Introduction

Dementias and other neurocognitive disorders rank with heart disease and cancer as the largest contributors to morbidity and mortality in America, with Alzheimer’s disease (AD) the most prominent of these illnesses. Currently, there are approximately 5.7 million Americans living with AD and it is predicted that 14 million will be diagnosed by 2050 [1]. AD is also an incredibly expensive illness, with $277 billion spent annually in direct care for AD patients [2]. Additionally, AD remains the sixth leading cause of death in the USA, and while the mortality rate from heart disease and many cancers have decreased over the last 20 years, deaths from AD have increased by 123% between 2000 and 2015 [1]. Alzheimer’s disease will only cause more devastation as the population continues to age, so advances in management are imperative.

One feature often overlooked yet inherent to the insidious nature of AD is the collection of associated neuropsychiatric symptoms (NPS). Ninety-seven percent of patients with AD will demonstrate at least one neuropsychiatric symptom, with apathy, depression, anxiety, and sleep disorders the most common [3••, 4] (see Table 1 for prevalence by type of NPS). Additionally, NPS can predict the course of the disease: A 2015 evaluation of subjects with AD involved in the Cache County Dementia Progression Study showed that aggression, psychotic symptoms (such as hallucinations and delusions), and mood symptoms (such as depression, anxiety) were all associated with shorter survival time from mild AD to death; aggression and psychotic symptoms were additionally associated with more rapid functional decline to severe dementia [5••].

Table 1 Neuropsychiatry symptoms (NPS) in Alzheimer’s disease, by order of prevalence (as measured in the 2016 Journal of Affective Disorders study listed as article 3 in the Reference section). Definitions in this table approximate the descriptions of the NPS in the Neuropsychiatric Inventory (NPI), which was used to identify the NPS in the study
Full size table

Identifying and managing NPS in AD remain challenging due to the ambiguity between which symptoms represent psychiatric disorders and which represent dementia. Even when a diagnosis can be made, there is a shortage of effective treatments that can be used safely in the typically elderly and frail AD patient population. Yet progress has been made in understanding the development of neuropsychiatric symptoms in AD patients and constructing therapies. This review, then, will focus on the most recent findings on NPS in AD, exploring the pathophysiology of NPS in an attempt to explain why these symptoms are so common and establish the framework for a discussion of the latest developments in the diagnosis and management of NPS in AD.

Pathophysiology of Neuropsychiatric Symptoms in Alzheimer’s Disease

Before delving into the mechanism of NPS, it is important to understand the pathophysiology of AD itself. Ideas of how AD develops have changed considerably over the decades since this illness was first studied. The neurotransmitter acetylcholine (ACh) served as a particular source of interest in early studies. Such studies include illustrations of how anticholinergic therapies impair memory, autopsies of the brains of AD patients revealing a depletion of enzymes responsible for the development of ACh, and application of acetylcholinesterase inhibitors—which block acetylcholinesterase, the enzyme responsible for the breakdown of ACh—improving cognition of test subjects. These findings provide the basis for the “cholinergic hypothesis,” which proposes that the deficiency in acetylcholine is responsible for the development of late-onset AD, with the severity of the dementia directly related to the degeneration of cholinergic neurons [6]. This hypothesis—or at least the idea that AD stems solely from a dearth of acetylcholine—has been critiqued because of the lack of efficacy of acetylcholinesterase inhibitors in treating AD.

Later studies argued that protein misfolding led to the development of amyloid plaques that cause neural damage and cognitive impairment. This hypothesis arose from observations that certain mutations leading to excessive production of amyloid (i.e., extra genetic copies of the apolipoprotein on chromosome 21, which comes in triplicate in patients with Down syndrome and increases their risk of developing AD) or its defective destruction (i.e., the APOΣ4 enzyme) significantly increased the risk of AD [7]. Another protein of interest is the tau protein. Normally responsible for neuronal communication, tau proteins are hyperphosphorylated in AD and subsequently aggregate into neurofibrillary tangles that destroy neurons [8] and cause atrophy of the hippocampi, amygdala, and temporal lobes [9].

