Skip to main content
NCBI home page
The Journal of Clinical Endocrinology and Metabolism logoLink to The Journal of Clinical Endocrinology and Metabolism
. 2020 Dec 30;106(4):e1139–e1149. doi: 10.1210/clinem/dgaa921

Association of Cognitive Function and Retinal Neural and Vascular Structure in Type 1 Diabetes

Ward Fickweiler 1,2, Emily A Wolfson 1, Samantha M Paniagua 1, Marc Gregory Yu 3, Atif Adam 1,3, Vanessa Bahnam 1, Konstantina Sampani 1,2, I-Hsien Wu 1, Gail Musen 1, Lloyd P Aiello 1,2,4, Hetal Shah 1,3, Jennifer K Sun 1,2,4, George L King 1,3,4,✉
PMCID: PMC7993575  PMID: 33378459

Abstract

Context

Cognitive dysfunction is a growing and understudied public health issue in the aging type 1 diabetic population and is difficult and time-consuming to diagnose. Studies in long duration type 1 diabetes have reported the presence of proliferative diabetic retinopathy was associated with cognitive dysfunction.

Objective

This study assessed whether structural and vascular abnormalities of the retina, representing an extension of the central nervous system, are associated with cognitive impairment and other complications of type 1 diabetes.

Methods

An observational cross-sectional study of individuals with 50 or more years of type 1 diabetes (Joslin Medalist Study) was conducted at a university hospital in the United States. The study included 129 participants with complete cognitive testing. Validated cognitive testing measures included psychomotor speed, and immediate, and delayed memory. Optical coherence tomography (OCT) and OCT angiography (OCTA) were performed to obtain neural retinal layer thicknesses and vascular density for superficial (SCP) and deep retinal capillary plexus (DCP). Multivariable modeling was adjusted for potential confounders associated with outcomes in unadjusted analyses.

Results

Decreased vessel density of the SCP and DCP was associated with worse delayed memory (DCP: P = .002) and dominant hand psychomotor speed (SCP: P = .01). Thinning of the retinal outer nuclear layer was associated with worse psychomotor speed both in nondominant and dominant hands (P = .01 and P = .05, respectively). Outer plexiform layer thickness was associated with delayed memory (P = .04).

Conclusion

These findings suggest that noninvasive retinal imaging using OCT and OCTA may assist in estimating the risks for cognitive dysfunction in people with type 1 diabetes.

Keywords: type 1 diabetes, diabetic complications, cognition, diabetic retinopathy, ocular imaging

The burden of diabetes and its complications, including cognitive dysfunction, is expected to rise as life expectancy increases. Cognitive impairment is difficult and time-consuming to diagnose (1). The development of interventions to address this burden has been limited by the lack of a detailed clinical assessment of cognitive dysfunction and other complications in a large cohort with long duration of type 1 diabetes.

Previously, in a subset of the Joslin Medalist Study, which enrolls patients with insulin-dependent diabetes of duration of 50 or more years, we reported that people with type 1 and type 2 diabetes had a comparable prevalence of cognitive dysfunctions, which were correlated to the history of cardiovascular disease (CVD) and presence of proliferative diabetic retinopathy (PDR) (2-4). A study of the Kaiser Permanente Northern California registry also reported that older people with type 1 diabetes had a higher risk for cognitive dysfunction than nondiabetic controls (5). The association between cognitive dysfunction and PDR in people with type 2 diabetes has been reported (6) and attributed to the shared origin of the retina and central nervous system (CNS) and the similarity of their microvascular structures, including the localization of mural cells that constitute an intrinsic part of the vascular wall and alterations in the blood-retina and blood-brain barrier in diabetes (7). In addition, diabetes-induced retinal and brain neurodegeneration may share common pathogenic pathways such as oxidative stress (8, 9). Thus, it is possible that a detailed characterization of the retinal neuronal and vascular structures may identify markers associated with early cognitive and CNS changes in type 1 diabetes.

Recent studies using retinal imaging techniques, such as optical coherence tomography (OCT) (10-12) and OCT angiography (OCTA) (13-17), suggest correlations between cognitive impairment and changes in retinal cell layers and vasculature. However, these relationships have not been examined in older individuals who have experienced type 1 diabetes over multiple decades. To address this gap, this study correlated retinal neural and vascular structure observed by OCT and OCTA with cognitive changes and other complications in a large cohort of older individuals with type 1 diabetes.

