Abstract
Aims/Introduction
Elevated concentrations of fasting plasma glucose (FPG) or hemoglobin A1c (HbA1c) are well‐established independent risk factors for progression to diabetes, cardiovascular comorbidities and mortality. Most previous studies on the relationships of anthropometric measures with hyperglycemia were carried out among adults and adolescents, but few data are available for the performance predication of the predictors for diagnosing elevated FPG or HbA1c among young children.
Materials and Methods
Involving 5,556 students of aged 7–9 years, a school‐based cross‐sectional survey was carried out between March and June 2019 in Shenzhen, China. Receiver operating characteristic curve analysis was utilized.
Results
The median was 4.6 (interquartile range [IQR] 4.3–4.8) mmol/L for FPG and 5.3% (IQR 5.1–5.5%) for HbA1c levels for all participants. For detecting elevated FPG, weight (0.651, IQR 0.583–0.719) and waist circumference (0.650, IQR 0.584–0.717) showed the highest area under the curve and 95% confidence interval, followed by body mass index and the z‐score of body mass index (both 0.635, IQR 0.567–0.703); other anthropometric measures showed poorer diagnostic efficiencies or no ability. For detecting elevated HbA1c, lower efficiencies for the Conicity Index (0.651, IQR 0.583–0.719), waist‐to‐height ratio, waist‐to‐hip ratio and waist‐to‐chest ratio were shown. The correlations of FPG and HbA1c levels with anthropometric indices were weak (Spearman’s r ≤ 0.179).
Conclusions
None of the evaluated anthropometric indicators approached an adequate predictive accuracy for the detection of elevated FPG or HbA1c levels in Shenzhen children aged 7–9 years. The current study did not recommend anthropometry screening for prediabetes in young children.
Keywords: Anthropometric, Children, Hyperglycemia
Few data are available for the performance predication for diagnosing elevated fasting plasma glucose or hemoglobin A1c among young children. A school‐based cross‐sectional survey involving 5,556 students aged 7–9 years was carried out in Shenzhen, China. None of the evaluated anthropometric indicators approached an adequate predictive accuracy for the detection of elevated fasting plasma glucose or hemoglobin A1c levels. The current study did not recommend anthropometry screening for prediabetes in young children.
Introduction
Elevated concentrations of fasting plasma glucose (FPG) or hemoglobin A1c (HbA1c), also known as intermediate hyperglycemia, are well‐established independent risk factors for progression to type 2 diabetes mellitus 1 , 2 , 3 , cardiovascular comorbidities 4 , 5 and mortality 6 , 7 , and significantly and independently related to future development of hypertension 8 , 9 . Screening for elevated blood glucose concentrations could assist in identifying individuals at risk for the above chronic conditions, but blood collection is inherently invasive and cannot be widely used for a large‐scale survey, especially for non‐adults. In contrast, simple measurements of anthropometric parameters, such as height, weight, waist circumference (WC), hip circumference (HC) and chest circumference (CC), are relatively quick, inexpensive and non‐invasive, and therefore can be applied for large crowds in a massive epidemiological investigation. If the discriminative accuracy of anthropometry to detect hyperglycemia is high and anthropometry is closely related to blood glucose concentrations, the anthropometric indices could be useful surrogate markers for blood glucose levels.
Numerous studies on the association of anthropometric measures with hyperglycemia have been carried out in some countries and regions. For example, several epidemiological studies suggested that body mass index (BMI), WC, waist‐to‐hip ratio (WHpR) or waist‐to‐height ratio (WHtR) has an effective predictive power for diabetes or impaired fasting glucose 10 , 11 , 12 , 13 , 14 , 15 . On the contrary, a study from Durban of South African Asian Indians did not confirm these results, showing that anthropometric measures had a low predictive accuracy for the detection of diabetes in Asian Indians 16 . For adolescents, BMI and WC estimation are easy‐to‐use tools, and can be carried out for early detection of prediabetes for both boys and girls in India 17 . However, whether the potential indices of the z‐score of BMI‐for‐age (BMI‐z), waist‐to‐chest ratio (WCtR), chest‐to‐hip ratio (CHpR) and a novel body index of Conicity Index (C‐Index) for predicting elevated FPG and HbA1c were comparable or superior to the above‐mentioned parameters remained a controversial issue up to now. Furthermore, although most previous investigations on the relationships of anthropometric measures with hyperglycemia were carried out among adults and adolescents, few data are available for the performance predication of the predictors for diagnosing elevated blood glucose among young children.
In response to the insufficiency of evidence, the present study was designed to examine the diagnostic performance of 14 anthropometric parameters and subsequent combinations of variables in discriminating elevated FPG and HbA1c among Shenzhen children aged 7–9 years. Early screening and diagnosis for hyperglycemia among children are vital in preventing or delaying progression to some chronic and metabolic diseases.
Methods
Study design and data collection
The school‐based cross‐sectional survey was carried out between March and June 2019, involving grade 2 students from 19 schools, randomly selected for cluster sampling from 134 local primary schools, in the Bao’an District of Shenzhen, China. The new metropolis of Shenzhen is a major financial center, and the world’s largest manufacturing base, with a gross domestic product of $27,100 per capita in 2017, ranking number one in mainland China 18 , 19 . This study was approved by the institutional review board of Bao’an Central Hospital of Shenzhen (IRB‐PJ‐2018‐002).
Basic personal characteristics, such as name, sex and national identification number (used to calculate age), were recorded. Anthropometric parameters of standing height (to the nearest 0.5 cm), weight (0.1 kg), WC (1 cm), HC (1 cm) and CC (1 cm) for each student were respectively measured by trained physicians. By using a measuring tape, WC was measured midway between the lower margin of the rib cage and iliac crest, and HC was dimensioned around the widest portion of the buttocks in a standing position.
