Advanced Glycation End Product Accumulation is Associated with Lower Cognitive Performance in an Older General Population: The Nagahama Study

Abstract

Accumulation of advanced glycation end products (AGEs) has been linked with cognitive decline as a risk factor based on the analysis in small populations. We investigated the association between skin autofluorescence of AGEs and global cognitive function in a Japanese older (≥60 years) population (n = 4,041). The AGEs quartiles were inversely associated with the Revised Hasegawa’s Dementia Scale score (Q1: reference, Q2: β= –0.011, p = 0.537, Q3: β= –0.043, p = 0.016, Q4: β= –0.064, p < 0.001) independent of major risk factors. Accumulation of AGEs was associated with lower cognitive performance in older adults.

Keywords

Advanced glycation end products cognitive function epidemiological studies population at risk

INTRODUCTION

Advanced glycation end products (AGEs) include a number of heterogeneous molecules that are non-enzymatically generated by glycation of proteins and fats [1]. Because AGEs form as covalent cross-links between proteins, extracellular matrices with long-lived protein, such as skeletal muscle, are highly susceptible to AGE formation [1]. Several cross-sectional studies [2, 3], including our previous study [4], found an inverse association between AGEs and weak grip strength [2], slow gait speed [3], and lower skeletal muscle mass [4].

AGEs may also be involved in the worsening of cognitive function [5]. An experimental study in mice reported increased AGEs accumulation in the brain and cognitive functional decline by oral supplementation of an AGE molecule [6], in which mechanism might be decreased expression levels of survival factor sirtuin-1 in brains. In humans, several immunohistochemical studies observed AGEs accumulation in the brain [7 –9] and suggested an involvement of AGEs accumulation in vascular [8] and Alzheimer’s type [9] dementia. The harmful effect of AGEs on cognitive function was also found in observational studies; urinary level of pentosidine, one of major fluorescent AGEs, was associated with longitudinal cognitive decline in older patients without dementia [10]. Decreased circulating levels of soluble receptor for AGEs in patients with mild cognitive decline [11] and dementia [12] also supported the association between AGEs and cognitive function. Accumulation of AGEs can also be assessed by skin autofluorescence technique (SAF-AGE) of which accuracy has been verified by comparing to the tissue levels of pentosidine and carboxymethyl-lysine, other major AGEs [13]. If easily measurable SAF-AGE was associated with cognitive function in a general population, it may be useful tool for risk assessment of cognitive decline.

In this cross-sectional study, we aimed to investigate the association between SAF-AGE and cognitive function by analyzing a dataset from the Nagahama study, a community-based cohort study, where data of SAF-AGE and cognitive function was available for analysis.

METHODS

Study participants

We analyzed a dataset of the follow-up investigation of the Nagahama study, which was conducted between 2013 and 2016. Participants in the Nagahama study were recruited between 2008 and 2010 from the general population of Nagahama, a rural city located in central Japan. Community residents, aged between 30 and 74 years, who were living independently without physical impairment or dysfunction, were eligible to participate. The baseline population comprised 9,764 individuals, which corresponded to approximately 14% of total residents at the age group.

Among the baseline population, 8,289 individuals participated in the second health survey performed 5 years after the baseline evaluation. By further recruiting 1,561 participants meeting the inclusion criteria, the dataset of the follow-up investigation of the Nagahama study comprised 9,850 participants. Participants recruitment details are listed elsewhere [14].

Of the 9,850 participants in the follow-up investigation, a total of 5,018 participants were aged 60 years or older. Among them, 4,041 participants were ultimately included in the analysis after excluding participants who met any of the following exclusion criteria: no assessment of cognitive function performed during 2013 and 2016 (n = 637), which was an optional examination for participants who agreed to perform the test, pacemaker implantation (n = 10), receiving hemodialysis therapy or severe renal functional decline (estimated glomerular filtration rate [eGFR] <30 ml/min/1.73 m² or urinary albumin ≥300 mg/day) (n = 64), no SAF-AGE measurement (n = 258), and a wide deviation or lack of clinical values required for the present study (n = 8).

