A Randomized-Controlled Trial Targeting Cognition in Early Alzheimer’s Disease by Improving Sleep with Trazodone (REST)

Abstract

Alzheimer’s disease (AD) is a leading cause of mortality and morbidity among aging populations worldwide. Despite arduous research efforts, treatment options for this devastating neurodegenerative disease are limited. Sleep disturbances, through their link to changes in neural excitability and impaired clearance of interstitial abnormal protein aggregates, are a key risk factor for the development of AD. Research also suggests that the neuroprotective effects of sleep are particularly active during slow wave sleep. Given the strong link between sleep disturbance and AD, targeting sleep in the prodromal stages of AD, such as in mild cognitive impairment (MCI), represents a promising avenue for slowing the onset of AD-related cognitive decline. In efforts to improve sleep in older individuals, several pharmacologic approaches have been employed, but many pose safety risks, concern for worsening cognitive function, and fail to effectively target slow wave sleep. Trazodone, a safe and widely used drug in the older adult population, has shown promise in inducing slow wave sleep in older adults, but requires more rigorous research to understand its effects on sleep and cognition in the prodromal stages of AD. In this review, we present the rationale and study design for our randomized, double-bind, placebo-controlled, crossover trial (NCT05282550) investigating the effects of trazodone on sleep and cognition in 100 older adults with amnestic MCI and sleep complaints.

Keywords

Alzheimer’s disease circadian rhythms hippocampus mild cognitive impairment sleep trazodone

INTRODUCTION

Alzheimer’s disease (AD) is a leading cause of mortality and morbidity among aging populations worldwide [1]. Sleep disturbances are a key risk factor for the development of AD and commonly present along with other neuropsychiatric symptoms in its prodromal stages, such as in mild cognitive impairment (MCI) [2]. Moreover, there is evidence of a bidirectional relationship between sleep disturbance and AD, wherein impaired sleep may initiate brain changes that accelerate AD pathogenesis, which in turn further disrupt normal sleep patterns [3, 4]. We are presenting the rationale and study design for the use of a commonly prescribed sleep medication (trazodone) to improve sleep and cognition in older persons with amnestic MCI (aMCI) and sleep complaints.

Currently, treatment options for AD are limited, and targeting sleep represents a promising avenue for slowing the onset of AD-related cognitive decline. Healthy sleep has been shown to be associated with hippocampal-dependent memory consolidation [3]. Conversely, sleep disturbances are linked to deleterious effects on vigilance, attention, and memory, mediated in part by disruptions in synaptic pruning and homeostasis [3, 5]. Two primary mechanisms— changes in neural excitability and clearance of interstitial amyloid-β (Aβ), tau and other metabolites— may link disturbed sleep to pathogenic pathways in AD. First, wakefulness increases neural excitability in insects, rodents, and humans that is renormalized during sleep, particularly slow wave sleep (SWS). We have observed increased hippocampal activation in individuals with aMCI during a memory task involving hippocampal encoding [6, 7]. Because poor sleep may be associated with increased hippocampal activation, improving sleep is a potential target for positively affecting cognition and disease progression in AD. Second, in animal model studies, Aβ and tau are cleared from brain interstitial fluid and cerebrospinal fluid (CSF) during SWS and accumulate when SWS is reduced, which has been proposed to occur via sleep-related glymphatic mechanisms. Analogous mechanisms may be involved in humans that coincide with sleep-related changes in neural activity. Thus, improving sleep in aMCI may improve cognition and mitigate AD pathogenic mechanisms.

The glymphatic system has been well-characterized in murine models. Flow through this system is composed of para-arterial CSF that moves intracellularly through a trans-astrocytic pathway to mix with interstitial fluid before being cleared along para-venous pathways [8]. Evidence from animal models suggests that heightened slow wave activity (SWA), power in the delta frequency characteristic of SWS, clears Aβ and tau proteins from brain interstitial fluid and CSF through sleep-related glymphatic mechanisms [9 –11]. For example, a mouse model of AD showed evidence of glymphatic impairments leading up to the significant accumulation of Aβ, suggesting that impaired glymphatic clearance could be a key mechanism underlying development of AD-related pathology [12].

The existence of analogous hemodynamic mechanisms in humans are speculated to exist but require further study. In cognitively normal older adults, age-related medial prefrontal cortex volume decreases were shown to be associated with reduced SWA, which mediated reduced overnight sleep-dependent memory retention, increased hippocampal activation, and reduced functional connectivity between the hippocampus and prefrontal cortex during an episodic memory task [13]. These data support an association between reduced SWA, increased hippocampal activation, and long-term memory impairment. Thus, increasing SWA may decrease hippocampal hyperactivity, which can serve to improve memory consolidation in cognitively normal individuals and in those with aMCI.

In efforts to improve sleep in older individuals, several pharmacologic approaches have been employed, including melatonin, second-generation antipsychotics, and sedating antidepressants. However, many of these medications have high safety risks or pose concern for worsening cognitive function [14]. Suvorexant, a recently approved dual orexin receptor antagonist with a promising safety profile, was shown to increase total sleep time, decrease wake after sleep onset, and reduce tau phosphorylation and Aβ concentrations in the central nervous system in cognitively normal human adults [15]. However, suvorexant did not impact time spent in SWS, emphasizing the need for a safe, pharmacological agent that improves SWS in older adults with aMCI/AD.

Trazodone (TRZ) is a generic antidepressant widely used off-label to treat sleep disturbance, particularly enhancing SWS. Reduced hippocampal excitability is one potential mechanism for this effect. While TRZ has been demonstrated to improve sleep in AD and potentially mitigate the risk of developing aMCI, its effect on sleep has not been rigorously studied in aMCI. TRZ, originally approved by the FDA as a treatment for major depressive disorder, is a widely used off-label sleep aid among patients with AD [16 –18]. Low doses of TRZ have been shown to be sufficient to induce its sleep promoting effects, and TRZ is distinct among sleep medications in its ability to induce SWS [18 –21]. Due to its favorable side effect profile, lack of anticholinergic activity, and low potential for dependence, TRZ is a promising therapy for older adults, who are at high risk of adverse drug reactions [22].