Materials and Methods

Study design

Using an exploratory, cross-sectional study design, this investigation included participants from a subset of the Joslin Medalist Study (“Medalists”), individuals with insulin-dependent diabetes for 50 or more years who have been followed longitudinally since 2004. A detailed description of the Medalist Study has been previously reported (2, 3), and the cognitive subset was recruited based on those Medalists who consented for cognitive testing. Briefly, Medalists are a diverse group of individuals (n = 1019) from all 50 US states who have been extensively characterized at Joslin Diabetes Center by questionnaires, clinical and ophthalmologic examinations, and biospecimen analysis for type 1 diabetes human leukocyte antigen risk alleles, autoantibodies, and mixed-meal tolerance studies for glucose and C-peptide levels. As of August 2019, 129 Medalists who completed cognitive testing and retinal imaging studies were included. Those with a clinical diagnosis of dementia, Alzheimer disease, or memory impairment as part of the Medalist Study were excluded (n = 1). The institutional review board approved this study at Joslin Diabetes Center, and consent was obtained from participants before study procedures began. Diabetic nephropathy (DN) was defined by an estimated glomerular filtration rate less than 60 mL/min/1.73 m2, calculated using the Chronic Kidney Disease Epidemiology Collaboration equation based on serum creatinine (18). CVD status was determined by self-report of coronary artery disease, myocardial infarction, angina, prior cardiac or leg angioplasty, or bypass graft surgery. Retinopathy status was graded within the 7 standard fields on fundus photography using guidelines from the Early Treatment of Diabetic Retinopathy Study (ETDRS). PDR was defined as ETDRS DR severity level of 60 or greater. The Michigan Neuropathy Screening Instrument was used to assess peripheral neuropathy; scores of 2 or greater were considered positive. Glycated hemoglobin A1c (HbA1c) was measured by high-performance liquid chromatography (Tosoh G7 and 2.2) and lipid profiles by standard enzymatic methods (Roche Diagnostics, Denka Seiken, and Asahi Kasei). Serum C-peptide was determined by radioimmunoassay (Beckman Coulter) and validated at the Northwest Lipid Research Laboratory at the University of Washington (19). Human leukocyte antigen genotyping and autoantibody titers of insulinoma antigen 2 and glutamic acid decarboxylase were performed at the Barbara Davis Diabetes Center (20).

Cognitive testing

Cognitive testing procedures were previously described by Musen et al (21). All study participants underwent a validated cognitive battery consisting of the Trail Making and Number-Letter-Switching subtest from the Delis-Kaplan Executive Function System, the Letter-Number-Sequencing from the Wechsler Memory Scale-III, the Rey Auditory Verbal Learning Test, and the Grooved Pegboard both for dominant and nondominant hands (22). The cognitive battery assessed executive function, working memory, immediate and delayed memory, and psychomotor speed, respectively. All cognitive testing was performed by trained research personnel.

Retinal imaging techniques

Retinal imaging was performed at the Joslin Beetham Eye Institute. For spectral-domain OCT (Spectralis, Heidelberg Engineering), 49-line horizontal high-resolution scans were obtained in a 6-mm × 6-mm grid centered on the fovea and segmented by layer using Heidelberg, version 6.0c automated segmentation software. The following layers were evaluated within the central 1-mm subfield: nerve fiber (RNFL), ganglion cell (GCL), inner plexiform (IPL), inner nuclear (INL), outer plexiform (OPL), outer nuclear (ONL), photoreceptor (PHOTO), and retinal pigment epithelium (RPE). OCTA images with a signal strength of 55 or greater were obtained on an Avanti RTVue-XR system (Optovue) using the split-spectrum amplitude-decorrelation angiography modality, with a transverse resolution of 15 μm, axial resolution of 5 μm, 304 × 304 A-scans with 2 repeats per B-scan. Each eye of the participant underwent high-definition scanning of the 3-mm × 3-mm area centered on the fovea. Capillary vessel density was obtained for superficial (SCP), and deep (DCP) retinal capillary plexuses using projection artifact-resolved automated software (Optovue, version 151).

Statistical analysis

Statistical analyses were performed using SAS/STAT, version 9.4, and R/RStudio, version 3.51 (R-project). Descriptive statistics are presented as mean (± SD), median (interquartile range), or frequencies, as appropriate. Logistic regression models were applied to test the cross-sectional associations between individual cognitive predictors and vascular complications, including CVD, DN, PDR, and peripheral neuropathy as dependent variables. To test the cross-sectional relationships between ophthalmologic outcomes (OCT/OCTA parameters) and cognitive predictors, as well as between these outcomes and vascular complications, generalized estimating equations with an unstructured correlation matrix were employed. Because each individual had 2 unique values per eye, this method was adjusted for within-subject correlations. When studying the association between cognitive tasks and OCT/OCTA parameters, psychomotor speed and executive function were both adjusted for visual acuity because they are visually based assessments. To control for confounding factors, models were adjusted for relevant clinical characteristics based on a priori knowledge as described in the legends of each table or figure, including, demographic characteristics, glycemic control and metabolic markers, exercise, smoking, and markers of β-cell function. In this exploratory analysis, the α significance level was set at P less than .05.

