Global cognition and executive functions of older adults with type 1 diabetes mellitus without dementia

Abstract

Aims/Introduction

This study aimed to characterize the global cognition and executive functions of older adults with type 1 diabetes mellitus in comparison with type 2 diabetes mellitus.

Materials and Methods

This study included 37 patients with type 1 diabetes mellitus aged ≥65 years and 37 age‐ and sex‐matched patients with type 2 diabetes mellitus. Patients with dementia scoring <24 on the Mini‐Mental State Examination were excluded. General cognition, memory, classic, and practical executive function were investigated.

Results

Patients with type 1 diabetes mellitus demonstrated lower psychomotor speed scores on Trail Making Tests A and B (P < 0.001, P < 0.013) than those with type 2 diabetes mellitus. The dysexecutive syndrome behavioral assessment revealed similar results in patients with types 1 and 2 diabetes mellitus. The Wechsler Memory Scale‐Revised verbal episodic memory and Montreal Cognitive Assessment Japanese version were similar in terms of general cognition, but worse delayed recall subset on the latter was associated with type 2 diabetes mellitus (P = 0.038). A worse Trail Making Test‐A performance was associated with type 1 diabetes mellitus and age (P < 0.004, P < 0.029).

Conclusions

Executive function of psychomotor speed was worse in older outpatient adults without dementia with type 1 diabetes mellitus than in those with type 2 diabetes mellitus but with no significant differences in the comprehensive and practical behavioral assessment of dysexecutive syndrome. Patients with type 1 diabetes had more severely impaired executive function, whereas those with type 2 had greater impaired memory than executive function.

This study aimed to characterize the global cognitive and executive functions of older adults with type 1 diabetes mellitus in comparison with type 2 diabetes mellitus. General cognition, memory, classic, and practical executive function tests were performed in 37 patients with type 1 and type 2 diabetes over 65 years. In type 1 diabetes, executive function of psychomotor speed was worse compared with type 2 diabetes mellitus and it revealed that patients with type 1 diabetes had more strongly impaired executive function than memory.

INTRODUCTION

The life expectancy of patients with type 1 diabetes mellitus has recently increased due to advancements in diabetes therapy, including glucose measurement devices, insulin preparations, and insulin delivery ¹ , ² . The Pittsburgh Epidemiological Study of Diabetes Complications reported a lower life expectancy at birth in participants diagnosed from 1950 to 1964 than in those diagnosed from 1965 to 1980 ³ . Early childhood deaths have decreased, as evidenced by more individuals with type 1 diabetes mellitus reaching older adulthood. Additionally, mortality from cardiovascular diseases, which significantly affected the life outcome of adults with diabetes, was decreased with the introduction of statins and technological changes such as percutaneous coronary intervention. The survival of patients with type 1 diabetes mellitus may increase in the future. Thus, age‐related problems, such as dementia and cognitive impairment, will become clear.

Diabetes mellitus increases the relative risk for Alzheimer's disease by 1.46 and for vascular dementia by 2.48 ⁴ . Several studies have focused on cognitive impairment in type 2 diabetes mellitus with insulin resistance and obesity ⁵ , ⁶ . However, the cognitive function of older patients with type 1 diabetes mellitus remains elusive. Self‐regulation and self‐injection of insulin are essential in type 1 diabetes mellitus for good glucose control as well as for their daily lives. Executive function is crucial, which controls human cognition and behavior and encompasses working memory, inhibition, and components of attention, and mental flexibility ⁷ . Executive functions are closely related to treatment adherence ⁸ , ⁹ . Therefore, older patients with type 1 diabetes mellitus must maintain memory and executive function.

Children and young adults with type 1 diabetes mellitus have impaired cognitive and executive functions ¹⁰ , ¹¹ , ¹² . An early meta‐analysis reported that young adults with type 1 diabetes mellitus had worse attention, cognitive flexibility, psychomotor efficiency, information processing speed, intelligence, and visual perception than nondiabetic controls ¹⁰ . Conversely, a study of patients with type 1 diabetes mellitus with an average age of 60.9 years indicated their significantly lower information processing speed than nondiabetic controls, but the same was not observed for memory ¹¹ . Although the prevalence and incidence of dementia increase exponentially at age ≥65 years ¹³ , ¹⁴ , few studies have focused on cognitive impairment in patients with type 1 diabetes mellitus >65 years. Participants in the Medalist Study, a study of 82 patients with type 1 diabetes mellitus, who have had the disease for ≥50 years, were ≥50 years old ¹⁵ . The Study of Longevity in Diabetes (SOLID) prospective multicenter cohort study on aging and diabetes looked at the cognitive abilities of 734 people with type 1 diabetes mellitus, 232 people with type 2 diabetes mellitus, and 247 people who did not have diabetes. The participants were all aged 60 or older ¹⁶ . The general cognition and executive functions in patients aged ≥65 years with type 1 diabetes mellitus and their differences from those with type 2 diabetes mellitus remain unclear ¹⁷ .