Findings from several studies provide support for the efficacy of TRZ in treating insomnia in older adults, with evidence supporting positive impacts on AD pathogenesis. A large retrospective cohort study followed 6,000 participants over 3–4 years and found that baseline sleep complaints were associated with an increased risk of incident MCI, but that this association was not present in TRZ users [18]. In a small randomized-controlled trial, Camargos et al. (2014) found that TRZ use improved total sleep time in patients with AD compared to placebo [23]. Beyond its effects on sleep, TRZ has also been shown to slow cognitive decline in older adults. For example, La et al. (2019) reported that regular TRZ users had a 2.6-fold decreased rate of decline on Mini-Mental Status Exam (MMSE) performance compared to non-TRZ users, which was associated with subjective improvement of sleep [24]. Conversely, a UK cohort study found that TRZ prescription use in 406 individuals with MCI did not significantly improve cognition [25]. However, only approximately 25% of subjects in this study reported sleep complaints at baseline, suggesting that the beneficial effects of TRZ on cognition may be particularly suited to those with sleep difficulties.

Together, these results suggest that TRZ, through its induction of SWS, may improve sleep in older adults and potentially affect AD pathophysiology. However, more rigorous research is needed to understand the effects of TRZ on sleep and cognitive functioning in individuals in the prodromal stages of AD, such as those with aMCI. Thus, we present the methods for a rigorous double-blind, randomized-controlled crossover trial of TRZ in 100 subjects with prodromal AD/aMCI and sleep complaints. To investigate the effects of TRZ on sleep, hippocampal-dependent memory and hippocampal excitability, we are conducting a rigorous randomized double-blind, placebo-controlled, crossover trial in 100 subjects with aMCI and sleep complaints.

Our primary aim is to assess the effect of TRZ on sleep parameters as measured by polysomnography (PSG), actigraphy, and self-report. We hypothesize that TRZ will improve total sleep time and proportion of time in SWS. To measure associated changes in cognition, we will administer a cognitive battery that assesses domains of cognition affected by aMCI and poor sleep [24 , 27]. Additionally, given findings that aMCI patients show performance deficits on dentate gyrus (DG)/cornu ammonis-3 (CA3)-mediated memory tasks in functional magnetic resonance (fMRI) analyses, participants will also complete three fMRI sessions that will assess hippocampal activation during performance a behavioral task [6 , 28]. With these data, our secondary aim is to examine the effect of TRZ on hippocampal hyperactivity and performance on the behavioral pattern separation objects task, which is designed to test hippocampal-dependent episodic memory function. We will also assess a broader range of cognitive domains and neuropsychiatric symptoms as tertiary outcomes. Finally, we will perform exploratory analyses to assess the impact of TRZ on changes in plasma biomarkers relevant to AD-related pathology and whether the relationships between TRZ and biomarker changes are mediated by increases in SWS [29].

METHODS

Study design

Our study, RCT targeting cognition in Early Alzheimer’s Disease by Improving Sleep with Trazodone (REST, NCT05282550), was approved by the Johns Hopkins Medicine Institutional Review Board (IRB). REST is a randomized, placebo-controlled, double-blind crossover study of TRZ in 100 participants with aMCI who report sleep complaints. The study is being conducted at the Johns Hopkins University of Medicine and at the Kennedy Krieger Institute in Baltimore, Maryland. Participants are randomized to receive either TRZ (50 mg at bedtime, provided in opaque capsules) or placebo (matching capsules containing inert filler) for 4 weeks each. Following completion of the 4-week treatment phases, both groups will undergo a 4-week intervening washout period, sufficient to allow elimination of TRZ given a half-life of 10–12 h [30, 31]. After the washout period, groups will be crossed-over such that participants who received TRZ for the first treatment phase will then receive placebo, and vice versa, for another 4 weeks. The crossover design will allow for subjects to act as their own controls and reduce the need for additional recruitment.

Prior to initiating in-person procedures, a telephone-based screening will be conducted during which the Hopkins Verbal Learning Test– Revised Version (HVLT-R) will be administered to assess potential eligibility of participants [32]. Individuals who qualify (see below) will be invited for a series of five in-person study visits (Fig. 1). Cognitive testing and neuropsychiatric assessment will be collected on all participants at the baseline visit (visit 1) and upon completion of both treatment phases (visits 3 and 5). After visit 1, eligible participants will complete an MRI session at visit 2, and again upon completion of both treatment phases (visits 3 and 5). Sleep measures including PSG, actigraphy, and self-report surveys will also be collected at all visits except visit 2. All human subjects research and procedures, including advertising and recruitment methods, will be carried out following the rules and regulations of the Johns Hopkins Medicine IRB and in accordance with HIPAA regulations.

Fig. 1

REST study design and visit timeline. The total study period is approximately 13–15 weeks and involves a series of five in-person visits for eligible participants. HVLT-R, Hopkins Verbal Learning Test-Revised; CDR, Clinical Dementia Rating; MMSE, Mini-Mental State Exam; NPI-C, Neuropsychiatric Inventory-Clinician Rating Scale; PSQI, Pittsburgh Sleep Quality Index; ADLs, Activities of Daily Living; ESS, Epworth Sleepiness Scale; SF-12, Short-Form 12 Health Survey; HST, Home Sleep Study.

