This article appeared in a journal published by Elsevier.

The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elseviers archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/authorsrights

Author's personal copy

Obesity Research & Clinical Practice (2013) 7, e218e229

ORIGINAL ARTICLE

Overweight is associated with low hemoglobin levels in adolescent girls

Ursula Viana Bagni a,, Ronir Raggio Luiz b, Gloria Valeria da Veiga a
Department of Nutrition, Federal University of Rio de Janeiro. Av. Carlos Chagas Filho 373, Bloco J, 2 andar, Ilha do Fundo, Rio de Janeiro, RJ - CEP 21941-902, Brazil b Department of Public Health Studies, Federal University of Rio de Janeiro. Prac a Jorge Machado Moreira 100, Cidade Universitria, Ilha do Fundo - Rio de Janeiro, RJ CEP 21941598, Brazil
a

Received 12 September 2011 ; received in revised form 17 December 2011; accepted 29 December 2011

KEYWORDS
Adolescent; Iron deciency anemia; Hemoglobin; Overweight; Body fat

Summary Objective: To verify the prevalence of iron deciency anemia according to sexual maturation stages and its association with overweight as well as excessive body fat in adolescents. Design: A school-based cross-sectional study was performed. Anemia was assessed by measuring the hemoglobin level (Hb). Nutritional status was dened by sex and age specic body mass index (BMI) cutoffs, and body fat (BF) was determined by bioelectrical impedance. Sexual maturation was assessed by breasts/genitalia and pubic hair development stages. Statistical analyses considered the effect of cluster sampling design (classes) and sampling expansion corrected by relative weight. Odds ratio and general linear modeling were used to assess the associations, regarding the value of p < 0.05 for statistical signicance. Setting: Public schools in the Metropolitan area of Rio de Janeiro, Brazil. Subjects: Probabilistic sample of 707 teenagers between 11.0 and 19.9 years old. Results: The prevalence of anemia among the adolescents was 22.8% (95%CI 16.730.2%), higher among girls than among boys (30.9% vs. 10.9%; p < 0.01). The chance of developing anemia did not change with the nutritional status according BMI or BF percentage, however, overweight girls presented lower Hb levels than those who were not overweight (12.2 g/dL vs. 12.8 g/dL, p < 0.01). In boys this association was not observed. Sexual maturation did not change the association of Hb and anemia with overweight and excessive body fat. Conclusion: The reduction of Hb levels points at overweight as a risk factor for the development of iron deciency among adolescents. 2012 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.

Corresponding author at: Instituto de Nutric o Josu de Castro, Centro de Cincias da Sade, Universidade Federal do Rio de Janeiro. Av. Carlos Chagas Filho 373, Bloco J, 2 andar, Ilha do Fundo, Rio de Janeiro, RJ - CEP 21941-902, Brazil. Tel.: +55 21 2562 6432. E-mail addresses: ursulaviana@gmail.com (U.V. Bagni), ronir@iesc.ufrj.br (R.R. Luiz), gvveiga@globo.com (G.V.d. Veiga).

1871-403X/$ see front matter 2012 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.

doi:10.1016/j.orcp.2011.12.004

Author's personal copy

Overweight and low hemoglobin levels e219 42.6% in northeastern [28] and northern cities of the country [29]. Only a few studies developed in Brazil investigated the problems related to the occurrence of iron anemia specically during adolescence, such as the inuence of sexual maturation stages, the occurrence of menarche, and the variations of the menstrual cycle. Besides, the results are still extremely controversial. Although a high prevalence of anemia and a low hemoglobin concentration during the growth spurt have already been identied among Brazilian boys [30] and girls [3032], some researchers did not nd any association [33,34]. Most of the studies addressing nutritional status [10,28,3537] focused on the association of anemia with malnutrition, and used only anthropometric indices for nutritional assessment. The association of anemia with obesity has not been deeply investigated, however, the current epidemiological scenario requires new investigations regarding this aspect, as obesity and iron deciency anemia have reached epidemic proportions around the world [1,38], including Brazil [25,39], and both can damage growth and development during adolescence. In this context, it is believed that the body composition assessment, in conjunction with anthropometric indices, could help clarify such an association. Thus, the current study was developed to reach two main objectives: (a) to describe the prevalence of iron deciency anemia and its distribution according to sexual maturation stages and the variables related to the menstrual cycle; and (b) to verify the association of anemia with overweight and body fat in adolescent students from public schools in the Metropolitan area of Rio de Janeiro, Brazil.