Results

Clinical characteristics

The clinical characteristics of Medalists who had OCT, OCTA, and cognitive testing vs the overall cohort (n = 1019) were comparable (Supplementary Table 1 [23]). All supplementary material and figures are located in a digital research repository (23). Mean ± SD scaled scores of immediate, delayed memory, and working memory were 53.4 ± 13.5, 51.3 ± 12.8, and 10.7 ± 3.0, respectively; psychomotor speed in nondominant and dominant hands were 150.0 ± 85.6 seconds and 121.0 ± 51.0 seconds, respectively; and executive function scaled score was 10.7 ± 2.7 (Supplementary Table 2 [23]). These cognitive scores are similar to age-matched people with type 2 diabetes and worse than age-matched controls without diabetes, as previously reported (21). Metabolic parameters in the cognitive subset of the Medalists showed a mean HbA1c of 7.0 ± 0.8% (53 ± 8.6 mmol/mol), low-density lipoprotein of 80.2 ± 24.4 mg/dL, high-density lipoprotein of 71.2 ± 21.0 mg/dL, and triglycerides of 72.5 ± 32.1 mg/dL, indicating excellent glycemic and lipid metabolic regulation similar to the overall Medalist cohort (24). Several clinical factors, including age, diabetes duration, triglycerides, and hypertension were associated with cognitive dysfunction (Supplementary Table 3 [23]). Cognitive dysfunctions were associated with the presence of PDR (psychomotor speed, P = .01) and CVD (immediate memory, P = .03), even in a multivariate analysis adjusted for potential confounders such as age, triglycerides, hypertension, body mass index, and HbA1c (Supplementary Table 4 and Supplementary Figure 1 [23).

Optical coherence tomography angiography findings

OCTA, which visualizes and quantitates SCP and DCP capillary density in the retinal microvasculature, was used to evaluate the associations between complication status and retinal small vessel abnormalities in the Medalists. For cognitive dysfunction, decreased vascular density of the SCP was associated with decreased psychomotor speed (point estimate [PE], –0.03; 95% CI, –0.06 to –0.01, P = .01) even after adjusting for DR severity, diabetes duration, and visual acuity (Fig. 1A). Decreased vessel density of the DCP was strongly associated with worse delayed memory (PE, 0.08; 95% CI, 0.03 to 0.14, P = .002) and showed a nonsignificant trend with worse immediate memory (PE, 0.07; 95% CI, –0.01 to 0.14, P = .06) even after adjusting for diabetes duration and DR severity (Fig. 1B). No relationship was found between the central retinal microvasculature visualized by OCTA and executive function or working memory when adjusting for potential confounders. Fig. 2A shows representative images of the SCP and DCP of the retinal microvasculature in Medalists with no to mild DR vs those with PDR. Decreased SCP vascular density was strongly associated with the presence of PDR (PE, –6.03; 95% CI, –8.60 to –3.47, P < .0001), but not diabetic peripheral neuropathy, CVD, or DN, either on a bivariate or multivariate level (Fig. 2B). In contrast, decreased DCP vascular density was associated with the presence of peripheral neuropathy (PE, –2.37; 95% CI, –4.26 to –0.48, P = .01), PDR (PE, –3.94; 95% CI, –6.12 to –1.76, P = .004), and CVD (PE, –2.28, 95% CI, –4.40 to –0.16, P = .04, Fig. 2C). These associations were relatively independent of each other as the presence of peripheral neuropathy (PE, –2.42; 95% CI, –4.16 to –0.69, P = .006), and PDR (PE, –3.10; 95% CI, –5.42 to –0.77, P = .009) remained associated with decreased DCP vascular density after adjustment for other vascular complications. No significant relationship was found between severity of DN and DCP in bivariate or multivariate analysis (Supplementary Table 6, (23)).

Figure 1.

Figure 1.

Open in a new tab

Forest plots demonstrating the relationship between cognitive tasks and vascular density of the A, superficial, and B, deep retinal capillary plexus. β Estimates from a mixed linear-effects model. All models are adjusted for proliferative diabetic retinopathy status and type 1 diabetes duration. Both psychomotor speed and executive function are also adjusted for visual acuity because they are visually based assessments.

Figure 2.

Figure 2.

Open in a new tab

A, Optical coherence tomography angiography images of the superficial and deep retinal capillary plexus and the presence of proliferative diabetic retinopathy. B and C, Forest plots demonstrating β estimates and 95% CIs for the association between diabetic complications in the 50-Year Medalist cohort and vascular density of the B, superficial, and C, deep retinal capillary plexus adjusting for type 1 diabetes duration. CVD, cardiovascular disease; PDR, proliferative diabetic retinopathy.

