Sleep in Independently Living and Institutionalized Elderly

Chapter 136 Sleep in Independently Living and Institutionalized Elderly




Abstract


Although not a mainstream source of patients in most sleep clinics, an ever-increasing proportion of the population residing in institutional settings necessitates that sleep medicine specialists be comfortable with and understand issues that affect sleep in those settings. Many patients residing in assisted living facilities or nursing homes meet criteria for dementia, and sleep-related syndromes such as sleep-disordered breathing, nocturnal incontinence, falls, agitation, and wandering are very common comorbidities in those patients. Medication to improve nocturnal sleep or to enhance daytime alertness should be considered, but it must be judiciously implemented and carefully monitored. Sleep apnea interventions, such as nasal continuous positive airway pressure, can also be considered in some cases, especially in the presence of an engaged caregiver or spouse. Appropriate expectations should be fostered regarding the value of treating sleep apnea in late life. In certain cases, the sleep medicine specialist plays a critical role at the institutional level by educating staff about identifying specific sleep disorders (sleep apnea, restless legs syndrome), relevant sleep pharmacology, and principles of nonpharmacologic management, including optimizing the sleep environment. Treatment of sleep and sleep disorders in nursing home patients can benefit both patients and their caregivers and significantly affect quality of life.


This chapter is organized around common clinical problems commonly seen by the sleep medicine specialist who encounters geriatric patients, including those living independently, but especially those residing in assisted living and nursing homes. For many older persons, loss of cognitive capacities in late life overlay the multiple medical morbidities and syndromes (e.g., nocturia, incontinence, wandering, agitation) that constitute problems with sleep in this population. Treatment considerations must be made viewing the frailty of the patient, evaluating the risk-to-benefit ratio, and, ultimately, the cost and feasibility of implementing treatment. Elucidation of central nervous system pathophysiology underlying disturbed sleep in specific dementing conditions is covered in greater detail elsewhere (see Chapter 91). This chapter focuses on issues related to interventions, including their efficacy and rationale.



Sleep Problems in Demented Nursing Home Patients


Clinicians encountering patients with sleep problems in institutional environments face innumerable challenges. Considerable heterogeneity exists in the range of dementing conditions that are present. Patients are most likely to carry a diagnosis of Alzheimer’s disease (AD), but dementia with Lewy bodies (DLB), frontotemporal dementia (FTD), Parkinson’s disease with dementia (PDD), and vascular dementia (VasD) are likely to be present in sizeable proportions of the nursing home population as well (see Chapter 91 for a more-detailed description of these conditions). Many patients are placed in long-term care without the benefit of well-documented neurologic evaluations and neuroimaging, and chart notes are unlikely to contain sufficient detail to ascertain which specific diagnoses may be present in a given patient. With the possible exceptions of DLB (see Nocturnal Wandering, later) and VasD (see Sleep Apnea, later), the specific diagnosis of dementia subtype is unlikely to affect treatment decisions.



The Nursing Home Environment as a Challenge to Sleep Integrity


A major challenge to the integrity of sleep–wake in the nursing home is the environment itself. The typical nursing home patient receives little or no high-intensity light to synchronize circadian rhythms, may be confined or put to bed for long periods, may be subject to little or no daytime stimulation, and may be subjected to noise and light disruptions at night. Transfer from noninstitutional to institutional settings is associated with longer time in bed, increased daytime napping, and earlier bedtimes, even within a period of several months,1 although there are huge differences between facilities in terms of how much time patients are placed in bed during the daytime hours.2


Lower ratios of staff to resident may be a relevant factor,3 but time in bed in the nursing home is a complex issue. Nursing home patients experiencing more pain may spend more time in bed but have poorer sleep.4 A study encompassing 53 nursing homes across the United States showed that facilities placing patients in bed during the daytime were reported to have residents who exhibited lower rates of agitated behavior when compared to facilities that that did not use daytime bedrest,5 which demonstrates the complexity of implementing good sleep hygiene in these environments. Apart from staffing issues, if nursing home personnel believe that their patients “need rest” for fatigue during the daytime hours6 and that such practices reduce disruptive behavior, it may be difficult to convince them otherwise. In the outpatient setting, daytime sleepiness in the dementia patient is seldom viewed as problematic by the caregiver.7 Institutionally based data have linked peak agitation to the time when nursing staff shift (typically 4 PM)8 and imply that staff change is disruptive for patient’s behavior.


