Abstract
Background
Sleep disturbances are associated with risk of cognitive decline but it is not clear if treating disturbed sleep mitigates decline. We examined differences in cognitive trajectories before and after sleep treatment initiation.
Method
Data came from the 2006–2014 Health and Retirement Study (HRS). At each of 5 waves, participants were administered cognitive assessments and scores were summed. Participants also reported if, in prior 2 weeks, they had taken medications or used other treatments to improve sleep. Our sample (N = 3 957) included individuals who at HRS 2006 were 50 years and older, had no cognitive impairment, reported no sleep treatment, and indicated experiencing sleep disturbance. We identified differences between those receiving versus not receiving treatment in subsequent waves and, among those treated (n = 1 247), compared cognitive trajectories before and after treatment.
Results
At baseline, those reporting sleep treatment at subsequent waves were more likely to be younger, female, Caucasian, to have more health conditions, to have higher body mass index, and more depressive symptoms (all ps ≤ .015). Decline in cognitive performance was mitigated in periods after sleep treatment versus periods before (B = −0.20, 95% CI = [−0.25, −0.15], p < .001 vs B = −0.26, 95% CI = [−0.32, −0.20], p < .001), and this same trend was seen for self-initiated and doctor-recommended treatments. Trends were driven by those with higher baseline cognitive performance—those with lower performance saw cognitive declines following sleep treatment.
Conclusions
In middle-aged and older adults with sleep disturbance, starting sleep treatment may slow cognitive decline. Future research should assess types, combinations, and timing of treatments most effective in improving cognitive health in later life.
Keywords: Cognitive decline, Sleep, Sleep treatment
There is an emerging interest in the role that sleep may have on neurodegeneration and cognitive aging. Sleep disturbances (which may include insomnia, obstructive sleep apnea [OSA], restless legs syndrome, circadian rhythm disorders, etc.) are highly prevalent in middle-aged and older adults with estimates ranging from 40% to 70% experiencing sleep disorders and/or transient difficulties sleeping (1–3). At the same time, there is evidence suggesting poor sleep to be associated with increased risk of cognitive impairment and cognitive decline (4–8) in epidemiological cohorts. There is some biological plausibility for this relationship—for example, sleep has been shown to serve as a clearance mechanism for neurotoxins built up throughout waking hours (9), and poor sleep is associated with accumulation of beta-amyloid burden in humans (10,11). Sleep apnea is characterized by intermittent hypoxia and reoxygenation which can lead to oxidative stress (12), one factor thought to be important in the pathogenesis of Alzheimer’s disease (AD) (13), and sleep fragmentation—inherent in sleep apnea—has been shown to increase deposition of amyloid and tau proteins in both rodent models and in humans (14,15). There is also some evidence from genetic analyses showing APOE ε4 gene is associated with both sleep apnea and AD, suggesting some possible genetic link (16). Taken together, considerable evidence exists to suggest an important role for sleep health in maintaining brain and cognitive health.
A variety of treatments exist for sleep disturbances (17), which may be recommended by a doctor or be self-initiated. For example, clinicians may prescribe sleep medications (eg, benzodiazepines and other hypnotics) or deliver behavioral treatments (eg, cognitive behavioral therapy for insomnia [CBT-I]) for the treatment of insomnia, or prescribe continuous positive airway pressure (CPAP) and mandibular advancement devices for the treatment of sleep apnea (17). While there is limited availability of behavioral sleep treatments like CBT-I (18), use of prescription hypnotics and CPAP is quite common in older persons. For example, previous estimates indicate that between 5% and 7% of older adults aged 65 and older have used sleep medications in the past month (19). There are also a number of other marketed treatments available directly to patients including herbal supplements (eg, melatonin) or other “over-the-counter” (OTC) medications. The actual prevalence of use of these patient-initiated treatments among older persons remains largely unknown, although one previous estimate showed 17.5% of older adults with sleep disturbance used OTC sleep aids (20). Past research has shown that sleep treatment may be more common among women (21), and those with more comorbid health conditions and mental health symptoms (eg, depressive symptoms) (22).
At present, the effect of treating sleep disturbance on cognitive function remains unclear. In theory, treatments that consolidate sleep might facilitate its restorative role, which may lead to improved cognition (23). For example, a study by Ancoli-Israel et al. found that CPAP among patients with mild–moderate AD and OSA improved cognitive functioning over 3 weeks (24). On the other hand, some insomnia treatments such as benzodiazepines and other hypnotics may worsen memory and lead to cognitive impairment (25). Indeed, benzodiazepines have been shown to be associated with AD risk in some (26–29) but not all studies (30–32). It should be noted that many OTC medications (eg, Benadryl), while commonly used, are not recommended for use to address sleep and there are limited data about their efficacy for sleep treatment (20). More research is needed to determine whether improving poor sleep could have the potential to mitigate cognitive decline, and whether there might be differences based upon the type of treatment (eg, self-initiated or doctor-recommended).
Based on this conceptual framework, the overall purpose of this study was to determine in a large community-based sample of older adults whether treatment for sleep disturbances was associated with subsequent improvements in cognitive performance and, therefore, whether sleep treatment might have the potential to modify cognitive performance trajectories over time. Specifically, we aimed (i) to identify participant demographic and health characteristics associated with obtaining sleep treatment and (ii) to determine the association between sleep treatment initiation and concurrent and subsequent trajectories of cognitive functioning. We hypothesized that receipt of sleep treatment would be greater among women, those with more health conditions, and those with elevated depressive symptoms, and that treatment for sleep issues would be associated with improved subsequent cognitive performance and mitigation in cognitive decline longitudinally.
Methods
Data Source and Participants
We examined data from the 2006, 2008, 2010, 2012, and 2014 waves of the Health and Retirement Study (HRS). The HRS is a biennial longitudinal cohort study of a nationally representative sample of older adults. The study was initiated in 1992 and has continued with follow-up interviews every 2 years. Each wave consists of approximately 20 000 older adults who are contacted via telephone or in-person interviews. Subjects were eligible for the study if they were noninstitutionalized adults older than age 50+ years and primary U.S. residents. The HRS is sponsored by the National Institute on Aging (U01AG009740) and is conducted by the University of Michigan.