Optical coherence tomography findings

OCT technology allows for the quantification of structural changes in individual retinal layers throughout the macula, which contains the highest density of photoreceptors in the retina and is critical for central visual acuity (Fig. 3A). Thinning of the ONL of the neural retina, which contains photoreceptor nuclei, was associated with lower performance across multiple cognitive tests including psychomotor speed both in nondominant and dominant hands (PE, –0.04; 95% CI, –0.07 to –0.01, P = .01 and PE, –0.08; 95% CI, –0.16 to 0.01, P = .05, respectively), immediate memory (PE, 0.72; 95% CI, –0.02 to 1.45, P = .05), and showed a nonsignificant trend with delayed memory (PE, –0.67; 95% CI, –1.45 to 0.11, P = .09), when adjusting for potential confounders of OCT retinal thickness including sex, diabetes duration, visual acuity, and DR severity (Fig. 3B). Thinning of the OPL was associated with delayed memory score (PE, –0.45; 95% CI, –0.87 to –0.03, P = .04). No significant relationship was found between retinal layer thickness and working memory (Supplementary Table 5 [23]). Associations were also observed between retinal layer thickness and other diabetic complications, including PDR, CVD, DN, and peripheral neuropathy. As expected, the presence of PDR was significantly associated with increased thickness of primarily inner retinal layers, including the RNFL (PE, 3.74; 95% CI, 1.04 to 6.43, P = .007), GCL (PE, 2.17; 95% CI, 0.55 to 3.79, P = .009), INL (PE, 4.04; 95% CI, 2.11 to 5.97, P < .0001), and OPL (PE, 1.78; 95% CI, 0.52 to 3.05, P = .006) (Fig. 3D). PDR was significantly associated with decreased thickness of the ONL (PE, –21.70; 95% CI, –40.06 to –3.34, P = .02), PHOTO (PE, –5.06; 95% CI, –6.26 to –3.86, P < .0001), and RPE (PE, –0.80; 95% CI, –1.60 to –0.01, P = .047). Even after adjusting for focal or scatter laser treatment, which in general are not targeting the macula, the presence of PDR remained associated with increased thickness of the RNFL (P = .02), GCL (P = .04), INL (P = .001), and decreased ONL (P = .03) and PHOTO (P < .0001). Similarly, the presence of PDR remained significantly associated with increased INL thickness (P = .02) and decreased PHOTO thickness (P = .01) after adjusting for scatter laser treatment. Thinning of the ONL was associated with the presence of CVD (PE, –4.55; 95% CI, –8.32 to –0.77, P = .02, Fig. 3C). DN was associated with a decreased thickness of the PHOTO(Fig. 3E, PE, –12.01; 95% CI, –23.97 to –0.23, P = .046). Thinning of the PHOTO was also associated with the presence of peripheral diabetic neuropathy (PE, –17.85; 95% CI, –34.69 to –1.02, P = .04). Even after adjustment for peripheral neuropathy or the presence of CVD, the associations between PDR and various retinal layer thicknesses remained significant, suggesting that the observed associations were relatively independent of each other and history of scattered laser treatment.

Figure 3.

Figure 3.

Open in a new tab

A, Retinal layer segmentation showing delineation of the outer nuclear layer using optical coherence tomography. B, Forest plot demonstrating beta estimate and 95% Cis for the association between outer nuclear layer thickness and cognition adjusting for sex, type 1 diabetes duration, and diabetic retinopathy severity. Both psychomotor speed and executive function are adjusted for visual acuity because they are visually based assessments. C to F, Forest plots demonstrating the association between retinal layer thicknesses and C, cardiovascular disease, D, proliferative diabetic retinopathy, E, diabetic nephropathy, or F, diabetic neuropathy, adjusting for sex and type 1 diabetes duration. D, dominant hand; GCL, ganglion cell layer; RNFL: retinal nerve fiber layer; INL, inner nuclear layer; IPL, inner plexiform layer; ND, nondominant hand; ONL, outer nuclear layer; OPL, outer plexiform layer; PHOTO, photoreceptor layer; RPE, retinal pigment epithelium.

Combined effects of optical coherence tomography angiography and optical coherence tomography parameters

We evaluated the combined effects of OCTA and OCT parameters on diabetic complications, including cognitive dysfunction. As expected, 100% of people in the lowest combined tertiles of vascular density of the SCP (< 36%) and ONL thickness (< 80 µm) had PDR, compared to less than 12.5% in the highest SCP tertile (> 42%, Fig. 4A). The combined lowest tertiles of vascular density of the DCP (< 43%) and ONL thickness (< 80 µm) exhibited a similar relationship with the presence of PDR (Fig. 4B). For cognitive dysfunction, psychomotor speed was decreased by more than 20% in the lowest combined tertile of vascular density of SCP (< 36%) and ONL thickness (< 80 µm) compared to the highest tertile of vascular density of the SCP (> 42%) and ONL thickness (> 94 µm, Fig. 4C). Similarly, delayed memory score was decreased by more than 30% in the lowest combined tertiles of vascular density of the DCP (< 43%) and OPL thickness (< 27 µm) compared to the highest tertile of DCP of the retinal microvasculature (> 47%, Fig. 4D).

Figure 4.

Figure 4.

