Abstract
This study investigated the relationship between diabetes-related eye problems and dementia as well as the impact of dementia on vision-related quality of life and activities of daily living for patients with Type 2 diabetes.
Dementia and Type 2 diabetes mellitus (T2DM) are multifaceted diseases and are critical global health concerns. Dementia not only affects memory but also causes changes in behavior and personality, thus affecting daily activities (World Health Organization [WHO], 2017). In 2019, the WHO estimated that 55 million people had dementia, with nearly 10 million new cases annually (Rizzi et al., 2014; WHO, 2023). Dementia is one of the major causes of disability and dependency among older people globally (WHO, 2017, 2023). According to the International Diabetic Federation, the number of people with diabetes worldwide will rise from 536.6 million in 2021 to 783.7 million in 2045 (Magliano & Boyko, 2021).
Several studies show that diabetes can increase the risk of dementia (Chen et al., 2022; Ninomiya, 2014; Zhang et al., 2017). According to Ninomiya’s (2014) meta-analysis, older people with diabetes are 1.5- to 2.5-fold more likely to have dementia. The increased risk may arise from cerebrovascular complications of diabetes, hyperglycemia, hypoglycemia, glycemic variability, and antidiabetic agents (Chen et al., 2022; Gómez-Guijarro et al., 2023; van Sloten et al., 2020). Diabetes increases the risk of stroke, which is the leading cause of vascular dementia (van Sloten et al., 2020). Several long-term follow-up studies have reported that patients with recurrent hypoglycemia, compared with people without it, have a 1.5 to 2 times higher risk of cognitive impairment (Gómez-Guijarro et al., 2023; van Sloten et al., 2020). Many studies have found that glycemic variability is also associated with dementia (Chen et al., 2022). Previous research has suggested that medications such as glucagon-like peptide-1 receptor agonist (GLP-1RA; Gejl et al., 2016), sodium–glucose cotransporter-2 inhibitors (SGLT-2is; Shaikh et al., 2016), and thiazolidinediones (TZDs; Tang et al., 2022) may affect or reduce the development of dementia. However, results from different studies are inconsistent (Tian et al., 2023).
Another potential risk factor associated with T2DM for the development of dementia is visual impairment (Chan et al., 2023). T2DM increases the risk of developing visual complications such as diabetic retinopathy (DR) and diabetic macular edema (DME), both of which can potentially lead to visual impairment and, in severe cases, blindness (Ruta et al., 2013). Several studies have suggested an association between DR and dementia (Chan et al., 2023; Cheng et al., 2021). In addition to DR and DME, other visual comorbidities, such as cataracts and glaucoma, or age-related conditions may increase the risk of low vision in patients with T2DM (Glassman et al., 2024). Low vision alone has been shown to limit a person’s ability to perform daily activities and social interaction and to accelerate cognitive decline (Deremeik et al., 2007; Elliott et al., 2014).
Regarding dementia, previous studies have shown a link between dementia and visual impairment. Visual impairment can include impaired visual acuity (VA), which refers to the sharpness or clarity of vision. It also includes decreased contrast sensitivity (CS), defined as the ability to detect subtle differences in intensity between lighter and darker regions of an image, pattern, or objects of different sizes (Cormack et al., 2000). Additionally, visual impairment increased visual crowding (Yong et al., 2014), a phenomenon in which the perception of an object is affected by the presence of other nearby objects. Visual crowding can affect central and peripheral VA, reading performance, and object localization, especially in cluttered environments (Yong et al., 2014). All of these factors affect vision-related quality of life (VRQoL) in people with dementia (Tyler et al., 2022). Poor vision may increase the risk of developing dementia, and additionally, dementia may further impair visual abilities, with both conditions interacting with each other. Therefore, it is necessary to understand the complex relationships between T2DM, dementia, and visual loss.
Previous research has primarily measured the impact of T2DM on visual complications and vision loss by assessing best corrected VA (Glassman et al., 2024). However, eye disease affects not only VA but also other important visual functions, functional vision, and VRQoL (van Nispen et al., 2020). Visual function refers to objective eye performance and is primarily measured by VA, CS, and visual field tests (Bennett et al., 2019; Colenbrander, 2003).
