The Association Between Diabetes Mellitus and Mild Behavioral Impairment Among Mild Cognitive Impairment: Findings from Singapore

Abstract

Background:

Mild behavioral impairment (MBI) describes persistent behavioral changes in later life as an at-risk state for dementia. While cardiovascular risk factors (CVRFs) are linked to dementia, it is uncertain how CVRFs are associated with MBI.

Objective:

To determine the prevalence of MBI and its association with CVRFs among cognitively normal (CN) and mild cognitive impairment (MCI) individuals in Singapore.

Methods:

172 individuals (79 CN and 93 MCI) completed the MBI-checklist (MBI-C). The prevalence of MBI and MBI-C sub-domain characteristics among CN and MCI were examined. Regression models evaluated the relationships between MBI-C sub-domain scores with CVRFs.

Results:

The prevalence of MBI and mean MBI-C total score were significantly higher among MCI than CN (34.4%versus 20.3%, p = 0.022 and 7.01 versus 4.12, p = 0.04). The highest and lowest-rated sub-domains among CN and MCI were impulse dyscontrol and abnormal thoughts and perception respectively. Within the MCI cohort, a higher proportion of individuals with diabetes mellitus (DM) had MBI compared to individuals without DM (28.1%versus 10.4%, p = 0.025). The interaction of DM and MCI cohort resulted in significantly higher mean MBI-C total, decreased motivation, emotional dysregulation, impulse dyscontrol, and abnormal thoughts and perception sub-domain scores.

Conclusion:

The prevalence of MBI is higher among a Singapore cohort compared to Caucasian cohorts. The associations of DM with both the presence and severity of MBI among MCI suggest that DM may be a risk factor for MBI. The optimization of DM may be a potential therapeutic approach to improve clinical outcomes among MCI with MBI.

Keywords

Behavior cardiovascular diseases cognitive dysfunction dementia diabetes mellitus

INTRODUCTION

Mild behavioral impairment (MBI) is characterized by the emergence of persistent behavioral changes in later life as an at-risk state for cognitive decline and dementia. For some, MBI is the initial manifestation of dementia [1 –3]. The concept of MBI is informed by the notion that neuropsychiatric symptoms can be non-cognitive markers of neurodegenerative disease [4], and operationalizes it for more specific case detection in preclinical and prodromal disease states [5]. It is imperative to study MBI in pre-dementia individuals as the prevalence of neuropsychiatric symptoms among mild cognitive impairment (MCI) is high, ranging from 35%to 85%[6] and the presence of neuropsychiatric symptoms is associated with increased risk of progression from MCI to dementia from 10%to 25%[5]. MBI is also linked to detrimental effects such as worse cognitive performance in older adults [7] and increased caregiver burden [8].

The MBI-Checklist (MBI-C) is the validated instrument to case-find MBI in accordance with the ISTAART-AA MBI criteria [5]. The MBI-C is explicit that symptoms emerge in later life representing a change from longstanding patterns of behavior and are present for a 6-month period, thus increasing specificity by reducing false positive cases due to transient symptoms, life stressors or other medical issues [3]. To date, only a few studies have used the MBI-C to operationalize MBI and are mostly limited to Caucasian populations [9]. A memory clinic in India determined the frequency of MBI to be 41.12%among a mix cohort of MCI and subjective cognitive decline (SCD) patients [10]. A population-based study in the United Kingdom determined the frequency of MBI to be 10%among cognitively normal (CN) individuals [11], and a primary care population in rural Spain determined the frequency of MBI to be 5.8%among SCD [12] and 14.2%among MCI patients [13]. However, the frequency of MBI in a Japanese psychiatric outpatient clinic population was 3.5%among a mix cohort of MCI and SCD patients [14]. The published frequencies of MBI vary between diagnoses and ethnicities and the data on the prevalence of MBI among a Singapore population focusing on CN and MCI remains undetermined.

Cardiovascular risk factors (CVRFs) such as diabetes mellitus (DM), hyperlipidemia, and hypertension were found to be associated with neuropsychiatric symptoms among Alzheimer’s disease (AD) and MCI patients. Among AD patients, elevated hyperlipidemia had been positively associated to neuropsychiatric symptoms [15], DM was found to be an independent risk factor for neuropsychiatric symptoms [16], and hypertension was found to be associated with neuropsychiatric symptoms such as depression, anxiety, and apathy [17]. Among MCI patients, neuropsychiatric symptoms such as agitation, anxiety, disinhibition, and irritability were found to be associated with increased CVRFs [18]. Another study demonstrated similar findings and found that CVRFs were associated with depression, anxiety, disinhibition, and irritability [19]. MCI and SCD individuals with DM were also found to have significantly higher MBI than the non-DM group [10]. Given that CVRFs are highly prevalent among individuals at later life, elucidating the prevalence of CVRFs in individuals with MBI and evaluating the associations between CVRFs with MBI are paramount as the optimization of CVRFs may be a potential therapeutic approach to improve outcomes in MBI.

Here in a real-world community and clinic setting, we compared the prevalence of MBI measured using the MBI-C among Singapore CN individuals in the community and MCI patients in a memory clinic. We further studied the prevalence of CVRFs in MBI and the associations of CVRFs with MBI. We hypothesized that the prevalence of MBI among a Singapore population will be higher among MCI cohort compared to CN cohort and that CVRFs will be associated with MBI.