Measures
Cognitive performance
All HRS participants were asked to complete several cognitive measures, and those older than age 65 years were additionally asked to complete a small subset of naming tasks. Consistent with previous literature (33), we focused only on those measures administered to the entire sample. For Immediate and Delayed Word Recall, the interviewer read aloud 10 nouns, and participants repeated as many words as possible after the 10 words were read. After 5 minutes of answering other survey questions, participants again recalled as many words as possible. To account for practice effects, 4 word-lists were rotated across successive interviews and 2 members of same household were never administered the same list in an interview. For Serial 7’s, participants were asked to start at 100 and subtract 7 consecutively, 5 times. Backwards Counting consisted of asking participants to count backwards for 10 consecutive numbers from 20 (point was given for correct response). Consistent with other research (33), we focused on the total cognition score (sum of all measures; 0–27) and the scores on individual tests: immediate word recall (up to 10 points), delayed word recall (up to 10 points), Serial 7’s (up to 5 points), and backwards counting (up to 2 points).
Sleep complaints
At each wave, subjects were asked to report how often (ie, “rarely or never,” “sometimes,” “most of the time”) they had “trouble falling asleep,” “trouble staying asleep,” “trouble with waking up too early and being unable to fall asleep again,” and “feeling really rested in the morning.”
Sleep treatment
Respondents were also asked at each wave “In the past 2 weeks, have you taken any medications or used other treatments to help you sleep?” with response options being “yes” or “no.” If a respondent gave an affirmative response, they were subsequently asked “Were these medications or other treatments recommended to you by a doctor?” (“yes” or “no”).
Other covariates
We also considered respondent age (years), gender (male, female), race (Caucasian, African American, other), Hispanic ethnicity, education (which we categorized as less than high school, high school diploma or equivalent, some college, higher than college degree), number of health conditions, body mass index (BMI; calculated from self-reported height and weight), depressive symptoms (as measured by the 8-item version of the Center for Epidemiologic Studies Depression Scale [CES-D] (34)), and current smoking status (yes vs no). Because alcohol use is associated with sleep quality (35) and may interact with sleep medications (36), we also assessed whether participants ever drink alcohol (yes vs no), the number of times per week they drank, and the average number of drinks per day they drank.
Participants
Our analyses focused on HRS respondents in the 2006 wave who had normal cognition (based on previous research (33) defined as a total cognition score between 12 and 27), who self-reported no sleep treatment, and who reported sleep difficulty. We defined “sleep difficulty” as the respondent indicating at baseline they experienced any of the sleep disturbance symptoms “all of the time” (or “rarely or never” for feeling rested in the morning). We then excluded those who were <50 years old. In total, our sample consisted of N = 3 957 participants.
Statistical Analyses
Analyses were conducted in 4 stages. We first sought to determine the baseline (HRS 2006) demographic and health characteristics that may predict future receipt of treatment for sleep disturbances. We compared baseline measures using multinomial logistic regression among those who in the years after 2006 reported (i) no future treatment (which served as the reference), (ii) future treatment not recommended by a doctor (hereafter referred as “self-initiated treatment”), and (iii) future treatment recommended by a doctor (“doctor-recommended treatment”). If they reported both self-initiated treatment and doctor-recommended treatment, we categorized them as receiving doctor-recommended treatment. For these models, the main outcome was treatment type (self-initiated and doctor-recommended treatment both compared to no future treatment) and the predictors were the various baseline measures in HRS 2006; analyses controlled for all variables assessed.
Second, we focused on those who received any sleep treatment and examined the association between treatment for sleep disturbance and cognitive performance at the same wave across years of our study (ie, HRS 2006–2014). In order to account for clustering of observations within subjects, we employed generalized estimating equations (37) with an identity link for continuous outcomes. Specifically, the cognitive performance scores served as the outcomes and treatment type (ie, no treatment [reference], self-initiated treatment, doctor-recommended treatment) was the main predictor of interest. These analyses controlled for baseline age, gender, race, ethnicity, education, number of health conditions, BMI, CES-D, current smoking status, ever alcohol use, and alcohol use frequency and quantity, all of which we allowed to be time varying. In order to isolate the effect of treatment on cognitive performance regardless of changes over time, we also controlled for HRS interview year.
Third, to determine whether there were differences in cognitive trajectories before and after sleep treatment initiation, we defined a “treatment initiation” variable that began with a value of 0 at each wave (ie, before treatment began), and changed to a value of 1 for the first and all subsequent waves in which treatment was reported. This variable was created for any treatment, self-initiated treatment, and doctor-recommended treatment. We stratified analyses for years before and years after initiation of sleep treatment and determined differences in these trajectories via an interaction term for time and treatment period while adjusting for confounds.
Finally, because the effect of sleep treatment on cognitive performance may depend on baseline cognitive functioning, we also repeated analyses for stages 2 and 3 stratified by those with the highest and lowest cognitive performance in the normal range at baseline. Specifically, among all respondents (whether treatment received or not), we divided the sample at the median by which half of the sample had the highest and half had the lowest cognitive performance. While the HRS provides variables and weights to account for the study’s complex sampling design, all analyses do not incorporate these variables as we were focused on examining the impact of treatment among those with sleep disturbance rather than nationally representative estimates. All analyses were conducted in Stata version 15 (StataCorp, College Station, TX).
Results
Baseline Demographic and Health Characteristics
Overall, 8.9% of respondents in our analysis sample reported future self-initiated treatment, and 22.7% reported future doctor-recommended treatment. Compared to those with no future treatment, those reporting self-initiated treatment were younger (adjusted odds ratio [AOR] = 0.98, 95% confidence interval [CI] = [0.97, 1.00], p = .010), more likely to be female (AOR = 1.37, 95% CI = [1.07, 1.75], p = .011), to have more depressive symptoms (AOR = 1.07, 95% CI = [1.01, 1.13], p = .015), and were less likely to be African American (AOR = 0.58, 95% CI = [0.37, 0.89], p = .014) versus Caucasian. Compared to no future treatment, those receiving doctor-recommended treatment were younger (AOR = 0.99, 95% CI = [0.98, 1.00], p = .013), more likely to be female (AOR = 1.49, 95% CI = [1.25, 1.76], p < .001), have a greater number of health conditions (AOR = 1.15, 95% CI = [1.08, 1.22], p < .001), and more depressive symptoms (AOR = 1.10, 95% CI = [1.06, 1.14], p < .001), and were less likely to be African American (AOR = 0.51, 95% CI = [0.38, 0.69], p < .001) versus Caucasian (Table 1).
Table 1.