In contrast, functional vision pertains to a person’s ability to perform vision-related activities such as reading, writing, and mobility and is primarily measured with patient-reported outcome measures (PROMs; Colenbrander, 2003). Although PROMs for assessing functional vision and VRQoL have become increasingly important in vision care (Braithwaite et al., 2019), few studies have applied PROMs to assess and compare people with T2DM, with or without dementia, which represents a gap in our understanding and application of these measures in this population.
To better understand the complex relationships between T2DM, dementia, visual function, and functional vision, we conducted two substudies involving database and real-world data collection simultaneously. The purpose of the first substudy was to identify predictors of cognitive dysfunction in people with T2DM, using a retrospective cohort study design. In the second substudy, we investigated the potential associations between visual function, vision-dependent activities of daily living (ADLs), and dementia in people with T2DM, using a nested case-control design.
Method
We obtained approval for data collection and analysis of the two substudies from the Taipei City Hospital Institutional Review Board, and all tests were performed in accordance with the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants in Substudy 2.
Data Source and Cohort Identification of the Two Substudies
In Taiwan, the Diabetic Shared Care System was established with the aim of enhancing the quality of care for patients with diabetes. On registration, patients with diabetes provide baseline data, and regular biochemical check-ups are mandated. At the Department of Endocrinology and Metabolism of Taipei City Hospital, visiting patients with diabetes are offered registration in the Diabetic Shared Care System, subject to the patients’ consent. From May 2002 to June 2023, 7,000 patients with diabetes became members of this system. Annually, the program attracts 300 to 500 participants, with over 6,000 patients actively receiving follow-up care.
For Substudy 1, we adopted a retrospective cohort study design and aimed to recruit patients with diabetes who were registered in the Diabetic Shared Care System at Taipei City Hospital. The recruitment window was from May 2002 to June 2023, targeting an enrollment of 4,000 patients.
For Substudy 2, we used a nested case-control study design. Cases were randomly selected from dementia cases within the cohort of Substudy 1. Control participants were chosen from among patients who had diabetes without dementia. For each case, we recruited two control participants from the Substudy 1 cohort, using specific frequency matching criteria such as age and sex.
Substudy 1: Study Design (Retrospective Cohort Study)
Participants
Cases were defined as patients with T2DM who met the criteria for dementia in this cohort. Identification of study participants with dementia was based on the International Classification of Diseases, 10th Revision (WHO, 2014), Codes F00.00 to F03.90; the International Classification of Diseases, Clinical Modification (Centers for Disease Control and Prevention, National Center for Health Statistics, 2014), Codes 290.00 to 290.40, 291.2, and 294.0; or prescription of medications for dementia. Control participants were defined as patients with T2DM who did not meet the criteria for dementia in this cohort.
Data Collection
Patients’ demographic data, including age, sex, duration of diabetes, socioeconomic status, education level, and smoking and drinking habits, along with relevant clinical variables such as body weight, body height, blood pressure, and waist circumference, were collected from medical histories and physical examinations. Laboratory data, including those on HbA1c, total cholesterol, triglycerides, low-density lipoprotein, high-density lipoprotein, and creatinine, were obtained through follow-ups conducted every 3 mo within the diabetes management program. Information regarding medications and comorbidities was also gathered through electronic medical records. In this study, HbA1c data were obtained at two specific data points: The first was obtained at baseline, and the second was the most recent measurement before the analysis. Other biochemical data obtained at baseline were used for analysis. Baseline HbA1c, recent HbA1c, other baseline biochemical data, recent medication usage, and recent comorbidity status were used in the regression analysis.
Substudy 2: Study Design (Nested Case-Control Study)
Participants
Cases were randomly selected from cases in the Substudy 1 cohort. Control participants were patients who had diabetes without cognitive impairment. For each case, we recruited two control participants from the Diabetic Shared Care System, using a set of frequency-matching criteria, including age and sex. Exclusion criteria were unwillingness or noncooperation with vision testing and a clinical history or evidence of eye or neurological disease that was not caused by diabetes, including trauma, multiple sclerosis, and stroke.
Procedure and Data Collection
After we obtained written informed consent from the participants, three visual function (distance VA, distance crowded VA, and CS) and two functional vision questionnaires were assessed by two trained occupational therapists. Visual function and functional vision were measured at a single clinical visit.
