The mediating effect of quality of life on alcohol consumption and the 10-year type 2 diabetes mellitus risk score in adult Korean men

Background: The rate of high-risk drinking, a major cause of type 2 diabetes mellitus (T2DM), is increasing among Korean men. This study descriptive survey whether quality of life (QoL) mediates the relationship between drinking and 10-year T2DM risk scores in adult Korean men. Methods: We further analyzed data from the 7th Korea National Health and Nutrition Examination Survey 7th (KNHANES VII) from 2016 to 2018, selecting 4134 men aged 40–69 years who were not diagnosed with T2DM. The data were analyzed using complex sample regression; after analyzing the mediation. Results: In step 1 of the regression, the drinking rate significantly predicted QoL (β = 0.08, p < 0.001), and in step 2, the drinking rate significantly predicted the 10-year T2DM risk score (β = 0.08, p < 0.001). In step 3, the drinking rate (β = 0.08, p < 0.001) and QoL (β = –0.09, p < 0.001) were found to significantly predict the 10-year T2DM risk score, confirming that QoL partially mediates the 10-year T2DM risk score (z = –3.62, p < 0.001). QoL was found to mediate the relationship between the drinking rate and 10-year T2DM risk score. Conclusions: Sex-specific risk factors must be considered to reduce the incidence of T2DM, and measures to promote healthy drinking habits need to be implemented to reduce the incidence of T2DM in men. Furthermore, these findings highlight the need to develop and actively implement various practical intervention strategies to enhance QoL.


Introduction
In 2019, there were approximately 463 million adults aged 20-79 years worldwide with type 2 diabetes mellitus (T2DM), and this number is expected to rise to 578 million in 2030, and 700 million in 2045 [1]. In 2018, the prevalence of T2DM in South Korea increased to one in seven adults aged 30 years or older [2]. Moreover, T2DM occurs more often in men (15.9%) than women (11.8%) (Korean Diabetes Association, 2020), highlighting the urgent need for interventions to target the specific causes of T2DM in men [2].
Risk factors associated with T2DM include obesity, reduced physical activity, increased intake of sugar-rich foods, smoking, heavy alcohol consumption, social and economic stress, and sleep disorders [3]. The blood testosterone concentration begins to drop by 1%-2% every year from the age of 40 in men, which increases the risk for T2DM [4,5]. Accordingly, T2DM prevalence increases with age in Korean men, as it reportedly afflicts 13.0%, 17.9%, and 25.7% of men in their 40s, 50s, and 60s respectively [2].
Alcohol consumption is a major risk factor for T2DM [6]. Korean and Asian men have a higher risk for T2DM than their western counterparts due to the prevalence of variants of the aldehyde dehydrogenase 2 (ALDH2) gene. In women, this gene does not affect the incidence of T2DM [7]. ALDH2 inhibits alcohol metabolism, and when ex-pressed in individuals that frequently consume alcohol, it can increase the risk of T2DM [7]; hence multilateral studies are warranted to determine the factors associated with drinking rates in men.
The monthly drinking rate (which is calculated by dividing the number of people who have had at least one drink per month in the past year by the total number of people) of Koreans in 2017 was 74% for men and 50% for women. Furthermore, 20.8% of Korean men engaged in high-risk drinking (>7 drinks at any given time), exceeding the global average of 18.2% [8][9][10], which underlines the need to promote healthy drinking habits. High-risk drinking is correlated with job stress in Korean men [11,12]. Additionally, social and cultural pressures also need to be taken into account, as Korean men may be encouraged to drink alcohol by friends or colleagues and their inability to refuse alcohol in social settings may affect their drinking habits [13], which is different to the findings for other countries [14].
The social pressure placed on Korean men to consume alcohol can increase the risk of T2DM [13,15] and reduce their quality of life (QoL) [16]. Some studies have reported that high blood sugar levels and the complications and duration of T2DM can also reduce QoL; hence, early diagnosis and treatment could help to improve QoL [17,18]. Data showing the direct impact of QoL on the onset of T2DM are lacking. Furthermore, QoL is negatively correlated with al-cohol consumption [19,20], i.e., lowering alcohol consumption has been shown to improve QoL [4]. Hence, it is important to examine the relationships among these factors to lower the incidence of T2DM attributed to high levels of alcohol consumption among Korean men.
This study examines the relationship between the drinking rate and QoL in Korean men and the impact of these factors on the estimated T2DM risk and investigates the mediating effect of QoL on the drinking behavior and the estimated 10-year risk of developing T2DM in adult Korean men. The results of this study will provide foundational data to develop intervention strategies that promote optimal drinking habits, establish a healthy drinking culture, and enhance QoL, thereby reducing the incidence of T2DM in men, and aid in the development of sex-specific lifestyle interventions, T2DM prevention programs, and help with diagnosis and treatment.

