Cardiovascular disease (CVD) is rare among women of reproductive age. However, it is a main cause of mortality in menopausal women.¹ Hormonal changes caused by menopause produce major physiological changes. Estrogen is a representative female hormone that is essential for female sexual development and growth.² Estrogen deficiency after menopause contributes to fat redistribution, decreases maximal oxygen consumption, and causes changes in lipid metabolism.^3,4 Furthermore, during the reproductive years, the ovaries produce estradiol (E2) in abundance using low-density lipoprotein cholesterol (LDL-C) as a substrate, and that plays an important role in protecting the cardiovascular system.⁵ After menopause, the expression of E2 decreases rapidly, which causes not only sleep disorders, depression,⁶ osteopenia, and osteoporosis⁷ but also cardiovascular and metabolic diseases and obesity.⁸ That series of adverse changes correlates closely with post-menopausal syndrome. The negative changes can be caused directly by ovarian failure or indirectly by the metabolic consequences of central fat redistribution that occurs with estrogen deficiency.⁹

Several studies have shown that estrogen deficiency increases total cholesterol (TC), LDL-C levels and decreases high-density lipoprotein cholesterol (HDL-C) levels in post-menopausal people.^10-15 Post-menopausal people also have high levels of apolipoprotein (Apo)-B, which is a strong marker of coronary atherosclerosis.^16,17 Changes in appetite and physical and psychological changes such as depression can all occur due to menopausal factors. Therefore, the risk of CVD in post-menopausal people can be influenced by several factors.¹⁸

Previous studies have shown that medical nutrition therapy (MNT) not only affects weight management but also regulates blood lipid levels.^19-21 Some studies^22-24 have discussed the effect of a low-fat diet on blood lipid levels in pre- and post-menopausal women. Those results showed that the same intervention had different effects based on menopausal status. However, most study designs used only one group without a control group. Moreover, few studies about the effects of MNT on blood lipid levels have considered menopausal status.

We hypothesized that the effects of MNT on CVD risk factors would differ according to menopausal status because menopause has hormonal, behavioral, and psychological effects. To test that hypothesis, we examined the effects of MNT, in terms of changes to CVD risk factors, and compared those effects between pre- and post-menopausal women with borderline dyslipidemia in Korea.

Study design and ethical approval

This was a 12-week designed-intervention study. Prior to study initiation, informed written consent was obtained from all participants. The study protocol was approved by the Institutional Review Board of Kyung Hee University Medical Center (approval no.: KMC IRB 1227-02).

Participants

Participants were recruited between October 2013 and January 2014 via posters on a bulletin board at Kyung Hee University Medical Center in Seoul, South Korea. The inclusion criteria were as follows: (1) pre- or post-menopausal women aged 20 to 60 years, with menopausal status assessed using a questionnaire; (2) borderline dyslipidemia according to the Korean dyslipidemia diagnostic standards (TC of 200–239 mg/dL or LDL-C of 130–159 mg/dL or triglyceride [TG] of 150–199 mg/dL). Participants with the following criteria were excluded: lipid-lowering medications; a history of chronic cardiovascular, kidney, or liver disease; metabolic disturbances with a thyroidal component; pancreatic diseases and diabetes; alcoholic disease; hormone-replacement therapy; pregnancy or breast feeding; and refusal to provide consent or follow all trial procedures. In total, 91 Korean women who filled out the consent form were recruited, and 15 participants were excluded according to the exclusion criteria. Therefore, 76 women participated in the study. Both pre- and post-menopausal participants were randomly assigned to the control group and MNT group. Two participants from the MNT group of pre-menopausal women withdrew their consent, so 74 people completed the study.

Intervention protocol

MNT was conducted using an intervention mapping method with the decision tree approach, which can help facilitate effective decision making among healthcare professionals during an intervention process. The targeted dietary message for each participant was designed based on her body weight, blood lipid level, and diet. For body weight, normal was defined as a body mass index (BMI) of <23 kg/m² and high as a BMI of ≥23 kg/m².²⁵ Blood lipid profile levels were classified as normal and high according to diagnostic criteria.²⁶ Participants with a TC level of ≥200 mg/dL, LDL-C level of ≥130 mg/dL or higher, or TG level of ≥150 mg/dL were classified as high. Diet was assessed according to food intake, meal patterns, and usual eating habits and classified as normal or low-quality based on the Dietary Reference Intake for Koreans (KDRIs) 2010.²⁷ Low-quality diets were characterized by excessive consumption of refined carbohydrates and high-fat foods or insufficient intake of vegetables and fruits.²⁷ After each participant’s weight, blood lipid level, and dietary status had been evaluated, the target message for her status was designed. Prior to the start of the study, trained nutritionists produced an educational booklet using information from the KDRIs²⁷ and the Korean Guidelines for the Management of Dyslipidemia, 2nd edition.²⁶ Six times during the study (baseline visit and in the 1st, 3rd, 6th, 9th, and 12th weeks), a nutritionist used the booklet to train the participants according to their individualized target messages, as shown in Table 1.

