Hippokratia 2015, 19(3):239-248
Tambalis KD1, Panagiotakos DB1, Arnaoutis G1, Psarra G1, Maraki M1, Mourtakos S1, Grigorakis D1, Sidossis LS1,2
1Department of Nutrition and Dietetics, Harokopio University, Athens, Greece, 2Department of Internal Medicine, Sealy Center on Aging, Institute for Translational Sciences and Shriners Hospital for Children, University of Texas Medical Branch at Galveston, Texas, USA
Background: We sought to establish cross-sectional curves for body weight, height, body mass index (BMI), and waist circumference for 4- to 18-year-old Greek boys and girls, using the empirical distribution and the Lambda Mu and Sigma (LMS) statistical method.
Methods: From March 2014 to May 2014, a total of 473,837 boys and girls aged 4 to 18 years who attended school in Greece were enrolled. The studied sample was representative, in terms of age-sex distribution and geographical region. Anthropometric indices measurements (i.e., height, weight and waist circumference) were performed and BMI was calculated and used to calculate normative values, using the percentiles of the empirical distributions and the LMS method.
Results: Updated growth references for 4- to 18-year-old Greek children tabulated as critical percentiles and LMS values from 3 to 97 (P3, P10, P25, P50, P75, P90, P97) and smoothed curves are presented. Positive secular trends of height, weight and waist circumference were observed in children and adolescents 4 to 18 years old (all p values <0.001). At all ages, boys had higher anthropometric measurements than girls (all p values <0.01). Compared to 1998 data, mean height and weight were greater in 2014 for boys and girls at all ages.
Conclusion: Current national percentile curves for anthropometric indices could provide a more accurate estimation to assess physical growth in Greek children and adolescents. Hippokratia 2015; 19 (3): 239-248.
Key words: Βody mass index, BMI, cutoff point, children, Greece
Corresponding author: Labros S. Sidossis, PhD, University of Texas Medical Branch at Galveston, 301 University Blvd, Galveston Texas, 77555-0177, USA, tel: 4092669690, e-mail: firstname.lastname@example.org
It has been widely adopted that children’s growth is determined by both genetic and environmental factors. The assessment of somatic growth by objective anthropometric indices (e.g., body weight, height and waist circumference) is essential in child’s care to evaluate the nutritional status and to identify growth failure. Specifically, growth references help public health professionals to evaluate the growth status of a child and to diagnose undernutrition, overweight and obesity, and other growth-related conditions among children1-2. The pattern of children’s growth changes with time and hence it is recommended that references should be updated regularly3. Moreover, various ecological, environmental and genetic factors have been associated with children’s growth. Therefore, it is important for each country to use growth curves specific to the population of interest.
In Greece, the need to develop appropriate updated growth references for children and adolescents has been motivated by two contemporary events: the public health concern over the rising incidence of childhood obesity4 and the lack of national representative epidemiologic data regarding anthropometric measurements of children for these age groups. The currently existing growth reference curves in use in Greece are based on data collected by the First Paediatric Clinic of the University of Athens in 20015. To the best of our knowledge, Greece lacks current national cut-offs for a wide age-range. In addition to the aforementioned considerations, during the past decades, different methodologies have been used to develop children’s growth curves, mainly based on the reference range of the distribution. However, when the measurement is strongly related to another factor, like height and age in children, the age-dependent, smoothed, centile curves looks much more appropriate since they incorporate the dynamic of the covariate. Another issue was the lack of normality. Cole et al assumed an underlying normal distribution of the anthropometric measurements, in which case a power transformation would render to a normal distribution6. Thus, a three-parameter method was proposed, the Lambda Mu and Sigma (LMS) statistical method, where L reflects the Box-Cox power lambda, M the arithmetic mean of the measurement and S the coefficient of variation σ7. This methodology has been adopted by the International Obesity Task Force (IOTF) in order to develop global growth curves for children and adolescents8.
Therefore, the aim of the present study was to establish updated age-sex specific normative values of anthropometric indices and to evaluate sex- and age-related differences, using both percentiles and Z-score values in a nationwide sample of 4- to 18-year-old children and adolescents.
Methods and Procedures
Population-based, representative data were derived from a nationwide school-based survey under the auspices of the Ministry of Education. Specifically, anthropometric, physical activity, nutrition, and physical fitness data along with information on age and sex were collected from March 2014 to May 2014. In total, 473,837 (51% boys and 49% girls) children aged 4 to 18 years old from public and private schools agreed to participate in the study (participation rate was 40% of the total population). The working sample was representative of the entire Greek population (chi-square p-value as compared to the current sample with the age-sex distribution of all Greek areas = 0.93). The participation rates assured the proportional enrolment of children based on the urban/rural areas student population distribution.
