*Article* **Association of Dietary Pattern with Cardiovascular Risk Factors among Postmenopausal Women in Taiwan: A Cross-Sectional Study from 2001 to 2015**

**Sabrina Aliné 1, Chien-Yeh Hsu 2,3, Hsiu-An Lee 4, Rathi Paramastri <sup>1</sup> and Jane C.-J. Chao 1,3,5,\***


**Abstract:** Unhealthy diet and inappropriate lifestyle contribute to an imbalance in cardiometabolic profiles among postmenopausal women. This research aimed to analyze the association between dietary pattern and changes in cardiovascular risk factors among postmenopausal Taiwanese women using binary logistic regression. This cross-sectional study involved 5689 postmenopausal Taiwanese women aged 45 years and above, and the data were obtained from Mei Jau Health Management Institution database between 2001 and 2015. The cardiovascular risk dietary pattern characterized by high intakes of processed food, rice/flour products, organ meat, and sauce was derived by reduced rank regression. Participants in the highest quartile of the cardiovascular risk dietary pattern were more likely to have high levels of systolic blood pressure (OR = 1.29, 95% CI 1.08–1.53), diastolic blood pressure (OR = 1.28, 95% CI 1.01–1.62), atherogenic index of plasma (OR = 1.26, 95% CI 1.06–1.49), triglycerides (OR = 1.38, 95% CI 1.17–1.62), and fasting blood glucose (Q3: OR = 1.45, 95% CI 1.07–1.97). However, this dietary pattern was not correlated with total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and C-reactive protein. Therefore, adherence to the cardiovascular risk dietary pattern increases the risk of having higher levels of blood pressure, triglycerides, fasting blood glucose in postmenopausal Taiwanese women.

**Keywords:** postmenopausal women; dietary pattern; cardiovascular risk factors; reduced rank regression

#### **1. Introduction**

Menopause is defined as the cessation of menstruation owing to a decrease in ovarian follicles and the further reduction of estradiol levels. It occurs mostly at a median age of 51 years [1]. The diagnosis of menopause is based on no menstrual period for 12 consecutive months in women [2,3]. Around 467 million postmenopausal women were registered in the world in the 1990s, and by 2030 the number of postmenopausal women is expected to be 1.2 billion with 47 million new postmenopausal women each year [4]. Several studies revealed that postmenopause was associated with increased inflammatory markers such as C-reactive protein (CRP), interleukin-1α (IL-1α), and tumor necrosis factor-α (TNF-α) and an imbalance in cardiometabolic profiles such as low levels of high-density lipoprotein cholesterol (HDL-C) and elevated levels of total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), triglycerides (TG), visceral fat, and blood glucose [5–8]. These imbalanced cardiometabolic profiles were favorable for the progression of atherosclerosis and an increased risk of cardiovascular disease (CVD) [9]. Cardiovascular disease was

**Citation:** Aliné, S.; Hsu, C.-Y.; Lee, H.-A.; Paramastri, R.; Chao, J.C.-J. Association of Dietary Pattern with Cardiovascular Risk Factors among Postmenopausal Women in Taiwan: A Cross-Sectional Study from 2001 to 2015. *Nutrients* **2022**, *14*, 2911. https://doi.org/10.3390/nu14142911

Academic Editors: Birgit-Christiane Zyriax and Nataliya Makarova

Received: 15 June 2022 Accepted: 14 July 2022 Published: 15 July 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

known as the leading cause of mortality worldwide between 1990 and 2019. This scourge claimed around 18.6 million individuals' lives in 2019 [10]. According to the report by the Ministry of Health and Welfare, Taiwan, heart disease was the second leading cause of death following malignant neoplasms in 2020 [11]. Shen et al. found that in Taiwanese women, early age at menopause between 45 and 49 years was linked to higher CVD death rate and all-cause mortality [12]. In 2018, the percentage of the elderly aged 65 years and above in Taiwan surpassed 14% and has become an aged society [13]. The prevalence of cardiovascular disease, diabetes, and cancer was high among the Taiwanese elderly during the past decade [14]. Aging and atherosclerosis can cause vascular wall damage and estrogen receptor loss, and a decrease in circulating estrogen also reduces estrogen receptors in both vascular endothelium and vascular smooth muscle cells [3]. Additionally, women with vasomotor symptoms have significantly higher blood pressure, elevated circulating total cholesterol levels, and greater body mass index (BMI) than women without vasomotor symptoms [3].

Evidence showed that postmenopausal Chinese women increased the risk of dyslipidemia after multiple adjustment as compared to premenopausal women probably due to the loss of endogenous estrogen after menopause [15]. Some studies also supported that menopause was associated with adverse changes of cardiometabolic profiles and increased risk and mortality of CVD [5,15,16]. Research conducted by Lin et al. demonstrated that compared to premenopausal women in North Taiwan, postmenopausal women had considerably greater odds of having central obesity, metabolic syndrome, high blood pressure, and high blood triglycerides [17]. In addition, diet has been associated with cardiovascular risk factors and other health-related outcomes. A healthy balanced diet plays a significant role in the prevention and mortality reduction of chronic diseases [18]. However, postmenopausal women consuming an unhealthy diet such as high intake of sodium, added sugar, trans fats, and red meat but low intake of fruit, whole grains, fibers, fish, nuts, and legumes were correlated with abnormal fasting blood glucose, high BMI, hypertension, and high blood cholesterol which are considered as risk factors of CVD among postmenopausal women [19]. Brazilian postmenopausal women who consumed a low-quality diet with an excessive intake of sodium and low intakes of vegetables and fruit had central obesity, higher blood pressure, and increased levels of blood lipids and fasting blood glucose [20].

The dietary pattern is considered as a new approach applied in nutritional epidemiology to assess the relationship between dietary factors and disease risk [21]. However, little is known about the outcomes resulting from the association between dietary patterns and CVD risk factors among postmenopausal Taiwanese women. Hence, the aim of this study was to analyze the association between dietary patterns and changes in cardiovascular risk factors such as blood pressure, blood lipids, blood glucose, and CRP among postmenopausal Taiwanese women.

#### **2. Materials and Methods**

#### *2.1. Study Population and Data Source*

This cross-sectional study was conducted using the database from 2001 to 2015, and the data were collected by the Mei Jau (MJ) Health Screening Centers which are located in Taipei, Taoyuan, Taichung, and Kaohsiung cities in Taiwan. All the subjects signed the consent form and agreed their data only for research use without their identity before their health check-up at the MJ Health Screening Center. While visiting the MJ Health Screening Center, all the subjects filled the questionnaires to collect information about their socio-demographic status, lifestyle, and dietary habits by the self-reported questionnaires. Blood samples were analyzed for biochemical parameters. The study was approved by the Joint Institutional Review Board of Taipei Medical University (TMU-JIRB N202007075). There were 377,124 subjects who visited the MJ Health Screening Center between 2005 and 2015. We included women aged ≥45 years who self-reported menopausal status after missing their menstrual period for at least 12 consecutive months using a questionnaire. We excluded 299,450 participants who were male, had disease conditions such as cancer, cystic fibrosis, lung disease, cirrhosis, kidney disease, or infectious disease, or used any forms of lipid-lowering drugs. In addition, we excluded 68,985 women who were nonpostmenopausal, aged less than 45 years, or failed to complete the questionnaire about their dietary habits. After excluding 3000 participants who had multiple entries between 2005 and 2015, a total 5689 postmenopausal women were retained in this study (Figure 1).

