Background
Alzheimer’s disease (AD) is a progressive, neurodegenerative disorder that initially exhibits cognitive symptoms, such as impaired memory, disorientation, and reduced executive ability, which are accompanied by deterioration in activities of daily living (ADL) [
1]. AD is the most common form of dementia and its prevalence increases exponentially with age. About 10 % of people over 65 years will experience some type of dementia, whereas one-third of those over 85 will be affected by this condition [
2].
The need of assistance starts early in patients with AD, and deficits in instrumental ADL tasks have been reported even in the stages of mild cognitive impairment and mild AD [
3,
4]. Their care needs develop constantly and about half of the individuals with dementia need help with personal care, whereas the remaining half will need such help over time [
5]. This results in an increasing demand for community-based services (e.g., home help and adult day care) and nursing-home placement (NHP), the costs of which rise dramatically with advancing disease severity [
6]. Currently, about 30 %–50 % of people with dementia in high-income countries receive care in nursing homes (NHs) [
5]. The annual cost of NH care in Sweden for one recipient in 2014 was SEK 610,000 (~USD 69,600, ~EUR 63,700) [
7]. In middle-income countries in particular, an increased demand for high-quality care is expected. It is necessary for health authorities to plan and allocate resources for different types of dementia care and special accommodation [
5].
The symptomatic treatment mainly recommended for mild-to-moderate AD is cholinesterase inhibitor (ChEI; donepezil, rivastigmine, and galantamine) therapy. ChEIs inhibit the degradation of the neurotransmitter acetylcholine by acetylcholinesterase, resulting in increased levels of acetylcholine in the synaptic cleft and increased availability for receptor absorption. This sustains cholinergic transmission and improves communication between neurons [
8]. The level of short-term response to ChEI and the long-term outcomes vary among persons with AD [
9]. No difference in effect has been detected among the three ChEI agents [
10,
11]. However, higher ChEI doses, regardless of drug agent, were associated with better longitudinal cognitive and functional capacities [
11‐
13]. Significant associations between a higher dose of ChEI and lower amount of home-help services [
14], postponed need for NHP [
15,
16], and longer life-span [
17] have also been observed. Moreover, we reported recently that a longer ChEI treatment duration was independently related to longer survival in AD [
17]. Whether different aspects of ChEI therapy alter the survival time in NHs has not been investigated.
Few previous studies have described the length of time spent in NHs for AD patients exclusively. Two American studies observed a median of 33 and 38.57 months, respectively [
18,
19], although the factors that might affect this parameter were not addressed. However, predictors of mortality in NH residents with AD have been reported, such as male sex, older age, and functional disability. The severity of dementia and the impact of comorbid illnesses yielded inconsistent results [
20].
Here, we aimed to identify sociodemographic and clinical factors, including various aspects of ChEI treatment, AD progression, and use of community-based services, that might predict survival time in NHs.
Discussion
In this longitudinal AD study performed in a routine clinical practice setting, we found that the mean survival time in NHs was about 4 years. Females spent 1.75 years (21 months) more in NHs than males. The general linear model showed that a shorter stay in NHs was independently associated with the interaction term male living with a family member, use of antihypertensive/cardiac therapy, use of anxiolytics/sedatives/hypnotics, and lower basic ADL capacity (but not IADL) at NHP. Cognitive ability did not affect the survival time in NHs. Males living with family spent ~2–2.5 years less in NHs compared with the other groups, despite the absence of significant differences regarding age, disease severity, or number of concomitant medications. Users of antihypertensive/cardiac therapy or anxiolytics/sedatives/hypnotics spent 1 year less, on average, in NHs than nonusers. Impairment in basic ADL might imply a shorter time spent in NHs of up to 1.5 years, depending on severity. No relationships between different aspects of ChEI treatment and survival time in NHs were detected.
The mean time spent in NHs (4.06 years) found in the present AD study is consistent with the 4.3 years observed for demented persons reported by the National Board of Health and Welfare, Sweden [
26]. A recent German study of dementia [
27] described a similar survival time in NHs for females (mean, 52.5 months), and a slightly longer time for males (37.1 months) vs our SATS (54.4 and 33.4 months, respectively). Two American studies of AD reported a shorter time spent in NHs by almost 1 year, median of 33 and 38.57 months, respectively [
18,
19], which might be explained by differences between health care systems. In most countries of northern Europe, NHP is predominantly based on the individual’s need for care, which is almost exclusively publicly funded and covers the entire population, irrespective of income or insurance [
28]. This was supported by our finding that the level of education did not affect survival time in NHs. Social-service systems that ease the family members’ economic burden of care can lead to a longer time spent in NHs.
Here, males living with a family member spent a shorter time (by ~2–2.5 years) in NHs than did the other participants, after controlling for sex, age, disease severity, and concomitant medications; hence, the usually shorter lifespan of males cannot explain this observation. A large number of previous studies of dementia investigated separately the potential effects of sex and living status on community-based service utilization; in contrast, analyses of the interaction term are scarce. We reported previously a longer time (mean of 6 months) to use of home-help services [
14] and a lower risk of NHP [
16] among male AD patients living with family vs the corresponding females. These differences are probably explained by the informal care provided by female spouses, who are more likely to provide care for a longer period than male caregivers [
29]. Moreover, the service providers may be less alert to the needs of female caregivers regarding formal help because they assume that these mostly older women can manage the care by themselves [
30]. Younger females may be more educated, less committed to caring for relatives, and more assertive; thus, the demand for formal care and NHs may increase in the future.