Blood samples were drawn into vacutainer tubes by trained nurses in the morning after children fasted for at least 10 h. FPG was measured using the hexokinase method with the automatic analyzer (Model 7170s; Hitachi, Tokyo, Japan), and the concentrations of serum HbA1c were measured utilizing the ion high performance liquid chromatography with Bio‐Rad VARIANT II (Hercules, CA, USA), which is standardized to the Diabetes Control and Complications Trial reference assay.
Data preprocessing
To set up a database, BMI (calculated in weight divided by the square of height; kg/m2), WHtR, WHpR, WCtR and CHpR were calculated for each participant 20 . C‐Index = WC / 0.109 / square root (weight / height), where WC is in meters, weight is in kg and height is in meters 21 . BMI‐z was calculated using the R macro (https://www.who.int/growthref/en/), based on an international norm of the 2007 World Health Organization growth reference for school‐aged children 21 .
A total of 6,192 students aged 7–9 years first agreed to participate in our survey, and written informed consent was obtained for each student and their parents before entering the study. We excluded 484 students due to missing data on any of the anthropometric measurements (i.e., identification number or age, height, weight, WC, HC and CC). Those with missing values on the concentrations of serum FPG or HbA1c (n = 148) were also excluded. Thus, effective statistical analyses were carried out in our investigation based on the remaining 5,556 participants.
Definition of elevated FPG or HbA1c levels
For the main analysis, in the current study, an elevation of serum FPG level among children was defined as FPG ≥5.6 mmol/L, and elevated HbA1c was defined as HbA1c ≥5.7%, with reference to the diagnostic criteria of prediabetes or intermediate hyperglycemia for adults 2 , 17 . As there was no consensus on what levels constituted an elevated serum FPG or HbA1c concentration for children and adolescents, we also reported a secondary analysis for the sake of comparison and to check the stability of our main findings, by changing the definitions of elevated FPG levels as equal to or greater than the 75th percentile (4th quartile) of FPG distribution (i.e., ≥4.8 mmol/L) and of elevated HbA1c levels as equal to or greater than the 75th percentile of HbA1c distribution (i.e., ≥5.5%) among all young participants.
Statistical analysis
The normal distribution for each variable was determined through the Shapiro–Wilk test, and all of the anthropometric and laboratory indices were non‐normal distribution (all P < 0.001). The non‐normally distributed measurements were described as the median with the interquartile range (IQR; 25–75th percentile), and significance for differences was evaluated by the Mann–Whitney U‐test. The receiver operating characteristic (ROC) curve analysis with area under the curve (AUC) was utilized to examine the diagnostic performance of each parameter as an indicator of FPG and HbA1c elevation 22 . The proposed cut‐off point for an anthropometric parameter was chosen by the highest Youden Index (sensitivity + specificity − 1) 23 . The Spearman’s rank correlation analyses for anthropometric measures with the levels of FPG and HbA1c among the skewed parameters were also carried out. All statistical analyses were processed by SPSS for Windows 21.0 (SPSS Inc., Chicago, IL, USA) results, with two‐tailed P < 0.05 considered statistically significant.
Results
Characteristics of participants
A total of 5,556 children aged 7–9 years were enrolled for analysis, of which 3,025 (54.4%) were boys. Table 1 shows the descriptive statistics for anthropometric and laboratory parameters. Each anthropometric index was significantly greater among boys than girls (all P < 0.05). Overall, the median was 4.6 (IQR 4.3–4.8) mmol/L for FPG and 5.3% (IQR 5.1–5.5%) for HbA1c levels.
Table 1.
Baseline descriptive statistics for anthropometric and laboratory parameters stratified by sex among children of Shenzhen aged 7–9 years in 2019
| Parameters | Total children (n = 5,556) | Boys (n = 3,025) | Girls (n = 2,531) | P |
|---|---|---|---|---|
| Age (years) | 8 (7–8) | 8 (7–8) | 8 (7–8) | 0.051 |
| Height (cm) | 130.0 (126.5–134.0) | 131.0 (127.0–134.5) | 129.5 (126.0–133.0) | <0.001 |
| Weight (kg) | 25.8 (23.2–29.5) | 26.5 (24.0–30.8) | 25.0 (22.7–28.2) | <0.001 |
| WC (cm) | 54 (51–58) | 55 (52–60) | 53 (50–57) | <0.001 |
| HC (cm) | 66 (63–70) | 66 (63–71) | 65 (62–69) | <0.001 |
| CC (cm) | 59 (57–63) | 60 (58–64) | 58 (56–61) | <0.001 |
| BMI (kg/m2) | 15.27 (14.21–16.79) | 15.51 (14.45–17.33) | 14.94 (13.92–16.27) | <0.001 |
| BMI‐z | −0.32 (−1.09 to 0.57) | −0.18 (−0.99 to 0.88) | −0.50 (−1.19 to 0.29) | <0.001 |
| WHtR | 0.418 (0.395–0.445) | 0.423 (0.400–0.454) | 0.411 (0.390–0.436) | <0.001 |
| WHpR | 0.826 (0.794–0.859) | 0.838 (0.806–0.870) | 0.815 (0.786–0.847) | <0.001 |
| WCtR | 0.914 (0.879–0.952) | 0.915 (0.879–0.955) | 0.912 (0.878–0.949) | 0.033 |
| CHpR | 0.905 (0.877–0.933) | 0.913 (0.886–0.940) | 0.892 (0.867–0.919) | <0.001 |
| C‐Index | 1.115 (1.075–1.158) | 1.123 (1.082–1.167) | 1.108 (1.068–1.148) | <0.001 |
| FPG (mmol/L) | 4.6 (4.3–4.8) | 4.6 (4.4–4.9) | 4.5 (4.2–4.7) | <0.001 |
| HbA1c (%) | 5.3 (5.1–5.5) | 5.3 (5.2–5.5) | 5.3 (5.1–5.5) | 0.136 |
Data are expressed as the median (interquartile range); the bold indicates a statistically significant association. BMI, body mass index; BMI‐z, z‐score of BMI‐for‐age; CC, chest circumference; CHpR, chest‐to‐hip ratio; C‐Index, Conicity index; FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; HC, hip circumference; WC, waist circumference; WCtR, waist‐to‐chest ratio; WHpR, waist‐to‐hip ratio; WHtR, waist‐to‐height ratio.
