Predicting three categories of Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items score‐based glycemic targets using the number of animal names recalled

Abstract

Aims/Introduction

The Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items (DASC‐8) assesses memory, orientation, instrumental activities of daily living and basic activities of daily living. Category I (DASC‐8 score ≤10), category II (11 ≤ DASC‐8 score ≤16) and category III (DASC‐8 score ≥17) have been defined. Based on these categories, the glycemic targets in diabetes patients aged ≥65 years have been proposed by the Japan Diabetes Society and the Japan Geriatrics Society Joint Committee. DASC‐8 is difficult to apply to patients without family members or supportive persons. We propose a verbal fluency test as the screening tool.

Materials and Methods

We enrolled 69 inpatients aged ≥65 years with type 2 diabetes, who were administered the DASC‐8 and VF tests, which included recalling animal names and common nouns starting with a specified letter in 1 min. The relationship between DASC‐8 and verbal fluency test scores was investigated.

Results

Animal fluency correlated with DASC‐8 scores after adjustment for patient characteristics. Animal scores correlated with orientation, instrumental activities of daily living and basic activities of daily living scores of DASC‐8, and tended to show a relationship with DASC‐8 memory scores. An animal score ≥8 predicted category I with a sensitivity of 89% and a specificity of 57%. An animal score ≤6 predicted category III with a sensitivity of 85% and a specificity of 67%.

Conclusions

Animal scores would be useful in predicting the categories of DASC‐8. Animal fluency could be a screening tool of DASC‐8 when a patient's family member or supportive person is absent.

Short abstract

We enrolled 69 inpatients aged ≥65 years with type 2 diabetes, who were administered the Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items and verbal fluency tests, which included recalling animal names and common nouns starting with a specified letter in 1 min. Animal fluency would be useful in predicting the categories of Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items.

INTRODUCTION

The Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items (DASC‐8) is a short version of the DASC 21‐items test, and assesses memory, orientation, instrumental activities of daily living (IADL) and basic ADL (BADL) of the participants. DASC‐8 assesses participant memory through item 1, orientation through item 2, IADL through items 3–5 and BADL through items 6–8. Participants were classified into three categories according to their DASC‐8 scores: category I (DASC‐8 score ≤10), those with intact cognitive functions and no impairment in activities of daily living; category II (11 ≤ DASC‐8 score ≤ 16), those with mild cognitive impairment to mild dementia or impairment of IADL with no impairment of BADL; and category III (DASC‐8 score ≥17), those with moderate or severe dementia or impairment of BADL ¹ .

Based on these categories according to DASC‐8 scores, the glycemic targets in diabetes patients aged >65 years have been proposed by the Japan Diabetes Society and Japan Geriatrics Society Joint Committee (JGS). ² Higher glycemic control targets (glycated hemoglobin [HbA1c] values) for patients in categories II and III were accepted by Japan Diabetes Society and Japan Geriatrics Society Joint Committee than those for category I patients, to avoid hypoglycemic complications ² , ³ .

DASC‐8 assessment has often been undertaken in clinical practice, because DASC‐8 scores recommended the glycemic targets for elderly diabetes patients, in addition to determine if support has been required for shopping, mobility and other daily living functions. These scores have often been based on information from family members or other persons supporting the patient; however, it can be difficult to glean information in certain circumstances; for example, when patients live alone. To establish glycemic targets in these elderly diabetes patients, it would be useful to have a screening tool to assess their DASC‐8 scores – a tool that does not rely on the input of family members or supportive persons.

We propose verbal fluency (VF) tests as a convenient screening tool that takes 1 min. We found that VF scores reflect executive function and memory, such as is required for insulin self‐injection ⁴ . Executive function is closely related to IADL function ⁵ . Hence, we hypothesized that VF scores correlate with the DASC‐8 scores of items 3–5 (regarding IADL) and of item 1 (regarding memory). In the present study, we aimed to investigate the association between VF scores and DASC‐8 scores.

