Non-pharmacologic and pharmacologic treatments for anxiety in long-term care: a systematic review and meta-analysis

Abstract

Background

older adults living in long-term care (LTC) commonly suffer from anxiety symptoms and disorders. We completed a systematic review and meta-analysis to identify efficacious treatments for anxiety symptoms for older adults living in LTC.

Methods

we searched five electronic databases (MEDLINE, Embase, PsycINFO, Cochrane Database of Systematic Reviews and Cochrane Central Register of Controlled Trials) to identify treatments for anxiety that have been trialled in LTC. Included studies had to be randomised trials, include residents of LTC, and measure anxiety symptoms as an outcome.

Results

the electronic search returned 6,617 articles, 519 were reviewed in full text, and 80 were included in the descriptive synthesis. Limited studies were meta-analysed (n = 10) due to differences in described treatment and comparator conditions. Limited clinically relevant evidence supporting the use of pharmacologic treatments for symptoms of anxiety in LTC was identified. Of the treatments trialled, music compared with usual care (standardised mean difference, SMD: −0.82; 95% confidence interval (CI): −1.31, −0.34), music compared with social interaction (SMD: −0.41; 95% CI: −0.72, −0.10) and massage compared with usual care (SMD: −4.32; 95% CI: −7.44, −1.19) were found to improve anxiety symptoms, however, significant heterogeneity was detected in two comparisons.

Conclusions

a range of non-pharmacologic treatments that improved anxiety symptoms were identified for use in LTC. Although limited evidence exists to support the use of particular treatments, most non-pharmacologic treatments were low-risk interventions that may be readily implemented. Further research is required to assess the treatment effect on residents of LTC with anxiety disorders or clinically relevant symptoms at baseline.

anxiety, treatment, long-term care, systematic review, older people

Key Points

• Many low-risk non-pharmacologic treatments for anxiety were identified for use in long-term care.

• There is a need for more research with residents of long-term care with clinically relevant anxiety symptoms.

• Identifying efficacious anxiety treatments supports treatment selection and may reduce suffering in the long-term care population.

Introduction

Global estimates have found up to 58.4% of older adult residents of long-term care (LTC) experience symptoms of anxiety [1]. Anxiety symptoms and disorders are often under-detected and under-treated in residents of LTC [2–4]. Anxiety disorders are those meeting standardised diagnostic criteria, whereas anxiety symptoms are measurable using anxiety detection tools but do not meet standardised diagnostic criteria. Anxiety in the LTC population has been associated with negative outcomes including poorer resident well-being and greater burden to caregivers [5, 6].

There is an immediate need to identify appropriate treatment options given how common and burdensome anxiety is in LTC. Evidence to support treatment approaches for anxiety in the LTC population is unclear and available guidelines do not contain information specific to the LTC population [7]. The lack of evidence available to inform practice represents a major knowledge gap for older adults living in LTC.

Pharmacologic treatments for anxiety such as anxiolytics and benzodiazepines have negative effects on older adults and may not be recommended for long-term use [8–10]. Antidepressants are commonly used to treat anxiety in older adults, however, efficacy and tolerability findings supporting their use in older adult populations may not translate to those in LTC [11]. Non-pharmacologic treatments such as cognitive behavioural therapy or music therapy are of interest for use, but clinical evidence supporting their use for residents of LTC is unclear and thus these programs are not regularly offered in LTC settings [12, 13]. Previous systematic reviews are limited by a lack of focus on LTC residents [14, 15], anxiety being one of many outcomes [16, 17] or a focus on single treatments [18–20].

We addressed these limitations in seeking to determine the most efficacious treatments for symptoms of anxiety specific to residents of LTC. The objective of the systematic review was to identify available treatment options for anxiety symptoms and disorders for residents of LTC.

Methods

Search strategy and selection criteria

The systematic review was registered with PROSPERO (CRD42020182006) and has been reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist (Supplementary Table S1, Supplementary data are available in Age and Ageing online).

Five electronic databases (MEDLINE, Embase, PsycINFO, Cochrane Database of Systematic Reviews (CDSR) and Cochrane Central Register of Controlled Trials (CENTRAL)) were searched on 14 March 2022. Grey literature sources searched included theses databases and websites of applicable associations. The MEDLINE and grey literature search strategies are reported in Supplementary Table S2 (Supplementary data are available in Age and Ageing online). No limits were applied to the date of publication, publication status or language. Following the removal of duplicates, articles were screened for inclusion by two independent reviewers at the level of title/abstract and full text. Reference lists of relevant review articles were hand-searched.