**Figure 1.** Flowchart of study participants.

#### *2.2. Dietary Assessment and Other Covariates*

A semi-quantitative food frequency questionnaire (FFQ) was developed, standardized, and validated by the MJ Health Management Institution, and used to assess dietary habits of the subjects. The FFQ questionnaire contained the closed-ended questions about 22 non-overlapping food groups with a total of 85 individual food items consumed by the participants in the past month [22]. The intake frequency was assessed in accordance with daily and weekly consumption. Each question was given the definition about one serving size of the food item, and presented 5 frequency response options as described previously [22]. Dietary data were collected for further frequency response options as described previously [22]. Dietary data were collected for further analysis to derive the dietary pattern using a reduced rank regression (RRR) model. The RRR model as a multivariable linear function was performed to derive the dietary pattern related to the disease of interest by a priori and a posteriori approaches based on the response variables for identifying a linear combination of the predictor variables [23].

Demographic data such as age, education (≤high school or >high school), and marital status (never married, married, or divorced/widowed) were collected. We also evaluated lifestyle data including smoking status (no or yes), drinking alcohol (no or yes), physical activity frequency (<150 min/week or ≥150 min/week), and sleep duration (<6 h, 6–8 h, or >8 h). Medical history regarding hypertension, diabetes mellitus, and CVD was recorded. All covariates were assessed using a self-reported questionnaire.

#### *2.3. Anthropometric, Clinical, and Biochemical Data*

Anthropometric parameters such as height, weight, waist circumference (WC), and waist-to-hip ratio (WHR) were assessed using an anthropometer with electronic scale at the MJ Health Screening Center. The values of BMI were calculated using weight (kg) divided by height (m2) [24]. To identify central obesity among the participants, WC (≥80 cm) [25] and WHR (≥0.85) [26] were measured and calculated. Blood pressure was measured twice at 10 min intervals using a standardized sphygmomanometer. Biochemical data such as total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglycerides (TG), fasting blood glucose (FBG), and C-reactive protein (CRP) were assessed after overnight fasting for 12–14 h by the central laboratory at the MJ Health Management Institution. Blood TC, HDL-C, TG, and FBG were evaluated using the commercial kits (Randox Laboratories Ltd., Antrim, UK). The levels of LDL-C were determined by Friedewald formula (LDL-C (mg/dL) = TC-HDL-C-TG/5) [27]. Atherogenic index of plasma (AIP) as an indicator for CVD risk was calculated by the following formula: AIP = log(TG/HDL-C) [28]. Inflammatory marker CRP was diagnosed by the reagent from Fortress Diagnostics (Antrim, UK). Cardiovascular disease risk factors were defined as: systolic blood pressure (SBP) ≥140 mmHg and/or diastolic blood pressure (DBP) ≥ 90 mmHg [20,28], AIP ≥ 0.24 with high risk of CVD [29], TC ≥ 5.17 mmol/L (200 mg/dL) [20], LDL-C ≥ 2.59 mmol/L (100 mg/dL) [20], HDL-C ≤ 1.29 mmol/L (50 mg/dL) [20], TG ≥ 1.69 mmol/L (150 mg/dL) [20], FBG ≥ 7.0 mmol/L (126 mg/dL) [28], and CRP ≥ 28.6 nmol/L (3 mg/L) [30].

#### *2.4. Statistical Analysis*

Statistical analysis was performed using SAS version 9.4 (SAS Institute Inc., Chicago, IL, USA) and IBM SPSS 20 (IBM Corp., Armonk, NY, USA). Kolmogorov–Smirnov test was used to determine the distribution of the data. To compare the differences between two groups, Mann–Whitney U test and chi-square test were used for categorical data. To compare data among multiple groups, one-way analysis of variance (ANOVA) and Kruskal– Wallis test were performed. We used binary logistic regression expressed as odds ratios (ORs) and 95% confidence intervals (CIs) to determine the association between the dietary pattern and cardiovascular risk factors. The dietary pattern was derived by RRR using PROC PLS function in SAS 9.4, and 22 food groups were considered as the predictors. After performing Pearson's correlation coefficient, triglycerides, systolic blood pressure, fasting blood glucose, and AIP were retained as the response variables (Figure 2). In compliance with previous investigation, to obtain the dietary pattern linked to CVD risk, the value of factor loading was set at ≥0.20 [31]. The dietary factor score for each food group was calculated by summing food frequency intake weighed by their factor loadings. Finally, we only retained the first dietary factor for further analysis because it explained the maximum variation of the response variables. The derived dietary pattern was then divided into quartiles according to the dietary factor score. The reference group for the cardiovascular risk dietary pattern was quartile 1 (Q1) which was the lowest quartile of the dietary factor score, and quartile 4 (Q4) represented the highest quartile of the dietary factor score. In binary logistic regression analysis, model 1 was unadjusted, model 2 was adjusted for age, BMI, WC, and WHR, and model 3 was adjusted for model 2 variables plus education, family income, smoking, drinking alcohol, physical activity frequency, and sleep duration. The *p*-value < 0.05 was considered statistically significant.

**Figure 2.** Cardiovascular risk dietary pattern derived from reduced rank regression model. TG: triglycerides, SBP: systolic blood pressure, FBG: fasting blood glucose, AIP: atherogenic index of plasma.

#### **3. Results**

#### *3.1. Characteristics of Study Participants*

Table 1 presents the demographic and lifestyle characteristics of the participants. The majority of postmenopausal women in this study had education below high school (81.5%), non-professional occupation (63.9%), low annual income (<NTD800,000: 69.1%), married status (70.3%), no smoking (97.9%), no drinking alcohol (95.5%), less physical activity frequency (<150 min/week: 55.5%), and sleep duration for 6–8 h (58.6%). The anthropometric, clinical, and biochemical data are shown in Table 2. The majority of postmenopausal women had normal BMI (44.2%), waist circumference (56.1%), and waistto-hip ratio (68.9%). However, 31.3% postmenopausal women were overweight, 22.7% subjects were obese, 43.9% subjects had central obesity, and 31.1% subjects had abnormal waist-to-hip ratio. The prevalence of hypertension, diabetes, and CVD was 11.3%, 17.5%, and 10.7%, respectively. The mean value of AIP (0.3 ± 0.3) was higher than 0.24 defined as a CVD risk factor. The mean values of TC (5.9 ± 0.8 mmol/L) and LDL-C (3.7 ± 0.8 mmol/L) were abnormal among postmenopausal women. Among 5689 participants, only 7.3% subjects had normal FBG level (<7.0 mmol/L, data not shown).

**Table 1.** Demographic and lifestyle characteristics of postmenopausal women aged ≥45 years (*n* = 5689) 1.



**Table 1.** *Cont.*

<sup>1</sup> Continuous data are presented as mean ± SD and categorical data are expressed as numbers (percentage).

**Table 2.** Demographic, clinical, and biochemical data of postmenopausal women aged ≥45 years (*n* = 5689) 1.


<sup>1</sup> Continuous data are presented as mean ± SD and categorical data are expressed as numbers (percentage).

#### *3.2. Cardiovascular Risk Dietary Pattern*

A dietary pattern identified as a "cardiovascular risk dietary pattern" was derived using the RRR model. Four food groups including processed food, rice/flour products, organ meat, and sauce showed a positive correlation (factor loading ≥0.20) with the cardiovascular risk dietary pattern, meanwhile food groups such as dairy products, fruits, whole grains, and sweet bread had a negative correlation with this dietary pattern (factor loading

≤−0.20) (Figure 3). The cardiovascular risk dietary pattern explained 6.6% cumulative percentage of variation and 1.7% of the total variation for the four response variables. The explained variation was 1.5% for TG, 1.6% for AIP, and 1.8% for both SBP and FBG.