Demographic variables, including age, gender, education, employment status, smoking, and alcohol consumption, were obtained from medical records or by interview. Collection of baseline or recent biochemical data was the same as for Substudy 1: through electronic medical records. The Mini-Mental State Examination (MMSE; Folstein et al., 1975) and Clinical Dementia Rating (CDR) scale (Morris, 1993) were administered by a neurologist or psychologist, and the scale scores were obtained through electronic medical records. The cutoff point for dementia was 24 for the MMSE and 0.5 for the CDR in this study.
Measure of Visual Function
Visual function assessment included presenting distance VA, crowded distance VA, and CS function (CSF). Both binocular and monocular conditions were used for distance VA, whereas crowded VA and CSF were tested binocularly (Liao et al., 2021). We administered the Freiburg Visual Acuity Test to determine distance and crowded VA at 3 m on a CTX EX951F 19-in. monitor (Bach, 1996, 2007), using, as parameters, Landolt Cs, four choices of gap direction, starting logarithm of the minimum angle of resolution (LogMAR), 40 s of display timeout, occasional easy trials, auditory feedback, and a white background color with a black foreground color. Crowded VA used rows of optotypes spaced two-gap distances apart. The participant pointed to the gap directions of the Landolt C optotype or provided a verbal response. VA results are presented in logMAR units.
For CSF, we used the CSV-1000 test at 2.5 m, which provides a fluorescent luminance chart at a standardized light level (85 cd/m2) to display circular patches of vertical sine-wave gratings at four spatial frequencies (SFs 3, 6, 12, and 18 cycles per degree [cpd]). Each SF is presented on a separate row; each row has 17 circular patches, with eight pairs of circular patches and a sample patch with the highest contrast grating and at the beginning of the row. A two-alternative quasi–forced-choice procedure is used to encourage the participant to decide or guess whether the upper or lower patch contains the grating. CS values are expressed in log units and range from 0.70 to 2.08, 0.91 to 2.29, 0.61 to 1.99, and 0.17 to 1.55 for 3, 6, 12, and 18 cpd, respectively. The area under the log CSF (AULCSF) was also determined by fitting a third-order polynomial to the CS values at four log SFs, providing a broad measure of CS across all SFs (Applegate et al., 1998).
Measure of Vision-Dependent ADLs and VRQoL
The Revised Self-Report Assessment of Functional Visual Performance (R–SRAFVP) encompasses 33 items across nine categories: personal care, oral care, meal and laundry preparation, financial management, telephone usage, personal preference activities, reading, writing, and functional mobility. It evaluates the capability of adults to carry out vision-dependent daily activities, primarily focusing on instrumental ADLs (IADLs) because they demand more vision than basic ADLs (Snow et al., 2018; Zemina et al., 2018). The rating scale is a 5-point Likert-type scale ranging from unable (0) to independent (4). The occupational therapist read aloud each R–SRAFVP item and description and asked the participants to rate their ability in terms of the level of difficulty they experience carrying out each daily activity. Higher scores indicate more independence. Some patients never complete a task or are no longer able to complete a task for nonvisual reasons. Therefore, the R–SRAFVP score is expressed as a percentage of the client’s ability to perform desired or relevant tasks (Snow et al., 2018; Zemina et al., 2018).
The 25-item National Eye Institute Visual Function Questionnaire (NEI–VFQ 25) is a well-known visual function questionnaire that is frequently used with people who have diabetes and low vision (Cusick et al., 2005; Khoo et al., 2019). It includes 12 vision-targeted subscales—general health, general vision, ocular pain, near activities, distance activities, social functioning, mental health, role difficulties, dependency, driving, color vision, and peripheral vision—to measure the impact of vision on multiple dimensions of VRQoL (Mangione et al., 2001). The standard algorithm is used to transfer the raw scores of each item into transferred item scores, subscale scores, and composite scores ranging from 0 to 100. In general, the scale scores of 11 subscales, excluding general health, are averaged to yield a composite score (Mangione et al., 1998).