It is this atrophy that arguably causes the NPS in AD. A 2015 overview of studies related to NPS in AD noted that certain areas of the brain—such as the hippocampus, frontal lobes, anterior and posterior cingulate, insula, and amygdala—tend to suffer a particularly excessive amount of atrophy. It makes sense that damage to these regions would lead to NPS, because collectively, they are associated with impulse control, motivated behavior, emotional expression, and mood regulation [10••]. A more recent fMRI study on the damage to the hippocampal-prefrontal cortex network in both AD and depression explained that although there are subtle differences in the types of changes undergone in the two disorders, both involve prominent volume reductions in the hippocampi. Such findings indicate the hippocampus’s function in managing memory and regulating emotion [11•], and the observation of amyloid-mediated damage of the hippocampus damage occurring early in the development of AD reinforces the perception of depression as a sequelae of AD [12].

Structural damage of the brain could also lead to NPS via an impact in the production of monoamines such as serotonin, norepinephrine, and dopamine. There is evidence for a monoamine deficiency in AD similar to that in depression and anxiety disorders [13•]. Given that these neurotransmitters are so closely linked to mood, energy, and reward, it would make sense that the depletion in their levels would lead to depression [14], aggression [15],⁠ and apathy [16]. One possible explanation for the decrease in monoamines is the amyloid-mediated damage to areas containing serotonergic, noradrenergic, and dopaminergic neurons, including the entorhinal cortex [17]⁠ and the prefrontal cortex [18].

There is also the possibility that NPS may have less to do with neuroanatomical changes and more to do with inflammation, which has been proposed as factor in the development of depression [19, 20], schizophrenia [21•], and AD [22]. It is unclear if AD is a cause or a consequence of the inflammation potentially contributing to NPS. One study found that amyloid had a pro-inflammatory function [23], which suggests that AD drives inflammation. Yet inflammation may also just increase with age [24], which would mean that AD and geriatric depression are both consequences of inflammation. Moreover, the studies identifying illnesses with strong inflammatory components such as infections [25], rheumatoid arthritis [26]⁠, and even depression itself [27], as potential risk factors for AD, additionally point to increased inflammation as one etiology of AD. The concept that inflammation causes, or at least exacerbates, AD was recently reinforced by a recent meta-analysis indicating the potential for NSAIDs to decrease the risk of developing AD, regardless of the presence of NPS [28••].

Diagnosing NPS in AD

Clinical Diagnosis

Distinguishing the NPS in AD from psychiatric disorders such as MDD and schizophrenia is a formidable challenge. Despite advances in imaging and lab studies that will be discussed later in the section, AD remains primarily a clinical diagnosis and an understanding of the most common psychiatric presentation of AD could allow observed NPS to aid in diagnosis rather than serve as a confounding variable. The Neuropsychiatric Inventory (NPI) is a popular tool among clinicians in identifying psychiatric symptoms in patients with dementia. Developed in 1994 [29•], this form consists of questions designed to test the presence of NPS in 12 different domains. The NPI has undergone a few updates since its inception: the NPI-Clinician (NPI-C) scale increased the number of questions, added sub-questions to address the frequency, severity and duration of each symptom, and included clinician ratings in addition to the caregiver ratings from the original NPI [30]. Alternate versions of the NPI have allowed for the questions to be administered in nursing homes and also by clinicians with only 5 min to assess for NPS [31, 32].

The NPI was crucial in many of the studies measuring NPS that were aggregated in a 2015 meta-analysis. This study identified apathy as the most prevalent NPS in AD, occurring in 49% of patients. Depression was the next most frequent at 42% [3••]. Apathy is defined as a lack of interest and motivation; while these symptoms are also frequently seen in depression, apathy does not involve the low mood or hopelessness seen in depression. Furthermore, whereas depression primarily derives from a deficiency in serotonin, apathy stems from impairment in dopamine and acetylcholine transmission [33, 34]. These discrete mechanisms are key to understanding why apathy will respond to different treatments than depression in Alzheimer’s disease, and it is thus immensely important that clinicians be able to distinguish between depression and apathy in order to choose the appropriate medication (see Treatment section below for further details). Lastly, while the Cache County Dementia Progression Study cited earlier did not find that apathy predicted a quicker functional decline in AD, there have been other studies that have found that apathy can indeed serve as a risk factor for a transition from mild cognitive impairment to AD [35], and that unlike depression, it will worsen as the disease progresses [36].