#### *3.3. Association between the Dietary Pattern and Cardiovascular Risk Factors*

The unadjusted and adjusted models for the association of the cardiovascular risk dietary pattern with SBP, DBP, and AIP among postmenopausal Taiwanese women are presented in Table 3. Model 1 was unadjusted, model 2 was adjusted for age, BMI, WC, and WHR, and model 3 was adjusted for model 2 variables plus education, family income, smoking, drinking alcohol, physical activity frequency, and sleep duration. The results showed that participants in the higher quartiles (Q3 and Q4) of the cardiovascular risk dietary pattern were more likely to increase the odds of having high SBP (OR = 1.40–1.84), high DBP (OR = 1.28–1.69), and high AIP (OR = 1.43–1.69) compared to those in the reference group (Q1) before adjustment. After adjusting variables in models 2 and 3, participants in the highest quartile (Q4) of the cardiovascular risk dietary pattern were still more likely to increase the odds of having high SBP (model 2: OR = 1.42, 95% CI 1.20–1.68, model 3: OR = 1.29, 95% CI 1.08–1.53), high DBP (model 2: OR = 1.43, 95% CI 1.13–1.79, model 3: OR = 1.28, 95% CI 1.01–1.62), and high AIP (model 2: OR = 1.29, 95% CI 1.09–1.52, model 3: OR = 1.26, 95% CI 1.06–1.49).

The association of the cardiovascular risk dietary pattern with TC, LDL-C, and HDL-C among postmenopausal Taiwanese women in the unadjusted and adjusted models is shown in Table 4. The cardiovascular risk dietary pattern was not correlated with the odds of high TC in all models. Participants in the Q2 quartile of the cardiovascular risk dietary pattern were more likely to decrease the odds of having high LDL-C in all models (model 1: OR = 0.63, 95% CI 0.47–0.83, model 2: OR = 0.68, 95% CI 0.51–0.91, model 3: OR = 0.71, 95% CI 0.53–0.94) compared to those in the Q1 quartile. Participants in the higher quartiles (Q3 and Q4) of the cardiovascular risk dietary pattern were more likely to decrease the odds of having low HDL-C in the unadjusted model; however, no association was found after adjustment in models 2 and 3.

Table 5 demonstrates the association of the cardiovascular risk dietary pattern with TG, FBG, and CRP in the unadjusted and adjusted models among postmenopausal Taiwanese women. Participants in the higher quartiles (Q2–Q4) of the cardiovascular risk dietary pattern were more likely to increase the odds of having high TG in all models (model 1: OR = 1.39–1.79, model 2: OR = 1.21–1.43, model 3: OR = 1.18–1.38) compared to those in the lowest quartile (Q1). Participants in the higher quartile (Q3) of the cardiovascular risk dietary pattern were more likely to increase the odds of having high FBG in all models (model 1: OR = 1.75, 95% CI 1.30–2.35, model 2: OR = 1.54, 95% CI 1.14–2.07, model 3: OR = 1.45, 95% CI 1.07–1.97). Participants in the higher quartiles (Q3 and Q4) of the cardiovascular risk dietary pattern were more likely to increase the odds of having high

CRP (Q3: OR = 1.38, 95% CI 1.13–1.67, Q4: OR = 1.51, 95% CI 1.25–1.83) only in the unadjusted model.

**Table 3.** Binary logistic regression for the association between the dietary pattern, systolic blood pressure (SBP), diastolic blood pressure (DBP), and atherogenic index of plasma (AIP) (*n* = 5689).


<sup>1</sup> High SBP, high DBP, and high AIP were defined as SBP ≥ 140 mmHg, DBP ≥ 90 mmHg, and AIP ≥ 0.24, respectively. <sup>2</sup> Model 1 was unadjusted. <sup>3</sup> Model 2 was adjusted for age, body mass index, waist circumference, and waist-to-hip ratio. <sup>4</sup> Model 3 was adjusted for model 2 variables plus education, family income, smoking, drinking alcohol, physical activity frequency, and sleep duration. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, significantly different from the reference group.

**Table 4.** Binary logistic regression for the association between the dietary pattern, total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C) (*n* = 5689).



**Table 4.** *Cont.*

<sup>1</sup> High TC, high LDL-C, and low HDL-C were defined as TC ≥ 5.17 mmol/L (200 mg/dL), LDL-C ≥ 2.59 mmol/L (100 mg/dL), and HDL-C ≤ 1.29 mmol/L (50 mg/dL), respectively. <sup>2</sup> Model 1 was unadjusted. <sup>3</sup> Model 2 was adjusted for age, body mass index, waist circumference, and waist-to-hip ratio. <sup>4</sup> Model 3 was adjusted for model 2 variables plus education, family income, smoking, drinking alcohol, physical activity frequency, and sleep duration. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, significantly different from the reference group.

**Table 5.** Binary logistic regression for the association between the dietary pattern, triglycerides (TG), fasting blood glucose (FBG), and C-reactive protein (CRP) (*n* = 5689).


<sup>1</sup> High TG, high FBG, and high CRP were defined as TG ≥ 1.69 mmol/L (150 mg/dL), FBG ≥ 7.0 mmol/L (126 mg/dL), and CRP ≥ 28.6 nmol/L (3 mg/L), respectively. <sup>2</sup> Model 1 was unadjusted. <sup>3</sup> Model 3 was adjusted for age, body mass index, waist circumference, and waist-to-hip ratio. <sup>4</sup> Model 4 was adjusted for model 2 variables plus education, family income, smoking, drinking alcohol, physical activity frequency, and sleep duration. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, significantly different from the reference group.

#### **4. Discussion**

#### *4.1. Association between the Dietary Pattern and Cardiovascular Risk Factors*

In this cross-sectional study of 5689 postmenopausal Taiwanese women, we derived the cardiovascular risk dietary pattern and found a positive association with several CVD risk factors such as SBP, DBP, AIP, TG, and FBG. Among the participants in the highest quartile of the cardiovascular risk dietary pattern, 60.4% of postmenopausal women were overweight or obese, and 64.6% were physically inactive (<150 min/week) (data not shown). The cardiovascular risk dietary pattern was recognized by high consumption of processed food, rice/flour products, organ meat, and sauce, but low intakes of dairy products, fruit, whole grains, and sweet bread. The cardiovascular risk dietary pattern reflected similar characteristics as the western dietary pattern recognized by high intakes of processed

food, meat, organ meat, rice/flour products, but low consumption of fruit, dark-colored vegetables, bread, and legume/soy products among Taiwanese middle-aged and elderly with chronic kidney disease [23]. Processed food and organ meat are often rich in calories, cholesterol, and/or saturated fat, and all of which could contribute to excessive energy consumption [23].

Low fiber and excessive salt and/or sugar in processed food as well as unbalanced saturated and unsaturated fats in animal food could be correlated with abnormal blood pressure, blood lipids, and blood glucose among Taiwanese middle-aged adults and elderly [23]. Highly refined carbohydrate in rice/flour products, a dietary component for high intake in the cardiovascular risk dietary pattern, could be associated with increases in cardiovascular risk and the development of atherosclerosis among middle-aged adults [32].