As shown in Table 2, children with an elevated FPG activity had greater anthropometric parameters than those with a normal FPG level (all P < 0.05); children with elevated HbA1c had higher WHtR, WHpR, WCtR and C‐Index values than those with normal HbA1c (all P < 0.05).
Table 2.
Baseline descriptive statistics for anthropometric parameters stratified by fasting plasma glucose (elevated fasting plasma glucose ≥5.6 mmol/L) and hemoglobin A1c (elevated hemoglobin A1c ≥5.7%) levels among children of Shenzhen aged 7–9 years in 2019
| Parameters | FPG | HbA1c | ||||
|---|---|---|---|---|---|---|
| Normal FPG (n = 5,495) | Elevated FPG (n = 61) | P | Normal HbA1c (n = 5,065) | Elevated HbA1c (n = 491) | P | |
| Age (years) | 8 (7–8) | 8 (8–8) | 0.414 | 8 (7–8) | 8 (7–8) | 0.787 |
| Height (cm) | 130.0 (126.5–134.0) | 133.0 (130.0–136.5) | <0.001 | 130.5 (126.5–134.0) | 130.0 (126.0–134.0) | 0.071 |
| Weight (kg) | 25.8 (23.2–29.4) | 29.4 (25.7–34) | <0.001 | 25.8 (23.3–29.5) | 25.5 (23.0–29.75) | 0.347 |
| WC (cm) | 54 (51–58) | 57 (54–63) | <0.001 | 54 (51–58) | 55 (51–59) | 0.186 |
| HC (cm) | 66 (63–70) | 68 (65–74) | <0.001 | 66 (63–70) | 66 (63–70) | 0.625 |
| CC (cm) | 59 (57–63) | 62 (59–65) | 0.001 | 59 (57–63) | 60 (57–63) | 0.504 |
| BMI (kg/m2) | 15.26 (14.2–16.76) | 16.13 (15.20–18.11) | <0.001 | 15.27 (14.21–16.76) | 15.26 (14.15–17.05) | 0.796 |
| BMI‐z | −0.33 (−1.09 to 0.57) | 0.16 (−0.41 to 1.24) | <0.001 | −0.32 (−1.08 to 0.56) | −0.34 (−1.13 to 0.72) | 0.767 |
| WHtR | 0.418 (0.395–0.444) | 0.432 (0.406–0.467) | 0.003 | 0.418 (0.395–0.444) | 0.421 (0.395–0.457) | 0.012 |
| WHpR | 0.826 (0.794–0.859) | 0.836 (0.806–0.879) | 0.021 | 0.825 (0.794–0.859) | 0.831 (0.799–0.871) | 0.029 |
| WCtR | 0.914 (0.879–0.952) | 0.919 (0.897–0.966) | 0.035 | 0.914 (0.879–0.951) | 0.919 (0.881–0.958) | 0.044 |
| CHpR | 0.905 (0.877–0.933) | 0.896 (0.879–0.925) | 0.592 | 0.905 (0.877–0.933) | 0.908 (0.875–0.933) | 0.691 |
| C‐Index | 1.115 (1.075–1.158) | 1.130 (1.097–1.169) | 0.056 | 1.114 (1.075–1.157) | 1.128 (1.078–1.171) | 0.001 |
Data are expressed as the median (interquartile range); the bold indicates a statistically significant association. BMI, body mass index; BMI‐z, z‐score of BMI‐for‐age; CC, chest circumference; CHpR, chest‐to‐hip ratio; C‐Index, Conicity index; FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; HC, hip circumference; WC, waist circumference; WCtR, waist‐to‐chest ratio; WHpR, waist‐to‐hip ratio; WHtR, waist‐to‐height ratio.
Univariate ROC curve analyses
Table 3 shows the information about the ROC curves for discriminating abilities of various anthropometric indices for screening elevated FPG and HbA1c separately. The anthropometric measures of weight (AUC 0.651, confidence interval [95% CI] 0.583–0.719) and WC (AUC 0.650, 95% CI 0.584–0.717) showed the highest AUC and 95% 95% CI for elevated FPG, followed by BMI and the z‐score of BMI (both AUC 0.635, 95% CI 0.567–0.703), but the differences were small. In addition, sex, height, HC, CC and WHtR had a poorer diagnostic efficiency in detecting abnormal FPG. Age, CHpR, and C‐Index had no ability to evaluate elevated FPG (all P > 0.05). For detecting elevated HbA1c, a low diagnostic performance for C‐Index (AUC 0.547, 95% CI 0.519–0.574) and lower efficiencies for WHtR, WHpR and WCtR were shown; the discriminatory power of other anthropometric indices in the prediction of elevated HbA1c was null (all P > 0.05). In the stratification analyses by sex, most diagnostic associations were similar, with poorer or null efficiencies (AUC 0.456–0.660 for boys and 0.437–0.644 for girls).
Table 3.