Demographic information of students (e.g., school, class, gender and date of birth) was obtained from each school headmaster. Children’s height, weight and waist circumference were measured in the morning, using a standardized procedure. Data collection in each school was completed in one working day. Children were instructed to wear little clothing and stand with feet close together, arms at the side so body weight was evenly distributed. The exact ages of the participants were calculated from birth and examination dates. Weight was measured in the standing upright position with electronic scales with a precision of 100 g. We determined standing height to the nearest 0.5 cm with the child’s weight being equally distributed on the two feet, head back and buttock on the vertical land of the height gauge. We calculated BMI as the ratio of body weight to the square of height (kg/m2). Waist circumference was measured at the midpoint between the lower margin of the least palpable rib and the top of the iliac crest, using a flexible measure to the nearest 0.1 cm. All measurements were repeated; if the measurements were within 1 cm of one another, the average was calculated. All anthropometric measurements were performed by trained professionals (teachers of Physical Education). Specifically, measurements were performed by one teacher of Physical Education in each class. All Physical Education professionals were instructed through a detailed and extended manual of operations and followed a standardized procedure of measurements in order to minimize the potential inter-rate variability among schools. The Physical Education teachers were first trained by a school advisor of Physical Education for accurate anatomical landmarks, subject positioning and measurement techniques. Verbal informed consent was obtained, for each child to participate in the measurements, from physical education teachers. As the measurements were included in an obligatory school program, verbal informed consent was considered sufficient. Ethical approval for the survey was granted by the Review Board of the Ministry of Education and the Ethical Committee of Harokopio University (decision No 37/20-02-2013).
Comparisons of the anthropometric measurement data between boys and girls were performed using the independent samples t-test, after testing for equality of variances using the Levene test. Comparisons of percentile values of anthropometric measurement data between both calculation methods were performed using the paired samples t-test. We performed the tests for sex-specific trends in the anthropometric measurements using linear regression analysis (with lag 0). We evaluated serial dependency using the partial autocorrelation function; no autocorrelation was observed for various lags tested. Age-sex specific distributions and percentiles were calculated using two methods: using the empirical distribution of the data to calculate the 3th, 10th, 25th, 50th, 75th, 90th, and 97th percentiles; also, using the LMS method proposed by Cole et al7. The LMS method was used in order to smooth the age-dependent skewness usually observed in anthropometric values. In this method, the data were normalized using the Box-Cox power transformation. The principle idea of the LMS method is to power-transform the measurement, i.e., height here, and to use the coefficient of variation (CV =standard deviation/mean) of the raw data. The optimal Box-Cox power λ is the one that gives the lowest CV7. Thus, the LMS method calculates the best power (L), the best mean (M) and CV (S) in each series of measurements at a specific age. All statistical analyses were performed using the SPSS program (version 18; SPSS Inc., Chicago, IL, USA). The LMSchartmaker9 and the LMSgrowth10 freeware packages were used to calculate L, M and S values at ages 4 to 18 based on Greek reference values.
Comparison with international growth references
The 3rd, 50th, and 97th percentiles of the constructed growth charts of the present study were selected to be compared with the corresponding percentiles of the United States’ Centers for Disease Control and Prevention 2000 growth charts11, based on comparability with regard to statistical methods used and on historical usage.
In Tables 1-4 are presented normative anthropometric indices (height, weight, waist circumference and BMI) for 4- to 18-year-old children in Greece, by gender and age, as tabulated critical percentiles and LMS values from 3 to 97 (P3, P10, P25, P50, P75, P90, P97). Also presented are the gender- and age-specific LMS values for the above anthropometric indices. Figure 1 shows the smoothened weight, height, BMI and waist circumference curves for Greek boys and girls using 3rd, 10th, 25th, 50th, 75th, 90th and 97th percentiles. A secular trend of increasing weight, height and waist circumference was observed in all Tables and Figures (all p-values <0.001). For each of the anthropometric indices, measurements were higher in boys compared with girls, at all ages (all p-values <0.01). In girls, height and waist circumference measurements seemed to tend to peak at about the age of 17 years, although the age range was deteriorated.