Key outcome measurements

Anthropometric measurements

Anthropometric measurements were recorded at baseline and in the 6th and 12th weeks. Height, weight, BMI, percent body fat, and fat free mass were evaluated via a bioimpedance analysis (Inbody 3.0; BioSpace) with light clothing and without shoes. Waist circumference (WC) was measured using a flexible non-elastic tape midway between the 10th rib and the iliac crest while the participant was standing, and hip circumference (HC) was measured at the widest horizontal diameter of the buttock. The waist-to-hip ratio (WHR) was derived from the WC and HC.

Blood pressure measurements

Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were evaluated at baseline and in the 6th and 12th weeks. All participants were evaluated twice per visit using an automated blood pressure monitor while in a sitting position after resting for more than 15 minutes. The mean value of the two measurements was used in the analyses.

Biochemical measurements

Blood samples were collected from the midarm vein after 12 hours of overnight fasting at baseline and 12 weeks. The samples were separated into ethylenediamine tetraacetic acid potassium anticoagulant tubes and serum separate tubes (SSTs). After being allowed to clot for 30 minutes, the SSTs were centrifuged (3,000 ×g, 4 °C, for 10 minutes), and the supernatant was used for analysis. The TG, TC, HDL-C, LDL-C, C-reactive protein, and glucose levels were analyzed using an automatic analysis system (Modular Analytics; Roche). Very-low-density lipoprotein cholesterol (VLDL-C) was assessed using the Friedewald equation: VLDL=TG/5.²¹ The atherogenic index (AI) is an important predictive biomarker of cardiovascular risk: AI=(TC–HDL-C)/HDL-C.²⁸ Insulin content and secretion were evaluated using a radioimmunoassay. Insulin resistance was estimated with the homeostatic model assessment for insulin resistance (HOMA-IR) formula: HOMA-IR=[fasting insulin level (μIU/mL)×fasting glucose level (mmol/L)]/22.5.

Dietary assessments

All participants were instructed to complete a 3-day (2 weekdays, 1 weekend) food record at baseline and 12 weeks. A nutritionist trained all participants how to write a food record, and they then recorded all information about their food intake throughout each day, including stocks, side dishes, snacks, fruits, and beverages. A nutritionist analyzed the food record through a 1:1 interview with each participant using food models, measuring spoons, and dimensional food containers. Dietary intakes were quantified by a nutritionist using a computer-aided nutritional analysis program (CAN pro version 4.0; Korean Nutrition Society).

Statistical analyses

The collected data were analyzed using Statistical Package for the Social Sciences software version 22.0 (IBM Cooperation). Categorical variables were compared with the χ² test. Continuous and discrete variables assessed via descriptive analyses are expressed as the mean±standard deviation or standard error and number (%), respectively. The normality of all variables was assessed before the statistical analyses using the Kolmogorov-Smirnov test. Two-sample t-testing was used to evaluate significant differences between groups at both examinations. Testing for changes from baseline to the end of the 12-week intervention was performed using paired-sample t-tests. Two-sided P-values of <0.05 were considered statistically significant.

Descriptive characteristics of the participants according to menopausal status

Table 2 shows the descriptive characteristics of the participants according to their menopausal status. Age, height, and body composition, such as percent body fat and WHR, differed significantly according to menopausal status. Post-menopausal participants (mean, 51.4±3.4 years [range, 45.0 to 59.0]) were significantly older than pre-menopausal participants (mean, 39.6±6.3 years [range, 22.0 to 48.0], P<0.001). Pre-menopausal participants were significantly taller than post-menopausal participants (159. 6±5.1 cm vs. 157.0±5.8 cm, P<0.01). Post-menopausal participants had a significantly higher body fat percentage than pre-menopausal participants (34.2%±4.1% vs. 30.5%±5.5%, P<0.01). The WHR (0.89±0.04 vs. 0.86±0.06, P<0.05) also differed significantly between the pre- and post-menopausal participants. Post-menopausal participants had a significantly higher TC (200.9±28.5 mg/dL vs. 214.1±20.1 mg/dL, P<0.05) and LDL-C (126.8±22.3 mg/dL vs. 137.5±20.8 mg/dL, P<0.05) than pre-menopausal participants. Post-menopausal participants also had significantly higher Apo-B than pre-menopausal participants (92.4±17.8 mg/dL vs. 101.6±17.2 mg/dL, P<0.05). In addition, post-menopausal participants had significantly higher glucose levels than pre-menopausal participants (89.3±9.4 mg/dL vs. 96.1±9.7 mg/dL, P<0.05). On the other hand, weight, BMI, fat free mass, fat mass, WC, HC, SBP, DBP, TG, HDL-C, VLDL-C, free fatty acid levels, Apo-A1, and insulin levels did not differ significantly between the pre- and post-menopausal participants.

