Introduction

At the beginning of the new millennium, new problems for public health have appeared; one of them is certainly the epidemic of obesity in childhood and adolescence, especially in industrialised countries. The increase in the prevalence of overweight and obesity worldwide has been proceeding at such a high extent that this phenomenon has been called “a crisis in public health” by the World Health Organization (WHO) and International Association for the Study of Obesity (IASO) [1, 2]. Unfortunately, the prevalence of overweight and obesity has not only been increasing in adulthood, but also in childhood and adolescence, and these problems tend to persist until adulthood [3, 4].

Therefore, earlier development of overweight and obesity represents an increased risk factor for premature metabolic and/or cardiovascular diseases in later life [5–7]. In addition, obesity in childhood has even been identified as a major risk factor for premature death in adulthood among American Indians [8] that underlines the life-threatening function of excessive overweight. To point the reasons for overweight and obesity in children and adolescents, genetic factors as well as environmental factors such as the increase of sedentary activities, the decrease in physical activity, hyper-caloric and fast/pre-packaged food, loneliness, social isolation and last but not least psychosocial/familiar problems could be identified [9].

Moreover, the impact of parental migration on the prevalence of overweight and obesity among Central European children and adolescents has been described in previous studies [10–12].

Nowadays, the occurrence of severe and extreme forms of obesity seems to be a major problem in the field of overnutrition. In this regard, an extreme form of obesity, also called morbid obesity, defined by a body mass index (BMI) higher than the 99.5th percentile has been described by the Arbeitsgemeinschaft Adipositas im Kindes- und Jugendalter (AGA) [13]. This category comprises children and adolecents at the highest risk for comorbidities in adulthood or even in childhood [14]. Children and adolescents suffering from extreme obesity urgently need medical observation and therapy starting from non-invasive treatments through to bariatric surgery, which should be performed as ultimate ratio under certain indications and exclusively in centres with extensive experience of such treatment in adults as well as in children [15].

Many national as well as international classifications for the weight status by means of the BMI are available, but they do not include a classification for extreme obesity in children and adolescents such as the International Obesity Task Force (IOTF) cut-off-points by Cole et al. [16] or the WHO growth standards for infants and young children [17].

In the present survey, the prevalence of extreme obesity was assessed for the first time by means of the 99.5th percentile according to the German national reference percentiles by Kromeyer-Hauschild et al. [13, 18]. The aim of this survey was to determine the prevalence of overweight, obesity and extreme obesity in a representative sample of 24,989 Viennese children and adolescents aged from 2 to 16 years and to investigate the possible impact of different mother tongues (as an indicator for migration background) on the prevalence of overweight, obesity and extreme obesity.

Patients, materials and methods

Subjects

Total 24,989 Viennese children which represent 8 % of children and adolescents in Vienna were recruited in the time frame from 2003 to 2004. Out of 24,898, 12,807 (51 %) were boys whereas 12,182 (49 %) were girls. The participating subjects were recruited from all Viennese public schools, some private kindergartens and randomly selected private schools in every city district. Consequently, data on weight and height of all children and adolescents were measured and collected by school physicians and pedagogues. In addition, data on age (in months) and gender were recorded within 1 day and therefore, only attendant children and adolescents could be examined. Two research papers based on subsamples of the present study population have already been published by Dietrich et al. [19] and Tsarmaklis et al. [20]. However, the prevalence of extreme obesity and the relationship to mother tongue in children and adolescents have not been investigated so far.

Anthropometry

Body height and weight were measured by the means of direct anthropometry in order to avoid measurement biases which may result from parent-reported values [21]. Standing height was measured barefoot to the nearest 0.01 m using stadiometer.

Weight was taken to the nearest 0.1 kg in underwear without shoes using a calibrated weighing scale.

Mother tongue

In addition, data on native language from 22,438 children and adolescents in this sample were collected. Although all children and adolescents were mainly resident in Vienna, 12,157 children and adolescents (54.2 %) reported to speak German as their first language, 2,881 (12.8 %) Turkish and 7,400 (33 %) another mother tongue than German or Turkish with a presumably high percentage of Serbocroatian native speakers. Unfortunately, more specific data for native language were not available in the third group due to the retrospective analysis of previously recorded parameters.