Recruitment

Individuals with self-identified memory and sleep complaints will be recruited primarily through automated electronic health record messages sent to potential participants based on age, zip code, and excluded sleep medications. Additional recruitment strategies will include patient referrals from the Alzheimer’s Disease Research Center and the Memory and Alzheimer’s Treatment Center at Johns Hopkins, advertisements placed in local newspapers, radio stations, community events, social media, and postings in primary care offices in the local catchment area.

Eligibility criteria

Individuals who express interest in the study will undergo a screening telephone assessment to determine study eligibility. Those who meet the study criteria during the phone screening will be invited to a baseline visit which will include medical, cognitive, neuropsychiatric, and assessments of daily function. Participants must meet diagnostic criteria for aMCI at baseline to be included in the study, as supported by a Clinical Dementia Rating Scale (CDR) equal to 0.5. The study physician will obtain a history, including information from a knowledgeable informant on each participant’s cognitive status, and review participant assessment data in order to make a clinical diagnosis.

Specific inclusion criteria for the study are as follows: 1) clinical diagnosis of aMCI as defined by Albert et al. [63] which includes subjective memory complaint and objective evidence of memory problems; 2) CDR of 0.5 with a Memory Box score of >0.5; 3) Evidence of sleep complaints with Pittsburgh Sleep Quality Index score of >5 (a cutoff observed in >40% of older persons); 4) Memory performance >1.0 SD below age-education norms on HVLT-R; 5) Adequate visual and auditory acuity for neuropsychological testing; 6) Good general health with no disease expected to interfere with participation in the study; 7) Ability to have an MRI scan; and 8) Availability of knowledgeable informant [33, 34].

The following criteria will be used as exclusion: 1) Less than 55 years of age; 2) Too frail or medically unstable to undergo study procedures; 3) Prior diagnosis of obstructive sleep apnea or evidence of moderate to severe obstructive sleep apnea on baseline PSG as evidenced by an apnea/hypopnea index of≥15 (as defined by Medicare criteria, i.e., apneas or hypopneas with >4% O₂ desaturations); 4) Moderate or severe dementia with CDR 2 or greater; 5) Cognitive complaints and deficits better explained by other medical/neurologic conditions; 6) Delirium; 7) Allergic to TRZ; 8) Taking sleep medications including TRZ and unwilling to undergo 4-week pre-study washout period; 9) Current substance abuse; 10) Current major depressive, manic, or acute psychotic episode; 11) Prior diagnosis of significant systemic illness or unstable medical condition which could lead to difficulty complying with the study protocol or represent alternate primary cause of memory problems beyond AD pathology; 12) Lack of available knowledge informant; 13) Prior diagnosis of QT (corrected for heart rate) prolongation (>470 ms in females or > 450 ms in males); and 14) Inability to provide informed consent.

Study measures

Medical history, clinical data, and demographics

During the screening telephone assessment, we will collect participant medical history and current medications. Participants taking medications prescribed for sleep including TRZ are asked to pause treatment to participate in the study and will undergo a minimum of a 4-week pre-study washout period prior to enrollment. At visit 1, we will collect participant demographic information including age, sex, race, ethnicity, and education. Clinical measures will be recorded at each visit, including vital signs, physical exams. Electrocardiogram will be performed only at Visit 1.

Cognitive battery

To characterize cognition in aMCI patients, a comprehensive battery will be administered that spans all major cognitive domains. The MMSE and the CDR will be administered during Visit 1 to assess baseline cognition [35, 36]. The MMSE was used to allow maximum comparability with published trial results [24]. Visits 1, 3, 4, and 5 will include tasks such as Digit Span, Trail Making Test, Letter (Phonemic) Fluency, and several tasks from the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS), including Coding, List Learning and Recall, and Story Memory and Recall [37 –40]. These tests collectively assess domains of attention, processing speed, declarative memory, and executive functioning. Practice effects are minimized with the use of alternate forms of each test.

Neuropsychiatric measures

Sleep and cognition can affect and be affected by neuropsychiatric symptoms other than sleep/wake problems, particularly depression and anxiety [41]. The Neuropsychiatric Inventory-Clinician rating scale (NPI-C) is administered at visits 1, 3, 4, and 5 to evaluate 14 neuropsychiatric symptoms common throughout the dementia spectrum. These domains include delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, elation/euphoria, apathy/indifference, disinhibition, irritability, aberrant motor behavior, sleep, and appetite/eating disorders [41]. The frequency of symptoms will be rated based on participant interview and responses from participants’ caregivers, and the study clinician will make an expert judgment as to the overall symptom severity coded from 0 (‘none’) to 3 (‘severe’). In addition to participant psycho-pathology, caregiver distress caused by each of the neuropsychiatric symptoms will be captured by the NPI-C.

Functional ability and quality of life

To determine participant functional ability and quality of life, participants and their caregivers complete the Alzheimer’s Disease Cooperative Study-Activities of Daily Living (ADCS-ADL, referred to as “ADLs” in Fig. 1), a 27-item instrument assessing the ability to perform activities of daily living [42]. A study clinician will also rate participant physical and mental well-being through the 12-Item Short Form Health Survey [43].

Sleep assessments

Participants are given Actiwatch Spectrum Plus devices (Philips Respironics, Murrysville, PA, USA) to wear throughout the 13- to 15-week study period beginning at Visit 1. The devices measure omnidirectional movement via an accelerometer with a piezoelectric bimorph element that generates a voltage when moved. The device determines activity by capturing voltage differences and will be set to record voltages in 1-min epochs. Participants or study partners also fill out sleep diaries daily to record several events, including when they 1) got into bed each night, 2) intended to begin sleeping, 3) got out of bed, and 4) removed the Actiwatch. Participants are instructed to wear the Actiwatch 24/7 and leave it on during showering, bathing, and swimming. Actiwatch will be replaced at 4-week intervals in the clinic. Participants will not change it themselves. If the battery runs out in less than 4 weeks, the data remain intact and retrievable.