All study procedures were approved by the ethics committee of Kyoto University Graduate School of Medicine and by the Nagahama Municipal Review Board. Written informed consent was obtained from all participants.

Measurement of SAF-AGE

SAF-AGEs were measured on the middle finger of the non-dominant hand using a prototype of the AGE sensor RQ-AG01J (SHARP Life science Co., Kobe, Japan). The fluorescence emission spectrum (440–460 nm) excited by a light-emitting diode (365 nm) was measured on a skin surface area approximately 0.38 mm in diameter using a 2,048-pixel charge-coupled device linear image sensor [15]. Measurements were taken in triplicate, and the mean value was calculated for the analysis. The coefficient of variation and intraclass correlation coefficient of the repeated measurements were 6.65±7.25% and 0.938, respectively.

Assessment of cognitive function

Cognitive function was assessed using the Revised Hasegawa’s Dementia Scale (HDS-R), which consists of 9 questions regarding orientation, registration, attention and calculation, recall, and language [16]. The HDS-R score ranges from 0 to 30 points, and ≤20 points are considered as the appropriate cut-off point to discriminate possible dementia. The correlation coefficient between HDS-R and Mini-Mental State Examination scores were reported as 0.94 [17]. Diagnostic accuracy of HDS-R for Alzheimer-type dementia and its mild type was reported to be comparable with that of the Mini-Mental State Examination [17].

Basic clinical characteristics

Basic clinical characteristics were obtained at the follow-up investigation. Smoking habits, alcohol consumption, education attainment, medication use, and a history of cardiovascular disease (symptomatic stroke or ischemic heart disease), were determined using a structured, self-administered questionnaire. Hypertension was defined as systolic blood pressure (BP) ≥140 mm Hg, diastolic BP≥90 mm Hg, and/or prescribed antihypertensive drugs. Diabetes was defined as glucose ≥126 mg/dl (fasting) or ≥200 mg/dl (non-fasting), hemoglobin A1c≥6.5%, or antihyperglycemic treatment. Renal function was assessed by eGFR calculated using the following formula: 194×serum creatinine^–1.094×age^–0.287 (×0.739 if female).

Statistical analysis

Values are presented as mean±standard deviation or frequency. The SAF-AGE quartiles calculations were stratified sex and measurement periods to avoid potential confounding. Group differences in numeric variables were assessed by analysis of variance, while the frequency difference was assessed using a chi-squared test. Multiple linear or logistic regression analyses were used to identify factors independently associated with the HDS-R score. Adjusted factors in the regression model are shown in the legend of Table 1. Statistical analyses were performed using the JMP 14.3.0 software (SAS Institute, NC, USA). p < 0.05 was considered statistically significant.

Table 1

Clinical characteristics of study participants (n = 4,041)

Age (y)	68.3±5.4
Sex (men, n (%))	1,462 (36.2)
BMI (kg/m²)	22.5±3.1
Smoking (current/past/never, %)	7.6/23.3/69.1
Alcohol consumption (Go/week)	3.7±7.4
Education attainment (n, (%))
≤9 y	1,201 (29.7)
≤12 y	1,856 (46.0)
>12 y	984 (24.3)
Cardiovascular disease (n, (%))	274 (6.8)
Systolic BP (mmHg)	131±17
Diastolic BP (mmHg)	73±10
Antihypertensive treatment (n, (%))	1,551 (38.4)
Hypertension (n, (%))	2,181 (54.0)
Glucose (mg/dl)	91±15
HbA1c (%)	5.7±0.5
Antihyperglycemic treatment (n, (%))	306 (7.6)
Diabetes (n, (%))	426 (10.5)
Creatine (mg/dl)	0.7±0.2
eGFR (ml/min/1.73 m²)	72.1±13.2
HDS-R score	28±2