Therefore, this study aimed to determine the differences in cognitive and executive functions between older adults with type 1 diabetes mellitus and age‐ and sex‐matched older adults with type 2 diabetes mellitus.

MATERIALS AND METHODS

Study population

Participants with type 1 diabetes mellitus aged ≥65 years who visited the Toranomon Hospital were recruited continuously from April 2018 to March 2020. Type 1 diabetes mellitus was defined as an absolute insulin deficiency due to autoimmune beta cell destruction ¹⁸ . The autoimmune indicators used to diagnose type 1 diabetes mellitus included islet cell autoantibodies, autoantibodies to GAD (GAD65), insulin, tyrosine phosphatases IA‐2, and occasionally, serum C‐peptide levels of <0.5 ng/mL. We recruited all 44 outpatients with type 1 diabetes mellitus admitted to our hospital from April 2018 to March 2020. Exclusion criteria were a neuropsychiatric disorder that could influence cognitive function and a history of alcohol and substance use. Participants with a history of asymptomatic stroke, severe renal failure with hemodialysis, and liver damage without hepatic encephalopathy were included. All participants were functionally independent at home, administered insulin injections, and monitored their blood glucose. Type 2 diabetes mellitus was defined as having hemoglobin A1C (HbA1c) levels of ≥48 mmol/mol (≥6.5%) or taking antidiabetic medication. We consecutively recruited individuals with type 2 diabetes mellitus and type 1 diabetes mellitus of the same age and sex. We excluded patients with dementia using Mini‐Mental State Examination [MMSE] scores <24.

Information on smoking behavior, alcohol consumption, educational level, and severe hypoglycemic events (requiring external help) was obtained through interviews and medical records. Hypertension was defined as a blood pressure of >130/80 mmHg or blood pressure‐lowering drug intake. The presence of retinopathy was diagnosed by ophthalmologists, whether diabetic simple retinopathy or more severe. Nephropathy was diagnosed by an estimated glomerular filtration rate (eGFR) of 45 mL/min/1.73 m^{2 15} . The Toranomon Hospital Ethics Committee approved this study. All participants signed written consent after being fully informed.

Neurological assessment

Global cognition and memory

The MMSE and the Montreal Cognitive Assessment‐J (MoCA‐J) were used to test general cognition ¹⁹ , ²⁰ .

The maximum score for each subitem is 30 points. The revised Wechsler Memory Scale (WMS‐R) was used to evaluate verbal memory ²¹ . We used revised scores for oral immediate recall (Logical Memory I) and delayed recall (Logical Memory II). The full score for each is 50 points. The Rey Osterrieth complex figure test assessed visual memory ²² . The copy and delayed tasks were evaluated after 3 and 30 min, respectively. There were 18 units with a maximum score of 2 points per item, yielding a total score of 36 points.

Executive function

The Trail Making Test (TMT)‐A,B ²³ , ²⁴ , and Victoria Stroop Test ²⁵ were used to evaluate classical executive function. The score reflected the time taken in seconds to complete the task, with higher scores representing worse psychomotor function. Behavioral Assessment of the Dysexecutive Syndrome (BADS) was used to evaluate the executive function required in daily life ²⁶ , ²⁷ (action program test and Key search test). Researchers performed the action planning test to assess the participants’ ability to plan a sequence of actions. During the test, subjects were required to pull a cork out of a tube while following specific rules and regulations. In the key search test, they were asked to assume that a square on the paper is a field and a key is lost in the field. The researchers asked the subject to draw a path in order to find the key in the field. The search pattern determines the scoring items, and the test requires the ability to search effectively and a plan of action. Each test had a score of 4 points. The third edition of the Wechsler Adult Intelligence Scale's digit span for forward and backward subjects was used to evaluate working memory ²⁸ , ²⁹ . The Digit Symbol test evaluated the information processing speed in the subjects. Raw scores were changed to evaluation scores (1–19), which were standardized to obtain a normal distribution with a mean and standard deviation of 10 and 3 in the general population, respectively.