Open in a new tab

A to D, Three-dimensional graphs showing the combined effect of optical coherence tomography and optical coherence tomography angiography on diabetic complications or cognition. A, Proportion of proliferative diabetic retinopathy by tertiles of retinal outer nuclear layer thickness and vascular density of the superficial capillary plexus. B, Proportion of proliferative diabetic retinopathy by tertiles of retinal outer nuclear layer thickness and vascular density of the deep capillary plexus. C, Mean psychomotor speed by tertiles of retinal outer nuclear layer thickness and vascular density of the superficial capillary plexus. D, Mean delayed memory scores by tertiles of retinal outer plexiform layer thickness and vascular density of the deep capillary plexus. D, dominant hand; DCP, deep capillary plexus; ND, nondominant hand; ONL, outer nuclear layer; OPL, outer plexiform layer; SCP, superficial capillary plexus.

Discussion

Abnormalities in retinal neural structure and capillary perfusion evaluated by OCT and OCTA were found to be strongly associated with specific cognitive dysfunctions in the Medalists, independent of retinopathy severity level. These findings suggest that noninvasive retinal imaging using OCT and OCTA may provide readily accessible, reproducible, and noninvasive surrogate markers to evaluate cognitive dysfunction in people with type 1 diabetes. Furthermore, these markers may be relevant across the spectrum of retinopathy severity. These findings have also confirmed an association between cognitive dysfunction, specifically delayed psychomotor speed and working memory, and the presence of PDR and history of CVD, respectively (3). This study provides the first comprehensive analysis of the relationship between retinal structure and capillary density with cognitive function and several major complications in a large group of people with long-duration type 1 diabetes. Results from the OCTA studies showed that decreases in SCP correlated with worse psychomotor speed and decreases in DCP with deficits in immediate and delayed memory, supporting the idea that the retina and CNS may exhibit similar capillary abnormalities due to the metabolic changes of diabetes. Importantly, although the presence of PDR was associated with decreases both in SCP and DCP, the associations with cognitive dysfunction remained significant even after adjustment for retinopathy severity.

The correlation of decreased DCP with worse immediate and delayed memory is consistent with the finding by OCT that ONL thinning was associated with worsening of delayed memory. These results suggest that hypoperfusion of the ONL and photoreceptor regions might be related to common factors of diabetes that are causing abnormal blood perfusion to the hippocampal area of the CNS, which has been correlated to delayed memory and where abnormal perfusion has been reported to be associated with memory deficit in type 2 diabetes (25, 26). Interestingly, ONL thinning correlated with psychomotor speed and immediate and delayed memory, which are the cognitive parameters that also strongly associated with CVD and PDR, suggests that common factors due to diabetes may be contributing to the degeneration of ONL in PDR, cognitive dysfunction and CVD. The OCT studies confirm previous reports of the association of thinning of ONL and photoreceptor layers but a widening of the inner layers of the retina with PDR (26). Although panretinal photocoagulation (PRP) can affect the thickness of the peripheral retinal, most reports indicate that PRP does not substantially affect central (fovea) retinal thickness even over the long term (27). Further, adjusting for the presence of PRP did not significantly change the effect of the associations between PDR and retinal structural changes in this study. The thinning of ONL and PHOTO in PDR has been suggested to be related to retinal dysfunction or degeneration, whereas the widening of inner retinal layers is potentially due to subclinical edema or preretinal or intraretinal fibrosis (28). Since DR can be viewed as a disease of the neurovascular unit that refers to the coupling and integration of neurons and vasculature to regulate normal retinal function (28-30), these findings highlight the need for a greater understanding of the common factors and interactions between neurons, glia, and vasculature in diabetes contributing to the development of retinal and brain neurodegeneration, and cognitive impairment.

Changes in multiple metabolic and hormonal factors induced by diabetes likely contribute to the abnormalities in the retina, CVD, and the CNS. Because cognitive dysfunction is associated both with PDR and CVD, abnormalities of the large vessels and capillaries are likely involved in cognitive deficits related to aging people with a chronic duration of type 1 diabetes. Hyperglycemia is an important risk factor since it is the major risk factor both for PDR and CVD in people with type 1 diabetes. Lacy et al have reported that glycemic control correlated with elevated risks of diagnosis of cognitive decline in people with type 1 diabetes of long duration (31). However, lack of associations between cognitive dysfunction and nephropathy or peripheral neuropathy, for which hyperglycemia is a known risk factor, suggests that the mechanisms of hyperglycemia’s adverse effects on the CNS and retina could be different from those in peripheral tissues. Loss of insulin action due to insulin resistance or deficiency may also be an important common risk factor for cognitive dysfunction, PDR, and CVD. Multiple clinical and basic reports have established insulin as having important functions in the CNS, such as the regulation of appetite and cholesterol metabolism (32-34). Insulin resistance is an established risk factor for Alzheimer disease in type 2 diabetes but its relationship to cognitive dysfunction in type 1 diabetes is not well established. Similarly, insulin has important effects on vasculature, such as the regulation of blood flow via endothelial nitric oxide synthase activation and capillary permeability, both of which are abnormal in CVD, PDR, and cognitive diseases (35-38). Insulin may also be important for neural maintenance and survival. Mice with global deletion of insulin receptor substrate 2, a critical signaling protein for insulin, exhibited gradual retinal degeneration and vessel loss with parallel degenerations of the CNS, even when development of diabetes was prevented (39, 40). Loss of insulin action in the retina has also been proposed as a contributory factor to DR (25). Our study did not evaluate the role of insulin resistance in the Medalist cohort specifically. However, the body mass index of the Medalists, who mostly have type 1 diabetes, was not elevated and their daily insulin dose was not exceptionally high, suggesting the absence of severe insulin resistance in the Medalists (4, 24). Similarly, the frequency of hypoglycemia has been associated with cognitive dysfunction in people with type 1 diabetes (5, 26, 31). Further studies on the effect of hypoglycemia and daily glucose variance in the Medalist cohort are now ongoing to address these questions.