MATERIALS AND METHODS

Study participants

One hundred and seventy-two participants (79 CN community-dwelling individuals and 93 MCI patients from National Neuroscience Institute of Singapore) aged 50 years and above were recruited for this study. Community-dwelling individuals were recruited via e-mails sent to senior activity centers, government organizations, and social development groups across Singapore. Diagnosis of MCI was made by cognitive neurologists based on Petersen’s criteria [20] and the National Institute on Aging-Alzheimer’s Association (NIA-AA) criteria [21]. Individuals with MCI presented with a predominant memory deficit but had normal ability to perform daily functions. Participants were excluded if they had prior diagnosis of major psychiatric diseases (e.g., schizophrenia, major depression) or comorbid neurodegenerative disease (e.g., dementia, Parkinson’s disease).

Ethics approvals and patients consents

Informed consent was obtained from all participants according to the Declaration of Helsinki and local clinical research regulations, and procedures used in the study were in accordance with ethical guidelines. The study was granted approval by the Singhealth Centralized Review Board.

Cognitive screening and assessment

Subjects underwent an interview comprising a set of standardized questions about demographic characteristics and CVRFs (presence of DM, hypertension, or hyperlipidemia). The presence of CVRFs were determined by the participant’s primary physician and the information on the duration and medication status of CVRFs were traced from the patient’s medical records.

Global cognitive functioning was assessed using the locally modified and validated Mini-Mental State Examination (MMSE) [22] and the Montreal Cognitive Assessment (MoCA) [23]. MBI was measured using the MBI-C self-report, is a 2-page questionnaire consisting of 34-items in 5 sub-domains: decreased motivation (apathy), emotional dysregulation (mood and anxiety symptoms), impulse dyscontrol (agitation, aggression, response inhibition, social inappropriateness (impaired social cognition), and abnormal thoughts and perception (psychotic symptoms). Changes in symptoms that were different from baseline behaviors had to be present for at least 6 months. The items were scored “yes” or “no”, followed by a severity rating of 1-mild, 2-moderate, and 3-severe [5]. Patients were stratified into MBI and no MBI using the recently published cut-off score of≥6.5 for MCI [13]. As the cut-off score for the diagnosis of MBI among CN individuals has not been established, we used the same cut-off score as SCD≥8.5 [12] for CN individuals given that their cognitive profile is comparable to those with SCD.

Statistical analysis

Baseline characteristics were compared between CN and MCI cohorts using two independent sample t-test and chi-square test for continuous and categorical variables, respectively. Multivariable linear and logistic regression models were performed to determine MBI prevalence and MBI-C sub-domains scores between CN and MCI cohorts, while controlling for confounders including age, sex and education. Group differences between individuals with MBI and without MBI in CN and MCI were examined for demographics and CVRFs using two independent sample t-test for continuous variables and chi-square test for categorical variables. Moderation analysis of CVRFs and cohorts (MCI versus CN) on outcomes MBI-C total score and MBI-C sub-domains scores were investigated using two-way interaction terms in multiple linear regression controlling for age, sex, education, duration of CVRFs, and medication status. Significance level was set at p < 0.05. Statistical analysis was performed using the IBM SPSS Statistics for Macintosh, Version 25.0 (IBM SPSS Statistics for Macintosh, IBM Corporation, Armonk, NY).

Data availability statement

The dataset analyzed during the current study are not publicly available but is available upon request from the corresponding author.

RESULTS

Demographic, CVRFs, and cognitive differences between CN and MCI

There were a total of 172 participants consisting of CN (n = 79) and MCI (n = 93). MCI had a significantly older age than CN (69.08±7.75 versus 63.86±7.79, p < 0.001) and MCI had significantly more males than CN (51.6%versus 34.2%, p = 0.022). The mean number of years of education was significantly lower among MCI than CN (10.44±4.65 versus 12.51±3.44, p = 0.002) (Table 1).

Table 1
Demographic and baseline clinical features between CN and MCI

CN (n = 79) MCI (n = 93) p ^*

Age (y) 63.86±7.79^† 69.08±7.75 < 0.001

Sex (Male) 27 (34.2) 48 (51.6) 0.022

Years of Education (y) 12.51±3.44 10.44±4.65 0.002

Hyperlipidemia 40 (50.6) 55 (59.8) 0.230

Duration of hyperlipidemia (y) 7.75±9.79 7.02±10.15 0.726

Hyperlipidemia medication (yes) 21 (52.5) 18 (32.7) 0.053

Hypertension 29 (36.7) 47 (51.1) 0.059

Duration of hypertension (y) 12.86±12.67 6.70±7.96 0.024

Hypertensive medication (yes) 21 (72.4%) 16 (34.0) 0.001

DM 11 (13.9) 16 (17.2) 0.556

Duration of DM (y) 8.73±6.56 12.31±9.13 0.275

DM medication (yes) 7 (63.6%) 5 (31.3) 0.096

MMSE 28.70±1.18 25.99±3.53 < 0.001

MoCA 27.70±2.10 24.78±4.32 < 0.001

	CN (n = 79)	MCI (n = 93)	p ^*
Age (y)	63.86±7.79^†	69.08±7.75	< 0.001
Sex (Male)	27 (34.2)	48 (51.6)	0.022
Years of Education (y)	12.51±3.44	10.44±4.65	0.002
Hyperlipidemia	40 (50.6)	55 (59.8)	0.230
Duration of hyperlipidemia (y)	7.75±9.79	7.02±10.15	0.726
Hyperlipidemia medication (yes)	21 (52.5)	18 (32.7)	0.053
Hypertension	29 (36.7)	47 (51.1)	0.059
Duration of hypertension (y)	12.86±12.67	6.70±7.96	0.024
Hypertensive medication (yes)	21 (72.4%)	16 (34.0)	0.001
DM	11 (13.9)	16 (17.2)	0.556
Duration of DM (y)	8.73±6.56	12.31±9.13	0.275
DM medication (yes)	7 (63.6%)	5 (31.3)	0.096
MMSE	28.70±1.18	25.99±3.53	< 0.001
MoCA	27.70±2.10	24.78±4.32	< 0.001

^†Numbers reported as mean±standard deviation and frequency (%) for numeric and categorical variables, respectively. ^*Two independent sample t-test and chi-square test for continuous and categorical variables, respectively. DM, diabetes mellitus; MMSE, Mini-Mental State Examination; MoCA, The Montreal Cognitive Assessment; CN, cognitively normal; MCI, mild cognitive impairment.

