Abstract
OBJECTIVE
To evaluate associations between plasma biomarkers of brain injury and MRI and cognitive measures in participants with type 1 diabetes (T1D) from the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) study.
RESEARCH DESIGN AND METHODS
Plasma amyloid-β-40, amyloid-β-42, neurofilament light chain (NfL), phosphorylated Tau-181 (pTau-181), and glial fibrillary acidic protein (GFAP) were measured in 373 adults who participated in the DCCT/EDIC study. MRI assessments included total brain and white matter hyperintensity volumes, white matter mean fractional anisotropy, and indices of Alzheimer disease (AD)–like atrophy and predicted brain age. Cognitive measures included memory and psychomotor and mental efficiency tests and assessments of cognitive impairment.
RESULTS
Participants were 60 (range 44–74) years old with 38 (30–51) years’ T1D duration. Higher NfL was associated with an increase in predicted brain age (0.51 years per 20% increase in NfL; P < 0.001) and a 19.5% increase in the odds of impaired cognition (P < 0.01). Higher NfL and pTau-181 were associated with lower psychomotor and mental efficiency (P < 0.001) but not poorer memory. Amyloid-β measures were not associated with study measures. A 1% increase in mean HbA1c was associated with a 14.6% higher NfL and 12.8% higher pTau-181 (P < 0.0001).
CONCLUSIONS
In this aging T1D cohort, biomarkers of brain injury did not demonstrate an AD-like profile. NfL emerged as a biomarker of interest in T1D because of its association with higher HbA1c, accelerated brain aging on MRI, and cognitive dysfunction. Our study suggests that early neurodegeneration in adults with T1D is likely due to non-AD/nonamyloid mechanisms.
Graphical Abstract
Introduction
Cognitive decline and dementia are established complications of type 1 diabetes (T1D), but the mechanisms and features remain poorly characterized (1–3). Although long-term dysglycemia and diabetes-related complications increase risk for cognitive decline in individuals with T1D (1), more research is needed to understand the pathways involved and establish whether T1D enhances risk for other aging-related causes of neurodegeneration such as Alzheimer disease (AD) (4). Currently, there is significantly more research that examines cognitive impairment associated with type 2 diabetes (T2D) relative to T1D, especially among older individuals, with studies demonstrating an increased risk of AD, vascular dementia, and mild cognitive impairment in patients with T2D (5). Plasma biomarkers of brain injury provide insight into mechanisms underlying cognitive dysfunction, including amyloid burden, neuroinflammation, and neurodegeneration in AD and other neurodegenerative diseases (6). A few studies in small T1D cohorts suggest associations between these plasma biomarkers and brain MRI abnormalities and/or cognitive dysfunction (7,8). However, the relationships between these biomarkers and cognitive impairment in T1D remain largely unknown.
The Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) study demonstrated that older adults with T1D have worse cognitive function, more brain volume loss, and increased vascular injury compared with similarly aged control participants without diabetes, and these factors are equivalent to 4–9 years of accelerated brain aging (1,9). Further, the pattern of atrophy in the DCCT/EDIC cohort was not typical of AD-related neurodegeneration (10). Because neuropathologic changes of AD may precede MRI-derived structural changes, we measured select plasma biomarkers in this cohort to determine whether profiles are consistent with AD pathophysiology prior to detectable MRI changes (11) or are more consistent with another etiology of neurodegeneration, thereby corroborating our prior findings that T1D predisposes to accelerated brain aging but not AD.
Plasma amyloid-β-40, amyloid-β-42, neurofilament light chain (NfL), glial fibrillary acidic protein (GFAP), and phosphorylated Tau-181 (pTau-181) are biomarkers associated with various pathways of brain injury; when interpreted as a panel along with MRI measures, biomarker profiles suggest the presence of AD or other neurodegenerative processes (6,12–19). Specifically, higher pTau-181 and a lower ratio of amyloid-β-42 to amyloid-β-40 (termed the amyloid-β-42/40 ratio) are markers of the amyloid deposition of AD, the first measurable brain change of AD (6,12,13,15). Higher NfL and GFAP are nonspecific markers of neurodegeneration and neuroinflammation, respectively (18,19). Brain atrophy measures, such as total brain volume (TBV) or AD signature region atrophy, provide additional measures of neurodegeneration, while white matter hyperintensity (WMH) volume is a biomarker of cerebrovascular disease. Typically, the presence of amyloid deposition (higher pTau-181 and lower amyloid-β-42/40 ratio) together with neurodegenerative biomarkers (higher NfL and brain atrophy, particularly in AD signature regions) suggests AD-related neurodegeneration and may influence cognitive symptoms, while higher NfL and brain atrophy without amyloid deposition suggest cognitive symptoms are due to non-AD/nonamyloid neurodegeneration (11,19,20). Positive amyloid biomarkers in the absence of neurodegeneration indicate early AD pathology but without downstream injury likely to result in cognitive change.
With extensive longitudinal data and standardized assessments of biomedical and cognitive outcomes, the DCCT/EDIC study provides an unparalleled opportunity to evaluate the association between plasma biomarkers and cognitive outcomes in a large T1D cohort. We hypothesized that plasma amyloid-β-42, amyloid-β-42/40 ratio, pTau-181, NfL, and GFAP would be associated with adverse MRI measures of neurodegeneration, including decreased TBV, increased white matter injury, advanced brain aging, lower cognitive function in the domains of memory and psychomotor and mental efficiency, and greater cognitive impairment based on consensus ratings.
Research Design and Methods
The DCCT/EDIC study was described previously (21,22). Briefly, between 1983 and 1989, the DCCT randomly assigned 1,441 participants with T1D (age range 13–29 years) but without history of cardiovascular disease, hypertension, hyperlipidemia, or neuropathy requiring medical intervention to intensive or conventional diabetes therapy to evaluate the effects of glycemia on the development and progression of diabetes-related complications (21). After a mean (range) of 6.5 (3–9) years of follow-up, intensive therapy markedly reduced the development and progression of complications (21); all participants were encouraged to adopt intensive therapy and transitioned to their health care practitioners for diabetes care. In 1994, 96% of the surviving DCCT cohort enrolled in the EDIC observational study (22).
In 2018–2019 (average follow-up 32 years), 425 of 1,190 active EDIC participants were randomly selected and invited to enroll in an MRI study (9). Exclusion criteria included end-stage renal disease; visual acuity worse than 20/40 corrected in both eyes; severe claustrophobia; pacemakers, implanted neurostimulators, or other known or suspected retained metallic foreign objects; or body weight >350 lb. The present analyses included 373 participants from the MRI study with complete plasma biomarker and concurrent cognitive measures. The characteristics of this subgroup were similar to those of the overall surviving cohort at the time of acquisition (9). The study was approved by institutional review boards at all institutions, and all participants provided written informed consent.