Included studies were randomised controlled trials that detailed a therapeutic treatment of any duration and follow-up period and measured anxiety symptoms as an outcome. Included studies looked at individuals with anxiety symptoms or disorders at baseline or measured anxiety symptoms as an outcome in the general LTC population. Anxiety symptoms may have a clinical impact on residents and are measurable using anxiety screening tools but do not necessarily meet established diagnostic criteria (e.g., Diagnostic and Statistical Manual of Mental Disorders criteria) for an anxiety disorder. Study populations were individuals living in LTC or equivalent settings (i.e. access to 24-h nursing support and assistance with activities of daily living). Studies in which the population combined LTC and other non-equivalent settings were included if anxiety outcomes were reported separately for residents of LTC.

Included studies had to describe the effect of at least one treatment condition and comparator condition. Treatments could be pharmacologic or non-pharmacologic and comparators could be any condition. Anxiety did not have to be the primary outcome of included studies. Studies were eligible for inclusion in the meta-analysis if symptoms of anxiety were measured at baseline and post-treatment and if complete data were available within the publication or through correspondence with study authors.

At the level of title/abstract screening, included studies had to describe a therapeutic treatment in a LTC population. At the level of full text screening, participants had to be randomised into treatment conditions and anxiety symptoms had to be measured as an outcome. Disagreements amongst reviewers at the level of title/abstract were resolved between reviewers, whereas disagreements at full text were resolved by a third reviewer.

Data extraction

Data from included studies were extracted in duplicate by two independent reviewers and compared for accuracy by one reviewer. Discrepancies in extracted data were resolved by referring to the source study. Data extracted from each study included study origin and funding information, participant characteristics, anxiety detection tool information, LTC or equivalent facility characteristics, treatment and comparator condition information and outcome data. The longest follow-up measure was extracted as the post-treatment anxiety measure. Google translate was used to extract data from included non-English language studies.

Quality assessment

The quality of included studies was assessed by two independent reviewers using the Cochrane RoB 2 tool [21]. The RoB 2 tool can be used to assess parallel, cross-over or cluster randomised controlled trials. Quality assessments were specific to the anxiety outcome measured within each study. Disagreements in quality assessments were resolved between reviewers.

Meta-analysis

Pairwise meta-analyses were completed to produce a pooled standardised mean difference (SMD) for each treatment comparison. The SMD was selected as the summary effect measure because anxiety outcome data were continuous and obtained using assorted detection tools.

Pairwise meta-analyses were completed using random effects (DerSimonian Laird) models. The effect size was calculated per Hedges’ g. The Q statistic and I² value were used to evaluate if heterogeneity was present between studies and to quantify the proportion of total variability from between-study heterogeneity, respectively. Analyses with I² values <25%, between 25 and 50% and over 50% were considered to have low, moderate and high degrees of heterogeneity present [22]. Forest plots were produced for each comparison. All meta-analyses were completed using the statistical software STATA version 16.0 [23].

Results

A total of 6,617 articles were returned from electronic database searches after duplicates were removed. In total, 519 articles were reviewed in full text. An additional 10 articles were added from reference lists of relevant review articles. A descriptive synthesis was completed with 80 articles meeting full text inclusion criteria. The percent agreement amongst reviewers was 93.9% at the level of title/abstract screening and 93.3% at the level of full text screening. Seventy-four of the 80 included articles were eligible for inclusion in meta-analyses. Ten articles were included in meta-analysis comparisons once similar treatment and comparator conditions were grouped (Figure 1).

Figure 1

PRISMA diagram of study flow throughout the systematic review process.

Study quality

Most studies were found to have concern for risk of bias (45/80) or high concern for risk of bias (35/80; Supplementary Table S3, Supplementary data are available in Age and Ageing online). Individually randomised parallel group trials commonly had high concern for risk of bias within the domain of missing outcome data (18/60). Cluster randomised trials often had high concern for risk of bias within the domain of randomisation (2/10) and cross-over trials within the domain of effect of adherence to intervention (5/10). Across designs, studies had low concern for risk of bias within the domain of outcome measurement (39/80).