Our results revealed that the cardiovascular risk dietary pattern was positively associated with blood pressure. We found that the prevalence of hypertension was only 11.3% among 5689 postmenopausal Taiwanese women. Unlike our results, the previous studies conducted among postmenopausal women reported that the prevalence of hypertension was 31.6% and 56.0% in Brazilian and Chinese postmenopausal women, respectively [20,33]. Weight gain and increased sensitivity to salt in the diet might occur due to hormonal changes after menopause and age-associated metabolic changes, which could lead to a raise in blood pressure [34]. Weight status and physical activity could also contribute to abnormal blood pressure. Postmenopausal women aged <65 years with overweight (33.3%) or obesity (42.9%) also had higher prevalence of high blood pressure (130 mmHg/85 mmHg) compared to those who had normal weight (18.8%), and those who did not do aerobic exercise tended to have higher prevalence of high blood pressure compared to those who did aerobic exercise actively (44.0% vs. 14.3%, *p* = 0.06) [20].

Our findings showed that the cardiovascular risk dietary pattern was correlated with an increase in AIP among postmenopausal women. Numerous studies demonstrated that AIP was an important cardiovascular risk factor and a better predictor for CVD [33,35,36]. The previous studies have reported that AIP was a better predictor of the fractional esterification rate of HDL-C which is a powerful predictor of CVD [35], and a more sensitive diagnostic marker for studies of CVD [35], and a more sensitive diagnostic marker for CVD among postmenopausal women compared to traditional lipid parameters [35,36].

Our results revealed that participants in the highest quartile (Q4) of the cardiovascular risk dietary pattern were more likely to increase the odds of having high CRP before adjustment, even the association between dietary pattern and CRP was not significant after adjustment. A previous study conducted in Southern Brazil among postmenopausal women also observed that participants with high CRP were positively correlated with BMI, WC, body fat, TG, glucose, sedentary lifestyle, and excessive dietary carbohydrate intake (>55% of total energy) [37].

Although the association between aberrant lipid profiles and certain nutrients or food groups has been established, few have demonstrated the association between dietary pattern or quality and blood lipids in postmenopausal women [32]. We found that the cardiovascular risk dietary pattern was positively correlated with increased odds of high TG among postmenopausal women after full adjustment. Brazilian postmenopausal women with a low-quality inadequate diet characterized by an excessive intake of sodium (>2400 mg/day) had increases in the prevalence of high TC and high LDL-C known as cardiovascular risk factors [20]. However, Tardivo and co-workers [32] showed that there was no significant association between diet quality determined by healthy eating index scores and blood lipids in Brazilian postmenopausal women. A study conducted among Korean women showed that postmenopausal women who consumed the western dietary pattern with high intakes of oil and fats, meat, eggs, fast food, and sweets but low intake of grains were correlated with hyper LDL-C [38]. Other studies conducted among Chinese women and Japanese women consuming a western dietary pattern with high intakes of milk, dairy products, and fast food but a low intake of rice or vegetables revealed an imbalance in lipid profile, especially increases in TC and LDL-C [39,40]. The abnormality

of various serum lipids was linked to hormonal changes, such as the rise in circulating androgen and the reduction in estrogen, during the menopausal transition period [39].

Adherence to a western type dietary pattern could be associated with the status of being overweight or obese and having high WC, which might contribute to metabolic alteration. The metabolic changes in postmenopausal women could explain the imbalance of CVD-related biochemical variables [41]. Because of estrogen deficiency, postmenopausal women could increase CVD risk factors including central obesity, elevated blood pressure, increased blood lipids, decreased glucose tolerance, and increased vascular inflammation [42]. Compared to premenopausal women, postmenopausal women were more prone to increase blood lipids, which could lead to increase the risk for the development of atherogenesis [16]. In addition, the dietary components could be correlated with abnormal CVD-related biochemical variables in postmenopausal women. High consumption of energy [43], saturated fatty acids [44,45], trans fats [45], cholesterol [46], and eggs [46] was associated with an increased risk of CVD or abnormal CVD-related biochemical variables among postmenopausal women. In contrast, a low-fat dietary pattern [45] or the dietary pattern with high consumption of plant food such as whole grains, vegetables, fruits, legumes, and nuts or seeds, but low intakes of processed food, red meat, sugar, and sodium [47] were correlated with a reduced risk of CVD among postmenopausal women. The cardiovascular risk dietary pattern identified in our study was characterized by high intakes of processed food, rice/flour products, organ meat, and sauce which were accompanied by a high amount of energy, saturated fats, trans fats, cholesterol, added sugar, and sodium. Although the underlying mechanism for the effects of dietary patterns or dietary components on CVD risk factors among postmenopausal women has not been fully understood, changes in lipid metabolism and the increased accumulation of visceral fat related to estrogen deficiency in postmenopausal women could partially contribute to the effect of the dietary pattern on CVD risk factors.

#### *4.2. Strengths and Limitations*

To our knowledge, the present study is the first one to identify the cardiovascular risk dietary pattern in postmenopausal Taiwanese women using the RRR model as a novel and powerful method. Additionally, the RRR model gave more explanation about the association between the dietary pattern and the disease of interest. Since the RRRderived dietary pattern was generated by a disease-specific response, the response variables were correlated to the disease of concern [48]. Instead of explaining the variation in significant biomarkers, principal component analysis only provided the explanation of the overall variation in food group intake [48]. Meanwhile, by maximizing the explained variation in the biomarkers for diet-related disorders, the RRR model could be able to predict dietary pattern scores. Researchers can also determine the percentage variance using the RRR approach from the predictor variables and response variables, and both of which contributed to the dietary component [48]. Both the corresponding response scores and the explained variation in the predictor variables could be used to evaluate the extracted factor scores [48]. The large study population collected for 15 years could be representative of postmenopausal Taiwanese women. We also included demographic, anthropometric, clinical, biochemical, and dietary data to explore the association between these variables. However, a number of methodological limitations need to be addressed. First, our study was a cross-sectional study which provided features of eating habits and other characteristics at a specific time point and could raise the possibility of reverse causation bias. Second, the information for FFQ used to identify dietary habits could have self-reported bias. Additionally, the FFQ could be used for an estimate of habitual food intake but not for actual nutrient consumption. Even though the analysis was adjusted for the majority of known confounding variables, the residual confounding bias due to unknown or unmeasured covariates could not be completely ruled out. A longitudinal study is needed to explore the association between dietary patterns and CVD risk factors

among postmenopausal Taiwanese women. Further research should be conducted to compare the association in premenopausal versus postmenopausal Taiwanese women.

#### **5. Conclusions**

The cardiovascular risk dietary pattern with a high intake of processed food, rice/flour products, organ meat, and sauce is associated with increased odds of high blood pressure, AIP, TG, and FBG among postmenopausal women. Our study suggests that choosing a healthier dietary pattern with a lower intake of processed food, rice/flour products, organ meat, and sauce could reduce the risk of CVD in postmenopausal Taiwanese women.

**Author Contributions:** Conceptualization, S.A. and J.C.-J.C.; data curation, C.-Y.H. and H.-A.L.; formal analysis, S.A. and R.P.; writing—original draft preparation, S.A. and J.C.-J.C.; writing—review and editing, J.C.-J.C.; supervision, C.-Y.H. and J.C.-J.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki, and approved by the Taipei Medical University–Joint Institutional Review Board (N202007075 and date of approval 11 September 2020).

**Informed Consent Statement:** All the participants signed the consent form authorized by the Mei Jau Health Management Institution.