Values are mean±standard deviation or a frequency. Alcohol consumption was described in Japanese traditional units of alcohol (Go), where 1 Go corresponds to 22 g of ethanol. Nine education years corresponded to a junior high school graduate, whereas 12 education years corresponded to a high school graduate. Cardiovascular disease includes symptomatic stroke and ischemic heart disease. Hypertension was defined as systolic blood pressure (BP) ≥140 mm Hg, diastolic BP≥90 mm Hg, and/or prescribed antihypertensive drugs. Diabetes was defined as glucose ≥126 mg/dl (fasting) or ≥200 mg/dl (non-fasting), glycated hemoglobin A1c (HbA1c) ≥6.5%, or antihyperglycemic treatment. Estimated glomerular filtration rate (eGFR) was calculated using the following formula: 194×serum creatinine^–1.094×age^–0.287 (×0.739 if female). BMI, body mass index; HDS-R, Revised Hasegawa’s Dementia Scale.

RESULTS

The clinical characteristics of the study participants are summarized in Table 1. There were no marked differences in the clinical characteristics between the study population and the total population (Supplementary Table 1).

Thirty-two participants had HDS-R score ≤20 points, while a total of 829 participants had ≤26 points—corresponding to lower 20 percentiles of the study participants. Quartiles of SAF-AGE were inversely associated with the HDS-R score (Fig. 1). Although participants with low HDS-R score were older (≤26 points: 70.6±5.4, ≥27 points: 67.7±5.3 years old, p < 0.001), more of male (≤26 points: 55.1%, ≥27 points: 31.3%, p < 0.001), and had lower educational attainment (≤9 years; ≤26 points: 42.9%, ≥27 points: 26.3%, p < 0.001), results of the linear regression analysis adjusted for these covariates identified SAF-AGE quartiles as an independent inverse determinant of HDS-R score (Table 2), even when hemoglobin A1c (HbA1c) was adjusted instead of diabetes (Q4: β= –0.065, p < 0.001). Furthermore, the association of SAF-AGE remained significant when it was included in the regression model as a continuous variable (β= –0.064, p < 0.001).

Fig.1

Frequency differences in SAF-AGE quartiles by HDS-R score. Quartiles of skin autofluorescence of advanced glycation end product (SAF-AGE) was calculated separately by sex and measurement periods to avoid potential stratification. HDS-R indicates the Revised Hasegawa’s Dementia Scale. Statistical significance was assessed by the chi-squared test (p < 0.001).

Table 2

Liner regression analysis for HDS-R score

	Total N = 4,041		<68 years old N = 1,969		≥68 years old N = 2,072
	β	p	β	p	β	p
Age (y)	–0.207	<0.001	–0.065	0.004	–0.132	<0.001
Sex (men)	–0.266	<0.001	–0.274	<0.001	–0.274	<0.001
BMI (kg/m²)	0.006	0.713	0.028	0.220	–0.011	0.623
Education attainment
>12 y	reference		reference		reference
≤12 y	–0.030	0.100	–0.046	0.063	–0.023	0.429
≤9 y	–0.132	<0.001	–0.166	<0.001	–0.118	<0.001
Cardiovascular disease	–0.032	0.033	–0.037	0.091	–0.031	0.145
Current or past smoking	0.035	0.099	0.078	0.012	0.012	0.703
Alcohol consumption (Go/week)	0.017	0.300	–0.037	0.138	0.054	0.025
Hypertension	–0.001	0.927	–0.027	0.239	0.011	0.617
Diabetes	–0.034	0.025	–0.019	0.394	–0.044	0.040
eGFR (ml/min/1.73m²)	–0.035	0.019	–0.015	0.508	–0.051	0.018
SAF-AGE quartiles
Q1	reference		reference		reference
Q2	–0.011	0.537	–0.001	0.955	–0.018	0.493
Q3	–0.043	0.016	–0.019	0.472	–0.064	0.014
Q4	–0.064	<0.001	–0.013	0.628	–0.100	<0.001

Quartiles of skin autofluorescence of advanced glycation end product (SAF-AGE) was calculated separately by sex and measurement periods to avoid potential stratification. Participants were subdivided at the median of age. β, standardized regression coefficient; HDS-R, Revised Hasegawa’s Dementia Scale; BMI, body mass index; eGFR, estimated glomerular filtration rate.