There are multiple limitations to this study. Because this is not a longitudinal study, it prevents conclusions concerning the predictive value of ocular imaging techniques for cognitive dysfunctions. In addition, we recognize that statistical significance is not equivalent to clinical relevance and that the cognitive testing performed in this study cannot characterize every aspect of cognitive function critical to daily life. Additionally, no adjustments for multiple testing were carried out for the following reasons: 1) This is an exploratory study testing one overarching hypothesis—that there is a relationship between cognitive function and retinal markers, not a confirmatory analysis of multiple hypotheses, for which one would require adjustment for possible α-level inflations. 2) Even if we do adjust for multiple comparisons, it is debatable how many actual independent tests there are given correlations between various cognitive parameters among themselves, or correlations among retinal markers or vascular outcomes. Using conventional Bonferroni or false discovery rate thresholds might result in missing important effects that do exist. Clearly, confirmatory investigations will be need to adjust for multiple comparisons. It is of note that only one of the 129 Medalists with completed cognitive testing presented with Alzheimer disease, which may suggest the presence of some protection in this group of selected participants. The Medalist group as a whole must have protective characteristics and factors that allowed a substantial proportion of them to be protected from advanced diabetic complications, including retinopathy (3). Given the association between cognitive dysfunction and retinopathy and the shared origin of the CNS and retina, it is possible that a shared mechanism, which can alter the progression of early stages of retinal and brain neurodegeneration, may provide protection against the development of PDR and Alzheimer disease in people with diabetes. However, we acknowledge the small sample size as a limitation in this study, but it is one of the very few studies examining the relationships between cognitive function and retinal imaging markers in aging people with type 1 diabetes. Future studies with larger sample sizes in aging populations with shorter duration of type 1 diabetes will be needed to validate the findings in this report and assess the potential of OCT and OCTA to function as a clinical tool for cognitive impairment. Finally, we did not assess the relationship between cognitive impairment and hypoglycemia and functional studies of retinal neurodysfunction. Thus, despite the listed shortcomings, and given that the Joslin Medalist Study is well suited to identify potential markers for cognitive dysfunction due to aging in people with type 1 diabetes, we view this study as a way to gain insight into the utility of noninvasive ocular imaging in this population to help design a larger study to fill in the gaps.

In summary, this is the first study to evaluate retinal neural and vascular structure in aging people with type 1 diabetes. Our findings suggest that noninvasive retinal imaging using OCT and OCTA may provide surrogate markers to evaluate cognitive dysfunction in people with type 1 diabetes. This is important because cognitive dysfunction is a growing and understudied public health issue in the aging population with type 1 diabetes. Future studies may provide information on the utility of repeated noninvasive ocular imaging as a screening tool for cognitive dysfunction in people with diabetes, and insights into pathophysiological relationships between cognitive dysfunction and neural and microvascular pathology of the CNS and retina in diabetes.

Acknowledgments

This study would not have been possible without the staff of the Joslin Clinical Research Center and the 50-Year Medalists and their families. The authors thank the Beetham Eye Institute (Joslin Diabetes Center) staff for ocular evaluation of the Medalist patients.

Financial Support: This work was supported by the Dianne Nunnally Hoppes Fund; the Beatson Pledge Fund; the National Institute of Diabetes and Digestive and Kidney Diseases, and the National Institutes of Health (grant Nos. P30-DK-036836, UL1-RR-025758-03, R24-DK-083957-01, 1DP3DK112192, and DP3-DK-094333-01); the Juvenile Diabetes Research Foundation (grant No. 17-2013-310); and the American Diabetes Association (grant No. 9-18-CVD1-005 to M.G.Y.).