Area under the and 95% confidence interval of the receiver operating characteristic curve to judge the discrimination ability of various anthropometric parameters for detecting elevated fasting plasma glucose (≥5.6 mmol/L) and hemoglobin A1c (≥5.7%) separately
| Parameters | Total children (n = 5,556) | Boys (n = 3,025) | Girls (n = 2,531) | ||||||
|---|---|---|---|---|---|---|---|---|---|
| AUC | 95% CI | P | AUC | 95% CI | P | AUC | 95% CI | P | |
| Elevated FPG (≥5.6 mmol/L) | |||||||||
| Age | 0.524 | 0.452–0.597 | 0.511 | 0.546 | 0.396–0.697 | 0.550 | 0.514 | 0.432–0.597 | 0.737 |
| Sex | 0.614 | 0.549–0.680 | 0.002 | – | – | – | – | – | – |
| Height | 0.629 | 0.561–0.698 | 0.001 | 0.660 | 0.517–0.803 | 0.039 | 0.598 | 0.518–0.677 | 0.021 |
| Weight | 0.651 | 0.583–0.719 | <0.001 | 0.603 | 0.469–0.737 | 0.182 | 0.638 | 0.559–0.718 | 0.001 |
| WC | 0.650 | 0.584–0.717 | <0.001 | 0.572 | 0.431–0.713 | 0.354 | 0.644 | 0.569–0.718 | 0.001 |
| HC | 0.632 | 0.562–0.702 | <0.001 | 0.550 | 0.408–0.693 | 0.517 | 0.638 | 0.560–0.715 | 0.001 |
| CC | 0.627 | 0.556–0.699 | 0.001 | 0.545 | 0.405–0.684 | 0.563 | 0.619 | 0.538–0.700 | 0.005 |
| BMI | 0.635 | 0.567–0.703 | <0.001 | 0.532 | 0.403–0.660 | 0.684 | 0.641 | 0.564–0.719 | 0.001 |
| BMI‐z | 0.635 | 0.566–0.703 | <0.001 | 0.528 | 0.405–0.650 | 0.721 | 0.642 | 0.565–0.718 | 0.001 |
| WHtR | 0.609 | 0.539–0.680 | 0.003 | 0.482 | 0.347–0.618 | 0.818 | 0.622 | 0.543–0.700 | 0.004 |
| WHpR | 0.586 | 0.517–0.655 | 0.021 | 0.536 | 0.424–0.648 | 0.640 | 0.570 | 0.487–0.653 | 0.099 |
| WCtR | 0.579 | 0.509–0.648 | 0.035 | 0.561 | 0.427–0.695 | 0.429 | 0.581 | 0.500–0.661 | 0.058 |
| CHpR | 0.480 | 0.411–0.549 | 0.592 | 0.491 | 0.362–0.619 | 0.904 | 0.437 | 0.355–0.519 | 0.139 |
| C‐Index | 0.571 | 0.503–0.640 | 0.056 | 0.535 | 0.395–0.674 | 0.655 | 0.560 | 0.482–0.638 | 0.155 |
| Elevated HbA1c (≥5.7%) | |||||||||
| Age | 0.503 | 0.475–0.531 | 0.828 | 0.502 | 0.461–0.542 | 0.940 | 0.504 | 0.466–0.542 | 0.820 |
| Sex | 0.502 | 0.475–0.529 | 0.891 | – | – | – | – | – | – |
| Height | 0.475 | 0.449–0.502 | 0.071 | 0.460 | 0.423–0.498 | 0.050 | 0.490 | 0.453–0.528 | 0.603 |
| Weight | 0.487 | 0.459–0.515 | 0.347 | 0.456 | 0.416–0.496 | 0.031 | 0.513 | 0.475–0.551 | 0.469 |
| WC | 0.518 | 0.490–0.546 | 0.187 | 0.485 | 0.445–0.526 | 0.470 | 0.547 | 0.509–0.584 | 0.012 |
| HC | 0.507 | 0.479–0.534 | 0.626 | 0.479 | 0.439–0.520 | 0.310 | 0.529 | 0.491–0.566 | 0.120 |
| CC | 0.509 | 0.481–0.537 | 0.505 | 0.468 | 0.428–0.509 | 0.120 | 0.544 | 0.507–0.581 | 0.017 |
| BMI | 0.504 | 0.476–0.531 | 0.796 | 0.466 | 0.425–0.506 | 0.089 | 0.534 | 0.497–0.571 | 0.065 |
| BMI‐z | 0.504 | 0.476–0.532 | 0.767 | 0.466 | 0.425–0.507 | 0.094 | 0.533 | 0.496–0.570 | 0.074 |
| WHtR | 0.534 | 0.507–0.562 | 0.012 | 0.501 | 0.461–0.541 | 0.958 | 0.564 | 0.527–0.601 | 0.001 |
| WHpR | 0.530 | 0.502–0.557 | 0.029 | 0.503 | 0.464–0.543 | 0.867 | 0.556 | 0.518–0.594 | 0.002 |
| WCtR | 0.528 | 0.500–0.555 | 0.044 | 0.512 | 0.472–0.552 | 0.561 | 0.539 | 0.502–0.577 | 0.033 |
| CHpR | 0.505 | 0.478–0.533 | 0.691 | 0.486 | 0.444–0.527 | 0.477 | 0.519 | 0.482–0.556 | 0.301 |
| C‐Index | 0.547 | 0.519–0.574 | 0.001 | 0.519 | 0.479–0.560 | 0.347 | 0.570 | 0.533–0.607 | <0.001 |
The bold indicates a statistically significant association. Null hypothesis: true area = 0.5. AUC, area under curve; BMI, body mass index; BMI‐z, z‐score of BMI‐for‐age; CC, chest circumference; CHpR, chest‐to‐hip ratio; CI, confidence interval; C‐Index, Conicity index; FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; HC, hip circumference; WC, waist circumference; WCtR, waist‐to‐chest ratio; WHpR, waist‐to‐hip ratio; WHtR, waist‐to‐height ratio.