In terms of CVRFs, there were significant differences in the duration and medication status of hypertension between MCI and CN. There were no significant difference in the frequency of hypertension and in the frequency, duration and medication status of hyperlipidemia and DM between MCI and CN (Table 1).

In terms of cognition, MCI had significantly lower MMSE and MoCA scores compared to CN (MMSE: 25.99±3.53 versus 28.70±1.18, p < 0.001; MoCA: 24.78±4.32 versus 27.70±2.10, p < 0.001) (Table 1).

Prevalence and characteristics of MBI between CN and MCI

The prevalence of MBI was significantly higher among MCI than CN (34.4%versus 20.3%, OR = 2.48, CI = 1.14 to 5.36, p = 0.022). The MBI-C total score and distribution of total score was significantly higher among MCI compared to the CN (MCI adjusted mean = 7.01, CI = 5.24 to 8.77 versus CN adjusted mean = 4.12, CI = 2.14 to 6.10, p = 0.040). The MBI-C sub-domain score for decreased motivation was significantly higher among MCI compared to CN (MCI adjusted mean = 1.92, CI = 1.34 to 2.49 versus CN adjusted mean = 0.87, CI = 0.22 to 1.51, p = 0.022). The MBI-C sub-domain score for emotional dysregulation was significantly higher among MCI compared to CN (MCI adjusted mean = 2.20, CI = 1.59 to 2.82 versus CN adjusted mean = 1.23, CI = 0.54 to 1.92, p = 0.047) (Table 2).

Table 2

MBI prevalence and MBI-C sub-domain differences between CN and MCI

	CN (n = 79)	MCI (n = 93)	Odds Ratio (95%CI)	p ^*
MBI Prevalence, n (%)	16 (20.3)	32 (34.4)	2.48 (1.14, 5.36)	0.022
	CN Adj. Mean (95%CI)	MCI Adj. Mean (95%CI)	Beta Coefficient (95%CI)	p ^*
MBI-C Total Score	4.12 (2.14, 6.10)	7.01 (5.24, 8.77)	2.89 (0.13, 5.65)	0.040
Decreased Motivation	0.87 (0.22, 1.51)	1.92 (1.34, 2.49)	1.05 (0.15, 1.95)	0.022
Emotional Dysregulation	1.23 (0.54,1.92)	2.20 (1.59, 2.82)	0.97 (0.01, 1.93)	0.047
Impulse Dyscontrol	1.66 (0.93, 2.38)	2.48 (1.83, 3.12)	0.82 (–0.19, 1.83)	0.110
Social Inappropriateness	0.26 (0.06, 0.45)	0.37 (0.19, 0.55)	0.11 (–0.16, 0.39)	0.419
Abnormal Thoughts and Perception	0.14 (–0.02, 0.30)	0.24 (0.10, 0.39)	0.10 (–0.13, 0.32)	0.385

^*Multivariate logistic/linear analysis (for binary and numeric outcomes respectively) controlling for age, sex and education. P values indicate whether frequency/adjusted means were significantly different for MCI in reference to CN. Adj, adjusted; CI, confidence interval; MBI, mild behavioral impairment; MBI-C, Mild Behavioral Impairment-Checklist; CN, cognitively normal; MCI, mild cognitive impairment.

We found that the impulse dyscontrol sub-domain was rated the highest across CN and MCI (CN adjusted mean = 1.66, CI = 0.93 to 2.38 and MCI adjusted mean = 2.48, CI = 1.83 to 3.12). The abnormal thoughts and perception sub-domain was rated the lowest across CN and MCI (CN adjusted mean =0.14, CI=–0.02 to 0.30 and MCI adjusted mean =0.24, CI = 0.10 to 0.39) (Table 2).

Prevalence of DM among individuals with MBI in CN and MCI

Within the MCI cohort, a significantly higher proportion of individuals with MBI had DM compared to those without MBI (28.1%versus 10.4%, p = 0.025). However, we did not find any significant associations between the duration and medication status of DM with the prevalence of MBI and MBI-C total and sub-domain scores (p > 0.05). There was no significant difference in the frequency of DM among individuals with or without MBI within the CN cohort. Additionally, there were no significant differences in demographics, frequency of hyperlipidemia and hypertension and cognitive scores among individuals with or without MBI in both the CN and MCI cohorts (Table 3).

Table 3

Differences in demographics between MBI positive and negative among CN and MCI

	CN			MCI
	MBI- (n = 63)	MBI+ (n = 16)	p ^*	MBI- (n = 61)	MBI+ (n = 32)	p ^*
Age (y)	64.05±7.11^†	63.13±10.30	0.740	69.66±8.06	67.99±7.12	0.328
Gender (Male)	23 (36.5)	4 (25.0)	0.386	31 (46.3)	18 (56.3)	0.353
Years of Education (y)	12.24±3.37	13.56±3.60	0.170	10.41±5.08	10.50±3.74	0.958
Hyperlipidemia	31 (49.2)	9 (56.3)	0.615	38 (56.7)	19 (61.3)	0.669
Hypertension	24 (38.1)	5 (31.3)	0.612	31 (46.3)	16 (51.6)	0.622
DM	9 (14.3)	2 (12.5)	0.854	7 (10.4)	9 (28.1)	0.025
MMSE	28.71±1.16	28.63±1.31	0.789	25.97±3.72	26.03±3.19	0.934
MOCA	27.66±2.16	27.87±1.92	0.737	24.52±4.81	25.30±3.22	0.396

^†Numbers reported as mean±standard deviation and frequency (%) for numeric and categorical variables, respectively. ^*Two independent sample t-test and chi-square test for continuous and categorical variables, respectively. MBI+, mild behavioral impairment positive group; MBI-, mild behavioral impairment negative group; DM, diabetes mellitus; CN, cognitively normal; MCI, mild cognitive impairment.