MATERIALS AND METHODS

Participants

We enrolled 69 consecutive inpatients aged ≥65 years with type 2 diabetes who answered the questionnaires in the present study between December 2019 and March 2022 at the Saiseikai Yokohamashi Nanbu Hospital, Yokohama, Japan. Participants were included if they had family members or persons who knew their DASC‐8 scores. Participants were excluded if they had diabetic ketoacidosis or were thought to be inappropriate for this study by physicians; for example, if they had concurrent psychiatric disease.

Study design and verbal fluency tests

This was a cross‐sectional observational study. The DASC‐8 was originally developed as a simple assessment tool for cognitive function in community‐dwelling people in Japan. The DASC‐8 consists of eight items categorized into four domains: orientation (item 1), memory (item 2), IADL (items 3–5) and BADL (items 6–8) ¹ . On admission, the first author administered DASC‐8 to all members of a patient's family or persons who had knowledge of their DASC‐8 items. The following VF tests were undertaken by all participants during their hospital stay and administered by the first author: (i) animal fluency, recalling animal names; and (ii) letter fluency, recalling common nouns starting with ‘a,’ ‘ka’ and ‘shi’ (Japanese letters), which were used in our previous study ⁴ . For each of the animal fluency and letter fluency tests, patients were instructed to recall as many of the terms as possible within 1 min. The total number of responses was noted, excluding any errors and repetitions. The sum of the number of nouns starting with ‘a,’ ‘ka,’ and ‘shi’ recalled was considered to represent the total letter fluency ⁶ . The total of the animal fluency responses was termed the ‘animal score,’ whereas the total of the letter fluency responses was named the ‘letter score.’

Ethical considerations

This study was registered with the University Hospital Medical Information Network (UMIN) (UMIN ID: 000038745). The study was approved by the ethics committee of the Saiseikai Yokohamashi Nanbu Hospital on 19 December 2019. The approval number is 2019 D‐35. This study was carried out in accordance with the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants.

Assessments of blood profiles and characteristics

To evaluate glucose metabolism and serum creatinine levels, overnight fasting blood samples were obtained from all patients on the second day after admission. HbA1c levels were measured using high‐performance liquid chromatography (Adams A1c HA‐8160; Arkray Inc., Kyoto, Japan). The fasting plasma C‐peptide levels were measured at a central clinical laboratory (SRL, Inc., Tokyo, Japan). Estimated glomerular filtration rate (eGFR) was calculated as eGFR (mL/min/1.73 m ² ) = 194 × serum creatinine^−1.094 × age^−0.287 (×0.739 for females). ⁷

We investigated all participant characteristics, such as age, sex, body mass index and duration of diabetes (years).

Statistical analysis

All of the statistical analyses were carried out using IBM SPSS 26 Software for Windows (IBM, Armonk, NY, USA). The mean ± standard deviation was reported for continuous variables, and the numbers for categorical variables. Three groups were identified by DASC‐8 scores: category I (DASC‐8 score ≤10), category II (11 ≤ DASC‐8 score ≤16) and category III (DASC‐8 score ≥17). Age, body mass index, HbA1c, eGFR, DASC‐8 total scores and VF test scores were compared among the three groups using the analysis of variance. VF scores were reported to follow a normal distribution ⁴ , ⁸ . Fasting plasma C‐peptide levels, duration of diabetes (years), memory, orientation, and the IADL and BADL scores of DASC‐8 were compared among the three groups using the Kruskal–Wallis test. The categorical variables among the three groups – such as sex and individual DASC‐8 item scores – were analyzed using the χ ² ‐test or Fisher's exact test, according to the data distribution. The relationships between animal score and each of the following: memory/orientation/BADL/IADL scores of DASC‐8 were investigated using univariate linear regression analysis. The relationships between letter score and each of the following: memory/orientation/BADL/IADL scores of DASC‐8 were also investigated using univariate linear regression analysis.