Statistical Analysis
We performed all statistical analyses using IBM SPSS Statistics software (Version 22). In the statistical testing, a two-sided p ≤ .05 was considered statistically significant. The distributional properties of continuous variables are expressed as M (SD), and categorical variables are presented as frequency (%). In univariate analysis, we examined the differences in the distributions of continuous variables and categorical variables between the dementia and control groups, using a two-sample t test and χ2 analysis. Next, we conducted multivariate analysis by fitting a linear regression model and a logistic regression model to estimate the adjusted effects of risk factors on continuous and binary outcomes. The stepwise variable selection procedure was applied to obtain the best candidate final regression model.
In Substudy 1, we examined dementia as a binary outcome variable. We used logistic regression models, incorporating all potential predictors as independent variables. In Substudy 2, we evaluated visual function test and vision-dependent ADL and VRQoL questionnaire scores as a continuous outcome. We used multiple linear regression models to assess the impact of dementia on these scores, controlling for all other potential predictors.
Results
Substudy 1
The first substudy had a total sample of 4,454 participants. The baseline characteristics of these patients are depicted in Table 1. On average, our patient group was 70.5 yr old and had been managing diabetes for approximately 13.2 yr. The mean blood pressure was recorded as 129/74.5 mmHg. The gender distribution was relatively balanced, with women constituting approximately 47% of our study population. The average body mass index at baseline was 26.35. The average HbA1c was 8.1% at baseline and dropped to 7.1% at the most recent measurement.
Substudy 1: Baseline Characteristics of All Participants in This Cohort
Note. N = 4,454. ACR = urine albumin to creatinine ratio; BMI = body mass index; GLP-1RA = glucagon-like peptide-1 receptor agonist; DBP = diastolic blood pressure; DPP-4i = dipeptidyl peptidase-4 inhibitor; LDL = low-density lipoprotein cholesterol; HDL = high-density lipoprotein cholesterol; GPT = glutamic–pyruvic transaminase; SBP = systolic blood pressure; SGLT-2i = sodium–glucose cotransporter-2 inhibitor; TZD = thiazolidinedione.
Pertaining to common eye conditions among our patients, we noted prevalences of 19.5% for DR, 52.9% for cataracts, and 6.1% for glaucoma. Meanwhile, 4.6% of the patients had dementia.
Considering prescribed diabetes management medications, the distribution was as follows: 44.7% of the patients were on SGLT-2is, 15.0% were on GLP-1RAs, 48.9% were on dipeptidyl peptidase-4 (DPP-4) inhibitors, and 49.3% were on TZDs. We expand on these figures in Table 2.
Substudy 1: Potential Risk Factors to Predict Dementia by Logistic Regression Models (Final Model)
Note. CI = confidence interval; DBP = diastolic blood pressure; DR = diabetic retinopathy; PDR = proliferative diabetic retinopathy; ref. = reference category; SGLT-2i = sodium–glucose cotransporter-2 inhibitor. p values were significant at <.05.
Using multiple logistic regression models, we established that age, lower educational level, and elevated diastolic blood pressure had significant correlations with elevated dementia risk. This is further detailed in Table 2, which presents the odds ratios (ORs) and 95% confidence intervals (CIs) for the variables age (OR = 1.11, 95% CI [1.09, 1.13]), higher education level (>16 yr of education, compared with <9; OR = 0.38, 95% CI [0.18, 0.80]), and diastolic blood pressure at 1.023 (1.003–1.045 per mmHg increase).
Baseline biochemical data, including HbA1c, lipid profile, and creatinine, exhibited no significant correlations with dementia. Among antidiabetic medications, only SGLT-2is demonstrated a significant association with dementia (OR = 1.73, 95% CI [1.10, 2.74]). Conversely, GLP-1RAs, DPP-4 inhibitors, and TZDs had no associations with dementia.
Last, proliferative diabetic retinopathy (PDR) had a borderline significant association (p = .098) with dementia, compared with no DR. Other ophthalmological complications, such as glaucoma and macular edema, seemed to enhance dementia risk, but the results failed to reach statistical significance, possibly because of the limited number of cases.