Comparison of Baseline Characteristics by Treatment Type, Health and Retirement Study, 2006–2014
| Characteristic | No Treatment n = 2 710 n (%) |
Self-Initiated Treatment n = 350 n (%) |
Comparison AOR (95% CI), p Value |
Doctor-Recommended Treatment n = 897 n (%) |
Comparison AOR (95% CI), p Value |
|---|---|---|---|---|---|
| Baseline age, mean (SD) | 67.3 (9.83) | 65.7 (9.14) | 0.98 (0.97, 1.00), p = .010 | 66.9 (9.55) | 0.99 (0.98, 1.00), p = .013 |
| Gender | |||||
| Male | 1 195 (44.1) | 123 (35.1) | Ref. | 301 (33.6) | Ref. |
| Female | 1 515 (55.9) | 227 (64.9) | 1.37 (1.07, 1.75), p = .011 | 596 (66.4) | 1.49 (1.25, 1.76), p < .001 |
| Race | |||||
| Caucasian | 2 305 (85.1) | 311 (88.9) | Ref. | 805 (89.7) | Ref. |
| African American | 289 (10.7) | 25 (7.1) | 0.58 (0.37, 0.89), p = .014 | 63 (7.0) | 0.51 (0.38, 0.69), p < .001 |
| Other | 116 (4.3) | 14 (4.0) | 0.67 (0.35, 1.27), p = .217 | 29 (3.2) | 0.67 (0.42, 1.06), p = .090 |
| Hispanic ethnicity | |||||
| No | 2 531 (93.4) | 323 (92.3) | Ref. | 843 (94.0) | Ref. |
| Yes | 179 (6.6) | 27 (7.7) | 1.18 (0.72, 1.91), p = .512 | 54 (6.0) | 0.85 (0.59, 1.22), p = .369 |
| Education | |||||
| <High school | 415 (15.3) | 41 (11.7) | Ref. | 162 (18.1) | Ref. |
| High school degree or GED | 1 033 (38.1) | 143 (40.9) | 1.37 (0.93, 2.04), p = .113 | 337 (37.6) | 0.82 (0.64, 1.04), p = .094 |
| Some college | 643 (23.7) | 93 (26.6) | 1.46 (0.96, 2.21), p = .077 | 214 (23.9) | 0.85 (0.65, 1.10), p = .215 |
| >College | 618 (22.8) | 73 (20.9) | 1.26 (0.81, 1.97), p = .305 | 184 (20.5) | 0.85 (0.65, 1.12), p = .254 |
| Number of health conditions, mean (SD) | 2.1 (1.39) | 1.9 (1.31) | 0.94 (0.86, 1.03), p = .184 | 2.3 (1.40) | 1.15 (1.08, 1.22), p < .001 |
| BMI, mean (SD) | 28.4 (5.97) | 28.9 (6.06) | 1.01 (0.99, 1.03), p = .379 | 28.6 (6.13) | 0.99 (0.98, 1.01), p = .467 |
| CES-D, mean (SD) | 1.7 (1.98) | 2.0 (2.30) | 1.07 (1.01, 1.13), p = .015 | 2.3 (2.25) | 1.10 (1.06, 1.14), p < .001 |
| Current smoking | |||||
| No | 2 293 (85.2) | 298 (85.6) | Ref. | 743 (84.0) | Ref. |
| Yes | 399 (14.8) | 50 (14.4) | 0.84 (0.60, 1.18), p = .318 | 142 (16.1) | 0.94 (0.75, 1.18), p = .608 |
| Ever drink | |||||
| No | 1 235 (45.6) | 162 (46.3) | Ref. | 438 (48.8) | Ref. |
| Yes | 1 475 (54.4) | 188 (53.7) | 1.03 (0.78, 1.35), p = .855 | 459 (51.2) | 0.99 (0.82, 1.20), p = .909 |
| # days drink per week, mean (SD) | 1.2 (2.16) | 1.0 (1.86) | 0.96 (0.89, 1.04), p = .334 | 1.0 (2.03) | 0.99 (0.94, 1.04), p = .705 |
| # drinks per day, mean (SD) | 0.7 (1.37) | 0.6 (1.05) | 1.00 (0.88, 1.13), p = .988 | 0.6 (1.26) | 1.00 (0.92, 1.08), p = .940 |
Notes: AOR = adjusted odds ratio; CI = confidence interval; BMI = body mass index; CES-D = Center for Epidemiologic Studies Depression Scale; GED = general equivalency diploma; SD = standard deviation. All percentages are column percentages unless otherwise specified. AORs control for baseline age, gender, race, ethnicity, education, number of health conditions, BMI, CES-D, smoking, drinking, # days drink/week, and # drinks/day. All bolded values denote statistical significance at p < .05.
Association Between Treatment and Concurrent Cognitive Performance
Among those who received treatment at any time point from 2006 to 2014, we observed no association between treatment type and concurrent cognitive performance on the total cognition score or on individual cognitive test scores, indicating that sleep treatment (both self-initiated and doctor-recommended treatment) did not correspond to differences in concurrent cognitive performance as compared to no treatment (Table 2).
Table 2.
Association Between Treatment Type and Concurrent Cognitive Performance, Health and Retirement Study, 2006–2014
| Cognitive Score | Total Cognition B (95% CI), p Value |
Immediate Word Recall B (95% CI), p Value |
Delayed Word Recall B (95% CI), p Value |
Serial 7’s B (95% CI), p Value |
Backwards Counting B (95% CI), p Value |
|---|---|---|---|---|---|
| No treatment | Ref. | Ref. | Ref. | Ref. | Ref. |
| Self-initiated treatment | 0.01 (−0.22, 0.25), p = .901 | 0.02 (−0.08, 0.12), p = .696 | 0.02 (−0.10, 0.14), p = .742 | −0.04 (−0.13, 0.05), p = .394 | 0.01 (−0.02, 0.04), p = .500 |
| Doctor-recommended treatment | −0.09 (−0.28, 0.10), p = .369 | −0.05 (−0.14, 0.03), p = .240 | −0.02 (−0.12, 0.08), p = .689 | −0.02 (−0.10, 0.05), p = .526 | 0.00 (−0.02, 0.03), p = .813 |
Notes: Beta coefficient comes from generalized estimating equations and corresponds to association between treatment type and each cognitive test averaged across all years. Results control for time, baseline age, gender, race, ethnicity, education, number of health conditions, body mass index, Center for Epidemiologic Studies Depression Scale, smoking, drinking, # days drink/week, and # drinks/day. All bolded values denote statistical significance at p < .05.