Figure 1: Growth curves for weight, height, body mass index and waist circumferences percentiles for boys (left) and girls (right) using the Lambda Mu and Sigma (LMS) method. * Age: completed age, e.g., 4 years = 4.00–4.99 years.
In order to investigate potential differences between percentile values from the two methods, comparisons by anthropometric measurement were performed. Data analysis did not reveal any significant differences between critical percentiles and LMS percentiles in height and weight, while in waist circumference, critical and LMS percentiles differed significantly (p =0.002).
Moreover, aiming to investigate potential differences between 2014 and 1998 in weight and height in Greek boys and girls aged 8, 9 and 10 years (data not shown)12, we compared mean values by gender. In 8-year-old boys, height and weight increased from 130.9 ± 6.1 cm and 30.0 ± 6.3 kg in 1998 to 134.2 ± 6.3 cm and 32.6 ± 7.5 kg in 2014, respectively (p-values <0.001). Correspondingly, mean values of height and weight for 8-year-old girls increased from 129.8 ± 5.9 cm and 29.3 ± 6.1 kg in 1998 to 133.2 ± 6.3 cm and 31.9 ± 7.3 kg in 2014, respectively (p-values <0.001). In 9-year-old boys, height and weight increased from 136.1 ± 6.5 cm and 33.5 ± 7.2 kg in 1998 to 139.6 ± 6.6 cm and 36.7 ± 8.5 kg in 2014, respectively while the correspondence values in girls increased from 135.0 ± 6.6 cm and 32.8 ± 7.2 kg in 1998 to 138.9 ± 6.9 cm and 35.9 ± 8.5 kg in 2014 (all p-values <0.001). In 10-year-old boys, height and weight increased from 141.1 ± 6.6 cm and 37.0 ± 8.1 kg in 1998 to 144.9 ± 7.0 cm and 41.9 ± 9.9 kg in 2014, respectively while in girls, height increased from 140.9 ± 7.0 cm to 145.5 ± 7.4 cm and weight from 36.5 ± 8.1 kg 41.0 ± 9.8 kg, over the same time period (all p-values <0.001). Furthermore, BMI increased from 17.4 ± 2.8 kg/m2, 18.0 ± 3.1 kg/m2 and 18.5 ± 3.2 kg/m2 in 1998 to 18.0 ± 3.1 kg/m2, 18.7 ± 3.4 kg/m2 and 19.4 ± 3.7 kg/m2 in 2014 in 8-year-old, 9-year-old and 10-year-old boys, respectively (all p-values <0.001), and from 17.3 ± 2.8 kg/m2, 17.9 ± 3.1 kg/m2 and 18.3 ± 3.2 kg/m2 in 1998 to 17.9 ± 3.2 kg/m2, 18.5 ± 3.4 kg/m2 and 19.2 ± 3.7 kg/m2 in 2014 in 8-year-old, 9-year-old and 10-year-old girls, respectively (all p-values <0.001).
Figure 2shows the 3rd, 50th, and 97th percentiles for height, body mass, BMI of Greek children against U.S.11 counterparts. Third percentile curves for weight, height and BMI of the present Greek sample were comparable with those of U.S. sample (CDC) 2000. On the other hand, as depicted in 50th and 97th percentiles, Greek boys were on average heavier, while only somewhat taller, resulting in having greater BMI than their U.S. counterparts. A similar trend was observed in Greek girls, however after puberty differences between Greek and U.S. girls are minimized.
Figure 2: Comparison of the age and sex-specific 3rd, 50th, and 97th weight, height and body mass index percentile curves for the United States’ Centers for Disease Control and Prevention 2000 reference (CDC) and present study (Greece) among boys (left) and girls (right) aged 4-18 years. * Age: completed age, e.g., 4 years = 4.00–4.99 years.
The aim of the present study was to develop up-to-date age- and sex-specific anthropometric indices of normative values for Greek children and adolescents aged 4-18 years and to compare specific percentile values from two widely-applied estimation methods: the frequency percentiles and the LMS smoothed percentiles. This study provides current information on normative values of somatic growth. These values could be used as approximate indicative values to compare anthropometric indices scores of children from other countries similar to Greece: i.e., a developed country with a population predominately Caucasian. These data can also be used as benchmark values for health screening and surveillance of children and adolescents 4 to 18 years old in Greece.