A broader perspective comes from a nursing home study that based ratings of negative and positive behavior based on observations made in the hours before and after sunset and noted that frequencies of both kinds of behavior appeared to increase in the hours before sunset.9 Such more-generalized behavioral activation might represent a manifestation of the wake-maintenance zone in cognitively impaired patients.


In the United States, many nursing facilities continue to interpret the Centers for Medicare and Medicaid Services (CMS) guidelines, established in the1987 Omnibus Budget and Reconciliation Act, as providing a mandate for nightly bedchecks and awakenings on a 2-hour basis for all nursing home patients at risk for developing pressure ulcers because of incontinence. The logic is that dementia patients, if not periodically awakened and repositioned and having their bedding changed, would more readily develop erythema and frank bedsores and be subject to their corresponding morbidity.10 However, 66% of incontinent bedridden patients demonstrated spontaneous mobility during the night, both at the shoulder and hip, at rates in excess of one turn per hour,11 thus potentially obviating the need to enforce awakenings in at least some patients. Not surprisingly, such staff-induced nocturnal awakenings have been shown to induce agitation in nursing home patients.1


More-profound levels of dementia, both within and outside of the nursing home setting, are associated with greater daytime sleepiness.12 This appears to be the case for assisted living facilities as well.13 Considerable descriptive evidence, relying upon behavioral observation14 and actigraphy15 in the institutional setting and relying on caregiver report16 in the home setting, suggest that daytime sleepiness and napping might have functional consequences for participation in social activities, occupational therapy, and even independence in selected activities of daily living. Lower amounts of REM sleep at night have been related to higher levels of observed sleep in nursing home patients.17 Actigraphic data also suggest that longer sleep durations per 24-hour period were associated with more-profound dementia in the nursing home,18 these data being compatible with loss of the wake-promoting function of the mammalian suprachiasmatic nucleus.19



Treatment



Pharmacologic



Sedative-Hypnotics


Controversy exists over the use of medication for sleep in the nursing home. Injurious falls and hip fracture remain one of the most serious risks ascribed to such medications. Reluctance of many physicians to prescribe certain types or classes of medication was codified with the publication of the consensually derived Beers list of undesirable medications for older patients. The modified Beers list20 includes several older benzodiazepine sedative-hypnotics, consisting of all dosages of flurazepam and temazepam (doses >15 mg) and triazolam (doses >0.25 mg). Both short-half life and intermediate half-life benzodiazepines are discouraged in geriatric patients with a history of syncope or falls. Gamma-aminobutyric acid A (GABAA) selective agonists (zolpidem, zolpidem MR, zaleplon, eszopiclone) were not specified on the list. Diphenhydramine was included, and a study of acutely hospitalized geriatric patients demonstrated an exceptionally high risk for delirium with its use.21 Anticholinergics have been known for years to cause delirium.22 A Finnish nursing home study suggested that as of 2003, the most commonly misused drug on the Beers list for nursing home patients was temazepam, used at doses greater than 15 mg.23 As of early 2009, there are no published randomized, placecbo-controlled clinical trials of the newer sedative-hypnotics (eszopiclone, ramelteon, sustained-release zolpidem) specifically studying demented, institutionalized patients. Some trials cited in this chapter have included zolpidem and zaleplon.


A widely cited meta-analysis24 looked at 24 randomized clinical trials of sedative-hypnotics in elderly persons living independently, in various types of senior living facilities, and in nursing homes; five of the studies used zolpidem or zaleplon encompassing about 15,800 patient nights. The meta-analysis concluded that although sedative hypnotics as a group improved sleep time by an average of about 25 minutes, such medication increased the risks for altered cognition (odds ratio [OR], 4.78) and daytime fatigue or sleepiness (OR, 3.82). The effects on falls and psychomotor impairment (OR, 2.25; 95% confidence interval [CI], 0.93-5.41) did not reach significance; no falls were associated with zolpidem or zaleplon. Of the eight falls occurring across all these trials, none specified the time of day when the fall occurred.