Differences in Cognitive Performance Trajectories Before and After Initiation of Sleep Treatment
Among those who subsequently received any type of treatment, the total cognition score decreased on average −0.25 points per year (95% CI = [−0.28, −0.23], p < .001) between 2006 and 2014. During periods before sleep treatment initiation, declines in cognition were similar to the overall trend (B = −0.26, 95% CI = [−0.32, −0.20], p < .001); however, during periods after sleep treatment, the trend was less pronounced (B = −0.20, 95% CI = [−0.25, −0.15], p < .001). This change in trajectories before and after sleep treatment initiation was statistically significant (interaction p value = .031), indicating that cognitive decline slowed after initiating sleep treatment. When examining individual cognitive tests (eg, immediate and delayed word recall, etc.), we saw similar patterns, although only scores on Serial 7’s showed statistically significant differences in trends before and after sleep treatment was initiated. For the most part, similar results were seen when focusing on either self-initiated or doctor-recommended treatment (Table 3).
Table 3.
Trajectories of Cognitive Performance Before and After Sleep Treatment, Health and Retirement Study, 2006–2014
| Cognitive Score | Before Sleep Treatment B (95% CI), p Value |
After Sleep Treatment B (95% CI), p Value |
Interaction p Value |
|---|---|---|---|
| Overall treatment | |||
| Total cognition | −0.26 (−0.32, −0.20), p < .001 | −0.20 (−0.25, −0.15), p < .001 | p = .031 |
| Immediate word recall | −0.10 (−0.13, −0.07), p < .001 | −0.08 (−0.10, −0.06), p < .001 | p = .197 |
| Delayed word recall | −0.10 (−0.13, −0.07), p < .001 | −0.08 (−0.11, −0.06), p < .001 | p = .265 |
| Serial 7’s | −0.05 (−0.07, −0.02), p < .001 | −0.03 (−0.05, −0.01), p = .002 | p = .036 |
| Backwards counting | −0.01 (−0.02, 0.00), p = .016 | 0.00 (−0.01, 0.00), p = .265 | p = .198 |
| Self-initiated treatment | |||
| Total cognition | −0.28 (−0.32, −0.25), p < .001 | −0.17 (−0.24, −0.10), p < .001 | p = .003 |
| Immediate word recall | −0.11 (−0.12, −0.09), p < .001 | −0.07 (−0.10, −0.04), p < .001 | p = .034 |
| Delayed word recall | −0.11 (−0.12, −0.09), p < .001 | −0.08 (−0.12, −0.05), p < .001 | p = .328 |
| Serial 7’s | −0.06 (−0.07, −0.04), p < .001 | −0.01 (−0.04, 0.01), p = .351 | p = .003 |
| Backwards counting | −0.01 (−0.01, −0.01), p < .001 | 0.00 (−0.01, 0.01), p = .472 | p = .012 |
| Doctor-recommended treatment | |||
| Total cognition | −0.25 (−0.28, −0.21), p < .001 | −0.20 (−0.26, −0.14), p < .001 | p = .028 |
| Immediate word recall | −0.10 (−0.11, −0.08), p < .001 | −0.08 (−0.11, −0.05), p < .001 | p = .237 |
| Delayed word recall | −0.11 (−0.13, −0.09), p < .001 | −0.07 (−0.10, −0.04), p < .001 | p = .010 |
| Serial 7’s | −0.04 (−0.05, −0.02), p < .001 | −0.03 (−0.06, −0.01), p = .004 | p = .252 |
| Backwards counting | 0.00 (−0.01, 0.00), p = .205 | −0.01 (−0.02, 0.00), p = .066 | p = .469 |
Notes: Beta coefficient comes from generalized estimating equations and corresponds to change in cognitive score with 1-y change in time. Results control for baseline age, gender, race, ethnicity, education, number of health conditions, body mass index, Center for Epidemiologic Studies Depression Scale, smoking, drinking, # days drink/week, and # drinks/day. All bolded values denote statistical significance at p < .05.
Stratified Results by Baseline Cognitive Functioning
Fifty-nine percent had lower cognitive functioning (n = 2 314, total cognitive score range from 12 to 17) and the remaining 39.2% had higher cognitive functioning (n = 1 551, total cognitive score range from 18 to 27); n = 92 (2.3%) were excluded as their data for HRS 2006 were reported by proxies. Overall, among those with lower cognitive functioning and who received treatment, self-initiated treatment was associated with higher concurrent cognitive functioning for the total cognition score and immediate word recall as compared to no treatment, while we did not see differences for doctor-recommended treatment. Among those with better cognitive performance at baseline, both self-initiated treatment and doctor-recommended treatment were associated with lower concurrent cognitive functioning (Table 4). When examining trajectories of cognition before and after treatment initiation, the periods after sleep treatment for those with the lower cognitive performance at baseline were actually associated with a faster decline in cognitive performance for overall treatment and self-initiated treatment, while there were no statistically significantly different changes for doctor-recommended treatment. Among those with better cognitive performance at baseline, decline in cognitive performance slowed substantially for overall, self-initiated, as well as doctor-initiated treatment (Table 5).
Table 4.
Association Between Treatment Type and Concurrent Cognitive Performance Stratified by Baseline Cognitive Functioning, Health and Retirement Study, 2006–2014
| Cognitive Score | Total Cognition B (95% CI), p Value |
Immediate Word Recall B (95% CI), p Value |
Delayed Word Recall B (95% CI), p Value |
Serial 7’s B (95% CI), p Value |
Backwards Counting B (95% CI), p Value |
|---|---|---|---|---|---|
| Lowest cognitive functioning | |||||
| No treatment | Ref. | Ref. | Ref. | Ref. | Ref. |
| Self-initiated treatment | 0.36 (0.05, 0.66), p = .022 | 0.17 (0.03, 0.30), p = .014 | 0.11 (−0.04, 0.27), p = .161 | 0.06 (−0.07, 0.19), p = .385 | 0.01 (−0.04, 0.05), p = .770 |
| Doctor-recommended treatment | 0.22 (−0.02, 0.46), p = .078 | 0.05 (−0.05, 0.16), p = .319 | 0.12 (−0.01, 0.24), p = .071 | 0.04 (−0.06, 0.15), p = .421 | 0.00 (−0.04, 0.04), p = .920 |
| Highest cognitive functioning | |||||
| No treatment | Ref. | Ref. | Ref. | Ref. | Ref. |
| Self-initiated treatment | −0.39 (−0.74, −0.05), p = .026 | −0.15 (−0.31, 0.00), p = .055 | −0.10 (−0.28, 0.08), p = .282 | −0.15 (−0.27, −0.04), p = .008 | 0.02 (−0.02, 0.05), p = .395 |
| Doctor-recommended treatment | −0.57 (−0.86, −0.28), p < .001 | −0.23 (−0.36, −0.10), p = .001 | −0.23 (−0.38, −0.08), p = .003 | −0.12 (−0.22, −0.02), p = .016 | 0.01 (−0.02, 0.04), p = .478 |
Notes: Beta coefficient comes from generalized estimating equations and corresponds to association between treatment type and each cognitive test averaged across all years. Results control for time, baseline age, gender, race, ethnicity, education, number of health conditions, body mass index, Center for Epidemiologic Studies Depression Scale, smoking, drinking, # days drink/week, and # drinks/day. All bolded values denote statistical significance at p < .05.