Though not as prevalent as apathy, depression is perhaps more frequently recognized as an associated psychiatric symptom of AD, likely due to the impaired cognition that can occur with depression. In fact, the sheer frequency of cognitive impairment in late-life depression has spawned the term “pseudodementia,” referring to cognitive deficits in the absence of dementia. Though poor memory on basic cognitive screens such as the MOCA or MMSE may hint at the presence of dementia, it requires more thorough neuropsychological testing to isolate pseudodementia from AD. Generally, patients with AD will do worse than depressed patients on tasks testing semantic memory [37•], executive function, speech tasks, and delayed recall [38]. One potential explanation for these differences is that AD involves a much more aggressive destruction of the CNS while the poor performance on cognitive tasks in depression stems from more minor deficits in attention and motivation; this could account for the observations that depressed patients will often be quick to answer questions on cognitive tests with “I don’t know” [39]. Yet while AD patients perform worse on cognitive tests than patients with pseudodementia, there is no agreed-upon score that can serve as a cutoff point between pseudodementia and AD. Further research should hopefully address this deficit.

Psychotic symptoms also appear in a substantial portion of patients with AD, and their presence can provide a marker for distinguishing AD from MCI given psychosis is so much more prevalent in AD than in MCI [40•]. Psychosis in AD differs from psychosis seen in other neurologic and psychiatric disorders. Firstly, the delusions in AD are typically not bizarre, as in schizophrenia [41]. Instead, they are usually paranoid in nature, centered on ideas of persecution and infidelity [42]. Furthermore, the hallucinations in AD—when they do occur—are more frequently visual than auditory—another distinction from schizophrenia [43]. The co-presentation of memory impairment with visual hallucinations also occurs in Dementia with Lewy bodies (DLB), the most prevalent neurocognitive disorder aside from AD. Psychosis in DLB can be distinguished from AD, though the differences are subtler than those between schizophrenia and AD. Visual hallucinations are more frequent in DLB (a recent study finding visual hallucinations in 62% of LBD patients versus 9.5% in AD patients [44]) and occur earlier in the disease course, often before memory impairment occurs [45]. Clinicians can also diagnose DLB by investigating for the core features of the illness—determined in the 2017 DLB Consortium’s consensus report—that are not seen in AD, such as Parkinsonian symptoms and a history of REM sleep behavioral disturbances, including individuals acting out their dreams due to a lack of the atonia that normally occurs in REM sleep [46].

A thorough physical exam can also help lead to a diagnosis of AD, with some research showing that an impairment in smell can signal early dementia [47]. This finding received further support when an evaluation of a cohort of elderly individuals with family history of AD identified an inverse relationship between olfaction strength and the amount of amyloid in the CSF: patients with more amyloid in the CSF had a weaker sense of smell [48]. This finding makes sense considering the entorhinal cortex both receives over half of all the olfactory inputs [49] while also receiving much of the amyloid burden early in the course of AD. At this time, though, the authors of these studies advocate that olfaction tests not be used as a way of diagnosing AD but instead function as a tool for further studies on AD prevention.

In addition to research into the types of NPS that present in AD, there have also been several studies focusing on the relationship between the presence of a certain NPS and the severity of the AD. Unfortunately, there does not appear to be consistent agreement across these studies about when these symptoms first develop. It is not even entirely certain if the severity of the NPS will parallel the severity of AD, as some studies have found that certain NPS will actually worsen in the middle stages of the dementia and then remit as cognitive decline worsens. The increased prevalence of apathy even in very mild AD is one of the more consistent findings, and it is recommended that physicians keep AD in their differential when confronted with AD [50, 51].

Imaging Studies

While a thorough history with appropriate cognitive tests remains essential for an accurate diagnosis of AD, imaging studies like CT and MRI scans have diagnostic utility by detecting neuroanatomical changes associated with atrophy in AD. MRI is the preferred test, but both MRI and non-contrast-enhanced CT scans can be helpful in clinical practice [52]. Since the hippocampus suffers damage early in AD, hippocampal atrophy has been proposed as a structural biomarker. Support for this hypothesis comes from an observation between an association of decreased hippocampus volume and heightened risk of AD [53]. Reductions in hippocampal volume are also associated with schizophrenia and major depressive disorders [54••, 55]—two psychiatric disorders that can be confused for AD—so any diagnosis of AD based on small hippocampi would be premature.

