PDF
Citation
Print
INTRODUCTION
The average life expectancy and aging population in Korea are increasing. As a result, there is a growing interest in personal health, and health management in middle-aged adults is becoming an important issue, as disease susceptibility is higher within this age bracket, and metabolic syndrome (MetS) is related to a heightened risk of major diseases, such as cardiovascular disease and type 2 diabetes, and mortality risk.
MetS is a syndrome marked by the presence of various cardiovascular risk factors, such as abdominal obesity, dyslipidemia, hyperglycemia, and hypertension. MetS is a primary public health problem worldwide. Based on data from Statistics Korea [1], the occurrence of MetS rises with age in adults aged 40–64 yrs, underscoring the importance of lifestyle management to reduce the risk of MetS [2].
Smoking, alcohol consumption, physical activity, sleep duration, and stress perception play significant roles in the rising prevalence of MetS. Smoking and alcohol consumption elevate the risk of MetS, and smoking is positively related to hypertriglyceridemia, hyperglycemia, and hypertension [34]. In contrast, regular physical activity reduces the risk of developing MetS. Insufficient or excessive sleep durations increase the risk of MetS by more than 45% in middle-aged adults [5], and stress is positively correlated with the risk of MetS [6]. Furthermore, MetS is closely related to lifestyle factors and diet. Since middle-aged individuals are in a period when lifestyle factors tend to stagnate, underscoring the need to develop strategies to improve lifestyle.
The impact of an individual’s lifestyle factors on MetS is complex [7]. Therefore, it is essential to consider these factors comprehensively in the prevention of MetS. The latent class analysis (LCA) is useful for analyzing these complex influences. It enhances the reliability of research results as it does not assume specific patterns in advance but identifies patterns based on actual data. By deriving latent clusters from observed categorical indicators, this method allows for the identification of the independent effects of each lifestyle factor and the complex patterns that arise from the combination of multiple factors [8].
Poor diet quality is related to a heightened risk of MetS [910]. An unhealthy combination of lifestyle habits is closely related to poor eating habits [11]. Thus, evaluating the impact of dietary habits according to lifestyle factors is crucial to developing effective strategies to manage MetS by changing these habits.
In Korea, numerous studies have been conducted on MetS, including some or all of the 5 lifestyle factors—smoking, alcohol consumption, sleep duration, physical activity, and stress perception—or analyzing them alongside other variables [1213]. Additionally, studies that dichotomize some of these lifestyle factors for analysis also exist [141516]. However, to the best of our knowledge, no study has dichotomized all 5 lifestyle factors to derive lifestyle patterns and analyzed their association with MetS.
Although the relationship between lifestyle factors and MetS has been well-established [3456], little is known about whether diet quality indicators mediate the relationship between lifestyle patterns and MetS. The goal of this study was to derive lifestyle patterns through LCA among middle-aged (age 40–64 yrs) individuals at increased risk of onset of MetS and analyzed the relationship between these patterns and MetS through the Korean Healthy Eating Index (KHEI).
SUBJECTS AND METHODS
Selection of research participants
This study analyzed data from the eighth Korea National Health and Nutrition Examination Survey (KNHANES) (2019–2021). Out of 22,559 participants, 8,549 participants aged 40 to 64 yrs were selected. The exclusion criteria were individuals who consumed fewer than 500 kcal or over 5,000 kcal per day (n = 1,602); those with cardiovascular diseases (stroke, myocardial infarction, angina pectoris) (n = 500); and those with missing data on any variable (n = 1,221). Participants with cardiovascular disease were excluded from the study because they were likely to already have major risk factors for metabolic syndrome, such as hypertension and dyslipidemia. This exclusion was made to prevent potential bias in analyzing the relationship between lifestyle patterns and MetS through the KHEI [17]. In this study, a total of 5,196 participants (2,062 male and 3,134 female) were involved in the analysis. The flowchart of patient selection is shown in Fig. 1. This study was conducted after receiving approval from the Institutional Review Board of Wonkwang University (WKIRB-202409-SB-062).
