxiety14, 3–6.

Hauri, P., Friedman, M. and Ravaris, C. (1989) Sleep in patients with spontaneous panic attacks. Sleep12, 323–337.

Maixner, S., Tandon, R., Eiser, A. et al. (1988) Effects of antipsychotic treatment on polysomnographic measures in schizophrenia: a replication and extension study. Am J Psychiatry155, 1600–1602.

Raskind, M., Peskind, E., Kanter, E. et al. (2003) Reduction of nightmares and other PTSD symptoms in combat veterans by prazosin: a placebo-controllled study. Am J Psychiatry160, 371–373.

Rodin, J., McAvay, G. and Timko, C. (1988) A longitudinal study of depressed mood and sleep disturbances in elderly adults. J Gerontol Psychol Sci43, 45–53.

Wulff, K., Dijk, D.J., Middleton, B. et al. (2012) Sleep and circadian rhythm disturbance in schizophrenia. Br J Psychiatry200 (4), 308–316.

If you find an error or have any questions, please email us at Thank you!