Introduction
Iron deciency anemia is the most prevalent nutritional disorder around the world, with pregnant women and children as the main risk groups [1,2]. Adolescents, however, also have a high vulnerability to anemia [3]. In this age group, anemia can result in impaired immune function and cause higher vulnerability to infections; reduction of physical resistance and tolerance to efforts, as well as damages to growth and intellectual capacity, hence, difculty to concentrate and memorize, which can lead to negative results in the learning development and school performance [2,4,5]. Anemia affects approximately 30% to 55% of adolescents of all over the world [6], and the vulnerability in this age group has been traditionally associated with the intense physical modications that happen during puberty, which impose an increasing need for energy, macronutrients, vitamins, and minerals, especially during the growth spurt [3]. The iron requirement increases more than twofold in boys and girls during the sexual maturation, due to the total blood volume increase, total body mass gain, myoglobin synthesis and muscle mass gain, respiratory capacity increase, as well as menstrual losses [3]. Inappropriate food habits also can be related to the occurrence of iron deciency anemia among adolescents. Habits frequently observed in this group, such as changing the principal meals for snacks, skipping breakfast, reducing the intake of fruits and vegetables, and the increasing consumption of sodas and high calorie foods [79], not only can result in of iron deciency due to insufcient consumption of its sources [10], but can also increase the risk of chronic diseases such as obesity [11,12], which has been reported as a risk factor for iron deciency anemia independently of food consumption [1315]. Many studies have veried a higher proportion of iron deciency and anemia among overweight adults and elders [13,14,1618], and also among overweight children and adolescents [13,1921], and such fact has been attributed to the mineral homeostasis damage caused by increased inammatory activity in the adipose tissue of overweight people [13,2224]. In Brazil, research on the prevalence of iron deciency anemia in adolescents, as well as its associated factors, are still scarce, and the investigation focusing on this part of the population is not yet included in large epidemiologic studies and national surveys [25]. However, a high prevalence of the problem has been observed in local studies with adolescents, reaching 24.4% and 43.0% in southeast [26] and southern [27] cities, as well as 35.7% and

Methods
A probabilistic sample including adolescent students aged 1119.9 years, from 5th grade of a public elementary school to 3rd year of high school, in the city of Niteri, Southeastern Brazil, was assessed in a cross-sectional study. The students were enrolled in 13 out of 34 public state schools who took part in a larger research, in which the main objective was to assess the evolution of overweight and obesity in teenagers by comparing two cross-sectional studies made between the years 2003 and 2008/2009. In Brazil, public schools are usually attended by students with low socioeconomic status. Niteri is a city located in the metropolitan east side of Rio de Janeiro, and it is approximately 15 km