KHEI
The KHEI contains 14 items divided into 3 categories: adequacy, moderation, and balance of energy intake. The “adequacy” category includes 8 items: “Have breakfast,” “Mixed grains intake,” “Total fruits intake,” “Fresh fruits intake,” “Total vegetable intake,” “Vegetables intake excluding Kimchi and pickled vegetables intake,” “Meat, fish, eggs and beans intake,” and “Milk and milk products intake.” The “moderation” category comprises 3 items: “Percentage of energy from saturated fatty acid,” “Sodium intake,” and “Percentage of energy from Sweets and beverages.” The “balance of energy intake” category contains 3 items: “Percentage of energy from carbohydrate,” “Percentage of energy from fat,” and “Energy intake.” Each item is scored on a scale of 5 or 10 points, with a maximum score of 100 points, with higher scores reflecting better dietary quality [18].
Diagnosis of MetS
The diagnosis of MetS followed the criteria established by the Modified National Cholesterol Education Program Adult Treatment Panel III of the American Heart Association and the National Heart, Lung, and Blood Institute. High triglyceride levels was defined as triglyceride levels of 150 mg/dL or higher. High blood pressure was determined as systolic blood pressure of ≥ 130 mmHg or diastolic blood pressure of ≥ 85 mmHg. Imparied fasting glucose was determined as a fasting blood glucose level of 100 mg/dL or higher. Low high-density lipoprotein cholesterol (HDL-C) was determined as HDL-C levels below 40 mg/dL in men and below 50 mg/dL in women. Abdominal obesity was determined as a waist circumference of ≥ 90 cm in men and ≥ 85 cm in women according to the criteria of the Korean Society for the Study of Obesity. MetS was determined as the occurrence of three or more of the following conditions: hyperglycemia or glycated hemoglobin (HbA1c) of ≥ 5.7% or a physician’s diagnosis of diabetes; hypertension or a physician’s diagnosis of hypertension; low HDL-C or a physician's diagnosis of dyslipidemia; abdominal obesity; and hypertriglyceridemia [19]. The MetS variable was dichotomized as ‘MetS’ (coded as 1) for individuals meeting 3 or more diagnostic criteria, and ‘not MetS’ for those meeting fewer than 3.
Lifestyle factors
Five lifestyle factors were analyzed: smoking, alcohol consumption, sleep duration, physical activity, and stress perception. These factors were dichotomized for analysis based on the study by Wang et al. [20]. Smokers were classified as people who have smoked more than 100 cigarettes (5 packs) in their lifetime. Alcohol drinkers were classified as persons who consumed 2 or more alcoholic beverages (for men) and 1 or more alcoholic beverages (for women) when they drank. Sleep duration of under 6 h or over 8 h per day during weekdays was considered unhealthy. Insufficient physical activity was considered as performing under 150 min of moderate-intensity physical activity or under 75 min of vigorous-intensity workouts weekly. Stress perception was assessed by posing a question regarding perceived stress, and those who responded “very much” or “a lot” were considered to be in a state of high stress.
General characteristics
Data on sociodemographic characteristics and lifestyle factors were analyzed. The sociodemographic factors included sex, age, body mass index (BMI), household income, and marital status. Household income levels were classified as “low,” “intermediate,” “high,” and “very high.” Marital status was defined as “married” or “unmarried.” Health behaviors comprised smoking, alcohol consumption, physical activity duration, sleep duration, and stress perception. For Alcohol consumption, individuals in the category “ever consumed alcohol” were classified as alcohol drinkers, while those in the categories “never consumed alcohol,” “not applicable,” or “do not know/no response” were classified as non-drinkers. Individuals who smoked daily or occasionally were classified as smokers, while those who answered “used to smoke but do not currently smoke,” “not applicable,” or “do not know/no response” were classified as non-smokers. The calculation of low-density lipoprotein cholesterol levels was performed using the Friedewald formula.
