Higher Total Cholesterol Concentration May Be Associated with Better Cognitive Performance among Elderly Females

Background: The brain contains the highest level of cholesterol in the body, and the total amount of serum cholesterol in the blood has a huge impact on brain aging and cognitive performance. However, the association of total serum cholesterol with cognitive function remains uncertain. This study determines whether there is an association between the total amount of cholesterol in the blood and cognitive performance in elderly females without a history of stroke. Methods: This population-based cross-sectional study was conducted on elderly (over 60 years old) females and males without a history of stroke from 2011 to 2014 in the US National Health and Nutrition Examination Survey (NHANES). The primary exposure was total blood cholesterol, and the main outcome was cognitive performance; this association was assessed with logistic regression analysis and restricted cubic splines. Results: 1309 female and 1272 male participants were included. In females, higher total cholesterol was significantly associated with higher cognitive scores, particularly in the digit symbol substitution test (OR 0.51, 95% CI (0.36–0.72)) and the animal fluency test (OR 0.64, 95% CI (0.45–0.91)). This association remained significant in models adjusted for age, race, smoking status, education level, and chronic conditions (OR 0.40, 95% CI (0.25–0.63)). This association was not significant in males, however. Conclusions: A higher concentration of total cholesterol measured in later life may be a protective factor for cognitive performance among females over 60 years old without a history of stroke. Further, this association was more pronounced among women with higher levels of education than women with lower or no education.


Introduction
Dementia, or Alzheimer's disease, is the fifth leading cause of death for individuals 65 years of age and older in the United States [1]. Worldwide, mild cognitive impairment affects 10-15% of the population aged 60 years or more [2]. Cognitive impairment results in various symptoms, including memory disorders, language decline, and execution ability disorder [3,4]. Cognitive impairments have become a key challenge for public health [5].
It has been hypothesized that lipid metabolism is associated with cardiovascular health problems [6]. However, lipid metabolism also plays an important role in cognitive health [7]. Docosahexaenoic acid (DHA) is a structural constituent of membranes in the development of the central nervous system and potentially influences cognitive decline in adult life [8]. The brain contains the highest level of cholesterol in the body [9], and cholesterol is important for brain activity. For example, the myelin, which provides an insulating layer for neurons to increase the brain's processing speed, contains 46% cholesterol [10]. Cholesterol is also associated with dopamine transport as it is present in the crystal structures of dopamine transporters (DATs) [11,12]. Meanwhile, low cholesterol levels are associated with a decrease in the number of serotonin receptors, leading to an overall reduction in serotonergic transmission in the brain [13]. Another related study revealed that a reduced

Study Population
The National Health and Nutrition Examination Survey (NHANES) is a survey research program conducted to assess the health and nutritional status of adults and children in the United States, tracking changes over time. It began in 1971 and became an annual event in 1999. In each cycle, approximately 5000 representative people are selected via a complex process of multistage probability sampling. All participants complete an informed consent document. The survey combines interviews, physical examinations, and laboratory tests. The participants first take part in an interview at home; physical examination and laboratory tests are conducted in the mobile examination center (MEC) before a call-back consultation is carried out within the next few days.
Two cycles (2011-2012 and 2013-2014) of data with the most recent information on cognitive function measures were collected and combined for our research. A total of 19,931 individuals participated in the NHANES from 2011 to 2014. A total of 2934 of the participants aged 60 or older were used in our research because only this sample of participants received a cognitive assessment. Among these participants, we further excluded those without total cholesterol measurements (N = 157) and participants with a history of stroke (N = 197). Next, the study enrolled 1309 female and 1272 male participants aged 60 or older ( Figure 1).

Cholesterol Measurement
An enzymatic assay method and the Trinder reaction (Roche Modular P chemistry analyzer) were used to measure total cholesterol. A heparin-manganese precipitation

Cholesterol Measurement
An enzymatic assay method and the Trinder reaction (Roche Modular P chemistry analyzer) were used to measure total cholesterol. A heparin-manganese precipitation method or a direct immunoassay technique was used to measure high-density lipoprotein cholesterol [31][32][33][34].

