Received: 24 March 2020 | Revised: 18 April 2020 | Accepted: 27 April 2020

DOI: 10.1111/jch.13895

ORIG INAL PAPER

Validity of anthropometric indices in predicting high blood

pressure risk factors in Iranian children and adolescents:
CASPIAN-V study

Mayram Yazdi PhD1 | Farahnak Assadi MD2 | Mostafa Qorbani PhD3,4 |

Seyede Shahbanoo Daniali PhD1 | Ramin Heshmat MD, PhD5 |
Mohammad Esmaeil Motlagh MD6 | Roya Kelishadi MD1
1
Child Growth and Development Research Center, Research Institute for Primordial Prevention of Non-Communicable Disease, Isfahan University of Medical
Sciences, Isfahan, Iran
2
Department of Pediatrics, Rush University Medical College, Chicago, IL, USA
3
Non-Communicable Diseases Research Center, Alborz University of Medical Sciences, Karaj, Iran
4
Chronic Research Center, Epidemiology and Metabolism Population Sciences Institute, Tehran University of Medical Sciences, Tehran, Iran
5
Chronic Diseases Research Center, Endocrinology and Metabolism Population Sciences Institute, Tehran University of Medical Sciences, Tehran, Iran
6
Department of Pediatrics, Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran

Correspondence
Farahnak Assadi, MD, Department of Abstract
Pediatrics, Rush University Medical College, Anthropometric indices have been used as indicators for predicting hypertension
445 E. North Water Street, Suite 1804,
Chicago, IL 60611. (HTN) in children and adolescents but it is not clear which anthropometric measures
Email: farahnak_assadi@[Link] are a better index for identifying elevated blood pressure (EBP) risk factors in pedi-
Roya Kelishadi,MD, Child Growth and atric population. Body mass index (BMI), waist circumference (WC), weight-height
Development Research Center, Research
ratio (WHR), a body shape index (ABSI) and blood pressure were measured in 14 008
Institute for Primordial Prevention of Non-
Communicable Disease, Hezarjerib Ave, children and adolescents aged 7-18 years in a national school-aged survey CASPIN
Isfahan, Iran
V. Hypertension (HTN) was defined according to the 2017 American Academy of
Email: Kelishadi@[Link]
Pediatrics guidelines, using the 95th percentile. The predictive power of anthropo-
Funding information
metric indices for HTN risk factors was examined using receiver operating charac-
This study was conducted as part
of a national surveillance program teristic (ROC) analyses. Multivariate logistic regression analysis was used to compare
in collaboration of the Ministry of
areas under ROC curves (AUCs) among the four anthropometric indices. BMI, WC,
Education and Training, Child Growth
and Development Research Center, WHR, and ABSI were significantly higher in adolescents than in children. EBP was
Isfahan University of Medical Sciences,
more prevalent in boys (7.2%) than girls (5.5%), whereas the prevalence of HTN was
Endocrinology and Metabolism Research
Institute, and Chronic Diseases Research higher in girls (11.3%) than boys 10.4%. Prevalence odds ratio was around 2 for BMI,
Center, Tehran University of Medical
WC, and WHR with AUCs scores of nearly 0.6 to predict EBP in both children and
Sciences.
adolescents of both sexes. Thus, the ability of BMI z-score, WC, WHR or ASBI to
identify Iranian children and adolescents at higher risk of EBP was week. WC, WHR
or ASBI in combination with BMI did not improve predictive power to identify sub-
jects at higher risk of EBP.

J Clin Hypertens. 2020;22:1009–1017. [Link]/journal/jch© 2020 Wiley Periodicals LLC | 1009

1010 | YAZDI et al.

1 | I NTRO D U C TI O N of the World Health Organization—Global School-based Student