Author's personal copy

e220 away from the state capital. It encompasses an area about 130,000 km2 , and is the sixth most populous city of the state, with approximately 480,000 inhabitants. The data analyzed in this article were obtained in the second study, collected between May 2008 and April 2009. The sampling size was estimated in 780 students distributed among 34 classes (with about 23 students each), taking into account 30% of non-response, as observed in the pilot study. With this sampling size, it would be possible to estimate 35% of anemia prevalence [34], at a 95% condence interval with 5% precision, and considering the cluster sampling in one stage (random classes). All adolescents in the established age group with regular attendance to classes were invited to take part in the study. The participation was voluntary, and only those who had consent forms signed by their parents were allowed to participate. The physically impaired adolescents who were unable to take part in the anthropometric assessment, pregnant teenagers, and the ones who reported, in the interview right before the blood test, having history of diseases leading to anemia not caused by iron deciency, or having iron accumulation diseases or capillary fragility and/or coagulation abnormality were considered ineligible. The presence of anemia was assessed by measuring blood hemoglobin (Hb) concentration, using a portable -hemoglobinometer (HemoCue ). Adolescents between 10 and 11 years old with Hb < 11.5 g/dL, between 12 and 14 years old with Hb < 12.0 g/dL, boys over 15 years old with Hb < 13.0 g/dL, and girls over 15 years old with Hb < 12.0 g/dL were considered anemic [2]. Weight was measured with an electronic platform scale (Kratos ), that has a 150 kg capacity and a 50 g resolution. Height was assessed using a portable anthropometer manufactured in Brazil (Alturexata ), which has a 213 cm amplitude and a 0,1 cm resolution. The height was measured twice; a 0.5 cm maximum variation was allowed between the two measures, and then the mean value was calculated. The adolescents wore light clothes, with no shoes on. Body mass index was calculated (BMI = weight/height2 ), and the nutritional status was classied according to the criteria of the World Health Organization [40], based on the following cutoff points: <2 Z-scores (underweight), from 2 to 1 Z-score (normal weight), >1 to 2 Zscores (overweight) and >2 Z-scores (obesity). Body composition was assessed by bioelectrical impedance analysis (RJL System , model 101Q), and the fat free mass was estimated by using the equation validated by Houtkooper et al. [41] for U.V. Bagni et al. adolescents from 10 to 19 years old. Body fat (BF) mass (kg) was calculated as the difference between weight (kg) and fat free (FF) mass (kg). The body fat percentage (BF%) was obtained, regarding as excess the values 25% for boys and 30% for girls, which were related to morbidity associated with BF excess in teenagers [42]. Sexual maturation stage was investigated, focusing on the development of breasts, genitalia, and pubic hair, in accordance with the criteria of Tanner [43], and using the self-evaluation technique validated by Saito [44]. Both measurements were included in the analysis, once the breasts and genitalia reect the activity of the hypothalamic-pituitary-gonadal axis, while the development of pubic hair is associated with activity of both the hypothalamic-pituitary-gonadal and hypothalamic-pituitary-adrenal axis. For both sexes, the prepubertal period was classied as Stage 1; the beginning of the growth spurt as Stages 2 and 3 for boys and Stage 2 for girls; the peak of growth velocity as Stage 4 for boys and 3 for girls; and growth deceleration as Stage 5 for boys and Stages 4 and 5 for girls. For the girls, the age of menarche occurrence, as well as the last menstruation date, was also obtained. The current use of supplements with iron, current smoking habits, current and/or past diagnosis of parasitic diseases and information related to the menstrual cycle were investigated using a questionnaire previously completed by the adolescents themselves. Statistical analyses were made regarding the cluster sampling design and the sampling expansion corrected by the relative weight, using the commands for complex samples (Complex Samples) of the software Statistical Package for the Social Sciences SPSS version 19.0. For the association analyses, general linear model procedures were used for continuous variables; and logistic regression was performed for categorical variables. All the analyses were stratied by sex, due to the recognized heterogeneity of anemia prevalence and the associated factors between boys and girls. An exploratory analysis of the dependent variables, hemoglobin and anemia, and the independent variables, nutritional status (underweight; adequate weight; overweight; obesity) and BF% classication (without excess; with excess), was conducted. An analysis of the covariates sex, age group (from 11 to 13.9 years old; 14 to 16.9 years old; 17 to 19.9 years old), sexual maturation stage (beginning of growth spurt; peak of growth velocity; growth deceleration), use of iron supplementation, current use of cigarettes, current and/or past diagnosis of parasitic diseases, menarche occurrence,