Cognitive Performance Assessment
The cognitive function of older participants (aged 60 or over) was estimated using four tests [35,36]: (1) the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) test, (2) the animal fluency test, (3) the digit symbol substitution test (DSST), and (4) the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) delayed recall test. The four tests are widely used in cohort studies to estimate memory, language, speech, and cognitive functions and to screen risk factors [37][38][39].
The CERAD [40] includes a word list recall test (immediate and delayed). The immediate word list recall measures mild cognitive impairment; the participant is asked to recall stimuli immediately after they are presented. This is used to identify mild dementia. The word list delayed recall is a verbal declarative memory test using a visually or verbally presented word list where participants are asked to remember as many words as possible after a specified delay interval in order to quantify memory performance [41,42].
The AFT is widely used to estimate the cognitive abilities of patients with various neurological diseases. The test requires participants to name as many animals as possible in a short period of time. A low AFT score reflects an impairment of language function [43,44]. It has been proven to identify patients with mild cognition impairment, dementia, and normal aging [45,46].
The DSST consisted of digit-symbol pairs followed by a list of digits. The participants aim to write the corresponding symbol from the 133 boxes that hold adjacent numbers as quickly as possible. The number of correct symbols within the allowed time (2 min) is the score [47].
Due to the ceiling and floor effect caused by a wide range of cognitive function in the elderly population, namely, the scale attenuation effect [48,49], we created a global cognitive score by averaging the standardized scores of the three cognitive tests scores (the CERAD, animal fluency, and Digit symbol substitution tests) [50,51]. As there are no standard cutoff points for the CERAD, animal fluency, and DSST tests to identify low or normal cognitive function, we used the 25th percentile of the scores among different groups of participants as the cutoff point, which followed the method used in previous studies [52]. Given the significant effect of age on cognitive function, the global score was stratified at different ages (60-69 years, 70-79 years, and ≥80 years) [53]. For each test or global score, the participants were divided into a low-cognitive-performance group and a normal-cognitive-performance group-the former group's scores were less than or equal to the corresponding cutoff values, while the latter group's scores were higher than the corresponding cutoff values.

Statistical Analysis
Statistical analysis was performed on the original datasets extracted from the NHANES and a complex sample analysis was performed using the stratification, cluster, and sample weight variables provides in the NHANES [56]. A new sample weight variable for the two combined datasets from 2011 to 2014 was created by taking one twice for the 2-year weight for each participant sampled from 2011 to 2012 and one twice for the 2-year weight for each person sampled from 2013 to 2014. We used R 4.1.0 programming for statistical analysis and the R survey package to appropriately weigh analyses for the complex, multistage sampling design of NHANES. Continuous variable data were expressed as mean ± standard deviation, and categorical variables were expressed as percentages. Pearson's chi-square or Fisher's exact tests were used to compare the categorical variables between low-and normal-cognitive-performance groups. A Student's t-test or Mann-Whitney U-test was used to compare different levels of continuous variables between groups. A two-sided p < 0.05 was considered statistically significant.
In this study, the total cholesterol values among female and male participants were categorized into four groups according to first quartile, medium, and third quartile. For the main outcome (low or normal cognitive performance), single-variable logistic regression was performed to analyze the association between cognitive performance and total cholesterol in an unadjusted model. Multivariable logistic regression was performed and controlled for the effects of the following covariates of cognitive function: age, education level, race, smoking and drinking status, poverty-income ratio, BMI, chronic conditions (including diabetes and hypertension), and marriage status. The variables including age, race, education level, marriage status, poverty-income ratio, and BMI were used for adjustment in Model 2; the variables, including Model 2 plus drinking status, smoking status, and chronic conditions (hypertension and diabetes), were used for adjustment in Model 3.