Statistical analysis
Mediating effects were analyzed using SPSS PROCESS Macro version 4.2, using Hayes’ model 4 as the mediating model. While the SPSS PROCESS Macro is an efficient tool for analyzing mediation effects, it does not directly account for the complex sample design of KNHANES. LCA was performed using Mplus version 8.11 to classify the study into latent classes according to lifestyle factors. The SPSS ver. 29.0 program (IBM Corp., Armonk, NY, USA) was utilized for the remaining statistical analyses. The significance level was established at 5%. The fit of the model was assessed as the number of latent classes increased from 1 to 5. The fit of the model was evaluated using the Akaike information criterion, Bayesian information criterion, and adjusted Bayesian information criterion, with lower values indicating a better fit. Entropy results nearer to 1 reflect a more precise classification of subjects into unique categories, while results closer to 0 suggest more ambiguity. A higher entropy value demonstrates more effective classification into the 2 latent classes, indicating better class separation. The significance of latent classes was evaluated using the Lo-Mendell-Rubin test (LMR) and the bootstrap likelihood ratio test (BLRT), and models with P < 0.05 were considered significant. Demographic characteristics were analyzed according to the patterns identified by LCA. Differences in general characteristics and KHEI scores among groups were analyzed by cross-tabulation analysis and analysis of variance. The role of the mediating effect of the independent variable on the dependent variable was investigated, and 95% confidence intervals (CIs) were used to assess statistical significance. The bootstrap method (with 5,000 bootstrap resamples) was utilized for estimating risks and CIs and enhance the reliability of the analysis of mediating effects [21].
RESULTS
Latent class analysis
To identify behavior patterns, the number of models was increased to 5, and model fit was analyzed. The best model was the 3-cluster model (Table 1). LMR and BLRT values were significant in the 2- and 3-cluster models, and a model with an entropy of 0.7 or higher was considered adequate. The 3-cluster model met all these conditions and was selected [22].
Table 1
Fitting indexes of models of different potential categories
AIC, Akaike information criterion; BIC, Bayesian information criterion; aBIC, adjusted Bayesian information criterion; LMR, Lo-Mendell-Rubin likelihood ratio test; BLRT, bootstrap likelihood ratio test.
LCA of lifestyle characteristics
The 3-cluster model contained three latent classes: ‘Low Activity Class’ (Type I), ‘Low Activity and Smoke Class’ (Type II), and ‘Multiple Risk Class’ (Type III) (Table 2). We named the classes based on key lifestyle factors that are associated with higher rates of unhealthy behaviors. This naming approach effectively highlights the representative behavioral traits of each class and emphasizes the differences between them.
Table 2
Conditional response probabilities for lifestyle in cluster 3
1) Type I (Low Activity Class): The middle-aged individuals in this class had lower rates of smoking, sleep, and stress compared to the other types but were characterized by lower levels of physical activity. Compared to other lifestyle factors, physical inactivity stood out as a particularly important feature. This class comprised 60.6% of all participants, making it the largest group. This class showed a relatively low.
2) Type II (Low Activity and Smoke Class): This class was characterized by low levels of physical activity and a higher rate of unhealthy smoking habits than Type I. This accounted for 32.7% of the total number of participants.
3) Type III (Multiple Risk Class): This class was characterized by unhealthy habits across all areas and accounted for 6.7% of the total participants.
General characteristics by lifestyle pattern
The analysis of general characteristics according to lifestyle pattern is shown in (Table 3).
Table 3
General characteristics of the study population
Values are presented as mean ± SE or n (%). Post-hoc comparisons were performed using the Scheffé test.
BMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; FBG, fasting blood glucose; HbA1c, glycated hemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TG, triglycerides; MetS, metabolic syndrome; KHEI, Korean Healthy Eating Index.
a-cDifferent superscripts are significantly different at the P-value < 0.05 by the Shaffe method used in the post hoc analysis.
