Abstract
We developed and evaluated the effect of Multimodal Cognitive Enhancement Therapy (MCET) consisting of cognitive training, cognitive stimulations, reality orientation, physical therapy, reminiscence therapy, and music therapy in combination in older people with mild cognitive impairment (MCI) or mild dementia. This study was a multi-center, double-blind, randomized, placebo-controlled, two-period cross-over study (two 8-week treatment phases separated by a 4-week wash-out period). Sixty-four participants with MCI or dementia whose Clinical Dementia Rating was 0.5 or 1 were randomized to the MCET group or the mock-therapy (placebo) group. Outcomes were measured at baseline, week 9, and week 21. Fifty-five patients completed the study. Mini-Mental State Examination (effect size = 0.47, p = 0.013) and Alzheimer’s Disease Assessment Scale-Cognitive Subscale (effect size = 0.35, p = 0.045) scores were significantly improved in the MCET compared with mock-therapy group. Revised Memory and Behavior Problems Checklist frequency (effect size = 0.38, p = 0.046) and self-rated Quality of Life – Alzheimer’s Disease (effect size = 0.39, p = 0.047) scores were significantly improved in the MCET compared with mock-therapy. MCET improved cognition, behavior, and quality of life in people with MCI or mild dementia more effectively than conventional cognitive enhancing activities did.
Keywords
INTRODUCTION
Nonpharmacological interventions have been widely used to treat or prevent dementia in combination with pharmacotherapies in various care settings [1, 2]. Neural plasticity and capacity for cognitive-deficit compensation may underlie the efficacy of nonpharmacological interventions [3]. Elderly individuals who engaged in a wider activity spectrum (comprising physical, cognitive, and socializing activities) were less likely to develop dementia than those engaging in only one type of activity, or none at all [4]. However, this does not mean that any simple combination of cognitive stimulations, exercises, or socializing activities would be beneficial or cost-effective to people with dementia.
Thus, there is a need for a multimodal nonpharmacological intervention (MNPI) that optimally consists of appropriate unimodal interventions. MNPIs consisting of a combination of two or more types of nonpharmacological interventions may show better efficacy than unimodal interventions. MNPIs were beneficial for cognition [5–12], activities of daily living (ADL) [7, 13–15], depression [8–11], behavioral and psychological symptoms (BPSD) [9], sleep disturbances [13], functional capacities [5], physical health [9], gait and posture [5, 9], and quality of life (QoL) [8] in people with dementia, and for decreasing care time of their caregivers [15].
However, the extent of evidence provided by most previous trials investigating MNPIs has been hindered by methodological limitations, including the following: a sample size below 50 subjects in the intervention group [5, 8–13]; non-randomization [5, 8]; absence of placebo intervention (i.e., usual care) [5–8, 13–15]; data unblinded to evaluators and participants [9]; or the use of single-blinded studies [9, 12]. Additionally, most trials did not adopt comprehensive outcome variables [5–7, 14]. Several MNPIs that included only a few [5, 6] or no cognitive activities [16] showed only modest efficacy. Some proven effective MNPIs are recommended to be administered daily for more than a year [7, 15]. Mild cognitive impairment (MCI) greatly increases risk for dementia, and is approximately two to three times more prevalent in the elderly population than dementia [17]. Moreover, patients with MCI may receive increased benefits from MNPIs compared with those with dementia [18] since their cognitive plasticity and learning potential may be more intact than dementia patients [19]. However, most previous MNPI trials targeted patients withmild-to-moderate dementia, rather than those with MCI [5, 6, 10, 14].
We developed an MNPI named “Multimodal Cognitive Enhancement Therapy (MCET) [20]” that consists of six nonpharmacological interventions (cognitive training [21], cognitive stimulations [22], reality orientation, physical therapy, reminiscence therapy, and music therapy) that received an “A-grade” recommendation for improving cognition, mood, behavioral symptoms, ADL, or QoL of people with dementia in a previous meta-analysis [23]. In this study, we conducted a multi-center, double-blind, randomized, placebo-controlled, two-period crossover clinical trial to compare the efficacy of the MCET with that of a mock-therapy that consists of conventionally employed cognitive stimulation activities, social activities, and exercises used in many Korean care facilities for improving cognitive function, BPSD, functional disability, and QoL among individuals with mild dementia or MCI.