Results
The demographic characteristics of all the participants (N = 2581) and of the female (N = 1309) and the male (N = 1272) participants in the final study sample were summarized and shown in Tables 1-3, respectively. The results showed that the decline in cognitive performance was significantly associated with race, ratio of family income to poverty, chronic conditions (including diabetes and hypertension), drinking status, education level, and total cholesterol (p < 0.01). Participants who were considered to have a cognitive impairment were more likely to be Mexican American, other Hispanic, non-Hispanic Black, and other races; have lower educational level and poverty-income ratio; have chronic conditions (including diabetes and hypertension); drink more; be widowed, divorced, separated, or never married; or have lower total cholesterol levels. For DSST, the prevalence of smoking (including currently smoking or previously smoked) in participants with low cognitive performance was significantly higher than that of people with normal cognitive performance. However, a higher concentration of total cholesterol was significantly associated with cognitive function only among older female participants; in older male participants, this association was not significant. Table 4 shows the relationship between the different dimensions of cognitive performance (including the animal fluency test, the digit symbol substitution test, and the CERAD test) and the different total cholesterol levels among older female participants.
Additionally, in Model 3, compared with low total cholesterol levels, those who had 201-229 mg/dL of total cholesterol had a multivariate-adjusted OR (95% CI) of 0.62 (95% CI, 0.39-0.99) for DSST, and 0.58 (95% CI, 0.37-0.90) for global cognitive performance, and those who had more than 229 mg/dL of total cholesterol had a multivariate-adjusted OR (95% CI) of 0.57 (95% CI, 0.38-0.85) for AST score, and 0.40 (95% CI, 0.25-0.63) for global cognitive performance. The trend test shown in Table 4 showed that cognitive performance (including global cognitive performance and the AST score) was associated with total cholesterol in a dose-dependent manner (p trend < 0.05) in an adjusted model, and that a higher dose of total cholesterol had a greater advantage for the protection of cognition.
However, Table 5 showed that, among older female participants, the association between total cholesterol and cognitive function was not obvious before or after adjustment for other variables in the logistic regression analysis. We performed a sub-groups analysis in terms of education level, as shown in Table 6. The result showed that higher total cholesterol was significantly associated with normal cognitive performance in people who have received higher education, while higher total cholesterol was normally associated with normal cognitive performance in people with a medium level of education.
Restricted cubic spline analyses between total cholesterol and the Animal Fluency Test score (A), the DDST (B), the CERAD test (C), and the global cognitive performance (D). The solid line and dashed lines represent the estimated ORs and their 95% confidence intervals (OR, odds ratio). Figure 2 described the results of the restricted cubic spline analyses between the Animal Fluency test, the DSST, the CERAD test, and the global cognitive performance, respectively, among female participants. The prevalence of impaired cognitive function decreased with the increasing value of total cholesterol and showed a potential nonlinear association.    Test score (A), the DDST (B), the CERAD test (C), and the global cognitive performance (D). The solid line and dashed lines represent the estimated ORs and their 95% confidence intervals (OR, odds ratio). Figure 2 described the results of the restricted cubic spline analyses between the Animal Fluency test, the DSST, the CERAD test, and the global cognitive performance, respectively, among female participants. The prevalence of impaired cognitive function decreased with the increasing value of total cholesterol and showed a potential nonlinear association.
Restricted cubic spline analyses between total cholesterol and the Animal Fluency Test score (A), the DDST (B), the CERAD test (C) and the global cognitive performance (D). The solid line and dashed line represent the estimated ORs and their 95% confidence intervals (OR, odds ratio). The baseline characteristics of the approximately 25 million persons represented by the 1309 female participants are shown in Supplementary Table S1, in which the result was similar to that in Table 2. Additionally, we found similar results (Supplementary Table  S2) that total cholesterol was a protective factor for cognitive performance with logistic regression, especially in the Animal Fluency Test, the Digit Symbol Substitution Test, and global cognitive performance. Furthermore, the trend analysis in Supplementary Table S2 suggested a dose-dependent protective role for total cholesterol in cognitive performance. Additionally, the missing data in the original dataset of the female participants are shown in Supplementary Table S3.