Comparison of MNT effects on dietary intake according to menopausal status

Table 3 shows the effects of MNT on dietary intake according to menopausal status. Among pre-menopausal participants, energy, carbohydrate, protein, fat, fatty acid, cholesterol, and micronutrient intakes did not differ significantly between the control and MNT groups at baseline. After 12 weeks of intervention, the intake of energy (from 1,782.6±586.3 to 1,454.3±300.9 kcal, P<0.05), fat (from 50.9±23.1 to 42.6±14.5 g, P<0.05), total fatty acids (TFA) (from 37.5±11.4 to 21.5±5.4 g, P<0.001), saturated fatty acids (SFA) (from 10.9±4.0 to 5.4±2.5 g, P<0.001), monounsaturated fatty acids (MUFA) (from 15.1±5.2 to 7.5±2.7 g, P<0.001), polyunsaturated fatty acids (PUFA) (from 11.5±3.9 to 8.6±3.4 mg, P<0.01), cholesterol (from 359.6±194.1 to 229.8±59.7 mg, P<0.05), sodium (from 4,042.3±1,327.2 to 3,284.2±1,031.5 mg, P<0.01), potassium (from 2,558.6±880.2 to 2,087.9±738.5 mg, P<0.05), and fiber (from 18.3±5.7 to 15.0±4.4 g, P<0.05) all decreased significantly in the MNT group. No significant changes were found in the control group. After the intervention, the MNT group had a lower intake of energy (1,727.8±358.4 kcal vs. 1,454.3±300.9 kcal, P<0.05), carbohydrates (259.1±41.4 g vs. 210.6±43.0 g, P<0.01), TFA (31.4±11.3 g vs. 21.5±5.4 g, P<0.01), SFA (9.9±5.5 g vs. 5.4±2.5 g, P<0.01), MUFA (12.5±5.1 g vs. 7.5±2.7 g, P<0.01), cholesterol (333.0±132.8 mg vs. 229.8±59.7 mg, P<0.01), and potassium (2,596.9±745.3 mg vs. 2,087.9±738.5 mg, P<0.05) than the control group.

Among post-menopausal participants, the intake of energy, carbohydrates, protein, fat, fatty acids, cholesterol, and micronutrients did not differ significantly between the control and MNT groups at baseline. After 12 weeks of intervention, the intake of energy, carbohydrates, and fat did not change significantly in either group. However, the intake of TFA (from 33.1±14.1 to 19.6±7.7 g, P<0.001), SFA (from 12.3±7.9 to 6.1±4.3 g, P<0.01), MUFA (from 13.8±7.7 to 7.7±4.8 g, P<0.01), PUFA (from 10.4±5.7 to 6.9±2.8 g, P<0.01), and cholesterol (from 276.6±132.1 to 182.0±96.6 mg, P<0.001) decreased significantly in the MNT group. Only the PUFA intake decreased (from 10.6±5.8 to 8.4±3.9 g, P<0.05) in the control group. After 12 weeks of intervention, the MNT group had a lower intake level of TFA (27.8±10.1 g vs. 19.6±7.7 g, P<0.01), MUFA (11.9±5.7 g vs. 7.7±4.8 g, P<0.01), and cholesterol (253.5±79.2 mg vs. 182.0±96.6 mg, P<0.01) than the control group.

Comparison of MNT effects on anthropometric measurements according to menopausal status

Table 4 depicts the effects of MNT on anthropometric measurements according to menopausal status. No significant differences in the anthropometric measurements were found between the control and MNT groups at baseline, regardless of menopausal status.

The WC of pre-menopausal participants decreased significantly (from 82.1±11.7 to 80.6±11.7 cm, P<0.05) after the 12-week intervention, but BMI, body fat percentage, and WHR did not change significantly in the MNT group. The BMI of post-menopausal participants did not change even after 12 weeks of intervention. However, the WC (from 83.9±5.0 to 82.2±5.1 cm, P<0.05), WHR (from 0.89±0.03 to 0.87±0.04, P<0.05), and body fat percentage (from 34.3%±3.8% to 33.4%±4.6%, P<0.01) all decreased significantly in the MNT group. Among the pre- and post-menopausal participants in the control group, none of the anthropometric measurements changed significantly after 12 weeks.

