Background
Central obesity is a medical condition in which excess abdominal fat accumulates resulting in increased waist size, while overweight/obesity occurs when weight is higher than what is considered healthy weight for a given height (
http://www.cdc.gov/obesity/adult/defining.html). Both of these conditions can result in various chronic non-communicable diseases (CNCDs), such as cardiovascular disease (CVD), metabolic syndrome (MS), and diabetes mellitus (DM) [
1‐
4]. Obesity has become a worldwide and major public health problem due to increasing prevalence. Data from World Health Organization (WHO) showed that in 2016, more than 1.9 billion adults were overweight [
5]. In China, the prevalence of obesity was 10.1% for Chinese urban adults in 2015 [
6]. Compared with hypertensive patients with normal weight, the probability of suffering from hypertension complications in patients with obesity and overweight is increased by 11.65% and 6.45%, respectively. The per capita annual medical expenses due to overweight and obesity are 1410 and 985 RMB respectively, accounting for 22.79% of the medical expenses of hypertension complications [
7]. For diabetes, the economic burden caused by diseases attributable to systemic obesity and central obesity are 11.2 and 38.8 million, respectively [
8].
Fortunately, people with obesity can be treated appropriately, thereby reducing the incidence of CNCDs and greatly improving the living quality of individuals. Therefore, it is necessary and helpful to screen individuals with obesity at early stages. Waist circumference (WC) is widely used to measure central obesity while body mass index (BMI) is used as an indicator of overweight/obesity [
9]. There are sustained efforts by researchers to find better indices for screening obese subjects due to the limitations of WC and BMI [
10,
11]. For example, the diagnosis of central obesity cannot be made on pregnant women and subjects with abdominal ascites or masses using WC, which can also be influenced by many other factors including meal, respiration, or health conditions. Furthermore, BMI cannot determine body fat distribution nor can it be used to distinguish between muscle and body fat mass, which is the reason why athletes tend to have higher BMI readings suggestive of overweight/obesity, even though their extra weights are due to increased muscle mass not fat.
Neck circumference (NC) is recognized as a screening measure for identifying obese individuals [
12‐
14]. In fact, recent studies have focused on its diagnostic accuracy for central obesity and overweight/obesity [
15‐
17]. However, the results of these studies have been inconsistent possibly due to differences in the study population, diagnostic criteria and lifestyle. Additionally, the sensitivity, specificity, diagnostic odds ratios (DORs) and the area under the receiver operating curve (AUC) of NC as a measure of central obesity and overweight/obesity have not been reported in a meta-analysis. Thus, this study evaluated the efficacy of NC as a measure of central obesity and overweight/obesity through a cross-sectional survey that was analyzed with other similar studies to form a meta-analysis using a hierarchical summary receiver operator characteristic (HSROC) model. This study was designed to explore a universally applicable and convenient method of screening central obesity and overweight/obesity, so that appropriate measures can be taken on time to slow down the progression of obesity and related disease. It is hoped that the findings can have huge implications on prevention of CNCDs.
Discussion
To our knowledge, the meta-analysis on the subject that explored the gender-specific relationship between NC and central obesity as well as overweight/obesity has never been reported. In this study, our own epidemiological research was included to compare the results with that of meta-analysis and improve the representativeness of the study population through acquisition of reliable data from a larger sample size. The results showed that NC may not be a good tool for screening individuals with central obesity. However, it may be a simple and valuable surrogate indicator for BMI, which is significant to identify overweight/obesity in big epidemiological research and in some special occasions or crowd, prevent the development of the overweight/obesity by taking appropriate measures at early stage, and then reduce the incidence or complications of CNCDs.
The prevalence of obesity is increasing at an alarming rate, and the negative implications of this are well known [
36]. The WC and BMI are commonly used to identify individuals with central obesity and overweight/obesity. NC reflects the deposits of adipose tissue in the neck, which can be used as an indicator of subcutaneous adipose tissue in the upper-body [
37]. The neck is at the junction between the head and the trunk, and is often not covered by clothing, making it easily accessible for measurements. Similarly, NC measurements are less intrusive than those of WC and less cumbersome than those of BMI.
The ROC analysis of this epidemiological study found that NC ≥ 37.1 cm for male and ≥ 32.6 cm for female were determined to be the best cut-off points for screening individuals with central obesity. The thresholds observed by Zhang et al. [
27] that studied 9740 participants in Jiangxi, China are similar to this study. Another epidemiological research from Caloocan, Philippines revealed that NC ≥ 40.0 cm for male and ≥ 33.8 cm for female were the best cut-off points for screening individuals with central obesity [
28]. In the case of overweight/obesity, NC ≥ 37.4 cm for male and ≥ 32.2 cm for female were determined to be the ideal cut-off points in this study, while the cut-off points reported in India were 36.0 and 32.0 cm for male and female, respectively [
29]. The different cut-off points in these populations may have been predetermined by genetic and environmental factors, such as different medical condition and dietary habit. In addition, the objects of our study are undergraduates, so the composition of age is different from other studies, which would cause the inconsistent results among these researches.