Discussion
With the increase in life expectancy, cognitive impairment is an important public health problem; furthermore, the incidence of cognitive impairment is increasing. A systematic review has reported that the global prevalence of cognitive impairment ranged from 5.1% to 41% [57]. The factors leading to cognitive decline vary, including increasing age, lower education, race, and chronic conditions such as stroke and diabetes [58]. However, it has not been confirmed whether the amount of total cholesterol measured in late life is a protective factor for cognitive function.
In this study, we combined data including NHANES 2011-2012 and 2013-2014. This research explored data provided by 1309 female Americans aged 60 years or older who had no history of stroke. The 1309 female participants are representative of approximately twenty-five million people in America. In the unadjusted model, a high concentration of total cholesterol was statistically associated with higher global cognitive and animal fluency test scores. In the model adjusted for age, race, education levels, income-poverty ratio, history of diabetes, and smoking status, total cholesterol remained associated with global cognitive and animal fluency scores.
To date, evidence for the relationship between total cholesterol and cognitive function in elderly populations has been equivocal and inconclusive, likely due to the effects of aging. Our findings on the relation between total cholesterol and cognition were consistent with some previous studies. Research based on 382 people studied the relationship between concentrations of total cholesterol and cognitive performance in participants ranging from 70 to 90 years old, revealing that a lower total cholesterol value is associated with poorer cognition in both non-dementia and dementia cases [59]. Another 70-year follow-up study including 1034 people showed that higher total cholesterol and HDL-C values were associated with higher cognitive scores among participants aged 70 [60], although the relationship between cholesterol levels and cognition were mostly no longer significant after considering IQ test scores recorded in early life.
There have also been studies revealing total cholesterol's harmful effects on cognition. A study found that higher blood concentrations of total cholesterol were related to faster global cognitive decline in a population of 1159 Chinese adults aged over 60 [61]; the associations between all lipids and cognitive decline appeared to be more significant among individuals over 100 years old. Another study found that higher circulating total cholesterol indicated an increased risk of mild cognitive impairment in elderly females [62]. However, these two studies enrolled participants who were different from the participants in our studies, and the cognitive scale they used was different from the scales used in our study [63]. The population enrolled in the former study investigated Chinese adults over 60 years who had only received a basic education [64]. However, our sub-groups analysis found that higher total cholesterol was only significantly associated with normal cognitive performance among the female population who had received a higher education. Moreover, our sample size was larger, with 1309 female individuals without a history of stroke; the weighted sample size was approximately twenty-five million people. A meta-analysis based on eight studies and over 21,000 individuals aged over 60 years found that there were no relationships between cholesterol and cognitive decline or dementia in older adult groups [65].
Some studies have suggested that this relationship could be age-dependent, i.e., the higher concentration of total cholesterol measured at midlife was a risk factor for late-life cognition [66], while the higher concentration of total cholesterol measured at late-life was a risk factor for late-life cognition [19][20][21]29,30,67].
Although genetic susceptibility and metabolic function determine an individual's lipid profile to a great extent, factors related to lifestyle may affect the clearance of cholesterol. Different kinds of diet may affect cognitive function [68,69]. In addition to dietary intake, smoking status [70], education level [71], family income [72], race, and chronic conditions were proven to be associated with cognition decline. As a result, we introduced these features as co-variables. Dietary cholesterol has been considered to increase the risk of cardiovascular disease and, for cardiovascular events, a low cholesterol level is advisable. A study reported that increased cholesterol levels were associated with a high CVD risk among young adults [73]. However, recently some studies found that total cholesterol had protective qualities for some non-cardiovascular conditions and hemorrhagic stroke [74,75], especially among the old-aged [76]. Higher total cholesterol was associated with a decrease in mortality among people over 85 years old. In addition, people who are old with a low concentration of total cholesterol had higher all-cause mortality [77]. As a result, we believe a higher concentration of serum total cholesterol signifies better health conditions in late life, in terms of both cognition performance and other non-cardiovascular and hemorrhagic stroke conditions.

Limitations
There are several limitations and strengths in our study. First, we chose the NHANES as our data source, which provided a sufficiently large sample size. Second, we used four standard scores of four cognitive tests to create a global compound cognitive score to minimize the ceiling and floor effects. Third, considering that age was the major risk factor for cognitive impairment, age was used to divide the population and to calculate the cut-off values (the lowest quartile) of each group. However, as a cross-sectional study, this research could not reveal the relationship between total cholesterol and cognitive performance for the population at different ages. Further longitudinal studies need to be conducted to investigate the nonlinear longitudinal relationship between total cholesterol and neuropsychological function and cognitive performance.
Investigating the relationship between total cholesterol and cognitive function can provide information that is significant for the prevention of cognition decline. Because this effect could be age-related, we might consider adopting different strategies in different ages in future work.

Conclusions
In America, a higher concentration of total cholesterol measured in late life may be a protective factor for cognitive performance among females over 60 years without a history of stroke. This association was more significant among females with higher education levels and was not as apparent in females with lower education levels or without an education. We postulate that a better total cholesterol concentration could be 200 mg/dL in elderly females without a medical history of strokes.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nu14194198/s1, Table S1: Characteristics of the study population after weighting; Table S2: Weighted odds ratios (95% confidence intervals) for four dimensions of cognitive performance across quartiles of total cholesterol in survey data; Table S3: Missing Data of  Variables for Table 2.
Author Contributions: W.O. and Y.T. designed the study; K.P. and C.L. organized data; C.L. examined the data; K.P., C.L. and Y.T. analyzed data and wrote the first draft of the manuscript; J.T. and S.H. revised the manuscript. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Patient consent was waived because the study was an analysis of a third-party anonymized publicly available database.

Data Availability Statement:
Restrictions apply to the availability of these data. Data were obtained from NHANES and are available at https://wwwn.cdc.gov/nchs/nhanes/Default.aspx (accessed on 26 March 2022) [78].