Statistical analysis

The independent sample, Fisher's exact t‐test, and Mann–Whitney U tests directly compared characteristics and neurological data between type 1 diabetes mellitus and type 2 diabetes mellitus groups. The correlation between the TMT‐A score and type 1 diabetes mellitus was determined using multiple regression analysis, adjusting for diabetic background (age, sex, duration of diabetes, and HbA1c) using a forced entry method. Differences in diabetic complications between patients with and without TMT‐A deteriorations were examined using Fisher's exact t‐test. TMT‐A deteriorations were determined from age‐ and education‐matched normative data (age 60–69: 39.6 s; 70–79: 45.6 s; 80–89: 56.4 s) ³⁰ . We considered patients to have attenuated TMT‐A if it was prolonged above this value. P‐values were two‐sided, and statistical significance was set at <0.05. Statistical Package for the Social Sciences for Windows, version 25.0 (IBM Corp., Armonk, New York, USA) was used for statistical analyses.

RESULTS

This study enrolled 44 outpatients with type 1 diabetes mellitus aged ≥65 years. Two patients diagnosed with dementia and four patients who refused to participate, one who scored <24 on MMSE were excluded from the study; thus, the study included 37 patients with type 1 diabetes mellitus. Overall, 51 patients with type 2 diabetes mellitus were recruited, 39 participated, and 2 who scored <24 on the MMSE were excluded from the study; finally, 37 patients were included.

Table 1 summarizes the demographic and biological characteristics of the two groups. The mean ± standard deviation (SD) of age was 72.2 ± 4.7 and 71.7 ± 4.5 years, and the years of education was 14.0 ± 2.4 and 14.6 ± 4.5 for type 1 diabetes mellitus and type 2 diabetes mellitus, respectively. Both groups were moderately well‐educated and without symptoms of depression. The type 1 diabetes mellitus group had a lower body mass index (BMI) than the type 2 diabetes mellitus group. Diastolic blood pressure was significantly lower in type 1 diabetes mellitus than in type 2 diabetes mellitus; however, neither group's systolic blood pressure was significantly lowered. Laboratory data indicated a significantly high total cholesterol, high‐density lipoprotein cholesterol, and low‐density lipoprotein cholesterol levels in type 1 diabetes mellitus.

Table 1.

Characteristics of type 1 and type 2 diabetes

	Type 1 diabetes mellitus (37)	Type 2 diabetes mellitus (37)	P Value
Age (years)	72.2 ± 4.7	71.7 ± 4.5	0.65
Sex (F/M)	16/21	18/19	0.41
Education level (years)	14.0 ± 2.4	14.6 ± 2.5	0.32
BMI (kg/m2)	21.5 ± 2.5	24.4 ± 4.3	0.032*
GDS	2.2 ± 2.3	2.2 ± 2.6	0.86
Systolic blood pressure (mmHg)	127.9 ± 12.0	131.2 ± 8.1	0.18
Diastolic blood pressure (mmHg)	66.1 ± 8.4	72.1 ± 8.2	0.004**
Hypertension (n (%))	22 (59.5)	26 (70.0)	0.47
ARB・ACE (n (%))	19 (51.0)	20 (54.0)	0.64
Statin (n (%))	16 (43.2)	24 (64.8)	0.10
Smoking (current/past) (n (%))	2 (5.4)/8 (21.6)	5 (13.5)/13 (35.1)	0.26/020
History of malignancy tumor (n (%))	7 (18.9)	7 (18.9)	1.00
BUN (mg/dL)	21 ± 11.3	17.4 ± 5.7	0.17
Cr (mg/dL)	1.2 ± 1.4	0.9 ± 0.3	0.59
eGFR (mL/min/1.73 m2)	59.9 ± 23.2	64.8 ± 17.5	0.32
Total cholesterol (mg/dL)	214.3 ± 31.4	182.1 ± 28.1	<0.001***
HDL cholesterol (mg/dL)	74.4 ± 16.6	55.0 ± 17.5	<0.001***
LDL cholesterol (mg/dL)	113.7 ± 26.5	96.0 ± 23.4	0.002**
Triglycerides (mg/dL)	112.4 ± 62.6	150.8 ± 110.3	0.16
Albumin (g/dL)	4.2 ± 0.2	4.1 ± 0.3	0.73
HbA1c (%)	7.6 ± 09	7.9 ± 1.5	0.60
Duration of diabetes (years)	23.7 ± 12.6	20.6 ± 10.2	0.25
Onset of diabetes (years)	48.4 ± 14.0	51.1 ± 8.9	0.34
Insulin dose (unit/kg)	0.49 ± 0.16	0.40 ± 11.4	0.11†
Retinopathy (n (%))	17 (45.9)	13 (35.1)	0.34
Nephropathy (eGFR 45 mL/min/1.73 m2)	6 (16.2)	4 (10.8)	0.73
History of IHD (n (%))	6 (16.2)	5 (13.5)	1.00
History of CI (n (%))	1 (2.7)	1 (2.7)	1.00
History of severe hypoglycemia (n (%))	6 (16.2)	2 (5.4)	0.26
C‐peptide depletion (<0.01 ng/mL) (n (%))	21 (56.8)	0(0)	<0.001***
Occasional serum C‐peptide (ng/mL)	0.3 ‡	2.4	<0.001***

