Different Effects of Metabolic Syndrome on Dementia According to Dementia Type: Analysis Based on the National Health Insurance Service Database of Gangwon Province in South Korea

Abstract

This study was approved by the Institutional Review Board of Chuncheon Sacred Heart Hospital, and all methods were performed in accordance with the approved guidelines and regulations.

De nition of dementia
For de ning dementia of the Alzheimer type (AD), the code of F00 or G30 was included but F01, F02, F03, F051, and G31 were excluded. For the de nition of vascular dementia (VD), codes of F02 were included.
De nition of Metabolic syndrome Individuals who met three or more of the ve components were de ned as having metabolic syndrome.
Five components were abdominal obesity, high TG level, reduced HDL-C level, elevated BP, and elevated blood glucose. [12] Individuals having abdominal obesity were de ned if waist circumference were over 90 cm in males and 80 cm in the female. Individuals having high TG level were de ned if the serum TG level was over 150 mg/dL. Individuals having low HDL-C level were de ned if serum HDL-C level was lower than 40 mg/dL in male and 50 mg/dL in the female. Individuals were de ned as having elevated blood pressure if anti-hypertensive medications were prescribed or systolic blood pressure more than 130 mmHg and/or diastolic blood pressure more than 85 mmHg was recorded. Individuals having high blood glucose were de ned if anti-diabetic drugs (insulins, sulfonylureas, metformin, meglitinides, thiazolidinediones, dipeptidyl peptidase-4 inhibitors, and α-glucosidase inhibitors) were prescribed or fasting serum glucose level was over 100 mg/dL.

Statistical analysis
The baseline characteristics based on the data from the NHIS database are presented as mean values ± standard deviation (SD) for continuous variables and percentages for categorical variables. Differences between the metabolic syndrome group and non-metabolic syndrome group were con rmed using the Student t-test for continuous variables and chi-square tests for categorical variables. The relationship of metabolic syndrome and each component of metabolic syndrome for dementia was evaluated using multiple logistic regression analysis. We ran three regression models. In model 1, we performed multiple logistic regression analysis with metabolic syndrome or each of the components as determinant and AD or VD as outcome variables after controlling for age and gender. In model 2, we performed multiple logistic regression analysis with age, gender, smoking, alcohol, physical inactivity, and metabolic syndrome or ve metabolic syndrome components (high TG, high BP, high blood glucose, abdominal obesity, and low HDL-C). In model 3, we performed multiple logistic regression analysis with age, gender, smoking, alcohol, physical inactivity, previous stroke, previous cardiac disease and metabolic syndrome or ve metabolic syndrome components (high TG, high BP, high glucose, abdominal obesity, and low HDL-C). Again, the relationship of other vascular risk factors for dementia was evaluated using multiple logistic regression analysis. We also ran three regression models in this analysis. In model 1, we performed multiple logistic regression analysis with vascular risk factors as determinant and AD or VD as outcome variables after controlling for age and gender. In model 2, we performed multiple logistic regression analysis with age, gender, smoking, metabolic syndrome, BMI, systolic BP, diastolic BP, fasting glucose, total cholesterol. In model 3, we performed multiple logistic regression analysis with age, gender, smoking, metabolic syndrome, BMI, systolic BP, diastolic BP, fasting glucose, total cholesterol, smoking, alcohol, physical inactivity, previous stroke, and previous cardiac disease. Disease risks were expressed as the odds ratio (OR) with 95% con dence interval (95% CI). We de ned statistical signi cance as p < 0.05. Statistical analyses were conducted with SPSS version 25 software (SPSS Inc., Chicago, IL, USA).

Data availability
The original anonymized data used in this analysis was obtained from the NHIS of South Korea. The dataset from NHIS is not publicly available due to restricted access. However, any researcher requiring access to the data can obtain it directly through a license agreement, including the payment of appropriate license fees.