Experimental design
This study used a descriptive survey and a secondary data analysis of raw data from the seventh Korea National Health and Nutrition Examination Survey, to examine the mediating effects of QoL on the relationship between drinking rate and the 10-year risk of T2DM in adult Korean men.

Participants
The KNHANES VII (2016-2018) raw data were obtained from the Korea Disease Control and Prevention Agency (KDCA), an affiliated organization of the Ministry of Health and Welfare. The survey used stratified two-stage cluster sampling, in which sample households were selected from 13,248 households in 576 enumeration districts nationwide in two stages. The sampling framework was stratified based on city/province, neighborhood/town/township (dong/eup/myeon), and house (single house, apartment). Factors such as living space ratio and education of the head of the household were criteria for implicit stratification.
Of the 24,269 participants in the survey, 10,586 adults aged 40-69 years were first extracted. Next, 972 participants who had been diagnosed with T2DM or were undergoing pharmacological therapy were excluded, resulting in a total of 9614 adults. Finally, 5480 women were excluded, resulting in 4134 men who were included in this study (Fig. 1).

Study variables 2.3.1 Sociodemographic characteristics
Sociodemographic and health-related characteristics such as age, family, living situation, and perceived health were surveyed. For central obesity, the waist-to-height ratio (WHtR) was calculated using waist circumference and height, and the participants were classified based on the Asian cutoff for cardiovascular risk (0.53) [21]. Low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides, glycated or glycosylated hemoglobin (HbA1c), blood sugar, and high-sensitivity C-reactive protein (hs-CRP) levels were obtained from laboratory results.

Monthly alcohol consumption
The monthly drinking rate was defined as the percentage of participants who have had at least one drink per month in the past year. The value is calculated from the raw data by dividing the number of participants who stated that they had had at least one drink per month in the past year by the total number of participants. High-risk drinking was defined as drinking twice or more per week, with an average of seven or more drinks per sitting. In the present study, drinking frequency in the past year was classified as abstinent, low (<1 drink/month to 2-4 drinks/month), and high (≥2 drinks/week) [22,23].
In this study, the EQ-5D index for each of the five subscales was computed using the following conversion formula that applies weighting for the Korean population: The EQ-5D Index ranges from -1 (worst health) to 1 (perfect health), where a higher index indicates a higher QoL [24].
The risk score for developing T2DM after 10 years was calculated using an equation to predict the risk of developing diabetes in Korea that was developed by Ha et al. [23]. The equation predicted the diabetes risk by applying the Cox proportional risk regression model using the National Health Insurance Corporation-National Health Checkup Cohort (NHIS-HEALS) [23,25]. The analysis included the following variables that are known to influence diabetes risk: age, family history of diabetes, alcohol intake, smoking status, physical activity, antihypertensive therapy, statin therapy, body mass index (BMI), systolic blood pressure (SBP), total cholesterol, and fasting blood glucose. BMI was classified according to Asianspecific criteria based on the World Health Organization Western Pacific Region [26] as follows: underweight (<18.5 kg/m 2 ), normal weight (18.5-22.9 kg/m 2 ), overweight (23.0-24.9 kg/m 2 ), obese class I (25.0-29.9 kg/m 2 ), and obese class II (≥30.0 kg/m 2 ).
In this study, according to Ha et al. [23], the variables used were first stratified and classified. Then coefficient values according to the stratified items were substituted. After the coefficient value for each variable was calculated, the total was calculated. The total number of variables was divided by the total score, to yield a risk score for developing T2DM after 10 years. The final value was obtained by making a calculation formula so that the subjects' risk of diabetes could be calculated statistically.

