Effects of Alzheimer’s Dementia Treatment on Agitation and Aggression: A Systematic Review

Abstract

Introduction—Aim: Agitation and aggression are common symptoms in people with Alzheimer’s disease and other dementias. This systematic review outlines the current evidence for the effect of antidementia treatments on agitation and aggression in patients with Alzheimer’s disease. Methods: The literature search was performed by manually checking articles published since 2000 in the following databases: PubMed, ResearchGate, and Google Scholar, following the PRISMA guidelines. Results: Nineteen double blinded placebo-controlled trials were included. Treatment with galantamine seems to provide more credible evidence for treating or preventing agitation/aggression. Studies on memantine often presented with an improvement of the neuropsychiatric inventory but not specifically on agitation/aggression. The trials on donepezil and rivastigmine either did not include enough information on agitation/aggression or did not provide compelling results. The incidence of agitation as an adverse event was not higher in antidementia treatments compared to placebo. Conclusions: Agitation and aggression are common symptoms of Alzheimer’s disease and have a significant impact on the patient, caregiver, and the broader healthcare system. The current literature lacks robust evidence on which of the antidementia treatments could be used to manage or prevent agitation and aggression in Alzheimer’s disease. In most included studies, no specific scores that assess those symptoms were used. Future research that specifically focuses on different disease phenotypes and behavioral profiles to enhance and facilitate the management of these symptoms is needed.

1. Introduction

The term dementia incorporates a spectrum of disorders about thinking and cognition, characterized mainly by a deterioration of the ability to learn and memorize, in language, in executive functions, and in attention, as well as in perceptual motor and social cognition [1,2]. Alzheimer’s disease is the most common type of dementia, which accounts for 60% to 80% of all dementia cases [3].

Agitation and aggression were fourfold more frequently observed in patients with dementia compared to same age participants with no dementia in a study of over 5092 participants [4]. Agitation and mood disturbance were linked not only to an earlier onset age [5] but also to advanced stages of the disease [6,7]. They are connected to worse outcomes, earlier in-patient admission, and poorer quality of life and are related to an increase in the cost of care [8]. They pose a great burden for families and caregivers [9,10]. The recognition and treatment of agitation and aggression are crucial parts of the management of patients with AD. However, limited work has systematically addressed agitation and aggression in patients with AD using valid questionnaires. The Cohen–Mansfield Agitation Inventory (CMAI) is a reliable and robust instrument [11] for assessing those symptoms and includes questions about physical and verbal aggressive behavior [12]. The Agitated Behavior in Dementia (ABID) scale is another specific tool for the assessment of those symptoms and was created to assess agitated behaviors in mildly to moderately demented patients who live at home and/or with their families [13]. Both scales were found to have high validity and had the advantage that they can be administered by a technician or even a caregiver instead of a clinician [14].

As there is no gathered information, to the best of our knowledge, about which treatment is suitable for patients with AD, when agitation/aggression is an issue, we aimed to address that gap by gathering and analyzing the effects, both positive and negative, of pharmacological treatments for dementia on agitation and aggression in patients with AD. Initially, agitation is evaluated as a potential adverse effect of antidementia medication. Secondly, NPI scores are collected to indicate overall behavioral changes in patients with dementia. Finally, information is gathered on which antidementia medication could lead to a positive change in agitation/aggression levels.

2. Methods

A systematic literature review was performed within the following databases: PubMed, ResearchGate, and Google Scholar.

Search terms were Alzheimer’s dementia, medication, donepezil, galantamine, rivastigmine, memantine, cholinesterase inhibitor, agitation, and aggression. The search focused on English- and German-language papers published since 2000 to ensure relevance to contemporary discussions, as we wanted to include important studies on this field of the early 2000s.

All included articles’ literature data were manually checked for further studies. This review was performed based on the 2020 updated Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [15].

Data about the number of patients, dose of the agent, CMAI, ABID, NPI scores, and incidence of agitation were extracted.

2.1. Quality Assessment

The risk of bias for each included study was assessed according to the GRADE guidelines [16]. Confounding, attrition, selection, and reporting biases were evaluated separately for each one of the included studies.