The results are presented as mean ± SD or n (%). Represents the significant differences between the groups (*P < 0.05, **P < 0.01, ***P < 0.001).

^† n = 7, others treated with oral hypoglycemic agent only.

^‡ Patients with depleted c‐peptide were considered 0 ng/mL.

ACE, angiotensin‐converting enzyme; ARB, angiotensin II receptor blocker; CI, cerebral infarction; GDS, Geriatric Depression Scale; IHD, ischemic heart disease.

Diabetes‐related data revealed that the onset age of type 1 diabetes mellitus and type 2 diabetes mellitus was 48.4 ± 14.0 and 51.1 ± 8.9 years, respectively; the duration of diabetes was 23.7 ± 12.6 and 20.6 ± 10.2 years, respectively. Most patients were diagnosed after adulthood. Onset modes of type 1 diabetes mellitus were acute in 35.1% (n = 13), late and slowly progressive (latent autoimmune diabetes in adults) in 59.5% (n = 22), and fulminant in 5.4% (n = 2). The prevalence of diabetic retinopathy, nephropathy, coronary heart disease, and cerebral infarction was 45.9%, 16.2%, 16.2%, and 2.7%, respectively, in type 1 diabetes mellitus. The two groups demonstrated no notable variations. Additionally, the occurrence of severe hypoglycemia demonstrated no discernible difference. All patients with type 1 diabetes mellitus and nine patients (24%) with type 2 diabetes mellitus underwent insulin therapy. The rate of depleted insulin secretion (serum C‐peptide <0.01 ng/mL) was 56.8% of type 1 diabetes mellitus, with no difference in the mode of onset. None of the patients with type 2 diabetes had depleted insulin secretion. The average occasional serum C‐peptide level was 0.3 ng/mL in type 1 diabetes mellitus and 2.4 ng/mL in type 2 diabetes mellitus (P < 0.001).

The average score (±SD) of the MMSE for type 1 diabetes mellitus and type 2 diabetes mellitus was 28.1 ± 1.9 and 27.6 ± 2.0, respectively. The total score or subdomains between the two groups were similar. The MoCA analysis revealed no discernible difference between the total scores (26.1 ± 2.9 and 24.9 ± 3.0, respectively), and 37.8% and 45.9% patients with type 1 diabetes mellitus and type 2 diabetes mellitus had a score of ≤25 points, respectively. Patients with type 2 diabetes mellitus demonstrated significantly worse performance on delayed recall than those with type 1 diabetes mellitus (P = 0.038). WMS‐R Logical Memory was similar between the two groups in total immediate recall (P = 0.76), in delayed recall (P = 0.78). The ROCFT Copy score was lower in type 1 diabetes mellitus (P = 0.09), but the difference in ROCFT Recall at 3 and 30 min (P = 0.18 and 0.14, respectively) was not significant (Table 2a).

Table 2.

(a) Raw cognitive test score of general cognition and memory. (b) Executive function and working memory