Apart from such meta-analyses, most knowledge of potential efficacy and adverse effects of sedative-hypnotics in demented nursing home populations derive from retrospective analyses of various administrative databases. Such data often represent relatively crude summaries of specific times when medications are ingested and when adverse events occur. They also cannot track when and for how long use of a sedative-hypnotic may be successful. For example, apparent lack of efficacy of sleep medications25 or even worse sleep26 with sedative-hypnotics in institutionalized patients have frequently been noted, though other databases suggest that hypnotics may be effective in many patients.27


Wang and colleagues28 examined 1-year risk of hip fracture in association with zolpidem in a retrospective case-control study of elderly community-dwelling patients undergoing surgical treatment for a hip fracture. Apparent zolpidem use was associated with nearly double the risk, but the specific ingestion of zolpidem related to the index event (hospital admission date) was inferred on the basis of adequate supply of zolpidem, not documented ingestion of a particular dose at specified time of day. The authors later argued further that residual confounding (i.e., patients most likely to be at risk for hip fracture were those most likely to ingest zolpidem) did not account for their findings,29 but they presented no data on time of drug intake or time of day of fall leading to hospital admission date to clarify their original finding.


Although a substantial amount of epidemiologic literature beginning in the late 1980s has demonstrated that psychotropic medications generally (and sedative-hypnotic medications specifically) are associated with increased risk for falls and hip fracture, these observational studies, often derived from either community-dwelling30 or institutionalized31 populations, cannot be considered definitive for several reasons. First, not all population-based studies uniformly have reported such effects,32 implying that the effects may be specific to population, medication, or database.


Second, virtually every study fails to account for the time of day the fall or injury occurs. This is highly relevant because such information on timing and dose of medication in relation to when the fall occurred is essential to establish whether a plasma level could reasonably be assumed to be present. An exception to this is the study of Ray and coworkers33 who examined medication dose and half-life and time of day (daytime versus nighttime) as predictors of falls in 2500 Tennessee nursing home patients and showed that acute use of short elimination half-life medications (primarily temazepam, and much smaller numbers using oxazepam, zolpidem, and triazolam) were associated with falls only in the nighttime hours.


Avidan and colleagues, using the Minimal Data Set (MDS), an administrative database of Medicare and Medicaid patients, examined successful and unsuccessful use of sedative-hypnotic medication in relation to falls in 34,000 Michigan nursing home patients.27 In these analyses, insomnia (derived from a single item on the MDS) and sedative-hypnotic use were examined as separate predictors of falls. Insomnia per se was associated with increased risk for falls regardless of medication use, and sedative-hypnotic use without concurrent insomnia (implying efficacious treatment) was not associated with falls (Table 136-1). Although such use of the MDS has been severely criticized and could not be validated by reference to actigraphic measurements in a separate analysis of about 180 patients,34 the Michigan nursing home data provide a broader perspective for interpretation of administratively derived databases.


Table 136-1 Prediction of Incident Falls in State of Michigan Nursing Homes (N = 34,163)



















BASELINE INSOMNIA OR 95% CL
Not using hypnotics 1.55 1.41-1.71
Using hypnotics 1.32 1.02-1.70
Using hypnotics (implied effective treatment) 1.11 0.94-1.31

CI, confidence interval; OR, odds ratio.


Data from Avidan AY, Bries BE, James ML, Szafara KL, Wright GT, Chervin RD. Insomnia and hypnotic use, recorded in the Minimum Data Set as predictors of falls and hip fractures in Michigan nursing homes. J Am Geriatr Soc 2005;53:955-962.


In that regard, perhaps the most compelling reasons to suspect that falls data may be more parsimoniously interpreted as bedrise episodes associated with an inability to sleep through the night, come from studies of nocturia and incontinence. Both nocturia35 and urge incontinence have been associated with falls,36 though the latter studies also fail to specify the time of day when falls occurred. An Australian study of elderly women demonstrated that daytime sleepiness and urge incontinence independently contributed to the likelihood of falling,37 and in a United States population, daytime sleepiness38 and short nocturnal sleep duration39 were associated with falls.