Study characteristics

The participant and LTC setting characteristics from included studies were summarised (Table 1 and Supplementary Table S4, Supplementary data are available in Age and Ageing online). Twenty countries were represented in the included literature, including the United States (n = 12), Australia (n = 11) and the United Kingdom (n = 7). Study sample sizes ranged from 11 to 659 enrolled participants. The mean age of study participants was between 62.5 and 89.3 years of age. Female participants made up between 30 and 100% of the study samples. The cognitive function of participants was commonly assessed using the Mini-Mental State Examination (MMSE; n = 36) with mean scores ranging from 4.1 to 29, representing severe dementia through to cognitively normal. Two studies reported an assessment of frailty within the study sample [24, 25]. Seven studies outlined inclusion criteria for baseline anxiety symptoms in participants [26–32], whereas all others looked at the general LTC population with anxiety symptoms measured as a primary outcome or as part of a global neuropsychiatric measure. Levels of cognitive function or dementia status of participants were an inclusion criterion for study participation in 70 studies. Although all studies measured anxiety symptoms, anxiety reduction was the primary treatment indication in 15 studies [26, 27, 29, 30, 32–42]. Studies ranged from a single LTC site to 35 different sites. Most studies did not report characteristics of the LTC site such as the level of care provided or the funding model.

Table 1

Demographic information from included studies

Study characteristic	Number of studiesa (% of total studies)
Country
United States	12 (15.00)
Australia	11 (13.75)
United Kingdom	7 (8.75)
Iran	7 (8.75)
Spain	6 (7.50)
The Netherlands	5 (6.25)
Taiwan	5 (6.25)
Other	27 (33.75)
Anxiety as primary treatment indication
Yes	15 (18.75)
No	65 (81.25)
Treatment type
Non-pharmacologic—other	21 (26.25)
Music	11 (13.75)
Multicomponent non-pharmacologic	10 (12.50)
Pharmacologic	9 (11.25)
Touch	6 (7.50)
Reminiscence	5 (6.25)
Robot and doll therapy	4 (5.00)
Exercise	4 (5.00)
Patient education	4 (5.00)
Psychotherapy	3 (3.75)
Cognitive behavioural therapy	3 (3.75)
Comparator type
Usual care control	41 (51.25)
>2 Comparator conditions	16 (20.00)
Social interaction control	9 (11.25)
Active treatment control	8 (10.00)
Placebo	6 (7.50)
Sample comorbidities reported
Dementia	33 (41.25)
Depression	4 (5.00)
Type 2 diabetes	1 (1.25)
Sleep disorder	1 (1.25)
Sample sizeb
<20	9 (11.25)
20–50	29 (36.25)
51–100	24 (30.00)
>100	18 (22.5)
% Female participants
<25%	–
25–50%	10 (12.50)
51–75%	28 (35.00)
>75%	32 (40.00)
Not reported	10 (12.50)
Ethnic majority Caucasian Taiwanese British African American Not reported	4 (5.00) 2 (2.50) 2 (2.50) 1 (1.25) 71 (88.75)
Age (mean)
<65	1 (1.25)
65−75	6 (7.50)
76−80	8 (10.00)
81−90	51 (63.75)
>90	–
Not reported	14 (17.50)
MMSE score (mean)
0−12	17 (21.25)
13−25	15 (18.75)
26−30	4 (5.00)
Not reported	44 (55.00)
Number of LTC sites
1	23 (28.75)
2−5	25 (31.25)
6−10	11 (13.75)
>10	11 (13.75)
Not reported	10 (12.5)

Note: LTC, long-term care; MMSE, Mini-Mental State Examination.

^aNumber of studies within each category or range.

^bNumber of individual participants enrolled in the study at onset.

Anxiety outcome assessment

Outcome measures focused on general anxiety questionnaires, which capture broad symptoms of anxiety. Sixty-nine studies measured participant anxiety as a symptom. Specific anxiety subtypes such as dementia worry (n = 1), death anxiety (n = 4), anxiety about upcoming events (n = 1) and combinations of anxiety and insomnia (n = 2), or anxiety/fear (n = 3) were assessed in other studies (Table 2).

