Waist-To-Height Ratio as an Indicator of Dyslipidemia in Brazilian School-Aged Children and Adolescents

Research Article

J Pediatr & Child Health Care. 2016; 1(1): 1007.

Waist-To-Height Ratio as an Indicator of Dyslipidemia in Brazilian School-Aged Children and Adolescents

Dornelles AZ¹, Bueno CT², Rotta LN¹ and Vitolo MR¹*

¹UFCSPA- Federal University of Health Sciences of Porto Alegre, Brazil

²ULBRA-Lutheran University of Brazil, Brazil

*Corresponding author: Vitolo MR, Federal University of Health Sciences of Porto Alegre, Brazil

Received: May 17, 2016; Accepted: June 26, 2016; Published: July 03, 2016


Objectives: To evaluate whether BMI-for-age Z-score (BAZ), Waist Circumference (WC), or Waist-to-Height ratio (WHtR) accurately identifies dyslipidemia in a school age population.

Study Design: This was a cross-sectional design study from a city of South of Brazil. A total of 592 children and adolescents from 11 schools were stratified into four groups: boys 6-11 years; boys 12-17 years; girls 6-11 years; girls 12-17 years. Area under the Curve (AUC) for ROC distribution evaluated anthropometric tools’ performance to identify two or more alterations on lipid profile (Total Cholesterol, Triglycerides, HDL-cholesterol, LDL-cholesterol). Youden Index was used to extract optimal cut-off points for each anthropometric tool.

Results: All anthropometric indices achieved moderate to high ability to identify dyslipidemia. WHtR was among the best performances for AUC in all groups, when compared to other anthropometric tools, and presented the strongest performance for boys at 12-17 years (0.902±0.04). According to best performances, optimal cut-offs points were: 0.74 for BAZ, 63.25 cm for WC, and 0.485 for WHtR in boys at 6-11. WHtR’s cut-off of 0.5 for boys at 12-17 presented the highest accuracy (0.757) when compared to all indices among all groups. Girls at 6-11 presented optimal cut-off values of 1.71 for BAZ, and 0.499 for WHtR, while the older ones presented 71.25 cm for WC and 0.498 for WHtR.

Conclusions: The three anthropometric indices were useful to identify dyslipidemia in school children and adolescents. Overall, WHtR can improve and simplify the screening process and should be encouraged as an index of cardiovascular risk factors.

Keywords: Body Mass Index; Waist Circumference; Waist-Height Ratio; Child; Cholesterol; Hdl; Lipids; Cardiovascular Diseases; Risk


AUC: Area under the Curve; BAZ: BMI-for-age Z-score; BMI: Body Mass Index; CVD: Cardiovascular Disease; DALYs: Disability Adjusted Life Years; ROC: Receiver Operating Characteristic; UFCSPA: Federal University of Health Sciences of Porto Alegre; WC: Waist Circumference; WHtR: Waist-to-Height Ratio


The increased prevalence of obesity and overweight is now considered to be a global pandemic [1-3]. In 2010, it was estimated that overweight and obesity caused 3-4 million deaths, a loss of 4% loss of life years, and 4% of Disability Adjusted Life Years (DALY) due to co-morbidities associated with excessive body fat mass [4]. Globally, the prevalence of overweight and obesity is 27.5% for adults is 47.1% for children. In absolute numbers, this population has increased from 857 million to 2.1 billion people, from 1980 to 2013 [5]. Thus, it is clear that obesity and overweight are serious public health problems and it is important not just to improve our understanding of how excess body weight increases the risk for chronic disease but how to prevent this scenario from growing.

Excess body weight and obesity are positively associated with Cardiovascular Diseases (CVD), the leading cause of death in the world [6,7]. The development of CVDs is associated with a number of risk factors, such as the consumption of high-fat diets and high levels of plasma lipids, which generally begin in childhood [8,9]. As reported by Berenson et al. [10], the level of cholesterol in infancy is a predictive factor of the cholesterol level in adulthood, a primary risk factor for CVDs. Atherosclerosis begins in infancy by the increase of plasma cholesterol and can be accelerated by obesity through a number of indicators, such as family history, physical inactivity, and hypertension [11]. Therefore, it is important to intervene as soon as possible to prevent risk factors from becoming more severe or numerous. Yet, it is unclear what anthropometric variable is most associated with abnormal lipid profiles, making the criteria for intervention difficult to determine.

The assessment of excess body weight and its relationship to other biological risk factors can be accomplished with relatively simple anthropometric markers. Two of the most common anthropometric assessments of overweight are Body Mass Index (BMI) and Waist Circumference (WC) which may explain part, but not all, of the potential risk for CVD in children. These two anthropometric indices have presented strong associations with fat mass for children and adults [12]. Waist-to-Height Ratio (WHtR) has been proposed as an alternative measure that concerns both longitudinal growth and central adiposity [13]. This index can be used to screen relevant changes in body composition and have already been used to identify risk for chronic diseases in adults and children [14,15]. However, how these three anthropometric indices are able to identify dyslipidemia in children of different ages and sex has not been well established. Therefore, the aim of this study was to compare the accuracy of BMI-for-age Z-score (BAZ), WC and WHtR to identify lipid profile alterations in school age children and adolescents from Sapucaia do Sul, Brazil.