Health Survey (WHO-GSHS) protocol. The methodology used for
Childhood obesity is an important risk factor for the development the survey is described elsewhere. 28
of hypertension (HTN) chronic kidney disease and cardiovascular Briefly, the study population consisted of healthy-looking school
diseases (CVD) during adulthood.1-4 Thus, early identification of students aged 7-18 years in urban and rural areas across 30 prov-
children at risk of HTN is essential to improve clinical outcome and inces of the country. They were selected using a multi-stage, strati-
prevent CVD later in life. fied sampling approach. Random sampling within each province was
Body mass index (BMI), waist circumference (WC), waist-to- carried out in proportion to the size of the population in urban or
height ratio (WHR), and waist-to-hip ratio are the most commonly rural areas and the school level (elementary, middle, and secondary).
5
used anthropometric indices for defining obesity-related HTN. A comprehensive verbal description of the nature and purpose of
Previous studies have found that BMI and WC were good indi- the study was given to the students and their parents and written
cators of predicting HTN in adults and adolescents. 6,7 However, informed consent was obtained from parents and oral assent from
findings from subsequent studies showed BMI may not accurately students. In accordance with standard protocols, a team of qualified
reflect visceral fat and fat distribution and is not a reliable indi- health professionals and physicians recorded data and performed a
cator in predicting HTN risk among children and adolescents and physical examination using standardized instruments. The Research
could underestimate the risk of HTN in short people or overesti- Ethics Committees of Isfahan University of Medical Sciences re-
mate the HTN risk tall subjects, since height is also a risk factor viewed and approved research protocols (Project Registration
8-13
for HTN. ISUMS 194049).
These inconsistent findings encouraged the investigators to
develop new anthropometric measures such as WHR and a body
shape index (ABSI) for the early prediction of HTN.14-19 Several 2.2 | Blood pressure measurement
studies from Asian countries found WHR is superior in predict-
ing HBP than BMI and WC among adolescents.13 However, a few Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were
studies from European countries showed conflicting results about measured in the right arm with a standardized mercury sphygmoma-
the usefulness of WHR in predicting HTN in both children and nometers using a stethoscope placed over the brachial artery pulse
14,15
adolescents. on the cubital fossa at heart level and appropriate sized cuff with an
There are also conflicting reports on the performance of ABSI inflammable bladder width of at least 40 percent of the arm circum-
as indictor of predicting HTN risk in both children and adults pa- ference at a point midway between the olecranon and the acromion
tients.19-27 ABSI, which is based on WC adjusted for height and with the child in a sitting position for at least 5 minutes rest.29 SBP
16,18,19,27
BMI, was first found to provide greater predictability of was determined by the onset of the appearance of Korotokoff sounds
mortality risk than WC and BMI.16,19,20 However, subsequent stud- (K1) and DBP by the disappearance of Korotokoff sounds (K5). Two
ies showed poor prediction of ABSI in identifying HTN and CVD risk measurements were recorded at 2-minute intervals, and this averaged
among adolescents and adult patients.17,20-22,24,25 In a more recent measurement was used for in the present study. The 2017 American
26
study, Tee et al compared BMI, MC, WHR, and ABSI and found Academy of Pediatrics (AAP) Clinical Practice Guideline for Screening
WHR provided a better prediction for screening HTN risk in children and Management of High Blood Pressure in Children and Adolescents
and adolescents. Thus, it remains unclear which of anthropometric was used to diagnose HTN.29 Normal blood pressure levels were
indices are the best predictor of hypertension in the young; more- defined as SBP and/or DBP ≤ 90th percentile, elevated blood pres-
over, ethnic difference should be considered in the interpretation of sure ((EBP) as SBP and/or DBP > 90th and ≤ 95th percentile or 120-
such associations. 129/<80 mm Hg (whichever was lower) and HTN as SBP and/or DBP
The aim of present study was to examine the associations be- 95th percentile or ≥ 130/89 mm Hg (whichever was lower).29
tween ABSI, BMI, WC, WCHR, and hypertension and to determine
which of these anthropometric indices are the best predictor of HTN
among Iranian children and adolescents. 2.3 | Anthropometric measurements

Weight and height were measured to the nearest 0.1 kg and 0.5 cm,
2 | M E TH O DS respectively, with participant in light clothing and without shoes.
WC was measured at a level midway between the lower rib margin
2.1 | Study population and the iliac crest to the nearest 0.5 cm with a flexible measuring
tape and the participants in a standing position. A WC >90th per-
This study was conducted as part of the fifth survey of a National centile was used as the cutoff point to define abdominal obesity.30
school-based project entitled Childhood and Adolescence WHR was calculated by dividing WC (cm) by height (cm), and a WHR
Surveillance and Prevention of Adult Non-Communicable Disease >0.50 was defined as abdominal obesity.31,32 The body mass index
(CASPIAN-IV). The study was conducted in 2015 in Iran on the basis (BMI) was calculated by dividing the weight (kg) to the square of
YAZDI et al. | 1011

TA B L E 1 Anthropometric indices and blood pressure values in the studied population

Girl Boy P 7-12 year 13-18 year P Total

Number 6917 7091 7388 6620 14 008

BMI z-score −0.177 ± 1.410 −0.211 ± 1.544 .182 −0.236 ± 1.502 −0.147 ± 1.452 <.001 −0.194 ± 1.479
Underweight 1016 (14.8) 1233 (17.5) 1273 (17.3) 976 (14.9) 2249 (16.2)
Normal weight 4443 (64.7) 4317 (61.3) 4555 (62.0) 4205 (64.0) 8760 (63.0)
Overweight 697 (10.2) 609 (8.6) 646 (8.8) 660 (10.0) 1306 (9.4)
Obese 710 (10.3) 885 (12.6) 867 (11.8) 728 (11.1) 1595 (11.5)
WC 65.881 ± 11.209 67.695 ± 12.879 .041 61.686 ± 10.136 72.505 ± 11.588 .030 66.799 ± 12.117
Normal 6176 (89.3) 6434 (90.7) 6728 (91.1) 5882 (88.9) 12 610 (90.0)
Abdominal 741 (10.7) 657 (9.3) 660 (8.9) 738 (11.1) 1398 (10.0)
obesity
WHR 0.455 ± 0.062 0.457 ± 0.067 <.001 0.457 ± 0.064 0.455 ± 0.064 <.001 0.456 ± 0.064
Normal 5447 (79.4) 5516 (78.3) 5816 (79.3) 5147 (78.4) 10 963 (78.9)
Abdominal 1412 (20.6) 1526 (21.7) 1519 (20.7) 1419 (21.6) 2938 (21.1)
obesity
ABSI 0.079 ± 0.010 0.080 ± 0.010 <.001 0.082 ± 0.011 0.078 ± 0.009 <.001 0.080 ± 0.010
BP classification
Normal (<90th 5754 (83.2) 5839 (82.3) 6096 (82.5) 5497 (83.0) 11 593 (82.8)
percentile)
EBP (>90th 379 (5.5) 513 (7.2) 415 (5.6) 477 (7.2) 892 (6.4)
and <95th
percentile)
HTN (>95th 784 (11.3) 739 (10.4) 877 (11.9) 646 (9.8) 1523 (10.9)
percentile)

Abbreviations: ABSI, a body shape index; BMI, body mass index; BP, blood pressure; EBP, elevated blood pressure; HTN, hypertension; WC, waist
circumference; WHR, waist-to-height ratio.

