Introduction
Agitation is a common and distressing behavioral symptom in patients with advanced dementia, particularly in hospice-eligible populations (Sadowska et al. Reference Sadowska, Turnwald and O’Neil2025). It includes increased motor activities, restlessness, aggressiveness, and mental distress. Agitation is commonly measured by the Cohen-Mansfield Agitation Index (CMAI), and higher scores have been associated with poorer quality of death, elevated caregiver burden, increased use of professional caregivers, and greater healthcare costs (Cohen-Mansfield, Reference Cohen-Mansfield1991, Reference Cohen-Mansfield1997; Eisenmann et al. Reference Eisenmann, Golla and Schmidt2020; Gerlach et al. Reference Gerlach, Martindale and Bynum2023; Jones et al. Reference Jones, Aigbogun and Pike2021; Moon et al. Reference Moon, McDermott and Kissane2018; Patrick et al. Reference Patrick, Gunstad and Martin2022). Hospice-eligible dementia patients are defined as individuals with advanced dementia and a prognosis of 6 months or less, according to Medicare hospice eligibility criteria or equivalent validated prognostic tools. In the terminal stage, patients may be non-verbal and bed-bound yet still experience significant agitation, affecting up to 71% of this population in the last years of life (Hendriks et al. Reference Hendriks, Smalbrugge and Galindo-Garre2015; Mintzer, Reference Mintzer2022).
Despite its prevalence, guidance for agitation management at the end of life remains limited. Approximately 73% of hospice-eligible patients with agitation receive psychotropic medications such as atypical antipsychotics and antidepressants (Kwon et al. Reference Kwon, Kim and Bruera2017), which can cause sedation, confusion, and extrapyramidal symptoms, potentially creating emotional disconnection between patients and caregivers (Marder, Reference Marder2007; Sadowska et al. Reference Sadowska, Turnwald and O’Neil2025). While nonpharmacological interventions and selected pharmacologic agents (e.g., SSRIs) have been tested in advanced dementia, few studies specifically target hospice-eligible populations, where altered pharmacodynamics, polypharmacy risks, and distinct care goals may influence efficacy and tolerability (Broers and Bianchi, Reference Broers and Bianchi2024; Chen et al. Reference Chen2023; Cummings et al. Reference Cummings, Lyketsos and Peskind2015; Hermush et al. Reference Hermush, Ore and Stern2022; Pautex et al. Reference Pautex, Bianchi and Daali2022; Porsteinsson et al. Reference Porsteinsson, Drye and Pollock2014; Schneider et al. Reference Schneider, Frangakis and Drye2016).
Previous reviews have examined palliative care in dementia broadly but have not synthesized evidence specific to agitation management in hospice-eligible patients (Agar, Reference Agar2020; Senderovich and Retnasothie, Reference Senderovich and Retnasothie2020). Given the high prevalence and clinical impact of agitation, a focused review of randomized controlled trials (RCTs) in this group is warranted. This review aims to identify and evaluate RCT evidence for agitation management in hospice-eligible dementia patients, highlight methodological and evidence gaps, and inform priorities for clinical practice and future research.
Methods
We conducted a structured literature search for published RCTs up to 2024 via MEDLINE, Pubmed, PsycINFO, Proquest, CINAHL, EBSCO, EMBASE, and Cochrane Library. Search terms included combinations of end-of-life, late-stage dementia, agitation, terminal restlessness, terminal delirium, and RCTs.
Studies were eligible if they: (1) included hospice-eligible dementia patients, defined as individuals with advanced dementia and a prognosis of ≤6 months, or a Functional Assessment Staging Tool (FAST) score ≥7a (Luchins et al. Reference Luchins, Hanrahan and Murphy1997; Sclan and Reisberg, Reference Sclan and Reisberg1992); (2) targeted agitation as a primary or secondary outcome; and (3) were RCTs. Hospice eligibility had to be explicitly defined using a hospice setting or a validated prognostic tool. We excluded non-RCT designs, studies without clear hospice eligibility measurement, non-English publications, and unpublished trials.
Of the 44 articles identified, 17 RCTs did not include terminally ill patients with dementia and agitation, 13 did not include agitation as an outcome variable, 6 were not RCTs, 2 were not completed or published trials, 2 were written in German, and 1 could not be retrieved.
Results
Three RCTs met the inclusion criteria, comprising 2 nonpharmacological and 1 pharmacological intervention (Froggatt et al. Reference Froggatt, Best and Bunn2020; Kovach et al. Reference Kovach, Taneli and Dohearty2004; Magai et al. Reference Magai, Kennedy and Cohen2000). All trials were conducted in residential care settings, and none were conducted in home hospice environments. Table 1 summarizes the study settings, designs, sample characteristics, interventions, and key findings.