Table 2

Description of anxiety outcome assessment including the anxiety detection tool(s) used and the anxiety status of participants

Anxiety assessment feature	Number of studiesa (% of total studies)
Anxiety type assessed
Anxiety symptoms	69 (86.25)
Anxiety/fear	3 (3.75)
Death anxiety	4 (5.00)
Anxiety and insomnia	2 (2.50)
Dementia worry	1 (1.25)
Anxiety about upcoming events	1 (1.25)
Anxiety symptoms at baseline
Yes	7 (8.75)
No	73 (91.25)
Outcome assessor blinding
Yes	29 (36.25)
No	13 (16.25)
Not reported	38 (47.50)
Anxiety screening tool usedb
Rating Anxiety in Dementia scale	11
State-Trait Anxiety Inventory	10
Neuropsychiatric Inventory	6
Neuropsychiatric Inventory-Nursing Home Version	5
Geriatric Anxiety Inventory	6
Depression Anxiety Stress Scales	5
Hamilton Anxiety Rating Scale	4
Beck Anxiety Inventory	5
Hospital Anxiety and Depression Scale	5
Otherc	27

^aNumber of studies within each category.

^bFour studies used two tools each, therefore % of total studies not reported.

^cSee Supplementary Table S4 for details.

Anxiety detection tools frequently used included the Rating Anxiety in Dementia (RAID) scale (n = 11), the State-Trait Anxiety Inventory (n = 10), the Neuropsychiatric Inventory (NPI) (n = 6) and the Geriatric Anxiety Inventory (n = 6). Four studies measured anxiety using multiple tools (Supplementary Table S4, Supplementary data are available in Age and Ageing online). Outcome assessors were blind to treatment allocation in 29 studies. All baseline and follow-up measures for anxiety were reported for the treatment and comparator conditions for each study (Supplementary Table S5, Supplementary data are available in Age and Ageing online). The longest reported follow-up measure for anxiety was of interest to identify lasting treatment effects and ranged between 3 minutes to 20 months from baseline assessment.

Treatments identified

Similar studies were grouped based on the described treatment. Treatment categories were established when three or more studies described similar interventions. Treatment categories and descriptions of the treatment conditions within each study were reported (Supplementary Table S6, Supplementary data are available in Age and Ageing online). The resulting 11 treatment categories were pharmacologic (n = 9), music (n = 11), touch (n = 6), multicomponent non-pharmacologic (n = 10), exercise or movement (n = 4), robotic animal or therapy doll (n = 4), patient education (n = 4), psychotherapy (n = 3), cognitive behavioural therapy (n = 3), reminiscence (n = 5) and other non-pharmacologic (n = 21). The other non-pharmacologic category included treatments with two or fewer studies and contained interventions like animal assisted therapy, cognitive stimulation therapy and staff consultation or education. Drug categories identified within pharmacologic treatments included antipsychotics, anxiolytics, beta-blockers, antidepressants and a pain management protocol. Non-pharmacologic treatments are displayed in Figure 2.

The comparator condition was described as usual care in 41 studies. Other studies used a social interaction control (9/80), active treatment control (8/80) or placebo (6/80) group. Sixteen studies employed designs of three or more conditions.

Figure 2

Non-pharmacologic intervention categories identified through the systematic review of treatments for anxiety symptoms and disorders.

Thirty-four studies reported significant between-group findings (P < 0.05) for anxiety symptom reduction favouring the treatment condition (Supplementary Table S4). An absence of significant findings was often attributed to small sample sizes, difficulty recruiting and retaining participants, floor effects and lack of treatment adherence monitoring.

Pharmacologic treatment studies describing the use of antipsychotics or anxiolytics reported significant reductions in anxiety compared with placebo or active pharmacologic comparators. One study each looking at beta-blockers compared with placebo, antidepressants compared with placebo and a pain management protocol compared with usual care did not report significant reductions in anxiety. Within non-pharmacologic treatments, cognitive behavioural therapy, psychotherapy and touch all compared with usual care or an active control, or patient education compared with usual care seemed to reduce anxiety. Exercise or movement compared with usual care or an active control, reminiscence compared with usual care or social interaction, a robotic animal or therapy doll compared with usual care or social interaction, and multicomponent interventions compared with usual care or active controls did not seem to reduce anxiety.

Meta-analysis

Meta-analysis comparisons were completed for music therapy compared with both usual care and social interaction conditions and for massage therapy compared with usual care (Figure 3). No further meta-analysis comparisons were completed as there were insufficient studies (<3) within groupings of comparable treatment and comparator conditions. All studies included in meta-analyses measured anxiety symptoms as an outcome and did not include participants with anxiety at baseline per inclusion criteria.

Figure 3

Forest plots of meta-analysis comparisons (A) music therapy and social interaction conditions; (B) music therapy and usual care conditions and (C) massage therapy and usual care conditions.

