Introduction

The prevention of metabolic syndrome (MetS), for which visceral fat accumulation and insulin resistance are considered upstream factors, has recently attracted the attention of the medical world as a useful approach to protect against lifestyle-related diseases typified by arteriosclerotic diseases [1–8]. Visceral fat accumulates for many reasons, including hyperalimentation and inadequate exercise, among others, and causes the abnormal functioning of fat cells and excessive secretion of hormones that are involved in various pathological conditions [9, 10]. Excessive secretion of these hormones is thought to act in combination with other factors to cause arteriosclerotic and other serious diseases, such as renal failure, blindness, lower limb amputation, cerebral apoplexy, cardiac arrest, and cerebrovascular diseases. The progression of conditions, from obesity into serious diseases, is sometimes referred as the metabolic domino effect [11, 12], and includes fatty liver disease.

Diagnostic criteria for MetS have been published by the World Health Organization [13], American National Cholesterol Education Programs, Adult Treatment Panel III (NCEP–ATP III) [14], and International Diabetes Federation (IDF) [15] for Asian countries, including Japan [16]. In Japan, the Examination Committee for Criteria of MetS introduced diagnostic criteria for Japanese metabolic syndrome (JMetS) [16], which are similar to the ones defined by IDF. The criteria essentially include central obesity and several other components, such as hypertension, hyperglycemia, and abnormal lipid metabolism. In Japan, the most prominent difference between the IDF and Examination Committee criteria for evaluating central obesity is in the cut-off point (COP) for waist circumference (WC), especially that for women: in all countries of the world, with the exception of Japan, the COP for WC is larger for men than that for women.

The relative newness of the MetS concept necessitates that the diagnostic criteria be updated as and when needed. The association between the diagnosis of MetS and downstream diseases in the metabolic domino needs to be addressed in prospective studies. In the study reported here, we applied several criteria to examine the association between metabolic status and concurrent fatty liver, which we used as a specific example of a disease in the metabolic domino. Our aim was to identify preliminary criteria and COPs for WC that can be used in diagnosing MetS.

Subjects and methods

Height, weight, and WC were measured, and abdominal ultrasonography was performed in 2,333 subjects (1,195 men and 1,138 women) of 2,428 subjects aged 40–79 years. These subjects underwent comprehensive medical examinations at the Kasugai City Medical Center during a 3-month period between April and July 2006. Patients receiving drug treatment(s) for liver diseases, hypertension, diabetes mellitus, or hyperlipidemia were excluded from the study. Height and weight were measured using an automatic scale (Tanita BF-220). The WC was measured in standing subjects with a tape measure placed horizontally at the level of the navel while the subject was gently exhaling. If the abdomen was protuberant and the navel was deviated downwards, the tape measure was placed at the midpoint level between the lower intercostal border and the anterior superior iliac spine.

Fatty liver was diagnosed after discussion with medical technologists (including ultrasound technicians), radiology technologists, and physicians and by taking fatty liver scores (as shown in Table 1) obtained at Kasugai City Medical Center into consideration. These scores were based on previous studies [17–20].

Table 1 Fatty liver score
Full size table

Blood pressure was measured on the right arm using a mercury sphygmomanometer; the subject was in a lying position and had rested for at least 5 min prior to the measurement. Venous blood samples were collected in the morning from subjects after a fasting period of 12 h. Triglyceride (TG) and serum high-density lipoprotein cholesterol (HDL-C) were measured by the direct enzymatic method, and fasting plasma glucose (FPG) was measured by the glucose oxidase method. Their concentrations were measured using an automated analyzer (model 7170S; Hitachi, Japan).

Current JMetS criteria require a central obesity (visceral adipose tissue area ≥100 cm2 or WC ≥85 cm for men and ≥90 cm for women) and two or more of the following three components: (1) high blood pressure, based on a systolic blood pressure ≥130 mmHg and/or diastolic blood pressure ≥85 mmHg; (2) hyperglycemia, based on FPG ≥110 mg/dl; (3) abnormal lipid metabolism, based on TG ≥150 mg/dl and/or HDL-C <40 mg/dl [16]. The Examination Committee for Criteria of MetS in Japan also defined a “risk group for MetS” (yobi-gun) consisting of people who have central obesity and one of the three components listed above (high blood pressure, hyperglycemia, or abnormal lipid metabolism). In our study, as in most epidemiological studies, only WC was considered in our evaluation of central obesity; the visceral adipose tissue area was not assessed.

Our primary aim was to identify and propose new MetS criteria based on our results. Our suggested criteria (our criterion 1) considers central obesity not to be an essential requirement for MetS but as only one of the components of MetS. Accordingly, we defined our patients as having MetS when they demonstrated three or more components of MetS, regardless of their central obesity status. Similarly, the risk group for MetS consisted of those individuals who demonstrated two components.