MRI Protocol
MRI studies were performed at 24 imaging centers (26 of 27 EDIC sites) using Siemens, Philips, and GE 3 Tesla scanners. The imaging protocol, quality control, and analysis methods were previously described (9). Briefly, MRI analysis was performed by readers masked to other participant data using semiautomated computational pipelines. TBV, a sum of all brain parenchymal volume, including cerebrum, cerebellum, and brainstem, was derived using a multiatlas, multiwarp label-fusion segmentation method (23) using T1 scans. A validated deep learning–based segmentation method was used to measure WMH volume (n = 339) (24). Total white matter mean fractional anisotropy (FA) was extracted from FA maps after registration with T1 (n = 323). Using a two-step harmonization, we derived two SPARE (spatial pattern for recognition) indices to measure AD-like atrophy (SPARE-AD) and predicted brain age (SPARE-BA) (10). Higher SPARE-AD values indicate more AD-like atrophy, and higher SPARE-BA values relative to chronologic age indicate an atrophy pattern of accelerated brain aging.
Cognitive Testing and Consensus Rating
EDIC cognitive test administration, scoring, and quality control procedures were previously described (1,25). Cognitive assessments included an abbreviated battery with psychomotor and mental efficiency tests sensitive to diabetes (25) and tests of memory sensitive to mild cognitive impairment/AD (26). A standardized z score was calculated for each test variable using the means and SD of the DCCT/EDIC cohort from the DCCT baseline evaluation (1983–1989). z scores in each domain were averaged to obtain a summary score. These standardized scores provide a unit-free measurement of the relative difference in performance compared with the total DCCT/EDIC cohort at the referent DCCT baseline assessment.
Cognition ratings were conducted by three senior neuropsychologists who independently classified each participant’s cognitive status based on review of available cognitive data but were masked to participant characteristics and biomedical data. A final cognitive rating determination was made by consensus. Additional details were previously reported and are presented in the Supplementary Material (27). While strict algorithms were not used to determine classification, the following guiding principles were used. Normal profiles were those in which current raw test scores fell within expected norms, without significant decline over time. Impaired profiles were those in which current test scores were lower on multiple tests (>1.5 to 2 SD below the cohort baseline means), with significant decline over time. Questionable profiles were those in which current test scores fell within expected norms, with significant decline over time, or were lower on multiple tests, without significant decline over time. With only 12 participants classified as impaired and 54 as questionable, these groups were combined for analyses.
Capillary blood glucose levels were measured immediately prior to cognitive testing and MRI to ensure absence of acute hypoglycemia. Participants with blood glucose levels ≤70 mg/dL received a snack; scanning and testing began when blood glucose values were ≥90 mg/dL and any symptoms had resolved.
Plasma Biomarkers of Brain Injury
Five biomarkers were evaluated by single-batch analysis of stored plasma samples coinciding most closely with the MRI/cognitive study visit (2018–2019). Specifically, amyloid-β-40, amyloid-β-42, NfL, and GFAP were measured in a multiplex assay (Neurology 4-Plex E kit) and pTau-181 was measured in a singleplex assay, using single-molecule array immunoassays (Simoa) on the Quanterix HD-X analyzer (Billerica, MA), from one sample per participant. Specimens were stored at −80°C until being thawed immediately before aliquoting into a 96-well plate; plated specimens were returned to −80°C until being thawed a second time immediately before assays were run. Interassay coefficients of variation (CVs) for the laboratory were 9.0–18.8% for amyloid-β-40, 6.8–14.6% for amyloid-β-42, 6.7–23.0% for GFAP, 9.6–25.6% for NfL, and 5.6–7.7% for pTau-181.
Evaluations, Risk Factors, and Complications
Risk factors were assessed quarterly using standardized methods during DCCT and annually during EDIC (22), including demographic and behavioral risk factors, medical outcomes, and measurements of height, weight, blood pressure, and pulse rate. HbA1c was measured centrally by high-performance liquid chromatography, and a cumulative measure was derived using the time-weighted mean of all follow-up values from DCCT baseline up to the biomarker study visit. Severe hypoglycemia was self-reported and defined as the cumulative number of events leading to coma or seizure within the 3 months prior to each DCCT/EDIC visit. The presence of kidney disease, proliferative diabetic retinopathy, neurologic complications, or cardiovascular disease was determined as previously described (28–31). All diabetes-related complications were defined as present based on any identification between the DCCT baseline and the biomarker study visit. APOE genotyping was performed as previously described (32); participants were classified as APOE ε4 carriers or noncarriers.
Statistical Analyses
Owing to skewed distributions, plasma biomarker measures were log transformed. Separate linear regression models were used to evaluate the associations of each biomarker with each of the MRI outcomes and cognitive domain z scores. Logistic regression models were used to evaluate the associations between each biomarker and cognitive ratings (normal vs. questionable/impaired cognition). In each model, we tested for an interaction between participant age (≤65 vs. >65 years) and biomarker. Since lower plasma amyloid-β-42 and amyloid-β-42/40 ratio are markers of increased amyloid burden, β-coefficients are presented as the change in outcome per 20% decrease in the biomarker. Similarly, since higher pTau-181 is a marker of increased amyloid burden in the brain and higher NfL and GFAP are markers of neurodegeneration and neuroinflammation, respectively, β-coefficients for these are presented as the change in outcome per 20% increase in the biomarker. Models were minimally adjusted for age, sex, education, intracranial volume (ICV) (for MRI models), and MRI scanner (for MRI models) and fully adjusted to include time-weighted mean HbA1c. Due to the skewed distribution for WMH volume, we applied an inverse hyperbolic sine transformation (asinh), which is similar to a log transformation but can accommodate values of zero (9). Associations of glycemia and diabetes-elated complications (independent variables) with plasma biomarkers (dependent variables) were assessed using linear regression models adjusted for age, sex, education, and time-weighted mean HbA1c. Since the natural log for each biomarker was used in these analyses, covariate effects are expressed as the percent difference in each plasma biomarker. In each model, we tested for an interaction between participant age (≤65 vs. >65 years) and complication status. All analyses were performed using SAS statistical software version 9.4 (SAS Institute). Given the exploratory nature of our analyses, results were not adjusted for multiple testing, and associations with P values <0.01 were considered nominally significant.
Data and Resource Availability
Data collected for the DCCT/EDIC study through 30 June 2017 are available to the public through the National Institutes of Diabetes and Digestive and Kidney Diseases (NIDDK) Central Repository (https://repository.niddk.nih.gov/studies/edic/). Data collected in the current cycle (July 2017–June 2022) will be available within 2 years after the end of the funding cycle.