Clinical trials on hospice-eligible patients with dementia and agitation
Abbreviation & Note:
FAST: Functional Assessment Staging Tool.
CMAI: Cohen-Mansfield Agitation Inventory.
QoL: Quality of Life.
BACE: Balancing Arousal Controls Excesses.
MMSE: Mini-Mental State Examination.
a. Studies were conducted in the United Kingdom.
b. Studies were conducted in the United States.
Nonpharmacological approach: Namaste Care (Froggatt et al. Reference Froggatt, Best and Bunn2020)
This UK-based cluster RCT enrolled 32 participants (mean age 79) with dementia and a less-than-three-month prognosis. Participants were predominantly White; only 1 African Caribbean participant was recruited. The intervention consisted of twice-daily sessions emphasizing person-centered activities, meaningful engagement, and sensory comfort to enhance quality of life, and consequently foster emotional connection and reduce behavioral symptoms. While CMAI scores fluctuated, qualitative observations suggested improved affect and interaction during the intervention.
Nonpharmacological approach: balancing arousal controls excesses (BACE) (Kovach et al. Reference Kovach, Taneli and Dohearty2004)
This US-based parallel-group RCT recruited 78 participants (mean age 86.6; 91% female) with moderate-to-severe dementia, including hospice-eligible individuals. The intervention scheduled individualized activities to balance high – and low-arousal states. CMAI scores decreased from 39.0 to 30.5 in the intervention group, compared to minimal change in controls (32.6–32.3).
Pharmacological approach: sertraline trial (Magai et al. Reference Magai, Kennedy and Cohen2000)
This US-based double-blind, placebo-controlled randomized trial enrolled 31 female nursing home residents (mean age 89) with late-stage Alzheimer’s disease and co-morbid depression. Medication was administered as follows: Weeks 1–2, 25 mg/day; Weeks 3–4, 50 mg/day; Weeks 5–8, 100 mg/day. Agitation, assessed as a secondary outcome using the CMAI, decreased over time in both groups, but between-group differences were not statistically significant.
Discussion
Of the studies identified through our systematic search, only 3 met the inclusion criteria, highlighting the scarcity of high-quality evidence for agitation management in hospice-eligible dementia patients. All trials were conducted in residential care settings, and none in home hospice environments, limiting generalizability to community-based care. Reporting of socio-demographic variables was incomplete (see Table 1); for example, Kovach et al. (Reference Kovach, Taneli and Dohearty2004) did not report race or ethnicity, and Froggatt et al. (Reference Froggatt, Best and Bunn2020) recruited predominantly White participants (96.9%) with only 1 African Caribbean participant, despite the ethnic diversity of the UK population (Office for National Statistics, 2022). Such gaps restrict the applicability of findings to diverse patient populations.
The nonpharmacological interventions varied in aims and design. BACE, a parallel-group RCT, directly targeted agitation reduction and showed measurable improvement in CMAI scores. Namaste Care, a cluster RCT, prioritized quality of life, with qualitative reports of improved affect and interaction but variable quantitative agitation outcomes. The single pharmacologic trial, a double-blind, placebo-controlled RCT of sertraline, found no significant benefit for agitation compared to placebo. Its all-female sample, depression-focused design, and focus on depressive symptoms rather than agitation as a primary outcome substantially limit the generalizability of findings to broader hospice-eligible dementia populations.
Collectively, these findings underscore the complexity of managing end-of-life agitation in dementia and the methodological challenges of conducting RCTs in this setting. Novel nonpharmacological strategies, such as the Hands of Comfort technique using warm gloves (Julião et al. Reference Julião, Bruera and Silva2023), and pharmacologic agents with preliminary promise (e.g., dextromethorphan-quinidine, cannabinoid-based compounds) have yet to be evaluated in hospice-eligible populations (Broers and Bianchi, Reference Broers and Bianchi2024; Cummings et al. Reference Cummings, Lyketsos and Peskind2015; Hermush et al. Reference Hermush, Ore and Stern2022; Pautex et al. Reference Pautex, Bianchi and Daali2022). In sum, only 3 RCTs in the last 34 years met our criteria for inclusion, and no treatments have been evaluated in over a decade. The small evidence base, combined with the limited diversity of study populations and variation in intervention aims, highlights an urgent need for future RCTs to identify, develop, and test both pharmacologic and nonpharmacologic interventions for managing end-of-life agitation in patients with dementia.
Competing interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Open access
