Biomarkers of Prevalent and Incident Cognitive Dysfunction in People with Dysglycemia: Data from the ORIGIN Trial

Abstract

Background:

Diabetes and cardiovascular disease increase the risk of incident cognitive dysfunction. Identification of novel biochemical markers for cognitive dysfunction may identify people at the highest risk while yielding insights regarding the pathophysiology of cognitive dysfunction.

Objective:

To identify cardiovascular biomarkers in serum that are independent predictors of cognitive dysfunction in individuals with dysglycemia.

Methods:

This analysis was conducted in 8,365 participants in the Outcome Reduction with an Initial Glargine Intervention (ORIGIN) trial whose stored serum was analyzed for 238 cardio-metabolic biomarkers and completed a baseline Mini-Mental State Examination (MMSE). Fine and Gray sub distribution hazard models accounting for the competing risk of death accounting for clinical risk factors and the baseline MMSE were used to identify biomarkers that predicted incident cognitive dysfunction (MMSE < 24 or dementia) using forward selection with an inclusion p-value < 0.0002 to account for multiplicity.

Results:

During a median follow-up period of 6.2 years, 939 individuals developed cognitive dysfunction. After accounting for 17 clinical risk factors, glargine allocation, and the baseline MMSE, three biomarkers (α-2 Macroglobulin, HR 1.19; 95% CI 1.12, 1.27; Macrophage Inflammatory Protein 1α, HR 1.11; 95% CI 1.06, 1.16; and Growth Hormone, HR 0.91; 95% CI 0.87, 0.96) independently predicted incident cognitive dysfunction (p < 0.0002). Addition of these biomarkers to a model that included clinical risk factors, however, did not improve the ability to predict cognitive dysfunction.

Conclusion:

Addition of independent biomarkers to clinical risk factors for cognitive dysfunction in people with dysglycemia did not predict incident cognitive dysfunction better than clinical risk factors alone.

Keywords

Biomarkers cardiovascular disease cognition cognitive dysfunction diabetes dysglycemia

Clinicaltrials.gov: NCT00069784

INTRODUCTION

Diabetes is a disease of accelerated cognitive aging. Compared to people without diabetes, those with diabetes have a 1.5 to 2-fold greater risk for experiencing an accelerated rate of cognitive decline and also for the development of mild cognitive impairment and dementia [1 –3]. Indeed, in the US one out of six cases of dementia is attributable to diabetes [4, 5].

Cardiovascular disease (CVD) and cardiovascular risk factors are emerging as important risk factors for cognitive dysfunction in people with diabetes. For example, epidemiological studies have demonstrated that in people with diabetes, both hypertension and a prior stroke were risk factors for incident dementia [6 –8]. Many pathophysiological mechanisms have been proposed to explain this relationship; however, the underlying pathogenic route is still unclear. Advances in measuring large numbers of serum biomarkers related to many pathophysiological pathways in small volumes of serum have enabled identification of new biomarkers for cardiovascular outcomes [9]. Assessing the relationship of these biomarkers to cognitive dysfunction and identifying novel biomarkers may potentially lead to novel insights regarding the pathophysiology of cognitive dysfunction in people with dysglycemia.

The ORIGIN trial, which assessed incident cognitive dysfunction over a median of 6.2 years and collected serum samples at baseline which were used to measure 238 biochemical markers, provides a unique opportunity to identify novel biomarkers for prevalent and incident cognitive dysfunction.

METHODS

The ORIGIN trial randomly assigned 12,537 middle-aged and older people with dysglycemia and additional cardiovascular risk factors to insulin-mediated normoglycemia versus standard care without insulin and to omega 3 fatty acid versus placebo. Participants were followed for a median of 6.2 years for the development of serious health outcomes, and cognitive function was assessed at baseline and during follow-up. The design, main results, and cognitive sub-study results of ORIGIN have been previously published [10, 11]. Briefly, individuals aged 50 years or older with impaired fasting glucose (IFG) (defined as a fasting plasma glucose (FPG) ≥110 and < 126 mg/dl (≥6.1 and < 7.0 mmol/L) without diabetes mellitus), impaired glucose tolerance (IGT) (defined as post prandial glucose (PPG) ≥140 and < 200 mg/dl (≥7.8 but < 11.1 mmol/L) with a FPG < 126 mg/dl (7.0 mmol/L) or early type 2 diabetes (defined as a FPG ≥126 mg/dl (7.0 mmol/L) or a PPG≥200 mg/dl (11.1 mmol/L) or a previous diagnosis of diabetes and either on no pharmacological treatment or taking one oral anti diabetic drug, who also had additional cardiovascular risk (defined as previous peripheral-cardio-cerebrovascular disease, microalbuminuria, or left ventricular hypertrophy) factors were recruited. Participants were asked to complete a Mini-Mental State Examination (MMSE) at baseline and at 3 additional time points during the trial. Incident cognitive dysfunction was also ascertained.

