Background
Childhood obesity has been found to be a precursor of health problems in adulthood [
1,
2], associated with insulin resistance [
3], abnormal glucose metabolism [
4], and elevated blood pressure [
5]. These abnormalities are identical to the characteristics of adult metabolic syndrome (MetS). Various groups have proposed several criteria of metabolic syndrome in children [
6‐
8] by consensus among experts [
9,
10]. Among those, the definition of the International Diabetes Federation (IDF) was most recently published, and has been widely used in research [
11‐
13]; although this definition still has several limitations with this definition [
14]. All criteria including the IDF definition, uniformly emphasize the importance of central obesity, defined as an excessive waist circumference (WC). In children and adolescents, percentiles, rather than absolute values of WC by age and sex, are more reliable measures of MetS, and they compensate for large variations throughout development. For example, the IDF defines a WC ≥ 90th percentile across and sex groups as central obesity in children and adolescents. The development of ethnic-specific WC percentiles has been strongly recommended to define central obesity in children across the world [
10]. To date, two sets of WC percentiles by age and sex in the Japanese pediatric population have been reported [
15,
16]. However, the notable drawback of these percentiles is that the definition of WC as measured at the site of maximum waist narrowing or the top of the iliac crest, which differs from the standard Japanese site, the umbilical position. In Japan the umbilical position is defined as a measurement site of WC for adult MetS criteria in Japan [
17]; these criteria are based on the relationship between individuals’ amount of abdominal fat, as measured by computed tomography, and their WC, both at the level of the navel [
18]. Although this method is not used internationally [
19‐
21], studies have indicated a strong correlation between the umbilical WC and the risk of MetS in the Japanese population [
22,
23]. Based on this background, the umbilical position has been used as the Japanese standard for evaluating WC in health examinations in both adults and children. Given the application of WC percentiles as a screening tool within the health checkup system, the development of a new set of WC percentiles for Japanese children is ideal, using the data collected from measurements at the umbilical position in Japanese children.
This study aimed to develop age- and sex-specific percentiles of WC and WC-to-height ratio (WC/Ht), the latter of which has been reported to be another sensitive indicator of MetS in adulthood and childhood [
24‐
27]. Furthermore, we tested the utility of the developed WC and WC/Ht percentiles to screen for childhood MetS in accordance with the IDF definition.
Discussion
This was the first trial to develop WC percentiles by age and sex in the Japanese pediatric population using a new dataset of measurements at the standardized Japanese site (the umbilical site). We showed that the new WC percentiles have the potential to distinguish children with MetS from obese children who are otherwise healthy according to the IDF criteria, which are used worldwide.
Currently, there is no consensus on the WC measurement site for the pediatric population, although several positions are used [
31]. In response to this situation, the international definitions of childhood MetS, such as the IDF criteria, do not define a measurement position for WC. Furthermore, the relationship between WC values and cardiometabolic risk factors do not appear to be affected by measurement site, including at the narrowest part of the waist, the midpoint between the lowest rib and iliac crest, or just above the iliac crest and the umbilical level [
32]. In Japan, the adult criteria for MetS included a WC measurement at the umbilical site for the diagnosis [
17], based on a previous study [
18]; therefore, the umbilical position is used as the standard position for all ages at health examinations. To date, two sets of WC percentiles by age and sex in the Japanese pediatric population have been reported [
15,
16]; however, those studies used WC measurements from the narrowest part of the waist or the top of the iliac crest, not at the umbilical site. Therefore, it should be emphasized that the new WC percentiles are more useful for screening childhood MetS through school health checkups.
Sex differences in visceral fat as well as WC are believed to appear at puberty [
33‐
37], with boys showing more visceral fat than do girls, due to sex differences in hormonal concentrations [
38]; however, in prepubescent children, visceral fat levels are roughly similar. It is noteworthy that the actual value of the boys’ WC at the 90th percentile in this study becomes greater than that of girls at around the age of six (Fig.
1 and Additional file
1); nevertheless, gender differences in sex hormones are thought to be negligible at this age. A study of WC percentiles in Japanese children reported similar results [
16], although the WC measurements differed from those in this study. Turkish boys also showed a predominance of central adiposity throughout early childhood [
39], indicating the presence of sex differences in visceral fat deposits before puberty, which is consistent with previous findings [
40]. This study underscores the importance of developing sex-specific WC percentiles by age from early childhood to puberty. It is worth mentioning that pubertal status can affect metabolic parameters, including blood pressure, lipid profile, and glucose metabolism [
41,
42]. In this study, the pubertal status of the participants was not considered. One reason is that pubertal status is not included in the IDF definition of MetS. Besides, because of the lack of individual privacy during the school health checkups, puberty status could not be assessed. Further studies on the screening of childhood MetS, so as to examine the strength of the effect of pubertal status on metabolic parameters are necessary.