Comparison of MNT effects on biochemical measurements according to menopausal status

Table 5 shows the effects of MNT on biochemical measurements according to menopausal status. In both pre- and post-menopausal participants, blood lipid profiles and blood glucose, insulin, and HOMA-IR did not differ significantly between the control and MNT groups at baseline.

TG in pre-menopausal participants decreased significantly after the 12-week intervention (from 107.1±40.0 to 91.2±32.6 mg/dL, P<0.05). TC, LDL-C, and HDL-C did not change significantly after the intervention, but VLDL-C (from 22.5±11.4 to 18.1±6.7 mg/dL, P<0.05) and AI (from 2.5±0.8 to 2.3±0.8, P<0.05) decreased. None of the other parameters changed significantly after the intervention.

The TC of post-menopausal participants decreased significantly after the 12-week intervention (from 214.6±24.8 to 200.6±25.2 mg/dL, P<0.05). Moreover, the MNT group had a lower TC (217.8±26.5 mg/dL vs. 200.6±25.2 mg/dL, P<0.05), LDL/HDL ratio (2.5±0.9 vs. 2.2±0.6, P<0.05), AI (2.7±1.0 vs. 2.5±0.7, P<0.05), and Apo-B/A ratio (0.6±0.2 vs. 0.6±0.1, P<0.05) than the control group. None of the other indicators changed significantly between the beginning and end of the intervention.

In this study, we investigated the effects of an MNT intervention on CVD risk factors in pre- and post-menopausal participants with borderline dyslipidemia. Interestingly, the results show that even when the same intervention was provided, the pattern of changes in dietary intake and anthropometric and biochemical indicators differed according to menopausal status.

The results of our study are consistent with those of a previous one²⁹ that showed that post-menopausal women had a higher risk of CVD than pre-menopausal women. At baseline, BMI did not differ significantly by menopausal status. However, post-menopausal participants had a higher body fat percentage (30.5%±5.5% vs. 34.2%±4.1%) and WHR (0.86±0.06 vs. 0.89±0.04) than pre-menopausal participants. In addition, post-menopausal participants had significantly higher TC (200.9±28.5 mg/dL vs. 214.1±20.1 mg/dL), LDL-C (126.8±22.3 mg/dL vs. 137.5±20.8 mg/dL), Apo-B (92.4±17.8 mg/dL vs. 101.6±17.2 mg/dL), and fasting blood glucose (89.3±9.4 mg/dL vs. 96.1±9.7 mg/dL) levels than pre-menopausal participants at baseline. Based on the characteristics of the participants, menopausal women are more at risk for CVD than pre-menopausal women.

After the intervention, energy and fat intake decreased significantly in our pre-menopausal participants. The MNT group had significantly lower energy (1,727.8±358.4 kcal vs. 1,454.3±300.9 kcal) and carbohydrate (259.1±41.4 g vs. 210.6±43.0 g) intakes than the control group. However, the percentage of calories from carbohydrates, protein, and fat (C:P:F ratio) was within the normal range both before and after the intervention. Among post-menopausal participants, by contrast, energy, carbohydrate, and fat intake did not change significantly after the intervention, and no intake changes occurred in the control group. The C:P:F ratio was within the normal range for post-menopausal participants regardless of the intervention. The significant decrease in energy intake after MNT in pre-menopausal participants alone was attributed to the fact that our post-menopausal participants already had a lower energy intake at baseline, and the number of participants was too low to find statistically significant changes after MNT. Therefore, the change after MNT was not significant. Fat intake did not decrease significantly after MNT in any group because it was not high at baseline. Recently, it has become clear that the quality of dietary fat is also important.³⁰ The pattern of changes in fatty acid intake between pre- and post-menopausal participants was thus examined. Although no changes in the patterns of fatty acid intake were found in the control group, the intake of TFA and SFA in the MNT group decreased and was within the recommended intake of 7% of total calories. The intake of MUFA and PUFA also decreased significantly, probably due to the decrease in TFA intake. No change in fatty acid intake patterns was found in the control group. Cholesterol intake decreased after the intervention in both pre-menopausal and post-menopausal participants, and the intake suggested by KDRIs, <300 mg/day,²⁹ was lower than found in the control group. Based on those results, the 12-week MNT had a positive effect on the overall dietary intake of the participants.

