Comparison of Slovak reference values for anthropometric parameters in children and adolescents with international growth standards: implications for the assessment of overweight and obesity

Aim To compare the national reference percentile values for body height, weight, and body mass index (BMI) of children and adolescents in Slovakia with international standards and to analyze growth trends in this population. Methods The study was designed as a repeated cross-sectional survey. Two nationwide anthropometric surveys (NAS) performed in 2001 and 2011 assessed body weight, height, and BMI of 38 692 children aged 7 to 18 years. Age- and sex-specifıc smoothed percentiles were generated with the lambda-mu-sigma method. Slovak standards were compared with World Health Organization (WHO) 2007 z-scores and International Obesity Task Force (IOTF) standards. Results Medians of body height corresponded to the 75th-85th percentile of the WHO 2007 standards. The secular trend of height increase was attenuated, and the final body height did not change between NAS 2001 and NAS 2011. The cut-off BMI values for obesity, set at the 97th percentile for age <14 years, were higher across age ranges than WHO 2007 standards but lower than IOTF standards. Obesity prevalence, relatively low in 2001 (<3%), doubled during the following decade (P < 0.001), with the highest values (4.8%-7.6%) observed in children aged up to 13 years. Conclusion NAS 2001 data were chosen as national growth standards, as these data were not influenced by the obesity rates increase in the period between the surveys. BMI cut-offs were lower than those in most European countries. Obesity proportions in prepubertal and pubertal boys might be overestimated when WHO 2007 cut-offs are used.

Child and adolescent growth and development are important indicators of nutritional status and health. To identify developmental deviations and impending health problems, it is necessary to have adequate standards and reference values. International and generally accepted standards, such as those set by the World Health Organization (WHO) Child Growth Standards (1), Centers for Disease Control and Prevention (CDC) 2000 (2), and International Obesity Task Force (IOTF) (3) were derived from cross-sectional surveys using representative samples. They enable the comparison of growth patterns of different populations or epidemiological data estimates of malnutrition, overweight, and obesity. While the WHO 2007 classification provides z-scores of body height, weight, and BMI for sex and age (1), IOTF approach bases age-and sex-specific BMI cut-off points for overweight and obesity in children on percentiles analogous to the adult BMI of 25 and 30 kg/m 2 , respectively (3). However, these do not consider growth rate and secular population trends, which is why there is a need for standards that would more objectively assess developmental trends and nutritional status at individual and population levels.
Slovakia had had a sustained tradition of growth studies of children long before WHO recommendations were made. The first nationwide anthropometric survey (NAS) of children and adolescents in former Czechoslovakia was conducted in 1951. Notable differences between Czech and Slovak populations of children and youths observed in the first representative studies have been gradually decreasing (4,5). In the meantime, several studies used growth data to assess the impact of social, environmental, and behavioral factors (6)(7)(8). After the dissolution of Czechoslovakia, transverse representative surveys in Slovakia have continued in 10-year intervals and served as a valid source of reference data for the population aged 0-18 years. They showed secular changes in child and adolescent anthropometric parameters (body height, weight, and circumferences) (9,10).
The latest NAS was carried out in 2011. Since NAS 2001 did not exclude outliers and references took no account of skewness of weight and BMI, we revised and recalculated these standards according to Cole (11). The aim of this study was to provide up-to-date national reference values for school-aged children and adolescents aged up to 18 years in Slovakia. Moreover, we compared our nationspecific references with other EU countries (12)(13)(14)(15) and international standards (1,3), to analyze time trends and compare overweight and obesity prevalence. Our hypothesis was that the latest BMI percentile distribu-tion of school-aged children and adolescents in Slovakia would be substantially modified by the increasing obesity prevalence in industrialized countries (16,17), and that Slovak-specific reference values for both sexes would differ systematically from the international standards across age ranges, as WHO 2007 standards were based on American population assessed 40 years ago, and IOTF standards assessment included only two European populations (British and Dutch) (3) and was based on different principles than WHO and Slovak standards assessments.
PaRtiCiPants and metHods

Participants
The study was designed as a repeated cross-sectional comparative survey using data from two surveys,  (20), which fits smooth centiles to reference data using the lambda-mu-sigma (LMS) method (21). Z-scores, percentile values (3rd, 10th, 25th, 50th, 75th, 90th, and 97th), and curves were set at 0.5 years intervals. They were compared to the z-scores according to WHO 2007 and IOTF grades by LMSgrowth 2.77, a Microsoft Excel add-in to access growth references (22). The same program was used to identify 10-year time trends in anthropometric parameters measured by NAS 2001 and NAS 2011.

statistical analysis
The data are presented as mean values and standard deviations (SD) and percentile values by age and sex. Prevalence rates are expressed as percentages. Differences in mean body height and weight between NAS 2001 and NAS 2011 and between boys and girls were evaluated by ANOVA. The prevalence of overweight and obesity in NAS 2001 and NAS 2011 by sex and age groups was compared using χ 2 test. The level of significance was set at P < 0.01. The statistical analysis was performed using Statgraphic Centurion (STATGRAPHICS® Centurion version XVI, Stat-Point Technologies, Inc., The Plains, VA, USA).

