Overweight and thinness frequently coexist. This phenomenon, termed the dual burden of body weight(Reference Doak, Adair, Bentley, Monteiro and Popkin1), occurs even within a single family, and is especially frequent in countries in nutritional transition(Reference Florencio, Ferreira, de Franca, Cavalcante and Sawaya2). However, it is also observed in developed countries such as the USA, where one overweight and one underweight person are simultaneously found in 7·9 % of households(Reference Doak, Adair, Bentley, Monteiro and Popkin1). The dual burden of body weight is of public health importance, because policies targeted at controlling obesity may have undesirable effects on the thinnest individuals.
There is abundant information on the childhood obesity epidemic(Reference Lobstein, Baur and Uauy3), but in developed countries trends in overweight and thinness among children are very seldom described jointly(Reference Wang, Monteiro and Popkin4). To our knowledge, the present study is the first on schoolchildren in a Mediterranean country which simultaneously examines the trends in overweight and thinness in the last decade.
Study design and population
The methodology of the study has been described elsewhere(Reference Martínez Vizcaíno, Salcedo Aguilar, Franquelo Gutiérrez, Torrijos Regidor, Morant Sánchez, Solera Martínez and Rodríguez Artalejo5). In 2004, a cross-sectional study was conducted among 1280 children attending the 4th and 5th grades at twenty-two public schools in the Province of Cuenca. Of these twenty-two schools, nineteen belonged to each of the most populated towns in the province and three were situated in the City of Cuenca. In 1992, 1996 and 1998, similar cross-sectional studies were conducted although only at schools in the City of Cuenca. The schools studied in 1992, 1996 and 1998 were also included in the 2004 study.
The four cross-sectional studies were authorized by the Cuenca Primary Health Care Research Committee. Also, the 2004 study was formally approved by the Clinical Research Ethics Committee of the ‘Virgen de la Luz’ Hospital in Cuenca.
Ethical and legal aspects
The study was approved by the Director and Board of Governors (Consejo Escolar) of each school. In addition, the parents gave written consent to the participation of their children in the study. Informative talks were then held class-by-class, where the schoolchildren were asked to collaborate. After the data were obtained, the parents were informed by letter of their children’s results.
In the four cross-sectional studies, the anthropometric measurements were obtained at the schools by trained and certified nurses. Weight was measured with subjects lightly dressed and in stockinged feet, taking the mean of two measurements using a digital scale with an accuracy of 100 g. Height was obtained as the mean of two determinations with a wall-mounted height rod, with children in stockinged feet, standing straight against the wall so that their spine was vertically aligned with the centre of the height rod. The head was placed with the chin parallel to the floor, and height was measured to the nearest millimetre. BMI was calculated as weight in kilograms divided by the square of the height in metres (kg/m2).
The following three criteria were used to classify the children as thin, overweight (including obesity) and obese: (i) the reference for BMI from the first National Health and Nutrition Examination Survey (NHANES I)(Reference Must, Dallal and Dietz6), which defines thinness, overweight and obesity according to the 5th, 85th and 95th sex- and age-specific percentiles, respectively; (ii) the International Obesity Taskforce (IOTF) criteria(Reference Cole, Bellizzi, Flegal and Dietz7, Reference Cole, Flegal, Nicholls and Jackson8), which provide sex- and age-specific BMI cut-offs based on percentiles passing at age 18 years through BMI 18·5 kg/m2 for thinness, BMI 25 kg/m2 for overweight and BMI 30 kg/m2 for obesity; and (iii) the Centers for Disease Control and Prevention (CDC) growth charts for BMI(Reference Kuczmarski, Ogden, Guo, Grummer-Strawn, Flegal, Mei, Wei, Curtin, Roche and Johnson9), which also define thinness, overweight and obesity based on 5th, 85th and 95th percentiles, respectively. Confidence intervals for frequencies of each BMI category were calculated using the Newcombe–Wilson method(Reference Newcombe10).
Gender comparisons of the frequency of each BMI category were performed with the Mantel-Haenszel χ 2 test. Also, overall frequencies of each category of BMI were standardized by age and sex, using the direct method and the Spanish population aged 9 to 10 years in the 2000 census as standard. Finally, to examine the trend in BMI categories from 1992 to 2004, the χ 2 test for trend was used. Statistical significance was set at P ≤ 0·05 (two-tailed). Analyses were performed with the SPSS for Windows version 15·0 (SPSS Inc., Chicago, IL, USA) and StatsDirect version 2·3·8 (StatsDirect Ltd, Cheshire, UK) statistical software packages.
