Ethics Statement

This study was approved by the institutional review boards from 17 participating institutions, which included First Affiliated Hospital of Fourth Military Medical University, China-Japan Friendship Hospital, Chinese People’s Liberation Army General Hospital, Third Affiliated Hospital of Sun Yat-sen University, People’s Hospital of Peking University, First Affiliated Hospital of Chinese Medical University, People’s Hospital of Shanxi Province, West China Hospital of Sichuan University, Affiliated Sixth People’s Hospital of Shanghai Jiao Tong University, Affiliated Drum Tower Hospital of Medical School of Nanjing University, Xinjiang Uygur Autonomous Region’s Hospital, Fujian Provincial Hospital, Qilu Hospital of Shandong University, Peking University First Hospital, Henan Province People’s Hospital, Second Affiliated Hospital of Harbin Medical University, and Xiangya Second Hospital. Written informed consent was gained from each participant prior to data collection. The 17 institutional review boards’ approvals covered every participant in the study.

2007-08 China National Diabetes and Metabolic Disorders Study

Data were obtained from the 2007-08 China National Diabetes and Metabolic Disorders Study, which was a population-based nationwide and multi-stage stratified sampling investigation performed between July 2007 and June 2008. Detailed information was previously described [27-31]. Briefly, adults aged >20 years with at least 2 years of residency were selected for the study. Signed informed consents were obtained from the participants before the study. After fasting for at least 10 hours at night, participants with no history of diabetes were given a standard 75-g glucose solution, whereas for safety reasons, participants with a self-reported history of diabetes were given a steamed bun that contained approximately 80g of complex carbohydrates [27]. In addition, all participants were subjected to other laboratory tests, such as fasting triglycerides, total cholesterol (TC), and high-density lipoprotein cholesterol (HDL-c) assessments. Professionally trained doctors or nurses measured anthropometric data including height, weight, waist circumference, and blood pressure [27]. A questionnaire was designed to collect demographic data and lifestyle risk factors on all participants. Educational level was categorized as college or above, secondary school, and elementary school or below. Cigarette smoking was defined as having smoked at least 100 cigarettes in one’s lifetime. Alcohol drinking was defined as consuming alcohol at least once per week for a year or more. Physical activity was defined as participating in moderate or vigorous activity for 30 minutes or more per day for at least 3 days a week. Questions in the questionnaire were used to investigate the diagnosis history of abnormal metabolic components, which included “What is your maximum weight? Please specify your age when you reached your maximum weight.” and “Have you ever had diabetes, hypertriglyceridemia, or hypercholesterolemia? If yes, please indicate the date of diagnosis”.

Definitions

The definition of MS referred to the International Diabetes Federation standard [3], with AO defined as waist circumference ≥ 90 cm (males) or ≥ 80 cm (females) plus any two of the following four abnormal metabolic components: 1) fasting triglycerides ≥ 1.69 mmol/l or the use of lipid medications; 2) systolic blood pressure ≥ 130 mmHg, diastolic blood pressure ≥ 85 mmHg, or the use of antihypertensive medications; 3) fasting plasma glucose ≥ 5.6 mmol/l or the use of diabetes medications; and 4) HDL-c < 1.04 mmol/l (male) or < 1.29 mmol/l (female).

Study Population

Among 46,399 participants in the survey, 27,353 participants without AO were excluded, while 19,046 participants with AO were selected as Sample 1 to study the prevalence of MS and its components. To investigate the process from the first appearance of an abnormal metabolic component to the development of MS, we further chose participants who met the following conditions: 1) AO; 2) a diagnosis history of at least one abnormal metabolic component, so that the first abnormal metabolic component had occurred. We used total cholesterol instead of HDL-c as one of the components, because diagnosis history of HDL-c was not available. People with TC ≤ 6.08 mmol/l or receiving cholesterol-lowering treatment were selected [32]. The feasibility was analyzed in detail, as described in the Discussion section; 3) To ensure that subjects had been in a state of obesity, previous maximum weight was required to be greater than the current weight; 4) To ensure that obesity appeared before the occurrence of the first abnormal metabolic component, the age at which maximum weight was reached was to be prior than the age when the first component happened; and 5) Two or more components could not first have been diagnosed at the same time. Otherwise, the MS may already have developed, and analyses could not be pursued. After the above selection, only 2,123 participants were chosen as Sample 2 for the study. (Figure 1)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Flow chart of study participants.

Abbreviations: MS: metabolic syndrome; AO: abdominal obesity.

https://doi.org/10.1371/journal.pone.0078270.g001

Data Analysis

Data were analyzed using SPSS 18.0 or Stata 11.0 software for Windows. Data were expressed as mean ± SD, median with interquartile range, or percentage as suitable. Comparisons between groups were analyzed by t-test or Mann-Whiteney U-test for measurement data, and chi-square test for enumeration data P < 0.05 was defined as the threshold for statistical significance.

The prevalence of MS and its components in Sample 1 was analyzed and gender differences were compared. Age-standardized point prevalence estimates and confidence intervals (95% CI) for MS and its components stratified by sex were calculated using Stata (version 11.0) svy commands to account for the multi-stage stratified random sampling design. The calculation was weighted on the basis of Chinese population data from 2006 [33]. The presence of two or more of four components of MS could be diagnosed as MS. Thus, people with AO have a total of 10 combinations of components, including six kinds of combinations of two components, three kinds of combinations of three components, and one kind of combination of four components. The distribution characteristics of each component at different ages, and combinations of different components were analyzed by gender.