Data collection
The KNHANES VII was conducted from January 1, 2016, to December 31, 2018, and the health surveys and examinations were performed in a mobile examination center. The health survey comprised a self-report questionnaire for health behaviors, including smoking status and drinking status, and an interview regarding medical history and medications. The examination consisted of anthropometry, observation, and sample analysis. Participants fasted for >8 hours before having their blood sampled. The professional surveyors were nurses, nutritionists, and public health majors, who were trained for 2-4 weeks before conducting the surveys. The surveyors' survey performance is managed through mandatory refresher training (7 times a year) and field quality control.

Statistical analysis
The collected data were analyzed using the SPSS version 23.0 (IBM SPSS Statistics, Chicago, IL, USA). The KNHANES used stratified two-stage cluster sampling, and following the analysis guidelines for the raw data, we used a complex sample analysis. As complex sample data, each variable had a missing value, so the subject was different for each variable. Missing values were replaced with significant values for the analysis of each variable. The general characteristics of adult Korean men and factors associ-ated with the 10-year T2DM risk score were analyzed using the t-test or χ 2 test. Furthermore, the mediating effect of QoL on the relationship between drinking rate and 10-year T2DM risk in adult men was analyzed using the three-step regression by Baron and Kenny [27]. The mediating effect of QoL and its significance were tested using the Sobel test (http://www.quantpsy.org/sobel/sobel.htm) with the calculated regression coefficient and standard error. Statistical significance was set at p < 0.05.

Patient characteristics
A total of 4134 participants were included in our analysis, with a mean age of 52.39 ± 0.15 years. Of the participants, 12.6% perceived themselves as having poor health. The monthly drinking rate was 0.74 ± 0.01, and the mean QoL score was 0.970. A total of 33.1% of the participants had a WHtR of 0.53 or higher. The laboratory results were as follows: 69.6% had LDL ≥100 mg/dL, 64.0% had HDL <50 mg/dL, 44.6% had triglycerides ≥150 mg/dL, 34.4% had an HbA1c of 5.7-6.4%, and the mean hs-CRP level was 1.23 ± 0.03 (Table 1).

Ten-year T2DM risk score of adult Korean men
Of all the participants, 23.4% had a family history of T2DM, and alcohol consumption rates were abstinent (15.3%), low (44.4%) and high (40.3%). A total of 37.8% of the participants were current smokers, and 24.7% claimed to engage in moderate-intensity physical activity. Of the participants, 20.1% were currently undergoing antihypertensive therapy, while 9.3% were currently undergoing statin therapy. A total of 43.4% of the participants had a BMI of ≥25.0 kg/m 2 . The mean 10-year T2DM risk score was 19.81% (Table 2).

QoL mediates the relationship between drinking rate and the 10-year T2DM risk
The three-step regression proposed by Baron and Kenny [27] was performed to analyze the mediating effect of QoL on the relationship between drinking rate and the 10-year T2DM risk. In step 1, the drinking rate (independent variable) had a significant effect on QoL (mediator) (β = 0.08, p < 0.001). In step 2, the drinking rate had a significant effect on the 10-year T2DM risk (dependent variable) (β = 0.08, p < 0.001). In step 3, the drinking rate (β = 0.08, p < 0.001) and QoL (β = -0.09, p < 0.001) significantly influenced the 10-year T2DM risk. In this step, the effect of the drinking rate on the 10-year T2DM risk remained statistically significant even after entering the mediator (QoL), confirming that QoL has partial mediation (Fig. 2). The Sobel test was performed to test the statistical significance of the partial mediation of QoL on the relationship between the drinking rate and 10-year T2DM risk, and the results confirmed that the partial mediating effect of QoL was significant (z = -3.62, p < 0.001) ( Table 3).