Demographics and baseline characteristics
Detailed demographic and clinical characteristics of the participants are presented in Table 1. 40.2% of participants had metabolic syndrome. The mean age of the metabolic syndrome group was higher than the non-metabolic syndrome group. The metabolic syndrome group had a higher proportion of females than the non-metabolic syndrome group. The non-metabolic syndrome group had a higher level of physical inactivity, however, they also more engaged in smoking and alcohol intake. The mortality rate of both groups was similar.
After eight years, of the 33,828 patients with metabolic syndrome, 1,380 patients converted to AD and 335 patients converted to VD. Of the 50,316 patients with the non-metabolic syndrome, 176 patients converted to AD and 380 patients converted to VD. Association between metabolic syndrome and AD  Association between metabolic syndrome and VD Table 3 shows the risk of VD according to the metabolic syndrome and component of the de nition of metabolic syndrome.
Metabolic syndrome was not associated with VD (OR 1.173, 95% CI 0.940, 1.465, p = 0.158) Among ve components of metabolic syndrome, high glucose was associated with VD independently. Patients with high glucose had 1.255-fold odds of VD (OR 1.255, 95% CI 1.008, 1.561, p = 0.042). High TG, high BP, obesity and low HDL-C were not associated with VD.

Discussion
We investigated how metabolic syndrome and related vascular risk factors are associated with the development of dementia after eight years. Metabolic syndrome was associated with AD, while it was not associated with VD. Rather than that, high fasting glucose and previous stroke history were found associated with VD. BMI, fasting glucose, smoking, and previous stroke were also found associated with AD.
The odds of metabolic syndrome were much greater than the sum of the odds of ve components in AD.
This might mean that the effect of metabolic syndrome, where risk factors coexist, on AD was greater than the sum of the effect of each component. Metabolic syndrome represented a chronic state of in ammation, hyperinsulinemia, dyslipidemia, dysglycemia, vascular injury and oxidative stress, which were linked to AD. [13] Amyloid beta (Aβ) deposition initiated an immune response, which was intended to clear the amyloid plaque. [14] However, this response also stimulated the cytokine cascade and reactive oxygen species (ROS), leading to neurodegeneration. [15] Furthermore, in ammatory cascade altered phosphorylation of tau protein along with the oxidative injury to the neurons. The immune response induced by amyloid, in addition to the chronic in ammatory state due to metabolic syndrome, exacerbated the AD pathogenic process, leading to a su cient etiological factor to progress AD. [16,17] Metabolic syndrome also induced oxidative stress and increased ROS production. The circulating lipid and glucose imbalance combined with ROS enhances lipoperoxidation, leading to the dysfunction of the antioxidant system. This causes vascular injury and blood-brain barrier (BBB) dysfunction, affecting amyloid and tau accumulation and chronic hypoperfusion, and leads to neuronal damage. [13] Among components of metabolic syndrome, only high glucose level was found associated with both AD and VD, while others show association only with AD. Many previous studies have reported that diabetes mellitus not only affects cognitive decline but is also associated with a higher risk of cerebrovascular disease including high white matter hyperintensities volume, and cerebral infarcts. Impaired insulin signalling and glucose metabolism in the brain are factors that are related to AD pathogenesis. Insulin modulates Aβ protein precursor expression and processing. [18] Insulin not only regulates glucose and lipid metabolism in the brain but also regulates neural development and neuronal activities, associated with learning and memory. [19] Insulin receptors are expressed in the brain, particularly in memory registration-related areas, such as the cerebral cortex, hippocampus, hypothalamus, and amygdala. [20] Therefore, impairment of the insulin signalling leads to the pathologic processes of AD. In addition, we assumed that glucose metabolism would be more closely related to the mechanism that caused cognitive impairment after the vascular event. Although previous studies have shown that other vascular risk factors, such as hypertension [21], dyslipidemia [22], and abdominal obesity [23], also increase ischemic stroke. Our study showed that these factors were not associated with VD. Similar to our study, another study showed that high glucose was the most signi cant component that was associated with cognitive impairment. [24] The previous stroke is also a factor that affects both AD and VD. The association between stroke and dementia has already been reported in several studies. Ischemic stroke was a risk factor for developing AD and VD. [25,26] Stroke doubles the risk of dementia and approximately 20% of stroke patients go on to develop cognitive dysfunction within 3 years. [27] Ischemic stroke leads