Users of antihypertensive/cardiac therapy or anxiolytics/sedatives/hypnotics exhibited a shorter life expectancy (by about 1 year) in our multivariate models. Cardiovascular disorders are well-known risk factors for increased mortality in NH residents with dementia [
20,
31]. The SATS patients who used antihypertensive/cardiac therapy were older and cognitively better at NHP, which might explain the similar age at death compared with nonusers. Furthermore, the MMSE scale was not adapted to measure altered executive function and mental speed, which are usually related to vascular pathology and might be expected in users of antihypertensive/cardiac therapy. Tests that evaluate these abilities specifically might have shown a greater cognitive impairment. Users of anxiolytics/sedatives/hypnotics were also older at NHP and had a similar age at death as the nonusers. A study of NH residents with dementia showed that the use of anxiolytics had a positive effect on basic ADL [
32], which might postpone NHP. Recently, we found an association between the use of anxiolytics/sedatives/hypnotics and a longer survival time [
17]. Improved sleep may lead to a more favorable outcome in AD [
33] and might be an additional reason for delayed NHP.
Cognitive ability at NHP or its rate of progression did not influence survival time in NHs in the current study, which is in accordance with most previous studies of mortality in NH residents with dementia [
20,
34], but not with all [
35]. Consistently, cognition was not an important predictor of use of home-help services [
14,
36] or time to NHP [
16,
37]. Rather, a lower cognitive status was associated with fewer hours of home help [
14]. This noteworthy observation suggests that the individuals’ cognitive ability was not considered when deciding the amount or level of community-based care needed, and that the recipients and family members might have been unable/unwilling to request additional help. The services provided must be better tailored to the requirements of persons with cognitive impairment.
In contrast, basic ADL (but not IADL) were an independent predictor of survival time in NHs in the present study. Worse basic ADL at NHP, a longer time between AD diagnosis and NHP, and, thus, a shorter time spent in NHs exhibited significant relationships, indicating that a longer period of care in the recipient’s home might reduce survival time in NHs. However, the burden of care for demented individuals with loss of many basic ADL must also be considered; this issue was not addressed in the SATS. Functional disability has been reported as an important predictor of mortality in many studies of NH residents with dementia [
20,
34]. The time to NHP seems to be more dependent on the participants’ total ADL or IADL capacity and their rates of decline [
16,
37], whereas basic ADL deficits affect life expectancy. Progression in IADL might have a lesser effect on survival vs the consequences of the loss of crucial functions during the later stages of AD, which may lead to poor personal hygiene, malnutrition, incontinence, and falls [
38]. Therefore, basic ADL might serve as an indicator of the remaining life-span in end-stage dementia.
Age was not a significant predictor of survival time in NHs in our multivariate models, suggesting that the remaining factors had a stronger effect. Furthermore, the amount of home-help services or adult day care at NHP, or the annual change in these services did not predict time spent in NHs. This indicates that community-based services cannot reduce the AD patients’ survival time in NHs and supports our finding that cognitive status was not related to service utilization. Home-help services cannot sufficiently address the needs associated with cognitive impairment, such as supervision, management of behavioral symptoms, and avoidance of dangerous situations [
36]. A study of adult day-care services observed that the recipients who used more care/week had an increased risk of NHP, even after controlling for disease severity and caregiver burden, which implies that adult day care serves more as a transitional period to NHP than as a respite, and, thus, indicates shortened time to admission [
39]. This information is essential for service providers and raises questions regarding the utility of the services available currently to community-dwelling people with dementia.
Here, different aspects of ChEI therapy, such as drug agent, dose, level of short-term response, or treatment duration, did not affect the survival time in NHs. A modest-sized mean improvement in cognitive ability for about 6–12 months [
11,
40] and a corresponding reduced decline in ADL [
41] after the initiation of ChEI therapy were described by previous AD studies. Moreover, higher ChEI doses have been associated with better longitudinal cognitive and functional outcomes [
11,
13,
42], a lower volume of community-based care [
14], delayed time to NHP [
15,
16], and a longer life-span [
17]. Taken together, the initial response to ChEI, the slowing of disease progression, and the postponement of the need for community-based services entail a 6–12 month positive shift in the disease. Hence, the unchanged length of stay in NHs observed indicates that this potential increase in life expectancy occurs in the mild-to-moderate stage of AD, when the persons are able to live in their homes.
The strengths of the SATS are its 6-month, prospective, well-organized evaluations of cognitive and functional performance and resource utilization (which is less commonly measured in most studies) after the onset of ChEI treatment. Survival time in this large cohort of “real-life” AD patients with comorbidities and concomitant medications from Swedish memory clinics has now been followed for 17 years. The community-based services in Sweden are publicly funded and, therefore, accessible to all residents, regardless of their socioeconomic status [
28]. Thus, we assume that the participants were representative of the general population and that their needs for formal care reflected their actual disabilities. Similar to other long-term observational studies of AD, the SATS is not placebo controlled (because of ethical concerns) or randomized with respect to ChEI agent. Another limitation is that somatic disorders and other health events, for instance, were not recorded between the last follow-up and death, which may have affected mortality.
Very few studies of AD exclusively have investigated the residents’ length of time spent in NHs and possible predictors that might affect this parameter, probably because of the long study time that would be required. No previous studies addressed the associations between type of ChEI agent, dose and duration of ChEI therapy, and survival time in NHs; thus, additional studies are warranted. The potential effect of sociodemographic factors and interaction terms—such as living status by sex, as well as various comorbidities and medications, such as anxiolytics/sedatives/hypnotics—on time spent in NHs needs further investigation.
Acknowledgements
We wish to thank all of the SATS patients and their relatives for their cooperation in this study. The authors are grateful to the staff from all of the different centers that took part in the management of the participants and provided administrative support to the study.