Results
Participant Characteristics
This ancillary study included 373 participants with a median age of 60 years (range 44–74) and median duration of T1D of 38 years (range 30–51) at the time of biomarker measurement, with 23% of participants >65 years of age (Table 1). Almost all were White. A total of 307 (82%) participants were classified as having normal cognition and 66 (18%) as questionable/impaired. Participants with questionable/impaired cognition were significantly older, with fewer years of education, higher prevalence of smoking (P < 0.05), higher systolic blood pressure (P < 0.05), higher mean HbA1c, higher prevalence of confirmed clinical neuropathy, lower TBV (P < 0.05) and total white matter mean FA (P < 0.05), greater WMH volume (P < 0.05), and higher brain age compared with those with normal cognition. AD signature region atrophy, however, was similar between the groups. Additionally, we observed a higher likelihood of severe hypoglycemia (55.4%) in those with questionable/impaired cognition compared with those with normal cognition (41.8%), although this finding was marginally significant (P = 0.05).
Table 1.
Participant characteristics at the time of the biomarker study visit, by normal versus questionable/impaired cognition
| Parameter | All | Normal | Questionable/impaired | P value |
|---|---|---|---|---|
| N | 373 | 307 | 66 | |
| Demographic | ||||
| Age at diagnosis of T1D (years) | 21.9 ± 7.9 | 21.3 ± 7.9 | 24.3 ± 7.3 | <0.01 |
| Age (years) | 59.8 ± 6.4 | 59.3 ± 6.3 | 62.5 ± 6.3 | <0.01 |
| Age (median [range]) | 60 (44, 74) | 59 (44, 73) | 63 (49, 74) | |
| Age >65 years (%) | 23.1 | 19.5 | 39.4 | <0.001 |
| Female sex (%) | 44.0 | 44.6 | 40.9 | 0.58 |
| White race (%) | 96.5 | 97.1 | 93.9 | 0.21 |
| Education (years) | 15.7 ± 1.9 | 15.8 ± 1.8 | 15.0 ± 2.0 | <0.01 |
| Professional or technical occupation (%) | 57.4 | 59.6 | 47.0 | 0.06 |
| Married or remarried (%) | 81.4 | 83.0 | 73.9 | 0.08 |
| Behavioral | ||||
| Current cigarette smoker (%) | 8.6 | 7.2 | 15.4 | 0.03 |
| Occasional or regular alcohol use (%) | 52.8 | 54.3 | 46.2 | 0.24 |
| Physical | ||||
| BMI (kg/m2) | 28.3 ± 5.2 | 28.4 ± 5.1 | 27.9 ± 5.3 | 0.48 |
| Waist circumference (cm) | 95.1 ± 14.0 | 94.9 ± 13.9 | 96.0 ± 14.2 | 0.63 |
| Blood pressure | ||||
| Systolic blood pressure (mmHg) | 123.4 ± 14.3 | 122.6 ± 13.4 | 127.3 ± 17.3 | 0.04 |
| Diastolic blood pressure (mmHg) | 68.8 ± 8.6 | 68.7 ± 8.5 | 68.9 ± 9.3 | 0.78 |
| History of hypertension (%) | 86.6 | 85.0 | 93.9 | 0.06 |
| Glycemia | ||||
| DCCT/EDIC mean HbA1c (%) | 7.8 ± 0.8 | 7.8 ± 0.8 | 8.1 ± 0.8 | 0.001 |
| Complications* | ||||
| Severe hypoglycemia (%) | 44.2 | 41.8 | 55.4 | 0.05 |
| Sustained AER ≥30 mg/24 h (%) | 25.1 | 24.3 | 28.8 | 0.44 |
| eGFR <60 mL/min/1.73 m2 (%) | 9.4 | 8.8 | 12.1 | 0.41 |
| Proliferative diabetic retinopathy (%) | 24.1 | 22.8 | 30.3 | 0.20 |
| Confirmed clinical neuropathy (%) | 29.6 | 26.5 | 43.9 | <0.01 |
| Cardiovascular disease (%) | 12.9 | 12.1 | 16.7 | 0.31 |
| MRI outcomes† | ||||
| TBV (cm3) | 1,210 ± 120 | 1,217 ± 118 | 1,175 ± 123 | 0.03 |
| WMH volume (cm3) | 2.86 ± 3.73 | 2.69 ± 3.71 | 3.77 ± 3.71 | 0.02 |
| Total white matter mean FA | 0.409 ± 0.024 | 0.410 ± 0.024 | 0.401 ± 0.026 | 0.03 |
| SPARE-AD | −2.06 ± 1.14 | −2.15 ± 1.11 | −1.64 ± 1.19 | <0.01 |
| SPARE-BA | 65.93 ± 8.96 | 64.81 ± 8.83 | 71.15 ± 7.63 | <0.001 |
| Cognitive outcomes† | ||||
| Immediate memory | −0.308 ± 0.911 | −0.096 ± 0.748 | −1.290 ± 0.961 | <0.0001 |
| Delayed recall | −0.218 ± 1.050 | −0.010 ± 0.941 | −1.183 ± 0.995 | <0.0001 |
| Psychomotor and mental efficiency | −1.038 ± 1.210 | −0.744 ± 0.917 | −2.402 ± 1.454 | <0.0001 |
| Plasma biomarkers (median [IQR]) | ||||
| Amyloid-β-42 (pg/mL) | 6.60 (5.50, 7.50) | 6.60 (5.60, 7.50) | 6.35 (5.20, 7.50) | 0.34 |
| Amyloid-β-42/40 ratio | 0.06 (0.06, 0.07) | 0.06 (0.06, 0.07) | 0.06 (0.06, 0.07) | 0.13 |
| NfL (pg/mL) | 15.0 (11.8, 19.9) | 14.6 (11.5, 19.0) | 18.4 (13.3, 26.8) | <0.001 |
| pTau-181 (pg/mL) | 2.20 (1.66, 3.11) | 2.19 (1.66, 2.95) | 2.39 (1.61, 3.79) | 0.19 |
| GFAP (pg/mL) | 104.3 (80.4, 136.2) | 104.0 (80.4, 130.4) | 104.3 (80.0, 151.0) | 0.24 |
| APOE ε4 allele (%) | 29.8 | 29.6 | 30.8 | 0.85 |
Data are means ± SD or percentages unless otherwise noted. Differences between normal vs. questionable/impaired cognition were tested using the Wilcoxon rank sum test for quantitative characteristics or χ2 test for categorical characteristics. Boldface type indicates statistical significance (P < 0.01). AER, albumin excretion rate.
Complications were defined as present based on any identification between the DCCT baseline visit and the biomarker study visit.