the height in meters and then graded on the basis of WHO growth area under the curve (AUC) was used to compare the predictive power
charts.32 Childhood overweight and obesity were defined as BMIs of different anthropometric indices for HBP. An AUC between 0.50
between the 85th and 95th percentile and ≥95th percentile by age and 0.70 was considered as poor with no discriminatory ability. An AUC
and sex groups, respectively.33,34 ABSI was calculated using the fol- between 0.70 and 0.80 regarded as acceptable, 0.80 to 0.90 viewed as
lowing formula ABSI = WC/(BMI2/3height1/2) with WC and height good, and >0.90 indicated as excellent performance.39 Cutoff values of
in meter.35 anthropometric indices to predict HTN were estimated on the highest
Since normal values for anthropometric indices vary considerably value of the Youden Index (specificity + sensitivity-1).40
by age and also by sex, we converted the individual BMI, WC, WHR, To examine the correlation of different levels of BMI, WC,
and ABSI indices into z-scores relative to age-and sex-specific means WCH, and ABSI with HTN risk factors, we stratified the studied
and SDs to simplify clinical interpretation for anthropometric mea- population into quartiles where second quartile included those
sures using to the National Health and Nutrition Examination Survey with mean z-score values. Logistic regression analysis was per-
(NHANES) for ages 3 and older.36 Conventionally, z-scores are classi- formed separately for boys and girls to examine associations
fied into five categories as poor (<−1.5), below normal (−1.5 to −0.5), between the four anthropometric indices with EBP and HTN. A
normal (−0.5 to 0.5), above normal (0.5-1.5), and much above normal two-tailed statistical measure was used with a p value of less than
37
(>1.5). 0.05 was considered significant. All statistical analyses were con-
ducted using the R software [Link]

2.4 | Statistical analysis

3 | R E S U LT S
The general characteristics of the participants including WC, BMI,
CHR, and ABSI were analyzed according to age- and sex-specific z- Data of 14 008 participants (49.37% females) ages between 7 to
scores. Categorical variables were expressed as numbers and percent- 12 years (52.74%) and 13 to 18 years (47.26%) were used in the cur-
age. Descriptive characteristics are presented as mean (SD). Receiver rent study. Table 1 summarizes the anthropometric characteristics
operative characteristic (ROC) curve and 95% confidence intervals (CIs) and BP measurements of all participants by age and sex. In the total
analyses were used to compare the predictive power of different an- population, the prevalence of obesity based on BMI was 11.5%; Boys
thropometric indices with EBP and HTN as previously described.38 The were found to be more obese (11.8%) than boys (10.3%) (P < .001).
1012 | YAZDI et al.

BMI, WC, WHR, and ABSI were significantly higher in adolescents In children (7-12 years), AUC of BMI (0.606) had the highest predic-
than in children. Prevalence of abdominal obesity based on WC tive power in predicting HTN, whereas, in adolescents (13-18 years),
index was higher in girls (10.7%) than boys (9.7%) (P < .03) and the WHR (0.603) showed the highest AUC for identifying HTN. Similar
prevalence of abdominal obesity based on WHR was greater in boys superiorities were also observed in girls and boys of both age groups.
(22.9%) than girls (20.6%) (P < .001).
EBP was more prevalent in boys than girls (5.5% vs 7.2%),
whereas, HTN was more prevalent in girls (11.5% vs 10.7%, re- 4 | D I S CU S S I O N
spectively). The anthropometric characteristics and blood pressure
measurements by sex and age groups (7-12 and 13-18 years) are pro- In this large nationwide study, we compared the predictive capabil-
vided in Table S1. ity of BMI z-score, WC, WCH, and ABSI indices for the early detec-
In the total population, BMI and WC had the largest AUC, al- tion of EBP and HTN in Iranian children and adolescents and found
though differences were small with overlapping 95% CI. ABSI had prevalence odds ratio around 2 with a nearly similar predictive
the smallest AUC in both EBP (AUC: 0.481, 95% CI: 0.462, 9.500) power for all markers except ABSI. WC, WHR, or ABSI in combi-
and HTN (AUC: 0.516, 95% CI: 0.500, 0.531). nation with BMI did not improve the predictive power to identify
We found the risk of HTN (BP ≥ 95th% percentile) increased with subjects with EBP.
increasing quartile of BMI and WC (Table 2). The highest quartiles of Data from our ROC analyses showed the four anthropometric
WHR, WC, and BMI compared with the reference quartile (1st) were indices performed poor accuracy for predicting EBP and HTN.
associated with HTN in age and sex-adjusted models. As presented Previous studies that have evaluated the effectiveness of dif-
in Table 2, only the fourth quartile of ABSI in boys showed a signifi- ferent anthropometric indices on early HTN detection in children
cant association with increased risk of HTN ((OR: 1.31 95% CI: 1.06- and adolescents have provided discrepant results. Li et al42 in a pro-
1.63) (Table 2). The same results were obtained for the association spective cross-sectional study compared the efficacy of BMI, WC,
of high EBP (BP ≥90th and ≤95th percentiles) with anthropometric WHT, ABSI, and waist-hip ratio indices for the early detection of car-
indices, but no statistical significant existed between risk of HTN (BP diometabolic risk factors including blood pressure in 15 698 Chinese
≥95th percentile) and increasing quartile of ABSI. children and adolescents aged between 6 and 17 years. They found
The results of the ROC analysis of various anthropometric in- anthropometric indices performed poor accuracy for identifying the
dices in predicting EBP (90th and <95th percentiles) in children risk of HBP even in the overweight and obese children. The same
and adolescents are shown in Table 3. In the total population, the conclusion is reported by Lo et al43 in a meta-analysis for AUCs of
AUCs of BMI, WC,WHR, and ABSI were 0.571,0.553,0.560, and BMI, WC, and WHR for identifying cardiometabolic risk factors
0.481, respectively, in predicting EBP among children and ado- among children and adolescents. Unlike the above mentioned stud-
lescents of both sexes. In the total population, ABSI showed the ies, Li et al and Lo et al, Tee et al26 in a cross-sectional study, com-
highest sensitivity level of 100% in detecting HTN followed by BMI pared the predictive power of BMI, WC, WHR, and ABSI for HTN
of 77.7%. Compared with girls, boys had higher sensitivity level in 513 Malaysian adolescents using the 90th and 95th percentiles
of WC (83.4% vs 77.8%), WHR (81.6% vs 72.2%), and BMI (77.7% blood pressure thresholds and found WHR followed by BMI pro-
vs 53.2%). Girls had higher sensitivity level of the ABSI than boys vided better prediction for screening HTN risk. They concluded from
(97.3 vs 74.1). the findings of their study that BP screening for early detection of
Table 4 shows the results of the ROC analysis for various an- HBP should be based on the 90th percentile cutoff values to prevent
thropometric indices for different anthropometric indices in relation underestimation of HTN among adolescents.
to HTN (BP ≥ 95th percentile). In the total population, the AUCs In another study, Ma et al reported that WC is a better index
of BMI (0.592%), WC (0.586), WHR (0.588), and ABSI (0.516) were than other anthropometric indices for predicting cardiovascular risk
nearly around 0.6 in detecting HTN (BP ≥ 95th percentile) in chil- factors in Chinese children.44 A few studies have shown that the
dren and adolescents of both sexes. In the total population, BMI magnitude of the associations between anthropometric indices and
showed the highest sensitivity level of 56.1% in predicting HTN in cardiovascular risk factors was greater in overweight and obese chil-
girls. ABSI showed the highest sensitivity level of 80.3% followed by dren and adolescent.45–48 The findings in our study are consistent
BMI (52.1%). In boys, BMI showed a sensitivity level of 54.1 followed with those reported by LI et al and Lo et al.42,43,49
by WC 50.1 for prediction of HTN. The optimal BMI cutoff points to The discrepancy between the findings in our study and the study
discriminate HTN were 68.5 in girls and 60.5 in boys, whereas WHR reported by others are likely the result of diversity in the study de-
cutoff points were 0.477 and 0.469 in girls and boys, respectively. sign and methodologies used by different investigators. The differ-
ABSI cutoff values to identify HTN were 0.477 in girls and 0.084 in ences in the ethnicity, age, and sex, lifestyle, genes, environmental
boys. Table 5 compares different combinations of 2 adiposity mark- influences, and socioeconomic status of the study participants en-
ers together for the prediction of EBP. The Pearson correlation coef- rolled in the different study settings can introduce bias that may af-
ficients among the 4 adiposity biomarkers are shown in Table 6. The fect the study findings. 24,47–50
ROC analyses for computing predictive power of anthropometric The major limitation of this study is the cross-sectional nature of
indices by sex and age groups are provided in the Tables S2 and S3. data. Moreover, we could not obtain the data about pubertal stage
 YAZDI et al.