2.2. Inclusion and Exclusion Criteria

Only randomized, double-blinded, and placebo-controlled trials were included. A trial was included when patients were diagnosed with AD according to the National Institute of Neurological and Communicative Disorders and Stroke–Alzheimer’s Disease and Related Disorders Association Criteria [17] and involved at least four weeks of treatment with either memantine or the following cholinesterase inhibitor: rivastigmine, galantamine, or donepezil. Neuropsychiatric outcomes were reported.

Duplicates and trials that did not include any information on agitation/aggression as well as trials about other types of dementia or trials that compared antidementia medication to antipsychotics or trials that did not include placebo comparison were excluded. Non-human studies were also excluded.

After identifying the relevant titles in all three databases, a total of 495 titles or abstracts in the literature search were initially examined to identify the eligibility for inclusion. In detail, 463 published papers were removed because they were not compatible with the inclusion criteria. Of the 32 trials screened for the inclusion criteria, 9 reported no data on agitation/aggression or NPI score, or the Alzheimer’s diagnostic criteria were not mentioned.

3. Results

Twenty-three studies fulfilled the inclusion criteria, but one was a duplicate, and one report was excluded because of the observational study design, and two reports were excluded because the comparison was against other agents but not against placebo, leaving nineteen trials studied for this review (see Figure 1).

Among the included studies, seven focused on donepezil, one on rivastigmine, eight on memantine, and three on galantamine. While all the studies provided data related either directly to agitation/aggression or to NPI, only a small number included data on CMAI, and none of them included data on ABID. These limited data did not allow for a meta-analysis on agitation/aggression.

3.1. Agitation as an Adverse Event of Antidementia Treatment

Ten studies including a total of 4926 patients provided data on agitation as an adverse effect of antidementia treatments. In these studies, patients with both mild and severe dementia were included, as indicated by their MMSE mean scores at the baseline, which varied from 19.6 (moderate impairment) to 7.5 (severe impairment) (see

Table 1. Studies for antidementia treatments stating agitation as an adverse event. The included studies were listed in chronological order. The dose of medication listed is the last dose after titration. MMSE was not provided in one study. The trials are listed firstly by agent and then chronologically.

Trial	Substance	Dose	Nr. of Patients	Incidence of Agitation on Treatment with Medication (%)	Incidence of Agitation on Treatment with Placebo (%)	MMSE at Baseline (Mean)
[21]	Galantamine	8, 16, or 24 mg/d	978	10%	9.4%	17.7
[22]	Galantamine	16 mg or 24 mg/d	971	6.5%	7%	17.95
[19]	Galantamine	24 mg/d	386	6.1%	0.8%	19.6
[23]	Memantine	20 mg/d	252	18%	32%	7.7
[24]	Memantine	20 mg/d	404	9.4%	11.9%	9.9
[25]	Memantine	20 mg/d	403	7.5%	5.9%	No data
[26]	Memantine	20 mg/d	350	9%	14%	10
[27]	Memantine	20 mg/d	470	1.6%	4.6%	18.6
[20]	Memantine	20 mg/d	369	8.2%	3.7%	11.9
[28]	Donepezil	10 mg/d	343	6.3%	6%	7.5

). In most trials, the incidence of agitation as an adverse event was not higher with antidementia treatments compared to the placebo (see Table 2). However, in two studies—one on galantamine [19] and one on memantine [20]—agitation occurred slightly more often in the treatment group than in the placebo group. The mean MMSE scores in both studies indicate a moderate level of cognitive impairment, and the incidence of agitation as an adverse event remained below 10%.

3.2. NPI Scores After Treatment

The questionnaires targeting the assessment of agitation/aggression in dementia, such as CMAI and ABID, are not commonly used in large-scale studies. However, the NPI does include a specific subitem on agitation/aggression. Therefore, studies that reported NPI scores were also collected and analyzed to provide an indirect assessment of overall behavioral changes after treatment.

A total of 18 studies including 10,339 patients are reported below (see Table 2). The mean baseline NPI in 13 out of 18 studies was ≥12, showing a moderate score of neuropsychiatric symptoms. In only two studies, the mean baseline NPI was ≥24, indicating a significant prevalence of neuropsychiatric symptoms. In one of those two studies, which was on memantine, a substantial improvement of neuropsychiatric symptoms was proven compared to placebo.