**Data Availability Statement:** The data that support the findings of this study are available from the Mei Jau Health Management Institution, but restricted for research use only. The data are not publicly available.

**Acknowledgments:** The authors thank the Mei Jau Health Management Institution for collecting and providing their database available for our study.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Nutritional Risk Factors Associated with Vasomotor Symptoms in Women Aged 40–65 Years**

**Alexandra Tijerina 1,†, Yamile Barrera 1,†, Elizabeth Solis-Pérez 1, Rogelio Salas 1, José L. Jasso 1, Verónica López 1, Erik Ramírez 1, Rosario Pastor 2,3, Josep A. Tur 3,4,5,\* and Cristina Bouzas 2,3,4,5**


**Abstract:** Vasomotor symptoms (VMS) are the most common symptoms among menopausal women; these include hot flashes and night sweats, and palpitations often occur along with hot flashes. Some studies in Mexico reported that around 50% of women presented with VMS mainly in the menopausal transition. It has been proven that VMS are not only triggered by an estrogen deficiency, but also by nutritional risk factors. Evidence of an association between nutritional risk factors and VMS is limited in Mexican women. The aim of this study is to identify nutritional risk factors associated with VMS in women aged 40–65 years. This is a comparative cross-sectional study, undertaken in a retrospective way. A sample group (*n* = 406 women) was divided into four stages according to STRAW+10 (Stages of Reproductive Aging Workshop): late reproductive, menopausal transition, early postmenopause, and late postmenopause. Hot flashes were present mainly in the early postmenopause stage (38.1%, *p* ≤ 0.001). Two or more VMS were reported in 23.2% of women in the menopausal transition stage and 29.3% in the early postmenopause stage (*p* < 0.001). The presence of VMS was associated with different nutritional risk factors (weight, fasting glucose levels, cardiorespiratory fitness, and tobacco use) in women living in the northeast of Mexico.

**Keywords:** vasomotor symptoms; risk factors; reproductive aging; menopause; women; Mexico

#### **1. Introduction**

Vasomotor symptoms (VMS) are the most common symptoms among menopausal women. These symptoms are short-term menopausal disorders, which include hot flashes and night sweats [1,2], and are often accompanied by palpitations [2]. According to STRAW+10 (Stages of Reproductive Aging Workshop) criteria, which is a "gold standard" that classifies women into stages of reproductive aging, VMS can appear in late menopausal transition but are more common in the early postmenopause [3,4]. Some studies in Mexico reported that VMS in women appear mainly in the menopausal transition, with a prevalence of 45% in women aged 45–55 years [5] and 47.63% in women aged 51–63 years [6].

It has been proven that VMS are not only triggered by estrogen deficiency, but also by different nutritional risk factors that can be modifiable and non-modifiable. Clinical practice guidelines for the menopausal stage [7–9], along with several other studies, delimitate that a body mass index higher or equal to 25, hypertension, and tobacco use are nutritional risk

**Citation:** Tijerina, A.; Barrera, Y.; Solis-Pérez, E.; Salas, R.; Jasso, J.L.; López, V.; Ramírez, E.; Pastor, R.; Tur, J.A.; Bouzas, C. Nutritional Risk Factors Associated with Vasomotor Symptoms in Women Aged 40–65 Years. *Nutrients* **2022**, *14*, 2587. https://doi.org/10.3390/ nu14132587

Academic Editors: Birgit-Christiane Zyriax and Nataliya Makarova

Received: 4 May 2022 Accepted: 20 June 2022 Published: 22 June 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

factors for VMS [10–12]. Their mechanism seems to relate to impaired heat conductance, blood flow, and hormones [12,13]. However, there is limited evidence concerning other factors such as fasting glucose ≥100 mg/dL, poor cardiorespiratory fitness, and excessive total fat intake.

There is also limited evidence concerning the nutritional factors associated with VMS in Mexican women. In Mexico, the prevalence of VMS has been reported but the associated risk factors have not [5,6]. A previous study has evaluated the social and nutritional factors associated with menopausal symptoms in Mexican women; however, there was no distinction of VMS, as nine different symptoms were arranged into one climacteric group [14]. These studies were carried out in central Mexico and there is limited evidence for the northeast region. The aim of this study is to identify nutritional risk factors associated with VMS in women aged 40–65 years in northeast Mexico.

#### **2. Materials and Methods**

#### *2.1. Design and Subjects*

This is a cross-sectional study which was carried out from 2015 to 2017. Women enrolled in this study were 40–65 years of age, living in the metropolitan area of Monterrey, in Nuevo León state, Mexico. They were apparently healthy, they voluntarily agreed to participate, and all provided written informed consent. Exclusion criteria included illnesses that affected their habitual eating habits and having undergone a hysterectomy. Women with incomplete data were eliminated from the study analysis (Figure 1). A total of *n* = 406 women were included in the study, representative of the study population (560,115 women aged 40–65 years) of Nuevo León state in the year 2020. This was according to a finite population equation, with 5% error and 95% confidence intervals, considering a VMS proportion of 50% [15].

**Figure 1.** Flowchart of the recruitment of the study population.

Women were invited to participate via physical flyers and social media. They were screened for the inclusion criteria via a telephone call and scheduled for an appointment at the Center for Research in Nutrition and Public Health of the Facultad de Salud Pública y Nutrición, Universidad Autónoma de Nuevo León (School of Public Health and Nutrition, Autonomous University of Nuevo León, translated into English).

This study followed the Declaration of Helsinki, and it was approved by the ethics committee of the Facultad de Salud Pública y Nutrición with protocol ID: 15–FaSPyN–SA–11.

#### *2.2. Stages of Reproductive Aging (STRAW+10 Criteria)*

A questionnaire was used to obtain data on menstrual cycles, including previous menstrual period, presence of blood flow or amenorrhea, and changes in length between cycles. According to the STRAW+10 criteria [3], women were classified into 4 stages of reproductive aging (Figure 1). Late reproductive was defined as "the presence of blood flow without changes or with short cycles." Menopausal transition was defined as "the presence of blood flow with long cycles or with at least one interval of amenorrhea ≥60 days," "amenorrhea <60 days and without changes," or "amenorrhea ≥60 days." Early postmenopause was defined as "amenorrhea ≥12 months, but ≤8 years," and late postmenopause was defined as "amenorrhea >8 years" [3].

#### *2.3. Vasomotor Symptoms*

Vasomotor symptoms (VMS) were reported in a questionnaire and included hot flashes and night sweats. Palpitations were also registered, as they are often present along with hot flashes [2]. Women had the option to answer yes (presence) or no (absence). After the information was collected, 4 categories were established: (1) absence, (2) presence of hot flashes, (3) presence of night sweats or palpitations, and (4) presence of 2 or more VMS, including hot flashes, night sweats, and palpitations.

#### *2.4. Nutrition Assessment*

A nutrition assessment of participants was performed according to the Nutrition Care Process of the Academy of Nutrition and Dietetics [16] as follows:

#### 2.4.1. Anthropometric Measurements

Body mass index (BMI) was determined by the formula BMI = weight (kg)/height<sup>2</sup> (m2), using a scale (Seca 874, ± 0.1 kg, Azcapotzalco, Mexico) for the weight and a digital stadiometer (Seca 274, ± 2 mm, Azcapotzalco, Mexico) for the height. BMI was classified as obese ≥30 kg/m2, overweight 25–29.9 kg/m2, or normal weight 18.5–24.9 kg/m2 [17].