TA B L E 2 Crude and adjusted odds ratios and 95% confidence intervals for HBP in quartiles of anthropometric parameters (BMI, WC, WHR, ABSI) by sex

Hypertension Elevated blood pressure

Girl Boy Girl Boy

OR(95% CI) aOR(95% CI) OR(95% CI) aOR(95% CI) OR(95% CI) aOR(95% CI) OR(95% CI) aOR(95% CI)

Quartiles of BMI
1th 1 1 1 1 1 1 1 1
2th 1.31 (1.03,1.67) 1.32 (1.03,1.68) 1.18 (0.92,1.50) 1.18 (0.92,1.50) 1.44 (1.16,1.80) 1.44 (1.16,1.80) 1.32 (1.07,1.62) 1.32 (1.07,1.62)
3th 1.70 (1.35,2.14) 1.70 (1.35,2.15) 1.52 (1.21,1.92) 1.52 (1.21,1.92) 1.67 (1.35,2.07) 1.67 (1.35,2.07) 1.59 (1.30,1.94) 1.59 (1.30,1.94)
4th 2.49 (1.99,3.10) 2.49 (2.00,3.11) 2.01 (1.61,2.51) 2.01 (1.61,2.51) 2.54 (2.07,3.11) 2.54 (2.07,3.11) 2.01 (1.66,2.44) 2.01 (1.66,2.44)
Quartiles of WC
1th 1 1 1 1 1 1 1 1
2th 1.26 (1.00,1.59) 1.28 (1.02,1.61) 0.97 (0.76,1.23) 0.97 (0.76,1.23) 1.43 (1.15,1.77) 1.43 (1.15,1.77) 1.03 (0.84,1.25) 1.03 (0.84,1.25)
3th 1.45 (1.16,1.82) 1.45 (1.16,1.82) 1.47 (1.18,1.83) 1.47 (1.18,1.83) 1.58 (1.28,1.95) 1.58 (1.28,1.95) 1.42 (1.18,1.71) 1.42 (1.18,1.72)
4th 2.28 (1.85,2.82) 2.30 (1.86,2.84) 1.88 (1.52,2.32) 1.88 (1.52,2.33) 2.51 (2.06,3.06) 2.51 (2.06,3.06) 1.73 (1.44,2.08) 1.73 (1.44,2.08)
Quartiles of WHR
1th 1 1 1 1
2th 1.32 (1.05,1.66) 1.32 (1.05,1.66) 1.17 (0.92,1.49) 1.17 (0.92,1.49) 1.35 (1.11,1.64) 1.35 (1.11,1.64) 1.35 (1.12,1.63) 1.35 (1.12,1.63)
3th 1.34 (1.06,1.69) 1.34 (1.06,1.69) 1.56 (1.23,1.96) 1.56 (1.23,1.96) 1.45 (1.19,1.75) 1.45 (1.19,1.75) 1.67 (1.39,2.01) 1.67 (1.39,2.01)
4th 2.17 (1.75,2.68) 2.17 (1.75,2.69) 2.22 (1.78,2.77) 2.22 (1.78,2.77) 2.03 (1.69,2.43) 2.03 (1.69,2.44) 2.14 (1.79,2.56) 2.15 (1.79,2.57)
Quartiles of ABSI
1th 1 1 1 1 1 1 1 1
2th 0.95 (0.78,1.17) 0.95 (0.78,1.17) 1.05 (0.84,1.32) 1.05 (0.84,1.32) 0.92 (0.76,1.12) 0.92 (0.76,1.12) 1.05 (0.87,1.27) 1.05 (0.87,1.27)
3th 0.87 (0.71,1.07) 0.87 (0.70,1.07) 1.06 (0.85,1.33) 1.06 (0.85,1.33) 0.92 (0.76,1.12) 0.92 (0.76,1.12) 1.11 (0.92,1.34) 1.11 (0.92,1.34)
4th 0.83 (0.67,1.02) 0.83 (0.67,1.02) 1.31 (1.06,1.62) 1.31 (1.06,1.63) 0.91 (0.75,1.10) 0.91 (0.75,1.10) 1.20 (0.99,1.44) 1.20 (0.99,1.44)