Classification of weight status and reference criteria for the definition of normalweight, overweight, obesity and extreme obesity

In this sample, the BMI [kg/m2] was used for the classification of the weight status by using the formula \(\displaystyle\text{BMI}=\frac{\text{body}\ \text{weight}\ [\text{kg}]\text{ }}{\text{(body}\ \text{lenghth}\ {{[\text{m}]}\text{)}^{\text{2}}}}.\)

Reference percentile curves by gender according to Kromeyer-Hauschild et al. [13, 18] based upon a Central European sample from Jena, Germany, were used to define normalweight and underweight (BMI < 90th percentile), overweight (90th–97th percentile), obesity (> 97th percentile) and an extreme form of body fatness called morbid obesity (³ 99.5th percentile).

Due to cultural, geographical and linguistical proximity between Austria and Germany and no availed Austrian BMI reference criteria, these percentile curves were applied for a Viennese sample. Another reason to use Kromeyer-Hauschild criteria is the additional category of morbid obesity (³ 99.5th percentile) that has not been defined by IOTF criteria [16], WHO growth standards [15] or the national reference criteria such as Centers for Disease Control and Prevention (CDC) growth charts [19] in the United States.

All data were entered into SPSS 17.0 (Statistical Package for the Social Sciences) and examined for statistical significance. Mean BMI and standard deviation were calculated for every subgroup according to gender and mother tongue.

Inter-gender differences in BMI were analysed by independent samples, t-tests and such between subgroups according to mother tongue by unifactorial ANOVA.

The difference in weight status distribution according to the criteria of Kromeyer-Hauschild et al. (BMI < P3 extreme underweight, BMI P3–P10 underweight, BMI P10–P90 normal weight, BMI P90–P97 overweight, BMI P97–P99.5 obesity and BMI ³ P99.5 morbid obesity) in different subgroups by gender, age group and native language was analysed for statistical significance by crosstabs (c2-test). Furthermore, the difference in the prevalence of extreme obesity between the subgroups according to gender, age group and mother tongue was compared by crosstabs (c2-test).

Results

Normalweight, overweight and obesity by mother tongue

The distribution of overweight, obesity and extreme obesity was calculated for every subgroup according to mother tongue and gender (Figs. 1 and 2). The highest prevalence of overweight and obesity was found in children and adolescents with Turkish as native language (14.3 % of boys vs. 12.6 % of girls were overweight, 6.3 % in boys vs. 7.5 % were obese), whereas the lowest one could be identified in those with German mother tongue (9.1 % of boys vs. 8.7 % of girls were overweight, 4.2 % of boys vs. 4.2 % of girls were obese).

Fig. 1
figure 1

Prevalence of overweight, obesity and extreme obesity in Viennese boys by native language. P percentile according to German national reference criteria by Kromeyer-Hauschild et al. Dark grey overweight (BMI P90–P97), light grey obesity (BMI P97–P99.5) and grey extreme obesity (BMI ³ P99.5)

Full size image
Fig. 2
figure 2

Prevalence of overweight, obesity and extreme obesity in Viennese girls by native language. P percentile according to German national reference criteria by Kromeyer-Hauschild et al. Dark grey overweight (BMI P90–P97), light grey obesity (BMI P97–P99.5) and grey extreme obesity (BMI ³ P99.5)

Full size image

The results for the prevalence of extreme obesity are described later (see section “Prevalence of Extreme Obesity by Mother Tongue”).

More boys than girls suffered from total overweight (BMI > 90th percentile) in every subgroup according to first language (p = 0.011). Within the same gender, children and adolescents with Turkish as native language showed the highest prevalence of total overweight (23.3 % in boys vs. 22.5 % in girls), followed by children and adolescents with another mother tongue than German or Turkish (20.6 % in boys vs. 17.3 % in girls) and finally German-native-speaking children and adolescents with the lowest prevalence of total overweight (15.1 % in boys vs. 14.6 % in girls). These differences in the prevalence of total overweight within the same gender were highly significant (p ≤ 0.01).

Therefore, the highest prevalence of total overweight (BMI ³ 90th percentile) was detected among boys with Turkish as native language (23.3 %), whereas the lowest one was identified among German-native-speaking girls (14.6 %).