BMI did not change significantly after the intervention in the pre- or post-menopausal participants. However, WC decreased significantly in both the pre- (from 82.1±11.7 to 80.6±11.7 cm) and post-menopausal participants (from 83.9±5.0 to 82.2±5.1 cm). Moreover, WHR (from 0.89±0.03 to 0.87±0.04) and body fat percentage (from 34.3%±3.8% to 33.4%±4.6%) decreased significantly in post-menopausal participants after the intervention. The significant change in WC and WHR could be attributed to the effects of MNT on fat distribution. WC is more closely correlated with dyslipidemia than BMI because it reflects abdominal and visceral fat.³¹ Post-menopausal participants had a higher WHR and body fat percentage than the pre-menopausal participants at baseline. Therefore, MNT was effective in reducing CVD risk. SBP and DBP did not change significantly after MNT in the pre- or post-menopausal participants, but all values were within the normal range.

We found differential effects from MNT on blood cardiovascular parameters according to menopausal status. In the pre-menopausal participants, TG (from 107.1±40.0 to 91.2±32.6 mg/dL), VLDL-C (from 22.5±11.4 to 18.1±6.7 mg/dL), and AI (from 2.5±0.8 to 2.3±0.8) decreased significantly after MNT. In the post-menopausal participants, TC (from 214.6±24.8 to 200.6±25.2 mg/dL), LDL/HDL (from 2.6±0.6 to 2.2±0.6), AI (from 2.9±0.8 to 2.5±0.7), and Apo-B/A (from 0.7±0.2 to 0.6±0.1) all decreased significantly after MNT. The decrease in blood TG in pre-menopausal participants was attributed to decreased dietary energy intake. After the intervention, LDL-C did not decrease in either pre- or post-menopausal participants. However, the LDL/HDL ratio, a major factor predicting the risk of CVD,³² decreased significantly in post-menopausal participants. AI is also a predictor of CVD, and it decreased after MNT in both pre- and post-menopausal participants.³³ Apo-A1 is the largest component of HDL. Apo-B is present in chylomicron, and it reflects the level of VLDL-C and LDL-C in the body.³⁴ Thus, Apo-B/A is a major clinical indicator of metabolic syndrome and CVD, and previous studies have shown that it is a better indicator of CVD than conventional lipid profiles.³⁵

Pre-menopausal women have a weak postprandial lipid response, which is attributed to the higher clearance capacity caused by an increase in lipoprotein lipase activity in that population.³⁶ After menopause, the postprandial lipid response is more prominent. Thus, because of their lower initial lipid levels and potentially attenuated lipid responses, the effect of interventions based on dietary fat intake might theoretically have fewer beneficial effects in post-menopausal women.³⁶ However, in this study, although the pattern differed according to menopausal status, the cardiometabolic risk factors improved, thereby indicating that MNT was effective.

This study had some limitations. The sample size was small, and dietary data might have been underreported. That is, because nutrition education was conducted within a certain period, individuals might have reported what they have learned rather than what they actually ate. In addition, this study examined only a 12-week period, so longer-term studies are needed to ascertain whether the results would continue to follow the same trends.

The main strength of this study is that it was an intensive intervention trial. This study has shown that lifestyle change interventions can be a cost-effective treatment that reduces the risk of dyslipidemia via permanent lifestyle changes.

In summary, a 12-week customized dietary intervention significantly improved the risk factors of CVD in Korean women with borderline dyslipidemia. After MNT, WC, TG, VLDL-C, and AI changed significantly in pre-menopausal participants. In the post-menopausal participants, WC, WHR, body fat percentage, TC, LDL/HDL ratio, AI, and Apo-B/A changed significantly. Therefore, dietary intervention is important for women with borderline dyslipidemia, and the approach should be individualized based on menopausal status.

The authors declare no conflict of interest.

This work was supported by National Research Foundation of Korea (NRF) grant funded by the Korea goverments (MIST) (NRF-2022R1A2C1003746).

Study concept and design: HL (Hansongyi Lee) and HL (Hyunjung Lim); acquisition of data: HL (Hansongyi Lee); analysis and interpretation of data: HL (Hansongyi Lee); drafting of the manuscript: HL (Hansongyi Lee) and KS; critical revision of the manuscript: HL (Hansongyi Lee), KS, IL, and HL (Hyunjung Lim); statistical analysis: HL (Hansongyi Lee); obtained funding: HL (Hyunjung Lim); administrative, technical, or material support: HL (Hansongyi Lee); and study supervision: HL (Hyunjung Lim).