Note: OR crude odds ratio; aOR1, adjusted odds ratios for age; CI, confidence interval. Bold typeface indicates significance.
Abbreviations: ABSI, a body shape index; BMI, body mass index; EBP, elevated blood pressure; HTN, hypertension; WC, waist circumference; WHR, waist-to-height ratio.
|
1013
1014 | YAZDI et al.

TA B L E 3 Area under ROC curve (AUC), optimal cut-off values, sensitivities, specificities, and Youden index of anthropometric indices in
predicting EBP (BP > 90th percentile)

Sensitivity (95% Specificity (95% Youden

AUC(95% CI) Cut-off CI) CI) index PPV% NPV% P¶ P¶

Total
BMI z-score 0.571 (0.552,0.590) -0.925 77.7 (75.0,80.5) 32.6 (31.7,33.4) 0.103 8.2 95.0 .118
WC Centile 0.553 (0.534,0.571) 30.5 76.7 (74.0,79.5) 32.3 (31.5,33.2) 0.09 8.0 94.8 .011 .173
WHR 0.560 (0.541,0.579) 0.432 71.5 (68.5,74.4) 38.8 (37.9,39.7) 0.103 8.3 94.6 .118
ABSI 0.481 (0.462,0.500) 0.032 100.0 0.2 (0.1,0.3) 0.002 7.2 100.0 <.0001 .012
(100.0,100.0)
Girl
BMI z-score 0.586 (0.556,0.615) 0.075 53.2 (48.1,58.2) 59.6 (58.3,60.8) 0.128 8.0 95.0 .042
WC Centile 0.568 (0.539,0.597) 31.5 77.8 (73.6,82.0) 33.5 (32.2,34.7) 0.113 7.1 95.8 .13 .289
WHR 0.560 (0.531,0.588) 0.432 72.2 (67.7,76.7) 38.7 (37.4,40.0) 0.109 7.3 95.5 0.042
ABSI 0.462 (0.432,0.492) 0.066 97.3 (95.7,99.0) 4.8 (4.2,5.3) 0.021 6.3 96.4 .003 .393
Boy
BMI z-score 0.561 (0.536,0.586) -0.975 77.7 (74.1,81.3) 32.5 (31.2,33.7) 0.102 9.2 94.3 .604
WC Centile 0.540 (0.515,0.566) 24.5 83.4 (80.2,86.6) 23.9 (22.8,25.0) 0.073 8.8 94.2 .023 .008
WHR 0.560 (0.535,0.585) 0.421 81.6 (78.3,85.0) 29.6 (28.4,30.8) 0.112 9.3 94.8 .604
ABSI 0.489 (0.464,0.514) 0.077 74.1 (70.3,77.9) 29.1 (27.9,30.3) 0.032 8.4 92.7 <.001 .025

Abbreviations: ABSI, a body shape index; AUC, area under the curve; BMI, body mass index; NPV, negative predictive value; PPV, positive predictive
value; ROC, receiver operating characteristic; WC, waist circumference; WHR, waist-to-height ratio.
¶
P for comparison of the AUC value of BMI z-score with the other anthropometric indices.
¶¶
P for the comparison of the AUC value of WHR with the other anthropometric indices.

TA B L E 4 Area under ROC curve (AUC), optimal cutoff values, sensitivities, specificities, and Youden index of anthropometric indices in
predicting HTN (BP >95th percentile)