MATERIALS AND METHODS
Patients with MCI or early dementia were recruited between December 2013 and March 2014 from dementia clinics of two university hospitals (Seoul National University Bundang Hospital and Konkuk University Medical Center) and two geriatric hospitals (Kyunggi Provincial Hospital for the Elderly and Jeonju City Welfare Hospital for the Elderly). Eighty-six patients who were informed about the study from dementia clinics, local dementia prevention centers, or community welfare centers contacted the study secretary in each research center. Among these, 77 patients were screened for eligibility, and 64 patients were finally enrolled (32 MCI and 32 dementia patients). Reasons for exclusion after screening included illiteracy or advanced dementia (eight cases), systemic illness (two cases), and refusal to participate (three cases) (Fig. 1). MCI was diagnosed according to the International Working Group on MCI criteria [24]. Of the 32 patients with MCI, 25 had amnestic single type, 6 had amnestic multiple type, and one had nonamnestic single type of MCI. Dementia was diagnosed according to Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) criteria [25]. Of the 32 patients with dementia, 28 had Alzheimer’s disease, three had vascular dementia, and one had frontotemporal dementia. All patients had a Clinical Dementia Rating (CDR) of 0.5 or 1.
Subjects with Axis I disorders listed in the DSM-IV [25], any neurological disorders that could affect cognitive function, or any physical condition that could preclude regular attendance and full intervention-program participation were excluded. Illiterate subjects were also excluded. Although participants were allowed to take cognitive enhancers, sedatives, and antidepressants during the study, they were encouraged to take stable medication doses for 3 months prior to join the trial and during the trial. All subjects were fully informed of the study protocol, and provided written informed consent, signed by the subjects or their legal guardians.
Study design
This study was a multi-center, double-blind, randomized, placebo-controlled, two-period, crossover trial (clinicaltrials.gov; NCT02350738) investigating MCET in individuals with MCI or mild dementia. Randomization was performed by the Medical Research Collaborating Center in Seoul National University Bundang Hospital (SNUBH) who had no contact with patients or caregivers, using a random code table and permuted-block randomization with varying block sizes [26], using SAS software, version 9.2 (SAS Institute Inc., Cary, NC, USA). Stratification variables included study center and clinical diagnosis (MCI or dementia). The allocation sequence was produced independently and concealed until patients had entered the trial. The trial consisted of two eight-week therapies that were crossed over. During the first period, patients were randomized to receive MCET (N = 32) or mock-therapy (N = 32). After a four-week washout period, each participant began an alternative treatment for the next eight weeks (i.e., the crossover). The clinical and neuropsychological responses were assessed at the beginning (week 0), and end of the first treatment period (week 9), in addition to the end of the second treatment period (week 21) by blind evaluators (Fig. 1). Patients and caregivers were blinded to intervention type. This study protocol was approved by the Institutional Review Board of SNUBH.
Interventions
MCET
MCET [20] consists of cognitive training [21], cognitive stimulations [22], reality orientation, physical therapy adopted from ROM dance [27], reminiscence therapy, and music therapy receiving an “A-grade” recommendation by the Scottish Intercollegiate Guidelines Network (SIGN) for efficacy in treating global cognition (cognitive stimulations, physical therapy, reminiscence therapy, and reality orientation), memory (cognitive training), depressive mood (physical therapy and reminiscence therapy), BPSD (reminiscence therapy), ADL (cognitive training and music therapy), QoL (cognitive stimulations), and social communication (cognitive training), in our previous systematic review and meta-analysis [23]. A comprehensive and uniform protocol was designed based on previous procedures used in well-designed, randomized, controlled trials demonstrating efficacy of each program. MCET was conducted in the form of a group therapy for eight participants, consisting of three 3-hour sessions per week for eight weeks. Each session consisted of a 30-minute physical exercise program, a 30-minute reality orientation program, a 30-minute cognitive training program, a 30-minute break, and a 60-minute program that provided reminiscence therapy, cognitive stimulation, and music therapy in turn (Fig. 2). To maintain therapeutic uniformity across study centers, we developed a 24-session MCET manual, and provided the research occupational therapists with a two-day MCET training program. Development and evaluation processes were implemented according to the new Medical Research Council guidance for developing and evaluating complex interventions [28].