Among the total of 5,196 participants, 2,062 were male and 3,134 were female, and the sex differences across the 3 patterns were tested, and the proportion of men in each pattern was as follows: 432 male (13.7%) in the ‘Low Activity Class,’ 1,452 male (85.5%) in the ‘Low Activity and Smoke Class,’ and 178 male (51.3%) in the ‘Multiple Risk Class,’ showing a significant sex difference in each group (P < 0.001). The KHEI scores were 63.1 ± 11.9 for the ‘Low Activity Class,’ 60.0 ± 12.1 for the ‘Low Activity and Smoke Class,’ and 57.0 ± 12.7 for the ‘Multiple Risk Class,’ with the ‘Low Activity Class’ showing the highest KHEI score (P < 0.001).
Analysis of mediating effect
The analysis of mediating effects using ‘Low Activity Class’ as the reference is shown in Table 4 and Fig. 2. The direct effect of ‘Low Activity and Smoke Class,’ ‘Multiple Risk Class’ on MetS was significant with an effect size of 0.2411 (95% CI, 0.0708–0.4115) and 0.4318 (95% CI, 0.1919–0.6717), respectively. The mediating effect of KHEI scores on MetS was significant in ‘Low Activity and Smoke Class,’ ‘Multiple Risk Class,’ with an effect size of 0.0205 (95% CI, 0.0062–0.0363) and 0.0420 (95% CI, 0.0133–0.0726), respectively. The mediating effect of the domain ‘adequacy’ on MetS was significant in the ‘Low Activity and Smoke Class’ and the ‘Multiple Risk Class,’ with effect sizes of 0.0357 (95% CI, 0.0184–0.0563) and 0.0662 (95% CI, 0.0364–0.6491), for the respective groups. In contrast, the mediating effect of the “moderation” category on MetS was not significant in the ‘Low Activity and Smoke Class’ and the ‘Multiple Risk Class,’ with 95% CIs of −0.0104 to 0.0125 and −0.0171 to 0.0146, for the respective groups. The mediating effect of “balance of energy intake” on MetS was not significant in ‘Low Activity and Smoke Class’ (95% CI, 0.0000–0.0164) but significant in ‘Multiple Risk Class,’ with an effect of 0.0189 (95% CI, 0.0044–0.0378).
Fig. 2
The mediating effect of healthy eating index in the mediation model chart.
KHEI, Korean Healthy Eating Index; MetS, metabolic syndrome.
*P < 0.01, **P < 0.001.
Table 4
Analysis of the mediating effect of KHEI on lifestyle and metabolic syndrome
Mediation effects were analyzed using bootstrapping with 5,000 resamples to estimate SEs and 95% confidence intervals.
KHEI, Korean Healthy Eating Index; CI, confidence interval; HEI, Healthy Eating Index; MetS, metabolic syndrome.
1)Tests were adjusted for sex and age.
DISCUSSION
This study classified middle-aged individuals who participated in the eighth KNHNES (2019–2021) into three latent classes based on lifestyle patterns: ‘Low Activity Class,’ ‘Low Activity and Smoke Class,’ and ‘Multiple Risk Class.’ We found that KHEI scores had a greater mediating effect in ‘Low Activity and Smoke Class’ and ‘Multiple Risk Class’ than in ‘Low Activity Class.’
Previous studies analyzing lifestyle factors through LCA, considering factors such as smoking, drinking, physical activity, obesity, hypertension, and sleep, have shown some variation in the final patterns depending on the characteristics or components of the subjects [2324], but drinking, smoking, and physical activity appeared to be the main components determining lifestyle patterns in middle-aged adults. In this study, the probabilities of engaging in each behavior across the three latent classes were as follows: smoking (0.000, 1.000, 0.584), alcohol consumption (0.575, 0.711, 0.824), and physical inactivity (0.970, 0.948, 0.916). These results indicate that smoking was the behavior that most distinctly differentiated lifestyle patterns within middle-aged individuals in Korea. In our study, smoking was distinctly classified as a behavior for healthy lifestyle factors. On the other hand, physical activity was generally lacking, indicating that even among groups practicing relatively healthy lifestyle factors, there is a need for strategies that emphasize the importance of physical activity.