Substudy 2: Study Population
This substudy included 33 cases with T2DM and dementia and 67 matched control participants with T2DM without dementia. The mean ages of the patients in the case and control groups were 78.7 and 76.6 yr, respectively. Approximately 55% of the patients in the case group and 46% of the patients in the control group were female. The mean MMSE and CDR scale scores in the case group were 21.0 (SD = 4.3) and 0.90 (SD = 0.52), respectively. The mean durations of diabetes were 16.8 yr (SD = 8.2) and 16.0 yr (SD = 8.9), and the mean HbA1c levels were 6.9% (SD = 1.2) and 7.8% (SD = 6.8) in the case and control groups, respectively. Demographic characteristics, baseline biochemical data, prevalence of hypertension, and ocular conditions of patients with and without dementia were compared, and there were no significant differences between the two groups (Table 3).
Substudy 2: Demographic Characteristics, Biochemical Data, and Ocular Conditions of the Participants
Note. Continuous variables are presented as mean (SD); the difference between the patients with diabetes with dementia and patients with diabetes without dementia was determined using the independent t test. Categorical variables are presented as n (%); the difference between patients with diabetes and control participants was determined using chi-square analysis. ACR = urine albumin-to-creatinine ratio; DME = diabetic macular edema; DR = diabetic retinopathy; eGFR = estimated glomerular filtration rate; GPT = glutamic–pyruvic transaminase; HbA1c = hemoglobin A1c; HDL = high-density lipoprotein; LDL = low-density lipoprotein; T-CHO = total cholesterol; TG = triglycerides.
*Significant at p < .05.
Visual Function in Patients With and Without Dementia
The descriptive data of VA, crowded VA, and CSF at 3, 6, 12, and 18 cpd of all patients with T2DM with dementia, and of those without dementia, are shown in Table 4. The differences between the patients who have diabetes with dementia and those without dementia for various visual functions, compared by t test first, are shown in Table A.1 in the Supplemental Appendix (available online with this article at https://research.aota.org/ajot). There were no significant differences in VA or crowded VA, so these variables were excluded from further regression analysis. The differences in various visual functions between patients who have diabetes with dementia and those without dementia were further tested, using the multiple linear regression model after controlling for demographic, biochemical, and ocular variables. Although the results showed no significant differences between cases and control participants in CSF in four types of SFs or the AULCSF, the group of patients with DM and dementia was numerically lower at 3 cpd (Table 4).
Substudy 2: Multiple Linear Regression Analysis on the Performance of Visual Function and Functional Vision Assessment in Patients With Diabetes
Note. AULCSF = area under log contrast sensitivity function; CI = conference interval; cpd = cycles per degree; NEI–VFQ 25 = 25-item National Eye Institute Visual Function Questionnaire; R–SRAFVP = Revised Self-Report Assessment of Functional Visual Performance; SF = spatial frequency.
*Significant at p < .05.
Vision-Dependent ADLs and VRQOL in Patients Who Developed Dementia or Not
Functional vision and VRQoL were assessed with two PROMs: the R–SRAFVP and the NEI–VFQ 25. Descriptive data of functional vision performance in the case and control groups are shown in Table 4. On the NEI–VFQ 25, more than 90% of cases and control participants no longer drove, and both groups achieved full scores on the Color Vision subscale. In addition to the General Health subscale, the Driving and Color Vision subscales were also excluded from the composite score in our study. As shown in Table A.1 in the Supplemental Appendix, several subscales of the R–SRAFVP and NEI–VFQ 25 were significantly different, so data from both tests were included in further regression analysis. As can be seen in Table 4, the total scores of the R–SRAFVP and the subscale scores for financial management, using the telephone, and reading were significantly lower in patients who developed dementia than in those who did not (p < .05). However, there was no significant difference in the composite score of the NEI–VFQ 25 between patients who developed dementia and those who did not (p = .067). Compared with the patients without dementia, those with dementia had significantly lower scores on the subscales that measured near activities, mental health, role difficulties, and dependency (Table 4).
Discussion
This is the first study to use measures of visual function, vision-dependent ADL, and VRQOL for patients who have T2DM with and without dementia to determine that dementia affects functional vision more than visual function. PROMs, rather than visual function tests, are useful tools for distinguishing visual performance between people with T2DM with and without dementia. In Substudy 1, age, lower education, and elevated diastolic blood pressure were significantly associated with dementia, and PDR showed a borderline significant association with dementia in people with T2DM after adjustment for risk factors. Although biochemical and ocular data were not correlated with dementia, the use of SGLT-2is revealed a significant association with dementia. Substudy 2 showed no significant difference in visual function between patients who have diabetes with and without dementia after adjustment for risk factors. Only a trend toward poorer performance at 3 cpd of the CSV-1000 test was found with the t test. However, people with T2DM and dementia scored lower on the R–SRAFVP and slightly lower on the NEI–VFQ 25. Several subscales of the R–SRAFVP and NEI–VFQ 25 were shown to be significantly different.