Table 5.
Trajectories of Cognitive Performance Before and After Sleep Treatment Stratified by Baseline Cognitive Functioning, Health and Retirement Study, 2006–2014
| Cognitive Score | Before Sleep Treatment B (95% CI), p Value |
After Sleep Treatment B (95% CI), p Value |
Interaction p Value |
|---|---|---|---|
| Lowest cognitive functioning | |||
| Overall treatment | |||
| Total cognition | −0.03 (−0.10, 0.04), p = .448 | −0.25 (−0.32, −0.19), p < .001 | p < .001 |
| Immediate word recall | −0.01 (−0.04, 0.03), p = .627 | −0.10 (−0.13, −0.08), p < .001 | p < .001 |
| Delayed word recall | 0.01 (−0.03, 0.05), p = .605 | −0.10 (−0.13, −0.07), p < .001 | p < .001 |
| Serial 7’s | −0.03 (−0.06, 0.01), p = .193 | −0.04 (−0.07, −0.02), p = .001 | p = .949 |
| Backwards counting | 0.00 (−0.01, 0.01), p = .702 | 0.00 (−0.01, 0.01), p = .741 | p = .906 |
| Self-initiated treatment | |||
| Total cognition | −0.18 (−0.23, −0.14), p < .001 | −0.26 (−0.35, −0.17), p < .001 | p = .071 |
| Immediate word recall | −0.07 (−0.08, −0.05), p < .001 | −0.11 (−0.15, −0.07), p < .001 | p = .036 |
| Delayed word recall | −0.06 (−0.08, −0.04), p < .001 | −0.12 (−0.17, −0.07), p < .001 | p = .006 |
| Serial 7’s | −0.05 (−0.07, −0.03), p < .001 | −0.03 (−0.07, 0.00), p = .081 | p = .397 |
| Backwards counting | −0.01 (−0.02, 0.00), p = .003 | 0.00 (−0.01, 0.01), p = .641 | p = .124 |
| Doctor-recommended treatment | |||
| Total cognition | −0.15 (−0.20, −0.11), p < .001 | −0.23 (−0.31, −0.15), p < .001 | p = .336 |
| Immediate word recall | −0.06 (−0.08, −0.04), p < .001 | −0.09 (−0.13, −0.06), p < .001 | p = .113 |
| Delayed word recall | −0.07 (−0.10, −0.05), p < .001 | −0.08 (−0.12, −0.04), p < .001 | p = .785 |
| Serial 7’s | −0.03 (−0.05, 0.00), p = .016 | −0.04 (−0.08, −0.01), p = .008 | p = .980 |
| Backwards counting | 0.00 (−0.01, 0.01), p = .801 | 0.00 (−0.02, 0.01), p = .469 | p = .743 |
| Highest cognitive functioning | |||
| Overall treatment | |||
| Total cognition | −0.58 (−0.66, −0.50), p < .001 | −0.12 (−0.19, −0.05), p = .001 | p < .001 |
| Immediate word recall | −0.24 (−0.28, −0.20), p < .001 | −0.05 (−0.08, −0.02), p = .003 | p < .001 |
| Delayed word recall | −0.25 (−0.30, −0.21), p < .001 | −0.06 (−0.09, −0.02), p = .004 | p < .001 |
| Serial 7’s | −0.08 (−0.10, −0.05), p < .001 | −0.01 (−0.03, 0.02), p = .470 | p < .001 |
| Backwards counting | −0.02 (−0.03, −0.01), p < .001 | −0.01 (−0.01, 0.00), p = .097 | p = .030 |
| Self-initiated treatment | |||
| Total cognition | −0.42 (−0.47, −0.37), p < .001 | −0.08 (−0.18, 0.03), p = .151 | p < .001 |
| Immediate word recall | −0.17 (−0.19, −0.15), p < .001 | −0.03 (−0.09, 0.02), p = .182 | p < .001 |
| Delayed word recall | −0.17 (−0.20, −0.15), p < .001 | −0.05 (−0.11, 0.00), p = .068 | p < .001 |
| Serial 7’s | −0.07 (−0.08, −0.05), p < .001 | 0.01 (−0.03, 0.05), p = .608 | p < .001 |
| Backwards counting | −0.01 (−0.02, −0.01), p < .001 | 0.00 (−0.01, 0.01), p = .510 | p = .018 |
| Doctor-recommended treatment | |||
| Total cognition | −0.38 (−0.43, −0.33), p < .001 | −0.14 (−0.23, −0.05), p = .003 | p < .001 |
| Immediate word recall | −0.15 (−0.17, −0.13), p < .001 | −0.06 (−0.10, −0.02), p = .005 | p < .001 |
| Delayed word recall | −0.17 (−0.20, −0.14), p < .001 | −0.05 (−0.09, 0.00), p = .057 | p < .001 |
| Serial 7’s | −0.05 (−0.07, −0.04), p < .001 | −0.02 (−0.05, 0.01), p = .214 | p = .019 |
| Backwards counting | −0.01 (−0.01, 0.00), p = .050 | −0.01 (−0.02, 0.00), p = .017 | p = .326 |
Notes: Beta coefficient comes from generalized estimating equations and corresponds to change in cognitive score with 1-y change in time. Results control for baseline age, gender, race, ethnicity, education, number of health conditions, body mass index, Center for Epidemiologic Studies Depression Scale, smoking, drinking, # days drink/week, and # drinks/day. All bolded values denote statistical significance at p < .05.