Although interventional studies at the level of health care systems are difficult to implement, a quasiexperimental study of hip fracture offered some perspective on the implementation of the New York State triplicate benzodiazepine prescription policy for elderly Medicaid enrollees by comparing hip fracture rates over a similar period for comparable Medicaid recipients in New Jersey.40 Although prescriptions showed a dramatic reduction in New York (unlike New Jersey), hip fracture rates showed few changes during that interval and were comparable between states, again raising the possibility that medications are not likely to be the primary cause of such events.



Cholinesterase Inhibitors, Antipsychotics, and Stimulants


It is well recognized that the most widely used class of medications for the cognitive impairments of AD, the cholinesterase inhibitors, have disrupted sleep as a side effect, which leads to higher than expected concurrent use of sedative-hypnotics.41 Many patients report increased frequency of unpleasant dreaming with this medication class, perhaps reflecting the stimulation of cholinergic systems controlling rapid eye movement (REM) sleep. Memantine, an N-methyl-D-aspartate (NMDA) receptor antagonist, is also approved for moderate to severe cognitive loss in AD, but no studies have specifically reported on its effects on sleep. Some retrospective data analyses derived from the MDS suggest that despite disrupting sleep, cholinesterase inhibitors were associated with lower use of antipsychotics.42 Another study demonstrated that although antipsychotics were not specifically prescribed for sleep in nursing home patients, their withdrawal was associated with lower sleep efficiency as measured with actigraphy.43


Judicious use of antipsychotics in dementia is strongly advised, because this constitutes an off-label use of such drugs. Essentially, promotion of nocturnal sleep by these agents represents an opportunistic use of the side effect of sleepiness (at a rate approaching three times that of placebo as estimated by meta-analysis),44 otherwise cast as an adverse effect. When administered to elderly dementia patients, older-generation antipsychotics (haloperidol, thioridizine)45 and newer atypical antipsychotics (olanzapine, risperidone, quetiapine, aripiprazole) increase risk for sudden death, and the FDA has issued a black-box warning for their use in such populations, the latter supported by meta-analysis of clinical trials46 and by a case-control study involving more than 13,000 matched patient pairs.47


Use of stimulant medication in dementia, targeting the daytime sleepiness and the apathy that characterizes dementia, has been the subject of several small-scale studies in nursing home patients48 and community-dwelling49 AD patients. Results suggested good tolerance and some improvement in apathy. A double-blind, placebo-controlled trial of methylphenidate for AD confirmed these improvements, though some adverse events (restlessness, agitation) were problematic,50 suggesting cautious use of this drug class. There are no published randomized trials with modafinil in institutionalized or noninstitutionalized AD patients, though a small case series in mixed-dementia patients suggested some benefit at 100 to 200 mg.51 Controlled trials in outpatients with PD at doses up to 400 mg have shown only mixed results in increasing daytime alertness.52



Nonpharmacologic


A large number of randomized clinical trials employing nondrug interventions to improve sleep in institutionalized patients have shown mixed success with such approaches. In contrast, multifaceted behavioral treatment programs involving engaged caregivers of dementia patients have shown better rates of success, extending out as long as 6 months subsequent to the intervention.53 The nursing home studies have employed parallel group designs involving hundreds of demented nursing home residents, have incorporated credible control treatments, and have employed a variety of measures attempting to assess sleep, using wrist actigraphy and systematic behavioral observations. Many studies have included secondary outcomes of importance such as mood, functional status, and nurses’ ratings of residents’ behavior.


Sloane’s group54 employed relatively high intensity (2500 lux) lights embedded in existing fixtures in common rooms, such as dining areas, in two geriatric facilities for 3 weeks across four lighting conditions: customary, bright light in the morning (7 AM to 11 AM), bright light in the evening (4 PM to 8 PM), bright light all day (7 AM to 8 PM). Sleep durations improved modestly (about 15 minutes) for morning and all-day exposures. The amplitude of wrist actigraphic rhythms did not change substantially, though some changes in phase were noted.