Participants undergo four separate overnight home sleep studies following visits 1, 3, 4, and 5 with the Sleep Profiler PSG2 (Advanced Brain Monitoring, Carlsbad, CA, USA). During visit 1, participants and their study partners are trained on the administration of the PSG2 device at home. The PSG2 records electrooculography, electromyography, and electrocardiography signals via leads that can be placed by the participant, or their care partner based on provided instructions [44]. The device allows for full PSG and measures nasal airflow, respiratory effort (chest and abdomen), pulse oximetry, body position, and snoring. This system enables high-quality measurements of sleep staging and amount, including SWS and rapid eye movement (REM) sleep, as well as other signals of interest including wake after sleep onset, sleep onset latency, and sleep efficiency. Signs of obstructive sleep apnea, such as snoring, will also be monitored during the first home sleep study. Any incidental findings will be reported to the participant and assessed as a potential exclusioncriterion.

Participant sleep habits are also evaluated through self-report questionnaires, including the Pittsburgh Sleep Quality Index (PSQI) and the Epworth Sleepiness Scale (ESS) [34, 45]. These questionnaires are administered on visits when an ambulatory PSG is provided (visits 1, 3, 4, and 5).

Functional magnetic resonance imaging encoding task

Participants complete three MRI sessions during visits 2, 3, and 5. Each MRI session includes task-based functional MRI scans as well as structural imaging (6 min). The task-based fMRI scan acquires high-resolution functional images with an isotropic resolution of 1.5 mm x 1.5 mm x 1.5 mm with no gap. Seventy triple-oblique slices are acquired parallel to the principal longitudinal axis of the hippocampus to cover almost the entire brain. In addition, a whole-brain MP-RAGE structural scan with 0.65 mm isotropic resolution is acquired. Techniques that maximize power and signal-to-noise ratio in the hippocampus will be utilized including manual segmentation of the hippocampal subfields following previously described methods [6 , 46] and the use of Advanced Normalization Tools, which implements a diffeomorphic algorithm, symmetric normalization to register the entire brain [47, 48].

During the task-based fMRI scans, participants complete the behavioral pattern separation objects task (BPS-O) [26–28 , 46]. This memory task aims to examine hippocampal subregion specific activity. Participants are presented with a series of pictures of everyday objects, and asked to determine whether each item is new, old (in the context of the task), or similar to a previously seen item. Objects that are merely similar but not identical to an item they’ve seen before are termed “lures,” and the correct identification of these merely “similar” or ‘lure” items have been shown to depend critically on DG/CA3 mediated pattern separation. Several cohort studies have reported that aMCI participants show a deficit in their ability to encode inputs with some degree of overlapping information (pattern separation), which is associated with increased activity in the DG/CA3 subregion on the fMRI scan [46].

Exploratory analyses: Plasma measures of Aβ, phosphorylated tau, neurofilament-light (NFL), and glial fibrillary acidic protein (GFAP)

Blood samples are collected during visits 1, 3, and 5 (baseline and after each treatment phase) in ethylenediaminetetraacetic acid (EDTA) tubes by venipuncture and centrifuged twice at 2500 xg for 15 min sequentially. 500 μl aliquots will be stored at –80°C with glycerol freezing component until ready for analysis. Genomic DNA will be extracted from a buffy coat culture using the QIAamp DNA Mini QIAcube Kit (Qiagen, Germantown MD). APOE genotyping will be performed via pre-made TaqMan single nucleotide polymorphism assays including C-905013-10 (rs405509), C-3084793-20 (rs429358), and C-904973-10 (rs7412) using an Applied Biosystems 7900HT Real Time PCR System. Plasma samples will be analyzed for biomarkers such as Aβ₄₀, Aβ₄₂, phosphorylated tau181, neurofilament light, and GFAP using single molecule array (SIMOA) on the Quanterix HDX platform, run at 1 : 4 onboard automated dilution in duplicate with the appropriate kit calibrators and kit controls. In addition, two control human plasma samples will be run on each plate with normal and low Aβ_42/40 to assess inter-plate variability.

Safety monitoring

An independent Data and Safety Monitoring Board (DSMB) will act in an advisory capacity to the REST investigators. Three experienced and knowledgeable investigators with expertise in dementia research, sleep research, and statistical support for trials will comprise the DSMB. A representative from the National Institute of Aging (NIA) will also be present. The DSMB will have access to unblinded results stratified by the treatment arm. Quality and Assurance Monitoring reports of the accumulating study data provided by the study’s Data Coordinating Center will be presented to the DSMB and will include comparisons of baseline characteristics, measures of disturbance, functional and cognitive outcomes variables, and occurrence of adverse events between treatment groups. The board may recommend for changes in study procedures or even pausing or stopping the trial if they observe convincing evidence of adverse treatment-associated safety issues.

Data analytic strategy

Power analysis

Sample size estimates were based on the pre-post change in the BPS-O score observed in a relevant drug trial for aMCI patients [27]. A one-sample t-test was used to estimate the power needed to see a treatment effect. To protect against a potential loss in power due to the influence of baseline covariates on outcomes, as well as data loss due to participants lost to follow-up, our estimated sample size is inflated by 24–45 patients above that needed to achieve 80% power, giving a total sample size of 100.

Aim 1: Sleep

In assessing the impact of TRZ on sleep, we will first compare sleep duration, percent time spent in SWS, SWA (0.5–4 Hz), sleep onset latency, and fragmentation indices (wake after sleep onset and sleep efficiency) between treatment arms. These outcome variables will be treated as continuous, assuming the normal approximation for binomial data. We will also conduct a repeated measures ANCOVA by performing a mixed-effects regression, whereby patients will act as their own controls. This analysis will allow for assessment of a period effect in which treatment order modifies the effect of treatment arms on outcome. Adjustments will be made for baseline covariates such as age, sex, APOE genotype, medical/mental health comorbidities, and sleep medication use requiring washout before study participation if they are imbalanced by treatment arm (alpha cutoff = 0.1).