Taking the number of MetS components listed above in consideration, we first calculated the odds ratios of fatty liver according to central obesity status in men and women by logistic regression. We then constructed receiver operating characteristic (ROC) curves to assess the detecting power of MetS criteria for concurrent fatty liver and calculated the areas under the curve (AUC) for diagnostic criteria. These procedures were repeated using the IDF COP for WC in the Japanese population, i.e., ≥90 cm for men and ≥80 cm for women (our criterion 2). We also calculated the COP for the largest AUC and suggested an optimal COP for men and women based on the study results. Statistical analyses were performed using the SAS system for Windows (release 9.1.3; SAS Institute, Cary, NC), and the AUC value was obtained to refer to the c statistic in PROC LOGISTIC output. All statistical tests were two-sided, and a P value <0.05 was considered to be significant. The study was approved by the ethics committee of Nagoya City University.

Results

Table 2 shows the number of subjects diagnosed with MetS according to the JMetS criteria and our newly proposed criteria, respectively. This diagnosis was based on the number of MetS components, other than central obesity, calculated by WC status in both men and women. Only 8.4% of the women satisfied the central obesity criterion of JMetS, whereas 26.7% men satisfied the criterion. When the COP for central obesity was changed to ≥80 cm, 36.6% of women satisfied the criterion. Among the 13 men and six women who were newly diagnosed with MetS based on our criteria using the same WC COP, seven men (53.8%) and five women (83.3%) had fatty liver. The prevalence of fatty liver was much higher than the total prevalence of fatty liver in men and women, i.e., 27.1 and 16.5%, respectively.

Table 2 Criteria of metabolic syndrome and number of subjects
Full size table

Table 3 shows the characteristics of the subjects diagnosed with MetS based on the application of several criteria. The prevalence of MetS using the JMetS criteria was 6.2% in men and 2.0% in women; based on our criteria using the JMetS COP for central obesity, MetS prevalence was 7.3 and 2.5%, respectively. When we applied the criterion for ≥80 cm COP for central obesity in women using our criteria, the prevalence of fatty liver increased to 4.9%. Similarly, the application of the COP increased the prevalence among the MetS risk group to 21.1%, which was close to that observed in men according to our criteria which include the ≥85 cm COP for central obesity. Since central obesity is an essential criterion for determining JMetS or the JMetS risk group, the subjects in these categories are much more obese than those falling in the normal category. The difference in WC and BMI between subjects in the MetS group and the normal group was 12.1 cm and 3.5 kg/m2, respectively, in men and 17.6 cm and 5.7 kg/m2 in women. When our criteria were used, these differences decreased to 10.4 cm and 3.0 kg/m2, respectively, in men and 14.5 cm and 5.0 kg/m2 in women. When the COP of ≥80 cm was applied, the differences decreased to 12.8 cm and 4.1 kg/m2, respectively.

Table 3 Characteristics of the subjects by MetS status
Full size table

Table 4 shows the odds ratios and 95% confidence interval (CI) for fatty liver according to the number of MetS components other than central obesity by WC status. Regardless of sex and the WC COP selected, a strong linear trend was observed for the association (trend P < 0.0001) with the number of components. The odds ratio for subjects without central obesity and with all three components of MetS was 9.69 (95% CI 3.1130.2) in men and 55.3 (6.34–483) in women. Using the ≥90 and ≥80 cm COP criterion for central obesity in men and women, respectively, the odds ratio was 55.3 (6.34–483) and 62.4 (6.23–626). These point estimates of odds ratios were higher than those of MetS subjects with two risk factors other than obesity among women, and even among men, they were higher than those of the risk group for MetS who satisfied the central obesity criterion.

Table 4 Odds ratio and 95% confidence interval for fatty liver according to the number of the components of MetS other than obesity by waist circumstance status
Full size table

Figure 1 shows the ROC curves for the diagnosis of fatty liver according to MetS status by the JMetS criteria and by our criteria. The AUC for the JMetS criteria and our criteria 1 and 2 in men was 0.638, 0.681, and 0.655, respectively. In women, the AUC for our criteria using ≥90 and ≥80 cm COPs for central obesity were 0.625 and 0.681, respectively, whereas that for the JMetS criteria was only 0.570. Based on the findings of our study, the largest AUC was recorded using our criterion 1 (≥85 cm) in men and our criteria 2 in women (≥80 cm). The shapes of the ROC curves of our criterion 2 for men and our criterion 1 for women were very similar, with the coordinates (false positive rate, true positive rate) for MetS and the risk group for MetS being (0.030, 0.188) and (0.204, 0.543), respectively, for men and (0.023, 0.181) and (0.205, 0.537), respectively for women. In addition, when WC was considered as a component, the COP for the largest AUC among men and women was ≥82 cm (0.701) and ≥77 cm (0.699), respectively. We therefore conclude that it would be both practical and appropriate to take WC into consideration, with WC COPs of ≥85 cm for men and ≥80 cm for women. In our study population, 26.7% of the men and 36.6% the women satisfied the criteria.

