The search for effective delirium treatment for persons with dementia in the post-acute setting

With the unprecedented rise in the proportion of older adults in most industrialized countries, the treatment of cognitive impairment among older adults has emerged as a leading healthcare priority. Delirium and dementia are each among the most common causes of cognitive impairment across clinical settings, but they also commonly coexist as delirium superimposed on dementia (DSD). An estimated 22–89% of hospitalized and community older populations with dementia develop DSD¹. The occurrence of DSD holds serious prognostic implications, including a worsening trajectory of cognitive decline,² substantial functional decline,³ increased length of stay in hospital,^{3, 4} and higher rates of rehospitalization,^{1, 2} institutionalization,^{5, 6} and death.^{1, 6} DSD is associated with a longer duration of delirium and slower recovery compared with patients without dementia who develop delirium.⁷ Moreover, health care costs and utilization are significantly higher in patients with DSD compared to patients with dementia alone, delirium alone, and patients with neither delirium nor dementia.⁴

Previous studies have found that DSD patients are more frequently diagnosed with hyperactive delirium, and exhibit symptoms such as agitation, aggression, and irritability.⁸ Such complex behavioral problems and functional needs often necessitate post-acute care following hospitalization. Delirium has been shown to be persistent in more than half of patients admitted with delirium to post-acute care (PAC) at one month.⁹ However, despite the increasing importance of PAC in DSD and other age-related conditions, this setting remains largely unexamined in research studies due to daunting logistical and ethical constraints in conducting high-quality research in vulnerable institutionalized populations with cognitive impairment, frailty, and multi-morbidity.

The present study¹⁰ conducted a single-blind randomized controlled trial of cognitively stimulating activities applied for up to 30 minutes daily, five days per week, in 283 older adults with DSD in 8 post-acute care facilities in Pennsylvania. The primary hypothesis was that individualized cognitively stimulating activities could reduce the duration and severity of delirium compared with usual care. However, the study demonstrated no significant differences in the primary outcomes of delirium duration or severity in the intervention group compared with usual care. Significant differences were found for secondary outcomes favoring the intervention, including a significant reduction in average length of stay from 53 days in usual care to 36 days in the intervention group (P=.01), and a statistically significant improvement in cognition, specifically on the executive control portion (CLOX1) of a clock-drawing task (p = 0.009). After adjusting for baseline difference, CLOX1 differences remained significant, however, no differences were found on the copy portion (CLOX2) (p=0.11). This finding is consistent with the notion that impairment in frontal systems controlling clock drawing will affect CLOX1 more than CLOX2.¹¹

Kolanowski and colleagues¹⁰ followed rigorous methodologic approaches, including a randomized design achieving balanced comparison groups, carefully blinded outcome assessment with assurance of maintenance of blinding at the end of the study, verification of high inter-rater reliability for all key study measures, adequate sample size and power to test hypotheses, and an intention-to-treat analysis approach. This is a high-quality randomized controlled study, with a high modified Jadad score of 5 out of 6 points; 1 point was lost for single rather than double blinding, which was not feasible with the intervention approach.¹²

The one major limitation of the study is the complexity of the intervention. This was an individualized “prescription” of 15 activities that were administered in individual sessions for up to 30 minutes each day, five days per week for 30 days or until discharge, by trained research assistants. Activities were selected from a large database¹³ that included activities such as describing steps to prepare a meal, discussing current events, or generating a list of three similar items. The selection of activities for each participant was adjusted depending on their level of cognitive performance for level of difficulty (“mild”, “moderate”, or “difficult”) and accounting for any functional impairment, such as low vision or impaired hearing. Activities were also selected based on the leisure interests of the participant. For example, activities for a person with a mild degree of cognitive impairment, without visual impairment, who enjoys movies might include identifying Famous Faces, a visually-based activity of moderate difficulty. This activity would not be appropriate for participants with severe dementia and macular degeneration; an alternative activity such as an easy verbal task of reciting the alphabet to the letter L would be selected instead. The authors chose to tailor the activities in this manner in order to engage and capture interest, however, such a personalized approach may prove quite difficult to reproduce across sites, and the potency -- the potential for effectiveness of each individualized intervention approach, analogous to the “dose” of a medication -- is difficult to standardize and quantify. In addition, while the missing data and study dropouts were carefully documented, description of the approaches to handle these missing data in the analyses was not fully detailed or well-explained. Typically, in this type of high-quality randomized trial, sensitivity analyses would be done using multiple approaches or imputation procedures to assess for any potential biases from the missing data, and to evaluate their potential impact on the study conclusions.