Aim 2: Memory

In assessing the impact of TRZ on memory performance, we will compare hippocampal activation as measured by fMRI and BPS-O scores between treatment arms. We will also determine whether improvements in these measures are impacted by time spent in SWS and SWA. Hippocampal activation is a continuous outcome and analysis of this outcome will mirror the procedure outlined in Aim 1. BPS-O scores will be measured as the difference in correct responses between the TRZ and placebo arms, and thus, positive values indicate memory improvements while taking TRZ. Treatment effect will be evaluated by the proportion of participants who improve on TRZ, and scores will be analyzed using a mixed-effects regression model to account for patient-specific differences in baseline BPS-O scores and hippocampal activation. Finally, a causal mediation analysis as described by VanderWeele (2014) will be performed to parse out the effect of TRZ on sleep measures listed above, the effect of sleep measures on hippocampal-dependent memory and other cognitive outcomes, and the direct effect of TRZ on hippocampal-dependent memory [49]. The mediation analysis with associated sensitivity analyses will help to determine whether the impacts of TRZ on memory may be mediated through its impact on sleep [50].

Exploratory analyses

We will also analyze whether the presence of the APOE4 allele or baseline plasma levels of Aβ_42/40 ratios, phosphorylated tau181, neurofilament light, and GFAP are associated with responses in the sleep and memory outcomes outlined above. ANCOVA regression models controlling for clinical and demographic covariates will be used to determine whether changes in plasma levels (Aβ_42/40 ratios, phosphorylated tau181, neurofilament light and GFAP) occur after exposure to TRZ in groups stratified by APOE4 genotype. Secondly, unblinded post-hoc analyses of samples collected after each treatment phase will determine whether pre- and post-measures of Aβ_42/40 ratios, phosphorylated tau181, neurofilament light, and GFAP are modified by exposure to TRZ in the presence or absence of APOE4 using a TRZ by APOE interaction term in the ANCOVA regression models.

DISCUSSION

Sleep disturbances and sleep fragmentation are relatively common in AD and aMCI individuals, beginning as early as the prodromal stages [3 , 51–53]. The bi-directional relationship between AD and poor sleep has been attributed to several potential mechanisms, including upregulation of the neurotransmitter orexin A, changes in neural excitability, and reduced glymphatic clearing of Aβ [54]. For example, in orexin knockout murine models, longer sleep cycles and decreased aggregation of Aβ were observed [54]. Consistent with these observations, the dual orexin receptor antagonist suvorexant was recently reported to reduce tau phosphorylation and Aβ in the CSF of non-AD human subjects [15]. Studies have also shown that reductions in SWS, during which the glymphatic system is particularly active, affect markers of AD-related pathogenesis and memory processes, particularly declarative memory consolidation [3 , 21]. In the context of AD, reductions in SWS and slowed glymphatic clearance may lead to accumulations of Aβ and tau and impaired memory performance. In fact, acute sleep loss has been reported to be associated with reduced CSF-to-blood clearances of AD biomarkers, and non-REM SWA has been recently described as a potential protective cognitive reserve factor in AD[55, 56].

Given the strong relationship between SWS and markers of AD-related pathogenesis, we describe a study on the use of TRZ for improving sleep and cognition in aMCI and assess its potential use as a disease modifying treatment. SWS has been shown to be reduced in many patients across the AD spectrum, and TRZ is a widely used and safe drug that has shown potential in inducing SWS in older adults and even in slowing memory decline [6, 17]. Therefore, this trial examines whether TRZ is effective as a pharmacological intervention to induce SWS and potentially disrupt the buildup of Aβ and tau protein aggregates in older adults with prodromal AD/aMCI and sleep complaints.

Alternative pharmacological approaches targeting sleep in the context of dementia and AD have produced mixed results. For example, a meta-analysis found that melatonin, a naturally occurring hormone with a strong safety profile, led to modest improvement in MMSE scores in patients with mild AD, but did not improve global cognition on the Alzheimer’s Disease Assessment Scale-Cognition (ADAS-Cog) irrespective of duration of treatment or stage of AD [57]. Benzodiazepines may have strong sedative effects but may also significantly decrease memory function, while second-generation antipsychotics do not have FDA indications for treatment of insomnia and are not recommended for routine use in older adults due to high risks of cardiovascular events and death [58, 59]. There is also limited data on using sedating antidepressants for sleep in the context of dementia, although one study found that mirtazapine did not benefit sleep after 2 weeks in AD patients but did increase daytime sleepiness [60]. Suvorexant, a recently approved dual orexin receptor antagonist with a promising safety profile, was shown to increase total sleep time and decrease wake after sleep onset but did not impact time spent in SWS [15]. Further PSG analyses are needed to explore mechanisms related to the efficacy of suvorexant in AD patients [61]. Finally, a recent review found that non-pharmacological interventions for sleep problems in patients with dementia, such as light therapy, physical and social activities, and caretaker interventions, failed to improve sleep with sufficient certainty [62].

In the context of limited efficacy and/or pertinent safety concerns of past pharmacological and non-pharmacological interventions for sleep difficulties in older adults with cognitive impairment, we propose the use of TRZ for improving sleep and cognition in aMCI. We have two major biological/mechanistic outcomes: 1) The potential benefit of TRZ on sleep and sleep architecture. To this end, we will measure sleep variables with actigraphy and home sleep testing; 2) The potential benefit of TRZ on hippocampal function and encoding mechanisms. To this end, we will assess the effect of TRZ on hippocampal activation during a pattern separation task as measured by fMRI, hypothesizing that TRZ will normalize hippocampal hyperactivity associated with aMCI during the task and improve memory outcomes [6]. Additionally, our exploratory aim is to analyze the presence of plasma biomarkers relevant to AD-related pathology, including Aβ_42/40 ratio, phosphorylated tau181, GFAP and neurofilament light, to understand whether levels of these AD biomarkers are modified in response to the impacts of TRZ on sleep.