Bio-banked serum was analyzed for a panel of cardio-metabolic biomarkers in 8,401 ORIGIN participants [9], and a baseline MMSE was recorded in 8,365 individuals. The biomarkers consisted of a combination of assays that were chosen based on their role in CVD: inflammation, coagulation, endothelial dysfunction, renal function, oxidative stress, adipocyte biology, angiogenesis, B-cell biology, tissue repair, and iron metabolism. The methodology used and the list of measured biomarkers appears in the Supplementary section of Gerstein et al. [9]

The MMSE is a measure of global cognitive function that has high sensitivity and specificity for the detection of dementia [12], and that contains 30 items pertaining to orientation, registration, attention and calculation recall, language, and visual-spatial ability [13]. In the ORIGIN trial, a validated translation was administered in the participant’s first language, and scored centrally. All forms were reviewed, and implausible scores were verified with the submitting sites utilizing a quality assurance scheme.

Incident cognitive dysfunction was defined as either reported dementia (defined as the first occurrence of a reported diagnosis of dementia since the last study visit) or an incident post-randomization MMSE score < 24 [11]. Participants were classified as having probable depression if they indicated during the randomization visit that they were feeling “sad, low in spirits or depressed for two or more weeks” and if they also indicated that during that time they “thought a lot about death or required treatment for depression”. Previous CVD was defined as a history of myocardial infarction, stroke, or previous revascularization; previous cerebrovascular disease was defined as a history of previous stroke; microvascular disease as a history of either retinal laser therapy, vitrectomy, or albuminuria; previous foot disease as a history of limb amputation, foot amputation, or foot infection requiring antibiotic treatment; and alcohol consumption was defined as more than two drinks per week.

Statistical analysis

Included in this analysis is data from 8,365 ORIGIN participants for whom both biomarker levels and a baseline MMSE score were available. Continuous variables were summarized as means with standard deviations (SD) or medians with interquartile ranges (IQR), and binary variables were summarized as counts and percentages. Differences in the distribution of the baseline variables were assessed using t-tests for means, nonparametric Wilcoxon tests for medians, and chi-square tests for counts (percentages).

Biomarkers that were independently associated with a baseline MMSE score < 24 at baseline after accounting for clinical risk factors were identified using logistic regression. The clinical risk factors included in the basic model were sex, hemoglobin A1c (HbA1c), glargine allocation, and 6 other variables that were previously identified as predictors of dementia in people with diabetes including age, education (≤8 years, 9–12 years, and > 12 years), depression, microvascular disease, previous CVD, and foot disease [14]. Additional risk factors included in a more comprehensive clinical model (i.e., the full model) included prior cerebrovascular disease, hypertension, dysglycemia status (diabetes, IFG/IGT), alcohol consumption, current smoking, LDL, systolic blood pressure, and creatinine values. A forward selection process was used to identify biomarkers that were independently associated with an MMSE < 24 at baseline. The p value for inclusion of biomarkers in these models was 0.05 divided by 238 (i.e., 0.00021) to account for the 238 comparisons.

Fine and Gray sub distribution hazard models accounting for the competing risk of death and accounting for clinical risk factors described above as well as the baseline MMSE were used to identify biomarkers that independently predicted incident cognitive dysfunction using a similar forward selection process. The analysis was repeated for the sub-set of individuals with a baseline MMSE below the lower interquartile range for the entire population (i.e.,<27).

Harrell’s C statistics [15] with 95% confidence intervals (CI) were calculated for the basic and full models with the clinical risk factors alone and with the risk factor models plus the identified biomarkers. The improvement in the ability of the model to predict the outcome was summarized using the net reclassification improvement (NRI) after classifying people into 4 categories of risk [16] defined by model-estimated probabilities of incident cognitive dysfunction of 0.05, 0.2 and 0.7. All statistical analyses were done using SAS version 9.4 for UNIX (SAS Institute Inc, Cary, NC).

