Sundown syndrome is a commonly occurring phenomena in geriatric medicine. There is uncertainty about the definition, the nature, the aetiology and the management of the syndrome. Nevertheless, there is general agreement on the adverse effect it can have on the patients and the carers. Drs Adrian Treloar, Subhi Yagoub and Ananth Puranik review the literature.

The observation that some patients become more confused and agitated at night was made by Hippocrates. While many researchers regard sundown syndrome as a useful diagnostic entity^1,2,3 , others have questioned this⁴. One study of the condition yielded a prevalence of 12 per cent (n=89) in a university, teaching nursing home¹. Others have quoted prevalences of 24 per cent³ and as high as 50 per cent². Sundown syndrome has been defined as the appearance or exacerbation of symptoms of confusion associated with the late afternoon or early evening hours¹.

Sundown syndrome does not appear to be associated with light. It occurs in the afternoon, but is not dusk related as one study found it to be most prevalent in the morning⁵. This might suggest that sundown syndrome is not a real, or single entity. However, where patients function well for part of the day, but are problematic at other times. The wish to preserve the period of wellness may conflict with treating the period of distress and this is a cogent reason why sundowning deserves serious study. We would define sundown syndrome as behavioural, emotional and cognitive problems associated with dementia which show a regular diurnal variation occurring most often in the afternoon and evening period.

A common and familiarly observed presentation consists of a change in behaviour, often of motor activity, ranging from minor degrees of restlessness to wandering, pacing, and changes in levels of anxiety as well as perceptual disorders mostly in the visual modality. Motor activity includes behaviours such as repeatedly picking objects and picking at one's clothes. Patients may engage in a quiet monologue, animated 'discussions' or scream, call, cry and shout. This is known as 'disruptive vocalisation⁶'.

Another study⁵ found that the most frequently observed behaviours were: making strange noises, requesting attention constantly, repeating mannerisms, throwing things, picking at things, pacing, and making strange movements. Some behaviours were manifested frequently, occurring almost constantly during waking hours; however agitated behaviours did not show a clear sundown syndrome (ie. agitation during the later hours).

Sundowning syndrome occurs in various types of dementing illnesses, including those associated with Alzheimer's disease, vascular and Lewy Body dementias. Parkinson's disease patients with cognitive impairment appear to display more florid forms of sundowning syndrome⁶and it has been described following cerebral injury⁸. Symptoms of sundowning also tend to be seen in the context of delirium. General medical illnesses and medications may also predispose to fluctuating mental states during the day. One study⁹ found that medications that were administered as needed in sundown syndrome were most often given in the early morning and mid afternoon, which they felt suggested a link to an environmental factor other than sunset.

Severe depressive disorder and psychosis have also been associated with sundown syndrome¹⁰. Psychosocial stressors and environmental changes have also been thought to be aetiological. Factors such as change of staff¹¹, change of care givers², premorbid life style¹², male gender, psychotropic medications¹⁰ number of sedatives³, sleep apnoea¹³ and fatigue have all been suggested.

Sleep disturbance, night restlessness and other circadian disturbances are frequently seen in demented patients. Changes in the suprachiasmatic nucleus and pineal gland have been suggested as the biological basis for these disturbances mediated via effects on melatonin level^14-18. Biochemical theories also abound in this field. The excess of sundowning in Parkinson's disease led to the raised possibility of dopaminergic dysfunction as an aetiological factor⁶. Anticholinergic drugs are also known to exacerbate cognitive dysfunction¹⁹. Other transmitters such as serotonin have also been implicated⁶. It is unlikely that any one receptor will provide the key to this syndrome.

Many management strategies for sundowning may be counter productive. The first priority should be to identify treatable causes of sundowning. A good drug history and behaviour profile together with questions asked of care assistants regarding care needs and behaviour patterns are essential first steps. Clear aetiologies may be elusive, however, and treatment is often needed before aetiology has been established.