descriptive statistics of basic anthropometric parameters
Distribution of participants included in NAS 2001 and NAS 2011 according to sex and age is presented in Table  1. Boys up to 12 years old grew yearly by an average of 5.3 cm. The growth velocity culminated between 12.5 and 14.5 years, with the height increment of 14.5 cm. Thereafter, the growth slowed down but continued until 18 years. Although growth acceleration in boys was more pronounced in NAS 2011 than in NAS 2001, the final mean height did not change. The median age at voice change in boys was 13 years (intersextile range 12-14 years). The median age at menarche was 12 years (intersextile range 11-13 years). The highest growth velocity in girls was observed from 10 to 12 years, with the increment of 12.8 cm. During this period, girls were significantly taller than boys according to NAS 2001, but not according to NAS 2011 (Table 2 and Table 3).

Comparison of percentile values
Medians of height in our population were by 2.5-3.0 cm higher than WHO 2007 standards, corresponding approximately to z-score 0.6-1 or the 75th-85th percentile (  overweight and obesity definitions The cut-offs for overweight and obesity in Slovakia were traditionally set at the 90th and 97th percentile of BMI, respectively. We compared these cut-offs with WHO 2007 and IOTF definitions expressed as z-scores and IU of BMI ( Figure 2). The cut-offs for overweight in boys were higher (0. The cut-offs for obesity in younger boys matched IOTF definitions. WHO 2007 cut-offs for boys up to 13 years of age were much lower (by 1.0-1.9 IU) than Slovak and IOTF cut-offs. These disparities decreased with age, and   the values from three classifications became similar, corresponding to z-scores of 1.8-2.2 ( Figure 3). Slovak cutoffs for younger girls were similar to WHO 2007 cut-offs, but cut-offs for older girls were lower by 1.5-2.5 IU than WHO 2007 and by 2.0-3.0 IU than IOTF limits (z-score 1.5-1.8) (Figure 3).

obesity prevalence
WHO 2007 obesity cut-offs for boys are set at lower BMI values (by 0.6-0.8 IU) than those for girls. Consequently, obesity prevalence in age groups <13 years was highest when WHO 2007 cut-offs were used ( Figure 4). When Slovak and IOTF cut-offs were used, it yielded similar results in boys, while in girls it was lowest when IOTF standards were used. In children aged up to 11 years, no sex differences were found when either of these cut-offs was used. However, obesity rates according to WHO 2007 standards in boys aged up to 13 years were nearly two times higher than in girls and much higher than IOTF or Slovak standards. In older age groups, they were highest when Slovak standards were used. Obesity prevalence in girls decreased with age, regardless of the standards used.  As Slovak standards (set at 90th and 97th percentile) were derived from NAS 2001 data, the overweight and obesity rates that year nearly met the theoretically expected values of 7% and 3%, respectively. In the following 10 years, these proportions doubled. For cross-country comparisons, we chose IOTF cut-offs ( Figure 5) to evaluate the changes during the 2001-2011 period. In NAS 2011, overweight prevalence was higher (P < 0.001) in all subgroups, but the trend rise for obesity was even steeper, as it more than doubled across all ages. The highest increase in overweight, including obesity, in girls was observed in the age group 10-12 years (by 10%), although in the oldest age group it was only 3.7%. In boys, the prevalence was higher than in girls (P < 0.001), but its age-related increment variations were lower (10%-13%) ( Figure 5).

disCussion
We confirmed our hypothesis that higher weight increments in NAS 2011 were not proportional to growth, which supports the choice of reference values based on the NAS 2001 data.

