Correlation between Body Mass Index and Cognitive Function in Patients with Atrial Fibrillation

Background Evidence regarding the relationship between body mass index (BMI) and cognitive function was limited. Therefore, the objective of this research is to investigate whether BMI is independently related to cognitive function in Chinese patients with atrial fibrillation after adjusting for other covariates. Methods The present study is a cross-sectional study. A total of 281 patients with atrial fibrillation who were hospitalized at the Affiliated Hospital of Jining Medical University in Shandong Province from January 2021 to November 2021 were included in the study. The target independent variable and the dependent variable were BMI and cognitive function in patients with atrial fibrillation, respectively. The patients' general information, BMI, past history, medication history, and other disease-related data were collected. The Montreal cognitive assessment scale (MoCA) was used to evaluate cognitive function. Results A total of 244 patients with atrial fibrillation were collected in this study, with an average age of (67.28 ± 10.33) years, of whom 55.3% were male. The average BMI was (25.33 ± 4.27) kg/m2, and the average cognitive function score was (19.25 ± 6.88) points. The results of the smooth curve fitting and threshold effect tests showed that there was a curve correlation between BMI and cognitive function score, and its inflection point was 24.56 kg/m2. To the left of the inflection point, the relationship was significant; the effect size and the confidence interval were 0.43 and 0.01–0.85, respectively. To the right of the inflection point, there was no significant correlation between BMI and cognitive function (P=0.152). Conclusion When BMI is lower than 24.56 kg/m2, the cognitive function score increases by 0.43 points for each unit increase in BMI in patients with atrial fibrillation. An increase in BMI at this time is a protective factor for cognitive function. Within the normal range of BMI, the higher the BMI in atrial fibrillation patients, the higher the cognitive function score. We encourage atrial fibrillation patients with normal BMI to maintain their current weight.


Introduction
Atrial fbrillation is a supraventricular tachyarrhythmia with uncoordinated atrial electrical activation and inefective atrial contractions [1]. Te lifetime risk of atrial fbrillation is 1 in 4 and increases to 1 in 3 in people over 55 years of age [2][3][4][5]. Atrial fbrillation will lead to impaired cognitive function in patients. Alonso et al. [6] showed that the incidence of dementia and mild cognitive impairment in patients with atrial fbrillation was 40%, and the risks was 2.25 times and 1.28 times that of the normal population, respectively. In addition, Chen et al. [7] showed that atrial fbrillation is a risk factor for dementia and cognitive impairment independent of stroke. Cognitive decline or even dementia in patients with atrial fbrillation will seriously reduce the quality of life of patients and increase the burden on family caregivers [8,9]. Studies have shown that obesity may lead to cognitive decline [10][11][12]. But there are also studies suggesting that obesity may be a protective factor for cognitive function [13,14]. At present, no studies have explored the relationship between obesity and cognitive function in patients with atrial fbrillation, so the relationship between the two is unclear. If the correlation between body mass index (BMI) and cognitive function in patients with atrial fbrillation can be clarifed, it can provide a basis for the prevention of cognitive dysfunction in patients with atrial fbrillation. Terefore, our study aimed to describe the association between BMI and cognitive impairment in patients with atrial fbrillation. Te results of this study can help clinicians determine the target value of weight control in patients with atrial fbrillation, and provide evidence for improving the cognitive function of patients with atrial fbrillation.

Study Design.
Tis study is a cross-sectional study. Te purpose of this study was to explore the relationship between BMI and cognitive function scores in patients with atrial fbrillation. Te independent variable was the BMI of patients with atrial fbrillation. Te dependent variable was cognitive function scores in patients with atrial fbrillation.