Figure 1 shows the ROC curves of significant indicators for identifying elevated FPG and HbA1c among total Shenzhen children. Based on the largest Youden Index, Table 4 shows the optimal cut‐points, and the probability of meaningful and significant indices in diagnosing abnormal FPG and HbA1c. The reasonable cut‐off points were 26.2 kg for weight and 55 cm for WC to differentiate elevated FPG, and 1.134 for C‐Index to recognize an elevation of HbA1c level.
Figure 1.
Receiver operating characteristic curve of significant anthropometric parameters for predicting elevated (a) fasting plasma glucose and (b) hemoglobin A1c among children of Shenzhen aged 7–9 years. BMI, body mass index; BMI‐z, z‐score of BMI‐for‐age; CC, chest circumference; CHpR, chest‐to‐hip ratio; C‐Index, Conicity Index; HC, hip circumference; WC, waist circumference; WCtR, waist‐to‐chest ratio; WHpR, waist‐to‐hip ratio; WHtR, waist‐to‐height ratio.
Table 4.
Optimal cut‐off points and the probability of significant anthropometric parameters for predicting elevated fasting plasma glucose (≥5.6 mmol/L) and hemoglobin A1c (≥5.7%) separately among children of Shenzhen aged 7–9 years in 2019
| Parameters | Cutoff | Sensitivity (%) | Specificity (%) | Youden index (%) |
|---|---|---|---|---|
| Prediction of elevated FPG | ||||
| Height | 132 cm | 67.2 | 60.0 | 27.2 |
| Weight | 26.2 kg | 72.1 | 53.4 | 25.6 |
| WC | 55 cm | 72.1 | 53.4 | 25.5 |
| HC | 69 cm | 45.9 | 74.1 | 20.0 |
| CC | 61 cm | 62.3 | 61.0 | 23.3 |
| BMI | 15.15 kg/m2 | 75.4 | 47.5 | 22.9 |
| BMI‐z | −0.45 | 77.0 | 46.0 | 23.1 |
| WHtR | 0.417 | 70.5 | 49.8 | 20.2 |
| WHpR | 0.800 | 86.9 | 27.7 | 14.6 |
| WCtR | 0.889 | 83.6 | 30.8 | 14.4 |
| Prediction of elevated HbA1c | ||||
| WHtR | 0.438 | 38.1 | 70.6 | 8.7 |
| WHpR | 0.865 | 29.9 | 77.2 | 7.1 |
| WCtR | 0.932 | 44.0 | 63.4 | 7.4 |
| C‐Index | 1.134 | 47.0 | 62.7 | 9.7 |
BMI, body mass index; BMI‐z, z‐score of BMI‐for‐age; CC, chest circumference; CHpR, chest‐to‐hip ratio; C‐Index, Conicity index; FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; HC, hip circumference; WC, waist circumference; WCtR, waist‐to‐chest ratio; WHpR, waist‐to‐hip ratio; WHtR, waist‐to‐height ratio.
Spearman’s rank correlation analyses
Table 5 presents the correlation coefficients of anthropometric parameters with FPG and HbA1c. Individual FPG levels were positively correlated with height, weight, WC, HC, CC, BMI and BMI‐z indices (r = 0.140–0.179, all P < 0.001). In contrast, the correlations of HbA1c levels with anthropometric indices were very weak (r ≤ 0.082).
Table 5.
Spearman’s rank correlation coefficients of anthropometric parameters with FPG and HbA1c among children of Shenzhen aged 7–9 years in 2019
| Parameters | FPG | HbA1c | ||
|---|---|---|---|---|
| r | P | r | P | |
| Age | 0.028 | 0.040 | −0.003 | 0.832 |
| Sex | 0.199 | <0.001 | 0.020 | 0.136 |
| Height | 0.140 | <0.001 | −0.002 | 0.869 |
| Weight | 0.179 | <0.001 | 0.017 | 0.197 |
| WC | 0.156 | <0.001 | 0.058 | <0.001 |
| HC | 0.152 | <0.001 | 0.029 | 0.031 |
| CC | 0.154 | <0.001 | 0.045 | 0.001 |
| BMI | 0.153 | <0.001 | 0.027 | 0.048 |
| BMI‐z | 0.146 | <0.001 | 0.028 | 0.038 |
| WHtR | 0.106 | <0.001 | 0.068 | <0.001 |
| WHpR | 0.067 | <0.001 | 0.069 | <0.001 |
| WCtR | 0.069 | <0.001 | 0.046 | 0.001 |
| CHpR | −0.006 | 0.657 | 0.030 | 0.023 |
| C‐Index | 0.059 | <0.001 | 0.082 | <0.001 |
The bold indicates a statistically significant association. BMI, body mass index; BMI‐z, z‐score of BMI‐for‐age; CC, chest circumference; CHpR, chest‐to‐hip ratio; C‐Index, Conicity index; FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; HC, hip circumference; WC, waist circumference; WCtR, waist‐to‐chest ratio; WHpR, waist‐to‐hip ratio; WHtR, waist‐to‐height ratio.
Sensitivity analyses
Sensitivity analyses were carried out by altering the definitions of elevated FPG levels as ≥4.8 mmol/L and of elevated HbA1c levels as ≥5.5%. All associations were similar, but with poorer diagnostic efficiencies or null ability in detecting abnormal FPG or HbA1c (Tables S1–S2). The AUCs of all anthropometric parameters and individual characteristics as predictive factors were ≤0.590 for discerning elevated FPG and ≤0.541 for elevated HbA1c.
