Prevalence of Metabolic Syndrome and Its Components among Japanese Workers by Clustered Business Category

The present study was a cross-sectional study conducted to reveal the prevalence of metabolic syndrome and its components and describe the features of such prevalence among Japanese workers by clustered business category using big data. The data of approximately 120,000 workers were obtained from a national representative insurance organization, and the study analyzed the health checkup and questionnaire results according to the field of business of each subject. Abnormalities found during the checkups such as excessive waist circumference, hypertension or glucose intolerance, and metabolic syndrome, were recorded. All subjects were classified by business field into 18 categories based on The North American Industry Classification System. Based on the criteria of the Japanese Committee for the Diagnostic Criteria of Metabolic Syndrome, the standardized prevalence ratio (SPR) of metabolic syndrome and its components by business category was calculated, and the 95% confidence interval of the SPR was computed. Hierarchical cluster analysis was then performed based on the SPR of metabolic syndrome components, and the 18 business categories were classified into three clusters for both males and females. The following business categories were at significantly high risk of metabolic syndrome: among males, Construction, Transportation, Professional Services, and Cooperative Association; and among females, Health Care and Cooperative Association. The results of the cluster analysis indicated one cluster for each gender with a higher prevalence of metabolic syndrome components; among males, a cluster consisting of Manufacturing, Transportation, Finance, and Cooperative Association, and among females, a cluster consisting of Mining, Transportation, Finance, Accommodation, and Cooperative Association. These findings reveal that, when providing health guidance and support regarding metabolic syndrome, consideration must be given to its components and the variety of its prevalence rates by business category and gender.


Introduction
Metabolic syndrome is a group of risk factors for cardiovascular disease and mortality that include central obesity, hypertension, glucose intolerance and dyslipidemia [1][2][3]. Furthermore, metabolic syndrome has, on a global scale, become one of the key challenges to the public health sector. Early studies reported that its prevalence was 20-30% in the U.S. [4,5], and recent Japanese studies have revealed a prevalence of 8-25% in male and 2-22% in female [6][7][8].
The prevalence of metabolic syndrome has recently been suggested to vary greatly depending on the subject's business category; high prevalence of metabolic syndrome has been reported among the retired, unemployed [9], bus drivers [10], university employees [11], and workers in the agricultural industry [12] oil industry [13], and health care sector [14].
However, these studies did not compare the prevalence between, nor did they indicate the common features of prevalence in, different business categories. This was due to a lack of categories covered. Moreover, the classifications of business categories used in these past studies were inconsistent. It is required to conduct a study using standard classification of business category and analyze the prevalence of metabolic syndrome between all categories.
The North American Industry Classification System (NAICS) is a widely used standard classification system of business categories [15], and has been applied in various occupational health studies [16,17]. Therefore, NAICS was considered to be suitable for the present study. We hypothesized that when business categories are clustered, the features of metabolic syndrome and its components can be elucidated by business category. Clustering may also contribute to the identification of the common and/or distinctive features of metabolic syndrome, which can essentially aid in understanding the background of this disease and its components.
Here, we reveal for the first time the prevalence of metabolic syndrome and its components, and describe the features of such prevalence among Japanese workers by clustered business category using big data

Study Sample
In 2012, health checkups of 161,362 workers were conducted in Fukushima Prefecture, Japan, by the Japan Health Insurance Association (JHIA), a national representative organization of insurance for laborers. Individuals who were aged 34 years or younger or aged 76 years or older, who had been under the insurance system for less than one year, or whose information regarding diagnostic criteria was unavailable, were excluded from the study. The JHIA health checkup included a questionnaire asking the subject's business category, and, together with the checkup data, was recorded in the JHIA database. Of the 161,362 subjects, those who underwent measurement of waist circumference, blood pressure, blood glucose, lipid, and metabolic syndrome were 120,100 (74.4%), 120,114 (74.4%), 120,090 (74.4%), 120,088 (74.4%), and 120,097 (74.4%), respectively.

Measurements
Diagnostic criteria. According to the Japanese Committee for the Diagnostic Criteria of Metabolic Syndrome in 2005, metabolic syndrome is defined as an excessive waist circumference (85 cm in men and 90 cm in women) as well as the presence of one or more of the following symptoms; hypertension, glucose intolerance, and dyslipidemia [18,19]. Hypertension is defined by systolic blood pressure 130 mmHg, or diastolic blood pressure 85 mmHg, or the use of antihypertensive drugs. Glucose intolerance is defined by fasting glucose 110 mg/dL, or the use of drugs for diabetes. Dyslipidemia is defined by neutral fat 150 mg/dL, or HDL cholesterol <40 mg/dL, or the use of antihyperlipidemic drugs. Business category. The business categories of the subjects were extracted from the JHIA database. The obtained information was classified based on the Japan Standard Industrial Classification, which is compatible with NAICS. Eighteen business categories were used in the present study: (1) Agriculture, (2) Mining, (3) Utilities, (4) Construction, (5) Manufacturing, (6) Wholesale Trade, (7) Transportation, (8) Information, (9) Finance, (10) Real Estate, (11) Professional Services, (12) Educational Services, (13) Health Care, (14) Arts, (15) Accommodation, (16) Cooperative Association, (17) Other Services, and (18) Public Administration. The details of the categories are shown in Table 1.