Materials and Methods

Study design and population

This cross-sectional study was based on a population of 592 children and adolescents of school age (6-17 years) from Sapucaia do Sul city, an urban city with an estimated population of 130,957 located in the state of Rio Grande do Sul, in the south of Brazil, enrolled in 2008. All public schools ruled by the city’s government were invited in a collaborative work with Municipal Department of Health. School’s directors that declared an interest and viability to participate in the study represented the eleven municipal schools. Groups were classified according to age and sex. Children and adolescents were stratified into four groups: Boys at 6 to 11.99 years; Boys at 12 to 17.99 years; Girls at 6 to 11.99 years; Girls at 12 to 17.99 years.

The staff that conducted the data collection was previously trained and informed according to the project’s technical and ethical aspects. Children of school age and regularly matriculated at the selected schools were invited to initiate in the study. Familiar responsible accompanied the children and signed the Informed Consent with the supervision of a previously trained member of the research’s staff. Data was collected in a regular school day by the morning period.


A single trained field worker obtained both subject’s weight and height. Weight was measured with no shoes and light clothes on using a digital weighing-machine (Techline®, Brazil), precision of 100g. Height was measured to the nearest 0.1 cm using a wallmounted stadio-meter (Seca®, Germany). Waist Circumference (WC) was obtained from the narrowest point of the subject’s chest with a non-extensive tape measure. Waist-to-Height Ratio (WHtR) value was obtained from the equation: Waist (cm)/Height (cm). Body- Mass-Index (BMI) value was obtained from the equation: weight (kg) / height (m2). Definition of overweight and obesity was, according to BMI-for-age Z-score (BAZ). Children who presented >1 Z-score were classified as overweight, and >2 Z-score were classified as obese [16]. Those who presented =90 percentile of WC were considered as central obese [17]. Values =0.5 for WHtR were considered high [13].

Blood analysis

Blood samples were obtained in 12 hours fasting state, with a vacuum system from the brachial vein. The samples were collected in tubes with no anti-coagulant for serum. The tubes were centrifuged at 3,500/10 minutes and analyzed immediately afterwards. Analysis were obtained using automation apparatus Lab max (Lab test, Brazil), after quality control (3 times for control and 3 times for pathologic serums), at The Laboratory of Clinics Analysis in Divina Providência Hospital- Porto Alegre. Total Cholesterol and Triglycerides were extracted by colorimeter enzymatic test (altered values were established at =170 and =130 mg/dL, respectively), HDL-cholesterol was extracted by precipitation (altered value was established at =45 mg/dL) and LDLcholesterol was calculated by the Friedwald equation (altered value was established at =130 mg/dL). Lipid concentrations were measured with Lab test kits. Two or more altered values in lipid concentrations were classified as dyslipidemia.

Statistical analysis

Data were analyzed with the software Statistical Package for the Social Sciences (SPSS) version 21.0. In order to evaluate the homogeneous distribution, Kolmogorov-Smirnov test was applied. Normal distribution variables were presented as mean and standard deviation, while non-parametric variables as median and interquartile intervals. Qualitative variables were presented as absolute and relative frequencies. One-way Analysis of Variance (ANOVA) was applied to compare the groups according to anthropometric and biochemistry data. Tukey’s post-hoc test was applied to express significant differences among groups, defined as <5%.

Receiver Operating Characteristic (ROC) analysis was used to explore anthropometric indices’ performance to associate lipid profile alterations. The accuracy of the anthropometric indices to identify dyslipidemia were assessed by the Area under the Operating Characteristic Curve (AUC); A 95% confidence interval was used (95% CI), based on two or more alterations in the lipid profile. In order to calculate the optimal cut-off point, Youden index (sum of sensitivity and specificity-1) was used [18].

Ethical aspects

The current study was submitted and approved by the Ethical and Scientific Committee of Universidade Federal de Ciências da Saúde de Porto Alegre - UFCSPA (716/08).


This study was a cross-sectional analysis of 592 low income children and adolescents from 6 to 17 years old. A significant number of families were of low-income, as 44.3% had a family annual income of under USD 3.141.52, which represents a monthly income equal to or lower than the national minimum wage (approximately USD 261.79 per month). Mothers spent an average of 6.79±2.8 years in school, while fathers spent 7.01±2.8 years in school.

Prevalence of overweight was 40.5% for the boys and 34.4% for the girls, while 20% of the boys and 12.9% of the girls were classified as obese. High WHtR was identified in 28.2% of the boys and 29.6% of the girls. Central obesity was observed in 22.2% of the boys and in 25.4% of the girls. There were no differences in anthropometric indices between the same age groups (Table 1). Boys showed lower levels of HDL-cholesterol, in mg/dL, at 12-17 years when compared to younger ones (48.7±12 vs 43.9±9.4) and when compared to girls of the same age (48.7±10.9 vs 43.9±9.4). The boys presented higher prevalence of triglycerides alterations at both age groups, when compared to girls (8.5% vs 5.8% at 6-11 and 11.4% vs 6.1% at 12-17) (Table 2). Dyslipidemia was identified in 11.8% of the boys and 10.5% of the girls