Yet there are subtle differences in the hippocampal changes between these disorders. Analyses of brain MRIs in patients with MDD and AD showed a reduction in the volume of the dentate gyrus (a part of the hippocampus that contributes to formation of new episodic memories) in patients with MDD. Patients with AD had normal dentate gyrus volumes but reduced volumes in the hippocampus’s CA1 region that connects to the entorhinal cortex [11•]. A comparison between the brain MRIs of individuals with “depressive pseudodementia (DPD)” and those with AD showed a relationship between left hippocampal volume and cognitive test results in patients with AD but not with DPD [37•]. A smaller left hippocampal volume was also seen in AD patients with the APOΣ4 gene compared to AD patients without the gene. This further reinforces a potential for left hippocampal volume as a structural biomarker because the APOΣ4 gene is the most significant genetic risk factor for later developing AD [56]. Interestingly, while AD may be associated with a decrease in left hippocampal volume, schizophrenia may be associated with a decrease in right hippocampal volume [57].

Other imaging techniques such as functional MRI (fMRI) and single-photon emission chromatography (SPECT) scans, both of which measure cerebral blood flow rather than structure, may also have use in diagnosing AD. A study of SPECT scans of patients with either depression or cognitive disorders including AD found that patients with depression had more cerebral perfusion than those with cognitive disorders, and that SPECT scans could distinguish depression from dementia with 86% accuracy [58•].

Positron emission tomography (PET) scans may provide the most accurate imaging modality for diagnosing AD. By detecting changes in tissue metabolism of glucose, PET scans have illustrated temporoparietal glucose hypometabolism in AD, with lower metabolism correlating with more severe dementia [59]. More recently, scientists have focused on the capacity for PET scans to detect the presence of brain amyloid. One study found that a combination of MRI with PET scans to detect amyloid in patients with mild cognitive impairment yielded accuracy of 76% when predicting the transition from mild cognitive impairment to AD [60].

In order to comprehensively evaluate the role of brain amyloid scanning in patient outcomes, the Imaging Dementia-Evidence for Amyloid Scanning (IDEAS) study was launched in 2016. The interim results, presented at the 2017 Alzheimer’s Association International Conference (AAIC), showed that amyloid plaques were detected in just 54.3% of individuals with MCI and 70.5% of patients with dementia diagnoses [61]. Changes in medical management stemming from the PET findings occurred in 67.6% of patients. PET scans, as well as SPECT scans, have mostly been used in research settings, but should future studies involving these imaging methods provide similar results, their implementation in clinical care may prove practical [62].

Lab Studies

Like imaging studies, lab tests hold promise for providing objective, quantifiable evidence of AD that would assist in distinguishing it from psychiatric disorders. In fact, confirmation of AD currently relies on detection of amyloid beta plaques, tau protein, and/or phosphorylated tau in the cerebrospinal fluid (CSF), with results 95% sensitive and 85% specific for AD [63••]. Blood draws are much simpler and safer than lumbar taps, so researchers have begun searching for serum biomarkers. The prospect of an amyloid blood test is promising, with one meta-analysis finding that amyloid could be measured in the blood and predict cognitive decline [64] and another trial finding that levels of amyloid in the blood correlated with CNS amyloid as diagnosed on PET scan [65]. A more recent case-control study did not find a relationship between amyloid blood levels and development of dementia, though [66]. There is also support for inflammatory markers, including TNF-alpha and C-reactive protein (CRP), to act as biomarkers for developing dementia [67, 68]. Such findings reinforce the hypothesized role of inflammation in the pathogenesis of AD, but considering the sheer amount of pathologies associated with increased inflammation—including depression—it would be worthwhile to determine the specificity of these biomarkers for AD.

Micro-RNAs (miRNAs) are also getting attention as biomarkers for AD, with several studies illustrating the downregulation of several miRNAs involved in regulation of apoptosis and transcription of amyloid, as well as the upregulation of other miRNAs in AD patients [69, 70]. Investigation into miRNAs pertinent to AD has made way for new studies focusing on the genes that are eventually transcribed into the miRNA [71•]. Though not currently used as routinely in clinical practice as imaging or blood tests, gene sequencing continues to attract interest for its capacity to make accurate diagnoses early enough for interventions to be effective; further studies into this technology for diagnosing AD should occur as it becomes more cost-effective and practical.