Glossary

Abbreviations

CNS

central nervous system

CVD

cardiovascular disease

DCP

deep retinal capillary plexus

DN

diabetic nephropathy

ETDRS

Early Treatment of Diabetic Retinopathy Study

GCL

ganglion cell layer

HbA1c

glycated hemoglobin A1c

INL

inner nuclear layer

IPL

inner plexiform layer

OCT

optical coherence tomography

OCTA

optical coherence tomography angiography

ONL

outer nuclear layer

OPL

outer plexiform layer

PDR

proliferative diabetic retinopathy

PE

point estimate

PHOTO

photoreceptor layer

PRP

panretinal photocoagulation

RNFL

retinal nerve fiber layer

RPE

retinal pigment epithelium

SCP

superficial retinal capillary plexus

Additional Information

Disclosures: The authors have nothing to disclose.

Data Availability

The data sets generated during and/or analyzed during the present study are available from the corresponding author on reasonable request.

References

  • 1. Cho NH, Shaw JE, Karuranga S, et al. IDF Diabetes Atlas: global estimates of diabetes prevalence for 2017 and projections for 2045. Diabetes Res Clin Pract. 2018;138:271-281. [DOI] [PubMed] [Google Scholar]
  • 2. Keenan HA, Costacou T, Sun JK, et al. Clinical factors associated with resistance to microvascular complications in diabetic patients of extreme disease duration: the 50-year Medalist Study. Diabetes Care. 2007;30(8):1995-1997. [DOI] [PubMed] [Google Scholar]
  • 3. Sun JK, Keenan HA, Cavallerano JD, et al. Protection from retinopathy and other complications in patients with type 1 diabetes of extreme duration: the Joslin 50-year Medalist Study. Diabetes Care. 2011;34(4):968-974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Tinsley LJ, Kupelian V, D’Eon SA, et al. Association of glycemic control with reduced risk for large-vessel disease after more than 50 years of type 1 diabetes. J Clin Endocrinol Metab. 2017;102(10):3704-3711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Whitmer RA, Biessels GJ, Quesenberry CP Jr, Liu JY, Karter AJ, Beeri M. Type 1 diabetes and risk of dementia in late life: the Kaiser Diabetes & Cognitive Aging Study. Alzheimers Dement. 2015;11(7S Pt 4):P179-P180. [Google Scholar]
  • 6. Ding J, Strachan MW, Reynolds RM, et al. ; Edinburgh Type 2 Diabetes Study Investigators . Diabetic retinopathy and cognitive decline in older people with type 2 diabetes: the Edinburgh Type 2 Diabetes Study. Diabetes. 2010;59(11):2883-2889. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Cogan DG, Kuwabara T. Comparison of retinal and cerebral vasculature in trypsin digest preparations. Br J Ophthalmol. 1984;68(1):10-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Simó R, Ciudin A, Simó-Servat O, Hernández C. Cognitive impairment and dementia: a new emerging complication of type 2 diabetes—the diabetologist’s perspective. Acta Diabetol. 2017;54(5):417-424. [DOI] [PubMed] [Google Scholar]
  • 9. Sundstrom JM, Hernández C, Weber SR, et al. Proteomic analysis of early diabetic retinopathy reveals mediators of neurodegenerative brain diseases. Invest Ophthalmol Vis Sci. 2018;59(6):2264-2274. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Coppola G, Di Renzo A, Ziccardi L, et al. Optical coherence tomography in Alzheimer’s disease: a meta-analysis. PloS One. 2015;10(8):e0134750. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Mutlu U, Colijn JM, Ikram MA, et al. Association of retinal neurodegeneration on optical coherence tomography with dementia: a population-based study. JAMA Neurol. 2018;75(10):1256-1263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Ko F, Muthy ZA, Gallacher J, et al. ; UK Biobank Eye & Vision Consortium . Association of retinal nerve fiber layer thinning with current and future cognitive decline: a study using optical coherence tomography. JAMA Neurol. 2018;75(10):1198-1205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Bulut M, Kurtuluş F, Gözkaya O, et al. Evaluation of optical coherence tomography angiographic findings in Alzheimer’s type dementia. Br J Ophthalmol. 2018;102(2):233-237. [DOI] [PubMed] [Google Scholar]
  • 14. O’Bryhim BE, Apte RS, Kung N, Coble D, Van Stavern GP. Association of preclinical Alzheimer disease with optical coherence tomographic angiography findings. JAMA Ophthalmol. 2018;136(11):1242-1248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Querques G, Borrelli E, Sacconi R, et al. Functional and morphological changes of the retinal vessels in Alzheimer’s disease and mild cognitive impairment. Sci Rep. 2019;9(1):63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Zhang YS, Zhou N, Knoll BM, et al. Parafoveal vessel loss and correlation between peripapillary vessel density and cognitive performance in amnestic mild cognitive impairment and early Alzheimer’s disease on optical coherence tomography angiography. PLoS One. 2019;14(4):e0214685. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Jiang H, Wei Y, Shi Y, et al. Altered macular microvasculature in mild cognitive impairment and Alzheimer disease. J Neuroophthalmol. 2018;38(3):292-298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Levey AS, Stevens LA, Schmid CH, et al. ; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) . A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604-612. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Nielsen LR, Rehfeld JF, Pedersen-Bjergaard U, Damm P, Mathiesen ER. Pregnancy-induced rise in serum C-peptide concentrations in women with type 1 diabetes. Diabetes Care. 2009;32(6):1052-1057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Yu L, Rewers M, Gianani R, et al. Antiislet autoantibodies usually develop sequentially rather than simultaneously. J Clin Endocrinol Metab. 1996;81(12):4264-4267. [DOI] [PubMed] [Google Scholar]