Fig. 1
figure 1

Receiver operating characteristic curves for fatty liver diagnosis by metabolic syndrome status of several criteria. JMetS Japanese metabolic syndrom, WC waist circumference

Full size image

Discussion

In the present study, we considered concurrent fatty liver to be a specific example of a disease in the metabolic domino of MetS and observed that the accumulation of MetS components was associated with higher odds ratios, even without the central obesity component. Taking these results as a whole, we observed stronger associations between MetS and fatty liver in men and women when we considered central obesity as a component rather than an essential requirement for the diagnosis of MetS. We therefore suggest that individuals with an accumulation of components should be regarded as having MetS even in the absence of central obesity, since fatty liver is a component of the metabolic domino. In addition, these individuals may belong to a risk group for other metabolic diseases, including cardiac arrest and cerebrovascular diseases. We also suggest that the optimal COP for WC should be ≥85 cm for men and ≥80 cm for women.

Although the main concepts of MetS are consistent, the COPs for defining central obesity for MetS are controversial, especially in Japan [21]. Several studies have been performed to elucidate the optimal COPs in which ROC analyses with obesity and two or more MetS components other than obesity [22–25] were used. The results suggested that the optimal cut-offs for men and women are 84–90 and 78–82 cm, respectively. Our results are consist with these reported values. However, these earlier studies were based on the internal consistency of obesity and MetS components other than obesity. Further ROC analyses need to be performed to establish the optimal COP for WC, and these should include certain diseases not currently included in MetS. This study is one such analysis.

An important question is whether central obesity should be considered as a requirement for the diagnosis of MetS or as a component of MetS. To answer this question, we need to examine the association between the number of MetS components and particular diseases stratified by central obesity. To date, there have been only two prospective cohort studies [26, 27] from Japan on cardiovascular diseases. Results from NIPPON DATA [26] show the existence of risk accumulation among non-obese subjects, whereas those from Hisayama-cho [27] indicate there is no risk accumulation in such subjects. Data from many studies, including those from our study, are required to facilitate further discussion on this question. However, before the absence of risk accumulation can be established among non-obese individuals, it is possible to treat central obesity as a component of MetS as a precautionary measure.

In general, if a factor is considered to be an essential requirement for the diagnosis of a certain disease, then that factor should not only be etiologically essential but also amenable to accurate measurement in practice; at the very least, the COP should be a sensitive measure. Otherwise, a considerable number of cases would not be detected by the criterion. In fact, the COPs based on the IDF criteria (≥94 cm for men and ≥80 cm for women), with central obesity as a requirement, are more sensitive than those of the NCEP-ATP III criteria (≥102 cm for men and ≥88 cm for women), wherein central obesity is considered a component. Although the JMetS definition is similar to the IDF definition, the JMetS COP for WC in women (≥90 cm) is much less sensitive than the COP of the IDF (≥80 cm). The COP for central obesity for the diagnosis of JMetS is based on the association between visceral fat area and WC [16]. The committee reported that simple correlation analysis of the regression line in women indicated that a WC corresponding to 100 cm2 of visceral fat was 92.5 cm. However, the correlation coefficient was only 0.65, and more than half of the women with a visceral fat area ≥100 cm2 would not be found using the WC COP of ≥90 cm (meaning that sensitivity is <0.5). The poor sensitivity of the WC in detecting abdominal adiposity is directly linked to the poor sensitivity of the JMetS criteria, in which WC is an essential requirement.

Conclusion

Based on the findings of our study, we suggest that a WC of ≥85 cm for men and ≥80 cm for women would be optimal COPs for central obesity for the diagnosis of MetS in the Japanese population. We also suggest that central obesity should be used as a component of MetS rather than an essential requirement for the diagnosis of MetS. No definite conclusion has yet been reached regarding the most appropriate diagnostic criteria for MetS. However, within the framework of our study in which fatty liver was considered to be an independent variable, we found that defining abdominal circumference as a component of MetS was less likely to cause errors of oversight and was thus more appropriate than considering abdominal circumference to be a required criterion. The challenge for the future is to identify pathologic conditions that are responsible for MetS and to find better diagnostic criteria through further similar studies that consider factors, other than fatty liver, involved in the metabolic domino effect [11, 12] as independent variables.