The factors predictive for the total scores of DASC‐8, such as age, HbA1c level, eGFR and VF tests scores, were assessed using univariate and multivariate linear regression analyses.

A receiver operating characteristic curve was used to determine the cut‐off animal scores predicting category I versus II + III and category I + II versus III.

RESULTS

Characteristics of all participants

The characteristics of the 69 study participants (47 men, 22 women) are shown in Table 1.

Table 1.

Patient characteristics and cognitive test scores

	Category I (n = 46)	Category II (n = 14)	Category III (n = 9)	P
Male/female	31/15	10/4	6/3	0.95
Age (years)	76.2 ± 5.8	78.9 ± 6.0	82.3 ± 7.5	<0.05
BMI (kg/m 2 )	22.8 ± 4.4	22.5 ± 4.4	23.0 ± 4.4	0.96
HbA1c in the hospital (%)	10.1 ± 2.0	10.1 ± 1.8	10.0 ± 1.7	0.99
Fasting serum CPR (ng/ml)	1.24 (0.79–1.84)	1.11 (0.76–3.28)	1.11 (0.76–3.28)	0.73
Duration of diabetes (years)	8 (1.75–18.5)	15 (11–30)	15 (11–30)	<0.05
eGFR (mL/min/1.73m 2 )	54 ± 19.1	49.2 ± 16	40.4 ± 14.6	0.116
Total scores of DASC‐8	8.4 ± 0.66	13.4 ± 2.02	20.1 ± 2.37	<0.001
DASC‐8 items
Memory
Number of Item 1 scores (1/2/3/4)	36/10/0/0	4/5/5/0	3/2/3/1	<0.001
Orientation
Number of Item 2 scores (1/2/3/4)	41/5/0/0	1/8/2/3	0/4/4/1	<0.001
IADL
Number of Item 3 scores (1/2/3/4)	44/2/0/0	5/7/2/0	0/1/3/5	<0.001
Number of Item 4 scores (1/2/3/4)	46/0/0/0	8/4/2/0	0/0/2/7	<0.001
Number of Item 5 scores (1/2/3/4)	44/2/0/0	4/3/6/1	0/0/3/6	<0.001
BADL
Number of Item 6 scores (1/2/3/4)	45/1/0/0	13/1/0/0	5/2/2/0	<0.01
Number of Item 7 scores (1/2/3/4)	46/0/0/0	14/0/0/0	8/0/1/0	0.13
Number of Item 8 scores (1/2/3/4)	46/0/0/0	13/1/0/0	5/4/0/0	<0.001
Verbal fluency tests
Animal score	13.0 ± 5.1	8.4 ± 3.8	4.8 ± 3.1	<0.05
Letter score	19.2 ± 10.2	11.9 ± 8.6	6.5 ± 4.8	<0.05

Data are expressed as numbers, median (first quartile,third quartile)or means ± standard deviation. The animal score is the number of animal names recalled in 1 min. The letter score is the total number of the common nouns starting with ‘a’, ‘ka’ and ‘shi’ (Japanese letters) each recalled in 1 min. BADL, basic activities of daily living; BMI, body mass index; HbA1c, glycated hemoglobin; CPR, C‐peptide immunoreactivity; DASC‐8, Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items; IADL, instrumental activities of daily living.

A total of 46 participants belonged to category I (DASC‐8 score ≤10), 14 to category II (11 ≤ DASC‐8 score ≤ 16) and nine to category III (DASC‐8 score ≥17.).

Each group showed poor glycemic control, with HbA1c levels of >10%. Participants in category II and III were significantly older, and the duration of diabetes in years was longer than those in category I. Participants in category I showed lower scores for each item of DASC‐8, except for item 7, presumably because of a ceiling effect. Animal and letter scores were higher among category I participants than category II and III participants. Three patients had functional impairments as result of their physical injuries (spinal stenosis n = 2, cervical spondylosis n = 1), and their DASC‐8 categories were I (n = 2) and II (n = 1).

