The risk of dementia in adults with abdominal aortic aneurysm

Abdominal aortic aneurysm (AAA) and dementia have similar epidemiological profiles and common pathogenic mechanisms. However, there have been few studies on the link between these two diseases. For this study, information from 2009 to 2015 was extracted from the Korean National Health Insurance system database. A total of 15,251 participants with a new diagnosis of AAA was included. Propensity score matching by age and sex with patients in whom AAA was not diagnosed was used to select the control group of 45,753 participants. The primary endpoint of this study was newly diagnosed dementia (Alzheimer’s disease (AD), vascular dementia (VD), or other type of dementia). The incidence of dementia was 23.084 per 1000 person years in the AAA group, which was higher than that of the control group (15.438 per 1000 person years). When divided into AD and VD groups, the incidence of AD was higher than that of VD, but the HR of AAA for occurrence of dementia was higher in VD (1.382 vs. 1.784). Among the various risk factors, there was an interaction of age, hypertension, and history of cardiovascular disease with incidence of dementia (p < 0.05). In the presence of hypertension, the HR for occurrence of dementia was high according to presence or absence of AAA (1.474 vs 1.165). In addition, this study showed higher HR in the younger age group (age < 65) and in the group with no history of cardiovascular disease [1.659 vs. 1.403 (age), 1.521 vs. 1.255 (history of cardiovascular disease)]. AAA was associated with increased risk of dementia regardless of AD or VD, even after adjusting for several comorbidities. These findings indicate that follow-up with AAA patients is necessary for early detection of signs and symptoms of dementia.


Methods
Data source. The Korean National Health Insurance (NHI) system consists of two major health care programs for universal coverage of all residents of Korea: NHI and Medical Aid (MA). Approximately 97% of the population is covered by NHI, and the remaining 3% is covered by MA 8 . The NHIS provides biannual health examinations for all insured Koreans and maintains an extensive data set of Koreans that includes patient demographics, medical treatment and procedures, and disease diagnoses according to the International Classification of Disease, 10th editions (ICD- 10). For this study, information from 2009 to 2015 was extracted from the NHIS database.
Sampled patients. Patients in the database who were newly diagnosed with AAA between 2009 and 2015 were enrolled (n = 45,767). The AAA group was defined as patients diagnosed with AAA codes (I71.3-I71.6, I71.8, and I71.9) more than twice at an outpatient department in the past year, who had been hospitalized with one of these AAA codes more than once, or who had undergone aneurysm repair surgery such as open surgical aneurysm repair (OSAR) or endovascular aneurysm repair (EVAR) defined by with the codes (O0223, O0224, O0234, M6611, and M6612). We excluded patients who had not undergone a health examination within 2 years prior to diagnosis of AAA (n = 26,121), were younger than 40 years (n = 431), had data missing (n = 250), or had a previous diagnosis of dementia (n = 1024). Among the remainder, a group of patients who were followed for more than one year was selected (n = 15,251). To select the control group, with a number of participants three times larger than that of the patient group (n = 45,753), we used propensity score matching by age and sex with patients in whom AAA was not diagnosed (Fig. 1). The primary endpoint of this study was newly diagnosed dementia (AD, VD, or other type of dementia).
Data collection and definition. We collected baseline data from the NHIS database of age, sex, smoking status, alcohol consumption, physical activity, waist circumference, body mass index (BMI), and income level. Data about comorbidities, including hypertension, diabetes, dyslipidemia, chronic kidney disease (CKD),  www.nature.com/scientificreports/ and history of cardiocerebrovascular disease (CVD), also were collected. Smoking status was classified as nonsmoker, ex-smoker, and current smoker. Alcohol consumption was classified as heavy drinker when the average daily alcohol intake was 30 g or more, mild to moderate drinker when the individual consumed less than 30 g daily, and non-drinker. Physical activity was divided into two groups, one performing moderate exercise more than 5 days per week or vigorous exercise more than 3 days per week; and all others were classified into the second group. Income level was classified into a group that received medical aid and was in the bottom 20% of income and a group of all others.

Results
Baseline characteristics according to presence of AAA . The characteristics of subjects by presence of AAA are presented in Table 1. Hypertension, and dyslipidemia are known risk factors for AAA and were more common in the AAA group. Also, there were more smokers in the AAA group, and BMI and abdominal circumference were higher in the AAA group. In addition, history of CVD was significantly higher in the AAA group, but alcohol consumption and exercise level were lower in the AAA group.
Incidence of dementia. The incidence of dementia was 23.084 per 1000 person years in the AAA group, which was higher than that of the control group. When all models were applied respectively, the incidence rate remained significantly higher in the AAA group; and the hazard ratio was 1.516, 1.625, 1.568, and 1.422 for models 1 through 4, respectively. When divided into AD and VD groups, the incidence of AD was higher than that of VD, but the HR of AAA for dementia was higher in the VD group (1.382 vs. 1.784 in model 4). ( Table 2).
The Kaplan-Meier plot shows a comparison with the control group. All types of dementia occurred more frequently in the AAA group, which also was observed when divided into AD and VD groups (log rank test p value < 0.001) (Fig. 2).
Interactions with AAA on occurrence of dementia. Among the factors, there was interaction of age, hypertension, and history of cardiovascular disease with AAA (p < 0.05). In the presence of hypertension, the HR for occurrence of dementia was high according to presence or absence of AAA (1.474 vs 1.165). These results were observed in the younger age group and in the group with no history of cardiovascular disease [1.659 vs. 1.403 (age), 1.521 vs. 1.255 (history of cardiovascular disease)] (Supplemental Table 1).