#### 2.4.2. Biochemical Data

Venous blood samples were collected at fast, centrifuged at 3500 rpm for 12 min, and serum was obtained. The serum was frozen at −80 ◦C until assays were performed with the glucose oxidase/peroxidase method. Fasting glucose was obtained using A25 autoanalyzer (software version 4.1.1) (CV = 1.2%) (BioSystems S.A, Barcelona, Spain), according to the Norma Oficial Mexicana NOM–253–SSA1–2012 [9]. Fasting glucose was classified as high ≥100 mg/dL or normal <100 mg/dL [18]. Women on treatments using hypoglycemic drugs were also considered to have high fasting glucose levels.

#### 2.4.3. Nutrition-Focused Physical Exam Findings

Systolic and diastolic blood pressure measurements were performed to the nearest 1 mmHg using a digital sphygmomanometer, according to the Norma Oficial Mexicana NOM–030–SSA2–2009 [19]. Two readings were taken 5 min apart and the average was calculated. Blood pressure was classified as hypertensive between ≥130 and ≥80 mmHg, elevated between 120–129 and <80 mmHg, or normal between <120 and <80 mmHg [20]. Women on treatments using antihypertensive drugs were considered to be hypertensive.

The cardiorespiratory fitness (CF) of the women was obtained by measuring the walking distance achieved in meters (m) during a six-minute (min) test, in a 15 m × 28 m field. The CF value was reported as meters per minute, using the formula CF = m/6 min. CF was determined as poor at <400 m/6 min or excellent at ≥400 m/6 min [21].

#### 2.4.4. Food and Nutrition-Related History

Total fat intake was assessed by a validated Food Frequency Questionnaire (FFQ) [22]. Women were asked to report the frequency and measurements of their intake of 136 items, including foods and beverages. Total fat and energy intake was analyzed using the software Food Processor® version 15.0 (ESHA Research, Salem, OR, USA) and was reported in grams per day (g/day) [23].

#### 2.4.5. Covariates

Women also reported tobacco use (yes or no), hormone use (yes or no), and date of birth to determine age (years). Daily physical activity was obtained from the Minnesota leisure-time physical activity (LTPA) questionnaire, in which women reported the frequency and time spent in different activities. Metabolic equivalents per day (MET/d) were calculated [24]. Intake of alcohol in milliliters per day (mL/day), and caffeine in milligrams per day (mg/d), was also analyzed from the FFQ answers (see Section 2.4.4).

#### *2.5. Statistical Analysis*

Data were analyzed for normality using the Kolmogorov–Smirnov test. Differences between groups (stages of reproductive aging) were determined using a chi-square test with the Marascuilo procedure to establish the categorical variables. The Kruskal–Wallis test with the Bonferroni adjustment was also used to determine the numerical variables (post hoc test). The presence of VMS in the 4 categories (absence of VMS, presence of hot flashes, presence of night sweats or palpitations, and presence of 2 or more VMS) were reported in frequency and percentage. Differences among categories were determined using a chi-square test with the Marascuilo procedure (post hoc test).

The dependent variable was the presence of VMS in the 4 categories (yes or no). Independent variables included BMI (obese ≥ 30 kg/m2, overweight 25–29.9 kg/m2, and normal 18.5–24.9 kg/m2), fasting glucose (high ≥100 mg/dL or normal <100 mg/dL), blood pressure (hypertensive between ≥130 and ≥80 mmHg, elevated between 120–129 and <80 mmHg, or normal between <120 and <80 mmHg), cardiorespiratory fitness (poor at <400 m/6 min or excellent at ≥400 m/6 min), total fat intake (excessive >30% from total kcal/d or adequate ≤30% from total kcal/d), and tobacco use (yes or no).

A multivariate logistic regression model was used to define statistical models of the nutritional factors (independent variables) associated with VMS (dependent variable). Odds ratios (OR) with 95% confidence intervals (95% CI) were calculated. Several models were proposed: an unadjusted model (Model 1); an adjusted model (Model 2) including covariates (age (years), hormone use (yes or no), and daily physical activity (MET/day)); and an adjusted model (Model 3) including covariates (stage of reproductive aging (late reproductive, menopausal transition, early postmenopause, or late postmenopause), hormone use (yes or no), daily physical activity (MET/d), alcohol intake (mL/day), and caffeine intake (mg/day)) [25–27]. Model 2 was adjusted to include age as a covariate because VMS follows a natural pattern according to lifespan [28]. Model 3 was adjusted to include the stage of reproductive aging because VMS can be present at different stages [4].

Prevalence ratios (PR) were calculated using the following formulas, derived from calculated odds ratios (OR) as previously proposed [29,30], where *p*<sup>1</sup> is the prevalence of VMS in the reference group and z is the coefficient of regression divided by its standard error.

$$\text{PR} = \frac{\text{OR}}{(1 + p\_1 \* [\text{OR} - 1])} \tag{1}$$

$$\text{\\$95\% CI} = \text{OR}^{\left(1 \pm \left(1.96/\text{z}\right)\right)}\tag{2}$$

A *p* value of <0.05 was considered to be statistically significant. All analyses were performed using IBM SPSS® Statistics software, SPSS Inc., Chicago, IL, USA (version 25).

#### **3. Results**

Table 1 compares the descriptive characteristics in women grouped by stages of reproductive aging. A total of 24.4% of women were in the late reproductive stage, 23.4% were in menopausal transition, 36.2% were in early postmenopause, and 16.0% were in late postmenopause. The age of menopause was 48.6 years in women at the early postmenopause stage (*n* = 147) and 43.4 years in women in the late postmenopause stage (*n* = 65), while the average menopausal age (*n* = 212) was 47.0 years (data not shown).

**Table 1.** Descriptive characteristics of women aged 40–65 years grouped according to *STRAW+10*.


BMI: body mass index; SD: standard deviation. Numerical data were analyzed using the Kruskal–Wallis test with the Bonferroni adjustment, and expressed as mean and standard deviation, and median and quartiles (Q1, Q3). Categorical data were analyzed using a chi-square test with the Marascuilo procedure, and expressed as cases and percentages, *n* (%). Superscripts a, b, c, d denote differences among groups. *p* < 0.05 denotes statistical significance.

There was a significant difference between stages of reproductive aging in terms of BMI (0.018), fasting glucose (*p* = 0.002), systolic blood pressure (*p* = 0.001), and total fat intake (*p* = 0.039). There were no differences between groups in cardiorespiratory fitness (*p* = 0.877) and tobacco use (*p* = 0.507). Women in the late reproductive stage showed a lower BMI (27.9 kg/m2), while women in menopausal transition presented a higher BMI (30.6 kg/m2). Fasting glucose was lower in women in the late reproductive stage (95.4 mg/dL) and highest in women in the late postmenopause stage (105.5 mg/dL). Women in the late postmenopause stage showed the highest systolic blood pressure (121.1 mmHg). Total fat intake was higher in women in the late reproductive stage (36.9%), and all groups had intakes above the recommendation of 30% of total energy from fat.

Prevalence of symptoms are reported as four categories (absence, hot flashes, night sweats or palpitations, and 2 or more VMS) (Figure 2). The study reveals that of all participants (*n* = 406), 41.4% experienced an absence of VMS and 29.6% presented with hot flashes only. An absence of VMS prevailed in women in the late reproductive stage (67.7%). Hot flashes were mainly reported in the early postmenopause stage (38.1%) followed by late postmenopause (33.8%). The presence of either night sweats or palpitations was reported mainly in women in the late reproductive stage (11.1%), although the presence of this category did not differ between different stages (*p* = 0.202). A combination of two or more VMS was reported by 23.2% of women in menopausal transition and 29.3% in early postmenopause; only 6.1% of women in the late reproductive stage presented with two or more VMS.