NPI was significantly lower in the observed period in 7/20 studies, 4 of which were on memantine, whereas there was no difference in the rest of them (see Table 3). A reduction in NPI may indicate improvements in symptoms other than agitation. Studies that included a subitem investigation of NPI domains are discussed below (s. paragraph: Agitation after treatment). Mean MMSE scores vary widely in the included studies, from 5.8 (severe impairment) to 21.1 (mild impairment) (see Table 2).

Table 2. Studies of antidementia medication that included NPI scores. This table presents an overview of the studies, including the medication used, dosage, number of patients, baseline neuropsychiatric inventory (NPI) scores, and any statistically significant differences in NPI between the medication and placebo groups after treatment, as well as the Mini-Mental State Examination (MMSE) scores. The trials are listed firstly by agent and then chronologically. * n.s: not significant.

Trial	Substance	Dose	Nr. of Patient	NPI Baseline (Mean)	NPI Statistical Change Observed Between Drug and Placebo	MMSE at Baseline (Mean)
[21]	Galantamine	8, 16, or 24 mg/d	978	11.9	* p < 0.05	17.7
[19]	Galantamine	24 mg/d	386	9.2	n.s.	19.6
[22]	Galantamine	16 mg or 24 mg/d	971	11.2	n.s.	17.95
[29]	Donepezil	5 mg/d or 10 mg/d	290	19.5	* p = 0.0083	11.72
[30]	Donepezil	10 mg/d	134	14.3	* p = 0.02	21.1
[31]	Donepezil	10 mg/d	208	21	n.s.	14.4
[32]	Donepezil	10 mg/d	2488	19	n.s.	5.8
[28]	Donepezil	10 mg/d	343	22.7	n.s.	7.5
[33]	Donepezil	10 mg/d	272	23.6	n.s.	8.2
[23]	Memantine	20 mg/d	252	21.4	n.s.	7.7
[24]	Memantine	20 mg/d	404	13.4	* p = 0.01	9.9
[25]	Memantine	20 mg/d	403	11.5	* p = 0.01	No data
[26]	Memantine	20 mg/d	350	20.3	n.s.	10
[34]	Memantine	20 mg/d	149	36.1	* p < 0.001	7.3
[27]	Memantine	20 mg/d	470	No data	n.s.	18.6
[35]	Memantine	28 mg/d	677	17.2	* p = 0.005	10.6
[20]	Memantine	20 mg/d	369	30.9	n.s.	11.9
[36]	Rivastigmine	9.5 mg/d (patch), 17.4 mg/d (patch), 12 mg/d (capsule)	1195	13.9	n.s.	16.6

Table 2. Studies of antidementia medication that included NPI scores. This table presents an overview of the studies, including the medication used, dosage, number of patients, baseline neuropsychiatric inventory (NPI) scores, and any statistically significant differences in NPI between the medication and placebo groups after treatment, as well as the Mini-Mental State Examination (MMSE) scores. The trials are listed firstly by agent and then chronologically. * n.s: not significant.

Trial	Substance	Dose	Nr. of Patient	NPI Baseline (Mean)	NPI Statistical Change Observed Between Drug and Placebo	MMSE at Baseline (Mean)
[21]	Galantamine	8, 16, or 24 mg/d	978	11.9	* p < 0.05	17.7
[19]	Galantamine	24 mg/d	386	9.2	n.s.	19.6
[22]	Galantamine	16 mg or 24 mg/d	971	11.2	n.s.	17.95
[29]	Donepezil	5 mg/d or 10 mg/d	290	19.5	* p = 0.0083	11.72
[30]	Donepezil	10 mg/d	134	14.3	* p = 0.02	21.1
[31]	Donepezil	10 mg/d	208	21	n.s.	14.4
[32]	Donepezil	10 mg/d	2488	19	n.s.	5.8
[28]	Donepezil	10 mg/d	343	22.7	n.s.	7.5
[33]	Donepezil	10 mg/d	272	23.6	n.s.	8.2
[23]	Memantine	20 mg/d	252	21.4	n.s.	7.7
[24]	Memantine	20 mg/d	404	13.4	* p = 0.01	9.9
[25]	Memantine	20 mg/d	403	11.5	* p = 0.01	No data
[26]	Memantine	20 mg/d	350	20.3	n.s.	10
[34]	Memantine	20 mg/d	149	36.1	* p < 0.001	7.3
[27]	Memantine	20 mg/d	470	No data	n.s.	18.6
[35]	Memantine	28 mg/d	677	17.2	* p = 0.005	10.6
[20]	Memantine	20 mg/d	369	30.9	n.s.	11.9
[36]	Rivastigmine	9.5 mg/d (patch), 17.4 mg/d (patch), 12 mg/d (capsule)	1195	13.9	n.s.	16.6