RESULTS

Table 1 lists the baseline characteristics of the 8,365 people with both biomarker data and a baseline MMSE score and tabulates them according to whether the score was < 24 or≥24 at baseline. Individuals with a lower baseline MMSE were older (66.1 versus 63.5 years; p < 0.001), had a lower BMI (29.1 versus 30.2 p < 0.001), were less likely to have more than 12 years of education (10.9% versus 37.3%; p < 0.001), were more likely to be non-white (67.4% versus 36.5%; p < 0.001), female (59.9% versus 31.5%; p < 0.001), have prior hypertension (85% versus 78.4%; p < 0.001), cerebrovascular disease (21.7% versus 11.7%; p < 0.001), microvascular disease (39.9% versus 31.5%; p < 0.001), and foot disease (5.2% versus 2.5%; p < 0.001), and less likely to be smokers (8.5% versus 12.9%; p < 0.001) and alcohol consumers (10.4% versus 26%; p < 0.001), and have prior CVD (47.1% versus 60.4%; p < 0.001). Supplementary Table 1 depicts the difference in characteristics of the 8,365 ORIGIN participants that were included in the biomarker analysis versus the 4,172 participants that were not.

Table 1

Baseline Characteristics of 8,365 with a Baseline MMSE Score

Variable	MMSE < 24	MMSE ≥24	p
N	709	7,656
Age	66.1 (8.6)	63.5 (7.8)	< 0.001
Females	425 (59.9)	2,409 (31.5)	< 0.001
Education (y)
< 8	533 (75.2)	2,676 (35.0)
9–12	99 (14.0)	2,123 (27.7)
> 12	77 (10.9)	2,857 (37.3)	< 0.001
Ethnicity
White	231 (32.6)	4,862 (63.5)
Black	96 (13.5)	260 (3.4)
South Asian	56 (7.9)	391 (5.1)
Other Asian	2 (0.3)	36 (0.5)
Latin	312 (44.0)	1,917 (25.0)
Other	12 (1.7)	190 (2.5)	< 0.001
Depression¹	110 (15.5)	1,194 (15.6)	0.955
Hypertension	603 (85.0)	6,003 (78.4)	< 0.001
Body mass index (Kg/m²)	29.1 (5.6)	30.2 (5.2)	< 0.001
Current smoker²	60 (8.5)	987 (12.9)	< 0.001
Alcohol consumer³	74 (10.4)	1,993 (26.0)	< 0.001
Prior CVD⁴	334 (47.1)	4,626 (60.4)	< 0.001
Prior cerebrovascular disease⁵	154 (21.7)	895 (11.7)	< 0.001
Microvascular disease⁶	283 (39.9)	2,414 (31.5)	< 0.001
Foot disease⁷	37 (5.2)	192 (2.5)	< 0.001
HbA1c (%)	6.52 (1.0)	6.50 (0.9)	0.512
FPG (mmol/l)	7.24 (2.3)	7.32 (2.0)	0.386
Allocated to Glargine	342 (48.2)	3,832 (50.1)	0.355
Allocated to Omega 3	342 (48.2)	3,852 (50.3)	0.290
MMSE	20.6 (2.7)	28.5 (1.7)	< 0.001
MMSE, median (IQR)	21 (19–23)	29 (27–30)	< 0.001

N (%); Mean (SD); ¹affirmative answer to “sad, low in spirits or depressed for 2 or more weeks” and also “thought a lot about death or required treatment for depression; ²report of current smoking; ³more than two drinks per week; ⁴MI, stroke or previous revascularization; ⁵report of a previous stroke; ⁶composite of laser/vitrectomy or any microalb/albuminuria; ⁷composite of limb or foot amputation OR foot infection/ulcer requiring antibiotic treatment. MMSE, Mini-Mental State Examination; CVD, cardiovascular disease; HbA1c, hemoglobin A1c; FPG, fasting plasma glucose.