  • 21. Musen G, Tinsley LJ, Marcinkowski KA, et al. Cognitive function deficits associated with long-duration type 1 diabetes and vascular complications. Diabetes Care. 2018;41(8):1749-1756. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Strauss E, Sherman EMS, Spreen O.. A Compendium of Neuropsychological Tests: Administration, Norms, and Commentary. 3rd ed. Oxford University Press; 2006. [Google Scholar]
  • 23. Fickweiler W, Wolfson EA, Paniagua SM, et al. Supplementary data for: “Association of cognitive function and retinal neural and vascular structure in type 1 diabetes.” Harvard Dataverse. 2020. Deposited November 24, 2020. 10.7910/DVN/T8JB6R [DOI] [Google Scholar]
  • 24. Yu MG, Keenan HA, Shah HS, et al. Residual β cell function and monogenic variants in long-duration type 1 diabetes patients. J Clin Invest. 2019;129(8):3252-3263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Arnold SE, Arvanitakis Z, Macauley-Rambach SL, et al. Brain insulin resistance in type 2 diabetes and Alzheimer disease: concepts and conundrums. Nat Rev Neurol. 2018;14(3):168-181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Starr JM, Wardlaw J, Ferguson K, MacLullich A, Deary IJ, Marshall I. Increased blood-brain barrier permeability in type II diabetes demonstrated by gadolinium magnetic resonance imaging. J Neurol Neurosurg Psychiatry. 2003;74(1):70-76. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Boynton GE, Stem MS, Kwark L, Jackson GR, Farsiu S, Gardner TW. Multimodal characterization of proliferative diabetic retinopathy reveals alterations in outer retinal function and structure. Ophthalmology. 2015;122(5):957-967. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Duh EJ, Sun JK, Stitt AW. Diabetic retinopathy: current understanding, mechanisms, and treatment strategies. JCI Insight. 2017;2(14):e93751. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Simó R, Hernández C; European Consortium for the Early Treatment of Diabetic Retinopathy (EUROCONDOR) . Neurodegeneration in the diabetic eye: new insights and therapeutic perspectives. Trends Endocrinol Metab. 2014;25(1):23-33. [DOI] [PubMed] [Google Scholar]
  • 30. Simó R, Stitt AW, Gardner TW. Neurodegeneration in diabetic retinopathy: does it really matter? Diabetologia. 2018;61(9):1902-1912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Lacy ME, Gilsanz P, Eng C, Beeri MS, Karter AJ, Whitmer RA. Severe hypoglycemia and cognitive function in older adults with type 1 diabetes: the Study of Longevity in Diabetes (SOLID). Diabetes Care. 2020;43(3):541-548. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Neth BJ, Craft S. Insulin resistance and Alzheimer’s disease: bioenergetic linkages. Front Aging Neurosci. 2017;9:345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Suzuki R, Lee K, Jing E, et al. Diabetes and insulin in regulation of brain cholesterol metabolism. Cell Metab. 2010;12(6):567-579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Kleinridders A, Ferris HA, Cai W, Kahn CR. Insulin action in brain regulates systemic metabolism and brain function. Diabetes. 2014;63(7):2232-2243. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Rask-Madsen C, King GL. Vascular complications of diabetes: mechanisms of injury and protective factors. Cell Metab. 2013;17(1):20-33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. King GL, Park K, Li Q. Selective insulin resistance and the development of cardiovascular diseases in diabetes: the 2015 Edwin Bierman Award Lecture. Diabetes. 2016;65(6):1462-1471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Park K, Mima A, Li Q, et al. Insulin decreases atherosclerosis by inducing endothelin receptor B expression. JCI Insight. 2016;1(6):e86574. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Li Q, Park K, Li C, et al. Induction of vascular insulin resistance and endothelin-1 expression and acceleration of atherosclerosis by the overexpression of protein kinase C-β isoform in the endothelium. Circ Res. 2013;113(4):418-427. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Schubert M, Brazil DP, Burks DJ, et al. Insulin receptor substrate-2 deficiency impairs brain growth and promotes tau phosphorylation. J Neurosci. 2003;23(18):7084-7092. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Yi X, Schubert M, Peachey NS, et al. Insulin receptor substrate 2 is essential for maturation and survival of photoreceptor cells. J Neurosci. 2005;25(5):1240-1248. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data sets generated during and/or analyzed during the present study are available from the corresponding author on reasonable request.

Articles from The Journal of Clinical Endocrinology and Metabolism are provided here courtesy of The Endocrine Society

RESOURCES