Statistical analyses
Age adjustment and calculation of standardized prevalence ratio with 95% CI. The subjects were classified into four age groups, and age adjustment was conducted by an indirect method based on the total subject population. The SPR was calculated as the ratio of observed prevalence to the expected prevalence for each business category. The expected prevalence for each business category was obtained by multiplying the number of people who fell into three specific categories (age group, business category, and abnormality) by the percentage of people who fell into the corresponding age group and abnormality categories of the total subject population. The 95% confidence interval (95% CI) of SPR was derived assuming a Poisson distribution for the observed numbers. Hierarchical cluster analysis. Hierarchical cluster analysis based on agglomerative statistics using Ward's method was conducted for the SPRs of metabolic syndrome components. The data were classified into three clusters of business categories for both males and females. The mean SPR of each metabolic syndrome component for each cluster was calculated, and the data were analyzed by SPSS statistics version 17.0.

Ethics
This study was approved by the Ethics Committee of Fukushima Medical University (Application No. 1703).

Results
The characteristics of the subjects are shown in Tables 2 and 3. Blood pressure abnormalities were most common in both males and females, at 53.9% and 34.9%, respectively. Approximately one-fifth of the male subjects had metabolic syndrome (22.2%); however, this was observed in very few females (4.4%). Of the business categories, (7) Transportation, (4) Construction, and (2) Mining showed the highest prevalences of metabolic syndrome at 25.7%, 21.0%, 20.5%, respectively, whereas (13) Health care, (18) Public Administration, (14) Arts, and (15) Accommodation showed the lowest prevalences at 8.7%, 11.4%, 12.1%, and 12.1%, respectively. The SPR of abnormalities by business category for the male and female subjects are shown in Tables 4 and 5. Among the male subjects, significantly higher prevalences of metabolic syndrome were seen in the following four business categories: (4)    The results of the cluster analysis of males are shown in Among the male subjects, Cluster MA had an SPR lower than 1 for all components while Cluster MB had an SPR greater than 1 for all components. The SPRs of excessive waist circumference and glucose intolerance in Cluster MB were particularly higher than those in the other clusters (1.14 and 1.12, respectively). Cluster MC had an SPR lower than 1 for hypertension and glucose intolerance, and an SPR greater than 1 for excessive waist circumference and dyslipidemia. The SPR of dyslipidemia in Cluster MC was higher than that in other clusters (1.17).
The results of the cluster analysis of the female subjects are shown in Among the female subjects, Cluster FA had an SPR lower than 1 for all components. Cluster FB had an SPR greater than 1 for excessive waist circumference, glucose intolerance and dyslipidemia (1.09, 1.02, 1.19, respectively), but not for hypertension. Cluster FC had an SPR greater than 1 for excessive waist circumference only (1.03); however, all other components had an SPR lower than 1, with glucose intolerance being particularly low (0.57).

Discussion
In the current study, we investigated the prevalence of metabolic syndrome and its components by clustered business category, using big data. We found that metabolic syndrome was significantly prevalent among the male workers in the (4) Construction, (7) Transportation, (11) Professional Services, and (16) Cooperative Association industries, and among the female workers in the (13) Health Care and (16) Cooperative Association industries. Furthermore, the results of the cluster analysis indicated a cluster with a higher prevalence of metabolic syndrome components; for the male subjects, a cluster consisting of (5) Manufacturing, (7) Transportation, (9) Finance, (16) Cooperative Association, and for the female subjects, a cluster consisting of (2) Mining, (7) Transportation, (9) Finance, (15) Accommodation, and (16) Cooperative Association. We believe that the present study can provide an essential contribution to the understanding of the background of metabolic syndrome and its components.
Our study has also revealed that workers in (7) Transportation have a higher prevalence of glucose intolerance, whereas past studies indicated that such workers were at high risk of obesity, hypertension, dyslipidemia, and metabolic syndrome [10,20].
Among the female subjects, those in the (13) Health Care and (16) Cooperative Association categories had significantly higher SPR of metabolic syndrome. In the present study, we found that dyslipidemia was prevalent among (13) Health Care workers, whereas past studies have reported that such workers have high prevalence of obesity, diabetes, and metabolic syndrome [14,16,21]. The female (13) Health Care workers had significantly higher SPRs of all abnormalities except for hypertension, suggesting that they may be unhealthier than their male counterparts.
We also used hierarchical cluster analysis to group the business categories into three clusters according to the SPR. Among the male subjects, Cluster MA had a mean SPR of less than 1 for all components of metabolic syndrome. Thus, it is assumed that this cluster is a relatively healthier group than the other male clusters of the current study. In contrast, Cluster MB had a mean SPR of higher than 1 for all components of metabolic syndrome. This cluster is considered to be an aggregation of unhealthier business categories.
Among the female subjects, Cluster FA was considered to be a relatively healthier group as the mean SPR was less than 1 for all components. In contrast, Cluster FB had a higher mean SPR for all components except for hypertension. This cluster is considered to be the unhealthiest of the female clusters, and dyslipidemia is particularly prevalent here.
A limitation of this study was that the subjects' precise occupations were unclear. As past studies have suggested, occupational factors affect the prevalence of metabolic syndrome and its components, which, for example, increases in workers whose work is sedentary [22]. Future studies should be designed to include such occupational factors as subcategories to business category.
In conclusion, we revealed the prevalence of metabolic syndrome and its components among Japanese workers by business categories and described the features of the prevalence. The business categories of (4) Construction, (7) Transportation, (11) Professional Services, and (16) Cooperative Association among the male subjects, as well as (13) Health Care and (16) Cooperative Association among the female subjects, were at a significantly high risk of metabolic syndrome. Furthermore, we were able to summarize the business categories into three clusters, based on the prevalence of the components of metabolic syndrome in both males and females. The Cluster MB in male and FB in female were identified as having a higher prevalence of metabolic syndrome components.
The findings of the present study show that the prevalence of metabolic syndrome or its components varies according to business category and gender, and must be taken into account when providing health guidance and support to patients with this disease.