Correlates of the total DASC‐8 score

Correlates of the total DASC‐8 score were investigated (Table 2). Univariate linear regression analysis showed that both VF scores, age and eGFR were correlates of total DASC‐8 scores; however, only animal score was a correlate when multiple linear regression analysis was used.

Table 2.

Correlates for the total score of Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items

	Univariate B (95% CI)	P	Multivariate B (95% CI)	P
Age (years)	0.252 (0.098 to 0.405)	<0.005	0.104 (‐0.05 to 0.258)	0.18
HbA1c in the hospital (%)	0.056 (−0.486 to 0.598)	0.84	0.049 (−0.396 to 0.494)	0.83
eGFR (mL/min/1.73m 2 )	−0.062 (−0.117 to −0.007)	<0.05	−0.023 (−0.073 to −0.027)	0.37
Verbal fluency tests
Animal score	−0.437 (−0.593 to −0.28)	<0.001	−0.314 (−0.51 to −0.117)	<0.01
Letter score	−0.183 (−0.273 to −0.093)	<0.001	−0.06 (−0.164 to −0.045)	0.26

The animal score is the number of animal names recalled in 1 min. The letter score is the total number of the common nouns starting with ‘a’, ‘ka’ and ‘shi’ (Japanese letters) each recalled in 1 min. CI, confidence interval; DASC‐8, Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items; eGFR, estimated glomerular filtration rate; HbA1c, glycated hemoglobin.

DASC‐8 scores per category

DASC‐8 consists of memory (item 1), orientation (item 2), IADL (items 3–5) and BADL (items 6–8). As shown in Figure 1, participants in category I showed lower memory scores than those in categories II and III. However, there were no significant differences between category II and category III in their memory scores (P = 0.98). Participants in category I showed lower orientation scores than those in categories II and III; however, there were no significant differences between category II and category III in their orientation scores (P = 0.60).

Figure 1.

Memory/orientation/instrumental activities of daily living (IADL)/basic activities of daily living (BADL) scores of Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items among the three categories.

Participants in category I showed lower IADL scores than those in category II, as did those in category II when compared with category III. Participants in category III showed higher BADL scores than those in categories I and II, without significant differences between categories I and II (P = 0.56).

Association between DASC‐8 items and animal and letter fluency scores

Whether DASC‐8 items correlated with both VF scores was investigated (Table 3). Animal scores correlated with orientation, IADL score and BADL score significantly, and tended to correlate with memory (P = 0.085); however, letter correlated with orientation and IADL scores.

Table 3.

Items of Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items correlating with animal and letter scores through univariate linear regression analysis

	Animal score B (95% CI)	P	Letter score B (95% CI)	P
DASC‐8 items memory scores	−0.029 (−0.062 to 0.004)	0.085	−0.011 (−0.029 to 0.006)	0.204
Orientation scores	−0.082 (−0.115 to −0.049)	<0.001	−0.035 (−0.054 to −0.017)	<0.001
IADL scores	−0.275 (−0.381 to −0.169)	<0.001	−0.121 (−0.181 to −0.061)	<0.001
BADL scores	−0.051 (−0.084 to −0.018)	<0.01	−0.016 (−0.034 to −0.003)	0.098

The animal score is the number of animal names recalled in 1 min. The letter score is the total number of the common nouns starting with ‘a’, ‘ka’ and ‘shi’ (Japanese letters) each recalled in 1 min. BADL, basic activities of daily living; CI, confidence interval; DASC‐8, Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items; IADL, instrumental activities of daily living.

Animal cut‐off score for predicting DASC‐8 category I or III

The animal cut‐off value and the area under the curve in the receiver operating characteristic curve were determined (Table 4). An animal score ≥8 predicted category I (DASC‐8 score ≤10) with a sensitivity of 89% and a specificity of 57% (area under the curve 0.83, 95% confidence interval 0.73–0.92, P < 0.001). An animal score ≤6 predicted category III (DASC‐8 score ≥17) with a sensitivity of 85% and a specificity of 67% (area under the curve 0.89, 95% confidence interval 0.80–0.97, P < 0.001).