Author's personal copy

Overweight and low hemoglobin levels menarche date (less than 2 years before the assessment; 2 or more years before the assessment), and last menstruation date (less than 7 days before the assessment; from 7 to 15 days before the assessment; more than 15 days before the assessment) was also performed, presenting the prevalence and 95% condence interval (95%CI) for categorical variables and means and standard error (SE) for continuous variables. The strength of the association between anemia and the exposure variables and co-variables was measured by odds ratio (OR) with its respective 95%CI, analyzing as dummy the categorical nonbinary variables. Multivariate analysis were made in order to investigate the association of both anemia and blood Hb concentration with nutritional status evaluated by BMI (not overweight; overweight) and BF% (without excess; with excess), adjusted for the effect of age and sexual maturation stages (both pubic hair and breasts/genitalia), as they are potential confounders of these associations, regardless of the statistical signicance in the bivariate analysis. We also adjusted for other possible confounding variables (FF mass, use of iron supplementation, current use of cigarettes, current and/or past diagnosis of parasitic diseases) that presented p < 0.20 in bivariate analysis. The multivariate analysis was made using the backward selection procedure, remaining in the model the variables age and sexual maturation stages (pubic hair, breasts/genitalia) regardless of their statistical signicance, and the possible confounding variables that demonstrated p < 0.05. For the covariates that demonstrated p < 0.05 in multivariate analysis, we tested the existence of interaction with nutritional status. The correlation of Hb concentration (g/dL) with the following variables was veried: age (years), BMI for age (Z-score), FF mass (kg), BF mass (kg), BF%, age at menarche period (years), date of the last menstruation (days), with a p-value < 0.05 indicating statistical signicance. The research was approved by the Committee for Ethics in Research of the Instituto de Pediatria e Puericultura Martago Gesteira of the Federal University of Rio de Janeiro (protocol number 27/08). e221 experiencing the growth deceleration (Table 1). The menarche had already happened among 94.1% (95%CI: 86.797.5%) of the girls, at the mean age of 12.0 years (95%CI: 11.912.2 years). In 82.9% (95%CI: 71.890.2%) of these participants, menarche occurred 2 or more years before the assessment date. The beginning of the last menstruation had happened in 51.4% (95%CI: 44.957.9%) of the teenagers 2 or more weeks before the assessment date, although among the other girls, it occurred in the week before (22.7%; 95%CI:19.326.6%) or during the week of the study assessment (25.9%; 95%CI: 20.132.6%). The nutritional status according to BMI was similar in both sexes, once its veried that the proportion of overweight being much higher than that of underweight in both boys (20.4% and 2.3%) and girls (21.8% and 0.9%). Due to the low prevalence of underweight (1.5%; 95%CI: 0.82.7%) and obesity (4.5%; 95%CI: 2.77.6%) in the population assessed, the nutritional status was presented in the remaining analyses as a binary variable (not overweight; overweight). Regarding BF%, girls presented a higher proportion of excessive body fat when compared to boys (29.6% vs. 10.1%; p < 0.01), whereas the FF mass showed no differences regarding sex (Table 1). Iron deciency anemia was present in 22.8% (95%CI: 16.730.2%) of the adolescents. The girls had a three-times-higher chance of developing anemia than the boys (OR = 3.64 95%CI 2.156.22), and both boys and girls from 14 to 16.9 years old had more chance of having have anemia. Among males, the chance of developing anemia was six times higher in boys experiencing the peak of growth velocity according to their pubic hair, when compared to those already experiencing growth deceleration (OR = 6.43 95%CI 1.2533.38) (Table 2). The mean Hb concentration was 13.1 g/dL (95%IC: 12.813.3), with minimum and maximum concentrations of 9.0 g/dL and 17.9 g/dL, respectively. The Hb was lower in girls than in boys (12.5 g/dL vs. 14.0 g/dL, p < 0.01), and, although it remained stable in all age groups among females, it presented a signicant increase with aging among males. Overweight girls presented lower Hb concentrations in comparison with the girls who were not overweight (12.2 vs. 12.6 g/dL, p < 0.01). As for the males, a tendency to lower Hb values was observed among the ones who were overweight and with excessive BF%. However, such a difference did not seem to be signicant (Table 3). In the multivariate analysis for boys, the association of both Hb and anemia with nutritional status according to BMI and BF% did not change and the

Results
From the total number of 707 participants, 59.3% (95%CI: 53.465.0%) were girls and a greater proportion of the adolescents were about 1416.9 years old. Most of the boys were experiencing the growth spurt and most of the girls were

Author's personal copy

e222 U.V. Bagni et al.