P values were adjusted for age, sex, education, ICV (for MRI models), and MRI scanner (for MRI models).
Plasma Biomarker Associations and Patterns
Median (interquartile range [IQR], 25th and 75th percentiles) plasma biomarker values were the following: amyloid-β-42, 6.6 (5.5, 7.5) pg/mL; amyloid-β-42/40 ratio, 0.06 (0.06, 0.07); NfL, 15.0 (11.8, 19.9) pg/mL; pTau-181, 2.2 (1.7, 3.1) pg/mL; and GFAP, 104 (80.4, 136) pg/mL (Table 1). NfL was significantly higher in participants with questionable/impaired cognition (18.4 [13.3, 26.8] pg/mL) compared with those with normal cognition (14.6 [11.5, 19.0] pg/mL) (P = 0.0005). Thirty percent of participants had the APOE ε4 allele, with no significant differences between those with normal versus questionable/impaired cognition.
Analysis of relationships between individual biomarkers showed significant but weak correlations between amyloid-β-42 and NfL (r = 0.221, P < 0.0001), amyloid-β-42 and GFAP (r = 0.213, P < 0.0001), NfL and pTau-181 (r = 0.222, P < 0.0001), and NfL and GFAP (r = 0.381, P < 0.0001) (Supplementary Table 1). Notably, there was essentially no correlation between plasma biomarkers of amyloid burden, specifically amyloid-β-42 and the amyloid-β-42/40 ratio did not correlate with pTau-181, suggesting the absence of a typical amyloid-positive profile. As a sensitivity analysis, we also examined the relationship between amyloid-β-42/40 ratio and pTau-181 with the APOE ε4 allele, a genetic risk factor for AD known to be associated with amyloid burden (33). The presence of the APOE ε4 allele was significantly associated with lower amyloid-β-42/40 ratio (P < 0.001) and slightly higher pTau-181 (not significant, P = 0.36), as expected, providing internal validation of the accuracy of plasma biomarker assays.
Associations of Plasma Biomarkers with MRI and Cognitive Outcomes
We observed several statistically significant associations between plasma biomarkers and MRI and cognitive outcomes. In minimally adjusted models, higher NfL was associated with an increase in predicted brain age according to SPARE-BA (β ± SE, 0.51 ± 0.14 years per 20% increase in NfL; P < 0.001) (Fig. 1 and Supplementary Table 2). This significant association persisted after adjustment for mean HbA1c (β ± SE, 0.54 ± 0.14, P < 0.001) (Table 2), sustained microalbuminuria or reduced estimated glomerular filtration rate (eGFR) (<60 mL/min/1.73 m2), or confirmed clinical neuropathy (data not shown). No biomarkers were significantly associated with TBV, WMH volume, white matter mean FA, or AD-like atrophy according to SPARE-AD. In addition, higher NfL and pTau-181 were associated with lower psychomotor and mental efficiency scores (β ± SE −0.11 ± 0.02, P < 0.0001, for NfL and β ± SE −0.07 ± 0.02, P < 0.001, for pTau-181) (Fig. 1 and Supplementary Table 2). These significant associations persisted after adjustment for mean HbA1c (β ± SE −0.08 ± 0.02, P < 0.001, for NfL and β ± SE −0.06 ± 0.02, P < 0.01, for pTau-181) (Table 2), sustained microalbuminuria or eGFR <60, or confirmed clinical neuropathy (data not shown). A 20% increase in NfL was also associated with a 19.5% increase in the odds of questionable/impaired cognition (P < 0.01) (Supplementary Table 2), which was attenuated after adjustment for mean HbA1c (15.3% increase in odds, P = 0.03) (Table 2). There were no significant differences in any of the associations between participants ≤65 and >65 years of age.
Figure 1.
Associations of plasma biomarkers of brain injury with MRI and cognitive outcomes. Elevations of NfL were associated with increased brain age (SPARE-BA) (left), while elevations of NfL and pTau-181 were associated with decreased psychomotor and mental efficiency (middle and right).
Table 2.
Associations of plasma biomarkers of brain injury with MRI and cognitive outcomes
| TBV (cm3) | WMH volume (cm3) | White matter mean FA | SPARE-AD | SPARE-BA | Immediate memory | Delayed recall | Psychomotor and mental efficiency | Questionable/impaired vs. normal cognition | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| β (SE) | P | β (SE) | P | β (SE) | P | β (SE) | P | β (SE) | P | β (SE) | P | β (SE) | P | β (SE) | P | % (SE) | P | |
| Amyloid-β-42 (pg/mL) | 2.56 (1.75) | 0.14 | −0.01 (0.04) | 0.84 | 0.0008 (0.001) | 0.52 | −0.02 (0.05) | 0.65 | −0.25 (0.30) | 0.39 | −0.05 (0.04) | 0.25 | 0.02 (0.05) | 0.68 | 0.03 (0.05) | 0.55 | 30.2 (16.5) | 0.04 |
| Amyloid-β-42/40 ratio | 2.49 (2.23) | 0.26 | −0.06 (0.05) | 0.26 | −0.0012 (0.002) | 0.41 | −0.06 (0.07) | 0.39 | −0.75 (0.37) | 0.05 | −0.01 (0.05) | 0.90 | 0.05 (0.06) | 0.42 | 0.06 (0.06) | 0.33 | 24.7 (20.0) | 0.17 |
| NfL (pg/mL) | −1.38 (0.87) | 0.11 | 0.02 (0.02) | 0.20 | −0.0012 (0.001) | 0.05 | 0.04 (0.03) | 0.09 | 0.54 (0.14) | <0.001 | −0.04 (0.02) | 0.07 | −0.04 (0.02) | 0.08 | −0.08 (0.02) | <0.001 | 15.3 (7.4) | 0.03 |
| pTau-181 (pg/mL) | 0.98 (0.74) | 0.18 | 0.01 (0.02) | 0.67 | 0.00001 (0.001) | 0.98 | 0.004 (0.02) | 0.85 | 0.05 (0.12) | 0.67 | 0.01 (0.02) | 0.41 | −0.02 (0.02) | 0.36 | −0.06 (0.02) | <0.01 | 0.3 (5.6) | 0.95 |
| GFAP (pg/mL) | −2.11 (0.98) | 0.03 | 0.002 (0.02) | 0.92 | −0.0010 (0.001) | 0.13 | 0.004 (0.03) | 0.89 | 0.23 (0.17) | 0.16 | −0.06 (0.02) | 0.02 | −0.03 (0.03) | 0.25 | −0.04 (0.03) | 0.13 | 6.2 (7.5) | 0.39 |
Data are from separate linear regression models evaluating the association of each plasma biomarker (on the log scale) with each MRI or cognitive outcome. For amyloid-β-42 and the amyloid-β-42/40 ratio, β-coefficients are presented as the change in outcome per 20% decrease in the biomarker, while for NfL, pTau-181, and GFAP, β-coefficients are presented as the change in outcome per 20% increase in the biomarker. The percent change in odds of cognitive impairment (vs. normal) per 20% decrease (amyloid-β-42 and amyloid-β-42/40 ratio) or increase (NfL, pTau-181, and GFAP) is presented from separate logistic regression models. Models were fully adjusted for age, sex, education, ICV (for MRI models), MRI scanner (for MRI models), and DCCT/EDIC mean HbA1c, characterized by the time-weighted mean of all follow-up values from the DCCT baseline up to the biomarker study visit. Biomarkers and outcomes without units are unitless. WMH was assessed in n = 339 participants; an inverse hyperbolic sine transformation was used to normalize the distribution (asinh). White matter mean FA was assessed in n = 323 participants and was not adjusted for ICV.