Because individuals with low HDS-R score (≤26 points) were significantly older than other participants (low HDS-R: 70.6±5.4, control: 67.7±5.3 years old, p < 0.001), we then performed age stratified analysis (Table 2). When participants were subdivided at the median of age, the association between the highest quartile of SAF-AGE and HDS-R score was observed only in the older subpopulation. The odds ratio of the highest SAF-AGE quartile for the HDS-R score ≤26 points in the older subpopulation calculated by a logistic regression model including the same covariates was 1.60 (95% CI: 1.20–2.13), p = 0.001.

Endogenous AGEs were suggested to be generated at higher rates in diabetic individuals due to altered glucose metabolism. In this study population, HbA1c did not show independent association with SAF-AGE (p = 0.358) in the analysis adjusted for age, sex, and body mass index. However, in the stratified analysis by the median of HbA1c (5.7%), the association between the highest SAF-AGE quartile and HDS-R score remained significant only in the higher HbA1c subgroup (n = 1,852, β= –0.089, p < 0.001) but not in the lower HbA1c subgroup (n = 2,189, β= –0.039, p = 0.118). The regression coefficient was therefore largest in the older population with higher HbA1c (n = 1,060, β= –0.121, p = 0.001).

DISCUSSION

In this cross-sectional study, we demonstrated that SAF-AGE was an independently inverse determinant of global cognitive function. To our knowledge, this is a novel study that showed the involvement of AGE accumulation in cognitive function in a large-scale community-dwelling older adult population.

Older age, male sex, and shorter years of education were factors strongly associated with low HDS-R score. Diabetes, another known risk factor for dementia, was also inversely associated with the HDS-R score, although the standardized regression coefficient of the highest SAF-AGE quartiles was considerably larger than that of diabetes. Despite that the mechanism of the accumulation of AGEs’ harmful influence on cognitive function is still unclear, we can firmly recognize AGEs as a factor associated with cognitive function. Very recently, ɛ4 genotype of apolipoprotein E was suggested to be associated with AGEs accumulation [18]. Results of the genotype analysis might be a clue to clarify the harmful effects of AGEs.

Dietary intake of AGEs is a major determinant of circulating AGE levels independently of total calories, nutrient consumption, and fat mass measures [19]. Another source of AGEs is endogenous AGEs, which was suggested to be generated at higher rates in diabetic individuals. In this study population, HbA1c did not show independent association with SAF-AGE possibly because of relatively lower glycemic levels than that in patients with diabetes. However, the association between SAF-AGE and HDS-R was observed only in the subpopulation with higher HbA1c, suggesting an existence of shared pathophysiology for cognitive functional decline between worse glycemic control and accumulation of AGEs. Given the growing frequency of dementia, further investigations in individuals with more cognitive impairment including dementia are required to clarify this issue.

We measured SAF-AGE on the middle finger. In our previous study in an independent population [20], we measured SAF-AGE on the inner surface of the forearm, and the SAF-AGE level was inversely associated with mild cognitive decline, independent of subclinical atrophy of the hippocampus. The association of SAF-AGE and cognitive function may not depend on the site of SAF-AGE measurement. Because the SAF-AGE values are usually expressed by arbitrary unit, and varies with measurements methods, standardization of SAF-AGE measurement is required for it to be used in the primary care setting.

A strength of this study is the large sample size with available SAF-AGE values. There were several limitations need to be mentioned. First, we did not assess detailed cognitive function. Although we used HDS-R instead of Mini-Mental State Examination for cognitive assessment, it would not cause serious bias because of a strong correlation between the score points [17, 18]. Second, several participants did not perform the cognitive functional test because the test was an optional examination for participants who agreed to perform it. Although participants with worse cognitive function were likely to refuse the test, this bias may act to weaken the relationship between SAF-AGE and HDS-R score because whether participants agreed to perform the cognitive functional test might be independent of the SAF-AGE level. Thus, the selection bias of study population, if any, might not distort the present findings. Third, the present study was a cross-sectional setting. We therefore could not clarify causality between SAF-AGE and HDS-R score. Further studies with a detailed assessment of cognitive function would strengthen the present findings. Also, longitudinal studies aimed to investigate a possible association between SAF-AGE and cognitive decline would provide a clue to clarify causality between them.