Mock-Therapy (MT)
Similar to MCET, MT also consisted of three 3-hour sessions per week for eight weeks. A single research occupational therapist provided both the MCET and MT at a given study site. The first and second MT sessions in a week included watching videotapes about general health issues for 90 minutes [12], gymnastic exercises following the videotape for 20 minutes, a break for 30 minutes, and unstructured conversation [12, 29] or recreation [30] for the last 40 minutes in turn. The third MT session in a week included watching movies and having conversations about such (Fig. 2). All MT components were passive cognitive stimuli, and provided neutral conditions based on previous controlled trials [12, 30]. To maintain therapeutic uniformity across study centers, we also developed a 24-session MT manual and provided research occupational therapists with a two-day MT training program.
Outcome measures
Primary outcome measures included Mini-Mental State Examination (MMSE) [31] and Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-Cog) scores to examine treatment effects on cognitive function [32]. Secondary outcome measures included the following: Revised Memory and Behavior Problems Checklist (RMBPC) scores, which measure BPSD in dementia [33]; Geriatric Depression Scale (GDS) scores, which measure the severity of depressive mood [34]; Disability Assessment for Dementia (DAD) [35] scores, which measure functional abilities, and QoL in Alzheimer’s Disease (QoL-AD), which was reported by the patients and their caregivers [36]. The RMBPC provides two kinds of indices: frequencies of problematic behaviors in patients with dementia (RMBPC-F) and reactions of caregivers to problematic behaviors (RMBPC-R). Each index consists of three subscores, including those measuring memory-related problems, depression, and disruption [33]. DAD total score was obtained by adding the rating for each question and converting this total score out of 100. DAD subscores, including the ADL score (17 items; DAD-ADL), and the Instrumental ADL (IADL) score (23 items; DAD-IADL), were derived in the same manner as the total score [37]. All outcome measures were administered at weeks 0, 9, and 21 by trained research psychologists who were blinded to allocation information (Fig. 1).
Sample size and statistical analyses
Assuming an attrition rate is 25%, sample size of 64 patients (8 per treatment arm among 4 centers) would provide more than 80% power detect a two-point mean difference in MMSE with a standard deviation (SD) of 2.187 (as in the Korean validation study [31]) at a two-sided type-I error of 0.05.
An intent-to-treat (ITT) sample was used for primary efficacy evaluation, which was defined as all randomized patients who provided at least one baseline efficacy assessment and attended at least one session. Nine-week and 21-week last-observation-carried-forward (LOCF) analyses were carried out for missing values. Baseline characteristics were summarized according to treatment sequences and compared using an un-paired t-test for continuous variables and a chi-squared test for categorical variables. MCET effects on cognition, behavioral symptoms, psychological symptoms, and functional level, were assessed using a linear mixed model. Period and treatment were included in the model as fixed effects. Patients were included in the model as a random effect. Models were adjusted for age, gender, education, current cognitive enhancer use, diagnosis (MCI or dementia), compliance to intervention, and research centers. Effect size was calculated as described by Cohen [38]. For all analyses, a two-sided p-value of <0.05 was considered statistically significant. All statistical analyses were performed using Predictive Analytics Software Statistics 18 software, Release Version 18.0.0 (SPSS, Inc., 2009, Chicago, IL).
RESULTS
Participants
Patients receiving MCET during the first period (N = 32) did not statistically differ by any demographic or clinical characteristics from those who first received MT (N = 32), as shown in Table 1. Progress for all patients is shown in Fig. 1. Four patients with Alzheimer’s disease dropped out in the first period (2 in MCET group and 2 in MT group); one patient allocated to the MCET group during the first period refused to get a baseline evaluation, and three patients discontinued due to transportation difficulties. After the 4-week washout, another three patients with Alzheimer’s disease could not participate in the second-period crossover therapy because their caregivers could not take them to the study centers. Another two patients with vascular dementia and MCI amnestic multiple type from MT group dropped out in the second period; one patient due to a new medical condition, and one due to a bereavement reaction. There were no adverse events related to MCET or MT.