Association between CVRFs and MBI-C scores

DM significantly interacted with MCI cohort compared to CN cohort on mean MBI-C total score (p = 0.008) (Table 4); where the MBI-C total score is significantly higher in MCI compared to CN (19.81 (13.73, 25.89) versus 8.30 (2.70, 13.91) in DM group; 6.91 (3.17, 10.65) versus 5.26 (1.55, 8.98) in No-DM group (Fig. 1).

Table 4

Moderation analysis of DM with cohorts (MCI versus CN) on MBI-C total and sub-domain scores

Variable	DM		MCI (versus CN)		p of the interaction term (DM X MCI)^*
	Beta Coefficient (95%CI)	p ^*	Beta Coefficient (95%CI)	p ^*
MBI-C Total Score	3.04 (–3.99, 10.07)	0.396	1.65 (–1.15, 4.44)	0.249	9.86 (2.59, 17.13)	0.008
Decreased Motivation	0.85 (–1.44, 3.14)	0.466	0.77 (–0.14, 1.68)	0.099	2.94 (0.58, 5.31)	0.015
Emotional Dysregulation	0.19 (–2.33, 2.70)	0.885	0.66 (–0.35, 1.66)	0.200	2.61 (0.01, 5.22)	0.049
Impulse Dyscontrol	2.06 (–0.52, 4.64)	0.118	0.33 (–0.70, 1.36)	0.528	2.81 (0.14, 5.48)	0.039
Social Inappropriateness	0.20 (–0.53, 0.93)	0.586	0.10 (–0.19, 0.39)	0.509	0.18 (–0.57, 0.94)	0.625
Abnormal Thoughts and Perception	-0.27 (–0.83, 0.30)	0.358	0.03 (–0.20, 0.25)	0.815	0.94 (0.36, 1.53)	0.002

^*Multiple linear regression controlling for age, sex and education, medication status, and duration of diabetes MBI, mild behavioral impairment; MBI-C, Mild Behavioral Impairment-Checklist; DM, diabetes mellitus; CN, cognitively normal; MCI, mild cognitive impairment; CI, confidence interval.

Fig. 1

Moderation analysis of DM with MCI (compared to CN) on mean MBI-C total score. Presence of DM significantly interacted with MCI cohort compared to the CN cohort on the mean MBI-C total score. DM, diabetes mellitus; MCI, mild cognitive impairment; CN, cognitively normal; MBI-C, Mild Behavioral Impairment-Checklist.

DM significantly interacted with MCI cohort compared to CN cohort on mean MBI-C sub-domain scores for decreased motivation (p = 0.015) (Table 4), where the MBI-C sub-domain score for decreased motivation is significantly higher in MCI compared to CN (6.21 (4.23, 8.18) versus 2.50 (0.68, 4.32) in DM group; 2.41 (1.20, 3.63) versus 1.65 (0.44, 2.86) in No-DM group (Fig. 2A).

Fig. 2

Moderation analysis of DM with MCI (compared to CN) on mean MBI-C sub-domain scores for Decreased Motivation, Emotional Dysregulation, Impulse Dyscontrol, and Abnormal Thoughts and Perception. DM significantly interacted with MCI cohort compared to CN cohort on mean MBI-C sub-domain scores for decreased motivation (A), emotional dysregulation (B), impulse dyscontrol (C), and abnormal thoughts and perception (D). DM, diabetes mellitus; MCI, mild cognitive impairment; CN, cognitively normal; MBI-C, Mild Behavioral Impairment-Checklist.

DM significantly interacted with MCI cohort compared to CN cohort on mean MBI-C sub-domain scores for emotional dysregulation (p = 0.049) (Table 4), where the MBI-C sub-domain score for emotional dysregulation is significantly higher in MCI compared to CN (5.12 (2.94, 7.29) versus 1.85 (–0.16, 3.86) in DM group; 2.32 (0.98, 3.66) versus 1.66 (0.33, 2.99) in No-DM group (Fig. 2B).

DM significantly interacted with MCI cohort compared to CN cohort on mean MBI-C sub-domain scores for impulse dyscontrol (p = 0.039) (Table 4), where the MBI-C sub-domain score for impulse dyscontrol is significantly higher in MCI compared to CN (6.30 (4.07, 8.54) versus 3.16 (1.10, 5.22) in DM group; 1.43 (0.06, 2.80) versus 1.10 (–0.27, 2.46) in No-DM group (Fig. 2C).

DM significantly interacted with MCI cohort compared to CN cohort on mean MBI-C sub-domain scores for abnormal thoughts and perception (p =0.002) (Table 4), where the MBI-C sub-domain score for abnormal thoughts and perception is significantly higher in MCI compared to CN (1.25 (0.76, 1.74) versus 0.78 (–0.17, 0.74) in DM group; 0.58 (0.27, 0.88) versus 0.55 (0.25, 0.85) in No-DM group (Fig. 2D).