Discussion
Our findings are important because they add to the existing literature regarding the influential role of sleep health on cognition (5–13,16,38,39). We sought to identify the role of sleep treatment on cognitive performance among older adults with sleep disturbances. While we did not find an association between sleep treatment and cognitive performance at the same time, we did find that, longitudinally, initiating treatment was associated with a mitigation in cognitive decline over an 8-year period, suggesting that short-term effects of treatment may be more subtle, whereas long-term effects may become more evident well after treatment. Our findings are clinically significant because we found an impact of treatment on cognition during an 8-year period, and given that dementia may take years if not decades to develop, it is possible our observed effect would be stronger when observed over longer time periods (eg, a decade or more). This mitigation we observed was seen for both self-initiated and doctor-recommended treatment and was particularly strong among those with the higher cognitive functioning at baseline as compared to those with functioning on the lower end. These findings are particularly notable, given the large burden of AD on modern society which has become more profound with the aging of the population (40).
Accelerated cognitive decline and development of dementia affects the lives of not only patients but also their families and caregivers. Thus, there have been efforts by the medical community to find a cure or therapy to prevent or effectively treat the disease. The search for pharmacological options, however, has not been fruitful to date. A number of promising drugs in development (eg, solanezumab) were deemed ineffective in larger clinical trials (41), and even medications which came to market and were approved for use in AD for the most part may be less effective. Donepezil, for example, has been approved for use in AD, although the observed benefits can be quite modest (42). Thus, the optimal therapy for AD and patients at risk of AD remains unclear.
As efforts to find pharmacotherapies for AD have yet to be successful, investigators have searched for “actionable” risk factors that may serve as targets for behavioral intervention efforts for AD, including physical activity, diet, and sleep. These lifestyle factors may account for 40% of dementia cases worldwide (43). While there is uncertainty about whether poor sleep is causal for developing AD, our new findings suggest that sleep issues may be important considerations when optimizing therapy for these patients to improve and/or preserve their cognitive health.
In the context of our main findings, it should be noted that only a third of our sample reported receiving sleep treatment. This result suggests an unmet need for therapies to improve the care of those with sleep disturbances. Findings indicated that individuals who received treatment were on average more likely to be female, Caucasian, report more depressive symptoms, have a higher BMI, and have greater number of health conditions. More research is needed to determine how best to offer these treatments so as to prevent health disparities. Efforts to make treatments more available may also further help in prevention of cognitive decline.
While we did not see associations between treatment type and concurrent cognitive performance, we saw a mitigation of cognitive decline longitudinally following initiating sleep treatment. This finding suggests the benefits of sleep treatment on cognition may accumulate long after treatment began, while the immediate effects are minimal. This warrants further investigation, specifically examining the timing of treatment delivery (whether treatments are used consistently, intermittently, or long-term), and type of treatment or combination of treatments used.
Paradoxically, the protective effect of sleep treatment was seen exclusively in those with higher cognitive functioning at baseline, whereas those with lower cognitive functioning actually showed precipitous declines in performance following treatment. Those with poorer cognition were likely those at greater risk for subsequent impairment, and cognitive declines after sleep treatment may be explained simply by worsening cognition due to a neurodegenerative disease that was also causing sleep disturbance. It is also possible that sleep treatment was initiated as part of a holistic strategy for memory-related problems introduced when cognitive impairment was significant—it should be noted that sleep problems are commonly seen in those with dementia, including AD (44).
A number of sleep treatments are available. Self-initiated treatments may include use of melatonin, OTC sleep aids, and smartphone applications including mindfulness and sleep tracking applications. While our findings suggest that self-initiated treatment could have the potential to slow cognitive decline, more research is needed to assess the efficacy of specific self-initiated treatments, as well as how use patterns of these treatments over time might also impact cognitive performance trajectories and cognition more broadly. It may also be important to assess characteristics of these treatments that are associated with sustained use over time and that impact on cognitive health. There are also a number of therapies overseen by a doctor that are available depending upon the underlying cause. For example, CBT-I is considered the first-line treatment for many patients with insomnia (45), particularly given the concerns about potential side-effects of pharmacotherapy. Nasal CPAP and mandibular advancement devices are efficacious for patients with OSA (46), although adherence can be a challenge particularly in older patients (47) and those with impaired cognition (48). The HRS question we relied upon to characterize sleep treatment was quite vague and there is considerable uncertainty about the type of treatment received or whether multiple therapies were used (eg, medications and CBT-I). Given the heterogeneity of treatments for various sleep conditions, we support further research into the optimal treatment strategy to preserve cognitive health in later life.
Our study had a number of strengths including a large population-based sample, with extensive follow-up, in a well-characterized population and with a detailed health profile including various domains of cognitive function. However, we acknowledge a number of limitations. First, we do not have granular data regarding sleep treatments as we relied on self-report and could not fully characterize what treatments were actually received. Moreover, the assessment period for receipt of sleep treatment was over the past 2 weeks, and it is possible treatments were received in the prior weeks. Thus, more research is needed to identify the sleep treatments and timing of these treatments that have greatest utility in modulating cognitive decline. Second, our study used observational data and we did not conduct a randomized controlled trial limiting our ability to assess causality. Nonetheless we view the findings as important since such data are critical to optimize the design of a robust randomized trial with adequate follow-up time. We are also supportive of “Big Data” approaches using electronic medical records and other large data sets to address these important questions. Third, although this study used carefully selected cognitive assessments, a number of cognitive tests of importance may also be important to characterize fully the cognitive impact of sleep treatment. As such, we are supportive of more extensive evaluations which may be possible using other data sets or via prospective data collection.
In summary, our study highlights the important role of sleep in cognitive outcomes and suggests that intervention for poor sleep, even in later adulthood, could have the potential to at a very minimum slow cognitive decline. While this study presents promising results, more research is needed to determine which sleep treatments are most effective in mitigating cognitive decline, and fuller assessment of the risks and benefits of various treatment options for other outcomes. These efforts could not only improve sleep health in these individuals but also could have the potential to contribute to AD prevention efforts.
Funding
This study received funding from the National Institutes of Health (K01AG061239 to C.N.K. and R01AG049810 to M.W.B.).
Conflict of Interest
S.A-I. has previously served as a consultant for Eisai, Merck, Biogen, and Pear. A.P.S. has received payment for serving as a consultant to Merck and from Springer Nature Switzerland AG for Guest Editing Special Issues of Current Sleep Medicine Reports. C.N.K., M.W.B., W.K.T., and A.M. declare no conflict of interest.