Discussion
The present large‐scale, school‐based, cross‐sectional survey provided detailed evidence for the diagnostic performance of anthropometric parameters in diagnosing elevated serum FPG and HbA1c concentrations among 5,556 young children of Shenzhen. Boys and individuals with elevated blood glucose had greater anthropometric parameters than girls and those with blood glucose, respectively. The results of the present data also disclosed that most anthropometric indices had poor diagnostic efficiencies or null ability in detecting elevated FPG and HbA1c for children aged 7–9 years.
A large body of evidence suggested that the excess risks of mortality, diabetes, cardiovascular and cerebrovascular morbidity were associated with increased HbA1c or FPG concentrations 5 , 24 , 25 , 26 . For instance, 10,232 Norfolk residents aged 45–79 years were followed up for 8 years, and the population‐based prospective cohort study showed that per 1‐percentage point increase of HbA1c was linked to a 24% (14–34%) increase in men and 28% (6–32%) in women for total mortality, 21% (13–29%) in men and 21% (11–31%) in women for cardiovascular disease, and 25% (16–34%) in men and 20% (7–34%) in women for the occurrence of coronary heart disease events 6 . Including 103 prospective cohort studies, a review on assessing the overall prognosis of individuals with intermediate hyperglycaemia for developing type 2 diabetes from the Cochrane database of systematic reviews also showed that the pooled hazard rate was 4.32 (95% CI 2.61–7.12) for elevated FPG (5.6–6.9 mmol/L), and 5.55 (95% CI 2.77–11.12) for elevated HbA1c (5.7–6.4%), as compared with normoglycaemia 2 . A previous survey from Shenzhen 27 pointed out that the crude prevalence was 1.57% (95% CI 1.21－1.94%) for elevated FPG of 5.6–6.9 mmol/L, and 8.89% (95% CI 8.08－9.70%) for elevated HbA1c of 5.7–6.4% among Shenzhen children aged 6–10 years. Taking into consideration the available evidence, we strongly suggest that practitioners and policymakers should be careful about the potential implications of any active intervention for individuals with elevated blood glucose concentrations, not only for adults, but also for children and adolescents.
Clearly, it is more convenient to collect anthropometric indices than biochemical indices to screen for junior grade school children in some investigations or clinical diagnoses, due to the low cost, simplicity, speed and convenience of the non‐invasive anthropometry. It is not clear which indices are the highest related to elevated blood glucose among children. Involving 526 Indian students aged 17–19 years, a previous cross‐sectional study 17 confirmed that the ROC analysis for BMI showed a good predictive power for an FPG of 5.6–6.9 mmol/L for both boys (AUC 0.828) and girls (AUC 0.838). The approximate efficiency for WC was also shown for both boys (AUC 0.804) and girls (AUC 0.795); the best cut‐offs were 22.8 kg/m2 for boys and 20.5 kg/m2 for girls for BMI, and 82.5 cm for boys and 80.3 cm for girls for WC. On the contrary, an ongoing cohort study followed up from 2003 to 2013 in central Iran suggested that neither the value of BMI (AUC 0.541) nor WC (AUC = 0.552) was a significant risk factor for prediabetes among Isfahan adults 14 . The previous study based on children and adolescents of Shenzhen also showed that the association of BMI and prediabetes was non‐linear, and followed a U‐shaped curve, indicating higher rates of prediabetes for underweight and obese people 27 . Overweight and obesity are highly related to type 2 diabetes in youth, and Asian people might have weaker insulin secretion compared with Western people 28 , 29 . In addition, the occurrence of diabetes in children and adolescents might relate to gestational diabetes risk of mothers, and is sometimes observed at the beginning of puberty; the occurrence of diabetes in girls might 1–2 years earlier than that in boys 30 . During puberty, glucose intolerance might be apparent as a result of the secretion of sex hormones, and the impact of puberty on glucose tolerance should be different between boys and girls, as some girls, but only few boys, start puberty at the age of 7–9 years. Thus, we highly recommended that several larger‐scale surveys involved with some factors (e.g., menstrual onset and Tanner classification of stages G1–G5 for young girls) related to blood glucose levels or anthropocentric indices should be investigated among children and adolescents.
With the advantage of not being dependent on sex‐ or age‐specific percentiles relative to a reference population, indices, such as WHtR, BMI‐z and C‐Index, could be used in adolescents and children, and might represent more practical tools to evaluate abdominal obesity or hyperglycemia, which might have relevant public health implications. However, the findings based on the young children of Shenzhen did not confirm the implication – the highest AUC to detect elevated FPG was just 0.651 (0.583–0.719) for weight, followed by WC, BMI and BMI‐z, and lower diagnostic efficiencies for C‐Index, WHtR, WHpR and WCtR to identify elevated HbA1c. For young children, especially for students aged 7–9 years in China, we do not recommend anthropometric indicators to screen for elevated FPG or HbA1c level, because of the poor diagnostic performance.
The current study first investigated the utility of the C‐Index as a predictive index of intermediate hyperglycemia for children, showing a low diagnostic performance of C‐Index for detecting an elevated HbA1c activity and a null discriminative power for elevated FPG. With the ability to provide estimates of centripetal concentration of body fat, the C‐Index had been proposed to assess cardiometabolic risk in a pediatric population, although its use was still limited to research purposes without a widespread use in the clinical and population settings. Among adolescents aged 12–15 years attending school in Jacarezinho, Brazil, the C‐Index showed a significant low capacity to predict metabolic syndrome (e.g., high FPG, altered blood pressure), with an AUC of 0.64 (95% CI 0.60–0.69) for boys and 0.68 (95% CI 0.64–0.72) for girls 31 .