3.3. Results for Agitation/Aggression After Treatment

Agitation/aggression as a result after treatment, either primary or secondary outcome, was only discussed in a few studies, and the CMAI score was also a very rare finding, which does not allow for a meta-analysis. However, all relevant findings were gathered and are discussed below in separate paragraphs for each agent.

3.3.1. Donepezil

One study [30] provided an analysis of specific subitems of NPI and reported a significant change in agitation in week 12 compared to the baseline and a significant decline in NPI in favor of donepezil compared to the placebo in week 24 (p = 0.02).

Another double blind study [31] of patients with AD, who lived in caregiving units whose neuropsychiatric symptoms were a primary outcome, reported that although no significant differences in NPI between the placebo and donepezil were observed, the agitation/aggression domain revealed a statistically significant difference in favor of donepezil (p = 0.044). More specifically, patients who had agitation/aggression at baseline which improved after the intervention was 46% (29/63) for the placebo group and 67% (46/69) for donepezil (p = 0.017).

A reduction in CMAI, but not a significant one, was also detected in a study about donepezil compared to the placebo [33].

Another study demonstrated an improvement in patients who received donepezil versus the placebo for most behavioral symptoms, such as anxiety, apathy/indifference, and irritability/lability, where the more significant improvement was observed, while for agitation, the outcome was not so notable [37].

3.3.2. Rivastigmine

In the conducted literature search, there were no placebo-controlled studies for rivastigmine referring specifically to agitation/aggression. In a trial of mild to moderate AD patients, the group that received rivastigmine patch monotherapy presented a reduction in non-aggressive agitated behaviors compared to baseline, but there was no comparison to the placebo [38].

3.3.3. Memantine

The efficacy of memantine on agitation in AD patients in a study by Fox et al. [34] proved no noteworthy results in CMAI compared to placebo at week 6 or 12, although there was a greater improvement in the total NPI score. However, the post hoc analysis favored memantine at week 6, but not at week 12, in NPI agitation/psychosis symptom grouping. This grouping of both agitation and psychosis may explain the difference reported on CMAI.

In another study on memantine [20], in which the CMAI score in patients with moderate to severe AD was assessed, there was no significant difference detected in the observed period of the study, although there was a clinical but not statistically significant improvement in NPI on the memantine group versus the placebo group.

3.3.4. Galantamine

A previous meta-analysis that used data from three extended placebo-controlled studies on galantamine found a statistically significant improvement in agitation at the studies’ endpoints for patients treated with galantamine when set side by side with the placebo (p = 0.05) [39]. Additionally, 57% of patients on galantamine who had agitation at the baseline showed a 30% or greater reduction in that symptom, compared to 45% of those on the placebo, an improvement that was also statistically significant (p = 0.004). This meta-analysis strongly indicated that galantamine is an effective therapeutic approach for managing behavioral and psychiatric symptoms, particularly agitation, in patients with mild-to-moderate Alzheimer’s disease.

3.4. Results of the Quality Assessment: The Risk of Bias

The studies included in the present review were individually evaluated for the risk of bias. Since all the studies had a double-blind, placebo-controlled design, they are generally considered high-quality evidence. Selection bias was generally rated as low due to the large sample sizes and multicenter design or as moderate in cases of single-center studies (see Table 3). Some studies exhibited confounding bias due to the concurrent use of other medications, such as antidepressants and neuroleptics, which could influence behavior, potentially influencing the results. Regarding attrition bias, most of the studies were seen as moderate (see Table 3). In those studies, the dropout range was between 10 and 20% for both the medication and placebo. In one study, there were no data on patients’ drop-out rate. In almost all studies, reporting bias was rated as low, as all data on patients and scores were provided, except for one study, in which a part of the important data was discussed but not provided in the publication.