Music therapy as a treatment was compared with usual care in four studies and produced a pooled SMD of −0.82 (95% CI: −1.31, −0.34). The I² value was 66.2% indicating a high level of heterogeneity was present. Music therapy was compared with social interaction conditions such as reading or entertainment activities in three studies. The pooled SMD was −0.41 (95% CI: −0.72, −0.10) and the I² value was found to be 0%. Massage therapy was compared with usual care in three studies. The pooled SMD was found to be −4.32 (95% CI: −7.44, −1.19). The I² value was 91.56%. Music and massage therapy compared with usual care had significant findings for the presence of heterogeneity (P < 0.05).

Discussion

Our systematic review identified efficacious treatments for anxiety symptoms used in older adults living in LTC. Music therapy compared with usual care or social interaction and massage therapy compared with usual care were identified as efficacious treatments, however, these findings should be interpreted with caution given the significant heterogeneity detected. A range of non-pharmacologic treatments identified are supported for use in LTC to reduce symptoms of anxiety considering the minimal associated risk. Further high-quality evidence is needed before the identified treatments can be clinically recommended for use in LTC.

Many non-pharmacologic treatments identified had evidence of benefit for symptoms of anxiety without any reported risk for participants and were reportedly low-cost [27, 37, 43, 44] or easy to implement [26, 36, 42, 45]. All studies with statistically significant findings reported an improvement in anxiety scores for residents relative to the comparator condition, however, it is difficult to understand the clinical significance of these treatments as we do not know the minimal clinically important difference of the included anxiety tools.

When looking at treatments where anxiety was measured as the primary outcome for those with dementia who had anxiety symptoms at baseline, music listening was identified as efficacious compared to social [26] and usual care controls [27]. Music listening treatments involved residents independently listening to music tailored to their personal preference and were described as being inexpensive and easy to apply [26, 27]. For treatments where anxiety was measured as the primary outcome in the general LTC population with anxiety symptoms at baseline, mindfulness-based cognitive therapy [30] improved symptoms of anxiety compared with a reminiscence control. Treatments tailored to anxiety reduction and trialled within populations experiencing anxiety symptoms at baseline may provide the best evidence for anxiety management in LTC.

Treatments, such as therapeutic touch compared with usual care or massage therapy [33, 40, 42], focused on reducing anxiety symptoms as the primary outcome but did not investigate the treatment effect in those with anxiety symptoms at baseline. Therapeutic touch and relaxation techniques were noted as being easily implemented and time effective. Future research could use the knowledge of anxiety symptom reduction established in therapeutic touch literature as a starting point to develop treatments tailored to anxiety reduction in the LTC population. Studies that focused on anxiety reduction as the primary outcome were largely completed with those without dementia, indicating findings may not apply to those living with dementia in LTC.

Music therapy compared with usual care and social interaction improved anxiety symptoms. Improvements in anxiety symptoms were also found for massage therapy compared with usual care. Heterogeneity in music and massage therapy versus usual care comparisons might be explained by clinical differences in the participants, setting or intervention. The small number of studies within comparisons and limited reporting of clinical characteristics prevented the exploration of potential sources of clinical or methodological heterogeneity through subgroup or stratified analyses. Although heterogeneity was detected, the treatment effect estimates still have useful application as a source of evidence to support the selection of treatments in LTC.

Very few studies were found of pharmacologic treatments where anxiety symptoms were measured in residents of LTC, with only one describing anxiety as the primary treatment indication [39]. Thus, it is difficult to interpret these studies and apply them to the general LTC population. Given the risk of harm associated with pharmacologic treatments in the LTC population, there is an immediate need for further research before these treatments can be recommended for use.

The majority of pharmacologic treatments identified in our review were antipsychotics, which are not recommended for either chronic or as-needed use in older adults with dementia or cognitive impairment [49, 50]. The studies investigating antipsychotics were indicated for psychosis or agitation in dementia and were not recent publications. Thus, these studies do not inform current antipsychotic prescribing recommendations. One study assessed the effect of antidepressants, a common treatment for anxiety symptoms and disorders. The lack of evidence suggests the efficacy of antidepressants has not yet been determined for the LTC population.