Biomarkers associated with prevalent cognitive dysfunction

After forcing in clinical risk factors for cognitive dysfunction, 6 biomarkers including heat shock protein 70, IgE, leptin, growth regulated α protein, receptor for advanced glycosylation end products (RAGE), and thrombomodulin independently improved the ability to discriminate between individuals with a baseline MMSE < 24 and those with a higher MMSE (Table 2). Thus, higher levels of heat shock protein 70, IgE, growth regulated α protein, and thrombomodulin were all associated with higher odds for a MMSE < 24, whereas higher levels of leptin and RAGE were associated with a lower odds for a MMSE < 24. C statistics for the model without and with the biomarkers were 0.76 (95% CI 0.74, 0.78) and 0.79 (95% CI 0.77, 0.81) respectively (NRI 0.11, 95% CI 0.07, 0.15). After forcing in additional clinical risk factors, only three biomarkers were independent determinants of an MMSE < 24 including heat shock protein 70, epithelial neutrophil activating peptide 78 (ENA-78), and IgE. Thus, higher levels of heat shock protein 70, ENA-78, and IgE were all associated with higher odds for a MMSE < 24. C statistics for the model without and with the biomarkers were 0.78 (95% CI 0.77, 0.80) and 0.80 (95% CI 0.78, 0.82) respectively (NRI 0.05, 95% CI 0.01, 0.09).

Table 2

Adjusted Risk for Prevalent Cognitive Dysfunction (Baseline MMSE < 24)

Variable /Name	Basic Model OR (95% CI)	Full Model OR (95% CI)
Age	1.29 (1.19, 1.41)	1.32 (1.21, 1.43)
Male	0.40 (0.32, 0.50)	0.67 (0.55, 0.83)
Education 9–12 y	0.33 (0.26, 0.42)	0.34 (0.27, 0.43)
Education > 12 y	0.26 (0.20, 0.33)	0.28 (0.22, 0.36)
Depression	0.90 (0.71, 1.41)	0.90 (0.70, 1.14)
Microvascular disease	1.04 (0.87, 1.24)	1.04 (0.87, 1.25)
Foot disease	1.25 (0.80, 1.93)	1.46 (0.94, 2.28)
Prior CVD	0.94 (0.79, 1.12)	0.63 (0.51, 0.78)
HbA1c	0.93 (0.85, 1.02)	0.93 (0.85, 1.02)
Glargine allocation	0.97 (0.82, 1.14)	0.95 (0.80, 1.12)
Cerebrovascular disease	N/A	2.65 (2.06, 3.40)
Hypertension	N/A	1.24 (0.97, 1.58)
Current smoking	N/A	0.86 (0.64, 1.15)
Alcohol consumption	N/A	0.58 (0.44, 0.76)
Diabetes	N/A	1.00 (0.74, 1.34)
LDL	N/A	0.97 (0.89, 1.05)
Systolic blood pressure	N/A	1.00 (1.00, 1.01)
Creatinine	N/A	1.00 (0.99, 1.00)
Heat-Shock protein 70	1.17 (1.10, 1.24)	1.20 (1.13, 1.27)
ENA-78	NA	1.25 (1.15, 1.37)
IgE	1.15 (1.09, 1.21)	1.15 (1.08, 1.21)
Leptin	0.84 (0.76, 0.92)	N/A
Growth-Regulated α protein	1.22 (1.12, 1.32)	N/A
RAGE	0.82 (0.75, 0.89)	N/A
Thrombomodulin	1.18 (1.09, 1.28)	N/A

Odds ratios and 95% confidence intervals (OR, 95% CI) from logistic regression models are shown. The OR denotes the odds of a Mini-Mental State Examination (MMSE) <24 for every 1 standard deviation higher level of the biomarker. ENA-78, Epithelial neutrophil activating peptide 78; RAGE, receptor for advanced glycosylation end products; ^*basic model included adjustment for the following variables: age, education (< 8 y, 9–12 y, > 12 y), depression, microvascular disease, prior cardiovascular event, diabetic foot; full model included additional adjustment for the following variables: prior cerebrovascular disease, hypertension, dysglycemia status (diabetes, IFG/IGT), alcohol consumption, current smoking, LDL, systolic blood pressure, and creatinine values.