Participants.
Te convenience sampling method was used to recruit patients with atrial fbrillation who were hospitalized in the Cardiology Department of a tertiary hospital in Shandong Province from January 2021 to November 2021 ( Figure 1). Inclusion criteria were: (1) age ≥18 years; (2) diagnosis of atrial fbrillation after 12-lead ECG and 24 hour Holter monitoring according to the diagnostic criteria of the 2020 ECS/EACTS Guidelines for the Diagnosis and Management of Atrial Fibrillation [15]; (3) signing the informed consent form. Te exclusion criteria were: (1) patients with severe comorbidities could not cooperate to complete the study; (2) patients with hearing or vision impairment were unable to complete the cognitive function assessment. Tis study was approved by the Ethics Committee of the Afliated Hospital of Jining Medical College (ethics batch number: 2020C009).

Variables.
Te covariates collected in this study included two parts: general patient data and clinical data. (1) General information of patients: including the patients' age, sex, marital status, education level, and living conditions. (2) Clinical data of the patients: including the patients' smoking history, drinking history, medication use within 1 month before admission, coronary heart disease history, history of myocardial infarction, history of cerebral infarction, diabetes, hypertension, valvular disease, hyperlipidemia, history of heart failure, systolic blood pressure (SBP), diastolic blood pressure (DBP), D-dimer, left ventricular blood fraction (LVEF), creatinine, free thyroxine, free triiodothyronine, and left atrial diameter. (MoCA). Te scale measures the following aspects: (1) short-term memory; (2) visuospatial ability; (3) executive ability; (4) attention, numeracy, and working memory; (5) language; and (6) orientation. We will add 1 point to the score for less than 12 years of education to correct for the efect of educational attainment (if the total score is less than 30 points) [16,17]. Te full score of MoCA is 30 points. Te higher the score, the better the patient's cognitive function. When the score ≥26 points, the patient's cognitive function is normal; otherwise, the patient will be recognized as having cognitive dysfunction. Te evaluation process takes approximately 10 minutes. Te MoCA had a high sensitivity of 0.94 [18].

Data Collection.
Patients who met the inclusion criteria signed the informed consent form after admission and completed the evaluation using the Montreal Cognitive Function Assessment Scale within 24 hours. Te cognitive function assessment was performed by specially trained nurses. In the process of cognitive function measurement, the patient was guaranteed to be in a calm state. If the patient has a sudden change in condition, the measurement is discontinued. Te nurses collected the general information and disease-related information of the patients. Te measurement of systolic and diastolic blood pressure uses a unifed measurement standard. Te patients were in a supine position when blood pressure was measured. We properly use a small towel to raise the upper arm so that it is at the same level as the heart. We measured blood pressure three times for the patients, and took the average of the three measurements as the fnal blood pressure value. During height and weight measurement, patients took of their coat and shoes and kept only light clothing. BMI was defned as weight in kilograms divided by height in meters squared (kg/ m 2 ). In the morning of the next day after admission, the nurses collected the fasting blood of the patients for the collection of laboratory indicators. Cardiac ultrasonography was performed by a qualifed sonographer, and the patients' left ventricular ejection fraction and left atrial diameter were collected.