	Raw score		P value
	Type 1 diabetes mellitus	Type 2 diabetes mellitus
(a)
General cognition
MMSE
Total score	28.1 ± 1.9	27.6 ± 2.0	0.17
Temporal orientation	4.9 ± 0.3	4.9 ± 0.3	0.72
Spatial orientation	4.9 ± 0.3	4.7 ± 0.5	0.22
Retention	3.0 ± 0.0	3.0 ± 0.0	1.00
Calculation/attention	3.7 ± 1.7	3.4 ± 1.8	0.49
Memory recall	2.8 ± 0.5	2.8 ± 0.5	0.79
Language	7.8 ± 4.2	7.7 ± 4.1	0.41
Visuospatial	1.0 ± 0.2	0.9 ± 0.2	0.56
MoCA‐J
Total score	26.1 ± 2.9	24.9 ± 3.0	0.10
Visuospatial	3.6 ± 0.8	3.8 ± 0.4	0.94
Executive	0.8 ± 0.4	0.8 ± 0.4	0.76
Naming	2.9 ± 0.3	2.9 ± 0.2	0.40
Short‐term memory	4.6 ± 0.6	4.6 ± 0.6	0.85
Attention	5.5 ± 0.8	5.2 ± 0.9	0.20
Language	1.6 ± 0.6	1.7 ± 0.6	0.79
Abstraction	1.6 ± 0.6	1.7 ± 0.6	0.79
Long‐term memory	3.6 ± 1.3	2.8 ± 1.6	0.038*
Orientation	5.8 ± 0.5	5.7 ± 0.5	0.40
Memory
WMS‐R
Logical memory I	22.7 ± 6.9	23.1 ± 6.9	0.76
Logical memory II	17.2 ± 6.6	17.5 ± 7.6	0.78
ROCF
Copy	35.0 ± 1.3	35.5 ± 0.9	0.09
Immediate recall (3 min)	17.2 ± 5.9	19.0 ± 5.9	0.18
Recognition (30 min)	16.6 ± 6.1	18.7 ± 5.8	0.14
(b)
Executive function
TMT‐A (s)	50.0 ± 18.7	38.0 ± 11.2	<0.001***
TMT‐B (s)	144.0 ± 58.3	125.3 ± 60.2	0.013**
Victoria Stroop test
Trial 1
Completion time	16.1 ± 3.8	15.0 ± 3.7	0.20
Miss	0.2 ± 0.7	0.03 ± 0.2	0.09
Trial 2
Completion time	21.5 ± 9.3	19.5 ± 5.0	0.64
Miss	0.2 ± 0.7	0.1 ± 0.3	0.76
Trial 3
Completion time	27.2 ± 8.8	27.5 ± 8.0	0.77
Miss	1.9 ± 3.7	1.3 ± 1.9	0.55
BADS
Key search test	2.2 ± 0.8	2.4 ± 0.9	0.45
Action program test	3.7 ± 0.6	3.8 ± 0.9	0.25
Working memory
WAIS‐III
Symbol	11.9 ± 1.9	14.0 ± 2.3	0.11
Digit span	13.8 ± 3.3	12.8 ± 2.7	0.13
Forward	7.0 ± 1.2	6.4 ± 1.4	0.06
Backward	4.6 ± 1.3	4.5 ± 1.1	0.63

The results are presented as mean ± SD or n (%). Represents the significant differences between the groups (*P < 0.05, **P < 0.01, ***P < 0.001).

BADS, Behavioral Assessment of the Dysexecutive Syndrome; MMSE, mini‐mental state examination; MoCA‐J, Japanese version of the Montreal cognitive assessment; ROCF, Rey‐Osterrieth Complex Figure Test; TAT, The Trail Making Test; WAIS, Wechsler Adult Intelligence Scale; WMS‐R, Wechsler Memory Scale‐Revised.

Table 2b presents the executive function. TMT‐A and TMT‐B were significantly different between type 1 diabetes mellitus and type 2 diabetes mellitus. The total time of TMT‐A and TMT‐B was significantly longer in type 1 diabetes mellitus than in type 2 diabetes mellitus (P < 0.001, P < 0.013). The Victoria Stroop test Part 1 revealed a trend toward more errors in type 1 diabetes mellitus; however, the difference was insignificant (P = 0.09). The complete times of the Victoria Stroop test, BADS on each Action program test and Key search test examination, and Digit span and digit symbol tests were not significantly different between type 1 diabetes mellitus and type 2 diabetes mellitus.

Global cognition, memory, and executive function were insignificantly different according to the mode of type 1 diabetes mellitus onset (Acute and fulminant type 1 diabetes mellitus vs LADA, data not shown).

Multiple linear regression analysis was conducted to distinguish the impacts of several diabetes factors on TMT‐A and the delayed recall of MoCA‐J. Type 1 diabetes mellitus and age were significantly associated with TMT‐A after adjusting for diabetes, HbA1c, and sex. The delayed recall subset of MoCA‐J was significantly associated with age (Table 3).

Table 3.