Cut- Sensitivity (95% Specificity (95% Youden

AUC(95% CI) off CI) CI) index PPV% NPV% P¶ P¶

Total
BMI z-score 0.592 (0.577,0.607) 0.025 56.1 (53.7,58.6) 57.9 (57.0,58.7) 0.14 14.1 91.5 .392
WC Centile 0.586 (0.570,0.601) 60.5 51.1 (48.6,53.6) 61.9 (61.0,62.7) 0.13 14.1 91.2 .21 .592
WHR 0.588 (0.573,0.603) 0.475 43.6 (41.1,46.1) 70.2 (69.3,71.0) 0.138 15.2 91.0 .392
ABSI 0.516 (0.500,0.531) 0.081 46.8 (44.3,49.4) 57.2 (56.4,58.1) 0.04 11.9 89.8 <.0001 <.0001
Girl
BMI z-score 0.600 (0.579,0.621) 0.245 52.1 (48.6,55.6) 62.8 (61.5,64.0) 0.149 15.3 91.1 .362
WC Centile 0.592 (0.571,0.613) 68.5 45.7 (42.2,49.2) 68.7 (67.5,69.9) 0.144 15.8 90.8 .362 .134
WHR 0.584 (0.562,0.605) 0.477 41.7 (38.3,45.2) 71.1 (69.9,72.2) 0.128 15.7 90.4 .064
ABSI 0.504 (0.483,0.525) 0.074 80.3 (77.5,83.1) 23.2 (22.1,24.2) 0.035 11.9 90.1 <.0001 <.0001
Boy
BMI z-score 0.584 (0.562,0.606) 0.075 54.1 (50.5,57.7) 59.6 (58.4,60.8) 0.137 13.6 91.7 .492
WC Centile 0.578 (0.556,0.601) 60.5 50.1 (46.5,53.7) 62.5 (61.3,63.7) 0.126 13.5 91.5 .354 .014
WHR 0.593 (0.571,0.615) 0.469 49.5 (45.9,53.1) 65.9 (64.7,67.0) 0.154 14.5 91.8 .492
ABSI 0.534 (0.511,0.556) 0.084 37.4 (34.0,40.9) 69.9 (68.8,71.0) 0.073 12.7 90.5 .006 <.0001

Abbreviations: ABSI, a body shape index; AUC, area under the curve; BMI, body mass index; NPV, negative predictive value; PPV, positive predictive
value; ROC, receiver operating characteristic; WC, waist circumference; WHR, waist-to-height ratio.
¶
P comparison of the AUC value of BMI z-score with the other anthropometric indices.
¶¶
P comparison of the AUC value of WHR with the other anthropometric indices.
YAZDI et al. | 1015

TA B L E 5 Prevalence of HTN in children and adolescents categorized by combination of BMI with WC, WHR or ABSI

Prevalence of Sensitivity (95% Specificity (95% Youden

HTN, n (%) AUC(95% CI) CI) CI) index PPV% NPV% Kappa

High BMI 885 (14.4) 0.578 (0.561,0.594) 56.2 (53.7,58.7) 59.3 (58.4,60.2) 1.155 14.4 91.7 0.068
High 643 (16.1) 0.577 (0.561,0.594) 42.3 (39.8,44.8) 73.2 (72.4,74.0) 1.155 16.1 91.2 0.090
BMI + High
WC
High 546 (16.1) 0.566 (0.55,0.581) 35.9 (33.5,38.3) 77.2 (76.5,78.0) 1.131 16.1 90.8 0.085
BMI + High
WHR
High 400 (16.9) 0.552 (0.538,0.567) 26.3 (24.1,28.5) 84.2 (83.6,84.8) 1.105 16.9 90.3 0.084
BMI + High
ABSI

Note:: High BMI, high WC, high WHR and High ABSI were defined according the age and sex specific cutoffs in Table S3.
Abbreviations: ABSI, a body shape index; AUC, area under the curve; BMI, body mass index; WC, waist circumference; WHR, waist-to-height ratio;
PPV, positive predictive value; NPV, negative predictive value.

TA B L E 6 Correlation coefficients
BMI WC Diastolic BP Systolic BP
among anthropometric indices and blood
z-score WHR centile ABSI z-score z-score
pressure
BMI z-score 1.00 0.542a 0.623a −0.341 0.181a 0.227a
WHR 1.00 0.794a 0.495a 0.103a 0.103a
a a
WC 1.00 0.369 0.166 0.197a
ABSI 1.00 −0.079a −0.127
Diastolic BP 1.00 0.680a
z-score
Systolic BP z-score 1.00

Note: Pearson's correlation coefficients between anthropometric parameters and blood pressure.
Abbreviations: ABSI; a body shape index; BMI, body mass index; BP, blood pressure; WC, waist
circumference; WHR, waist-to-height ratio.
a
Correlation is significant at the level of 0.01 (2-tailed).

and the duration of overweight and obesity, which could affect the AC K N OW L E D G M E N T
BP readings. The major strength of our study is recruiting a large pe- We forward our sincere thanks to the large team working in 30 uni-
diatric population with diverse lifestyle factors genetic background versities of the provinces, in which the study was conducted. We
living in urban and rural areas of 30 provinces of a vast country, would like to thank all organizations contributed to this survey, as
which allowed comparing the predictive ability of different anthro- well as students and their parents who participated in this study.
pometric indices for the early detection of children and adolescents
at risk of developing HTN. C O N FL I C T O F I N T E R E S T
Authors acknowledge that the submitted article is original, has not
been published previously in whole or part, and is not currently
5 | CO N C LU S I O N under review elsewhere. Authors further declare that there is not
any financial support or relationships that may pose a conflict of in-
The ability of BMI z-score, WC, WHR, and ABSI to identify Iranian terest regarding the content of this article.
children and adolescents at higher risk of EBP was week. Addition
of WC, WHR, or BSI to BMI did not improve the prediction of EBP AU T H O R S ' C O N T R I B U T I O N S
to BMI alone. MY and FA contributed to the conception and design of the work,
Of the four anthropometric indices, BMI, WC, and WHC show data acquisition, analysis, interpretation, drafting and revising the
a nearly similar predictive power for identifying EBP or HTN risk article critically for important intellectual content. RK contributed
factors among Iranian children and adolescents. The findings of this to the concept and design of the study, conducted the project with
study can be used in future international surveys in various ethnic MQ, ShD, and MEM and edited the final version of the manuscript.
groups of children and adolescents. Statistical analyses were performed by MY. All authors reviewed and
1016 | YAZDI et al.