Biomarkers predicting incident cognitive dysfunction

During a median follow-up of 6.2 years, 7,822 of the 8,365 individuals who had a baseline MMSE had at least one follow-up MMSE. Information regarding the presence or absence of reported dementia was available in the 7,822 plus the remaining 543 people with no follow-up MMSE score. A total of 939 out of 7,822 (12%) individuals developed incident cognitive dysfunction based on incident post-randomization MMSE score < 24 or incident dementia. Of these individuals, 1,326 died and 113 developed cognitive dysfunction before death.

After forcing in common clinical risk factors (Table 3), three biomarkers including α2-macroglobulin (A2M), macrophage inflammatory protein 1α (MIP-1α), and ENA-78 independently predicted incident cognitive dysfunction. Higher values of these were associated with an increased risk for incident cognitive dysfunction. When the process was repeated after forcing in additional clinical risk factors, three biomarkers including A2M, MIP-1α, and growth hormone independently predicted cognitive dysfunction. Thus, higher values of A2M and MIP-1α were associated with an increased risk, whereas higher values of growth hormone were associated with a reduced risk for incident cognitive dysfunction respectively. These biomarkers did not improve the ability to predict incident cognitive dysfunction compared to the clinical risk factors alone (Table 5).

Table 3

Adjusted Hazard for Incident Cognitive Dysfunction Accounting for the Competing Risk of Death

Variable/Name	Basic Model HR (95% CI)	Full Model HR (95% CI)
Age	1.26 (1.18, 1.35)	1.32 (1.23, 1.43)
Male	1.00 (0.86, 1.17)	1.06 (0.91, 1.25)
Education 9–12 y	0.68 (0.58, 0.81)	0.67 (0.57, 0.80)
Education > 12 y	0.48 (0.40, 0.58)	0.48 (0.40, 0.57)
Depression	0.99 (0.82, 1.19)	0.99 (0.82, 1.19)
MMSE at baseline	0.92 (0.90, 0.94)	0.92 (0.90, 0.95)
Microvascular disease	0.97 (0.84, 1.12)	0.99 (0.85, 1.14)
Foot disease	0.85 (0.54, 1.31)	0.89 (0.57, 1.38)
Prior CVD	0.90 (0.79, 1.04)	0.84 (0.72, 0.98)
A1C	1.00 (0.93, 1.08)	0.98 (0.91, 1.05)
Glargine allocation	0.92 (0.81, 1.05)	0.92 (0.80, 1.05)
Cerebrovascular disease	N/A	1.36 (1.11, 1.67)
Hypertension	N/A	1.17 (0.98, 1.41)
Current smoking	N/A	1.08 (0.87, 1.33)
Alcohol consumption	N/A	0.77 (0.64, 0.92)
Diabetes	N/A	1.33 (1.04, 1.70)
LDL	N/A	1.02 (0.96, 1.09)
Systolic blood pressure	N/A	1.00 (1.00, 1.00)
Creatinine	N/A	0.99 (0.99, 1.00)
A2M	1.20 (1.13, 1.27)	1.19 (1.12, 1.27)
MIP-1α	1.10 (1.05, 1.15)	1.11 (1.06, 1.16)
ENA-78	1.16 (1.08, 1.25)	N/A
Growth hormone	N/A	0.91 (0.87, 0.96)

Hazard ratios and 95% confidence intervals (HR, 95% CI) from Cox regression models are shown. A2M, α2 macroglobulin; MIP-1α, macrophage inflammatory protein 1 α; ENA-78, Epithelial neutrophil activating peptide 78; ^* basic model included adjustment for the following variables: age, education (≤8 y, 9–12 y, > 12 y), depression, microvascular disease, prior cardiovascular event, diabetic foot; full model included additional adjustment for the following variables: prior cerebrovascular disease, hypertension, dysglycemia status (diabetes, IFG/IGT), alcohol consumption, current smoking, LDL, systolic blood pressure, and creatinine values.

Table 5

Model Performance of the Cognitive ORIGIN Biomarker study

Incident Cognitive Dysfunction	C statistic (Clinical only)	C statistic (+Biomarkers)	NRI
Basic Model
All Participants	0.69 (0.68, 0.70)	0.70 (0.69, 0.72)	0.00 (0.00, 0.01)
Baseline MMSE < 27	0.56 (0.53, 0.58)	0.57 (0.55, 0.59)	0.03 (0.00, 0.06)
Full Model
All Participants	0.69 (0.68, 0.71)	0.70 (0.69, 0.71)	0.07 (0.04, 0.09)
Baseline MMSE < 27	0.56 (0.54, 0.58)	0.57 (0.55, 0.59)	0.08 (0.05, 0.10)

Finally, when the analyses were repeated in the subset of 1,994 people whose baseline MMSE was < 27), two biomarkers including tumor necrosis factor 1 (TNF-1) and MIP-1α independently predicted cognitive dysfunction. When the process was repeated using an expanded clinical model, only MIP-1α was an independent risk factor (Table 4). However, the addition of these biomarkers did not improve the C statistic (Table 5).