Childhood obesity is a growing public health problem worldwide. More than 20% of children are classified as overweight/obese according to the International Task Force Organisation, and percentages seem to be rising12-13. The prevalence of paediatric overweight among school-aged children is high across Europe, as well as in other developed countries, with a particularly worrying prevalence in Greece. Indicatively, 29.5% of 10- to 12-year-old boys and girls were classified as overweight according to a cross-sectional, population-based survey from Greece14. The data from our study confirm this troubling prevalence in our country, indicating a secular trend of increasing weight and waist circumference among school-aged children over the long term. The alarming magnitude of the paediatric overweight/obesity problem highlights the need for country-specific policies and interventions in order to confront this epidemic in European schoolchildren.
Interestingly, the distribution of childhood obesity differs significantly between northern and southern European countries, with the countries of the south (e.g., Greece, Italy, Cyprus, Spain) showing the highest proportions15-16. Despite the economic crisis of recent years, there is an inverse relationship between the prevalence of overweight and obesity in children and family income or financial status13-14. Moreover, the finding for each of the anthropometric indices – that the measurements were higher in boys compared with girls at all ages – is in accordance with other relative studies, where weight and height also increased with age and boys were also taller and heavier than girls in all, similar, age groups15-17. Likewise, it is a common finding throughout the last three decades that weight and height have shown an increasing trend in developed and developing countries worldwide, beginning from early childhood. Indicatively in Greece, for 8-year-old boys and girls, height and weight increased significantly from 1998 to 2014 (e.g., 97th percentiles of height and weight for boys were 143 cm and 45 kg, respectively, in 1998, increasing to 146 cm and 50 kg in 2014, while corresponding values for girls were 141 cm and 43 kg in 1998, increasing to 145 cm and 48.2 kg in 2014), suggesting that these increases contributed to the aforementioned rise in the corresponding BMI values (23.9 vs. 25.3 kg/m2 for boys and 23.6 vs. 24.9 kg/m2 for girls). Interestingly, large cohort studies indicate continuous gains in both of the aforementioned anthropometric characteristics at all ages, with peak values observed in younger ages, especially in developed countries18. Finally, it is well accepted that a child’s growth is an indicator of the health and wellbeing of a society19. Monitoring growth to identify health- or nutrition-related problems is an important task of health care professionals. Unfortunately, Greece has no representative national sample of children which has been regularly updated. Therefore, the results from the present study provide a very useful public health tool by which health care professionals can assess and evaluate the growth status of children nationwide. However, as previously suggested20, using the present developed BMI curves may underestimate the true prevalence of overweight and obesity in Greece from now on, as the reference population used is not the “ideal” population due to the increasing obesity and overweight trends in Greece. This is a common problem in developing growth curves; however, our data are unique in Greece, regarding age, national gender and geographical distribution and methodology used. “Freezing” our curves from now on, as suggested for UK charts, may minimize these limitations in the future21.
The present study has several strengths. The sample is representative of the national gender and geographical distribution, as we studied almost 40% of the total population of 4- to 18-year-old children and adolescents in Greece. The study was performed in children 4-18 years. This age range is an advantageous period in the life of children and adolescents at which to apply effective strategies to prevent and improve obesity levels. The use of the widely-adopted LMS method for growth curves and the huge, nationwide representative sample of about 475,000 participants aged 4 to 18 years old used here guarantee the generalization of the results. Finally, the presented data were derived using the same standardized procedures in all schools. There are also limitations in our study design. Although a common, validated protocol was used to evaluate anthropometric indices measurements in all schools, a large number of experienced, professional physical educators participated as evaluators in the study and different brands of electronic scales and stadiometers were used. In order to minimize the variability among the different experimenters, all educators were instructed through a detailed and extended manual of operations and followed a standardized procedure of measurement. Even so, some variability in measurement may still exist. Finally, the cross-sectional design of our study cannot provide causal relationships, but only hypotheses for further research.
In conclusion, we established sex- and age-specific normative anthropometric indices values for children and adolescents aged 4 to 18 years old living in Greece. These findings may help policy makers to evaluate somatic growth and obesity levels and to design appropriate health-related educational programs for the young.
Conflict of interest
Authors declare that they have no conflicts of interest.
This study was supported by the Hellenic Ministry of Education and Religious Affairs, OPAP S.A., Nestlé Hellas S.A., the Department of Internal Medicine and the Sealy Centre on Aging, University of Texas Medical Branch at Galveston, and the Department of Nutrition and Dietetics Graduate Program, Harokopio University of Athens.