Results for the prevalence of total overweight, underweight and normalweight, overweight, obesity and extreme obesity according to gender and mother tongue are presented in Table 1.

Table 1 Weight status by gender and mother tongue according to BMI reference percentiles by Kromeyer-Hauschild et al.
Full size table

Prevalence of extreme obesity by gender

In total, 2.1 % of all Viennese children and adolescents in the survey were classified extremely obese (BMI ³ 99.5th percentile) according to the Kromeyer-Hauschild criteria.

More boys (2.3 %) than girls (1.9 %) were affected by extreme obesity (p = 0.48). This inter-gender difference was also observed in all subgroups according to mother tongue. The prevalence of extreme obesity by gender and native language is shown in Fig. 3. The largest inter-gender difference was found in the subgroup of children and adolescents with another native language than German or Turkish (3.4 % in boys vs. 2.4 % in girls), whereas the smallest one was observed among this sample with German as first language (1.8 % in boys vs. 1.6 % in girls).

Fig. 3
figure 3

Prevalence of extreme obesity (BMI ³ 99.5th percentile by Kromeyer-Hauschild et al.) according to gender and native language. Dark grey German, light grey Turkish and grey Other

Full size image

Prevalence of extreme obesity by mother tongue

The prevalence of extreme obesity was calculated for every subgroup according to first language (German, Turkish and Other; Fig. 3).

The highest one was found in the subgroup of children and adolescents with another native language than German and Turkish, namely 2.9 % followed by the Turkish-native-speaking subgroup (2.5 %).

The lowest one was found in German-native-speaking children and adolescents (1.7 %). Those remarkable differences in the prevalence of extreme obesity between the subgroups according to mother tongue were statistically highly significant (p = 0.0001) and indicate an increased risk for developing extreme body fatness in children and adolescents who do not speak German as their native language.

When the prevalence of extreme obesity was examined for inter-gender differences within the different subgroups by mother tongue, following results were obtained: Turkish-native-speaking girls showed the highest prevalence (2.4 %) compared with girls with another mother language (German, another native tongue than German and Turkish), whereas the highest prevalence of extreme obesity in boys was found in those with another first language than German or Turkish (3.4 %). The lowest prevalence of overweight (9.1 vs. 8.7 %), obesity (4.2 vs. 4.2 %) and extreme obesity (1.8 vs. 1.6 %) was found in German-native-speaking boys and girls, respectively.

Prevalence of extreme obesity by age group

All data with and without additional information about mother tongue (n = 24,989) were subdivided into five age groups:

  1. 1.

    2–4 years (toddlers)

  2. 2.

    4–7-year-old children

  3. 3.

    7–10-year-old children

  4. 4.

    10–13-year-old adolescents

  5. 5.

    13–16-year-old adolescents

Prevalence of extreme obesity was calculated for every age group and following age-group distribution became evident: The highest prevalence of extreme obesity was identified among the age of 7–10 years (2.5 %), followed by 4–7 years (2.4 %), 2–4 years (1.8 %), 10–13 years (1.3 %) and 13–16 years (1.2 %).

These remarkable differences between the age groups in the prevalence of extreme obesity were statistically significant (p = 0.001; Fig. 4).

Fig. 4
figure 4

Prevalence of extreme obesity (BMI ³ 99.5th percentile by Kromeyer-Hauschild et al.) by age group

Full size image

Discussion

This survey, as far as we know, provides a completely new insight into the prevalence of extreme obesity in childhood and adolescence based on a representative population aged between 2 and 16 years in a Central European city such as Vienna. To our knowledge, extreme obesity has not been described so far in an entire group of 24,989 children and adolescents in Europe. Therefore, we define children and adolescents with a BMI at or above the 99.5th percentile according to the criteria by Kromeyer-Hauschild et al. as morbidly/extremely obese.