Findings that TRZ does not negatively impact memory and/or cognitive function would encourage several areas of future research. First, future trials with longer treatment phases of TRZ would be warranted to observe changes in cognitive functioning. Additionally, a null primary outcome would motivate efforts to identify covariates that predict response, such as the moderating effects of severity of cognitive impairment, age, or medical history.

There has been increasing interest in targeting sleep to improve cognitive and functional outcomes throughout the AD spectrum. Thus, our clinical trial of a promising and safe drug (TRZ) to address these issues in very early AD, with a thorough set of mechanistically oriented biological markers, may increase potential treatment response. We regard this as a promising first step in targeting sleep to improve AD outcomes.

Footnotes

ACKNOWLEDGMENTS

All authors wish to thank our current and future participants and their study partners. Additionally, we would like to acknowledge the following contributors: National Institute of Aging (NIA); Alzheimer’s Disease Research Center (ADRC) for their recruitment efforts; Kennedy Krieger Institute: Doris Lin, MD, PhD, study radiologist; Julia Fletcher; Hannah Miller; and Karthik Kartha: MRI coordinators; Johns Hopkins Bayview and Johns Hopkins School of Medicine: Laurie Bernie, PharmD, study pharmacist; Meghan Schultz, RN, MSN, senior study nurse; Frank P. Sgambati; Joshua East; and Nicholas Bienko, MA; Data Collection: Samantha Horn, MA; Samantha Schultz, BS; Mersania Jn Pierre, BS; and Phoebe Clark, BS.

FUNDING

This study is funded by the National Institute of Aging (NIA), Grant R01 AG071522.

CONFLICT OF INTEREST

P. Rosenberg received support in the form of grant funding from the National Institute of Aging Alzheimer’s Association, Lilly, Functional Neuromodulation, Vaccinex, the Alzheimer’s Disease Cooperative Study (ADCS), Alzheimer’s Disease Trials Research Institute (ATRI), and the Alzheimer’s Clinical Trials Consortium (ACTC); he has served as a consultant to GLG, Leerink, Otsuka, Avanir, ITI, IQVIA, Food and Drug Administration, Cerevel, Bioxcel, Sunovion, and Acadia.

A. Spira received payment for serving as a consultant for Merck and received honoraria from Springer Nature Switzerland AG for guest editing special issues of Current Sleep Medicine Reports.

DATA AVAILABILITY

The data from this trial will be made available to the field as a whole following NIH guidelines, with the current plan being to make them publicly available one year after the last data are collected.

References

(2023) 2023 Alzheimer’s disease facts and figures. Alzheimers Dement19, 1598–1695.

Brzecka

, Leszek

, Ashraf

, Ejma , Ávila-Rodriguez , Yarla

, Tarasov

, Chubarev

, Samsonova

, Barreto

(2018) Sleep disorders associated with Alzheimer’s disease: A perspective. Front Neurosci12, 330.

Mander

, Winer

, Jagust

, Walker

(2016) Sleep: A novel mechanistic pathway, biomarker, and treatment target in the pathology of Alzheimer’s disease?Trends Neurosci39, 552–566.

Wang

, Holtzman

(2020) Bidirectional relationship between sleep and Alzheimer’s disease: Role of amyloid, tau, and other factors. Neuropsychopharmacology45, 104–120.

Xiao

, Liu

, Lu

, Sha

, Xu

, Yu

, Lee

(2022) Possible neuropathology of sleep disturbance linking to Alzheimer’s disease: Astrocytic and microglial roles. Front Cell Neurosci16, 875138.

Bakker

, Krauss

, Albert

, Speck

, Jones

, Stark

, Yassa

, Bassett

, Shelton

, Gallagher

(2012) Reduction of hippocampal hyperactivity improves cognition in amnestic mild cognitive impairment. Neuron74, 467–474.

Bakker

, Albert

, Krauss

, Speck

, Gallagher

(2015) Response of the medial temporal lobe network in amnestic mild cognitive impairment to therapeutic intervention assessed by fMRI and memory task performance. Neuroimage Clin7, 688–698.

Scullin

, Gao

(2018) Dynamic contributions of slow wave sleep and REM sleep to cognitive longevity. Curr Sleep Med Rep4, 284–293.

YES

, Ooms

, Sutphen

, Macauley

, Zangrilli

, Jerome

, Fagan

, Mignot

, Zempel

, Claassen

(2017) Slow wave sleep disruption increases cerebrospinal fluid amyloid-β levels. Brain140, 2104–2111.

10.

Lucey

, Hicks

, McLeland

, Toedebusch

, Boyd

, Elbert

, Patterson

, Baty

, Morris

, Ovod

(2018) Effect of sleep on overnight cerebrospinal fluid amyloid β kinetics. Ann Neurol83, 197–204.

11.

Ooms

, Overeem

, Besse

, Rikkert

, Verbeek

, Claassen

(2014) Effect of 1 night of total sleep deprivation on cerebrospinal fluid β-amyloid 42 in healthy middle-aged men: A randomized clinical trial. JAMA Neurol71, 971–977.

12.

Peng

, Achariyar

, Li

, Liao

, Mestre

, Hitomi

, Regan

, Kasper

, Peng

, Ding

(2016) Suppression of glymphatic fluid transport in a mouse model of Alzheimer’s disease. Neurobiol Dis93, 215–225.

13.