Table 1 Characteristics of the adolescents regarding the biological and health variables and nutritional status Niteri, Rio de Janeiro, Brazil, 2008/2009 (n = 707). Boys (n = 288) n Age 11 to 13.9 years old 14 to 16.9 years old 17 to 19.9 years old Sexual maturation stage (breasts/genitalia) Beginning of growth spurt Peak of growth spurt Growth deceleration Sexual maturation stage (pubic hair) Beginning of growth spurt Peak of growth spurt Growth deceleration Anemia Yes No Nutritional status Underweight Normal Weight Overweight Obesity Body fat percentage Without excess With excess Free fat mass <85th Percentile 85th Percentile Use of iron supplementation Yes No Use of cigarretes Yes No Diagnostic of parasitic diseases Yes No n Hemoglobin (g/dL) Body mass index - BMI (kg/m2 ) BMI-for-age z-score Free fat mass (kg) Fat mass (kg) Body fat percentage (%) 288 288 288 263 263 263 288 55 150 83 288 113 153 22 287 79 144 64 288 32 256 288 7 222 46 12 263 237 26 263 227 36 288 22 265 287 24 263 283 106 177 % 19.0 52.3 28.8 95%CI 9.334.8 41.363.0 17.842.9 Girls (n = 419) n 419 103 185 131 416 16 185 215 418 17 87 314 419 129 290 418 4 323 72 20 411 289 122 411 367 44 417 25 392 419 26 393 413 182 231 n 419 418 418 411 411 411 % 24.5 44.2 31.2 95%CI 13.440.6 32.856.3 19.546.0 0.36 p*

39.3 53.1 7.6

32.446.6 45.760.4 4.911.7

3.9 44.4 51.7

1.211.7 38.550.5 44.858.5

<0.01

27.5 50.4 22.1 10.9 89.1 2.3 77.3 16.1 4.3 89.9 10.1 86.2 13.8 7.8 92.2 8.2 91.8 37.6 62.4 Mean 12.5 20.6 0.05 49.5 9.7 15.7

18.239.3 42.857.9 16.129.6 6.418.1 81.993.6 1.14.7 70.583.0 11.422.2 1.99.2 83.594,1 5.916.5 78.891.3 8.721.2 5.211.6 88.494.8 4.115.6 84.495.9 27.848.5 51.572.2 95%CI 12.312.7 19.921.3 0.250.15 46.752.3 8.710.7 14.517.0

4.0 20.9 75.1 30.9 69.1 0.9 77.2 17.1 4.7 70.4 29.6 89.3 10.7 6.0 94.0 6.2 93.8 44.1 55.9 Mean 14.0 21.2 0.14 39.6 14.6 26.1

2.17.7 15,028.4 37.152.1 23.040.1 59,3 0.42.2 72.281.6 13.321.7 2.68.6 63.676.3 23.736.4 84.092.9 7.116.0 4.09.0 91.096.0 3.510.8 89.296.5 37.551.0 49.062.5 95%CI 13.714.4 20.721.7 0.020,26 38.640.6 13.615.6 25.027.3

<0.01

<0.01

0.51

<0.01

0.22

0.38

0.33

0.29 p* <0.01 0.02 0.06 <0.01 <0.01 <0.01

95%CI, 95% condence interval. * Comparison between boys and girls.

chance of having low hemoglobin levels or presenting anemia did not differ between overweight and not overweight and between with or without excess body fat adolescents. In girls, the association between Hb and nutritional status according

to BMI remained signicant, and the adjustment for the covariates did not change the strength of the association of both Hb and anemia with BF%. We also observed interaction of nutritional status according to BMI with the sexual maturation stage,