A population‐based study carried out in a southwest rural region of China by Zhao et al. 11 in 2010 showed that only WHpR (r = 0.151, P < 0.001) and WHtR (r = 0.153, P < 0.001), but not BMI (r = 0.015, P = 0.622), had significant correlations with the concentrations of FPG, controlling for WC in the partial correlation analysis; WHpR and WHtR are common indices of central obesity, which is closely linked to insulin resistance and diabetes. However, the levels of FPG in the present study were positively correlated with BMI (r = 0.153), whereas weakly linked to WHtR (r = 0.106) and WHpR (r = 0.067), all P < 0.001. In terms of HbA1c, the correlation coefficients of all anthropometric indices became weaker.
The present study was the first, as far as we know, to examine the diagnostic performance of 14 anthropometric parameters for the detection of an elevated FPG or HbA1c activity, carried out with 5,556 Shenzhen children aged 7–9 years. Several shortcomings of the present study should be considered. First, sampling based on schools might weaken the representativeness of the population or children, but the school attendance rate in the prosperous city of Shenzhen appears to be nearly 100%. Second, the age range of 7–9 years was narrow, and the results based on a cross‐sectional design from a single city of China should be carefully generalized to other regions and other age groups of children. Among children aged 7–9 years in Shenzhen, the median was 130.0 cm (IQR 126.5–134.0 cm) for height, 25.8 kg (23.2–29.5 kg) for weight and 15.27 kg/m2 (14.21–16.79 kg/m2) for BMI. In an international norm from the 2007 World Health Organization growth reference for school‐aged children, the median (−1 standard deviation to standard deviation) of BMI‐for‐age of 8‐year‐olds (96 months) was 15.7 kg/m2 (14.4–17.4 kg/m2) for boys, and 15.7 kg/m2 (14.1–17.7 kg/m2) for girls 21 . BMI values of our Shenzhen students were similar to those of the World Health Organization, and our sample might show representativeness of these age groups in the world population. Third, the present study did not take into account some potential confounders of hyperglycemia for school children (e.g., leisure time physical activity, sedentary behavior, dietary, and family socioeconomic position and living settings). Another limitation was our inability to consider the 2‐h postprandial oral glucose tolerance test among a large sample of children for the definition of hyperglycemia, which might lead to a bias in assessing an elevated blood sugar activity.
In conclusion, the present study showed that all of the checked anthropometric indices did not approach adequate predictive accuracies for the detection of elevated FPG or HbA1c levels in a group of Shenzhen children aged 7–9 years. We do not recommend anthropometric indicators to screen for prediabetes in young children, because of the poor diagnostic performance.
Disclosure
The authors declare no conflict of interest.
Supporting information
Table S1 | Baseline descriptive statistics for anthropometric parameters stratified by fasting plasma glucose (elevated fasting plasma glucose ≥4.8 mmol/L) and hemoglobin A1c (elevated hemoglobin A1c ≥5.5%) levels among children aged 7－9 years of Shenzhen, 2019.
Table S2 | Area under the curve and 95% confidence interval of the receiver operating characteristic curve to judge the discrimination ability of various anthropometric parameters for detecting elevated fasting plasma glucose (≥4.8 mmol/L) and hemoglobin A1c (≥5.5%) separately.
Acknowledgment
The study was supported by the Metabolic Disease in Prevention and Treatment Program of Occupational Groups and Adolescents (2019–2020), under the Bao’an Government of Shenzhen.
J Diabetes Investig 2021; 12: 781–789
References
- 1. Schmidt MI, Bracco PA, Yudkin JS, et al. Intermediate hyperglycaemia to predict progression to type 2 diabetes (ELSA‐Brasil): an occupational cohort study in Brazil. Lancet Diabetes Endocrinol 2019; 7: 267–277. [DOI] [PubMed] [Google Scholar]
- 2. Richter B, Hemmingsen B, Metzendorf MI, et al. Development of type 2 diabetes mellitus in people with intermediate hyperglycaemia. Cochrane Database Syst Rev 2018; 10: Cd012661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Shahim B, De Bacquer D, De Backer G, et al. The prognostic value of fasting plasma glucose, two‐hour postload glucose, and HbA1c in patients with coronary artery disease: a report from EUROASPIRE IV: a survey from the European Society of Cardiology. Diabetes Care 2017; 40: 1233–1240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Lu J, He J, Li M, et al. Predictive value of fasting glucose, postload glucose, and hemoglobin A1c on risk of diabetes and complications in Chinese adults. Diabetes Care 2019; 42: 1539–1548. [DOI] [PubMed] [Google Scholar]
- 5. Menke A, Rust KF, Cowie CC. Diabetes based on 2‐h plasma glucose among those classified as having prediabetes based on fasting plasma glucose or A1c. Diab Vasc Dis Res 2018; 15: 46–54. [DOI] [PubMed] [Google Scholar]
- 6. Khaw KT, Wareham N, Bingham S, et al. Association of hemoglobin A1c with cardiovascular disease and mortality in adults: the European prospective investigation into cancer in Norfolk. Ann Intern Med 2004; 141: 413–420. [DOI] [PubMed] [Google Scholar]
- 7. Kreider KE, Feinglos M. The relationship of baseline fasting plasma glucose and cardiovascular morbidity and mortality. Eur J Intern Med 2018; 54: e25–e26. [DOI] [PubMed] [Google Scholar]