(a) Multiple linear regression analysis of TMT and several diabetes‐related factors. (b) Multiple linear regression analysis of delayed recall of MoCA‐J and several diabetes‐related factors

	Standard partial regression coefficient β	P value	95% CI
(a)
Age (year)	0.25	0.029*	0.10 to 1.71
Sex	−0.83	0.46	−10.12 to 4.66
Type 1 diabetes	0.33	0.004**	3.54 to 18.01
Duration	0.11	0.35	−0.17 to −0.48
HbA1c (%)	−0.03	0.79	−3.34 to 2.54
(b)
Age (year)	−0.33	0.004**	−0.34 to −0.05
Sex	0.08	0.49	−1.13 to 1.58
Type 1 diabetes	0.27	0.016*	−0.06 to 2.59
Duration	−0.11	0.34	−0.17 to −0.48
HbA1c (%)	−0.14	0.21	−0.90 to 017

The predictor variables were age, sex, diabetes type (type 1 or type 2), diabetes duration, and HbA1c (%). Represents the significant differences between the groups (*P < 0.05, **P < 0.01, ***P < 0.001).

Pearson's chi‐square test or Fisher's exact test was used to examine the relationships between diabetes complications and TMT‐A deterioration for the age equivalent average. Significantly more severe hypoglycemia events occurred in the group with deteriorating TMT‐A (P = 0.007). There was no significant difference in complications of retinopathy, nephropathy, stroke, or myocardial infarction with or without TMT‐A deterioration (Table 4). The MoCA‐J scores were also not associated with diabetic complications.

Table 4.

(a) Deterioration of TMA‐A and diabetic complications. (b) Score of MoCA‐J and diabetic complications

	TMT‐A		P value
	Deterioration (−) (n = 52) % (n)	Deterioration (+) (n = 22) % (n)
(a)
Retinopathy	40.4 (21)	40.9 (9)	0.97
Nephropathy	11.5 (6)	18.2 (4)	0.45
History of IHD	15.4 (8)	13.6 (3)	0.58
History of CI	2.2 (1)	4.5 (1)	0.52
History of severe hypoglycemia	3.8 (2)	28.3 (6)	0.007*
Hypertension	69.2 (36)	54.5 (12)	0.27
ARB・ACE	51.9 (27)	54.5 (12)	0.97
Statin	57.7 (30)	45.5 (10)	0.33
Smoking (current/past)	9.6/25.0 (5/13)	9.0/31.8 (2/7)	0.65/0.58

	MoCA‐J > 25	MoCA‐J ≤ 25	P value
	(n = 43) % (n)	(n = 31) % (n)
(b)
Retinopathy	37.2 (16)	45.2 (14)	0.49
Nephropathy	7 (3)	22.6 (7)	0.06
History of IHD	9.3 (4)	22.6 (7)	0.11
History of CI	0 (0)	6.5 (2)	0.17
History of SH	11.6 (5)	9.6 (3)	0.55
Hypertension	55.8 (24)	77.4 (24)	0.06
ARB・ACE	44.2 (19)	64.5 (20)	0.10
Statin	58.1 (25)	48.4 (15)	0.28
Smoking (current/past)	7.0/20.9 (3/9)	12.9/35.4 (4/11)	0.33/0.18

Represents the significant differences between the groups (*P < 0.05, **P < 0.01, ***P < 0.001).

DISCUSSION

This study confirmed that outpatients aged ≥65 years with type 1 diabetes mellitus without dementia were mostly diagnosed after adulthood and had a short duration of type 1 diabetes mellitus. Older patients with type 1 diabetes mellitus had impaired executive function, especially in psychomotor speed, compared with those with type 2 diabetes mellitus. Type 1 diabetes mellitus patients had a higher complete time for TMT‐A and B and a lower score for other classical executive functions, such as the Victoria Stroop test, than patients with type 2 diabetes mellitus. Multiple linear regression analysis revealed that these changes were dependent on age and type 1 diabetes mellitus but independent of diabetic duration and blood glucose control. However, BADS, which was used to assess the degree of executive dysfunction and functional impairment in daily life, did not worsen, and the scores between the two groups were similar. Deterioration in psychomotor function was a characteristic of type 1 diabetes mellitus executive function; however, the impact on daily living might be insignificant.

Montreal Cognitive Assessment‐J is more accurate than MMSE in assessing memory function to identify mild cognitive impairment and dementia in patients with diabetes mellitus ³¹ . The total MoCA score was lower and a delayed recall score in the sub‐study was significantly lower in type 2 diabetes mellitus than in type 1 diabetes mellitus. However, episodic memory in WMS‐R was not significantly different between the two groups. These results indicated that patients with type 1 diabetes mellitus without dementia had inferior executive function and similar verbal memory to those with type 2 diabetes mellitus.