approved the final version for submission. All authors accept the re- 13. Song L, Shen L, Li H, et al. Height and prevalence of hypertension in
a middle-aged and older Chinese population. Sci Rep. 2016;6:39480.
sponsibility of the manuscript content.
14. Jung C, Fischer N, Fritzenwanger M, et al. Anthropometric indices
as predictors of the metabolic syndrome and its component in ado-
C O M P L I A N C E W I T H E T H I C A L S TA N DA R D S lescents. Pediatr Int. 2012;52:402-409.
A comprehensive verbal description of the nature and purpose of 15. Bauer KW, Marcus MD, Ghormli EL, et al. Cardiometabolic risk
screening among adolescents: understanding the utility of body
the study was given to the students and their parents and written
mass indices waist circumference and waist-to-height ration.
informed consent was obtained from parents and oral assent from Pediatr Obes. 2015;10:329-337.
students. After getting the approval of relevant national regula- 16. Krakauer NY, Krakauer JC. A new body shape index predicts
tory organizations, the Research and Ethics council of Isfahan mortality hazard independently of body mass index. PLoS One.
University of Medical Sciences approved the study (Project number: 2012;7(7):e39504.
17. Wang H, Liu A, Zhao T, et al. Comparison of anthropometric indices
194049). The ethical committees of the Medical Sciences University
for predicting the risk of metabolic syndrome and its components
of Tehran and the Medical Sciences University of Isfahan reviewed in Chinese adults: a prospective longitudinal study. BMJ Open.
and approved research protocols. 2017;7:e016062.
18. Wang S, Liu Y, Li F, et al. A novel body shape score for detecting
association between obesity and hypertension in China. BMC Public
ORCID
Health. 2015;15:7.
Mayram Yazdi [Link] 19. Liang J-I, Chen Y-J, Jin Y, et al. Comparison of adiposity measures
Farahnak Assadi [Link] in the identification of children with elevated blood pressure in
Roya Kelishadi [Link] Guangzhou, Cina. J Hum Hpertens. 2015;29:732-739.
20. Xu Y, Yan W, Cheung YB. Body shape indices and cardiometabolic
risk in adolescents. Ann hum biol. 2015;42(1):70-75.
REFERENCES 21. Maessen MFH, Eijsvogels TMH, Verheggen RJHM, et al. Entering a
1. Parker ED, Sinaiko AR, Kharbanda EO, et al. Change in weight status new era of body indices: the feasibility of a body shape index and
and development of hypertension. Pediatrics. 2016;137(3):e20151662. body roundness index to identify cardiovascular health status. PLoS
2. Ogden CL, Carroll MD, Lawman HG, et al. Trends in obesity preva- One. 2014;9:e107212.
lence among children and adolescents in the United States, 1988– 22. Mameli C, Krakaur NY, Krakauer JC, et al. The association between
1994 through 2013–2014. JAMA. 2016;315(21):2292-2299. a body shape index and cardiovascular risk in overweight and obese
3. Vernooij JWP, van der Graaf Y, Visseren FL, Spiering W, SWART children and adolescents. PLoS One. 2018;13:e0190426.
study group. The prevalence of obesity-related hypertension and 23. Giudici KV, Martini LA. Comparison between body mass indes
risk for new vascular events in patients with vascular disease. and a body shape index with adiponectin/leptin ratio and mark-
Obesity (Silver Spring). 2012;20(10):2118-2123. ers of glucose metabolism among adolescents. Ann Hum Biol.
4. Urbina EM, Khoury PR, Bazzano L, et al. Relation of blood pres- 2017;44:489-494.
sure in childhood to self-reported hypertension in adulthood: 24. Chang Y, Guo X, Guo L, et al. The feasibility of two new anthro-
THE International Childhood Cardiovascular Cohort Consortium. pometric indices to identify hypertension in rural China. Medicine.
Hypertension. 2019;73(6):1224-1230. 2016;95:e5301.
5. Wang M, Kelishadi R, Khadilkar A, et al. Body mas index percen- 25. Ji M, Zhang S, An R. Effectiveness of body mass shape index (ABSI)
tiles and elevated blood pressure among children and adolescents. in predicting chronic diseases and mortality: a systemic review and
J Hum Hypertens. 2020;34(4):319-325. meta-analysis. Obes Rev. 2018;19:737-759.
6. Deng WW, Wang J, Liu MM, et al. body mass index compared 26. Tee JYH, Gan WY, Lim PY. Comparisons of body mass index, waist
with abdominal obesity indicators in relation to prehypertenstion circumferenece, waist-to-height ration and a body shape index
and hypertension in adults: the CHPSNE study. Am J Hypertens. (ABSI) in predicting high blood pressure among Malaysian adoles-
2013;26:58-67. cents: a cross-sectional study. BMJ Open. 2020;10:e032874.
7. Beck CC, AdaS L, Pitanga FJG. Anthropomeric indictors as pre- 27. Wang D, Li Y, Lee SG, et al. Ethnic differences in body composition
dictors of high blood pressure in adolescents. Arq Bras Cardiol. and obesity related risk factors: study in Chinese and white males
2011;96:126-133. living in China. PLoS One. 2011;6:e19835.
8. Li W-C, Chen I-C, Chang Y-C, et al. Waist-to-height ratio, waist 28. Motlagh ME, Ziaodini H, Qorbani M, et al. Methodology and early
ciecumference, and body mass index as indices of cardiometabolic findings of the fifth survey of childhood and adolescence sur-
risk among 36,642 Taiwanese adults. Eur J Nutr. 2013;52:65-75. veillance and prevention of adult noncommunicable disease: the
9. Carpenter CL, Yan E, Chen S, et al. Body fat and body mass index CASPIAN-V Study. Int. J Prev Med. 2017;2017(8):4.
among a multiethnic sample of college-age men and women. J Obse. 29. Flynn JT, Kaelber DC, Baker-Smith CM, et al. Clinical practice guide-
2013;2013:1-7. line for screening and management of high blood pressure in chil-
10. Nightinger CM, Rudnika AR, Owen CG, et al. Patterns of body size dren and adolescents. Pediatrics. 2017;140(3):e20171904.
and adiposity among UK children of South Asia, black African- 30. Poh BK, Jannah AN, Chong LK, et al. Waist circumference percen-
Caribbean and White European origin: child heart and health study tile curves for Malaysian children and adolescents aged 6.0-16.9
in England (Case study). Int J Epidemiol. 2011;40:33-44. years. Int J Peditr Obes. 2011;6:229-236.
11. Schneider HJ, Klotsche J, Silber S, et al. Measuring abnormal obe- 31. Etahed HS, Kelishadi R, Qorbani M, et al. Utility of waist circumfer-
sity: effects of height on distribution of cardiometabolic risk factor ence-to-height ration as a screening tool for generalized and central
using waist circumference and waist-to-height ratio. Diabetes Care. obesity among Iranian children and adolescents: the CASPIAN-V
2011;34:7. study. Pediatr Diabetes. 2019;20:530-537.
12. Hsieh O, Yoshinaga H. Do people with similar waist circumference 32. Li C, Ford ES, Mokdad AH, Cook S. Recent trends in waist circum-
share similar health risks irrespective of height? Tohoku J Exp Med. ference and waist-height ratio among US children and adolescents.
1999;188:55-60. Pediatrics. 2006;118(5):e1390-e1398.
YAZDI et al. | 1017