Author Contributions
C.N.K. conceived the study, analyzed the data, and wrote the manuscript. A.M. wrote the manuscript. M.W.B., W.K.T, A.P.S., and S.A-I. provided feedback on manuscript and analyses.
References
- 1. Foley DJ, Monjan AA, Brown SL, Simonsick EM, Wallace RB, Blazer DG. Sleep complaints among elderly persons: an epidemiologic study of three communities. Sleep. 1995;18(6):425–432. doi:10.1093/sleep/18.6.425 [DOI] [PubMed] [Google Scholar]
- 2. Jaussent I, Dauvilliers Y, Ancelin ML, et al. Insomnia symptoms in older adults: associated factors and gender differences. Am J Geriatr Psychiatry. 2011;19(1):88–97. doi:10.1097/JGP.0b013e3181e049b6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Olson LG. A community survey of insomnia in Newcastle. Aust N Z J Public Health. 1996;20(6):655–657. doi:10.1111/j.1467-842x.1996.tb01083.x [DOI] [PubMed] [Google Scholar]
- 4. Yaffe K, Laffan AM, Harrison SL, et al. Sleep-disordered breathing, hypoxia, and risk of mild cognitive impairment and dementia in older women. J Am Med Assoc. 2011;306(6):613–619. doi:10.1001/jama.2011.1115 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Spira AP, Chen-Edinboro LP, Wu MN, Yaffe K. Impact of sleep on the risk of cognitive decline and dementia. Curr Opin Psychiatry. 2014;27(6):478–483. doi:10.1097/YCO.0000000000000106 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Blackwell T, Yaffe K, Laffan A, et al. Associations of objectively and subjectively measured sleep quality with subsequent cognitive decline in older community-dwelling men: the MrOS Sleep Study. Sleep. 2014;37(4):655–663. doi:10.5665/sleep.3562 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Osorio RS, Gumb T, Pirraglia E, et al. Sleep-disordered breathing advances cognitive decline in the elderly. Neurology. 2015;84(19):1964–1971. doi:10.1212/WNL.0000000000001566 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Blackwell T, Yaffe K, Laffan A, et al. Associations between sleep-disordered breathing, nocturnal hypoxemia, and subsequent cognitive decline in older community-dwelling men: the Osteoporotic Fractures in Men Sleep Study. J Am Geriatr Soc. 2015;63(3):453–461. doi:10.1111/jgs.13321 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Xie L, Kang H, Xu Q, et al. Sleep drives metabolite clearance from the adult brain. Science. 2013;342(6156):373–377. doi:10.1126/science.1241224 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Shokri-Kojori E, Wang GJ, Wiers CE, et al. β-Amyloid accumulation in the human brain after one night of sleep deprivation. Proc Natl Acad Sci USA. 2018;115(17):4483–4488. doi:10.1073/pnas.1721694115 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Spira AP, Gamaldo AA, An Y, et al. Self-reported sleep and β-amyloid deposition in community-dwelling older adults. JAMA Neurol. 2013;70(12):1537–1543. doi:10.1001/jamaneurol.2013.4258 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Yamauchi M, Nakano H, Maekawa J, et al. Oxidative stress in obstructive sleep apnea. Chest. 2005;127(5):1674–1679. doi:10.1378/chest.127.5.1674 [DOI] [PubMed] [Google Scholar]
- 13. Butterfield DA, Halliwell B. Oxidative stress, dysfunctional glucose metabolism and Alzheimer disease. Nat Rev Neurosci. 2019;20(3):148–160. doi:10.1038/s41583-019-0132-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Zhang X, Zhou K, Wang R, et al. Hypoxia-inducible factor 1alpha (HIF-1alpha)-mediated hypoxia increases BACE1 expression and beta-amyloid generation. J Biol Chem. 2007;282(15):10873–10880. doi:10.1074/jbc.M608856200 [DOI] [PubMed] [Google Scholar]
- 15. Sun X, He G, Qing H, et al. Hypoxia facilitates Alzheimer’s disease pathogenesis by up-regulating BACE1 gene expression. Proc Natl Acad Sci USA. 2006;103(49):18727–18732. doi:10.1073/pnas.0606298103 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Uyrum E, Balbay O, Annakkaya AN, Gulec Balbay E, Silan F, Arbak P. The relationship between obstructive sleep apnea syndrome and apolipoprotein E genetic variants. Respiration. 2015;89(3):195–200. doi:10.1159/000369560 [DOI] [PubMed] [Google Scholar]
- 17. Ancoli-Israel S, Ayalon L. Diagnosis and treatment of sleep disorders in older adults. Am J Geriatr Psychiatry. 2006;14(2):95–103. doi:10.1097/01.JGP.0000196627.12010.d1 [DOI] [PubMed] [Google Scholar]
- 18. Koffel E, Bramoweth AD, Ulmer CS. Increasing access to and utilization of cognitive behavioral therapy for insomnia (CBT-I): a narrative review. J Gen Intern Med. 2018;33(6):955–962. doi:10.1007/s11606-018-4390-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Chong Y, Fryer CD, Gu Q. Prescription sleep aid use among adults: United States, 2005–2010. NCHS Data Brief. 2013;127:1–8. [PubMed] [Google Scholar]
- 20. Albert SM, Roth T, Toscani M, Vitiello MV, Zee P. Sleep health and appropriate use of OTC sleep aids in older adults—recommendations of a Gerontological Society of America Workgroup. Gerontologist. 2017;57(2):163–170. doi:10.1093/geront/gnv139 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Kaufmann CN, Spira AP, Alexander GC, Rutkow L, Mojtabai R. Trends in prescribing of sedative-hypnotic medications in the USA: 1993–2010. Pharmacoepidemiol Drug Saf. 2016;25(6):637–645. doi:10.1002/pds.3951 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Dunietz GL, Chervin RD, Burke JF, Braley TJ. Obstructive sleep apnea treatment disparities among older adults with neurological disorders. Sleep Health. 2020;6(4):534–540. doi:10.1016/j.sleh.2020.01.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Diekelmann S, Born J. The memory function of sleep. Nat Rev Neurosci. 2010;11(2):114–126. doi:10.1038/nrn2762 [DOI] [PubMed] [Google Scholar]
- 24. Ancoli-Israel S, Palmer BW, Cooke JR, et al. Cognitive effects of treating obstructive sleep apnea in Alzheimer’s disease: a randomized controlled study. J Am Geriatr Soc. 2008;56(11):2076–2081. doi:10.1111/j.1532-5415.2008.01934.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Lader M. Benzodiazepines revisited—will we ever learn? Addiction. 2011;106(12):2086–2109. doi:10.1111/j.1360-0443.2011.03563.x [DOI] [PubMed] [Google Scholar]