Associations Between Diabetes-Related Factors and Plasma Biomarkers
Diabetes-related factors and complications showed significant associations with plasma biomarkers. pTau-181 was 13.3% higher among participants originally randomized to DCCT conventional therapy compared with those from the original intensive therapy group (P < 0.01) (data not shown); however, this significant association was attenuated after adjustment for mean HbA1c (10.1%) (P = 0.04). Additionally, in adjusted models, a 1-percentage-point increase in mean HbA1c was associated with a 14.6% higher NfL and 12.8% higher pTau-181 (both P < 0.0001) (Table 3 and Supplementary Fig. 1).
Table 3.
Associations between glycemia and diabetes-related complications and plasma biomarkers of brain injury
| Amyloid-β-42 (pg/mL) | Amyloid-β-42/40 ratio | NfL (pg/mL) | pTau-181 (pg/mL) | GFAP (pg/mL) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| % | P | Mean | % | P | Mean | % | P | Mean | % | P | Mean | % | P | Mean | |
| Glycemia (per 1%) | |||||||||||||||
| DCCT/EDIC mean HbA1c (per 1%) | 2.9 | 0.07 | 6.4 | 0.5 | 0.68 | 0.06 | 14.6 | <0.0001 | 16.4 | 12.8 | <0.0001 | 2.4 | −0.03 | 0.99 | 105.2 |
| Complications (yes vs. no) | |||||||||||||||
| Severe hypoglycemia | −0.04 | 0.99 | 6.4 | 3.3 | 0.11 | 0.06 | −1.3 | 0.75 | 15.4 | −1.4 | 0.78 | 2.3 | −9.6 | <0.01 | 101.2 |
| Sustained AER ≥30 mg/24 h | 5.2 | 0.11 | 6.6 | 5.3 | 0.04 | 0.07 | 4.1 | 0.45 | 16.1 | 16.0 | 0.02 | 2.6 | 7.0 | 0.15 | 112.8 |
| eGFR <60 mL/min/1.73 m2 | 17.8 | <0.001 | 7.4 | −1.1 | 0.76 | 0.06 | 35.5 | <0.0001 | 20.5 | 21.7 | 0.02 | 2.8 | 12.5 | 0.07 | 119.2 |
| Proliferative diabetic retinopathy | 3.4 | 0.29 | 6.5 | −2.7 | 0.28 | 0.06 | 8.4 | 0.12 | 16.6 | 10.0 | 0.12 | 2.5 | 3.6 | 0.45 | 110.0 |
| Confirmed clinical neuropathy | 4.2 | 0.16 | 6.6 | 1.9 | 0.40 | 0.06 | 2.5 | 0.60 | 15.8 | 20.8 | <0.001 | 2.6 | −1.5 | 0.72 | 105.8 |
| Cardiovascular disease | −0.4 | 0.92 | 6.4 | 3.0 | 0.32 | 0.06 | −1.5 | 0.81 | 15.4 | 16.1 | 0.04 | 2.6 | −6.1 | 0.26 | 101.4 |
Data were obtained from separate linear regression models with HbA1c or complication status as the independent variable and the log of each plasma biomarker as the dependent variable. For HbA1c, the percent change in each biomarker per 1-percentage-point increase in HbA1c is presented along with the mean biomarker level for HbA1c 8.2%. For complication status, the percent difference in the geometric means for each biomarker between participants with vs. without each complication is presented along with the geometric mean for participants with the complication. Models were adjusted for age, sex, education, and, for the complications models, DCCT/EDIC mean HbA1c, characterized by the time-weighted mean of all follow-up values from the DCCT baseline up to the biomarker study visit. Boldface type indicates statistical significance (P < 0.01). AER, albumin excretion rate.
After adjustment for age, sex, education, and mean HbA1c, there were no significant associations between sustained microalbuminuria, proliferative diabetic retinopathy, and cardiovascular disease with plasma biomarkers. Amyloid-β-42 was 17.8% higher (P < 0.001) and NfL was 35.5% higher (P < 0.0001) in participants with versus without a history of kidney disease, defined by eGFR <60 mL/min/1.73 m2, while pTau-181 was 20.8% higher in participants with versus without confirmed clinical neuropathy (P < 0.001) (Table 3). Unexpectedly, GFAP was 9.6% lower in participants with versus without a history of severe hypoglycemia (P < 0.01). We also observed a significant interaction between participant age (≤65 vs. >65) and confirmed clinical neuropathy. Among participants >65 years of age, NfL was 33.3% higher and GFAP was 30.9% higher in those with versus without confirmed clinical neuropathy (data not shown) (geometric means 21.6 vs. 16.2 for NfL, P < 0.0001; 153.1 vs. 117.0 for GFAP, P = 0.0012). There were no significant differences in NfL or GFAP among participants ≤65 years of age.
Conclusions
The recent surge in research on plasma biomarkers of brain injury provides a foundation for our study to apply accumulating knowledge regarding these biomarkers to T1D. We couple plasma biomarker measurements with neuroimaging and cognitive assessments to noninvasively study potential mechanisms for poor cognitive function in T1D. We found significant associations between higher NfL and accelerated brain aging assessed by MRI (SPARE-BA) and between higher NfL and pTau-181 with lower psychomotor and mental efficiency. In addition, we demonstrated significant associations between higher HbA1c and higher NfL and pTau-181.