Restriction of day-time sleep
Researchers have looked into the effect of manipulating the environment on such behaviour. Restriction of day-time sleep may improve night-time sleep and reduce sundowning behaviour²⁰. One study¹⁵ used restricted day-time sleep augmented by mild activity schedules, increased contact between staff, talking to patients and accompanying them in walks out-doors for 2 hours early morning and 1 hour in the afternoon. A 30 per cent reduction in the aggressive events was also linked to the walkers group²². In addition, scheduling sleep by avoiding early bedtimes and excessive time spent in bed may help²³, as may limiting food and drink that contains caffeine and alcohol²⁴. However, in a double-blind study²⁵ it was found that there was no benefit from reduced caffeine intake. Homeopathic use of valerian²⁶ or aromatherapy using oil of lavender²⁷ may give better sleep quality comparable to that obtained with hypnotic medication.

Controlling light exposure
As one might expect, controlling light exposure has been tried. Decreased exposure to bright light may give worse sleep quality²⁸, which results in sundowning behaviour¹. Some investigators found that properly timed bright light exposure may induce shifts of the circadian pacemaker and may augment the amplitude of the circadian rhythm²⁹. Evening bright light pulses of 2 hours may ameliorate sleep/wake cycle and thus improve sleep especially among those patients with the most severe sundowning^30,17.

One study subjected 14 inpatients with dementia, showing sleep and behavioural disorder, and 10 control elderly people to morning bright light and observed them for 2 months. Morning bright light therapy significantly increased total and nocturnal sleep time and significantly decreased daytime sleep time. This indicates that morning bright light is a powerful sychronizer that can normalize disturbed sleep and substantially reduce the frequency of behavioural disorder in elderly people with dementia. Others have found no significant benefit from light therapy³¹.

Memory notebook
In another study a memory notebook was tried to decrease problematic behaviour in patients with Alzheimer's disease³². Using the Multidimensional Observational Scale for elderly subjects, interviews were conducted with family and staff members. It was found that using a notebook reduced anger, agitation, and confusion during sundowning and helped to engage residents, who had withdrawn from all activities, to participate in activities again. It also helped to reduce physical aggressiveness during bath time.

Visual, auditory and olfactory stimuli
When visual, auditory and olfactory stimuli is added to a nursing home unit environment to simulate two types of environments: a home environment and an outdoor nature environment, participants seemed to prefer the enhanced environments by choosing to spend more time in them33. There were less agitated behaviours and staff and family members express a preference of enhanced environment over the regular one.

Touch massage
When a 'Touch-Massage' action was used in patients with Alzheimer's disease it was found that this decreased abnormal behaviour (79per cent during the session, and 70 per cent fifteen minutes after the session)³⁴. In a double-blind study it was shown that low voltage (6-8V) transcranial electrostimulation alleviated behaviours such as wandering and nocturnal delirium³⁵.

Barriers
The effect of visual stimuli upon Alzheimer's patients, who wander and exit from rooms, has also been studied using a two dimensional grid pattern that patients tend to perceive as a barrier³⁶. The use of a horizontal grid was found to be effective and reduced exit door contact by up to 97 per cent for patients with a diagnosis of Alzheimer's disease, but this was less effective for the non-demented wanderers and for patients with other types of dementia.

Other types of visual barriers have also been used such as window blind, cloth barriers and a combination of the two³⁷. The cloth barrier was the most effective solution, reducing exiting by 96per cent. The closed blind reduced exiting by 44 per cent. The combination of the blind and cloth barrier reduced exiting by 88 per cent. Of course, just keeping people in their rooms is not the final arbiter of success in managing such behaviours.

When non-pharmacological strategies for managing behaviour problems were reviewed it was concluded that the evidence does support their use in dementia^38,39. However, the finding⁴⁰ that walking, talking, feeding and proper personal time and support for patients with delirium improves outcome should spur us on to try harder to provide quality environments as a response to sundowning. In America restraints may be used, although we find it hard to see how such measures really fit well with quality environments.

Antipsychotic medications are probably the most widely used drugs in sundowning. It must be admitted that the evidence for the use of psychotropics in the behavioural complications of dementia is not good^41-43. Indeed, side-effects may lead to worsening wandering behaviour¹⁰ and cognition. Despite this they represent a mainstay and often first response to such problems by doctors. It may be that this is because non doctors have tried all other possibilities before asking doctors, and that doctors are not good at suggesting behavioural strategies. It may be that the drugs do have some utility and are therefore rationally used. Most worrying is the possibility that drugs are easier and cheaper than non pharmacological strategies.