Of the 1456 schoolchildren aged 9 to 10 years invited in 2004, 1166 (80·1 %) participated in the study. Of these, 609 (52·2 %) were girls and 557 (47·8 %) boys. No statistically significant differences in age were observed between the two sexes. In the 1992, 1996 and 1998 studies, 271 (acceptance rate 88·7 %), 233 (83·6 %) and 249 (84·5 %) schoolchildren participated, respectively. The proportion of boys and girls in each study was approximately 50 %.
Table 1 shows the prevalence of BMI categories (overweight, normal weight and thinness) among schoolchildren in the Province of Cuenca in 2004. Overall prevalence of overweight varied from 30·8 % with the IOTF criterion to 32·9 % with the NHANES I criterion, and was slightly higher in boys than in girls in all three reference criteria. In general, frequency of overweight was higher with the NHANES I and CDC than with the IOTF criteria, although the differences were small. The overall prevalence of thinness varied from 4·4 % with the CDC criterion to 9·3 % with the IOTF criterion. Furthermore, it was greater among children aged 9 years than among those aged 10 years, and in boys aged 9 years and girls aged 10 years. The highest prevalence of thinness was obtained using the IOTF criterion, which was approximately double that obtained with NHANES I and CDC criteria.
NHANES I, reference for BMI from the first National Health and Nutrition Examination SurveyReference Must, Dallal and Dietz(6), which defines thinness, overweight and obesity according to the 5th, 85th and 95th sex- and age-specific percentiles, respectively; IOTF, International Obesity Taskforce criteriaReference Cole, Bellizzi, Flegal and Dietz(7, Reference Cole, Flegal, Nicholls and Jackson8), which provide sex- and age-specific BMI cut-offs based on percentiles passing at age 18 years through BMI 18·5 kg/m2 for thinness, BMI 25 kg/m2 for overweight and BMI 30 kg/m2 for obesity; CDC, Centers for Disease Control and Prevention growth charts for BMIReference Kuczmarski, Ogden, Guo, Grummer-Strawn, Flegal, Mei, Wei, Curtin, Roche and Johnson(9), which also define thinness, overweight and obesity based on 5th, 85th and 95th percentiles, respectively.
Figure 1 depicts the trend in BMI categories over the period 1992–2004. The overall prevalence of overweight went from 24·4 % in 1992 to 30·9 % in 2004 (P = 0·07), increasing from 21·2 % in 1992 to 32·0 % in 2004 (P = 0·03) among boys and from 27·7 % to 29·8 % (P = 0·67) among girls. The overall prevalence of thinness was 2·7 % in 1992 and 9·2 % in 2004 (P < 0·001); in the same period, thinness prevalence went from 1·9 % to 9·0 % (P = 0·10) among boys and from 3·7 % to 9·5 % (P < 0·01) among girls. In 2004, the prevalence of overweight at the three schools in the City of Cuenca where the previous studies had been conducted was 33·7 % in boys and 23·4 % in girls; prevalence of thinness was 8·1 % in boys and 10·3 % in girls.
Our study shows that the prevalence of overweight among schoolchildren in the Province of Cuenca is very high, and has risen substantially in boys in recent years. The prevalence of thinness increased in both sexes during the last decade.
Overweight prevalence in Cuenca schoolchildren in 2004 (30·8 %) ranked among the highest worldwide. Such prevalence is similar to that found in Greece (31 %) and Italy (36 %), higher than in France (19 %) and Germany (16 %), and even higher than in Central and Northern European countries, such as Denmark (15 %)(Reference Lobstein, Baur and Uauy3). It is also in excess of the US figure (25·6 %)(Reference Lobstein, Baur and Uauy3). Regardless of the present level, however, the frequency of overweight has increased among children in most developed countries(Reference Lobstein, Baur and Uauy3).