- 8. Geva M, Shlomai G, Berkovich A, et al. The association between fasting plasma glucose and glycated hemoglobin in the prediabetes range and future development of hypertension. Cardiovasc Diabetol 2019; 18: 53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Heianza Y, Arase Y, Kodama S, et al. Fasting glucose and HbA1c levels as risk factors for the development of hypertension in Japanese individuals: Toranomon hospital health management center study 16 (TOPICS 16). J Hum Hypertens 2015; 29: 254–259. [DOI] [PubMed] [Google Scholar]
- 10. Talaei M, Sadeghi M, Marshall T, et al. Anthropometric indices predicting incident type 2 diabetes in an Iranian population: the Isfahan Cohort Study. Diabetes Metab 2013; 39: 424–431. [DOI] [PubMed] [Google Scholar]
- 11. Zhao X, Zhu X, Zhang H, et al. Prevalence of diabetes and predictions of its risks using anthropometric measures in southwest rural areas of China. BMC Public Health 2012; 12: 821. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Stevens J, Couper D, Pankow J, et al. Sensitivity and specificity of anthropometrics for the prediction of diabetes in a biracial cohort. Obes Res 2001; 9: 696–705. [DOI] [PubMed] [Google Scholar]
- 13. Lee BJ, Ku B, Nam J, et al. Prediction of fasting plasma glucose status using anthropometric measures for diagnosing type 2 diabetes. IEEE J Biomed Health Inform 2014; 18: 555–561. [DOI] [PubMed] [Google Scholar]
- 14. Haghighatdoost F, Amini M, Feizi A, et al. Are body mass index and waist circumference significant predictors of diabetes and prediabetes risk: results from a population based cohort study. World J Diabetes 2017; 8: 365–373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Carba DB, Bas IN, Gultiano SA, et al. Waist circumference and the risk of hypertension and prediabetes among Filipino women. Eur J Nutr 2013; 52: 825–832. [DOI] [PubMed] [Google Scholar]
- 16. Prakaschandra R, Naidoo DP. Fasting plasma glucose and the HbA1c are not optimal screening modalities for the diagnosis of new diabetes in previously undiagnosed Asian Indian community participants. Ethn Dis 2018; 28: 19–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Pandey U, Midha T, Rao YK, et al. Anthropometric indicators as predictor of pre‐diabetes in Indian adolescents. Indian Heart J 2017; 69: 474–479. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Zhan ZY, Yu YM, Qian J, et al. Effects of ambient temperature on ambulance emergency call‐outs in the subtropical city of Shenzhen, China. PLoS One 2018; 13: e0207187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Chen D, Zhang F, Yu C, et al. Hourly associations between exposure to ambient particulate matter and emergency department visits in an urban population of Shenzhen, China. Atmos Environ 2019; 209: 78–85. [Google Scholar]
- 20. Ding Z, Zhang J, Deng C‐Y, et al. Association of body mass index with serum alanine aminotransferase in Chinese adolescents: a school‐based cross‐sectional study. Gastroenterol Rep 2020; 8: 119–124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. de Onis M, Onyango AW, Borghi E, et al. Development of a WHO growth reference for school‐aged children and adolescents. Bull World Health Organ 2007; 85: 660–667. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Obuchowski NA, Bullen JA. Receiver operating characteristic (ROC) curves: review of methods with applications in diagnostic medicine. Phys Med Biol 2018; 63: 07tr01. [DOI] [PubMed] [Google Scholar]
- 23. Asif M, Aslam M, Altaf S. Evaluation of anthropometric parameters of central obesity in Pakistani children aged 5–12 years, using receiver operating characteristic (ROC) analysis. J Pediatr Endocrinol Metab 2018; 31: 971–977. [DOI] [PubMed] [Google Scholar]
- 24. Nakagami T, Tanaka Y, Oya J, et al. Associations of HbA1c and fasting plasma glucose with incident diabetes: Implications for pre‐diabetes thresholds in a Japanese population. Prim Care Diabetes 2016; 10: 407–414. [DOI] [PubMed] [Google Scholar]
- 25. Lv Y, Yao Y, Ye J, et al. Association of blood pressure with fasting blood glucose levels in Northeast China: a cross‐sectional study. Sci Rep 2018; 8: 7917. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Vijayakumar P, Nelson RG, Hanson RL, et al. HbA1c and the prediction of type 2 diabetes in children and adults. Diabetes Care 2017; 40: 16–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Yang C, Ding Z, Zhou H, et al. Prevalence of prediabetes by the fasting plasma glucose and HbA1c screening criteria among the children and adolescents of Shenzhen, China. J Diabetes 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Davis CL, Pollock NK, Waller JL, et al. Exercise dose and diabetes risk in overweight and obese children: a randomized controlled trial. JAMA 2012; 308: 1103–1112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Association AD: Type 2 diabetes in children and adolescents. Diabetes Care. 23. 2000; 381–389. [DOI] [PubMed] [Google Scholar]
- 30. Leppänen MH, Raitanen J, Husu P, et al. Physical activity and body composition in children and their mothers according to mother’s gestational diabetes risk: a seven‐year follow‐up study, Medicina 2019; 55: 635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Oliveira RG, Guedes DP. Performance of anthropometric indicators as predictors of metabolic syndrome in Brazilian adolescents. BMC Pediatr 2018; 18: 33. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1 | Baseline descriptive statistics for anthropometric parameters stratified by fasting plasma glucose (elevated fasting plasma glucose ≥4.8 mmol/L) and hemoglobin A1c (elevated hemoglobin A1c ≥5.5%) levels among children aged 7－9 years of Shenzhen, 2019.
Table S2 | Area under the curve and 95% confidence interval of the receiver operating characteristic curve to judge the discrimination ability of various anthropometric parameters for detecting elevated fasting plasma glucose (≥4.8 mmol/L) and hemoglobin A1c (≥5.5%) separately.