Although neuroleptics are frequently prescribed for sundown syndrome, available evidence suggests that neuroleptics are more effective than placebo alone in only 18per cent of dementia sufferers with behavioural problems⁴⁴. Despite their minimal effect in a cross sectional study it was found that use rose from ⁴².1 per cent in 1988, to 47.6 per cent in 1992⁴⁵. Over that time there was a decrease in the consumption of tranquillizers and hypnotics.

In a double-blind controlled study using chlorpromazine a slight improvement compared to placebo was reported⁴⁷. The effect of the typical neuroleptics has also been investigated⁴⁸ and it was found that haloperidol and zuclopenthixol showed minimal improvement. In 5 placebo controlled studies of thioridazine in non psychotic patients with dementia it was concluded that the improvement with the drug therapy was not impressive⁴⁹. However, using the Time based Behavioural Disturbance Questionnaire (TBDQ) completed by caregivers another study concluded that traditional neuroleptics such as chlorpromazine and haloperidol and atypical neuroleptics such as risperidone and olanzapine have moderate efficacy²⁴. Though side-effects may restrict the use of typical neuroleptics, atypical neuroleptics may be better in the risk benefit ratio. Risperidone, in particular, at a low dose has been studied and seems to help behavioural symptoms in dementia⁴⁶.

In recent years, three double-blind clinical trials compared tiapride an atypical antipsychotic which is licensed in Europe, to chlorpromazine, lorazepam, and olanzapine. The results were in favour of tiapride as more effective in treating aggressiveness, agitation and wandering and with a better safety profile⁵⁰. Using clozapine to treat nocturnal behaviour in patients with Parkinson's disease revealed an excellent response, but this effect was not observed in patients with primary degenerative dementia^4,6,51.

The second most frequently used drug group for sundown syndrome is the benzodiazepines. Although they were reported to show significant sleep improvement⁵² the effect on the sundowning behaviour seems to be minimal4,24. Chlormethiazole may be useful in the treatment of sundown syndrome^53-55 and the non benzodiazepine hypnotic zolpidem has been shown to improve night -time wandering⁵⁶. Selective serotonin reuptake inhibitors have also been suggested to be useful in managing the behaviour complications of dementia as well as having a better side-effect profile than tricyclic antidepressants⁵⁷.

Although several reports suggest that intermediate to high doses of propranolol (80Ð160mg and 200-600mg/day) can effectively treat aggressive behaviour in dementia, hypotension may be a problem. In order to minimize these side-effects, a study⁵⁸ followed-up 12 demented patients with aggressive behaviour treated with low dose propranolol monotherapy. The aggression ratings showed that low dose propranolol was effective in 8 of 12 patients (67 per cent) within 2 weeks of treatment.

Based on the suggestion that the circadian rhythm disturbance could be the basis of sundown syndrome, melatonin was used for the treatment of sleep disorders and of sundown syndrome in patients with dementia. Studies gave conflicting results for sleep disorders^59-61 but results for sundowning behaviour were encouraging^62-63. The use of antidementia drugs remains, in our view experimental and unclear. In a study of 1,640 nursing home residents, wandering and physical abuse were reduced by the use of tacrine⁶⁴ (now not in use because of hepatotoxity). Others have also reported a reduction in behaviour problems with such drugs⁶⁵, but their use in such circumstances was not considered by the recent NICE guidelines in the UK⁶⁶.

There is a huge amount to learn about the management of sundown syndrome. Underlying biochemistry and physiology is poorly understood and pharmacological management protocols poorly validated. Interestingly, there is perhaps a slightly better evidence-base for non pharmacological treatments. The best advice at present, is to use non pharmacological methods first followed by small amounts of medicine a few hours before the problem behaviour starts. There is a real need for these practices to be subjected to proper studies.

Dr Adrian Treloar is Consultant and Senior Lecturer in Old Age Psychiatry, Kings College and Memorial Hospitals, London. Dr Subhi Yagoub is Specialist Registrar in Old Age Psychiatry, Guy's Kings and St Thomas Hospital Rotation. Dr Ananth Puranik is Consultant in Old Age Psychiatry, Invicta NHS Trust, Maidstone, Kent

References are available from the editorial office

return to main menu