overweight and obesity definitions
Sex-and age-specific BMI cut-offs based on NAS 2001 were lower than other nation-specific and IOTF standards. Cut-offs derived from 2011 data may be modified by an increase in obesity (14,15,(23)(24)(25) and are relatively high compared with Slovak or Czech references (12). Although our hypothesis about differences between nation-specific reference values and international standards was confirmed, we did not expect such great discrepancies in sex and age, mainly from WHO 2007 cut-offs. These discrepancies cannot be explained only by body height variations and secular trends in respective age groups and the entire population. We have recently pointed out these inconsistencies in a study on 2795 children aged 7 years under the WHO European Childhood Obesity Surveillance Initiative (COSI) project in Slovakia (26). The obesity prevalence in boys in Slovakia and in other countries participating in the COSI study (27) was higher compared to girls only using WHO 2007 standards. Such differences were not demonstrated when IOTF or nation specific references were used. In 8 out of 13 countries, obesity prevalence according to IOTF standards was even higher in girls. The present study showed that this also applied to older age groups.
WHO 2007 standards are derived from data recorded about 40 years ago, so they may have become outdated due to secular trends and growth acceleration. The contemporary school-age children in industrial Europe are by 3-6 cm taller than WHO standards (10,(28)(29)(30)(31)(32), which was also documented by consistently positive mean height z-scores. Taller children systematically showed higher BMI values, while shorter children showed lower median BMI values (30,32).
According to WHO 2007 definitions of overweight and obesity (1), boys under 13 years have BMI lower by 0.8 IU than girls of the same age. For example, a 9.5-year-old boy with the weight of 38.1 kg and height of 135 cm would be classified as obese, but a girl of the same age and height weighing 40 kg would be classified as overweight. Likewise, a girl weighing 34 kg would be classified as normal, but a boy weighing 33.2 kg would be classified as overweight.
According to the WHO classification, the portion of boys up to the age of 13 years who were identified as obese was implausibly higher compared with girls of the same age. According to Slovak national cut-offs, 12 228 boys and 13 724 girls were considered obese, while according to WHO cut-offs, 24 368 boys (two times more) and 14 841 girls (an almost equal number) were considered obese. Such great sex differences were not observed when obesity was assessed according to IOTF.
Since girls aged 11-12 years are only slightly taller and heavier than boys (12,14,15,28,30,31), the question arises if there is a need for stricter BMI limits for prepubertal boys.
In both 2001 and 2011 surveys, the prevalence of overweight, including obesity, in girls started to differ approximately at the age of 13-14 years and lowered with age. Girls start quite early with an effort to remain slender and are at risk of harming their health with inappropriate diet.
In contrast to IOTF standards (3), Slovak BMI-age curves do not pass through the BMI of 30 for obesity at the age of 18, as the 97th percentile represents much lower values, especially in girls. Therefore, they may overestimate the obesity prevalence in older age groups. On the other hand, Slovak BMI references match IOTF standards for children up to 10 years, and so they yield similar obesity rates. When applying IOTF criteria, our results are in good agreement with the recent Organisation for Economic Co-operation and Development statistics, which show a relatively low prevalence of obesity in Slovaks younger than 20 years (16).

obesity prevalence
Previous nationwide anthropometric surveys (4,(7)(8)(9)(10) have shown secular trends in all anthropometric parameters, mainly in body height. Toward the end of the 20th century, the population was getting slimmer (7). In our study, prepubertal and pubertal age groups grew slightly during the first decade of the new millennium, but the final body height (of the oldest age group) remained the same. Slovak girls reached their "adult" body height at 16 years, which is a trend first revealed in 1991 (7,9). At the same time, intensive weight gain was observed, especially among boys, which markedly increased the overweight and obesity proportions. Although overweight and obesity prevalence remain among the lowest in Central and Western European countries (16), the steep obesity rise is worrisome.
Regardless of its limitations as a measure of adiposity in children and adolescents (33)(34)(35)(36)(37), BMI is widely accepted as the simplest tool to unveil global time population trends and detect developmental deviations on an individual level. These goals can best be achieved using nation-specific reference values for the assessment of growth and proportionality of physical characteristics. In Slovak children, the "obesity epidemics" started approximately at the beginning of the new millennium, which is 15-20 years later than in other industrial and economically developed European countries. However, this initially slightly rising trend became more abrupt during the period 2001-2011, regardless of the references used.
When obesity started to rise in Slovakia, it plateaued or declined in many other European countries (38)(39)(40)(41)(42)(43). The NAS in 2021 will show whether Slovakia will face a similar trend. However, the most recent research (26) has shown the culmination of obesity rise approximately 6 years ago.
The strengths of the current study are uniform internationally accepted methodology and the nationally representative samples obtained in both surveys, with relatively high participation rates proportional to the population size in all parts of Slovakia (21). However, the study also has some limitations. Both surveys recruited voluntary participants, meaning that adolescents, especially girls, who refused to be measured might have been more obese. This could affect not only the overall obesity prevalence, which was lower in adolescents than in prepubertal children, but also the cut-off BMI values, which were lower than in most of the European countries in girls older than 16 years. Additionally, compulsory education in Slovakia ends at the age of 16, so the surveys did not include older adolescents who did not continue their education after that age.
The comparison of long-term obesity prevalence trends showed that although obesity prevalence doubled since the NAS 2001, it is still lower than in the majority of European countries. Obesity definition according to WHO 2007, set at much lower BMI values in boys than in girls, may overestimate the prevalence in boys. The revealed differences in the growth rate of Slovak children and adolescents justify the creation of a national reference of anthropometric data. declaration of authorship All authors conceived or designed the study; JH and HJ collected and checked the data and wrote the protocol. VR and ĽŠ analyzed and interpreted the data, and drafted the manuscript. All authors revised and approved the version to be submitted.
term changes in body weight, bmi, and adiposity rebound among children and adolescents in the Czech Republic. econ Hum biol.