Table 4

Adjusted Hazard for Incident Cognitive Dysfunction Accounting for Competing Risk of Death in the Subset with Baseline MMSE < lower IQR Range (<27)

Variable/Name	Basic Model HR (95% CI)	Full Model HR (95% CI)
Age	1.29 (1.16, 1.44)	1.29 (1.15, 1.44)
Male	0.82 (0.66, 1.02)	0.99 (0.78, 1.27)
Education 9–12 y	0.75 (0.56, 0.98)	0.75 (0.57, 1.00)
Education > 12 y	0.47 (0.34, 0.67)	0.45 (0.32, 0.64)
Depression	1.04 (0.78, 1.40)	1.05 (0.78, 1.41)
MMSE at baseline	1.21 (1.15, 1.28)	1.24 (1.17, 1.30)
Microvascular disease	0.88 (0.70, 1.11)	0.86 (0.68, 1.08)
Foot disease	0.65 (0.32, 1.33)	0.62 (0.30, 1.26)
Prior CVD	0.90 (0.72, 1.11)	0.76 (0.59, 0.98)
A1C	1.03 (0.92, 1.15)	1.01 (0.91,1.13)
Glargine allocation	0.91 (0.74, 1.12)	0.93 (0.75, 1.15)
Cerebrovascular disease	N/A	1.53 (1.11, 2.10)
Hypertension	N/A	1.08 (0.80,1.47)
Current smoking	N/A	0.92 (0.63, 1.33)
Alcohol consumption	N/A	0.78 (0.56, 1.08)
Diabetes	N/A	1.09 (0.74, 1.60)
LDL	N/A	1.01 (0.91, 1.13)
Systolic blood pressure	N/A	0.99 (0.99, 1.00)
Creatinine	N/A	0.99 (0.99, 1.00)
TNF Receptor 1	0.82 (0.74, 0.91)	N/A
MIP-1α	1.16 (1.09, 1.24)	1.14 (1.07, 1.21)

Hazard ratios and 95% confidence intervals (HR, 95% CI) from Cox regression models are shown. MIP-1α, Macrophage inflammatory protein 1 α; TNF, tumor necrosis factor; ^* basic model included adjustment for the following variables: age, education (< 8 y, 9–12 y, > 12 y), depression, microvascular disease, prior cardiovascular event, diabetic foot; the full model includes additional adjustment for the following variables: prior cerebrovascular disease, hypertension, dysglycemia status (diabetes, IFG/IGT), alcohol consumption, current smoking, LDL, systolic blood pressure, and creatinine values.

DISCUSSION

This analysis of 238 biomarkers conducted in 8,365 ORIGIN participants identified several biomarkers that were each independent predictors of cognitive dysfunction in people with dysglycemia. However, the discriminative ability of a model that included these biomarkers plus the clinical risk factors was similar to that of a model with just the clinical risk factors. Thus, although these biomarkers may be directly or indirectly relevant to the pathophysiology of cognitive dysfunction, they are unlikely to add clinically useful information regarding an individual’s risk of this outcome. The two biomarkers that were identified using both the basic and full model were MIP-1α and A2M. MIP-1α is a chemokine found in the central nervous system with receptors found on astrocytes, microglia, and neurons. It is thought to play a role in neurodegenerative processes. Thus, expression of MIP-1α on neurons and glial cells has been reported in postmortem brains of patients who had AD [17]. Higher levels of MIP-1α have also been found in lymphocytes of AD patients when compared to normal controls [18]. In animal models, intracerebral injection of MIP-1α resulted in impaired spatial memory and long-term memory and reduced basal synaptic transmission that was reversed with the injection of an antagonist [19].