Treatment of NPS in AD

Antipsychotics

The search for effective treatments of NPS in AD has been as formidable as the search for a treatment of AD itself, but existing psychiatric medications may have some benefit. Antipsychotics have received perhaps the most attention for their capacity to manage agitation and psychosis in patients with AD, even serving as the focus of the seminal clinical study, the Clinical Antipsychotic Trials of Intervention Effectiveness—Alzheimer’s Disease (CATIE-AD). This 2006 double-blind trial randomly assigned olanzapine, quetiapine, risperidone, or placebo to AD patients with psychosis or aggressive behaviors, measuring time to discontinuation as a primary outcome with clinical improvement in symptoms as a secondary outcome. Though there was no significant difference in the primary outcome between any of the agents, olanzapine and risperidone contributed more to extrapyramidal symptoms (EPS) compared to quetiapine (2%) or placebo (1%) [72]. Numerous other studies that have focused on the use of antipsychotics in managing agitation have all reached a similar conclusion: antipsychotics have at best a modest benefit in treating psychotic symptoms in AD but there are poor safety outcomes (such as increased risk for cerebrovascular events and falls) [73], and therefore should not be used routinely for treatment [74].

Yet the aggression from AD patients can pose such a hazard to the health of themselves and others that the benefits of immediate treatment with antipsychotics often do outweigh the risks. This is particularly true in the hospital setting, which has sharps and other hazards. When antipsychotics are indicated, atypical antipsychotics at low doses are recommended because there is a dose-dependent increase in mortality risk. Quetiapine is often used in AD patients because it is safer than other antipsychotics [75, 76], and in some studies, quetiapine did not yield any increase in mortality [77, 78]. Unfortunately, quetiapine has frequently not been found to be as effective for managing psychosis or aggression as olanzapine or risperidone [79, 80], two antipsychotics with a higher mortality risk. Aripiprazole is an atypical antipsychotic that has undergone less research but there is nevertheless evidence for its efficacy in managing NPS in AD [81, 82••]. A 2007 double-blind trial focusing on NPS changes in response to fixed doses of aripiprazole showed that a dose of 10 mg a day yielded significant reductions in aggression, though there remained a dose-dependent increase in cerebrovascular events [83]. Aripiprazole also has a higher risk of akathisia and insomnia compared to quetiapine.

Given the lack of clear indications for antipsychotics in patients with dementia, the American Psychiatric Association released practice guidelines in 2016 for their appropriate prescriptions. The guidelines identified risperidone as an especially effective treatment for agitation and aripiprazole as useful for the management of psychological symptoms related to dementia. Haloperidol should not be used as a first-line treatment. The guidelines reaffirm that the effect sizes remain small and potential side effects are significant, so antipsychotics should be prescribed only after a firm diagnosis of psychosis in dementia with no benefit from non-pharmacologic treatments (discussed later on) [84, 85••].

Though not specifically mentioned in the guidelines, quetiapine is often used first-line based on clinical experience because of its lower side effect profile while apparently still having some capacity to manage psychosis. Olanzapine will often be used next if agitation is not controlled with quetiapine both because of the evidence for its increased efficacy compared to quetiapine and the availability of an intramuscular (IM) formulation should the patient not be able to take medications orally (a common feature among agitated patients). Some sources state that olanzapine has fewer side effects and a lower mortality risk compared to risperidone, particularly with respect to cerebrovascular events including strokes [86•]. Haloperidol will typically be used in patients who are medically admitted because of its intravenous (IV) formulation, but this is at most a third-line agent for agitation because of its increased mortality risk. The indications, side effects, and dosing for the most popular antipsychotics is summarized in Table 2. It should be addressed that these doses apply to geriatric patients and not solely patients with AD, but the table should still apply to the vast majority of patients with AD who are over age 65.

Table 2 Comparing and contrasting the most popular neuroleptics for managing agitation and psychosis in geriatric patients, including those with Alzheimer’s dementia. The included information is derived from articles cited in the main text, specifically articles 69–83. The suggested dosages were verified with the online database Facts and Comparisons (St. Louis, MO: Wolters Kluwer Health, Inc. http://fco.factsandcomparisons.com.proxy.lib.umich.edu/lco/action/home. Accessed July 23rd, 2018). EPS, extrapyramidal symptoms; IM, intramuscular
Full size table

Acetylcholinesterase Inhibitors and Memantine

As the medications with FDA approval for AD treatment, cholinesterase inhibitors (galantamine, rivastigmine, and donepezil) and the NMDA receptor antagonist memantine have likewise been studied for their potential to manage NPS. Any benefit from these medications in managing NPS could theoretically prevent the need for additional drugs to manage psychiatric symptoms.