Mander

, Rao

, Lu

, Saletin

, Lindquist

, Ancoli-Israel

, Jagust

, Walker

(2013) Prefrontal atrophy, disrupted NREM slow waves and impaired hippocampal-dependent memory in aging. Nat Neurosci16, 357–364.

14.

Musiek

, Holtzman

(2016) Mechanisms linking circadian clocks, sleep, and neurodegeneration. Science354, 1004–1008.

15.

Lucey

, Liu

, Toedebusch

, Freund

, Redrick

, Chahin

, Mawuenyega

, Bollinger

, Ovod

, Barthélemy

(2023) Suvorexant acutely decreases tau phosphorylation and Aβ in the human CNS. Ann Neurol94, 27–40.

16.

Ashford

(2019) Treatment of Alzheimer’s disease: Trazodone, sleep, serotonin, norepinephrine, and future directions. J Alzheimers Dis67, 923–930.

17.

Burke

, Hu

, Spadola

, Li

, Naseh

, Burgess

, Cadet

(2018) Mild cognitive impairment: Associations with sleep disturbance, apolipoprotein e4, and sleep medications. Sleep Med52, 168–176.

18.

Buysse

, Tyagi

(2017) Clinical pharmacology of other drugs used as hypnotics. In Principles and Practice of Sleep Medicine (Sixth Edition), Kryger

, Roth

, Dement

, eds. Elsevier, pp. 432–445.e7.

19.

Wichniak

, Wierzbicka

, Walęcka

, Jernajczyk

(2017) Effects of antidepressants on sleep. Curr Psychiatry Rep19, 63.

20.

Diekelmann

, Born

(2010) The memory function of sleep. Nat Rev Neurosci11, 114–126.

21.

Kam

, Parekh

, Sharma

, Andrade

, Lewin

, Castillo

, Bubu

, Chua

, Miller

, Mullins

(2019) Sleep oscillation-specific associations with Alzheimer’s disease CSF biomarkers: Novel roles for sleep spindles and tau. Mol Neurodegener14, 10.

22.

Schroeck

, Ford

, Conway

, Kurtzhalts

, Gee

, Vollmer

, Mergenhagen

(2016) Review of safety and efficacy of sleep medicines in older adults. Clin Ther38, 2340–2372.

23.

Camargos

, Louzada

, Quintas

, Naves

, Louzada

, Nóbrega

(2014) Trazodone improves sleep parameters in Alzheimer disease patients: A randomized, double-blind, and placebo-controlled study. Am J Geriatr Psychiatry22, 1565–1574.

24.

, Walsh

, Neylan

, Vossel

, Yaffe

, Krystal

, Miller

, Karageorgiou

(2019) Long-term trazodone use and cognition: A potential therapeutic role for slow-wave sleep enhancers. J Alzheimers Dis67, 911–921.

25.

Sommerlad

, Werbeloff

, Perera

, Smith

, Costello

, Mueller

, Kormilitzin

, Broadbent

, Nevado-Holgado

, Lovestone

(2022) Effect of trazodone on cognitive decline in people with dementia: Cohort study using UK routinely collected data. Int J Geriatr Psychiatry37, doi: 10.1002/gps.5625.

26.

Killgore

(2010) Effects of sleep deprivation on cognition. Prog Brain Res185, 105–129.

27.

Roth

, McCall

, Liguori

(2011) Cognitive, psychomotor and polysomnographic effects of trazodone in primary insomniacs. J Sleep Res20, 552–558.

28.

Kirwan

, Stark

(2007) Overcoming interference: An fMRI investigation of pattern separation in the medial temporal lobe. Learn Mem14, 625–633.

29.

Leuzy

, Mattsson-Carlgren

, Palmqvist

, Janelidze

, Dage

, Hansson

(2022) Blood-based biomarkers for Alzheimer’s disease. EMBO Mol Med14, e14408.

30.

Bryant

, Ereshefsky

(1982) Antidepressant properties of trazodone. Clin Pharm1, 406–417.

31.

Kumi

, Baweka

(2009) Center for Drug Evaluation and Research Application Number: 22-411 Clinical Pharmacology and Biopharmaceutics Review(s), Food and Drug Administration.

32.

Benedict

, Schretlen

, Groninger

, Brandt

(1998) Hopkins Verbal Learning Test– Revised: Normative data and analysis of inter-form and test-retest reliability. Clin Neuropsychol12, 43–55.

33.

Ryan

, Woods

, Murray

, Shah

, Britt

, Reid

, Wolfe

, Nelson

, Lockery

, Orchard

(2021) Normative performance of older individuals on the Hopkins Verbal Learning Test-Revised (HVLT-R) according to ethno-racial group, gender, age and education level. Clin Neuropsychol35, 1174–1190.

34.

Buysse

, Reynolds

III CF

, Monk

, Hoch

, Yeager

, Kupfer

(1991) Quantification of subjective sleep quality in healthy elderly men and women using the Pittsburgh Sleep Quality Index (PSQI). Sleep14, 331–338.

35.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res12, 189–198.

36.

Berg

(1998) Clinical dementia rating (CDR). Psychopharmacol Bull24, 637–639.

37.

Wechsler

(1997) Digit Span Subtest: Wechsler Adult Intelligence Scale. The Psychological Corporation, San Antonio, TX.

38.

Reitan

, Wolfson

(1993) The Halstead-Reitan Neuropsychology Battery: Theory and Clinical Interpretation, Tucson, AZ, Neuropsychology Press-(1995): “Category test and trail making test as measures of frontal lobe functions”. Clin Neuropsychol9, 50–56.

39.

Borkowski

, Benton

, Spreen

(1967) Word fluency and brain damage. Neuropsychologia5, 135–140.

40.

Randolph

(2012) RBANS update: Repeatable battery for the assessment of neuropsychological status. PsychCorp.