Furthermore, we evaluated the efficacy of NC for screening central obesity and overweight/obesity by epidemiological research and meta-analysis. For central obesity, the efficacy of our survey (18–25 years old) was better than that of meta-analysis (≥18 years old); for overweight/obesity, the efficacy of our survey was similar to that of meta-analysis. That is to say, the efficacy of NC as a tool for screening central obesity might be different in each age stage, which may be because age can substantially impact on NC measurements. In this epidemiological study, the efficiency of NC for screening individuals with both central obesity and overweight/obesity in female was higher than in male. ROC curves showed that the best threshold of NC for screening individuals with central obesity and overweight/obesity might be different between male and female. Meta-analysis showed that NC predicted overweight/obesity better than central obesity, which suggested that NC may be a simple and valuable surrogate indicator for BMI, especially in female group. Reasons for the different efficacy between genders and types of obesity might be due to differences in body composition, sex hormone, distribution of adipose tissue and activity intensity between male and female. Studies have suggested that sex hormones may regulate body fat distribution [
38]. Androgen plays a key role in visceral adipose tissue accumulation in abdominal, while, estrogen can promote abdominal visceral adipose tissue transfer to subcutaneous and peripheral region [
39]. The main adipose in the neck is the subcutaneous adipose tissue. Studies have indicated that the correlation coefficient of subcutaneous adipose tissue and BMI was larger than that of subcutaneous adipose tissue and WC [
40,
41]. Therefore, NC could be a better indicator for screening overweight/obesity, especially in female, which was consistent with most of included publications. Furthermore, the results of meta-analysis revealed that there was substantial heterogeneity between studies and the HSROC curves were asymmetric except the curve of overweight/obesity for female. There might be several reasons that can explain these phenomena. (1) The number of satisfactory studies for meta-analysis was limited due to the less research for NC screening central obesity and overweight/obesity. (2) The variations in sample sizes and participants’ characteristics of each study may introduce heterogeneity between studies. (3) The critical value of WC and BMI for determining central obesity and overweight/obesity was diverse in different studies, which may affect the HSROC curves of meta-analysis.
In aggregate, the results of the present epidemiological research and meta-analysis suggested that NC may not be a good tool for screening individuals with central obesity. However, it may be a valuable tool for screening individuals with overweight/obesity, especially in female. From a public health perspective, it is valuable to be able use NC to assess overweight/obesity because it saves time and allows clinicians and researchers to increase the number of subjects investigated, especially in some special occasions, such as for expectant mothers, athletes and patients with ascites. Thus, NC might be a better surrogate index for screening overweight/obesity. Compared to WC and BMI, there are several unique advantages for NC. The measuring tool of NC is simple and can be carried conveniently. In winter, the use of thick clothing may erroneously give larger than actual WC and BMI values. The NC can be measured easily without considering the thickness of an individual’s clothes [
34,
42]. Additionally, NC cannot be affected by factors like meal, respiration or health conditions. For expectant mothers, NC could evaluate the levels of obesity better than WC and BMI, which can prevent the development of gestational diabetes mellitus and pregnancy-induced hypertension syndrome by taking appropriate measures when a lager NC was observed. Besides, NC is associated with MS, obstructive sleep apnea and cardiometabolic risk factors [
14]. Study showed that the relationship between MS and NC was stronger than that with WC [
30]. Therefore, identification of obesity in early stage, controlling weight and improving lifestyle on time will certainly permit drastic reductions in risk of MS and other CNCDs. Unfortunately, measurement of NC is not suitable for patients with certain diseases, such as goiter or neck tumor, otherwise it may overrate the prevalence of obesity. However, this study has some limitations. First, the number of studies included in the meta-analysis was small. In addition, all of the included participants were from Asia except the study by Coelho et al. Therefore, subgroup analysis based on age groups and continents were not conducted, and the cut-off values of NC we got cannot be generalized to a larger population. Second, the critical values of WC and BMI for determining central obesity and overweight/obesity were different in some studies, which resulted in the inevitable heterogeneity in the meta-analysis. Third, there was potential publication bias in the studies that used NC screening for central obesity in females, possibly due to different characteristics and limited number of included researches. Therefore, it would be helpful to examine these findings in other ethnic groups using larger samples, and increase the number of studies for meta-analysis, while subgroup analyses as well as meta-regression could be performed for age, region, and cut-off values of central obesity and overweight/obesity in future studies.