Two other studies failed to find markedly beneficial effects of bright light in nursing home patients. Alessi and colleagues55 combined 5 consecutive days of morning outdoor light exposure (documented at >10,000 lux) with an intensive program to limit time in bed during the day, increase daytime physical activity, and reduce nursing home noise and light at night. Effects were dramatic reductions in daytime sleep but no significant change in hours of nighttime sleep or number of awakenings. Dowling and coworkers56 used a much longer treatment period (10 weeks of active treatment), but combined light treatment (1 hour morning exposure at 2500 lux) with 5 mg melatonin. As in the Alessi55 study, the most dramatic effects were seen in reducing daytime sleep, but nighttime sleep did not show differential improvement relative to control condition.


Dowling’s group56 demonstrated marked improvements in the actigraphically measured rest–activity rhythm (e.g., amplitude, improved cosine goodness of fit), which the authors interpreted as compatible with functional improvements.


The most ambitious attempt at improving sleep in demented nursing home patients was a 3.5-year trial comparing fixture-based light alone (1000 lux administered from 10 AM to 6 PM), melatonin alone (2.5 mg medium-fast release), combined light and melatonin versus a dim light (300 lux), and a placebo-control condition.57 Although the study design was straightforward, the analyses were complicated to interpret because residents entered and left the study at various periods; only 7 of 189 residents completed the entire study. Many participants entered the protocol after the baseline period was completed, and only about 30% completed even the 6-month follow-up. Intent-to-treat analyses employing last observation carried forward, an approach of questionable utility in dementia studies,58 were employed, but analyses were offered only for 3.5- and 1.5-year follow-ups. Given these constraints, the authors’ results were difficult to interpret. For example, light significantly improved sleep duration by 10 minutes and melatonin significantly improved sleep duration by 27 minutes, but the combined interventions did not produce a significant effect on sleep time. Other notable nonsleep outcomes were decreases in mood and greater behavioral withdrawal associated with melatonin and improved mood and cognition with light. The latter extrapolated to an improvement of less than 1 point over 3.5 years on the Mini Mental State Examination (MMSE). Taken together, these results did not suggest dramatic improvements.


The melatonin results in particular are difficult to interpret given the large scale (n = 157) multisite National Institutes of Health study of AD outpatients,59 which demonstrated few beneficial effects on sleep of melatonin at a dose of 2.5 mg (sustained release) or 10 mg (immediate-release) in an 8-week trial using actigraphic measurements. This trial had few drop outs (6%) and showed no change in Hamilton Depression Rating Scale score. Similar negative results were reported in a smaller randomized clinical trial using a single dose of combined 8.5 mg immediate release and 1.5 mg sustained-release melatonin administered for 10 days.60


Perhaps the most intriguing manipulation attempting to improve sleep was an individualized social activity intervention applied daily for 3 weeks in 147 demented nursing home patients.61 The intervention was tailored thoughtfully to the specific background and (often former occupational) interests of the residents and compared to usual nursing home care. Because of broadly defined entry criteria, many patients slept more than 7 hours at baseline (as recorded by actigraph). When analyses were limited to subjects with low baseline sleep efficiencies (<50%), total nocturnal sleep times increased by about 40 minutes and daytime sleep times and nocturnal sleep latencies both decreased by 40 minutes each. A unique aspect of that study was that the cost of such an intervention (if several patients were engaged in an activity at the same time) was estimated at about $10 per patient per day, probably not an unrealistic expense given today’s health care cost burden.