In summary, the SAF-AGE was an independent determinant of global cognitive function in community-dwelling older adults. The SAF-AGE may be a potentially non-invasive measure for the assessment of the potential risk of dementia.

Footnotes

ACKNOWLEDGMENTS

We are extremely grateful to the Nagahama City Office, and the nonprofit organization, Zeroji Club, for their assistance in analyzing the Nagahama study. We also thank the editors of Crimson Interactive Pvt. Ltd. for the English language review.

The Nagahama study was supported by a university grant, the Center of Innovation Program, the Global University Project, and a Grant-in-Aid for Scientific Research (25293141, 26670313, 26293198, 17H04182, 17H04126, 17H04123, 18K18450) from the Ministry of Education, Culture, Sports, Science and Technology of Japan; the Practical Research Project for Rare/Intractable Diseases (ek0109070, ek0109070, ek0109196, ek0109348), the Comprehensive Research on Aging and Health Science Research Grants for Dementia R&D (dk0207006, dk0207027), the Program for an Integrated Database of Clinical and Genomic Information (kk0205008), the Practical Research Project for Lifestyle-related Diseases including Cardiovascular Diseases and Diabetes Mellitus (ek0210066, ek0210096, ek0210116), and the Research Program for Health Behavior Modification by Utilizing IoT (le0110005), from Japan Agency for Medical Research and Development (AMED); the Takeda Medical Research Foundation; Mitsubishi Foundation; Daiwa Securities Health Foundation; and Sumitomo Foundation.

Authors’ disclosures available online ().

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

References

Semba

, Nicklett

, Ferrucci

(2010) Does accumulation of advanced glycation end products contribute to the aging phenotype? J Gerontol A Biol Sci Med Sci 65, 963–975.

Dalal

, Ferrucci

, Sun

, Beck

, Fried

, Semba

(2009) Elevated serum advanced glycation end products and poor grip strength in older community-dwelling women. J Gerontol A Biol Sci Med Sci 64, 132–137.

Semba

, Bandinelli

, Sun

, Guralnik

, Ferrucci

(2010) Relationship of an advanced glycation end product, plasma carboxymethyl-lysine, with slow walking speed in older adults: The InCHIANTI study. Eur J Appl Physiol 108, 191–195.

Tabara

, Ikezoe

, Yamanaka

, Setoh

, Segawa

, Kawaguchi

, Kosugi

, Nakayama

, Ichihashi

, Tsuboyama

, Matsuda

; Nagahama Study Group (2019) Advanced glycation end product accumulation is associated with low skeletal muscle mass, weak muscle strength, and reduced bone density: The Nagahama Study. J Gerontol A Biol Sci Med Sci 74, 1446–1453.

Haddad

, Perrotte

, Landri

, Lepage

, Fülöp

, Ramassamy

(2019) Circulating and extracellular vesicles levels of N-(1-Carboxymethyl)-L-Lysine (CML) differentiate early to moderate Alzheimer’s disease. J Alzheimers Dis 69, 751–762.

Cai

, Uribarri

, Zhu

, Chen

, Swamy

, Zhao

, Grosjean

, Simonaro

, Kuchel

, Schnaider-Beeri

, Woodward

, Striker

, Vlassara

(2014) Oral glycotoxins are a modifiable cause of dementia and the metabolic syndrome in mice and humans. Proc Natl Acad Sci U S A 111, 4940–4945.

Kimura

, Takamatsu

, Ikeda

, Kondo

, Miyakawa

, Horiuchi

(1996) Accumulation of advanced glycation end products of the Maillard reaction with age in human hippocampal neurons. Neurosci Lett 208, 53–56.