There was no significant association between DM and MCI cohort compared to CN cohort on mean MBI-C sub-domain score for social inappropriateness. Both hypertension and hyperlipidemia were not significantly associated with MCI cohort compared to CN cohort on mean MBI-C total or sub-domain scores.

4 DISCUSSION

In this study, we determined the prevalence and the characteristics of MBI among Singapore CN and MCI cohorts using the self-report MBI-C. We demonstrated that the prevalence of MBI was significantly higher in MCI compared to the CN cohort (34.4%versus 20.3%). The MBI-C total adjusted mean score was also significantly higher in MCI compared to the CN cohort (7.01 versus 4.12). The MBI-C impulse dyscontrol sub-domain was rated the highest among CN and MCI. Abnormal thought and perception sub-domain was rated the lowest among CN and MCI. We further found that within the MCI cohort, a significantly higher proportion of individuals with MBI had DM compared to those without MBI. DM interacted with MCI in reference to CN cohort to result in higher MBI-C total score and MBI-C sub-domain scores for decreased motivation, emotional dysregulation, impulse dyscontrol and abnormal thoughts and perception. In addition to supporting the current literature that MBI is prevalent in both CN and MCI cohorts, our findings further demonstrate that DM is associated with both the presence and the severity of MBI.

Our findings support MBI-C as an instrument to case-find MBI and further contribute to the emerging conceptual framework of MBI as a neurobehavioral at-risk state for incident cognitive decline and dementia. MBI was more prevalent in MCI (34.4%) compared to CN (20.3%). These findings are consistent with previous literature suggesting that MBI frequency increases with cognitive decline [12 , 24]. An additional reason for this finding may be that the CN cohort was a community sample, while the MCI cohort was from a specialty clinic. Evidence has suggested that neuropsychiatric symptoms are more prevalent in clinical versus community samples, emphasizing the clinical significance of these symptoms, and the likelihood that they result in presentation to clinical care [25].

Our prevalence estimates of MBI in both MCI and CN were higher than the corresponding rates from the Spanish primary care samples [12, 13]. The underlying reason as to why a Singapore MCI cohort had a higher prevalence of MBI than Caucasian populations is unclear. However, MBI in a Singapore cohort may be linked to both modifiable risk factors (such as CVRFs) and non-modifiable risk factors (such as genetics), which require further clarification. In this regard, previous studies have found greater frequency of psychiatric symptoms in Asian versus Caucasian populations [26]; this corresponds to our finding of higher prevalence of MBI in a Singapore population. Differences in Singapore and Caucasian populations may also be related to cultural differences. For instance, apathy is reported to be more difficult to identify in Asian populations [27] and Asian populations are more likely to express psychological distress as physical complaints compared to Caucasian populations [28]. Additionally, the higher prevalence of MBI among a Singapore population may be due to sample differences; the Caucasian MCI population was recruited from a rural primary health care center with a mean year of education for MCI to be 7.89 (±4.07) [13], whereas the Singapore MCI population was recruited from an urban specialized memory clinic with a higher mean education year 10.44 (±4.65). Previous literature have demonstrated that more educated people were more likely to report more depressive symptoms [29]. Therefore, the higher frequency of MBI among the Singapore population may be due to the higher mean years of education in the cohort.

Another difference between the Singapore sample and the previously reported Caucasian sample relates to methodology. In the Singapore sample, MBI was captured using a self-report of symptoms, in contrast to an informant report in the Spanish study. Differences between self- and informant-reported MBI have not been well explored in the literature; however, in an MBI-C validation and factor analysis, self- and informant-reported MBI-C had a similar factor structure, but somewhat different psychometric properties [11]. Another methodological difference between our study and previous MBI prevalence studies has to do with use of the MBI-C. We determined that the frequency of MBI using the MBI-C was lower compared to MBI frequencies generated using a single time point administration of the neuropsychiatric inventory, which were considered to be inflated due to the one month reference range of the neuropsychiatric inventory [30]. These findings support the MBI-C as a more specific instrument for MBI, consistent with its a priori development plan, relative to more traditional measures of psychiatric symptomatology.

A more comparable approach to ours for MBI case ascertainment was used in the United Kingdom PROTECT cohort, in which self-reported MBI-C prevalence was 10%in CN older adults, but still lower than our Singapore sample prevalence of 20.3%[24]. In PROTECT, CN individuals who had MBI exhibited greater decline in cognitive performance in attention and working memory over one year from baseline, indicating that MBI may be an early non-cognitive manifestation of an underlying neurodegenerative disease in the pre-dementia stages [24]. These findings are consistent with more recent evidence demonstrating that MBI in CN older adults is associated with higher odds of incident MCI at 3 years [2]. Consequently, MBI-C may be used as a tool to case-find MBI in a CN cohort who may be at-risk for dementia and hence early prevention strategies may be instituted [24].

The most common MBI-C sub-domain reported among CN and MCI individuals in our Singapore cohort were different compared to a Caucasian CN cohort [11]. The characteristics of MBI-C sub-domains among a Singapore population may serve to inform clinicians on what to identify as a precursor to MCI. Similar to the study done in India [10], impulse dyscontrol was the most common MBI-C sub-domain among CN and MCI. It is postulated that impulse dyscontrol was the highest rated sub-domain among Singapore CN and MCI individuals because the symptoms under this sub-domain could adversely affect the relationship between patient and caregiver, who is typically the spouse. This postulation is supported by a study of individuals with dementia which exemplify how disruptive behaviors such as irritability and aggression negatively impact the relationship between patient and caregiver [31]. Further studies have also demonstrated that among a Caucasian sample of MCI individuals at most risk of developing AD, depression and apathy were the highest rated neuropsychiatric symptoms [32]. Therefore, among a Singapore CN and MCI cohort, clinicians may identify changes in behavior related to impulse dyscontrol such as irritability, argumentative, and impulsivity as risk symptoms of MBI. Moreover, to support why impulse dyscontrol sub-domain was the most common domain among the MCI cohort, a recent network analysis of the MBI-C impulse dyscontrol sub-domain among MCI in a cognitive neurology clinic sample determined the core features to be rigidity, agitation/aggressiveness, and argumentativeness to be the most frequent and core features of this sub-domain [33]. The least common MBI-C sub-domain among the Singapore cohort was abnormal thoughts and perception, which was also homogenously reported in an Indian population [10] as well as in a Caucasian cohort [3, 11], indicating that hallucinations and delusions are unanimously the least common MBI symptoms among a Singaporean, Indian, and Caucasian populations.