33. Onis M. WHO Child Growth Standards based on length/height, waist-to-height ratio for identifying caediometabolic risk in children
weight and age. Acta Paediatr. 2007;95:76-85. and adolescents. PolS One. 2016;11:e0149351.
34. de Onis M, Onyango AW, Borgghi E, et al. Development of a WHO 46. Expert panel on integrated guidelines for cardiovascular health
growth reference for school-aged children and adolescents. Bull and risk reduction in children and adolescents: Summary report.
World Health Organ. 2007;85:660-667. Pediatrics. 2011;128(Suppl 5):S213-S256.
35. Krakauer NY, Krakauer JC. An anthropometric risk index based on 47. Morandi A, Miraglia Del Giudice E, Martino F, Martino E, Bozzola
combining height, weight, waist, and hipp measurements. J Obese. M, Maffeis C. Anthropometric indices are not satisfactory predic-
2016;2016:8094275. tors of metabolic comorbidities in obese children and adolescents.
36. Heo M, Faith MS, Pietrobelli A, Heymsfield SB. Percentage J Pediatr. 2014;165:1178-1183.
of body fat cutoffs by sex, age, and race-ethnicity in the US 48. Chieolero A, Paradis G, Maximova K, Bumier M, Bovel P. No use for
adult population from NHANES 1999–2004. The Am J Clin Nut. waist-for-height ration in addition to body mass index to identify
2012;95(3):594-602. children with elevated blood pressure. Blood Press. 2013;22:17-20.
37. Krakauer NY, Krakuer JC. Dynamic association of mortality hazard 49. Cui J, HopperJL HSB. Genes and family environment explain correla-
with body shape. PLoS One. 2014;9:e88793. tions between blood pressure and body mass index. Hypertension.
38. Zou KH, O’Malley AJ, Mauri L. Receiver-operating characteris- 2002;40:7-12.
tic analysis for evaluating diagnostic tests and predictive models. 50. Merlo J, Li Y, Lee SG, et al. Ethnic differences in body composition
Circulation. 2007;115(5):654-657. and obesity risk factors: study in Chinese and White males in China.
39. Hosmer D, Lemeshow A. Applied Logestic Regression. Hoboken , NJ: PLoS One. 2011;6(5):e19835.
John Ailey and Sons; 2000.
40. Fluss R, Faraggi D, Reiser B. Estimation of the Youden index and its
associated cutoff point. Biom J. 2005;47:458-472. S U P P O R T I N G I N FO R M AT I O N
41. Core R, Team RR. A Language and Environmental for Statistical Additional supporting information may be found online in the
Computing. Vienna, Austria: R Foundation for Statistical Computing;
Supporting Information section.
2013.
42. Li Y, Zou Z, Luo J, et al. The predictive value of anthropometric in-
dices for cardiometablic risk factors in Chinese children and ado-
lescents: an nationalmulticenter schoool-based study. PolS One. How to cite this article: Yazdi M, Assadi F, Qorbani M, et al.
2020;15(1):e0227945. Validity of anthropometric indices in predicting high blood
43. Lo K, Wong M, Khalechelvam P, Tam W. Waist-to-height, body mass pressure risk factors in Iranian children and adolescents:
index and waist circumference for screening pediatric cardiometa-
CASPIAN-V study. J Clin Hypertens. 2020;22:1009–1017.
bolic risk factors: a meta-analysis. Obes Rev. 2016;17:1268-1275.
44. Ma L, Cai L, Deng L, et al. Waist circumference is better than other [Link]
anthropometric indices for predicting cardiovascular risk fac-
tors in Chinese Children-a cross-sectional study in Guanngzhous.
J Atheroscler Thromb. 2016;23:320-329.
45. Sardinha LB, Santos DA, Silva AM, Grontved A, Anderson LB,
Ekelund U. A comparison between BMI, waist circumference, and