A2M is a neuronal injury marker and an acute phase protein a major component of the innate immune system and functions as a pan-protease inhibitor and is a chaperone protein. Studies suggest that it may be an early marker for the development of cognitive dysfunction. Indeed, in an analysis of both the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and the Predictors of Cognitive Decline among normal individuals (BIOCARD) study samples, males with higher plasma levels of A2M at baseline were at higher risk for the development of mild cognitive impairment or Alzheimer’s disease. In the BIOCARD study, A2M concentrations in blood were also reported to be correlated with cerebrospinal fluid concentrations of the neuronal injury markers tau and phosphorylated tau [20]. Data from 293 cognitively intact participants of the Health and Aging Brain among Latino Elders (HABLE) study demonstrated an association between higher levels of A2M and subjective cognitive decline, a possible marker of pre-clinical AD [21].

Strengths of these analyses include the large sample size, the inclusion of different ethnicities, the standardized collection, processing, and storage of serum, and the large number of biomarkers assessed. They are limited by the fact that the MMSE (an excellent tool for detecting dementia) has only modest sensitivity and specificity for cognitive impairment [22 –24]. This, and the absence of a follow-up MMSE in 527 (6%) of the participants, means that the incidence of cognitive dysfunction was likely an underestimate [25]. They are also limited by the fact that the participants were generally cognitively intact, with a baseline median MMSE score of 29. Whether similar findings would have come from an older, less cognitively intact population is unknown. Finally, the MMSE variability introduced by the many trial sites and varying conditions under which it was administered may have reduced its ability to clearly detect cognitive dysfunction.

These findings link MIP-1α and A2M plus several additional biomarkers to cognitive dysfunction in people with dysglycemia. Whether they have pathophysiological significance remains unknown.

Footnotes

ACKNOWLEDGMENTS

This work was supported by Sanofi. Clinicaltrials.gov: NCT00069784

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article:

References

Cukierman

, Gerstein

, Williamson

(2005) Cognitive decline and dementia in diabetes–systematic overview of prospective observational studies. Diabetologia 48, 2460–2469.

Reijmer

, van den

, Ruis

, Kappelle

, Biessels

(2010) Cognitive dysfunction in patients with type 2 diabetes. Diabetes Metab Res Rev 26, 507–519.

Biessels

(2013) Sweet memories: 20 years of progress in research on cognitive functioning in diabetes. Eur J Pharmacol 719, 153–160.

National Center for Chronic Disease Prevention & Health Promotion, Division of Diabetes Translation (2017) National Diabetes Statistics Report, 2017. Estimates of Dia-betes and its burden in the United States.

Alzheimer’s Association (2014) 2014 Alzheimer’s disease facts and figures. Alzheimers Dementia 10, e47–e92.

, Qiu

, Wahlin

, Winblad

, Fratiglioni

(2004) Diabetes mellitus and risk of dementia in the Kungsholmen project: a 6-year follow-up study. Neurology 63, 1181–1186.

Umegaki

, Iimuro

, Shinozaki

, Araki

, Sakurai

, Iijima

, Ohashi

, Ito

, Japanese Elderly Diabetes Intervention Trial Study Group (2012) Risk factors associated with cognitive decline in the elderly with type 2 diabetes: pooled logistic analysis of a 6-year observation in the Japanese Elderly Diabetes Intervention Trial. Geriatr Gerontol Int 12 (Suppl 1), 110–116.

Feinkohl

, Keller

, Robertson

, Morling

, Williamson

, Nee

, McLachlan

, Sattar

, Welsh

, Reynolds

, Russ

, Deary

, Strachan

, Price

, Edinburgh Type 2 Diabetes Study (ET2DS) Investigators (2013) Clinical and subclinical macrovascular disease as predictors of cognitive decline in older patients with Edinburgh type 2 diabetes: the Type 2 Diabetes Study. Diabetes Care 36, 2779–2786.

Gerstein

, Pare

, McQueen

, Haenel

, Lee

, Pogue

, Maggioni

, Yusuf

, Hess

, Outcome Reduction With Initial Glargine Intervention Trial Investigators (2015) Identifying novel biomarkers for cardiovascular events or death in people with dysglycemia. Circulation 132, 2297–2304.

10.