Our findings on the interaction of DM with MCI cohort on higher MBI-C sub-domain scores are novel, and elucidates the mechanisms of association between DM with MBI found in the study done in the Indian memory clinic [10]. DM was found to interact with MCI cohort in reference to CN cohort to result in higher MBI-C total score and higher sub-domain scores on decreased motivation, emotional dysregulation, impulse dyscontrol, and abnormal thoughts and perception. Previous literature had shown than DM may result in fatigue which may indirectly affect motivation levels [34]. Current literature also demonstrates that DM and psychiatric disorders share a bidirectional association; DM is a risk factor for the development of psychiatric disorders and vice versa. For instance, anxiety is associated with a higher prevalence of DM [35], and the presence of anxiety toward insulin needles and hypoglycemic episodes further worsen glycemic control. Depressive symptoms such as lack of physical activity and poor diet can lead to the increased risk of DM [36], and hyperglycemia can lead to reduced energy and increased sadness and anxiety [37]. Additionally, insulin resistance have been found to increase impulsivity and reduce self-regulation of behavior [38].

Despite being lower frequency symptoms, hallucinations and delusions in MCI are associated with a substantially greater risk of dementia [39]. The association between DM and MBI psychosis may offer an opportunity to explore prevention in this extremely high-risk sample. While the pathophysiology of DM in MBI requires further clarification, insulin signaling deficiencies and insulin resistance may be the potential mechanism underlying this relationship. The relationship between insulin resistance and cognitive impairment has been demonstrated in previous studies. Insulin is known to be responsible for regulating neurons in the central nervous system and thus, insulin signaling disruption can cause amyloid-β proteins to clump together, leading to toxicity to the brain associated with neurodegeneration [40]. Insulin resistance can also induce tau hyperphosphorylation and cognitive decline in human and in animal models [41]. Moreover, hyperglycemia in the blood produces the formation of lactic acid which can cause cellular acidosis, damaging the cells in the brain and impairing memory over time [42]. In terms of behavioral symptoms, brain insulin resistance has been shown to alter dopamine turnover and induces anxiety and depressive-like behaviors [43] while schizophrenia is associated with abnormalities in glucose that may lead to insulin resistance and a 3-fold higher incidence of type II DM [44]. A recent review describes overlapping mechanisms such as oxidative stress, neuroinflammation and hypothalamic-pituitary-adrenal axis dysregulation between insulin resistance with cognition and neuroprogression in bipolar disorder [45]. Further research is needed to elucidate the common pathophysiology between MBI and DM, specifically that of insulin resistance and the potential role of optimizing glycemic control in DM to improve cognitive and MBI outcomes.

There are limitations in our study. Firstly, our study has a relatively small sample size especially among individuals with DM and MBI. As this is a sub-group analysis to explore the association between DM and MBI, the findings will need to be confirmed in larger studies. Secondly, we do not have information on the individual’s glycemic control. Hence, we are not able to study the magnitude and analyze the association of the severity of DM with MBI. Thirdly, our current study used self-reported MBI-C instead of informant-reported MBI-C as with some previous studies. Our results may not be generalizable to other studies that use informant-reported MBI-C. Lastly, as our study is cross-sectional in design, future longitudinal studies are needed to evaluate the clinical progression of MCI individuals with a more pronounced MBI at baseline.

CONCLUSION

We demonstrate that the prevalence of MBI is high in a Singapore CN and MCI cohort and the value of MBI-C in case-finding MBI in CN and MCI populations. Our findings also highlight DM as a risk factor for MBI among MCI, particularly in the sub-domains of decreased motivation, emotional dysregulation, impulse dyscontrol, and abnormal thoughts and perceptions. While future studies should clarify the link between DM and MBI, optimizing DM among MCI may offer an important opportunity to improve clinical outcomes.

Footnotes

ACKNOWLEDGMENTS

We thank the participants, National Neuroscience Institute-Health Research Endowment Fund (NNI-HREF), Singapore (Reference Number: 991016) and SingHealth Central Institutional Review Board (CIRB) (Reference Numbers: 2010/063/A and 2017/2550) for supporting our study.

Authors’ disclosures available online ().

References

Ismail

, Smith

, Geda

, Sultzer

, Brodaty

, Smith

, Aguera-Ortiz

, Sweet

, Miller

, Lyketsos

(2016) Neuropsychiatric symptoms as early manifestations of emergent dementia: Provisional diagnostic criteria for mild behavioral impairment. Alzheimers Dement 12, 195–202.

Ismail

, McGirr

, Gill

, Hu

, Forkert

, Smith

(2021) Mild behavioral impairment and subjective cognitive decline predict cognitive and functional decline. J Alzheimers Dis 80, 459–469.