Discussion
The current social expenditure of dementia is US$ 604 billion/year worldwide, and it is projected that the number of dementia patients will triple by 2050. Therefore, it is possible to estimate the global socio-economic impact of dementia 6,9 . For treatment of dementia, a variety of pharmacologic and non-pharmacologic approaches has been attempted, but their efficacy has not been confirmed. Rather, treatment is focused on symptom control, and prevention is considered the top priority 1,2,6,10 . A recent review by Livingston and coworkers suggests that approximately 35% of dementias is attributable to nine modifiable risk factors 2 . To prevent dementia, it is necessary to manage these risk factors including diabetes, hypertension, smoking, and obesity.
These risk factors are also related to AAA in various mechanisms and there are both commonalities and differences in the mechanisms by which the factors act on the disease [11][12][13][14][15][16][17][18][19][20] . The relationship between dementia and AAA was verified through this study, which might be due to the association of these risk factors. Although this effect cannot be completely offset, it was controlled as much as possible by analyzing through model that www.nature.com/scientificreports/ was adjusted using various variables including these risk factors, and there was a relationship between AAA and dementia. Since the association between these two diseases has not been revealed, we would like to propose our hypothesis through a pathophysiological approach. The first such association is related to atherosclerosis. Traditionally, it was thought that AAA develops as a pathological response to aortic atherosclerosis. Until half a century ago, the term "atherosclerotic aneurysm" was used. This view still is favored by some researchers 21,22 . Atherosclerotic plaque growth leads to a compensatory arterial response. In other words, due to arterial narrowing, hemodynamic changes such as those in shear stress occur. The endothelium detects this and changes the phenotype of vascular smooth muscle cells, allowing remodeling through secretion of proteolytic enzymes such as metalloproteinase (MMP) [21][22][23] . Atherosclerosis typically is widespread throughout the vasculature, and the carotid artery is not exempt 24 . Carotid atherosclerosis www.nature.com/scientificreports/ and stiffness consequently cause brain microcirculation transformation and increase blood-brain barrier (BBB) permeability, leading to cognitive impairment 25 . Similarly, associations between descending thoracic aortic plaque and acceleration of brain aging have been demonstrated, and trends in accelerated progression of brain atrophy and progression of cerebrovascular lesions have been reported 26,27 . www.nature.com/scientificreports/ The second association is inflammation and matrix degradation. Aortic inflammation is believed to lead to destruction of aortic media and vascular smooth muscle cell apoptosis and dysfunction. Release of a range of proteolytic enzymes, such as MMP and cysteine proteases, produces reactive oxygen, cytokines, and related products 21 . The tissue inhibitors of MMP (TIMP) are increased in the wall of the aneurysm 28 ; however, the balance between TIMP and MMP seems to favor proteolysis 29 , and elastin and collagen fibers are degraded. Also, due to the paracrine effect of VSMC, a protective effect is needed to maintain homeostasis from proteolysis and inflammatory reactions 30 . However, VSMC apoptosis prevents this protective effect. These MMPs (MMP-2) induce breakdown of the BBB, disrupt oxidative homeostasis in AD 31 , and play a role in the impaired Aβ peptide metabolism responsible for progression of dementia 32 . Reactive oxygen species (ROS) affect AD by inducing oxidation of lipids, proteins, and nucleic acids and impairing Aβ clearance by the low density lipoprotein receptor-related protein (LPR1) through its oxidation 33 .
The hazard ratio of AAA for dementia in this study was 1.422, and the reason for this was deduced based on the mechanisms of AAA occurrence. However, since this deduction is only a possibility, further research is needed. In addition, we examined whether there was a difference in dementia incidence according to the treatment method of AAA, that is, the difference between the surgical group and non-surgical group or the OSAR group and EVAR group, but there were no significant differences. (Tables 3 and 4).
Also, as documented in Supplemental Table 1, there were interactions of age, hypertension, and history of cardiovascular disease with AAA. The HR of dementia was high according to the presence or absence of AAA in the presence of hypertension, in the younger age group, and in the group with no history of cardiovascular disease. In this study, all of these factors were related to dementia, but there were differences in the directions of the interactions. Synergy was present with hypertension because it uniquely influenced dementia with AAA.  www.nature.com/scientificreports/ However, other factors influenced the occurrence of dementia in various ways, indicating a common denominator with AAA. As a result, the HR of AAA for dementia was decreased statistically. Based on the results revealed in this study, the relationship between AAA and Dementia should be confirmed with a prospective controlled study in the future, and it should be studied what mechanism such a relationship is caused by. In addition, although it could not be conducted in this study, it seems necessary to study whether there is a difference in the incidence of dementia in ruptured AAA and unruptured AAA patient groups.
The present study has several limitations. First, this study was a retrospective analysis. To overcome this, the influence of each variable was adjusted through a multivariate logistic regression model. However, control of the confounders among variables was not achieved. Second, diagnosis of AAA and dementia and identification of other risk factors solely were based on diagnostic codes, possibly introducing bias. For example, in the case of dementia, there is no way to confirm whether it is an accurate diagnosis made by neurologist. Third, considering the time of onset, a rough causal relationship might be inaccurately inferred because of the it is a retrospective nature of the study. Also, the degree of cognitive impairment and issues such as AAA diameter are not known, so the correlation between these could not be demonstrated clearly. Finally, due to the lack of data, the genetic factor could not be considered.
However, our study also has a number of strengths. First, we believe that this is the first study to not only demonstrate the relationship between AAA and dementia, but also to illustrate the effect of AAA on dementia in relation to each risk factor. Second, this study used a large, national sample with a relatively long follow-up period.