Table 4.

Cut‐off value and the area under the curve in the receiver operating characteristics curve with the animal score for predicting Dementia Assessment Sheet for Community‐based Integrated Care System 8‐items cut‐off score

	Cut‐off value	AUC (95% CI)	P
Animal score	8 (Category I)	0.83 (0.73–0.92)	<0.001
	6 (Category III)	0.89 (0.80–0.97)	<0.001

The animal score is the number of animal names recalled in 1 min. AUC, area under the curve.

DISCUSSION

In the present study, we showed that the number of animal names recalled in 1 min was negatively correlated with the total score of DASC‐8 using multiple linear regression analysis. Patients belonging to category I showed lower animal scores than those in category II, as did patients in category II when compared with those in category III. Our study proposed that animal scores ≥8 predicted category I (DASC‐8 score 10) and animal scores ≤6 predicted category III (DASC‐8 score ≥17).

DASC‐8 assesses memory (item 1; 1–4 points), orientation (item 2; 1–4 points), IADL (items 3–5; 3–12 points) and BADL (items 6–8; 3–12 points) ² . IADL was reported to reflect executive function ⁵ , ⁹ . We had hypothesized that animal and letter scores correlated with memory and IADL (items of DASC‐8), respectively, because we found that both scores could reflect executive function and memory, such as was required for insulin self‐injection ⁴ . As described in the results, animal scores correlated with orientation, IADL score and BADL score better than letter fluency did. The exception was memory. Therefore, animal fluency could be an alternative to DASC‐8.

Some underlying mechanisms were predicted, which explained that animal scores were closely related to DASC‐8 scores. Semantic fluency tasks, such as recalling animal names, require memory and executive function, which are sets of mental skills related to attention, inhibitory control and problem solving ¹⁰ . The temporoparietal lobe and prefrontal lobe are activated during a semantic fluency task ¹¹ . Memory involves the temporoparietal lobe, which contains the hippocampus, and the prefrontal lobe is thought to be activated while searching names in a specific group ¹² , ¹³ . Letter fluency requires executive function more than memory, because patients had difficulty searching names beginning with a specific letter. Letter fluency is known to reflect executive function of the prefrontal lobe, not the tempoparietal lobe, according to functional magnetic resonance imaging and studies of prefrontal lobe lesions ¹⁰ , ¹⁴ . Thus, animal scores might correlate with DASC‐8 scores better than letter scores, because DASC‐8 required memory (item 1) and IADL (item 3–5; which was closely related to executive function ⁵ , ⁹ ).

Animal scores did not correlate with item 1 (memory) of DASC‐8, contrary to our hypothesis, although the scores had a tendency to correlate with item 1 (P = 0.085). Small sample size might affect these relationships statistically.

As shown in Figure 1, category I versus II + III resulted from memory, orientation and IADL impairment. Animal scores were strongly associated with orientation and IADL scores (P < 0.001, respectively), thereby, animal fluency discriminated between category I and categories II + III. Orientation reflected function of the medial parietal cortex areas of the brain, which were activated during an animal fluency task ⁵ , ¹⁵ ; therefore, there is thought to be a relationship between orientation and animal fluency.

As shown in Figure 1, category I + II versus III resulted from IADL and BADL impairment. Animal scores could also discriminate category I + II versus III through BADL scores in addition to IADL scores (which had strong associations with animal fluency, as described before). BADL is independent of cognitive ability, such as executive function. It was reported that most people with BADL impairment had IADL disorders, and that those with IADL, as well as BADL dysfunction, had lower cognitive tests scores than those with IADL dysfunction only ¹⁶ . Patients with IADL impairment were thought to develop BADL dysfunction and cognitive deficits gradually. Thus, a cognitive test, such as animal fluency, might correlate with BADL scores through cognitive dysfunction accompanied by BADL dysfunction. Animal fluency was significantly associated with BADL scores (P < 0.01) through these underlying mechanisms and, therefore, could be thought to discriminate category I + II from category III.