**Figure 2.** Prevalence of vasomotor symptoms. Data were analyzed using a chi-square test with the Marascuilo procedure (post hoc test). They are expressed as percentages (%). \* *p* < 0.01, \*\* *p* < 0.001.

The association between nutritional risk factors and VMS is shown in Table 2 for unadjusted and adjusted models. In Model 1 (unadjusted model), an overweight BMI (25–29.9 kg/m2) denoted a risk for the presence of hot flashes (PR 2.92, 95% CI: 1.66–6.32) (OR 3.24, 95% CI: 1.66–6.33, *p* = 0.001). The presence of either night sweats or palpitations was associated with high levels of fasting glucose (≥100 mg/dL) (OR 2.63, 95% CI: 1.09–6.37, *p* = 0.031) (PR 2.49, 95% CI: 1.09–6.33) and poor cardiorespiratory fitness (OR 15.01, 95% CI: 1.94–115.62, *p* = 0.009) (PR 8.03, 95% CI: 1.95–115.62). In addition, the presence of two or more VMS was associated with high fasting glucose levels (≥100 mg/dL) (OR 2.27, 95% CI: 1.29–3.99, *p* = 0.004) (PR 1.98, 95% CI: 1.29–3.98) and tobacco use (OR 3.19, 95% CI: 1.25–8.11, *p* = 0.015) (PR 2.25, 95% CI: 1.26–8.10).





After adjusting for age, hormone use, and physical activity level, Model 2 showed very similar results to Model 1 (unadjusted). Excessive weight (BMI 25–29.9 kg/m2) was a nutritional risk for the presence of hot flashes (PR 2.93, 95% CI: 1.63–6.51) (OR 3.26, 95% CI: 1.63–6.52, *p* = 0.001). Women were at risk of presenting with either night sweats or palpitations if fasting glucose levels were ≥100 mg/dL (PR 2.38, 95% CI: 1.02–6.11) (OR 2.50, 95% CI: 1.02–6.11, *p* = 0.045) and if they had poor cardiorespiratory fitness levels (PR 7.63, 95% CI: 1.63–112.92) (OR 13.57, 95% CI: 1.63–113.01, *p* = 0.016). Presence of two or more VMS was also associated with high fasting glucose levels (≥100 mg/dL) (PR 1.81, 95% CI: 1.13–3.63) (OR 2.03, 95% CI: 1.13–3.63, *p* = 0.017) and smoking habits (PR 2.12, 95% CI: 1.11–7.55) (OR 2.89, 95% CI: 1.10–7.57, *p* = 0.030).

Model 3 was adjusted to include the stage of reproductive aging, hormone use, physical activity level, alcohol intake, and caffeine intake (Table 2), denoting consistency in results from Models 1 and 2. The presence of hot flashes was associated with an overweight BMI (25–29.9 kg/m2) (PR 2.93, 95% CI: 1.63–6.52) (OR 3.26, 95% CI: 1.63–6.54, *p* = 0.001). Levels of fasting glucose ≥100 mg/dL (PR 2.53, 95% CI: 1.05–6.80) (OR 2.67, 95% CI: 1.04–6.84, *p* = 0.039) and poor cardiorespiratory fitness (PR 8.33, 95% CI: 2.03–129.05) (OR 16.17, 95% CI: 2.02–129.11, *p* = 0.009) are risk factors associated with the presence of night sweats or palpitations. High fasting glucose ≥100 mg/dL (PR 1.75, 95% CI: 1.08–3.48) (OR 1.94, 95% CI: 1.08–3.50, *p* = 0.027) and tobacco use (PR 2.19, 95% CI: 1.15–8.11) (OR 3.05, 95% CI: 1.14–8.13, *p* = 0.025) denoted a risk for the presence of two or more VMS.

There was no association between blood pressure and total fat intake and the presence of VMS in any proposed model (unadjusted or adjusted).

#### **4. Discussion**

This observational study in women aged 40–65 years from Nuevo León state, in northeast Mexico, determined the association between nutritional risk factors and the presence of vasomotor symptoms. Menopausal onset was at 47.0 years; this is younger than women living in central Mexico, as previous studies have found that women from Queretaro started menopause at 49.1 years [5] and those from Mexico City at 50.0 years [6]. The onset of menopause at earlier ages, defined as early menopause (<45 years old), [31] could be due to some of the trigger factors that women reported in this study, such as tobacco use and an overweight or obese BMI [32]. Moreover, it has been demonstrated that menopause at early ages is related to a higher risk of cardiovascular disease and mortality, especially in women aged 50–78 years [33].

The experience of VMS is similar among Mexican women and follows a pattern that predominates in the early postmenopause stage. In this study, 29.6% of all women showed hot flashes, while 38.1% of women in the early postmenopause stage reported having this specific VMS. In a previous study in central Mexico, 21.3% of women aged 40–60 years reported having hot flashes, mainly during menopausal transition [34]. In other countries, such as India, the United Kingdom, Australia, and the United States of America, hot flashes were more frequently reported during the postmenopausal stage [35,36].

The presence of two or more VMS was reported in 22.2% of all women in the current study and was most frequently reported by early postmenopausal women (29.3%). However, previous studies have shown greater incidence of two VMS (hot flashes and night sweats) at the postmenopausal stage; this was found in 50.0% of women aged 24–44 years from the United States of America [11] and in 53.3% of women aged 40–65 years from India [35]. Therefore, in this study, the higher frequency of two or more VMS agrees with the STRAW+10 criteria, which suggests greater a likelihood of symptoms occurring during the early postmenopausal stage [3].

In this study, several women in the late reproductive stage also reported the presence of VMS; 15.2% had hot flashes and 11.1% had either night sweats or palpitations. Previous studies have reported night sweats in 40.0% and hot flashes in 29.0% of late reproductive women aged 35–55 years [37]. It is important to note that VMS can begin at earlier stages, such as the late reproductive stage, although in a low proportion due to estrogen reduction

through constant aging and death of follicles [38]. Thus, assessment and diagnosis of VMS is relevant at earlier stages.

The association between nutritional factors and the presence of VMS was analyzed using three regression models to obtain the odds ratio (OR), and a formula using OR values to obtain the prevalence ratio (PR). These demonstrate the risks of women being overweight, having fasting glucose levels above 100 mg/mL, having a cardiorespiratory fitness below 400 m/6 min, and being a smoker. Adjustment of Models 2 and 3 for covariates did not affect the association. Poor cardiorespiratory fitness was the only risk factor with subtle changes in Model 3 that resulted in slightly higher risk and a wider 95% CI.

Hot flashes were associated with being overweight in this study population (BMI 25–29.9 kg/m2) at PR 2.92–2.93 and OR 3.24–3.26 (*p* = 0.001 in all models). Overweight and obese BMIs have previously been associated with the presence of VMS, as reported in the clinical practice guidelines [7,8] and in some studies from Scotland [10], North America [27,39], Australia [40], and South Korea [26]. Women showing a higher BMI tend to have an excess of body fat; visceral fat increases by up to 20% during the postmenopausal stage [41]. In this study, obesity was not associated with hot flashes, which may be because more obese women were at later stages, such as the late postmenopause stage, in which hot flashes were less frequently reported, similar to a previous study [39]. Excessive fat does not allow heat conduction through the skin; therefore, the body tries to release it by maximizing vasodilation, which increases the central body temperature beyond the sweating threshold [13,42]. However, our findings may also suggest that the mechanisms of estrone, associated with a decrease in hot flashes, are naturally occurring in our study of postmenopausal women [39,43].