One pharmacologic study identified anxiety as the primary outcome measure [39]. All other studies measured anxiety as a secondary outcome, which means that findings may not be relevant to the treatment of anxiety symptoms in residents of LTC. All pharmacologic studies treated those with dementia, indicating the findings from these studies may be more applicable to those living with dementia in LTC. The Canadian Clinical Practice Guidelines for anxiety, post-traumatic stress and obsessive-compulsive disorders do not guide anxiety management in older adults living in LTC but do highlight drug-related safety concerns for the older adult population [7]. Harm must be considered when initiating pharmacologic treatments, and options more likely to cause harm based on existing criteria should be avoided [49]. The appropriateness of pharmacologic treatments must be considered on an individual basis before initiating treatment.

Treatment categories tailored to those with dementia (e.g. robot and therapy doll interventions, music therapy, staff education or consultation and cognitive stimulation therapy) are important given the prevalence of dementia in LTC [51]. Studies completed with persons with dementia that measured anxiety as a secondary outcome often utilised global measures like the NPI. The NPI is not specific to anxiety, therefore, the accuracy when used to assess for anxiety alone is unclear and may be further complicated by symptom overlap with agitation or depression [52]. Studies that measured anxiety as the primary outcome using a tool validated for use with persons with cognitive impairment like the RAID scale produce the best efficacy evidence for anxiety treatments in persons with dementia [53].

Few studies required anxiety symptoms or disorders as participant inclusion criteria. The treatments trialled in the general LTC population may be reasonable and show benefit for use in LTC but may not be generalizable to those with defined anxiety disorders or clinically relevant anxiety symptoms. Future research that focuses on a source population of residents of LTC with anxiety symptoms or disorders will provide better evidence supporting the therapeutic benefits of treatments that reduce anxiety symptoms.

Included studies did not discuss or present processes for intervention implementation in individual LTC settings. LTC sites considering implementing identified interventions should use evidence-backed implementation processes like the Knowledge to Action framework to guide the process [54]. Studies included in this systematic review were often limited by small sample sizes and attrition. Small study sample sizes could have limited the ability to detect small but significant changes in anxiety symptoms. Many studies lacked blinding and assessments of treatment fidelity concerning both treatment implementation and participant compliance. Study findings may not be generalizable to other settings as the study population was often a small number of participants obtained from a single LTC site. Limited information was provided in most studies about the care environment from which participants were sampled and the characteristics of participants including ethnicity and comorbidities making the generalizability of findings difficult to interpret.

Only randomised controlled trial designs were included in the systematic review to identify the highest level of causal evidence. Sample size and participant recruitment in included trials may have been impacted by barriers specific to the LTC setting including issues with resident recruitment, consent, treatment group preference, and attrition in addition to issues related to the investigator, caregivers, family and/or facility operator [55]. Using the SMD as the summary estimate of treatment effect was the best way to compare similar treatments across studies that occurred internationally and utilised various anxiety detection tools, however, this estimate may be difficult to interpret and apply in a clinical setting. The summary estimates produced may provide limited insight into the true treatment effect due to the small number of studies within each meta-analysis comparison. There were a limited number of identified studies included in meta-analysis comparisons. Our ability to quantitively compare many identified studies was limited by the variation in treatment and comparator conditions, the heterogeneous nature of the LTC population, and different sampling frames within studies.

In conclusion, non-pharmacologic treatment options for anxiety symptoms such as music and massage therapy that are feasible, low-risk and low-cost can readily be adopted in LTC. The heterogeneous population, particularly the impact of cognitive impairment, within LTC must be considered when interpreting research findings. Few studies described treatments tailored to anxiety reduction which underscores that the prevalence and burden of anxiety symptoms in LTC is under-recognised. To better understand the therapeutic effect of many treatments there is a need for future trials with residents of LTC that have anxiety disorders or clinically relevant anxiety symptoms at baseline.

Acknowledgements

We would like to thank Dr Fariborz Goodarzi for the valuable support he provided in translating articles for the present review.

Declaration of Conflicts of Interest

Zahra Goodarzi—All funding is independent peer reviewed research funding, from Canadian Institute for Health Research, Canadian foundation for Healthcare improvement, University of Calgary, Hotchkiss Brain Institute, O'Brien Institute for Public Health and Brenda Strafford Foundation. Zahinoor Ismail—Funding from Canadian Institutes of Health Research, Brain Canada, Canadian Consortium on Neurodegeneration in Aging, Weston Foundation and the National Institutes of Health. Has received honoraria from Lundbeck/Otsuka, unrelated to this work.

Declaration of Sources of Funding

None.