A 2013 meta-analysis of trials studying the impact of the acetylcholinesterase inhibitors (AChE-Is) on behavioral symptoms in patients with AD found that donepezil, galantamine, and memantine (but not rivastigmine) all had some efficacy in managing NPS [87]. A 2014 meta-analysis showed that only donepezil at doses of 10 mg daily and galantamine 24 mg daily had positive effects on NPS, and an even more recent meta-analysis did not find any significant benefits from donepezil or memantine but reaffirmed earlier results indicating the efficacy of galantamine [88]. Both meta-analyses cautioned that AChE-Is are associated with adverse effects—including gastrointestinal symptoms, dizziness, and infections—and an increase in mortality. This risk is lower than that of antipsychotics, but one additional drawback with AChE-Is is that their use is primarily indicated for patients with mild-to-moderate AD. It thus remains unclear whether they would be able to manage the extreme agitation mostly seen in advanced AD as effectively as antipsychotics.

Memantine, unlike the AChE-Is, is more effective for treating severe AD than it is for mild and moderate forms of the disease. Select analyses have found that memantine at doses between 5 and 20 mg a day could treat behavioral symptoms without causing as many side effects as AChE-Is [89]. Yet these same meta-analyses have also found conflicting opinions on its effectiveness. More relevant trials would need to occur before it could be implemented routinely into clinical practice.

Antidepressants, Mood Stabilizers, and Stimulants

Several other classes of medications have been evaluated for the treatment of NPS in AD. Antidepressants such as SSRIs should theoretically alleviate the depression seen in AD, but so far, antidepressants have overall failed to yield significant clinical benefits [90]. Of the antidepressants that have been studied, citalopram has arguably the most evidence supporting its use in AD patients. At a target dose of 30 mg a day, citalopram reduced agitation and hallucinations but also correlated with an increase in cognitive decline and nighttime behavior disorders [91, 92].

Like antidepressants, anticonvulsant mood stabilizers such as sodium valproate lack firm support for their use in managing neuropsychiatric symptoms [93]. Lithium has garnered interest for its observed capacity to halt the amyloid formation and tau hyperphosphorylation that drives AD even at doses too small to cause significant side effects [94]. There is a dearth of evidence for its use in managing NPS in AD, though.

There have also been studies illustrating the noteworthy ability of methylphenidate for managing apathy in AD [95] with the Alzheimer’s Disease Methylphenidate Trial (ADMET) illustrating reduced apathy in AD patients with methylphenidate at 20 mg a day [96]. A recent meta-analysis did not identify any one medication having a significant effect on improving apathy [97], though the results of the upcoming ADMET2 and its larger population size could amend that conclusion [98]. Other proposed treatments for NPS in AD requiring more studies before being put into practice include omega-3 unsaturated fatty acid supplementation [99], calcium channel blockers [100], and electroconvulsive therapy (ECT) [101].

Non-pharmacological Interventions

Considering the numerous side effects associated with the medications discussed above and no universal agreement on when such medications should be used, non-pharmacological therapies could offer clinical benefits without the side effects and risks that come with pharmaceuticals. Behavior management techniques decreased agitation as much as haloperidol, even in advanced AD, with fewer side effects [102]. Meanwhile, regular exercise and cognitive behavioral therapy reduced depression in patients with early dementia [103, 104]. Music therapy likewise reduced depression and anxiety, but not agitation [105].

One psychotherapy modality that appears to be specifically helpful for patients with dementia is reminiscence therapy. Centered on a discussion of past activities and events with the use of aids such as photographs and music from the pertinent area to help realize the memory, reminiscence therapy has some benefit in improving mood and cognition [106, 107]. Aroma-massage therapy, which involves caregivers delivering a massage with oils that have anxiolytic and sedating properties (lavender and lemongrass are two of the most commonly used oils), has also been studied for its therapeutic benefits in treating NPS in AD [108]. A 2016 cohort study actually found that aroma-massage therapy could effectively manage agitation and treat depression while reminiscence therapy could not, possibly because of the individual focus inherent with aroma-massage therapy [109]. This result is encouraging because aroma-massage therapy is more intuitive than reminiscence therapy or CBT, and would likely be easier to apply in busy healthcare settings.