The most notable biomarker results from this study are with NfL, the only biomarker significantly associated with accelerated brain aging in both minimally and fully adjusted models. NfL was also significantly elevated in individuals with T1D with cognitive impairment compared with individuals with T1D with normal cognition. NfL is a cytoskeletal filament protein in neurons that is released in large quantities secondary to neuropathological processes involving axonal damage or neurodegeneration (34). Notably, NfL has both diagnostic utility and prognostic value, as it correlates with disease progression and treatment efficacy for multiple disorders, including multiple sclerosis and AD (34). We previously showed that participants with T1D in the DCCT/EDIC cohort had a significant increase in brain aging (approximately 6 additional years of aging) relative to control participants without diabetes and a pattern of change that was not typical of AD (1,9). We now show that NfL relates to this brain aging signature on MRI, suggesting that NfL is a useful plasma biomarker in T1D because of its correlation with accelerated brain aging. Future studies should assess the longitudinal trajectory of NfL in aging individuals with T1D to determine whether elevated levels precede, and can predict, cognitive decline around the age of 60 years and accelerated aging seen in older adults with T1D.
The biomarker patterns observed in the DCCT/EDIC cohort provide insights into the pathophysiology of cognitive dysfunction in T1D. Both elevated NfL and pTau-181 were found to be significantly associated with lower psychomotor and mental efficiency. Interestingly, phosphorylated tau isoforms like pTau-181 are considered specific markers for amyloid burden in AD (35), but in our T1D cohort, pTau-181 did not correlate with atrophy typical of AD (SPARE-AD) or with our second measure of amyloid burden, the amyloid-β-42/40 ratio, which together indicate that AD is not a primary factor for current cognitive impairments (10). Additionally, we did not observe significant, consistent correlations between the amyloid-β-42/40 ratio and cognitive symptoms or MRI measures. However, we did find expected correlations between pTau-181 and amyloid-β-42/40 ratio with the AD risk allele APOE ε4, indicating that these biomarker findings were not due to erroneous assay measurements. NfL, while being related to SPARE-BA, did not show significant associations with other potential markers of neurodegeneration (SPARE-AD and TBV) or cerebrovascular disease (WMH), possibly because of absent or early subclinical stages of such pathologies. Thus, the lack of associations between markers of cerebral amyloidosis and MRI measures of neurodegeneration suggest that non-AD/nonamyloid factors are driving the observed MRI changes and cognitive dysfunction in T1D in this age-group (1). These biomarker findings are consistent with a preliminary report of associations between plasma NfL and pTau-181 (but not amyloid-β-42/40 ratio or GFAP) with intraindividual cognitive variability in a separate small adult T1D cohort (7). These findings also extend our own recent discovery that there were no differences on MRI in the degree of AD-like neurodegeneration between T1D participants in the DCCT/EDIC cohort and control participants without diabetes (10).
Our finding of significant associations for pTau-181 measures but no associations for amyloid-β measures with lower psychomotor and mental efficiency in T1D are of even greater interest when placed in the context of what is known regarding amyloid-β in individuals with T2D and cognitive dysfunction. Previous studies assessing AD-related amyloid-β in individuals with T2D found no increase in amyloid-β deposition despite their observed increased risk for a clinical syndromal diagnosis of AD (5), which is consistent with our current study demonstrating no associations between amyloid-β biomarkers and cognitive or MRI outcomes in T1D. One study that reported no relationship between T2D and amyloid-β deposition did demonstrate a significant association between T2D and increased phosphorylated tau (36). A second study that used positron emission tomography (PET) tracers to assess amyloid-β and tau load observed that 81% of T2D patients with dementia were positive for tau, but only 39% were amyloid-β positive, a pattern not typical for AD (37). Therefore, our study’s finding that lower cognitive function in a T1D cohort is associated with pTau-181 but not with amyloid-β is consistent with the current literature examining these biomarkers in T2D. This suggests that T1D and T2D share mechanisms or pathways involving phosphorylated tau that impact cognitive dysfunction. Further research is needed to understand why individuals with diabetes and cognitive dysfunction have increases in pTau-181, thought to be specific for amyloid burden, that do not correlate with amyloid-β.
Lastly, NfL and pTau-181 may be candidate biomarkers for cognitive dysfunction due to poor glycemic control. We observed that higher mean HbA1c was significantly associated with both elevated NfL and pTau-181. Additionally, pTau-181 was associated with the diabetes complication of confirmed clinical neuropathy. These findings are of interest given a recent study demonstrating that diabetes treatment status among patients with T2D can affect the cerebrospinal fluid (CSF) tau load. Specifically, untreated adults with T2D were found to have elevated CSF phosphorylated and total tau and more rapid progression to dementia relative to euglycemic control adults, whereas individuals with treated T2D (defined as use of one or more T2D medications) did not differ from euglycemic control adults in tau load or dementia progression (38). Therefore, given our study’s finding that higher HbA1c in T1D is associated with elevated NfL and pTau-181, further research is needed to longitudinally measure the cognitive function and biomarker trajectories for participants, and assess the impact of insulin therapy and tighter glucose control, to determine whether these two biomarkers are accurate for assessing the impact of treatment on diabetes-related neuropathology and tau load.
Study limitations include lack of plasma biomarker measurements in participants without T1D for comparison, lack of racial/ethnic diversity in this largely non-Hispanic White cohort, and the low expected prevalence of AD due to the age of the cohort (median age of 60 years). Therefore, we cannot evaluate whether our biomarker findings are accelerated at this age relative to those of control participants. Additionally, prior studies have demonstrated NfL elevation secondary to peripheral polyneuropathy (39), which is common in T1D, thus we cannot differentiate between peripheral and central contributions to NfL elevation. Indeed, our analysis found a significant association between elevated NfL and confirmed clinical neuropathy among participants aged >65 years. However, the significant associations between NfL and MRI and cognitive outcomes remained unchanged after adjustment for confirmed clinical neuropathy. Our study used plasma biomarkers rather than CSF biomarkers, which are better validated and recommended by current diagnostic guidelines (11,20). Cognitive function assessments relied on targeted research measures instead of full neurocognitive assessments. We did not include 18F-fluorodeoxyglucose positron emission tomography, another means to characterize neurodegeneration patterns. Lastly, prior studies have identified possible confounding of plasma biomarker measures by impaired renal function (40). Although we found significantly higher concentrations of amyloid-β-42 and NfL in individuals with kidney disease, our significant results correlating biomarkers with MRI and cognitive outcomes remained unchanged after adjustment for kidney function.
Altogether, plasma biomarkers of brain injury from the aging adult DCCT/EDIC T1D cohort demonstrate a neurodegeneration pattern that is distinct from the biomarker profile expected for AD (1,9,10), although we cannot exclude possible accelerated preclinical amyloidosis. Additionally, our results, along with literature on neurocognition in T2D, suggest common pathways for diabetes-related cognitive dysfunction involving phosphorylated tau but not amyloid-β deposition. Future repeat neuroimaging, cognitive, and plasma biomarker assessments in the DCCT/EDIC cohort will be invaluable to further elucidate mechanisms as participant’s age into a stage with higher expected prevalence of AD or other neurodegenerative diseases.