Table 3. Assessment of the risk of bias of the included studies. Selection, confounding, attrition, and reporting biases were assessed and rated as low, moderate, and high, according to the GRADE guidelines. When there was no information, bias was stated as unknown. Drugs are listed by substance and then chronologically. Green is for low, yellow for moderate and orange for high risk of bias.

References	Selection Bias	Confounding Bias	Attrition Bias	Reporting Bias
[21] Galantamine	Low	Low	Moderate	Low
[19] Galantamine	Low	Low	Moderate	Low
[22] Galantamine	Low	High	Moderate	Low
[29] Donepezil	Low	High	Low	Low
[31] Donepezil	Low	High	Moderate	Low
[37] Donepezil	Low	High	Low	Low
[30] Donepezil	Low	High	Low	Low
[32] Donepezil	High	High	Unknown, no information	Low
[28] Donepezil	Low	High	Moderate	Low
[33] Donepezil	Low	High	Low	Low
[23] Memantine	Low	Moderate	Moderate	Low
[24] Memantine	Low	High	Moderate	Low
[25] Memantine	Low	Moderate	Low	Moderate
[26] Memantine	Low	Moderate	Moderate	Low
[27] Memantine	Low	Moderate	Moderate	Low
[34] Memantine	Moderate	Low	Low	Low
[20] Memantine	Low	High	Moderate	Low
[35] Memantine	Low	Moderate	Low	Moderate
[36] Rivastigmin	Low	Moderate	Moderate	Low

Table 3. Assessment of the risk of bias of the included studies. Selection, confounding, attrition, and reporting biases were assessed and rated as low, moderate, and high, according to the GRADE guidelines. When there was no information, bias was stated as unknown. Drugs are listed by substance and then chronologically. Green is for low, yellow for moderate and orange for high risk of bias.

References	Selection Bias	Confounding Bias	Attrition Bias	Reporting Bias
[21] Galantamine	Low	Low	Moderate	Low
[19] Galantamine	Low	Low	Moderate	Low
[22] Galantamine	Low	High	Moderate	Low
[29] Donepezil	Low	High	Low	Low
[31] Donepezil	Low	High	Moderate	Low
[37] Donepezil	Low	High	Low	Low
[30] Donepezil	Low	High	Low	Low
[32] Donepezil	High	High	Unknown, no information	Low
[28] Donepezil	Low	High	Moderate	Low
[33] Donepezil	Low	High	Low	Low
[23] Memantine	Low	Moderate	Moderate	Low
[24] Memantine	Low	High	Moderate	Low
[25] Memantine	Low	Moderate	Low	Moderate
[26] Memantine	Low	Moderate	Moderate	Low
[27] Memantine	Low	Moderate	Moderate	Low
[34] Memantine	Moderate	Low	Low	Low
[20] Memantine	Low	High	Moderate	Low
[35] Memantine	Low	Moderate	Low	Moderate
[36] Rivastigmin	Low	Moderate	Moderate	Low

4. Discussion

4.1. Agitation as an Adverse Event

The first outcome of the current systematic review suggests that agitation is not a significant adverse event of antidementia medication, as the incidence is very similar to the one of the placebo. In most studies, agitation was among the most frequently reported adverse events of both the medication and placebo, which can be explained by the fact that agitation probably appears in the course of the disease regardless of the medication. Only two studies—one on memantine [20] and another on galantamine [19]—reported higher agitation rates in the treatment group compared to the placebo group. However, in the memantine study, agitation appeared as a worsening of a pre-existing condition, occurred in combination with acute medical problems, or emerged within the first four weeks of treatment. In the galantamine study, most cases of agitation were observed during the dose escalation phase, implying that a gradual increase in dosage might help minimize this side effect. Overall, those findings imply that agitation is not a major concern as an adverse event for the antidementia medication.