The prevalence of MetS was higher in individuals with unhealthy lifestyle patterns. Smoking, alcohol consumption, sleep, and stress [25262728] are closely related to MetS. Furthermore, these factors combined further increase the risk of MetS [29], consistent with our findings, indicating that the likelihood of MetS was greater in the group with multiple risk factors. Nevertheless, the reason for developing tailored strategies for improving MetS based on each pattern is that, although the prevalence of MetS does not differ significantly between the ‘Low Activity and Smoke Class’ and the ‘Multiple Risk Class,’ the distribution of these groups among the total participants shows that the ‘Low Activity and Smoke Class’ accounts for 32.7% and the ‘Multiple Risk Class’ for 6.7%, making individualized strategies more efficient in terms of treatment, care, and participant acceptance.
This study analyzed the mediating effect of the KHEI in the relationship between lifestyle patterns and MetS. The results showed that the relationship between dietary endpoints and MetS was more strongly mediated in the ‘Low Activity and Smoke Class,’ and “Multiple Risk Class” than in the ‘Low Activity Class,’ which may be due to the interaction between lifestyle factors. In other words, ‘Low Activity Class,’ which has a higher proportion of healthy lifestyle factors, tends to have relatively appropriate eating habits, whereas the two other groups, with higher levels of unhealthy lifestyle factors, are more likely to engage in inappropriate eating habits, leading to a greater mediating effect of diet quality. Although the mediating effect of KHEI is not strong, diet is a major contributor to life in general, so the greater the tendency to certain lifestyle patterns, such as low activity and smoking class in the study, the greater the influence of diet on MetS. Diet is one of the modifiable factors that significantly influence the development of MetS, and improving dietary habits can help prevent MetS. Since diet affects lifestyle factors such as physical activity levels and smoking, it can serve as an important tool to help mitigate the risks associated with MetS. Therefore, healthy eating habits not only contribute to health management but also mitigate the complex lifestyle factors that increase the risk of MetS and aid in its prevention.
The KHEI domains “adequacy” and “balance of energy intake,” but not “moderation,” had significant mediating effects on MetS. This result may be due to the characteristics of the items in the “moderation” domain (intake of saturated fats, sodium chloride, sugar, and beverages) and the specific characteristics of middle-aged individuals. These items had relatively higher scores than did other items (Supplementary Table 1), likely because middle-aged individuals manage their health more proactively than do younger individuals and are more aware of the health risks associated with these behaviors, increasing the scores in this category and reducing its impact on MetS. Consequently, the impact of moderation on MetS may have appeared relatively weaker. Therefore, further analyses of the mediating effects of subcomponents of the KHEI on MetS are necessary to develop dietary strategies for the prevention and management of MetS.
This study has several limitations. Defining lifestyle factors as dichotomous variables may result in missing complex interactions or intermediate states between factors [30], and this simplified variable definition could influence the study outcomes. Additionally, the use of a 24-h recall to evaluate participants’ dietary intake may not accurately reflect daily nutritional intake or account for potential reporting biases. The analysis also did not incorporate a weighted composite sample, which may limit the representativeness of the findings for the broader population. Furthermore, changes in dietary and lifestyle factors over time could affect the validity and applicability of the findings to current or future populations. Despite these limitations, this study effectively classified the cohort into latent classes based on lifestyle patterns and examined the relationship between KHEI scores and the effect of these patterns on MetS. The findings suggest that improving lifestyle factors, including dietary habits, can help middle-aged adults in Korea manage their health more effectively.
XML Download