Discussion
Among adult Korean men, high-risk drinking rates and the incidence of T2DM are both increasing, therefore, confirming that the drinking rate and QoL of adult men could predict the 10-year T2DM risk. This study suggests that strategies to lower drinking rates and improve QoL are required to lower the prevalence of T2DM within 10 years in men.
Our results confirmed that both the drinking rate and QoL were predictors of the 10-year T2DM risk in men, as when QoL decreased the 10-year T2DM risk scores and drinking rate increased. Furthermore, QoL was found to partially mediate the relationship between the drinking rate and 10-year T2DM risk. These results are consistent with the results of a Danish cohort study in which the incidence of T2DM was relatively higher among those who drank 14 drinks per week when compared to those who consumed no alcohol [28]. This study supported the results of an epidemiological survey that involved a 10-year follow-up cohort study of 39,259 Chinese adults aged 18 to 79 years. In this study, the incidence of diabetes increased in line with alcohol consumption, and those who did not drink alcohol for 10 years showed no association with the occurrence of 2TDM [29]. The 2010-2011 KNHANES data for adults aged 40 years or older showed that excessive alcohol consumption is associated with a lower QoL than moderate and no alcohol consumption, which supports our findings [16]. In contrast to foreign drinking cultures, Korean drinking culture involves heavy drinking for reasons such as fostering friendship, resolving interpersonal problems, and relieving stress [30]. A study on working men in Korea reported that men frequently experience extended work hours, stress at work, and frequent group dinners at work. Furthermore, Korean workers find it difficult to manage their stress and engaging in self-care; as 52.9% of workers drink heavily, 86.2% do not exercise, and 31.3% smoke [31], suggesting that the unhealthy drinking culture has society-wide effects. In addition, Koreans are more lenient towards the consumption of alcohol and high-calorie foods than their foreign counterparts [32]. Such excessive alcohol consumption is a risk factor for T2DM and may induce or exacerbate T2DM [33,34]. According to the 2019 Liquor Market Trend Report in Korea, Korean adults prefer mixed beers and soju drinks (69.8%). Soju contains 18-24% alcohol and therefore must be consumed with precaution [35].
Previous studies have reported that drinking and the risk for T2DM and its complications demonstrated a Ushaped relationship, where moderate drinking lowers the risk for T2DM and facilitates better metabolic regulation [6,19]. However, such benefits of drinking may differ based on age, sex, BMI, ethnicity, and type of alcohol [6].
One study has reported that one-or two-level reductions in the WHO risk drinking substantially improved physical health and QoL in 1142 individuals with alcohol dependence (68.8% men, 32.2% women) [4], illustrating the benefits of moderate alcohol consumption. Our results support previous findings that excessive alcohol consumption diminishes QoL [16,19,20]. In a study of 1594 participants in the United States, it was shown that moderate drinking increased the QoL of the adults [36], showing contradictory results to the results of this study. The results of this study did not correct for the occurrence of diseases, so it is difficult to generalize. In Helsinki's cohort study of 5301 women and 1230 men, excessive alcohol consumption was found to affect mental functioning and QoL [37]. Furthermore, 4592 participants in a 2.3-year follow-up study in Hong Kong, showed that drinking had an effect on mental well-being, indicating that caution is required when making recommendations regarding moderate alcohol intake [38]. Therefore, it is necessary to study various aspects of the relationship between the quantity and frequency of drinking intake and the quality of life. In this study, the 10-year T2DM risk increased with a decrease in QoL, which supports earlier findings that patients diagnosed with T2DM have a five-fold reduced QoL score when compared to their healthy counterparts [39] and that QoL significantly differs between the T2DM and normal groups [40], which is in line with our results. However, the participants in this study were limited to those without T2DM, while previous studies have mostly examined QoL in patients with T2DM. Cohort studies should thus be conducted to better investigate the effects of QoL on drinking rate and the incidence of T2DM. It has been reported that sensible drinking, i.e., fewer than two drinks per day, in men [41] is associated with a higher QoL and lower incidence of T2DM [42], suggesting that measures to promote an awareness of healthy alcohol drinking habits are warranted to improve QoL and decrease the incidence of T2DM.
In this study, 34.1% of the participants had a 10-year T2DM risk of 20% or higher. The high-risk drinking rate was 40.3%, highlighting the urgency of planned intervention strategies that restrict high-risk drinking and establish healthy drinking habits. However, QoL was identified as a predictor of drinking rate; hence, measures to reduce the drinking rate should focus on improving the QoL, to ultimately reduce the incidence of T2DM in men.
This study has several limitations. First, the data for lifestyle and medications were collected using a self-report questionnaire. As the responses were dependent on the participants' recall, they may differ from the actual values. Subsequently, cohort studies covering a longer study period for the incidence of T2DM in men are needed. Second, the KNHANES is a cross-sectional survey that focuses on understanding the relationships among the variables. This hinders the establishment of causality between the 10-year T2DM risk and other independent variables. Third, the discussion of the results of this study is limited due to the lack of epidemiologic studies on QoL as a factor influencing the predictive score of 2TDM occurrence after 10 years in men. Therefore, it is suggested that an expanded study should be conducted to confirm the effect on QoL and the occurrence of diabetes in subjects without diabetes mellitus. Thus, we suggest that longitudinal studies that implement interventions for drinking habits and examine the prevalence of T2DM should be conducted.
Despite these limitations, this study makes notable contributions as we have confirmed that QoL is a partial mediator in the relationship between drinking and 10-year T2DM risk in Korean men, using nationally representative data for the Korean people's health and nutritional status. The results of this study highlight the need to implement sex-specific approaches to lower the incidence of T2DM. Our findings provide new insights for the design of intervention programs to lower the prevalence of T2DM by exploring measures to enhance QoL in men.
Furthermore, to lower the incidence of diabetes in middle-aged men, policy control on alcohol marketing to lower the drinking rate, which affects QoL and sociocultural changes, is necessary to specifically lower the frequency and intensity of alcohol consumption.
These changes must first change the public perception of harmful drinking through active publicity via the TV, internet, and social networks to improve good drinking habits. In addition, the socio-cultural habits that promote high levels of drinking due to stress at work also need to be changed. To this end, measures to reduce job stress through efficient work management should be sought. In addition, changes  in workplace culture and improvement of company managers' awareness and policies should be established.
Medical staff believe that multidisciplinary cooperation is required to actively improve behavioral therapy and lifestyle to reduce risk factors for middle-aged men who have increased ten-year risk scores for the incidence of diabetes.