Southern

, Williams

, Esiri

(2007) Immunohistochemical study of N-epsilon-carboxymethyl lysine (CML) in human brain: Relation to vascular dementia. BMC Neurol 7, 35.

Valente

, Gella

, Fernàndez-Busquets

, Unzeta

, Durany

(2010) Immunohistochemical analysis of human brain suggests pathological synergism of Alzheimer’s disease and diabetes mellitus. Neurobiol Dis 37, 67–76.

10.

Yaffe

, Lindquist

, Schwartz

, Vitartas

, Vittinghoff

, Satterfield

, Simonsick

, Launer

, Rosano

, Cauley

, Harris

(2011) Advanced glycation end product level, diabetes, and accelerated cognitive aging. Neurology 77, 1351–1356.

11.

Chen

, Cai

, Chen

, Liang

, Lin

, Li

, Lin

, Yao

, Wen

, Huang

(2011) Serum level of endogenous secretory receptor for advanced glycation end products and other factors in type 2 diabetic patients with mild cognitive impairment. Diabetes Care 34, 2586–2590.

12.

Emanuele

, D’Angelo

, Tomaino

, Binetti

, Ghidoni

, Politi

, Bernardi

, Maletta

, Bruni

, Geroldi

(2005) Circulating levels of soluble receptor for advanced glycation end products in Alzheimer disease and vascular dementia. Arch Neurol 62, 1734–1736.

13.

Meerwaldt

, Graaff

, Oomen

PHN

, Links

, Jager

, Alderson

, Thorpe

, Baynes

, Gans

ROB

, Smit

(2004) Simple non-invasive assessment of advanced glycation endproduct accumulation. Diabetologia 47, 1324–1330.

14.

Tabara

, Matsumoto

, Murase

, Nagashima

, Hirai

, Kosugi

, Nakayama

, Wakamura

, Chin

, Matsuda

; and the Nagahama study group (2018) Seasonal variation in nocturnal home blood pressure fall: The Nagahama study. Hypertens Res 41, 198–208.

15.

Yamanaka

, Matsumura

, Ohno

, Fujiwara

, Shinagawa

, Sugawa

, Hatano

, Shirakawa

, Kinoshita

, Ito

, Sakata

, Araki

, Nagai

(2016) Non-invasive measurement of skin autofluorescence to evaluate diabetic complications. J Clin Biochem Nutr 58, 135–140.

16.

Imai

, Hasegawa

(1994) The Revised Hasegawa’s Dementia Scale (HDS-R) – Evaluation of its usefulness as a screening test for dementia. J Hong Kong Coll Psychiatr 4, 20–24.

17.

Kim

, Lee

, Jhoo

, Youn

, Suh

, Jun

, Seo

, Woo

(2005) Diagnostic accuracy of Mini-Mental Status Examination and revised Hasegawa dementia scale for Alzheimer’s disease. Dement Geriatr Cogn Disord 19, 324–330.

18.

Deo

, Dhillon

, Chua

, Thomas

, Fenech

(2019) APOE ɛ4 carriers have a greater propensity to glycation and sRAGE which is further influenced by RAGE G82S polymorphism. J Gerontol A Biol Sci Med Sci, doi: 10.1093/gerona/glz259.

19.

Uribarri

, Cai

, Woodward

, Tripp

, Goldberg

, Pyzik

, Yee

, Tansman

, Chen

, Mani

, Fayad

, Vlassara

(2015) Elevated serum advanced glycation endproducts in obese indicate risk for the metabolic syndrome: A link between healthy and unhealthy obesity? J Clin Endocrinol Metab 100, 1957–1966.

20.

Igase

, Ohara

, Igase

, Kato

, Okada

, Ochi

, Tabara

, Kohara

, Ohyagi

(2017) Skin autofluorescence examination as a diagnostic tool for mild cognitive impairment in healthy people. J Alzheimers Dis 55, 1481–1487.