Our results support the previously reported results that linked type 1 diabetes mellitus in children and young adults to a slowing of psychomotor speed ³² , ³³ . Younger adults with type 1 diabetes mellitus (mean age = 33 years) performed worse than controls in terms of intelligence, information processing speed, psychomotor efficiency, attention, cognitive flexibility, and visual perception; however, memory and learning were unaffected ¹⁰ . Older adults demonstrated a significantly worse performance in the speed of information processing, but memory was spared ³⁴ . These studies compared patients with type 1 diabetes mellitus and nondiabetic patients. Our study was conducted with patients with type 2 diabetes mellitus, and not nondiabetics, as a control. However, the completion time of TMT‐A in our patients with type 1 diabetes mellitus was longer than age‐ and education‐matched normative data ³⁰ . Thus, type 1 diabetes mellitus may result in decreased psychomotor function compared with people without diabetes mellitus. A few studies on older adults revealed how type 1 diabetes mellitus and type 2 diabetes mellitus affect cognition in direct comparison ¹⁵ , ¹⁶ . In a Medalist Study, long duration patients with type 1 diabetes mellitus (medalist) and type 2 diabetes mellitus performed worse on memory and psychomotor function compared with nondiabetic controls. However, memory and psychomotor skills between type 1 diabetes mellitus and type 2 diabetes mellitus were similar ¹⁵ . In the SOLID study indicated that patients with type 1 diabetes mellitus had worse scores on verbal episodic memory, psychomotor speed, and executive function than those with type 2 diabetes mellitus and nondiabetics ¹⁶ . In psychomotor function, our study and the SOLID study revealed it had deteriorated in older adults with type 1 diabetes mellitus than in those with type 2 diabetes mellitus, following past studies for young and middle‐aged adults compared with no diabetes. However, a Medalist Study revealed no difference between both types of diabetes. It was revealed that the deterioration of psychomotor speed in Medalists was correlated with proliferative diabetic retinopathy; type 2 diabetes mellitus was not considered. The cause of the decline in psychomotor function in type 2 diabetes mellitus to the same extent as in the Medalist study was not clear; however, the author speculates that the underlying mechanism might be different. Higher and progressively increasing functional connectivity indices on functional MRI are associated with type 1 diabetes mellitus in response to visuospatial working memory load ³⁵ . However, the underlying mechanisms of the deterioration of psychomotor function in type 1 diabetes mellitus remain poorly understood.

The results of the verbal memory test in older adults with type 1 diabetes mellitus varied between the above‐mentioned studies. In the SOLID study, those who had type 1 diabetes mellitus performed worse on immediate and delayed recall compared with those with type 2 diabetes mellitus. Our study and the Medalist study were not inferior in type 2 diabetes mellitus. However, the SOLID study indicated that differences in verbal episodic memory between type 1 diabetes mellitus and type 2 diabetes mellitus disappeared in the logistic regression analysis directly comparing those with type 1 diabetes mellitus and type 2 diabetes mellitus. Our study excluded dementia, which may not adequately reflect memory impairment in type 1 diabetes mellitus. Furthermore, in the SOLID study, acute metabolic events such as hypoglycemia (30%) and DKA (29%) were highly prevalent. The authors speculate that these factors contribute to cognitive decline in type 1 diabetes mellitus. Other studies indicated that severe hypoglycemia might affect cognitive function and memory with type 1 diabetes mellitus ³⁶ , ³⁷ . The risk of hypoglycemia was higher in the presence of dementia ³⁸ . The rate of severe hypoglycemia in type 1 diabetes mellitus in our study was 16%, which is relatively low compared with other studies. This might be affected by the exclusion of dementia. A low rate of severe hypoglycemia might cause a protective effect on memory function in our type 1 diabetes mellitus patients. In our study, participants with a deterioration of psychomotor function had statistically more frequent severe hypoglycemic events than those with other diabetic complications. In contrast, the group with delayed recall impairments did not show the same association. Our results revealed that hypoglycemia was associated with psychomotor function, not memory, which was consistent with the 32 year follow‐up of DCCT/EDIC ³⁹ . The effect of severe hypoglycemia on cognitive function varies from study to study, which might be due to differences in the evaluation area of cognition. Variables, such as the age at diabetes onset, the disease duration, glycemic control, micro‐and macrovascular problems, could influence the outcomes ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ . Ferguson et al. ⁴⁵ reported that early‐onset type 1 diabetes mellitus had a statistically significant lower intelligence and more prevalent brain atrophy than those with late‐onset type 1 diabetes mellitus. Learning and memory skills are more affected by type 1 diabetes mellitus with early onset rather than late onset ⁴⁶ . Disease duration might be another important factor in cognition in type 1 diabetes mellitus. Gao et al. ⁴⁰ indicated that delayed recall task performance is related to disease duration. Most of our patients were late onset, and the duration of type 1 diabetes mellitus was shorter than that of juvenile‐onset type 1 diabetes mellitus. In the early days, patients with juvenile‐onset type 1 diabetes mellitus received treatment with primitive insulin and glucose monitoring systems. Current patients taking advanced insulin and technology may manage their disease more easily; therefore, the impact on memory and executive functions might be small. These findings indicate that late‐onset and shorter durations of diabetes may conserve the cognitive and memory abilities of patients.