4.2. Behavioral Changes After Treatment

The second outcome of the study was the significant improvement of NPI scores after treatment with antidementia medication in seven studies, four of which were on memantine. Interestingly, in one trial [34], memantine led to an improvement in the NPI scores among patients with significant neuropsychiatric symptoms and severe cognitive impairment, and this was the only study that demonstrated better outcomes in patients with moderate to severe dementia.

Behavioral and neuropsychiatric symptoms can manifest differently among individuals, ranging from a single episode to an alternating pattern with relapses, or even a stable, constant course. As a result, the prevalence of neuropsychiatric symptoms at any given time does not always reflect their overall incidence. This variability complicates research in this area, as assessments typically occur during a single visit rather than a continuous observation over an extended period, making it challenging to capture the full course of these symptoms [40].

4.3. Agitation/Aggression After Treatment

The third outcome of this review favors the use of galantamine in treating agitation in AD patients after a positive outcome in four large-scale studies compared to the placebo. In contrast, only one out of seven studies on donepezil showed convincing evidence that agitation improved in the population treated with the drug compared to the placebo. Data on rivastigmine and memantine do not offer compelling evidence that would reinforce their role in treating or preventing agitation/aggression in AD patients. The benefits of antidementia medications may be more related to delaying the onset rather than improving pre-existing symptoms.

A few studies used the CMAI, but none of the included studies used the ABID, which assesses the caregiver’s level of distress regarding agitating behavior, which could be more trustworthy regarding the fact that those symptoms may not always be present during assessments, and only caregivers can help to assess this valuable information.

The agents against agitation and aggression in AD that are mostly studied and mostly prescribed in acute situations are the typical and atypical antipsychotics. Although they contribute to statistically and clinically significant improvements, their use has been shown to increase the risk of cerebrovascular events and mortality, and furthermore, they have been linked not only to extrapyramidal manifestations but also to metabolic side effects and disorders of cognitive function as well [41].

4.4. Limitations of the Study

Only a few studies reported specific agitation scores using tools like the Cohen–Mansfield Agitation Inventory (CMAI) or the Agitated Behavior in Dementia (ABID) scale, which made a meta-analysis on agitation/aggression not feasible to perform. While the neuropsychiatric inventory (NPI) can reflect general behavioral changes and serve as an indirect measure of agitation, it does not accurately capture the specific levels of agitation. This presents a significant limitation, as agitation may not be a uniform clinical phenomenon and can vary widely among patients.

The studies that were finally included in this review presented patient populations that were diverse in terms of their agitation levels, overall neuropsychiatric symptoms, and cognitive impairment level. In most studies, patients’ concomitant medication included hypnotics and neuroleptics, which can severely influence the results. Those facts complicate the ability to draw reliable conclusions about how effective some treatments on agitation/aggression can be.

Additionally, the variation in Mini-Mental State Examination (MMSE) scores across studies poses another limitation. Previous research has established a correlation between the seriousness of dementia and the level of agitation [6,7], meaning that differences in MMSE scores could influence the outcomes and comparability of the studies.

5. Conclusions and Future Perspectives

Agitation and aggression are symptoms that significantly impact the quality of life of patients with Alzheimer’s disease and their caregivers while also posing a challenge to the healthcare system. This review presented the recent data on the effectiveness of anti-dementia treatments in mitigating these challenging behaviors.

The current literature offers limited evidence on which agents are most effective in managing agitation in patients with Alzheimer’s disease (AD). This lack of robust data makes it challenging for clinicians to suggest the most proper control of agitation in AD patients. As a result, variations in care and treatment outcomes are evident.

In order to address this gap of knowledge, the use of validated agitation assessment tools should be considered. These tools would provide a precise evaluation of the severity and impact of agitation, leading to more accurate comparisons across different studies and treatments. Additionally, there is a lack of evidence on patients who are on no concomitant medication. Future research should include these individuals to draw safe conclusions and isolate the impact of antidementia medication on those symptoms. Future studies should also explore the behavioral profiles specific to AD. This investigation, along with the biological background, could help researchers identify patterns and triggers in the earlier stages of the disease, which may lead to the development of targeted and likely more effective therapies.

Deeper knowledge of the disease phenotypes and behavioral profiles will provide clearer guidance on which agents or interventions are most beneficial for controlling agitation in AD patients, improving both patient outcomes and quality of life.