Ultimately, these therapies hold promise as treatments for NPS in AD patients, yet they remain underutilized. The Describe, Investigate, Create, Evaluate (DICE) approach was developed in an effort to increase the implementation of non-pharmacologic interventions. In applying the DICE approach to a specific NPS, the caregiver will first describe to providers the specific problem at hand (i.e., aggression, paranoia, depression) in as much detail as possible. The providers will then investigate the potential causes and contributors to these behaviors. Usually, this will involve medication reconciliation and labs, but it is also important to evaluate the environment and the integrity of the relationship between the caregiver and the patient as contributors to the problem symptoms. If the patient can verbally communicate, it is helpful to speak with him or her to understand the reasons driving the behavior as well as any psychological issues leading to the behaviors (see Table 3 for one possible method of looking for underlying causes of NPS).

Table 3 Potential method of investigating and managing etiologies for NPS in patients with dementia, organized into an AEIOU acronym
Full size table

Once all of the possible etiologies have been isolated, the providers and caregiver work together to create a plan for addressing the causes. The goal would be amelioration—or at least minimization—of the target NPS. Only through evaluation of the patient could it be determined if this goal was reached. If the issue remains following the DICE approach, then psychotropic medications would be appropriate. Though ideally psychotropics should come second to the non-pharmacologic interventions espoused in the DICE approach, the clinicians who helped develop this treatment method acknowledged that psychotropics should be used as first-line treatments when the patient with dementia is posing a safety risk to him/herself or others [110••].

Conclusion

In terms of mortality, morbidity, and cost, Alzheimer’s disease remains one of the most destructive illnesses in the world and the extent of its damage is only increasing. Part of the impact of AD stems from its neuropsychiatric symptoms (NPS). The etiology of these symptoms is somewhat unclear; hypotheses include atrophy of the hippocampus and other neurologic structures involved in regulating mood and behavior, a deficit of monoamines, and a result of underlying inflammation that may also be causing AD. These NPS, including apathy, depression, and agitation, can cause confusion between AD and psychiatric disorders such as major depressive disorder and schizophrenia. A keen understanding of the specificities defining the NPS in AD, as well as use of the NPI, can help physicians clinically distinguish the disorders. Imaging and lab studies, including PET scans and blood tests to check for amyloid presence, are beginning to be recognized for their capacity to more objectively diagnose AD when clinical diagnosis remains ambiguous. And while antipsychotic medications and cholinesterase inhibitors remain the most well-established treatments for agitation in AD, researchers continue to look into the safest and most effective uses of these medications along with developing insights into other psychiatric medications—including antidepressants, stimulants, and behavioral therapy—to target the apathy and depression relatively untouched by antipsychotics. The benefits of non-pharmacological interventions are also become acknowledged and efforts have developed to incorporate these treatments into clinical practice.

For all of these advances, there are a number of questions that would need to be addressed before any study or treatment could be implemented into routine clinical practice. As mentioned earlier in the review, there is not a clear score in the NPI that can serve as a cutoff between pseudodementia and AD, and there is also limited evidence for the utility of non-pharmacological treatments in patients with severe AD. Despite numerous studies on the epidemiology of NPS in AD, there is still no consistent conclusion on the time-course of NPS and thus no definitive understanding of how the NPS can prognosticate the illness. The imaging and lab studies discussed have primarily been used in a research setting, so it is uncertain if these cutting-edge technologies can be affordably used outside of academic institutions, or how the sensitivity and specificity of these studies would compare to a thorough and predictably more cost-effective clinical exam. And with the growing evidence supporting a view of AD as an inflammatory illness, perhaps ongoing studies should assess the ability of NSAIDs and other anti-inflammatory agents to reduce NPS rather than restricting a focus on psychiatric medications.

Identifying such topics should not undermine the accomplishments made in the studies discussed but instead reinforce the challenge of diagnosing and managing AD and its component NPS, and offer direction for the future studies that could finally offer effective patient care.