Furthermore, chronic hyperglycemia might increase the risk of dementia and low executive function ⁴¹ , ⁴² , ⁴³ . A meta‐analysis of six adults with type 1 diabetes mellitus revealed that poor glycemic control (HbA1c of >8%) caused a moderate but significant decline in memory function.

Moreover, diabetic microvascular and macrovascular complications were associated with decreased cognition and executive function deterioration ³² , ⁴⁷ , ⁴⁸ , ⁴⁹ , while others found no relationship ⁵⁰ . Our study revealed that patients with a decreased MoCA score had a higher rate of hypertension and nephropathy, although the association between microangiopathy and cognitive deterioration was not statistically significant.

Our study has significant limitations. First, this was a single‐center cross‐sectional study with a small sample size. The HbA1c levels were only determined once, which could not indicate the long‐term glucose level. We could not consider several diabetic complications and cognitive outcome in multiple regression analysis because of the small sample size; hence, the clinical relationship between diabetes‐related factors and cognition was not established. Furthermore, the patients were all those who could independently visit our hospital via public transportation. Patients with impaired activity of daily living due to severe diabetic complications were possibly excluded. Our results may not apply to the general type 1 diabetes mellitus population and survival bias might have affected our results. The duration of type 1 diabetes mellitus in our study was shorter than that in previous studies. Other studies found statistically significant differences in the disease duration between juvenile‐onset type 1 diabetes mellitus and type 2 diabetes mellitus, although our study included a homogeneous population with type 1 diabetes mellitus and type 2 diabetes mellitus. This might be a strength of this study. Despite the homogeneous patient population, there were significant differences in C‐peptide levels, which represents insulin secretion, and the BMI, which affected insulin resistance in our study, between type 1 diabetes mellitus and type 2 diabetes mellitus. The differences in cognitive and executive functions between type 1 diabetes mellitus and type 2 diabetes mellitus might have been influenced by insulin deficiency in type 1 diabetes mellitus and insulin resistance in type 2 diabetes mellitus ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ . Furthermore, cognitive function in the elderly with type 1 diabetes mellitus has been reported in Europe and the USA ¹⁰ , ¹⁵ , ¹⁶ . This is the first study to clarify cognitive and executive functions in Japanese elderly patients with type 1 diabetes mellitus.

Older adults with type 1 diabetes mellitus without dementia had a late onset and a short duration of type 1 diabetes mellitus, and their executive function and psychomotor speed were worse than those with type 2 diabetes mellitus. However, their comprehensive and practical executive function and memory were similar; therefore, the decline in the cognitive and executive performance of type 1 diabetes mellitus might have minimal real‐life effects. Further prospective and longitudinal studies with an increased number of cases are needed to clarify the cognitive and executive functions of type 1 diabetes mellitus and related factors.

AUTHOR CONTRIBUTIONS

Kimio Matsumura: Data collection. Takayasu Uchida: Data collection. Yuya Suzuki: Data collection. Akihiro Nishimura: Data collection. Minoru Okubo: Data collection. Yukifusa Igeta: Data collection. Tetsuro Kobayashi: Data collection. Takashi Sakurai: Conceived and designed the analysis. Yasumichi Mori: Conceived and designed the analysis.

DISCLOSURE

The authors declare no conflict of interest.

Approval of the research protocol: The suitably constituted Ethics Committee of the Toranomon Hospital (Approval No. 1517) has approved the research protocol for this project, which conforms to the provisions of the Declaration of Helsinki.

Informed consent: All participants provided informed consent and signed informed consent forms.

Registry and the registration no. of the study/trial: February 22, 2024 and UMIN: R000033469.

Animal studies: N/A.