Lussier

, Pascoal

, Chamoun

, Therriault

, Tissot

, Savard

, Kang

, Mathotaarachchi

, Benedet

, Parsons

, Qureshi

MNI

, Thomas

ÉM

, Shin

, Dion

, Massarweh

, Soucy

, Tsai

, Vitali

, Ismail

, Rosa-Neto

, Gauthier

(2020) Mild behavioral impairment is associated with β-amyloid but not tau or neurodegeneration in cognitively intact elderly individuals. Alzheimers Dement 16, 192–199.

Wise

, Rosenberg

, Lyketsos

, Leoutsakos

(2019) Time course of neuropsychiatric symptoms and cognitive diagnosis in National Alzheimer’s Coordinating Centers volunteers. Alzheimers Dement (Amst) 11, 333–339.

Ismail

, Agüera-Ortiz

, Brodaty

, Cieslak

, Cummings

, Fischer

, Gauthier

, Geda

, Herrmann

, Kanji

, Lanctôt

, Miller

, Mortby

, Onyike

, Rosenberg

, Smith

, Sultzer

, Lyketsos

(2017) The Mild Behavioral Impairment Checklist (MBI-C): A rating scale for neuropsychiatric symptoms in pre-dementia populations. J Alzheimers Dis 56, 929–938.

Monastero

, Mangialasche

, Camarda

, Ercolani

, Camarda

(2009) A systematic review of neuropsychiatric symptoms in mild cognitive impairment. J Alzheimers Dis 18, 11–30.

Rouse

, Small

, Schinka

, Loewenstein

, Duara

, Potter

(2021) Mild behavioral impairment as a predictor of cognitive functioning in older adults. Int Psychogeriatr 33, 285–293.

Sheikh

, Ismail

, Mortby

, Barber

, Cieslak

, Fischer

, Granger

, Hogan

, Mackie

, Maxwell

, Menon

, Mueller

, Patry

, Pearson

, Quickfall

, Sajobi

, Tse

, Wang

, Smith

(2018) Prevalence of mild behavioral impairment in mild cognitive impairment and subjective cognitive decline, and its association with caregiver burden. Int Psychogeriatr 30, 233–244.

Fan

, LIang

, Yun

, Pei

, Hu

, Ismail

, Yang

, Xu

(2020) Mild behaviour impairment is related to frailty in cognitively normal older adults: A cross-sectional study. BMC Geriatr 20, 510.

10.

Rao

, Chatterjee

, Thakral

, Dwivedi

, Dey

(2020) Behavioural issues in late life may be the precursor of dementia- A cross sectional evidence from memory clinic of AIIMS, India. PLoS One 15, e023451.

11.

Creese

, Griffiths

, Brooker

, Corbett

, Aarsland

, Ballard

, Ismail

(2020) Profile of mild behavioral impairment and factor structure of the Mild Behavioral Impairment Checklist in cognitively normal older adults. Int Psychogeriatr 32, 705–717.

12.

Mallo

, Ismail

, Pereiro

, Facal

, Lojo-Seoane

, Campos-Magdaleno

, Juncos-Rabadán

(2019) Assessing mild behavioral impairment with the mild behavioral impairment checklist in people with subjective cognitive decline. Int Psychogeriatr 31, 231–239.

13.

Mallo

, Ismail

, Pereiro

, Facal

, Lojo-Seoane

, Campos-Magdaleno

, Juncos-Rabadán

(2018) Assessing mild behavioral impairment with the Mild Behavioral Impairment-Checklist in people with mild cognitive impairment. J Alzheimers Dis 66, 83–95.

14.

Matsuoka

, Ismail

, Narumoto

(2019) Prevalence of mild behavioral impairment and risk of dementia in a psychiatric outpatient clinic. J Alzheimers Dis 70, 503–511.

15.

Hall

, Wiechmann

, Johnson

, Edwards

, Barber

, Cunningham

, Singh

, O’Bryant

, Doody

, Roundtree

, Pavlik

, Chan

, Massman

, Darby

, Evans

, Williams

, Schrimsher

, Dentino

, Orozco

, Fairchild

, Knebl

, Mains

, Alvarez

, Adams

, Rosenberg

, Weiner

, Quiceno

, Reisch

, Huebinger

, Xiao

, Svetlik

, Werry

, Smith

, Royall

, Palmer

, Polk

(2014) Total cholesterol and neuropsychiatric symptoms in Alzheimer’s disease: The impact of total cholesterol level and gender. Dement Geriatr Cogn Disord 38, 300–309.

16.

Shi

, Zhou

, Mei

, Yang

, Li

(2020) The effect of type 2 diabetes mellitus on neuropsychological symptoms in Chinese early Alzheimer’s disease population. Neuropsychiatr Dis Treat 16, 829–836.

17.

Hwang

, Baek

K-H

, Han

, Ha

S-W

, Yang

(2017) Hypertension and neuropsychiatric symptoms in patients with drug-naïve Alzheimer’s disease. Dement Neurocogn Disord 16, 78.

18.

Kortebi

, Karameh

, Kumar

, Gallagher

, Golas

, Schweizer

, Munoz

, Barfett

, Butters

, Bowie

, Flint

, Rajji

, Herrmann

, Pollock

, Mulsant

, Mah

, Fischer

(2019) Examining the impact of cardiovascular risk factors on neuropsychiatric symptoms in the PACT-MD study. Am J Geriatr Psychiatry 27, S202–S203.

19.

Fischer

, Kortebi

, Karameh

, Kumar

, Gallagher

, Golas

, Munoz

, Barfett

, Butters

, Bowie

, Flint

, Rajji

, Herrmann

, Pollock

, Mulsant

, Schweizer

, Mah

(2019) Examining the link between cardiovascular risk factors and neuropsychiatric symptoms in mild cognitive impairment and major depressive disorder in remission. J Alzheimers Dis 67, 1305–1311.

