Blood pressure measurements in children require trained professionals to carry out and interpret the readings [9]. As this may be difficult in a school-setting, the use of anthropometry, carried out as part of a routine school physical examination, would be beneficial to identify children at risk of high blood pressure. Commonly used obesity indicators such as waist-to- height ratio (WHtR), body mass index (BMI) and waist circumference (WC) have been examined as alternate indicators of HBP in adults and children [10-14]. WHtR has been found to be a simple, easy and accurate index, suitable as a screening tool for obesity in children and adolescents from India [15]. However, studies on the association between WHtR and childhood high blood pressure [14,16,17], are unavailable for Indian children. The aim of the present study was to examine the association of WHtR, BMI, and WC with high blood pressure in urban children aged 6 to 16 years from Southern India.

Demographic data on date of birth, medical history, birth weight, parental education, occupation, income, parental height and weight were collected from the parents. The questionnaires used in this study were used in the previous PEACH study [15], and have been pre-tested to ensure that both the children and parents understood all questions and provided reliable data. Comprehensive questionnaires capturing physical activity patterns, tuition time (hours per week), time spent in sedentary activities at home (watching television, playing computer games) and average sleep duration per night were completed by the parents of students below the 6th grade and by the students in the 6th grade and above. Physical activity patterns were assessed by asking about time spent playing games outside school and the type of sports in which the children regularly engaged.

Anthropometric measurements of body weight and height were performed by trained staff using standardized procedures [18]. Body weight was measured to the nearest 0.1 kg using a calibrated digital scale (Tanita, Tokyo, Japan); children were asked to remove their shoes and socks, and to change into light clothing. Height measurements were taken to the nearest 0.1 cm using a portable stadiometer (Seca 213, Germany); children were asked to remove shoes and socks. For measurement of WC, children were asked to remain in a standing position and measurements were taken using a non-stretchable tape during end-tidal expiration, exerting the same standard pressure on the tape at the midpoint of the lowest ribcage and the iliac crest [19]. WHtR was calculated as WC (cm) divided by height (cm) for all children. BP measurements were taken using a mercury sphygmomanometer as recommended in American Heart Association guidelines [9]. Measurements were taken in a quiet room while the child was sitting with their arm resting on a table. Efforts were made to eliminate factors which may affect BP such as anxiety, crying or prior exercise. The average of two consecutive readings was used in the analysis. All BP measurements were taken by the same trained professional.

The different sports and games reported were classified into three categories (mild, moderate, vigorous) based on the metabolic equivalent (MET) score [20]. Activities below 3 METS were considered as light, 3-6 MET were considered as moderate, and greater than 6 METS as vigorous. The total duration of physical activity (in MET hours) was then calculated by multiplying the total duration of games played in each category with its average MET score value. Time spent watching television and time spent on the computer/video games was grouped together as screen time.

The discriminative capacity of BMI Z scores, WC Z scores and WHtR in their ability to determine high SBP and high DBP was examined and compared by plotting the Receiver Operating Characteristic (ROC) curves for each indicator and comparing the area under the curve (AUC). The AUC of WHtR was compared with that of BMI and WC using a previously described method [24].

The association of high WHtR, overweight/obese BMI and high WC with high systolic and diastolic BP was examined using log binomial regression. Multivariable log binomial models were constructed with the same covariates as mentioned above. Crude (RR) and adjusted risk ratios (ARR) are reported along with 95% CI. Each of these obesity indicator was considered as independent variable in separate regression analyses.

Two-sided P-values <0.05 were considered statistically significant. Statistical analyses were performed using SAS program (version 9.2; SAS, Cary, N.C., USA) and SPSS (IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY: IBM Corp.). Log binomial regression analysis was performed using the PROC GENMOD program in SAS.

A total of 1913 children (1111 males) were analyzed. Mean age of the children was 12 years. The proportion of children with high SBP and high DBP were 8.0% (108) and 2.7% (51) respectively. High WHtR was observed in 14.1% (269) of the children, while 15.1% (289) were overweight/obese and 21.7% (415) had a high WC. WHtR was strongly correlated with BMI Z score and WC percentiles (r=0.89 and 0.74, respectively, both P<0.001). Table I depicts the demographic, behavioral and anthropometric characteristics along with the BP of the children in the study.

TABLE  I	Demographic, Anthropometric and Behavioral Characteristics of Children  (N=1913) in the Study

The multivariable linear regression model adjusted for the following factors based on their significance in univariable analysis with either/both BP percentiles: maternal BMI (P=0.091 DBP), screen time (P<0.001 SBP, P=0.001 DBP), tuition time (P=0.021 SBP, P=0.040 DBP), physical activity (P=0.004 SBP, P=0.019 DBP) and sleep duration (P=0.010 DBP). Similarly, log binomial regression analysis adjusted for age, sex and sleep duration (P<0.001 SBP) based on univariable P values. Age and sex were included in all models regardless of P value as they directly affect BP.

TABLE II	AUCs of Waist-to-Height Ratio, BMI and Waist Circumference for Risk of High Systolic or 
Diastolic Blood Pressure (BP) (N=1913)

Some of the limitations of the present study were that all readings were taken on the same day and there was no follow-up to confirm hypertension. Additionally, since the dataset represents urban children attending schools in Bangalore, these results need to be tested in a more diverse setting, before they can be generalized to the population of children in India. WHtR has a low sensitivity, and thus some children may be missed by this initial screening tool. However, due to its ease of administration with only a tape measure, simplicity of measurement and application of a single cutpoint across all ages, we suggest it be used as a simple screening tool in Indian populations until a more sensitive tool can be found. Given the large sample size of this study, with little inter-observer variation in BP readings, and the consistency of results with available literature from other populations, it is reasonable to conclude that simple anthropometric measures may be used to initially screen school going children for risk of HBP in an urban Indian setting. The WHtR has the added advantage of being easy to calculate and interpret, and hence can be used widely during routine school physical examination.

Acknowledgements: Deepa P Lokesh, Mamatha Philip, Jini V Aravind, Jabin Dany Raj R, and Jayakumar J for help in data collection. We acknowledge the contribution of schools and teachers for their support.

Contributions: PEM, LS, TT, RK, AVK: conceptualizing the study, analyzing data and writing the manuscript; CD, RB, CMM: analyzing data and writing the manuscript. AVK: will be the guarantor of the study. The final manuscript was approved by all authors.

Funding: Some aspects of this study were funded by ICMR grant no 5/4/8-10/NCD-II. The rest was funded by intramural sources. Christopher Duggan was supported in part by NIH/NICHD K24 HD058795.

Competing interests: None stated.