This article contains supplementary material online at https://doi.org/10.2337/figshare.25810831.
Article Information
Acknowledgments. The DCCT/EDIC Research Group owes its scientific success and public health contributions to the dedication and commitment of the DCCT/EDIC participants.
Funding. Support for this DCCT/EDIC collaborative study was provided by grant DP3 DK114812. The DCCT/EDIC has been supported by cooperative agreement grants (1982–1993, 2012–2017, and 2017–2022), and contracts (1982–2012) with the Division of Diabetes Endocrinology and Metabolic Diseases of the National Institute of Diabetes and Digestive and Kidney Diseases (current grant nos. U01 DK094176 and U01 DK094157) and by the National Eye Institute, the National Institute of Neurologic Disorders and Stroke, the General Clinical Research Centers Program (1993–2007), and Clinical Translational Science Center Program (2006–present) (Bethesda, MD).
The NIDDK project scientist was not a member of the writing group of this article. The opinions expressed are those of the investigators and do not necessarily reflect the views of the funding agencies.
Duality of Interest. A.B.K. has received research support from Siemens Healthcare Diagnostics and Kyowa Kirin Pharmaceutical Development, has served as an external consultant for Roche Diagnostics, and has received speaker honoraria from the National Kidney Foundation, American Kidney Fund, American Society of Nephrology, and Yale University Department of Laboratory Medicine, all unrelated to this article. Industry contributors have had no role in the DCCT/EDIC study but have provided free or discounted supplies or equipment to support participants’ adherence to the study: Abbott Diabetes Care (Alameda, CA), Animas (Westchester, PA), Bayer Diabetes Care (North America Headquarters, Tarrytown, NY), Becton, Dickinson, and Company (Franklin Lakes, NJ), Eli Lilly (Indianapolis, IN), Extend Nutrition (St. Louis, MO), Insulet (Bedford, MA), Lifescan (Milpitas, CA), Medtronic Diabetes (Minneapolis, MN), Nipro Home Diagnostics (Ft. Lauderdale, FL), Nova Diabetes Care (Billerica, MA), Omron (Shelton, CT), Perrigo Diabetes Care (Allegan, MI), Roche Diabetes Care (Indianapolis, IN), and Sanofi (Bridgewater, NJ). No other potential conflicts of interest relevant to this article were reported.
Author Contributions. A.B.K. and I.M.N. led writing of the manuscript. B.H.B. conducted the statistical analyses. B.H.B., J.A.L., C.M.R., I.B., V.A., M.H., R.A.G.-K., N.C., G.J.B., and A.M.J. wrote sections of the manuscript and reviewed and edited the manuscript. B.H.B. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Handling Editors. The journal editors responsible for overseeing the review of the manuscript were Elizabeth Selvin and Alka M. Kanaya.
Funding Statement
Support for this DCCT/EDIC collaborative study was provided by grant DP3 DK114812. The DCCT/EDIC has been supported by cooperative agreement grants (1982–1993, 2012–2017, and 2017–2022), and contracts (1982–2012) with the Division of Diabetes Endocrinology and Metabolic Diseases of the National Institute of Diabetes and Digestive and Kidney Diseases (current grant nos. U01 DK094176 and U01 DK094157) and by the National Eye Institute, the National Institute of Neurologic Disorders and Stroke, the General Clinical Research Centers Program (1993–2007), and Clinical Translational Science Center Program (2006–present) (Bethesda, MD).
Footnotes
Clinical trial reg. nos. NCT00360815 and NCT00360893, clinicaltrials.gov
A.B.K. and I.M.N. are co-first authors.
A complete list of the members of the DCCT/EDIC Research Group can be found in the supplementary material online.
This article is part of a special article collection available at diabetesjournals.org/collection/2296/DCCT-EDIC-40th-Anniversary-Collection.
This article is featured in a podcast available at diabetesjournals.org/care/pages/diabetes_care_on_air.
References
- 1. Jacobson AM, Ryan CM, Braffett BH, et al.; DCCT/EDIC Research Group . Cognitive performance declines in older adults with type 1 diabetes: results from 32 years of follow-up in the DCCT and EDIC study. Lancet Diabetes Endocrinol 2021;9:436–445 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Ryan CM, van Duinkerken E, Rosano C. Neurocognitive consequences of diabetes. Am Psychol 2016;71:563–576 [DOI] [PubMed] [Google Scholar]
- 3. Zilliox LA, Chadrasekaran K, Kwan JY, Russell JW. Diabetes and cognitive impairment. Curr Diab Rep 2016;16:87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Gaspar JM, Baptista FI, Macedo MP, et al. Inside the diabetic brain: role of different players involved in cognitive decline. ACS Chem Neurosci 2016;7:131–142 [DOI] [PubMed] [Google Scholar]
- 5. Biessels GJ, Nobili F, Teunissen CE, Simo R, et al. Understanding multifactorial brain changes in type 2 diabetes: a biomarker perspective. Lancet Neurol 2020;19:699–710 [DOI] [PubMed] [Google Scholar]
- 6. Teunissen CE, Verberk IMW, Thijssen EH, et al. Blood-based biomarkers for Alzheimer’s disease: towards clinical implementation. Lancet Neurol 2022;21:66–77 [DOI] [PubMed] [Google Scholar]
- 7. Fonseca LM, Kivisakk P, Hawks ZW, et al. Plasma Alzheimer’s disease biomarkers and intraindividual cognitive variability measured by ecological momentary assessment in adults with type 1 diabetes. Alzheimers Dement 2023;19(Suppl. 15):e080319 [Google Scholar]
- 8. Sampedro F, Stantonyonge N, Martínez-Horta S, et al. Increased plasma neurofilament light chain levels in patients with type-1 diabetes with impaired awareness of hypoglycemia. BMJ Open Diabetes Res Care 2020;8:e001516 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Jacobson AM, Braffett BH, Erus G, et al.; DCCT/EDIC Research Group . Brain structure among middle-aged and older adults with long-standing type 1 diabetes in the DCCT/EDIC study. Diabetes Care 2022;45:1779–1787 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Habes M, Jacobson AM, Braffett BH, et al.; DCCT/EDIC Research Group . Patterns of regional brain atrophy and brain aging in middle- and older-aged adults with type 1 diabetes. JAMA Netw Open 2023;6:e2316182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Jack CR Jr, Bennett DA, Blennow K, et al.; Contributors . NIA-AA Research Framework: toward a biological definition of Alzheimer’s disease. Alzheimers Dement 2018;14:535–562 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Janelidze S, Mattsson N, Palmqvist S, et al. Plasma P-tau181 in Alzheimer’s disease: relationship to other biomarkers, differential diagnosis, neuropathology and longitudinal progression to Alzheimer’s dementia. Nat Med 2020;26:379–386 [DOI] [PubMed] [Google Scholar]