Despite these limitations, this study is groundbreaking and highly significant for addressing a topic of utmost clinical importance with noteworthy methodologic rigor in the difficult and highly relevant setting of post-acute care. This randomized clinical trial was conducted despite the constraints and challenges of studying a highly frail, vulnerable, and cognitively impaired population in the post-acute care setting across 8 facilities. Despite the success of delirium prevention trials, particularly of multicomponent nonpharmacologic intervention strategies,¹⁴ successful treatment of delirium once it occurs has lagged. To date, studies of delirium treatment across all patients have been largely disappointing, with little impact on delirium duration, severity, or important clinical outcomes in trials with antipsychotics,^{15, 16} cholinesterase inhibitors,^{17, 18} and other medications.¹⁹ Further, DSD patients who are undergoing management of delirium with antipsychotics have lower rates of delirium resolution than patients with delirium alone.²⁰ Unfortunately, the current study is consistent with previous work having negative outcomes for these primary outcomes.

The demonstration in this study of significant findings for improvement in executive functioning and shortening length of stay are noteworthy and clinically relevant. Because executive functioning is a key cognitive domain involved in delirium-- both as a risk factor and disease marker-- cognitively-stimulating activities may have a role in future delirium prevention trials by both improving cognitive performance and strengthening cognitive reserve as a protection against delirium risk factors. In addition, the length of stay benefit suggests that perhaps even the subtle changes noted in delirium severity and duration-- along with the improvement in executive functioning—may be enough to result in a positive impact in this highly vulnerable population.

The findings in this study are not surprising, since the treatment of delirium is recognized to be especially challenging because of its fluctuating course, affecting patients in a variety of clinical settings, and its multifactorial etiology. Successful delirium treatment likely requires targeted approaches to multiple etiologic contributors simultaneously -- rather than a single unified intervention strategy as tested here. It is likely that delirium treatment will be even more complex in persons with underlying dementia, where baseline cognitive status is already impaired, and clinical symptomatology may arise from the underlying dementia, the acute delirium, or both.

The next steps in searching for effective treatment of DSD will be to replicate findings from high-quality studies like this one in other populations, and to investigate if treatment effectiveness can be increased by combining cognitively-stimulating activities with other targeted approaches for the multiple etiologic contributors to delirium in persons with underlying dementia (Table 1). Other strategies, such as targeting shared pathophysiological mechanisms between delirium and dementia, bolstering cognitive reserve, and investigating interventions in different care settings, are equally critical areas of potential treatments to be considered. This study, by showing the potential for effective treatment of DSD, lays an important foundation for this future work.

Table:

Future directions for treatment of delirium superimposed on dementia

Design Considerations
Consideration	Target	Goal
Patient selection	Moderate to high risk for delirium superimposed on dementia	Develop predictive model Choose group most likely to benefit from intervention
Study setting	Home, outpatient, inpatient, ICU, surgery, post-acute care, long-term care	Treatment across settings may be needed to maximize effectiveness
Intervention type	Single versus multicomponent Pharmacologic versus non-pharmacologic	Choose number and type of intervention with the best therapeutic: toxic ratio in population chosen
Intervention potency	Assure adequate potency to impact on primary outcomes	Select intervention with greatest evidence of effectiveness
Timing and duration of intervention	Evaluate best timing to start intervention, and adequate duration	Hospital versus postacute care; sufficient duration of intervention to assure effect can be achieved
Intervention Strategies
Strategy	Target	Goals
Prevention of cognitive decline in DSD	(1) Earlier recognition of DSD (2) Targeted interventions towards etiologic factors (3) Prevention of complications	• Improve identification of delirium—diagnosis, phenomenology, severity, and subtypes • Targeted interventions towards etiologic contributors (e.g., drugs, infections, metabolic derangements) • Prevent complications, including falls, aspiration, pressure ulcers, functional decline
Investigate pathophysiological overlap between delirium and dementia	Inflammation Neuronal stress Acceleration of dementia pathology	Identify potential disease-modifying treatment strategies to lay groundwork for intervention development studies • Identify inflammatory biomarkers for delirium • Investigation of novel neuronal injury markers associated with endothelial damage and blood-brain barrier disruption in delirium • Explore impact of delirium on dementia markers
Boost cognitive reserve	Brain plasticity and resilience	Physical and/or cognitive intervention trials Transcranial Direct Current Stimulation trials
Targeted approaches aimed at multiple contributors to delirium	Multiple risk factors/contributors to delirium	Combined approaches to management, (i.e. multi-component strategies such as the Hospital Elder Life Program, HELP) Limited use of pharmacologic approaches