Although maintaining good glycemic control is fundamental in managing T2DM to reduce the risk of end-organ damage, the evidence affirming the benefits of rigorous blood glucose control on cognitive outcomes remains inconclusive. The MIND substudy of the ACCORD trial found no difference in cognitive function between standard and intensive glucose control groups (Launer et al., 2011). A meta-analysis of five trials (n = 24,297) showed no clear cognitive benefits from strict glycemic control (Tuligenga, 2015). In our study, baseline A1c and current A1c seemed not to be associated with dementia. Given the concerns about hypoglycemia risks, the A1c target was adjusted higher for the older people, especially those with a heightened risk of hypoglycemia. Stricter blood sugar control can lead to more instances of hypoglycemia, which, in turn, increases the risk of dementia. These are the possible explanations for the lack of an association between glycemic control and dementia found in our study.
Some may argue that the choice of glucose-lowering medication significantly influences the development of dementia (Chen et al., 2020; Gejl et al., 2016; Shaikh et al., 2016; Tang et al., 2022; Tian et al., 2023). However, our research was an observational study, not an interventional one. The medication usage and dementia status were also recorded simultaneously, making it difficult to establish a clear cause-and-effect relationship. As such, the use of specific glucose-lowering treatments may be a consequence, rather than a precursor, of dementia.
We found an association between the use of SGLT-2is and an increased risk of dementia. However, on the basis of current guidelines, SGLT-2is are prescribed mainly for patients with atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease, all of which are significant comorbidities associated with dementia. Thus, the association between SGLT-2i usage and dementia might be coincidental rather than causal. Although, usually, T2DM is not the primary reason for referral to occupational therapy, its high incidence and numerous complications, such as cardiovascular disease, cerebrovascular accidents, and visual loss, often necessitate occupational therapy services. Therefore, it is essential for occupational therapists to be familiar with glucose-lowering medications, glycemic control strategies, and related chronic conditions to help people with diabetes self-manage their diabetes, maintain their independence, and improve their overall quality of life.
Our data also suggested a potential association between PDR and dementia, supporting our hypothesis that visual impairment may increase the risk of dementia in patients with diabetes, possibly because of sensory deprivation. Other eye conditions, such as cataracts, glaucoma, and macular edema, did not show a clear association with dementia, possibly because of the small number of such cases in our sample.
Research on older adults consistently demonstrates an association between visual impairment and dementia onset. The correlation is attributed to factors such as age-related risk, decreased visual input to the brain, overlapping pathologies, or diminished social and physical activity participation (Shang et al., 2021). Assessing visual function by measuring VA is a common method for analyzing the association between diabetes and dementia. In addition to assessing VA, our Substudy 2 also evaluated crowded VA as a measure of visual function. Crowded VA was assessed because neurodegenerative diseases may lead to increased difficulty in discriminating objects in crowded environments (Yong et al., 2014). No significant differences in VA or crowded VA were noted between the dementia and nondementia groups, possibly because of the study’s limited sample size.
CS testing is commonly used to evaluate visual function in patients with cognitive impairment (Hong et al., 2020; Risacher et al., 2013). Patients with Alzheimer’s disease show a notable reduction in CS, which correlates with higher dementia risks in older women (Ward et al., 2018). A decrease in CS across SFs ranging from 1.5 to 18 cpd, as supported by the AULCSF, is also considered an early indication of cognitive decline in patients with Parkinson’s disease (Hong et al., 2020). Our research, pioneering the use of CS to differentiate patients with T2DM with or without dementia, revealed a significant difference at a low SF (3 cpd) and a marginal difference at a high SF (18 cpd) and in the AULCSF. However, these effects were not present in the following regression analysis after controlling for related variables. Compared with VA findings, assessing CS may be a potentially sensitive method to identify the early decline of cognitive impairment in people with T2DM. However, validation will require larger studies.