Author's personal copy

Overweight and low hemoglobin levels e223

Table 2 Odds ratio (OR) and 95% condence interval (95%CI) of anemia according to biological and health variables, and nutritional status Niteri, Rio de Janeiro, Brazil, 2008/2009. Boys % Nutritional status Not overweight 11.0 Overweight 10.6 Body fat percentage Without excess 10.8 With excess 9.1 Age 11 to 13.9 years. 7.0 14 to 16.9 years. 15.0 17 to 19.9 years. 6.1 Sexual maturation stage (breasts/genitalia) Beginning of growth spurt 10.5 Peak of growth spurt 12.4 Growth deceleration 2.8 Sexual maturation stage (pubic hair) Beginning of growth spurt 8.7 Peak of growth spurt 15.7 Growth deceleration 2.8 Menarche occurence Yes No Menarche date <2 years before the assessment 2 years before the assessment Last menstruation date <7 days before the assessment 7 to15 bays before the assessment 15 days before the assessment Free fat mass < 85th Percentile 11.8 85th Percentile 3.2 Use of iron supplementation Yes 8.1 No 11.1 Use of cigarretes Yes 12.4 No 10.8 Diagnostic of parasitic diseases Yes 13.5 No 9.6
*

Girls Crude OR 1.00 0.96 1.00 0.82 1.16 2.69 1.00 4.98 4.11 1.00 3.26 6.43 1.00 3.99 1.00 1.00 1.42 1.17 1.00 1.47 1.00 95%CI % 29.5 35.6 32.5 26.9 31.1 36.9 22.2 0.0 33.2 31.6 49.3 22.8 32.3 31.7 17.8 32.7 31.5 31.9 30.4 30.7 30.0 31.0 51.8 29.7 46.4 29.9 31.4 31.3 Crude OR 1.00 1.32 1.00 0.77 1.58 2.04 95%CI

0.392.41

0.722.43

0.213.30 0.284.76 1.046.93

0.441.33 0.623.99 1.073.91

0.6339.71 0.6647.62

1.08 1.00 2.04 0.62 1.00 2.14 1.00 1.00 0.94 1.06 0.99 1.00 1.05 1.00 1.00 0.39 2.03 1.00 1.01 1.00

1.081.08

0.7614.05* 1.2533.38

0.508.37* 0.351.11

0.528.85

0.342.58 0.552.05 0.541.82

0.8019.99*

0.333.31

0.326.23 0.255.52

0.121.30* 0.835.01*

0.772.78

0.681.47

p < 0.20.

suggesting that the association of nutritional status with Hb depends on the stage of sexual maturation stage (breasts) in which the teenager is (Table 4). Collectively, the Hb concentration showed a positive correlation with the FF mass (r = 0.36; p < 0.01); and negative correlations with the BMIfor-age Z-score (r = 0.11, p = 0.02), BF% (r = 0.31; p < 0.01) and BF mass (r = 0.18; p < 0.01). However, when the analysis was stratied by sex, it was veried that the Hb did not present any signicant correlation with these variables among females. As

for the males, positive correlations were veried only with FF mass (r = 0.30; p < 0.01). In Fig. 1, the correlations of Hb concentrations with the variables of nutritional status and body composition according to sex are presented.

Discussion
The prevalence of anemia found in this study (22.8%) can be classied as a moderate public

Author's personal copy

e224 U.V. Bagni et al.

Table 3 Mean and 95% condence interval (95%CI) of hemoglobin concentration to biological, health and nutritional status Niteri, Rio de Janeiro, Brazil, 2008/2009. Boys Crude mean (g/dL) Nutritional Status Without overweight 14.1 13.8 With Overweight Body fat percentage Without excess 14.1 With excess 13.5 Age 11 to 13.9 years. 13.3c 14.0b 14 to 16.9 years. 17 to 19.9 years. 14.5a Sexual maturation stage (breasts/genitalia) Beginning of growth spurt 14.0 14.1 Peak of growth spurt 14.0 Growth deceleration Sexual maturation stage (pubic hair) 13.6b Beginning of growth spurt Peak of growth spurt 14.1a Growth deceleration 14.4a Menarche occurence Yes No Menarche date <2 years before the assessment 2 years before the assessment Last menstruation date <7 days before the assessment 7 to15 bays before the assessment 15 days before the assessment Free fat mass < 85th Percentile 13.9 85th Percentile 14.4 Use of iron supplementation Yes 13.9 No 14.0 Use of cigarretes Yes 14.2 No 14.0 Diagnostic of parasitic diseases Yes 13.7 No 14.1
*