The frequency of thinness among Cuenca schoolchildren is lower than in countries in nutritional transition, such as Brazil or China(Reference Wang, Monteiro and Popkin4), similar to that in the USA(Reference Wang, Monteiro and Popkin4), France(Reference Rolland-Cachera, Castetbon, Arnault, Bellisle, Romano, Lehingue, Frelut and Hercberg11) and Poland(Reference Matusik, Malecka-Tendera and Klimek12), and far lower than in developing countries(Reference de Onis, Blössner, Borghi, Frongillo and Morris13), although the rates are inevitably based on different definitions of thinness. Forecasts for developed countries as a whole over the period 1990–2015 predict a decline in the frequency of thinness(Reference Rolland-Cachera, Castetbon, Arnault, Bellisle, Romano, Lehingue, Frelut and Hercberg11), probably due to a shift to the right of the entire distribution of BMI associated with the obesity epidemic. However, there are differences between countries. Whereas the frequency of thinness during childhood has been observed to decline in the USA(Reference Wang, Monteiro and Popkin4), it has risen in recent decades in Russia(Reference Wang, Monteiro and Popkin4) and there is also indirect evidence of an increase in France(Reference Rolland-Cachera, Castetbon, Arnault, Bellisle, Romano, Lehingue, Frelut and Hercberg11). Indeed, we also observed this phenomenon in our study. While our results do not inform on the causes of the increase in thinness, we can speculate that this might be due to earlier sexual maturation(Reference Cole14), greater participation in sports, and keener concern about controlling body weight in children.
As in previous studies(Reference Wang, Monteiro and Popkin4, Reference Rolland-Cachera, Castetbon, Arnault, Bellisle, Romano, Lehingue, Frelut and Hercberg11, Reference Flegal, Ogden, Wei, Kuczmarski and Johnson15), only small differences were observed in the frequency of overweight when the three reference criteria were used. Again as in another study(Reference Rolland-Cachera, Castetbon, Arnault, Bellisle, Romano, Lehingue, Frelut and Hercberg11), the NHANES I and CDC criteria yielded similar prevalences of thinness. The IOTF criteria, however, provided a frequency of thinness that was approximately double that of the US criteria. Given that the IOTF thinness criteria were published very recently, we found no other study which used them. The different results on thinness obtained with the IOTF criteria are probably due to differences in the date of data collection, the characteristics of the countries of origin of the data, the design of the criteria, and the smoothing techniques used in the US criteria. Had we used the grade 2 cut-offs (linked to BMI 17 kg/m2) of the IOTF criteria, the prevalence of thinness would have been lower. However, the low number of cases of thinness according to grades 2 and 3 of the IOTF criteria (only one child in 1992 and fifteen children in 2004) precluded a meaningful analyses with these cut-offs.
A correct interpretation of our results requires some methodological comments. Because we only studied children aged 9–10 years in Cuenca, extrapolation of the results to other age and geographic settings should be done with caution. It is of note, however, that our analyses are based on data which are comparable over time in terms of population representativeness and measurement techniques, and are unique in Spain for examining trends in overweight and thinness among children. Furthermore, the prevalence of obesity in Cuenca is similar to that reported by other cross-sectional studies in different Spanish provinces whose data were collected at the same dates as ours(Reference Serra-Majem, Aranceta Bartrina, Pérez-Rodrigo, Ribas-Barba and Delgado-Rubio16).
In conclusion, the frequency of normal weight has declined among children in Cuenca in the last decade, while the dual burden of body weight (overweight and thinness) has risen. Population-based policies addressing childhood obesity, which is the most frequent problem, should not increase the risk of thinness.
We thank the schools, families and children for their enthusiastic participation in the study.
The study was funded mainly by La Consejería de Sanidad de Castilla-La Mancha (grant GC03060-00). Additional funding was obtained from the Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III, Red de Investigación en Actividades Preventivas y de Promoción de Salud (grant RD06/0018/0038).
Conflicts of interest: none declared.
Contributors: V.M.-V. was the principal researcher. V.M.-V. and F.S.A. designed the study together with F.R.-A., who contributed to refine the study design. V.M.-V. and F.S.A. coordinated the data collection. M.S.L. and B.N.P. carried out the measurements in the schools. M.S.M. and P.M.M. performed the statistical analysis, and all authors contributed to interpretation of results. V.M.-V. and F.R.-A. drafted the manuscript, and all authors critically revised it for scientific content and approved the final version.