1. Cole TJ. Growth references and standards. Cameron N (ed) Human growth and development. Academic Press, London, 2006, 383-413.
2. de Onis M, Habicht JP. Anthropometric reference data for international use: recommendations from a World Health Organiza-tion Expert Committee. Am J Clin Nutr. 1996; 64: 650-658.
3. Buckler JMH. Children Growth. BMJ 1st ed. Growth disorders in Children. BMJ Publishing Group, London, 1994, 100-197.
4. Tambalis KD, Panagiotakos DB, Kavouras SA, Kallistratos AA, Moraiti IP, Douvis SJ, et al. Eleven-year prevalence Trends of obesity in Greek children: first evidence that prevalence of obesity is levelling off. Obesity (Silver Spring). 2010; 18: 161-166.
5. Newsletter. First Pediatric Clinic of University of Athens. 2003; 50: 136-156.
6. Cole TJ, Freeman JV, Preece MA. Body mass index reference curves for the UK, 1990. Arch Dis Child. 1995; 73: 25-29.
7. Cole TJ, Green PJ. Smoothing reference centile curves: the LMS method and penalized likelihood. Stat Med. 1992; 11: 1305-1319.
8. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ. 2000; 320: 1240–1243.
9. Pan H, Cole TJ. LMS chartmaker, a program to construct growth references using the LMS method. Version 2.43, 2010. Available at: http://www.healthforallchildren.co.uk/, last accessed: 22/7/2014.
10. Pan H, Cole TJ. LMS growth, a Microsoft Excel add-in to access growth references based on the LMS method. Version 2.71, 2011. Available at: http://www.healthforallchildren.co.uk/, last accessed: 22/7/2014.
11. Kuczmarski RJ, Ogden CL, Guo SS, Grummer-Strawn LM, Flegal KM, Mei Z, et al. 2000 CDC Growth Charts for the United States: methods and development. Vital Health Stat 11. 2002; 246: 1-190.
12. Tambalis KD, Panagiotakos DB, Kavouras SA, Papoutsakis S, Sidossis LS. Higher prevalence of obesity in Greek chil-dren living in rural areas despite increased levels of physical activity. J Paediatr Child Health. 2013; 49: 769-774.
13. Ahrens W, Pigeot I, Pohlabeln H, De Henauw S, Lissner L, Molnár D, et al. Prevalence of overweight and obesity in European children below the age of 10. Int J Obes (Lond). 2014; 38 Suppl 2: S99–S107.
14. Farajian P, Panagiotakos DB, Risvas G, Karasouli K, Bountziouka V, Voutzourakis N, et al. Socio-economic and de-mographic determinants of childhood obesity prevalence in Greece: the GRECO (Greek Childhood Obesity) study. Public Health Nutr. 2013; 16: 240–247.
15. Wijnhoven TM, van Raaij JM, Spinelli A, Starc G, Hassapidou M, Spiroski I, et al. WHO European Childhood Obe-sity Surveillance Initiative: body mass index and level of overweight among 6-9-year-old children from school year 2007/2008 to school year 2009/2010. BMC Public Health. 2014; 14: 806-822.
16. Brug J, van Stralen MM, Te Velde SJ, Chinapaw MJ, De Bourdeaudhuij I, Lien N, et al. Differences in weight status and energy-balance related behaviors among schoolchildren across Europe: the ENERGY-project. PLoS One. 2012; 7: e34742.
17. Kułaga Z, Litwin M, Tkaczyk M, Palczewska I, Zajączkowska M, Zwolińska D, et al. Polish 2010 growth references for school-aged children and adolescents. Eur J Pediatr. 2011; 170: 599-609.
18. Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, et al. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2014; 384: 766-781.
19. Kulaga Z, Litwin M, Tkaczyk M, Rózdzyńska A, Barwicka K, Grajda A, et al. The height-, weight-, and BMI-for-age of Polish school-aged children and adolescents relative to international and local growth references. BMC Public Health. 2010; 10: 109.
20. Christoforidis A, Dimitriadou M, Papadopolou E, Stilpnopoulou D, Katzos G, Athanassiou-Metaxa M. Defining overweight and obesity among Greek children living in Thessaloniki: International versus local reference standards. Hip-pokratia. 2011; 15: 141-146.
21. Wright CM, Booth IW, Buckler JM, Cameron N, Cole TJ, Healy MJ et al. Growth reference charts for use in the United Kingdom. Arch Dis Child. 2002; 86: 11-14.