In the majority of national and international paediatric reference criteria for BMI, no cut-off-points have been elaborated for the classification of extreme obesity [16, 17, 22]. However, a definition for extreme obesity was provided by Willi et al. [23] with > 200 % of the ideal body weight (IBW) based on the medium frame from the Metropolitan Life Insurance Tables [24]. Furthermore, Freedman et al. [25] considered the 99th percentile of CDC Growth Charts 2000 to be “associated with a greatly increased frequency of biochemical abnormalities” and to constitute “a high predictive value for adult BMI levels of ³ 35 kg/m2”. Children and adolescents with a BMI at or above the 99th percentile were classified as “very obese” which made up approximately 4 % of the US sample.

In our survey, 2.1 % of children and adolescents in this sample were identified as being extremely obese and, therefore, threatened by extreme body fatness, which may increase the risk for metabolic and cardiovascular illnesses dramatically in their later lives [7]. This high number of children and adolescents suffering from extreme obesity illustrates the problematic condition in Vienna and emphasises the need for an intensive multi-disciplinary approach for medical and dietary treatment and, as ultimate ratio, invasive treatments such as bariatric surgery in order to prevent the persistence or even the increase of excessive body fat until adulthood. In addition, the high percentage of physically inactive children [26] as a predisposing factor for obesity and reduced physical fitness [27] could be avoided by fostering ambitious promotion programmes already in the early childhood for a more active, healthier and sporty lifestyle.

For the first time, the influence of native language on the prevalence of extreme obesity has been examined in this survey: Children and adolescents with another mother tongue than German (Turkish, other) and therefore with a migration background showed a remarkably higher prevalence of extreme obesity in both genders compared with German-native-speaking children and adolescents (p = 0.0001). Unfortunately, we could not provide more detailed data according to native language in the third group as mentioned above. However, the majority may be Serbocroatian native speakers according to population-based data provided by “Statistik Austria” [28]. Different socio-economic circumstances [29], eating habits, activity levels [30], parental income and education [31, 32] may be some possible reasons for a higher prevalence of overweight, obesity and extreme obesity in migrant children and adolescents even if the exact reason for this phenomenon could not be identified in this survey. This negative effect of migration status on the prevalence of overweight and obesity in Vienna has already been described by Kirchengast et al. in two separated longitudinal surveys for boys and girls [10, 11]. The prevalence of overweight and obesity was assessed at the age of 6, 10 and 15 years in two identical samples of boys and girls originating from families with a generally low socio-economic status (SES). These samples were divided into three subgroups according to native language: German, Turkish and Serbocroatian. Across all three age groups, the highest prevalence of total overweight was found in boys [10] and girls [11] with Serbocroatian mother tongue, with a maximum of 38.2 vs. 34.6 % at the age of 10 years, respectively. Turkish-native-speaking boys and girls followed with a maximum of 23.7 % at the age of 15 years and 29.9 % at the age of 10 years, respectively. The lowest prevalence of total overweight was found in German native speakers in both genders with a minimum of 17.2 vs. 20.1 % in 6-year-old children. Consistently to our study, a significantly lower prevalence of overweight and obesity could be demonstrated in German-native-speaking children and adolescents. Within the subgroups with positive migration status, our study detected the highest prevalence in the Turkish-native-speaking subgroup in contrast to the findings by Kirchengast et al. However, all three studies agree on the fact that a positive migration background or another native language than German increase the risk for being overweight and obese. Finally, a lower prevalence of overweight and obesity was found in the present sample which might be the result of a rather socially balanced study population in contrast to subjects with a generally low SES in the other two surveys.

Interestingly, marked inter-gender differences in the prevalence of morbid obesity have become evident and indicate a significantly higher risk for extreme obesity in boys. They also show a higher prevalence in overweight and obesity than girls with the same native language (p = 0.011). Our data are consistent with another survey by Popkin et al. [33] about obesity in adolescents with and without migration history in the United States.

The economical burden for the treatment of excessive body weight is remarkable. More than US$ 100 billion have been spent on medical expenses, lost incomes, diet foods, products and programmes to lose weight for the treatment of obesity in adulthood in the United States [9]. It can be expected, if the number of obesity among children and adolescents kept increasing, additional costs would rise for healthcare system and financial problems could emerge. Political, educational and public health measures against the epidemic of overweight and obesity in childhood and adolescence should be undertaken urgently in order to reduce the prevalence of overweight, obesity and extreme obesity.