41.

de Medeiros

, Robert

, Gauthier

, Stella

, Politis

, Leoutsakos

, Taragano

, Kremer

, Brugnolo

, Porsteinsson

(2010) The Neuropsychiatric Inventory-Clinician rating scale (NPI-C): Reliability and validity of a revised assessment of neuropsychiatric symptoms in dementia. Int Psychogeriatr22, 984–994.

42.

Perneczky

, Pohl

, Sorg

, Hartmann

, Komossa

, Alexopoulos

, Wagenpfeil

, Kurz

(2006) Complex activities of daily living in mild cognitive impairment: Conceptual and diagnostic issues. Age Ageing35, 240–245.

43.

Ware

Jr , Kosinski

, Keller

(1996) A 12-Item Short-Form Health Survey: Construction of scales and preliminary tests of reliability and validity. Med Care34, 220–233.

44.

Levendowski

, Ferini-Strambi

, Gamaldo

, Cetel

, Rosenberg

, Westbrook

(2017) The accuracy, night-to-night variability, and stability of frontopolar sleep electroencephalography biomarkers. J Clin Sleep Med13, 791–803.

45.

Johns

(1991) A new method for measuring daytime sleepiness: The worth Epsleepiness scale. Sleep14, 540–545.

46.

Bakker

, Kirwan

, Miller

, Stark

(2008) Pattern separation in the human hippocampal CA3 and dentate gyrus. Science319, 1640–1642.

47.

Avants

, Tustison

, Song

, Cook

, Klein

, Gee

(2011) A reproducible evaluation of ANTs similarity metric performance in brain image registration. Neuroimage54, 2033–2044.

48.

Klein

, Andersson

, Ardekani

, Ashburner

, Avants

, Chiang

, Christensen

, Collins

, Gee

, Hellier

(2009) Evaluation of 14 nonlinear deformation algorithms applied to human brain MRI registration. Neuroimage46, 786–802.

49.

VanderWeele

(2014) A unification of mediation and interaction: A four-way decomposition. Epidemiology25, 749–761.

50.

VanderWeele

, Arah

(2011) Bias formulas for sensitivity analysis of unmeasured confounding for general outcomes, treatments, and confounders. Epidemiology22, 42–52.

51.

Bubu

, Brannick

, Mortimer

, Umasabor-Bubu

, Sebastião

, Wen

, Schwartz

, Borenstein

, Wu

, Morgan

(2017) Sleep, cognitive impairment, and Alzheimer’s disease: A systematic review and meta-analysis. Sleep40, zsw032.

52.

, Langhough

, Hermann

, Jonaitis

, Betthauser

, Cody

, Mueller

, Zuelsdorff

, Chin

, Ennis

(2023) Associations between self-reported sleep patterns and health, cognition and amyloid measures: Results from the Wisconsin Registry for Alzheimer’s Prevention. Brain Commun5, fcad039.

53.

Falgàs

, Walsh

, Yack

, Simon

, Allen

, Kramer

, Rosen

, Joie

, Rabinovici

, Miller

(2023) Alzheimer’s disease phenotypes show different sleep architecture. Alzheimers Dement19, 3272–3282.

54.

Zhao

, You

, Wang

, Zhou

, Chai

, Yan

, Jin

, Wang

, Sun

(2022) Orexin A peptidergic system: Comparative sleep behavior, morphology and population in brains between wild type and Alzheimer’s disease mice. Brain Struct Funct227, 1051–1065.

55.

Liu

, Barthélemy

, Ovod

, Bollinger

, He

, Chahin

, Androff

, Bateman

, Lucey

(2023) Acute sleep loss decreases CSF-to-blood clearance of Alzheimer’s disease biomarkers. Alzheimers Dement19, 3055–3064.

56.

Zavecz

, Shah

, Murillo

, Vallat

, Mander

, Winer

, Jagust

, Walker

(2023) NREM sleep as a novel protective cognitive reserve factor in the face of Alzheimer’s disease pathology. BMC Med21, 156.

57.

Sumsuzzman

, Choi

, Jin

, Hong

(2021) Neurocognitive effects of melatonin treatment in healthy adults and individuals with Alzheimer’s disease and insomnia: A systematic review and meta-analysis of randomized controlled trials. Neurosci Biobehav Rev127, 459–473.

58.

Ettcheto

, Olloquequi

, Sanchez-Lopez

, Busquets

, Cano

, Manzine

, Beas-Zarate

, Castro-Torres

, García

, Bullo

(2020) Benzodiazepines and related drugs as a risk factor in Alzheimer’s disease dementia. Front Aging Neurosci11, 344.

59.

Maust

, Kim

, Seyfried

, Chiang

, Kavanagh

, Schneider

, Kales

(2015) Antipsychotics, other psychotropics, and the risk of death in patients with dementia: Number needed to harm. JAMA Psychiatry72, 438–445.

60.

Scoralick

, Louzada

, Quintas

, Naves

, Camargos

, Nóbrega

(2017) Mirtazapine does not improve sleep disorders in Alzheimer’s disease: Results from a double-blind, placebo-controlled pilot study. Psychogeriatrics17, 89–96.

61.

Herring

, Ceesay

, Snyder

, Bliwise

, Budd

, Hutzelmann

, Stevens

, Lines

, Michelson

(2020) Polysomnographic assessment of suvorexant in patients with probable Alzheimer’s disease dementia and insomnia: A randomized trial. Alzheimers Dement16, 541–551.

62.

Wilfling

, Calo

, Dichter

, Meyer

, Möhler

, Köpke

(2023) Non-pharmacological interventions for sleep disturbances in people with dementia. Cochrane Database Syst Rev1, CD011881.

63.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

, Snyder

, Carrillo

, Thies

, Phelps

(2011) The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease, Alzheimers & Dementia7(3)m 270–279.