Statistical
Analysis. Continuous variables were tested for normality using the K-S method. Normally distributed variables were expressed as mean ± standard deviation (x ± s), and one-way ANOVA was used. Non-normally distributed variables were expressed as medians (minimum, maximum), and the Kruskal-Wallis H test was used. Categorical variables were expressed as frequencies or percentages (%) using the χ test. First, univariate analysis was performed to compare the baseline characteristics of patients with atrial fbrillation in diferent BMI groups. Univariate analysis was performed on the occurrence of cognitive impairment. Te relationship between BMI levels and cognitive impairment in diferent models was then analyzed. We built a total of 3 models. Model 1: Rough model without adjustment for any covariates; Model 2: Adjusted for sociodemographic data only; Model 3: Model adjusted for all confounders. We addressed the non-linear relationship between BMI and the occurrence of cognitive impairment using threshold efect tests and smooth curve ftting. All the analyses were performed with the statistical software packages R (https://www.R-project.org, the R Foundation) and EmpowerStats (https://www.empowerstats. com, X&Y Solutions, Inc, Boston, MA). P values less than 0.05 (two-sided) were considered statistically signifcant. (Table 1). A total of 244 patients with atrial fbrillation were collected for this study, with an average age of (67.28 ± 10.33) years, of whom 55.3% were male. Te average BMI was 25.33 ± 4.27 kg/m 2 . Te average cognitive function score of the research subjects was 19.25 ± 6.88 points. Te incidence of cognitive dysfunction was 80.33%. We divided 244 subjects into three groups equally: the BMI ≤ 23.3 kg/m 2 group, the 23.3 < BMI ≤ 26.6 kg/m 2 group, and the BMI >26.6 kg/m 2 group. Te number of patients in each group was essentially the same. In diferent groups, there were signifcant diferences in whether or not people took aspirin for more than 1 month, diabetes, hypertension, valvular disease, hyperlipidemia, coronary heart disease, heart failure, SBP, free thyroxine, and other indicators before medication (P < 0.05). (Table 2). Univariate analysis showed that age (F � 1.522, P � 0.025), sex (F � 14.440, P < 0.001), smoking history (F � 3.848, P � 0.010), drinking history (F � 4.544, P � 0.012), educational level (F � 17.404, P < 0.001), and hyperlipidemia (F � 7.795, P � 0.006) were signifcantly associated with the occurrence of cognitive impairment in patients with atrial fbrillation. (Table 3). In this study, three models were constructed to explore whether there was a linear correlation between BMI and cognitive function scores in patients with atrial fbrillation. Modelbased efect sizes were interpreted as the change in the MoCA score for each additional unit of BMI value. Only in Model 1, without adjustment for other variables, was BMI positively associated with cognitive function scores. In model 2 and model 3, the correlation between BMI and the cognitive function score was not statistically diferent. Tere was no linear correlation between BMI and cognitive function scores in patients with atrial fbrillation.

Te Nonlinear Relationship between BMI and Cognitive
Function Score (Table 4, Figure 2). In the non-linear relationship analysis of BMI and cognitive function scores, smooth curve ftting and threshold efect test results showed that there was a curvilinear relationship between the two. Using multiple regression and recursive algorithms, the infection point was calculated to be 24.56 kg/m 2 . To the left of the infection point, the relationship was signifcant, with efect sizes and 95% CI of 0.43 and 0.01-0.85, respectively. To the right of the infection point, there was no signifcant correlation between the two (P � 0.152). A second infection point can be observed in Figure 2, but we do not fnd statistical signifcance.