Compliance
Among the 60 patients who completed the allocated treatment during the first period, 59 showed an attendance rate above 75%. The remaining patient in the MCET group showed an attendance rate of 58%, who could not attend some sessions due to caregiver-related personal reasons. Among the 55 patients who completed the treatment during the second period, 54 showed an attendance rate above 85%. The remaining patient in the MCET group showed an attendance rate of 54.2% due to an accidental fall in his residence.
Efficacy
Changes in MMSE, ADAS-cog, RMBPC, GDS, DAD, and QoL-AD scores are presented in Table 2 and Fig. 3. MCET was more beneficial than MT in global cognitive function measures; the effect between MCET versus MT was 0.85±2.02 versus –0.17±2.32 points (effect size = 0.47, p = 0.013) for the MMSE, and – 1.95±3.68 versus – 0.64±3.85 points (effect size = 0.35, p = 0.045) for the ADAS-cog. In the MCET group, 58.3 and 70.0% of subjects showed improvement in MMSE and ADAS-cog scores, respectively, whereas, in the MT group, significantly fewer subjects showed improvement in MMSE (43.3%, p = 0.036 for linear-by-linear association) and ADAS-cog (48.3%, p = 0.035 for linear-by-linear association) scores. Among secondary measures, MCET was more beneficial than MT in the RMBPC-F and patient-rated QoL-AD assessments. Greater improvements were found for RMBPC-F measures in the MCET compared with MT group (–1.73±7.00 versus 0.93±7.12, effect size = 0.38, p = 0.046). Among the RMBPC-F subscores, greater improvements were found in the MCET group for depression score (–0.98±3.41 versus 0.82±3.52 points; effect size = 0.52, p = 0.006) and disruption factor (–0.50±1.95 versus 0.52±1.98 points; effect size = 0.52, p = 0.007). However, changes in the memory scores were comparable between the MCET and MT groups. There were no significant differences in changes of RMBPC-R measures between the MCET and MT groups. Changes in GDS measures were not significantly different.
No significant differences in DAD scores were found between the MCET and MT groups. Significant differences were also absent when we analyzed DAD-ADL and DAD-IADL scores separately.
MCET was also more beneficial than MT in measures of patient-rated QoL-AD (1.47±5.53 versus –0.38±3.97 points; effect size = 0.39, p = 0.047) but not caregiver-rated QoL-AD. Carryover effects were not significant in all primary and secondary outcome measures.
DISCUSSION
MCET, an eight-week, tri-weekly, half-day MNPI program for individuals with MCI or mild dementia (consisting of cognitive training, cognitive stimulations, reality orientation, physical therapy, reminiscence therapy, and music therapy) was more effective in improving measures of global cognition, BPSD, and self-rated QoL than placebo therapy.
MCET was superior to MT in both primary outcome measures of global cognition (the MMSE and ADAS-cog). The effect sizes were 0.47 for the MMSE and –0.35 for the ADAS-cog, which were comparable to those observed in a previous meta-analysis [22]. However, the observed MCET effect size was smaller than that of six-month cholinesterase inhibitor treatment (1.37 and –2.37 for the MMSE and ADAS-cog, respectively) [39]. The difference in effect size between MCET and cholinesterase inhibitors may be attributed, in part, to at least two factors. First, we employed an active mock therapy for the comparison group, rather than an inactive placebo. Second, cholinesterase inhibitors were provided every day for six months, whereas MCET was provided tri-weekly for only two months.
Improvement of mood measured by GDS in MCET group was higher than that of MT group, but the group difference was not significant. Because increase in social attention and interaction has been noted to improve mood and behavior in demented elderly, so mood gain could be a nonspecific benefit [11] from both MT and MCET. The effect on global cognition in MCET might result from MCET’s proper composition not from social interaction.