to pathophysiological processes that contribute to ischemic cell damage. Stimulation of the in ammatory process, free radical production, excitotoxicity, disruption of sodium and calcium in ux, enzymatic changes, endothelin release, delayed coagulation, activation of platelets and leukocytes, and endothelial dysfunction are the pathophysiological reactions resulting from the onset of stroke. [28] Several studies reported a synergistic relationship between ischemic stroke and AD. Postmortem studies have shown that individuals with AD pathology with cerebral infarction had a markedly increased risk of dementia compared to those with AD pathology without infarcts. [29,30] Stroke has been suggested as a contributing factor to AD pathological changes including selective brain atrophy and accumulation of abnormal protein such as Aβ. [31] A previous study also provided evidence that stroke leads to cognitive dysfunction more rapidly in patients with AD. [32] Besides, VD is the severest form of vascular cognitive impairment, [8] and it results from the subclinical vascular brain injury and stroke. Not only major stroke but also minor stroke and even transient ischemic attack are known to increase the risk of dementia.
As per our study, patients with higher BMI were less prone to AD. This result was consistent with so-called obesity paradox that high BMI in the elderly is associated with a lower risk of developing AD. Although obesity is linked to cardiovascular disease, the relationship between obesity and dementia is still not clear, especially in late-life. In previous studies, lower BMI was found associated with Aβ deposition, tau accumulation and cognitive decline. [33][34][35] It was also reported that high levels of leptin, a hormone synthesized by adipocytes, enhanced memory function. [36] In the present study, however, patients with obesity were diagnosed with more AD than patients without obesity. Therefore, the trend that patients with higher BMI were less likely to have AD might be because patients with lower BMI had a higher risk of AD.
AD was also found associated with smoking. It is controversial whether smoking has any harmful effect on degenerative diseases. Previously some studies reported that smoking had a protective effect on degenerative diseases. [37][38][39] There are even researches which have shown that nicotine has a neuroprotective and anti-ageing effect. [40,41] However, nowadays, smoking has been attributed as a risk factor for Alzheimer's disease. [42] Smoking increases oxidative stress and might have indirect effects on several vascular, in ammatory and degenerative processes. [43,44] If smoking particles are inhaled, they stimulate ROS production and enters the brain via blood. Smoking-related cerebral oxidative stress is a potential mechanism to accelerate AD pathology and increase risk for AD. [45] Smoking also impairs nitric oxide synthesis in cerebral vascular endothelial cells leading to interference with cerebral blood ow and glucose metabolism in the brain. It induces cerebral hypoperfusion and promotes the synthesis of Aβ. [46] Smoking stimulates the release of proin ammatory cytokines and immune system-mediated products causing an increase in Aβ accumulation and tau phosphorylation, hallmarks of the AD pathology. [47] Our study has some limitations. First, because our study used claim data, we classi ed the types of dementia, but the actual amyloid burden was unknown. However, in order to distinguish the pure dementia type, those with codes of both AD and VD were excluded. Second, since we only investigated the presence of dementia eight years later, there was no consideration what happened during the eight years and when dementia occurred during the eight years. Besides, there was no consideration for efforts to overcome metabolic syndrome. Some might have tried to treat metabolic syndrome, others might not, but the effect of improvement of metabolic syndrome was not known in our study. Third, it is also a cross-sectional study, there was no consideration for the duration of metabolic syndrome. The association with dementia might vary depending on how long components of metabolic syndrome have been existed, which were not known in our article. Therefore, further longitudinal study is required. Finally, as the cohort only included the speci c region of Korea, further investigation in other region or population is needed to generalize these ndings.
Despite these limitations, we investigated the effects of metabolic syndrome in different types of dementia and the association between vascular risk factors related to metabolic syndrome and dementia using population-based data. Therefore, we hope to provide new clinical insight into the association between metabolic syndrome and type of dementia with implications for considering different pathophysiology.

Conclusions
Metabolic syndrome was associated with AD, while it was not associated with VD. Therefore, the effect of metabolic syndrome on dementia would be different depending on the type of dementia.  Figure 1 Flow chart of the study population