20.

Petersen

, Smith

, Waring

, Ivnik

, Kokmen

, Tangelos

(1997) Aging, memory, and mild cognitive impairment. Int Psychogeriatr 9, 65–69.

21.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

, Snyder

, Carrillo

, Thies

, Phelps

(2011) The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 270–279.

22.

Feng

, Chong

, Lim

, Ng

(2012) The modified mini-mental state examination test: Normative data for Singapore Chinese older adults and its performance in detecting early cognitive impairment. Singapore Med J 53, 458–462.

23.

Nasreddine

, Phillips

, Bédirian

, Charbonneau

, Whitehead

, Collin

, Cummings

, Chertkow

(2005) The Montreal Cognitive Assessment, MoCA: A brief screening tool for mild cognitive impairment. J Am Geriatr Soc 53, 695–699.

24.

Creese

, Brooker

, Ismail

, Wesnes

, Hampshire

, Khan

, Megalogeni

, Corbett

, Aarsland

, Ballard

(2019) Mild behavioral impairment as a marker of cognitive decline in cognitively normal older adults. Am J Geriatr Psychiatry 27, 823–834.

25.

Ismail

, Elbayoumi

, Fischer

, Hogan

, Millikin

, Schweizer

, Mortby

, Smith

, Patten

, Fiest

(2017) Prevalence of depression in patients with mild cognitive impairment: A systematic review and meta-analysis. JAMA Psychiatry 74, 58–67.

26.

Chen

, Chen

(2002) Best practice. West J Med 176, 239–244.

27.

Fuh

, Lam

, Hirono

, Senanarong

, Cummings

(2006) Neuropsychiatric inventory workshop: Behavioral and psychologic symptoms of dementia in Asia. Alzheimer Dis Assoc Disord 20, 314–317.

28.

Kramer

, Kwong

, Lee

, Chung

(2002) Culture and medicine cultural factors influencing the mental. West J Med 176, 227–231.

29.

Bracke

, Pattyn

, Von

dem Knesebeck O

(2013) Overeducation and depressive symptoms: Diminishing mental health returns to education. Sociol Health Illn 35, 1242–1259.

30.

Mortby

, Ismail

, Anstey

(2018) Prevalence estimates of mild behavioral impairment in a population-based sample of pre-dementia states and cognitively healthy older adults. Int Psychogeriatr 30, 221–232.

31.

Cheng

(2017) Dementia caregiver burden: A research update and critical analysis. Curr Psychiatry Rep 19, 64.

32.

Teng

, Lu

, Cummings

(2007) Neuropsychiatric symptoms are associated with progression from mild cognitive impairment to Alzheimer’s disease. Dement Geriatr Cogn Disord 24, 253–259.

33.

Saari

, Smith

, Ismail

(2021) Network analysis of impulse dyscontrol in mild cognitive impairment and subjective cognitive decline. Int Psychogeriatr, doi: 10.1017/S1041610220004123.

34.

Kalra

, Sahay

(2018) Diabetes fatigue syndrome. Diabetes Ther 9, 1421–1429.

35.

Kruse

, Schmitz

, Thefeld

(2003) On the association between diabetes. Diabetes Care 26, 1841–1846.

36.

Kota

, Meher

, Jammula

, Krishna

SVS

, Kota

, Modi

(2012) Neuropsychiatric screening in type 2 diabetes mellitus. Indian J Endocrinol Metab 16, 37.

37.

Sommerfield

, Deary

, Frier

(2004) Acute hyperglycemia alters mood state and impairs cognitive performance in people with type 2 diabetes. Diabetes Care 27, 2335–2340.

38.

Eckstrand

, Mummareddy

, Kang

, Cowan

, Zhou

, Zald

, Silver

, Niswender

, Avison

(2017) An insulin resistance associated neural correlate of impulsivity in type 2 diabetes mellitus. PLoS One 12, e0189113.

39.

Dietlin

, Soto

, Kiyasova

, Pueyo

, De Mauleon

, Delrieu

, Ousset

, Vellas

(2019) Neuropsychiatric symptoms and risk of progression to Alzheimer’s disease among mild cognitive impairment subjects. J Alzheimers Dis 70, 25–34.

40.

Rad

, Arya

, Karimian

, Madhavan

, Rizwan

, Koshy

, Prabhu

(2018) Mechanism involved in insulin resistance via accumulation of β-amyloid and neurofibrillary tangles: Link between type 2 diabetes and Alzheimer’s disease. Drug Des Devel Ther 12, 3999–4021.

41.

Gonçalves

, Wijesekara

, Fraser

, De Felice

(2019) The link between tau and insulin signaling: Implications for Alzheimer’s disease and other tauopathies. Front Cell Neurosci 13, 17.

42.

Kim

H-G

(2019) Cognitive dysfunctions in individuals with diabetes mellitus. Yeungnam Univ J Med 36, 183–191.

43.

Kleinridders

, Cai

, Cappellucci

, Ghazarian

, Collins

, Vienberg

, Pothos

, Kahn

(2015) Insulin resistance in brain alters dopamine turnover and causes behavioral disorders. Proc Natl Acad Sci U S A 112, 3463–3468.

44.

Zhao

, Ksiezak-Reding

, Riggio

, Haroutunian

, Pasinetti

(2006) Insulin receptor deficits in schizophrenia and in cellular and animal models of insulin receptor dysfunction. Schizophr Res 84, 1–14.

45.

Cuperfain

, Kennedy

, Gonçalves

(2020) Overlapping mechanisms linking insulin resistance with cognition and neuroprogression in bipolar disorder. Neurosci Biobehav Rev 111, 125–134.