BPSD were reported in 35 to 85% of patients with dementia, which increase caregiver burden and institutionalization risk [40]. In a previous meta-analysis, unimodal nonpharmacological interventions (such as cognitive training or stimulation) were not effective in improving depression or BPSD [22, 41]. However, in the present study, patients in the MCET group showed greater improvements in RMBPC-F scores than those in the MT group. These data indicate that MCET may be effective in reducing BPSD in individuals with MCI or mild dementia. When we analyzed the effect of MCET on RMBPC-F subscores separately, MCET effectively improved depression and disruption scores. The effect size of the MCET was 0.38, which was much larger than values reported in previous studies (0.06–0.20) [22, 41]. However, MCET did not significantly improve RMBPC-R in any of the three subscores. The observed discrepancy in the effects of MCET on RMBPC-F and RMBPC-R scores may be attributed to a few factors. First, RMBPC-R baseline subscores were lower than corresponding RMBPC-F subscores, potentially leading to a ceiling effect among measures. Since all participants were diagnosed with MCI or early stage dementia, symptoms might not have changed greatly enough to be recognized by caregivers. Secondly, a lack of interventions for caregivers in the MCET group might have also played a role in this discrepancy. Although the mechanisms underlying the observed MCET efficacy for these BPSD are unknown, RMBPC-F scores were significantly worse in the MT group. Thus, these results may not be due to a nonspecific benefit from groupinterventions.
MCET was not superior to MT in improving DAD scores. This result is consistent with previous studies that reported insignificant or low effect sizes for cognitive training or stimulations on ADL outcomes (0.05–0.21) [22, 41]. A comprehensive stimulation program for patients with AD is hypothesized to enhance neuroplasticity processes, reduce cognitive loss, and help patients improve functional independence via increased cognitive performance [11]. However, there are no grounds for assuming that cognitive benefits will generalize to result in observable benefits in daily life [41]. Many tasks and evaluations designed to directly assess skills targeted in cognitive training intervention might produce significant benefits in task outcomes [41]. Thus, specific measures of daily function, rather than overall measures, might reflect actual MCET benefits. Since only individuals with MCI or mild dementia were included, ADL impairments among our participants were more modest than those found in previous studies; this may have additionally limited statistical power for detecting DAD score changes.
Although QoL-AD was not extensively assessed as an outcome measure in previous trials on MNPIs, QoL may be a useful efficacy outcome measure, as it reflects overall improvement via “real-world” treatment effects on factors such as mood, cognition, social relations, functional activities, energy level, and volition [8]. In this study, the MCET effect size on self-rated QoL-AD was 0.39, the second largest among those on all outcome measures. Recently developed MNPIs have also improved self-rated quality of life [42, 43]. MNPIs, such as MCET, require substantial time and effort on behalf of patients and caregivers. However, patients report improved quality of life following intervention, suggesting that therapeutic benefits outweigh the costs[42, 43].
The current study has several strengths. First, we developed MCET using only nonpharmacological interventions with “grade-A” recommendations, as determined by a meta-analysis [23]. This may have contributed to the excellent efficacy observed by MCET, despite administration tri-weekly for only eight weeks. Second, the MCET was effective with less frequent sessions and shorter durations than those for previously-reported MNPIs [5–11, 15]. Second, we standardized MCET and provided handbooks for therapists. MCET was therefore easily implemented in various settings, requiring little therapist preparation time. Third, we minimized ascertainment bias by adopting a double-blind design; we also minimized the potential confounding effect of simple social interactions or non-specific cognitive stimulations by employing mock therapy rather than usual care for the comparison group. Lastly, a major shortcoming of crossover trials is the “carry over” effect, which alters responses to subsequent treatments during the study. A useful strategy to overcome such pitfalls is to allow a washout period between consecutive treatments long enough to allow treatment effects to wear off. We believe our study had sufficiently long washout periods (four weeks); however, there has been no consensus on the exact washout duration to avoid potential treatment interactions in this population.
The present study also had several limitations, which should be noted. First, as in other MNPI studies, we could not evaluate the efficacy of each intervention component separately. Second, although we adjusted the effect of concomitant medications by including the use of cognitive enhancers and medications for BPSD in the analysis, we could not adjust medication dosages. Third, the small sample size may have reduced the statistical power of the study. Fourth, given the diagnostic heterogeneity of the type of dementia and MCI included in this study, the effect of MCET in each specific type of dementia or MCI remains unclear. Finally, the relatively short period of intervention may have limited the effect of MCET. It is necessary to determine the optimal treatment period that has a long-term effect infuture.
In conclusion, MCET improved global cognition, BPSD, and QoL of people with MCI or mild dementia more effectively than the conventional cognitive enhancing activities that are widely employed in care facilities for dementia patients in Korea. It is necessary to encourage care facilities for people with dementia to provide an evidence-based MNPI like the MCET.