The 90th percentile for WC has been reported as a reliable cutoff for screening metabolic abnormalities in children [
43‐
45]. Because there were no MetS subjects in boys and girls with a WC < 90th percentile, we demonstrated that the cutoff point of WC at the 90th percentile involved a certain level of reliability for distinguishing children with MetS, according to the IDF criteria, from obese children who were otherwise health, When the WC cutoff was set at the 97th percentile, some of the participants below that percentile were found to have multiple metabolic abnormalities. This finding indicated that the WC at the 97th percentile was inadequate in distinguishing children with MetS from those who were obese but were otherwise healthy. The validity of WC at the 90th percentile was supported by the results obtained from the statistical analysis, although data from all the participants were integrated, due to extremely small number of non-obese subjects. These results are compatible with previous reports that a slight increase in metabolic risk factors begins at the 75th WC percentile, while a considerable increase begins at the 90th percentile [
46,
47]. Further studies are needed to clarify specific cutoffs by ethnicity, while satisfying sensitivity and specificity criteria to screen for childhood MetS.
The WC/Ht ratio has the potential to be a better screening index for metabolic risks in adults than does WC [
27]. Since WC/Ht does not change dramatically in accordance with the ages of children and adolescents, specific references are unnecessary, giving it an advantage in screening for childhood MetS. The WC/Ht cutoff was 0.5, which was created based on adult studies showing an association between WC/Ht and metabolic risk [
25,
26], and the same 0.5 cutoff can be used across all ages in pediatric populations [
27,
48]. In our study, the WC/Ht 90th percentile of the boys was constant at 0.5 across age. Therefore, it can be used as a simple screening parameter to predict a cluster of metabolic risk factors in boys. The 90th percentile curves of the girls in our study were not constant at 0.5 across age, indicating that the value of 0.5 WC/Ht was not a good screening tool for MetS in girls. This result is consistent with a study showing that the WC/Ht cutoff was not constant at 0.5 for all ages in [
49] suggesting the need for specific age and sex references.
There are some limitations in the present study. The first limitation is that the new percentile curves were based on a sample from only one prefecture in Japan, not from randomly selected regions across the entire country, which restricts their generalizability as national references. Therefore, it might be difficult to apply the results of this study to Japanese children in the general population. However, mean heights and weights were similar to those of Japanese children of the same age in the general population. Furthermore, several studies reported that the distributions of height and weight in school-aged children of Shizuoka prefecture were similar to those in the national survey [
50,
51]. Therefore, the new WC percentiles from this study have the potential to be of great value in their application by clinicians to the general population of Japanese children. Studies that use data collected from several other regions of Japan are needed to establish national references. The second limitation of the present study is that the samples used for the utility evaluation of the new WC percentiles were obtained from only one site, Osaka prefecture, and did not include participants who were not overweight, resulting in a relatively small number of children being classified below the WC cutoff values (90th or 97th percentiles). Therefore, we cannot provide firm evidence supporting the use of the new set of WC percentiles to screen for childhood MetS (in accordance with the IDF definition) throughout the country. It should also be taken into account that the validity of the WC and WC/Ht percentiles is retrospective (using the pooled data collected before the establishment of those percentiles). Considering that the incidence of childhood obesity in Japan has either remained the same or decreased since 2000 [
52], this trial of validity should be acceptable. It must be mentioned that the result of poor AUC values and low specificity caused by relatively small number of MetS subjects is a significant weakness in this study. Given that the WC 90th percentile cutoff showed a good sensitivity (1.00) with relatively high AUC values among three cutoff points (75th, 90th, or 97th), it might be possible that this cutoff values has a potential for the screening of childhood MetS. Future studies using a data set of newly recruited subjects, including non-obese subjects from the entire country are necessary to ensure the reliability of the WC 90th percentile cutoff for screening childhood MetS in Japan.
Competing interests
We have no competing interests to declare.
Authors’ contributions
R.M. conceived the original idea for the study, collected the data, and analyzed them; T.I. made the percentile curves; R.T. collected the data from the participants who were obese; E.S., R.Y., K.K., E.N., S.S., and T.N. obtained data to evaluate the percentiles; Y.N., T. Ohzeki, and T. Ogata supervised this project; R.M. and Y.F. wrote the paper; and Y. F. will act as guarantor for the paper. All authors read and approved the final manuscript.