Investigators

, Gerstein

, Bosch

, Dagenais

, Diaz

, Jung

, Maggioni

, Pogue

, Probstfield

, Ramachandran

, Riddle

, Ryden

, Yusuf

(2012) Basal insulin and cardiovascular and other outcomes in dysglycemia. N Engl J Med 367, 319–328.

11.

Cukierman-Yaffe

, Bosch

, Diaz

, Dyal

, Hancu

, Hildebrandt

, Lanas

, Lewis

, Marre

, Yale

, Yusuf

, Gerstein

, Investigators

(2014) Effects of basal insulin glargine and omega-3 fatty acid on cognitive decline and probable cognitive impairment in people with dysglycaemia: a substudy of the ORIGIN trial. Lancet Diabetes Endocrinol 2, 562–572.

12.

Tombaugh

, McIntyre

(1992) The mini-mental state examination: a comprehensive review. J Am Geriatr Soc 40, 922–935.

13.

Folstein

, Folstein

, McHugh

(1975/11) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

14.

Exalto

, Biessels

, Karter

, Huang

, Katon

, Minkoff

, Whitmer

(2013) Risk score for prediction of 10 year dementia risk in individuals with type 2 diabetes: a cohort study. Lancet Diabetes Endocrinol 1, 183–190.

15.

Kremers

(2006) Concordance for survival time data including time-dependent covariates accounting for ties in predictor and time. Technical Report Series #80. Mayo Foundation.

16.

Cook

, Ridker

(2009) Advances in measuring the effect of individual predictors of cardiovascular risk: the role of reclassification measures. Ann Intern Med 150, 795–802.

17.

Xia

, Qin

, Wu

, Mackay

, Hyman

(1998) Immunohistochemical study of the beta-chemokine receptors CCR3 and CCR5 and their ligands in normal and Alzheimer’s disease brains. Am J Pathol 153, 31–37.

18.

Man

, Ma

, Shang

, Zhao

, Li

, Guo

, Fang

, Zhu

, Chen

(2007) Peripheral T cells overexpress MIP-1alpha to enhance its transendothelial migration in Alzheimer’s disease. Neurobiol Aging 28, 485–496.

19.

Marciniak

, Faivre

, Dutar

, Alves Pires

, Demeyer

, Caillierez

, Laloux

, Buee

, Blum

, Humez

(2015) The chemokine MIP-1alpha/CCL3 impairs mouse hippocampal synaptic transmission, plasticity and memory. Sci Rep 5, 15862.

20.

Varma

, Varma

, An

, Hohman

, Seddighi

, Casanova

, Beri

, Dammer

, Seyfried

, Pletnikova

, Moghekar

, Wilson

, Lah

, O’Brien

, Levey

, Troncoso

, Albert

, Thambisetty

(2017) Alpha-2 macroglobulin in Alzheimer’s disease: a marker of neuronal injury through the RCAN1 pathway. Mol Psychiatry 22, 13–23.

21.

Hall

, Wiechmann

, Johnson

, Edwards

, O’Bryant

(2019) Levels of alpha-2 macroglobulin in cognitively normal Mexican-Americans with subjective cognitive decline: a HABLE study. Curr Neurobiol 10, 22–25.

22.

Hoops

, Nazem

, Siderowf

, Duda

, Xie

, Stern

, Weintraub

(2009/11/24) Validity of the MoCA and MMSE in the detection of MCI and dementia in Parkinson disease. Neurology 73, 1738–1745.

23.

Arevalo-Rodriguez

, Smailagic

, Roque

IFM

, Ciapponi

, Sanchez-Perez

, Giannakou

, Pedraza

, Bonfill Cosp

, Cullum

(2015) Mini-Mental State Examination (MMSE) for the detection of Alzheimer’s disease and other dementias in people with mild cognitive impairment (MCI). Cochrane Database Syst Rev, CD010783.

24.

Dong

, Lee

, Basri

, Collinson

, Merchant

, Venketasubramanian

, Chen

(2012) The Montreal Cognitive Assessment is superior to the Mini-Mental State Examination in detecting patients at higher risk of dementia. Int Psychogeriatr 24, 1749–1755.

25.

Pendlebury

, Klaus

, Thomson

, Mehta

, Wharton

, Rothwell

, Oxford Vascular

(2015) Methodological Factors in determining risk of dementia after transient ischemic attack and stroke: (III) Applicability of cognitive tests. Stroke 46, 3067–3073.