Conclusions
We investigated the effects of QoL on the relationship between drinking rate and the 10-year T2DM risk in Korean men aged 40-64 years. The drinking rate and QoL were identified as predictors of the 10-year T2DM risk. QoL decreased as the drinking rate increased and had a partial mediating effect on the relationship between drinking and the 10-year T2DM risk. Furthermore, the 10-year T2DM risk increased with the decreasing QoL and increasing drinking rate. Therefore, to lower the incidence of T2DM in Korean men, healthy alcohol consumption habits must be promoted, and continuous educational interventions are required. Moreover, lifestyle modification approaches, which influence the drinking rate, should be tailored to each sex, and active education and promotion, and practical intervention strategies are urgently needed. The findings of this study will contribute to the development of individualized T2DM prevention programs and increase the awareness of the risk factors for diabetes in men.

Abbreviations
BMI, Body mass index; DM, Diabetes mellitus; HDL, High-density lipoprotein; IRB, Institutional Review Board; KDCA, Korea Disease Control and Prevention Agency; LDL, Low-density lipoprotein; QoL, Quality of life; SB, Systolic blood pressure.

Author contributions
MRB, KAK, and JLS conceived and designed the experiments. MRB analyzed the data. MRB, KAK, and JLS wrote and approved the final manuscript for submission.

Ethics approval and consent to participate
The KNHANES VII is a survey and examination of public health behaviors, prevalence of chronic diseases, and food and nutritional intake stipulated by Article 16 of the National Health Promotion Act, and it contains government-designated statistics grounded in Article 17 of the Statistics Act. The KDCA provides only de-identified data in compliance with the Personal Information Protection Act and Statistics Act, and the survey was conducted with approval from the Korea Centers for Disease Control and Prevention Research Ethics Review Committee (2018-01-03-P-A). We downloaded the raw data after reading the "Rules on the raw data disclosure by the KCDC" from the KNHANES website (https://knhanes.cdc.go.kr/) and used the data after reading the "User guidelines for the KN-HANES" and "Analysis guidelines for the KNHANES raw data". In addition, we obtained approval for secondary data analysis from the Institutional Review Board (IRB) of our affiliated university (IRB: 1041455-202106-HR-006-01).