Nocturnal Wandering and Restless Legs Syndrome


As a particular form of agitation, wandering, particularly when occurring during the evening and overnight hours, represents a significant safety problem in the care of the aged demented patient and is potentially one of the most troubling symptoms for caregivers,62 though one not always validated by actigraphic measurement of the caregivers’ own sleep.63 So great is the threat to welfare of senior citizens that many communities have implemented a “Safe-Return” program to deal with the problem.64 Alarms and egress warning systems have been implemented in institutional settings and in home environments. The prevalence of nocturnal wandering remains uncertain, though estimates range as high as 52% to 69% at some point in the course of the dementia patient’s illness.65


Although environmental controls are meaningful, the sleep medicine specialist may offer consultative input at several levels. Obtaining an accurate history of when and how often the episodes occur may be revealing on several fronts. If episodes arise after several hours of sleep, the possibility that the wandering represents an exacerbation of dream-enactment behavior reflecting Parkinsonism or DLB (see Chapters 87 and 91) should be considered. This would lead to treatments such as clonazepam or over-the-counter melatonin, which have been tested as treatments for dream-enactment behavior (though not specifically wandering) in open-label trials (see Chapter 95). No randomized, placebo-controlled clinical trials for wandering have specifically tested the efficacy of newer, prescription-based melatonin receptor agonists (ramelteon, tasimelteon, agomelatine).


By contrast, wandering behavior commencing in the early evening hours could be a presentation of restless legs syndrome (RLS). The 2002 NIH Development Conference on RLS suggested that RLS in the dementia patient could be manifest as early evening wandering and pacing.66 At least one videographic, observational study of “travel behavior” in nursing home patients demonstrated that such activities increased in frequency between 7 PM and 9 PM67 and were probably independent of shift change, thus lending some credence to this possibility.


Recorded PLMS, often a correlate of RLS (see Chapter 90), has been reported to be an independent cause of sleep disruption in community populations,68 and it is common in dementia patients, where it contributes significantly to reductions in total sleep time.69 Both caregiver-reported twitching and restless legs differentiated AD patients from such self-reported symptoms in controls,70 but they might not have generalized to caregiver’s report of “sleep problems” in those patients.71


Use of selective serotonin or noradrenergic reuptake inhibitors or the presence of musculoskeletal pain conditions tended to be related to a periodic leg movement index of more than 15 events per hour in community-dwelling dementia patients,72 and use of antidepressants or antipsychotics was associated with a greater likelihood of past injury from wandering in AD patients.73


Positron emission tomography and single photon emission computed tomography data suggest that AD patients who wander had reduced D2 receptors throughout the caudate and putamen74; these data were compatible with neuroimaging findings in RLS (see Chapter 90). Taken together, these diverse findings are compatible with the idea that wandering in at least some AD patients could represent RLS.


Diagnosis may be problematic. It would seem obvious that cognitively compromised dementia patients might not have full language faculties to allow complete verbalization of symptoms so aptly articulated by many patients with RLS. In that situation, careful inquiry should be made into medical risk factors for RLS in such patients, including anemia, kidney disease, neuropathy, and musculoskeletal conditions (see Chapter 90), because these would serve to corroborate a diagnosis of RLS. Another potentially important source of information comes from family members who might reveal a history of lower limb discomfort in the patient (or perhaps, in themselves or other family members), which might steer the clinician to consider RLS as a basis for the wandering. Wandering dementia patients were reported to be more likely to have used walking as a preferred behavior for coping with stress.75


Before attempting new pharmacologic intervention, careful examination of current medications might yield information. Atypical antipsychotics, often producing at least partial dopamine blockade, are sometimes used as drugs of last resort to control agitated behavior (including wandering) in dementia patients. In the case of a demented patient with medical or genetic risk for RLS, such mediations might worsen, rather than ameliorate, wandering. In this situation, a trial suspension of such medications should be entertained. In the absence of such medication and in the presence of corroborating evidence compatible with RLS (e.g., low serum ferritin), symptomatic treatments for presumed RLS in the dementia patient who wanders in the early evening might be undertaken cautiously, bearing in mind the potential adverse effects of dopamimetics in this population. As in any case of suspected RLS, careful examination of iron markers and use of iron supplementation may be considered. A more-complete description of issues involved in the wandering as a marker of RLS is found elsewhere.76

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Mar 13, 2017 | Posted by in NEUROLOGY | Comments Off on Sleep in Independently Living and Institutionalized Elderly

Full access? Get Clinical Tree

Get Clinical Tree app for offline access