Discussion
Te results of this study showed a non-linear relationship between BMI and cognitive function scores in patients with atrial fbrillation after adjustment for other covariates. Te turning point of the relationship between the two is 24.56 kg/ m 2 and on the right side of the infection point, there is no signifcant correlation between the two. On the left side of the infection point, cognitive function scores increased by 0.43 points for every one-unit increase in BMI in patients with atrial fbrillation.
Previous studies have shown that obesity is a risk factor for cognitive impairment [10,19,20]. In children and adolescents, obesity leads to impaired cognitive function [21], and middle age is also an important factor leading to cognitive decline [22]. Meta-analysis results showed that executive function, short-term memory, and other dimensions of cognitive function are negatively correlated with BMI [23,24]. Te possible reasons for the decline of cognitive function caused by obesity are: (1) Obesity leads to neurological atrophy in patients. For example, the increase in the BMI index is related to the decrease of the brain volume, gray matter atrophy, and white matter decrease in patients [25]; (2) obesity leads to changes in neural activity. Elevated BMI led to a decrease in regional blood fow in the frontal lobes of patients. Activity in cortical regions related to episodic memory (hippocampus, angular gyrus, and dorsolateral prefrontal cortex) is also associated with obesity and insulin        resistance [26]; (3) obesity leads to brain aging, and middleaged obesity leads to the most severe brain aging. In obese patients, adiponectin levels are reduced [27]. Adiponectin prevents infammation promotes cell proliferation and supports the upregulation of energy metabolism and other adipokines [28]. However, there is still an "obesity paradox," that is, obesity is a protective factor for cognitive function. Ely et al. [29] explored the relationship between cognitive function and obesity in patients with heart failure. Studies have shown that non-obese patients have a higher risk of cognitive dysfunction than obese patients. Qizilbash et al. [30] conducted a study on the relationship between BMI and the occurrence of dementia. Te study included nearly 2 million subjects. Te results showed that compared with healthy weight people, people with lower body weight (BMI <20 kg/ m 2 ) sufered from the risk of dementia was 34% higher (95% CI 29-38). Te study followed the population for 15 years, and the obesity paradox persists. In addition, the incidence of dementia continued to decline with each grade-level increase in BMI. People who were very obese (BMI >40 kg/m 2 ) had a 29% lower risk of dementia compared to people of healthy weight. Te reasons for the existence of the obesity paradox may include diet, exercise, genetics, and other factors, but the specifc mechanism is not fully understood.
Diferent from the results of the above two aspects, the results of this study found a specifc threshold for the impact of BMI on cognitive function. When the BMI value of patients with atrial fbrillation was less than 24.56 kg/m 2 , the cognitive function score increased by 0.43 points for each unit of increase in the BMI of patients with atrial fbrillation. An increase in BMI at this time is a protective factor for cognitive function, which is consistent with the fndings of Qizilbash et al. [30] discussed above. However, when BMI was greater than 24.56 kg/m 2 , the relationship between BMI and cognitive function was not signifcant, which was diferent from the results of the previous studies described above. Te reason may be that the population of this study is patients with atrial fbrillation, and the mechanism of cognitive decline in patients with atrial fbrillation is more complex, including blood hypercoagulability, the formation of infammatory factors, decreased cerebral blood perfusion, and genetic changes [31]. Te population of this study difers from previous studies, which may explain the existence of a threshold between BMI and cognitive function scores in patients with atrial fbrillation. For the Asian population, the standard BMI range is 18.5-23.9 kg/m 2 , and the threshold obtained in this study was 24.56 kg/m 2 . Within the normal range of BMI, the higher the BMI, the higher the cognitive function score in patients with atrial fbrillation.
Decreased cognitive dysfunction in patients with atrial fbrillation leads to a decline in the quality of life of patients and seriously increases the burden on family caregivers, so its prevention and management are crucial. From the perspective of the relationship between BMI index and cognitive function score, this study explored the nonlinear relationship between the two after adjusting for covariates, which can provide an important basis for the prevention of cognitive dysfunction based on weight management. However, this study also has certain limitations. (1) Tis study only included a single-center AF patient population in southwestern Shandong Province, so the results may be diferent in other races, which may have some limitations in the external applicability of the results; (2) Tis study was a cross-sectional study, and no follow-up of patients was carried out. In the future, long-termfollow-up and tracking of patients' BMI and cognitive function level can be carried out to further explore the longitudinal relationship between the two; (3) Obesity can be refected by many indicators including BMI, waist circumference, abdominal circumference, and waist-to-hip ratio. Only BMI was used in this study. Regarding the relationship between obesity and cognitive function in patients with atrial fbrillation, more comprehensive indicators can be used to refect the degree of obesity in the future.

Conclusions
Te results of this study showed that when the BMI index of patients with atrial fbrillation was lower than 24.56 kg/m 2 , the cognitive function score increased by 0.43 points for each unit of increase in BMI of patients with atrial fbrillation. Te increase of BMI at this time was a protective factor for cognitive function. However, the relationship between the two was not signifcant when the BMI exceeded 24.56 kg/m 2 . Within the normal range of BMI, the higher the BMI of patients with atrial fbrillation, the higher their cognitive function scores. We encourage atrial fbrillation patients with a normal BMI to maintain their current weight. However, for patients with a BMI of more than 24.56 kg/m 2 , it is necessary to consider multiple clinical perspectives to guide patients and maintain a healthy weight.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that there are no conficts of interest.