- 26. Lagnaoui R, Bégaud B, Moore N, et al. Benzodiazepine use and risk of dementia: a nested case-control study. J Clin Epidemiol. 2002;55(3):314–318. doi:10.1016/s0895-4356(01)00453-x [DOI] [PubMed] [Google Scholar]
- 27. Islam MM, Iqbal U, Walther B, et al. Benzodiazepine use and risk of dementia in the elderly population: a systematic review and meta-analysis. Neuroepidemiology. 2016;47(3–4):181–191. doi:10.1159/000454881 [DOI] [PubMed] [Google Scholar]
- 28. Gallacher J, Elwood P, Pickering J, Bayer A, Fish M, Ben-Shlomo Y. Benzodiazepine use and risk of dementia: evidence from the Caerphilly Prospective Study (CaPS). J Epidemiol Community Health. 2012;66(10):869–873. doi:10.1136/jech-2011-200314 [DOI] [PubMed] [Google Scholar]
- 29. Billioti de Gage S, Moride Y, Ducruet T, et al. Benzodiazepine use and risk of Alzheimer’s disease: case-control study. Br Med J. 2014;349:g5205. doi:10.1136/bmj.g5205 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Tapiainen V, Taipale H, Tanskanen A, Tiihonen J, Hartikainen S, Tolppanen AM. The risk of Alzheimer’s disease associated with benzodiazepines and related drugs: a nested case-control study. Acta Psychiatr Scand. 2018;138(2):91–100. doi:10.1111/acps.12909 [DOI] [PubMed] [Google Scholar]
- 31. Richardson K, Mattishent K, Loke YK, et al. History of benzodiazepine prescriptions and risk of dementia: possible bias due to prevalent users and covariate measurement timing in a nested case-control study. Am J Epidemiol. 2019;188(7):1228–1236. doi:10.1093/aje/kwz073 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Gray SL, Dublin S, Yu O, et al. Benzodiazepine use and risk of incident dementia or cognitive decline: prospective population based study. Br Med J. 2016;352:i90. doi:10.1136/bmj.i90 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Crimmins EM, Kim JK, Langa KM, Weir DR. Assessment of cognition using surveys and neuropsychological assessment: the Health and Retirement Study and the Aging, Demographics, and Memory Study. J Gerontol B Psychol Sci Soc Sci. 2011;66(suppl. 1):i162–171. doi:10.1093/geronb/gbr048 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Measure. 1977;1(3):385–401. doi:10.1177/014662167700100306 [Google Scholar]
- 35. Stein MD, Friedmann PD. Disturbed sleep and its relationship to alcohol use. Subst Abus. 2005;26(1):1–13. doi:10.1300/j465v26n01_01 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Mets MA, Volkerts ER, Olivier B, Verster JC. Effect of hypnotic drugs on body balance and standing steadiness. Sleep Med Rev. 2010;14(4):259–267. doi:10.1016/j.smrv.2009.10.008 [DOI] [PubMed] [Google Scholar]
- 37. Liang K-Y, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika. 1986;73(1):13–22. doi:10.1093/biomet/73.1.13 [Google Scholar]
- 38. Gozal D, Kheirandish-Gozal L. Cardiovascular morbidity in obstructive sleep apnea: oxidative stress, inflammation, and much more. Am J Respir Crit Care Med. 2008;177(4):369–375. doi:10.1164/rccm.200608-1190PP [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Javaheri S, Barbe F, Campos-Rodriguez F, et al. Sleep apnea: types, mechanisms, and clinical cardiovascular consequences. J Am Coll Cardiol. 2017;69(7):841–858. doi:10.1016/j.jacc.2016.11.069 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. GBD 2016 Dementia Collaborators. Global, regional, and national burden of Alzheimer’s disease and other dementias, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol. 2019;18(1):88–106. doi:10.1016/S1474-4422(18)30403-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Oxford AE, Stewart ES, Rohn TT. Clinical trials in Alzheimer’s disease: a hurdle in the path of remedy. Int J Alzheimers Dis. 2020;2020:5380346. doi:10.1155/2020/5380346 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Birks JS, Harvey RJ. Donepezil for dementia due to Alzheimer’s disease. Cochrane Database Syst Rev. 2018;6:CD001190. doi:10.1002/14651858.CD001190.pub3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Livingston G, Huntley J, Sommerlad A, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet. 2020;396(10248):413–446. doi:10.1016/S0140-6736(20)30367-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Moran M, Lynch CA, Walsh C, Coen R, Coakley D, Lawlor BA. Sleep disturbance in mild to moderate Alzheimer’s disease. Sleep Med. 2005;6(4):347–352. doi:10.1016/j.sleep.2004.12.005 [DOI] [PubMed] [Google Scholar]
- 45. Qaseem A, Kansagara D, Forciea MA, Cooke M, Denberg TD, Clinical Guidelines Committee of the American College of Physicians. Management of chronic insomnia disorder in adults: a clinical practice guideline from the American College of Physicians. Ann Intern Med. 2016;165(2):125–133. doi:10.7326/M15-2175 [DOI] [PubMed] [Google Scholar]
- 46. Patil SP, Ayappa IA, Caples SM, Kimoff RJ, Patel SR, Harrod CG. Treatment of adult obstructive sleep apnea with positive airway pressure: an American Academy of Sleep Medicine systematic review, meta-analysis, and GRADE assessment. J Clin Sleep Med. 2019;15(2):301–334. doi:10.5664/jcsm.7638 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Martinez-Garcia MA, Valero-Sanchez I, Reyes-Nunez N, et al. Continuous positive airway pressure adherence declines with age in elderly obstructive sleep apnoea patients. ERJ Open Res. 2019;5(1). doi:10.1183/23120541.00178-2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Sawyer AM, Gooneratne NS, Marcus CL, Ofer D, Richards KC, Weaver TE. A systematic review of CPAP adherence across age groups: clinical and empiric insights for developing CPAP adherence interventions. Sleep Med Rev. 2011;15(6):343–356. doi:10.1016/j.smrv.2011.01.003 [DOI] [PMC free article] [PubMed] [Google Scholar]