Fasting glucose levels above 100 mg/dL were associated with either night sweats or palpitations and with two or more VMS. A Swedish study in women aged 50–64 years reported night sweats as the only VMS associated with high glucose levels (*p* < 0.05) [44]. In a longitudinal study of Australian women aged 45–50 years, it was stated that there was a significant association between night sweats and diabetes (OR 1.91, 99% CI: 1.08–3.35, *p* < 0.001) in an adjusted model including similar covariates as this study: age, educational level, length of time, BMI, physical activity level, tobacco use, alcohol intake, menopausal status, and hormone use [45]. Therefore, chronic hyperglycemia could be a strong associated factor. High levels of glucose have been associated with insomnia, because the hypothalamic-pituitary-adrenal axis is altered [46]. This axis also regulates steroid and adrenal secretions [47], so its alteration impacts VMS occurrence.

The association between high blood pressure and VMS could be due to the increased activity of the sympathetic nervous system [48], and thus the increased activity of the adrenalin and sweat glands [49]. Although models in this study showed non-significant results when systolic and diastolic blood pressures were elevated or at hypertension levels, other authors have reported significant associations [44,50]. A study from Sweden in postmenopausal women, aged 50–64 years, reported an association between the presence of night sweats and systolic blood pressure (OR 2.07, *p* < 0.001) [44]. In addition, women in the United States of America aged 45–54 years who were under antihypertensive treatment had 1.80 times greater risk of presenting with hot flashes (*p* <0.05) [50].

Fat intake was not associated with the presence of VMS in any model (unadjusted or adjusted). Intake of foods high in fat, especially saturated fat obtained from animal sources such as red or white meat, can increase levels of LDL cholesterol (low-density lipoprotein) [51], and therefore increase the presence of VMS [44]. This is supported by findings of studies in Australian women aged 45–50 years, in which those who followed a high-fat diet presented a significant risk for the presence of both hot flashes and night sweats (unadjusted model, OR 1.16, *p* = 0.002) [52]. Additionally, a study of women aged 40–85 years in the United States of America, who were in menopausal transition and postmenopause, VMS were higher among those who consumed more red and white meat, seafood, and dairy, while an absence of VMS was presented among women consuming a plant-based diet [53].

It has been suggested that a decrease in VMS may be observed with improving cardiorespiratory fitness [54,55]. Cardiorespiratory fitness was associated with improved health (*p* < 0.001), emotions (*p* = 0.05), and occupational quality of life (*p* = 0.03), suggesting a positive effect on reduced menopausal symptoms in women aged 45–60 years in the United States of America [55]. However, there has been no reported association with hot flashes and night sweats in Spanish women aged 45–60 years [56]. In the present study, poor cardiorespiratory fitness was the nutritional factor with the highest association to either night sweats or palpitations (PR 7.63–8.33; OR 13.57–16.17, *p* < 0.05). Women who had been physically active since menopausal transition had more protection from VMS. A higher cardiorespiratory fitness level decreases the activation of the sympathetic nervous system, which narrows the blood vessels of the body; in turn, the thermoneutral zone maintains homeostasis [57].

Tobacco use has been reported by different studies as a nutritional risk factor for VMS. Cigarettes are composed of different metals that serve as endocrine disruptors which alter the hormonal balance, so estrogens can be affected until they trigger VMS [12,58]. Women that have been smoking for many years may be at increased risk of having VMS. In a longitudinal study, current smokers aged 24–44 years showed 2.5 times greater risk of having hot flashes and night sweats (95% CI: 1.5–5.3, *p* <0.05) after adjusting for age, hormone level, BMI, hormone use, marital status, and parity [11]. A study of women in the late reproductive stage and the menopausal transition stage from the United States of America, aged 42–52 years, showed an association between the presence of VMS (hot flashes and night sweats), the number of cigarettes smoked (unadjusted model, OR 1.6, 95% CI: 1.3–1.9, *p* < 0.05), and passive smoke exposure (unadjusted model, OR 1.3, 95% CI: 1.2–1.4, 0.05) [59].

#### *Strengths and Limitations*

The main strength of the current study is that it has determined the association between nutritional risk factors and the presence of vasomotor symptoms in women aged 40–65 years from Nuevo León state, in northeast Mexico. The analysis of prevalence and risk is also a strength of this study as it avoids over- or underestimation of this association; however, data could not be compared against other publications. A limitation of this study is the lack of precision in the responses of presence of VMS, as women were questioned in a dichotomous way (yes or no) without considering frequency or intensity. The analysis of some nutritional variables, such as blood pressure and cardiorespiratory function, was limited due to an insufficient number of cases when assessing the categories of VMS. Intakes of subtypes of fat were not analyzed. For future research, it is recommended that women present a similar intake of alcohol and caffeine for more uniformity in those covariates, which may achieve greater precision in adjusted models. In addition, longitudinal studies are needed to infer specific causes or to determine strong risk factors for this specific population.

#### **5. Conclusions**

The presence of VMS was associated with different nutritional risk factors (weight, fasting glucose levels, cardiorespiratory fitness, and tobacco use) in women living in northeast Mexico. This association was independent of covariates including age, stage of reproductive aging, hormone use, reported physical activity, alcohol intake, and caffeine intake. This evidence supports the need for updating these nutritional risk factors in clinical practice guidelines and its application as an instrument in primary health care services, which assist women at different stages of reproductive aging, at local and national levels.

**Author Contributions:** A.T., Y.B., E.S.-P. and J.A.T. designed the study; A.T. and Y.B. wrote the protocol; A.T. recruited the participants and collected samples; Y.B. and J.L.J. conducted the statistical analysis; A.T., Y.B. and J.A.T. wrote the first draft of the manuscript, and all authors (A.T., Y.B., E.S.-P., R.S., J.L.J., V.L., E.R., R.P., C.B. and J.A.T.) commented on previous versions of the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** C.B. and J.A.T. are funded by the Instituto de Salud Carlos III through the CIBEROBN CB12/03/30038, which is co-funded by the European Regional Development Fund. The funding sponsors had no role in the design of the study, in the collection, analysis, or interpretation of the data, in the writing of the manuscript, or in the decision to publish the results.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the ethics committee of the Facultad de Salud Pública y Nutrición of the protocol ID: 15–FaSPyN–SA–11.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study. The results and writing of this manuscript followed the Committee on Publication Ethics (COPE) guidelines on how to deal with potential acts of misconduct, maintaining integrity of the research and its presentation following the rules of good scientific practice, the trust in the journal, the professionalism of scientific authorship, and the entire scientific endeavor. Written informed consent has been obtained from the patient(s) to publish this paper.

**Data Availability Statement:** There are restrictions on the availability of data for this trial, due to the signed consent agreements around data sharing, which only allow access to external researchers for studies following the project purposes. Those wishing to access the trial data used in this study can make a request to pep.tur@uib.es.

**Acknowledgments:** The authors especially thank the participants for their enthusiastic collaboration and the personnel for outstanding support and exceptional effort. CIBEROBN is an initiative of Instituto de Salud Carlos III, Spain.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