Despite the known association between T2DM and dementia, disease-specific VRQoL measures are absent. Our data showed that patients who had T2DM with dementia scored significantly lower on several IADL subscales of the R–SRAFVP, aligning with prior research indicating that cognitive impairment affects IADLs before basic daily tasks. ADLs encompass various types of abilities, one of which is cognitive function (Mlinac & Feng, 2016). Cognitive impairment initially affects IADLs, such as managing finances, before affecting basic ADLs. Our findings corroborate previous research on dementia (Mlinac & Feng, 2016). Moreover, the performance of ADL is strongly correlated with measures of cognition and dementia severity (Galasko et al., 1997; Potashman et al., 2023). Further studies should explore whether visual deficits in T2DM influence cognitive and ADL performance, because knowledge of such influences could aid in early detection and intervention.
The NEI–VFQ 25 is a widely used tool for assessing VRQoL in people with low vision, including those with DM (Cusick et al., 2005; Khoo et al., 2019). Because a large number of participants either did not drive or had never driven, and all participants reported no color vision issues, the subscale scores for driving and color vision were excluded from the computation of composite scores. The results revealed significantly lower scores on subscales measuring near activities, mental health, role difficulties, and dependency in the dementia group. Our study results showed that patients who had T2DM with dementia had lower QoL scores in the socioemotional dimension than in the visual function dimension (Abe et al., 2019), in contrast to the results of another study (Tyler et al., 2022). A possible explanation is that the mean VA, expressed in decimal notation, was greater than 0.4 among patients with dementia. The NEI–VFQ 25 was originally designed for patients with low vision, so the test may not be able to accurately measure the impact of the visual function dimension on VRQoL in patients with dementia. However, related studies, including studies on the selection of VRQoL tools in patients with T2DM, are limited; therefore, it is difficult to reach a conclusion.
In this study, we combined data from two substudies to examine the interaction among diabetes, dementia, visual function, functional vision, and VRQoL. To our knowledge, this is the first study to address these important chronic conditions simultaneously. Additionally, we assessed not only VA but also CS.
However, our study has several limitations. In Substudy 2, a primary limitation was the small sample size, which resulted from the challenges associated with recruiting patients with dementia who were able to complete the comprehensive set of tests. Another limitation was the lack of near VA assessment. Although both distance VA and near VA offer valuable insight into visual function, we focused on distance VA because of feasibility constraints. Regardless of the limitations, because occupational therapists have contributed to the care of patients with T2DM (Clarke et al., 2019; Kravitz et al., 2021), the results may provide some useful evidence to guide routine clinical practice.
Implications for Occupational Therapy Practice
T2DM is a growing health concern around the world. The results of this study have the following implications for occupational therapists providing services to patients with diabetes: Occupational therapists working with patients with T2DM should be aware that age, education, medication use, and high diastolic blood pressure may increase the risk of dementia. Understanding these risk factors can assist occupational therapists in health education and prevention and guide them in assessment and intervention planning. Patients with T2DM and dementia show a significant decline in functional vision. Assessment of vision-dependent ADLs, IADLs, and VRQoL needs be emphasized in clinical practice.
Conclusion
To understand the complex relationships between T2DM, dementia, visual function, and functional vision performance, we conducted two substudies simultaneously. The results indicate a marginally significant correlation between PDR and dementia while also identifying age, lower education, and elevated diastolic blood pressure as significant risk factors for dementia. Although dementia does not significantly alter VA, VA in crowded conditions, or CS, it does have a significant impact on the vision-dependent ADLs and VRQoL. In this study, we administered both vision-dependent ADL and VRQoL questionnaires to patients who have T2DM with and without dementia. In the clinical setting, health care professionals should pay close attention to the assessment of vision-dependent ADL and VRQoL in the management and care of this population.
Supplemental Material
Supplementary material for Influence of Dementia on Vision-Related Functional Performance Among Patients With Type 2 Diabetes
Supplementary material, sj-pdf-1-aot-10.5014_ajot.2025.050631.pdf for Influence of Dementia on Vision-Related Functional Performance Among Patients With Type 2 Diabetes by Li-Ting Tsai, Chung-Sen Chen, Chia-Wei Hung, I-Mo Fang and Kuo-Meng Liao in The American Journal of Occupational Therapy
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