Survival analysis was used to study participants with AO in Sample 2 from the occurrence of the first abnormal component to the development of MS. The initial event was defined as the first occurrence of an abnormal component, while the end event was defined as the development of MS. The study time was defined as the time period between the initial event and the end event. If MS had not yet developed, it was defined as a censored value. Gender differences in prevalence were compared between participants aged < 50 years and those aged ≥ 50 years. A survival curve was plotted using Kaplan-Meier method and cumulative prevalence was calculated by life table method; gender differences were analyzed by Cox regression analysis. Andersen 95% confidence intervals for the median survival time of groups were constructed by a forward stepwise method.

Characteristics of study subjects

In Samples 1 and 2, the age and maximum weight age of males were significantly lower than those of females (p < 0.05). There was significant gender difference in educational level and lifestyle risk factors as cigarette smoking and alcohol drinking. The height, current and maximum weight, body mass index, waist circumference, diastolic blood pressure and triglyceride in males were significantly higher than those in females (p < 0.05), while the HDL-c in females was significantly higher than that in males (p < 0.05). There was no gender difference in TC in Sample 1 or in fasting blood glucose or systolic blood pressure in Sample 2 (p > 0.05). However, TC was significantly higher in females than that in males in Sample 2 (p < 0.05). (Table 1)

Gender difference in the prevalence of individual components and MS

The age-standardized prevalence of MS in people with AO was 50.5%. The prevalence in males (73.7%) was significantly higher than that in females (36.9%). Hypertension as one of the MS component had the highest age-standardized prevalence (48.9%), followed by hypertriglyceridemia (43.2%), HDL-c (34.0%), and hyperglycemia (32.4%). In males, HDL-c was the component with the highest prevalence (67.7%), followed by hypertension (61.3%) and hypertriglyceridemia (55.7%) showing moderate prevalence, and hyperglycemia having the lowest prevalence (36.4%). In females, the MS component of hypertension had the highest prevalence (41.3%), while HDL-c had the lowest prevalence (14.8%). (Table 2)

Gender difference in the prevalence of MS component combinations

Males had the highest prevalence of MS with the combination of hypertension plus hypertriglyceridemia plus HDL-c components (14.5%), while females had the highest prevalence with the combination of hypertension plus hypertriglyceridemia (10.4%) and hyperglycemia plus hypertension plus hypertriglyceridemia (10.3%). In addition, the proportions of combinations of three and four components were significantly higher in males than those in females (36.4% vs. 30.2% and 18.4% vs. 5%, respectively). (Figure 2)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Prevalence of metabolic syndrome components combination in male and female participants with abdominal obesity.

Abbreviations: BG: raised fasting plasma glucose, BP: raised blood pressure, TG: raised triglycerides, HDL: reduced high-density lipoprotein cholesterol.

https://doi.org/10.1371/journal.pone.0078270.g002

Gender difference in the development from the initial component to MS

There were 1,730 individuals that developed MS among the 2,123 participants with AO in Sample 2, within the total observation time. MS developed fast at first and then became slow down later. The 5-year and 10-year cumulative prevalence of MS were 58% and 83%, respectively. Half of the participants with AO developed MS after 3.9 years (95% CI: 3.7 - 4.1), whereas 75% developed MS after 7.7 years (95% CI: 7.5 - 7.9) (Figure 3). After adjusted for ethnics, educational level, cigarette smoking, alcohol drinking, and physical activity, males developed MS significantly faster than females in the group aged < 50 years (RR: 1.533, 95% CI: 1.178–1.993), whereas there was no significant difference in the development of MS between males and females aged ≥ 50 years (RR: 0.974, 95% CI: 0.837–1.134) (Figure 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Kaplan-Meier curves of the cumulative incidence rates of metabolic syndrome for all participants in sample 2.

Abbreviations: MS: metabolic syndrome.

https://doi.org/10.1371/journal.pone.0078270.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Kaplan-Meier curve of effect of gender for participants in sample 2with ages below 50 years (A) and above 50 years (B) divided by gender.

Abbreviations: MS: metabolic syndrome.

Male as reference. The relative ratio (RR) and 95% confidence intervals (CI) were calculated by cox regression after adjustment for ethnics, educational level, cigarette smoking, alcohol drinking, and physical activity.

https://doi.org/10.1371/journal.pone.0078270.g004

Limitations

The main limitation of our study is that TC was used as one component instead of HDL-c, because we could not obtain diagnostic history of HDL-c in China. There are two reasons for choosing TC to replace HDL-c. One reason is that the purpose of studying MS was to screen people at risk of CVD and diabetes, while TC has already been demonstrated to be associated with CVD [43-46]. The other reason is that it was still a feasible method, although we might screen different populations using TC as an alternative component. Previous studies have shown that about 21% of the high-risk population was filtered out as MS using three official criteria. However, only 9% of the high-risk population was screened by all three standards; in other words, about half of the screened population was different for each standard [3]. In fact, although not intending to compare, we discovered that multiple metabolic indexes (e.g. waist circumference and blood pressure) were higher in Sample 2 than in Sample 1 in the present study, which indicated that we might have overestimated the rate of development of MS to some extent. However, due to a lack of prospective studies, the use of TC as an alternative component is reasonable and feasible for studying the characteristics of MS development.

The second limitation of the study is recall bias, including previous weight and diagnostic history of abnormal components, which were unavoidable for a retrospective study. Another concern is whether people with abnormal components will be diagnosed in time. Thus, we randomly selected some of our subjects and called them back to ask about their previous history. It was found that the consistency of memory was 87% and the Kappa value was 0.78. Considering that AO people experiencing the occurrence of the first abnormal component will be concerned about their health status and take regular physical examinations, we believe that the above reasons could dispel any concerns about recall bias to some extent.

In addition, the number of females was higher than that of males because females had a relatively higher response rate than males, which might cause a potential selection bias [27].