- 13. Palmqvist S, Janelidze S, Stomrud E, et al. Performance of fully automated plasma assays as screening tests for Alzheimer disease-related β-amyloid status. JAMA Neurol 2019;76:1060–1069 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Mattsson N, Cullen NC, Andreasson U, Zetterberg H, Blennow K. Association between longitudinal plasma neurofilament light and neurodegeneration in patients with Alzheimer disease. JAMA Neurol 2019;76:791–799 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Giudici KV, de Souto Barreto P, Guyonnet S, Li Y, Bateman RJ; MAPT/DSA Group . Assessment of plasma amyloid-β42/40 and cognitive decline among community-dwelling older adults. JAMA Netw Open 2020;3:e2028634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. He L, de Souto Barreto P, Aggarwal G, et al.; MAPT/DSA Group . Plasma Aβ and neurofilament light chain are associated with cognitive and physical function decline in non-dementia older adults. Alzheimers Res Ther 2020;12:128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Shekhar S, Kumar R, Rai N, et al. Estimation of tau and phosphorylated tau181 in serum of Alzheimer’s disease and mild cognitive impairment patients. PLoS One 2016;11:e0159099. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Abdelhak A, Foschi M, Abu-Rumeileh S, et al. Blood GFAP as an emerging biomarker in brain and spinal cord disorders. Nat Rev Neurol 2022;18:158–172 [DOI] [PubMed] [Google Scholar]
- 19. Gaetani L, Blennow K, Calabresi P, Di Filippo M, Parnetti L, Zetterberg H. Neurofilament light chain as a biomarker in neurological disorders. J Neurol Neurosurg Psychiatry 2019;90:870–881 [DOI] [PubMed] [Google Scholar]
- 20. Frisoni GB, Festari C, Massa F, et al. European intersocietal recommendations for the biomarker-based diagnosis of neurocognitive disorders. Lancet Neurol 2024;23:302–312 [DOI] [PubMed] [Google Scholar]
- 21. Nathan DM, Genuth S, Lachin J, et al.; Diabetes Control and Complications Trial Research Group . The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977–986 [DOI] [PubMed] [Google Scholar]
- 22. Epidemiology of Diabetes Interventions and Complications (EDIC) Research Group . Epidemiology of Diabetes Interventions and Complications (EDIC). Design, implementation, and preliminary results of a long-term follow-up of the Diabetes Control and Complications Trial cohort. Diabetes Care 1999;22:99–111 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Doshi J, Erus G, Ou Y, et al.; Alzheimer’s Neuroimaging Initiative . MUSE: multi-atlas region segmentation utilizing ensembles of registration algorithms and parameters, and locally optimal atlas selection. Neuroimage 2016;127:186–195 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Nasrallah IM, Pajewski NM, Auchus AP, et al.; SPRINT MIND Investigators for the SPRINT Research Group . Association of intensive vs standard blood pressure control with cerebral white matter lesions. JAMA 2019;322:524–534 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Jacobson AM, Musen G, Ryan CM, et al.; Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Study Research Group . Long-term effect of diabetes and its treatment on cognitive function. N Engl J Med 2007;356:1842–1852 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Strauss E, Sherman EMS, Spreen O. A Compendium of Neuropsychological Tests: Administration, Norms, and Commentary. Oxford, U.K., Oxford University Press, 2006 [Google Scholar]
- 27. Chaytor NS, Trapani VR, Braffett BH, et al. Utility of the NIH Toolbox Cognition Battery in middle to older aged adults with longstanding type 1 diabetes: the DCCT/EDIC study. Clin Neuropsychol 2023;1007–1027 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Perkins BA, Bebu I, de Boer IH, et al.; Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC) Research Group . Risk factors for kidney disease in type 1 diabetes. Diabetes Care 2019;42:883–890 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Hainsworth DP, Bebu I, Aiello LP, et al.; Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC) Research Group . Risk factors for retinopathy in type 1 diabetes: the DCCT/EDIC study. Diabetes Care 2019;42:875–88230833368 [Google Scholar]
- 30. Braffett BH, Gubitosi-Klug RA, Albers JW, et al.; DCCT/EDIC Research Group . Risk factors for diabetic peripheral neuropathy and cardiovascular autonomic neuropathy in the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) study. Diabetes 2020;69:1000–1010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) Research Group . Risk factors for cardiovascular disease in type 1 diabetes. Diabetes 2016;65:1370–1379 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Jacobson AM, Paterson AD, Ryan CM, et al.; DCCT/EDIC Research Group . The associations of apolipoprotein E and angiotensin-converting enzyme polymorphisms and cognitive function in type 1 diabetes based on an 18-year follow-up of the DCCT cohort. Diabet Med 2010;27:15–22 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Corder EH, Saunders AM, Strittmatter WJ, et al. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in late onset families. Science 1993;261:921–923 [DOI] [PubMed] [Google Scholar]
- 34. Coppens S, Lehmann S, Hopley C, Hirtz C. Neurofilament-light, a promising biomarker: analytical, metrological and clinical challenges. Int J Mol Sci 2023;24:11624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Karikari TK, Pascoal TA, Ashton NJ, et al. Blood phosphorylated tau 181 as a biomarker for Alzheimer’s disease: a diagnostic performance and prediction modelling study using data from four prospective cohorts. Lancet Neurol 2020;19:422–433 [DOI] [PubMed] [Google Scholar]
- 36. Moran C, Beare R, Phan TG, et al. Type 2 diabetes mellitus and biomarkers of neurodegeneration. Neurology 2015;85:1123–1130 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Takenoshita N, Fukasawa R, Ogawa Y, et al. Amyloid and tau positron emission tomography in suggested diabetes-related dementia. Curr Alzheimer Res 2018;15:1062–1069 [DOI] [PubMed] [Google Scholar]
- 38. McIntosh EC; Alzheimer’s Disease Neuroimaging Initiative . Importance of treatment status in links between type 2 diabetes and Alzheimer’s disease. Diabetes Care 2019;42:972–979 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Fridman V, Sillau S, Ritchie A, et al. Plasma neurofilament light chain concentrations are elevated in youth-onset type 2 diabetes and associated with neuropathy. J Peripher Nerv Syst 2023;28:460–470 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Pajewski NM, Elahi FM, Tamura MK, et al. Plasma amyloid beta, neurofilament light chain, and total tau in the Systolic Blood Pressure Intervention Trial (SPRINT). Alzheimers Dement 2022;18:1472–1483 [DOI] [PMC free article] [PubMed] [Google Scholar]