The three patients with BADL dysfunction, but intact IADL function, who had body functional impairments (spinal stenosis n = 2; cervical spondylosis n = 1) should be taken into consideration, because they showed high animal scores despite their BADL dysfunction. However, three such patients out of 69 were not thought to affect the present results due to their small number.

We also proposed cut‐off values: animal score ≥8 predicting category I (DASC‐8 score ≤10) and animal score ≤6 predicting category III (DASC‐8 score ≥17). These cut‐off values can be considered appropriate, although the present study was carried out in only our hospital and the sample size was small. We reported that a score of 11 animal names was able to predict the non‐demented patients' ability to appropriately carry out insulin self‐injection ⁴ . In the present study, animal scores of 8 predicted category I versus II + III (distinguishing non‐demented from marginal cognitive impairment) and animal scores of 6 predicted category I + II versus III (distinguishing marginal cognitive impairment from demented).

Hanyu et al. ¹⁷ showed the score obtained by differentiating 13 animals by the demented patients from control patients among non‐diabetics. It was concluded that type 2 diabetes leads to mild‐to‐moderate deficiency in all the measured cognitive abilities ¹⁸ , and the cut‐off value for animal names that discriminated demented patients from control patients was thought to be lower among patients with diabetes. Evidently, an animal score of 8 predicting category I versus II + III (distinguishing non‐demented from marginal cognitive impairment) might be appropriate.

The present study had several limitations. First, this study was carried out in one hospital in Yokohama City, Japan, and our sample size was small. For this reason, selection bias might have affected our results. However, the animal score cut‐off values proposed in this study to discriminate category I (non‐dementia) from category III (dementia) were acceptable, based on our previous study ⁴ and the Hanyu et al. report ¹⁶ described above. More studies are required to confirm and validate the present findings.

Second, hyperglycemia might have affected the present results, because VF tests were undertaken during hospitalization for hyperglycemia. It is unclear whether poor glycemic controls affect cognitive function. For example, Naor et al. ¹⁹ reported that the cognitive test scores regarding executive function improved when the HbA1c level was reduced from 12 to 9.5% in 6 weeks using Trail Making Test‐A. Furthermore, Meneilly et al. ²⁰ reported that the scores regarding executive function and learning abilities improved when the HbA1c level was reduced from 8.6 to 7.5% in 6 months using Trail Making Test‐A, Stroop Color and Word Test, and Modified Cued Recall Test. Gradman et al. ²¹ reported that the scores regarding memory improved when the HbA1c level is reduced from 11 to 10% in 4 months using the Buschke Selective Reminding Test. VF tests required memory and executive function ⁴ , thereby poor glycemic controls might deteriorate VF tests scores. However, Hewer et al. ²² reported that VF test scores did not change even if glycemic control was improved within 2 weeks.

Animal fluency could be tested in 1 min and might be useful in predicting the DASC‐8 category of patients. Animal fluency could be a feasible screening tool of DASC‐8 when a patients' family members or supportive persons are absent and cannot assist in establishing DASC‐8 category, which determines the target HbA1c levels.

DISCLOSURE

The authors declare no conflict of interest.

Approval of the research protocol: The study was approved by the ethics committee of Saiseikai Yokohama‐Shi Nanbu Hospital and Odawara Municipal Hospital on 19 December 2019. The approval number is 2019 D‐35.

Informed consent: Written informed consent was obtained from all participants in accordance with the tenets of the Declaration of Helsinki.

Registry and registration no. of the study/trial: Date: 19 December 2019. No. UMIN ID: 000038745.

Animal studies: N/A.