Girls 95%CI p* Crude mean (g/dL) 12.6a 12.2b 12.5 12.6 12.5 12.4 12.7 13.5a 12.4b 12.5b 12.2 12.6 12.5 12.5 12.8 12.4 12.5 12.3 12.5 12.5 12.5 12.4 12.1 12.5 12.3 12.5 12.5 12.4 95%CI p*

13.714.5 14.514.2 13.714.5 12.914.0 12.913.6 13.614.4 14.314.8 13.614.4 13.714.5 13.514.5 13.313.9 13.614.5 14.014.8 13.614.3 14.114.7 13.114.6 13.714.4 13.714.7 13.614.4 13.214.2 13.814.5

0.31

12.312.8 11.912.5 12.312.7 12.212.9 12.013.0 12.112.6 12.412.9 12.814.1 12.112.7 12.312.6 11.512.8 12.213.0 12.312.7 12.312.7 12.013.7 11.713.0 12.312.6 12.112.5 12.112.8 12.312.8 12.312.7 12.012.8 11.512.6 12.312.7 11.812.7 12.312.7 12.212.8 12.312.6

<0.01

0.07

0.53

<0.01

0.09

0.83

0.02

<0.01

0.26

0.33

0.80

0.17

0.02

0.66

0.53

0.11

0.34

0.29

0.06

0.75

Comparison between categories, where a > b > c.

health severity problem according to the World Health Organization [2]. Although there are recent publications on the anemia in Brazilian adolescents [10,2637,45,46], in which the prevalence varies from 4.6% in the city of So Paulo, found in the southeastern region of the country [37] to 41.3% in Londrina, a city located in the southern Brazil [36], such studies present a high level of methodological heterogeneity, especially in relation to the age group

involved (many included non-adolescents in the analysis), study design, place where the teenagers were studied, number of participants and their representativeness in the population studied, as well as instruments, techniques and criteria for anemia diagnosis, which make a comparison with the results obtained in this study impossible. In the investigation of the sexual maturation variables that could be potentially associated with anemia in adolescence, it was identied that

Author's personal copy

Overweight and low hemoglobin levels

Table 4 Multivariate analysis for association of anemia (adjusted odds ratio and 95% condence interval) and hemoglobin concentration (adjusted mean and 95% CI) with nutritional status evaluated by body mass index (model 1) and by % body fat (model 2). Niteri, Rio de Janeiro, Brazil, 2008/2009. Model 1 Anemia Boys OR Nutritional status* Not overweight 1.00 0.84 Overweight Age 11 to 13.9 years. 1.12 14 to 16.9 years. 2.23 17 to 19.9 years. 1.00 Sexual maturation stage (breasts/genitalia) Beginning of growth spurt 2.35 Peak of growth spurt 2.87 1.00 Growth deceleration Sexual maturation stage (pubic hair) Beginning of growth spurt 2.97 Peak of growth spurt 5.08 Growth deceleration 1.00 Model 2 OR 95%CI Girls OR 1.00 1.23 2.25 2.06 1.00 1.01 1.00 5.88 0.57 1.00 OR 1.00 0.77 2.16 1.98 1.00 0.93 1.00 5.84 0.55 1.00 95%CI Hemoglobin concentration (g/dL) Boys Mean 13.9 13.8 13.2 13.9 14.4 14.1 13.9 13.6 13.7 13.8 14.0 Mean 13.9 13.6 13.1 13.8 14.3 14.0 13.7 13.4 13.5 13.8 13.9 95%CI 13.614.2 13.414.1 12.913.6 13.614.2 14.114.7 13.714.5 13.714.1 13.114.1 13.314.0 13.514.1 13.714.4 95%CI 13.614.2 13.114.1 12.713.5 13.414.2 13.814.7 13.714.4 13.413.9 12.914.0 13.514.3 13.414.1 13.514.3 p 0.65 Girls Mean 12.8 12.2 12.4 12.4 12.7 13.1 12.2 12.2 12.1 12.8 12.7 Mean 12.7 12.8 12.6 12.6 12.9 13.6 12.2 12.3 12.2 13.0 12.9 95%CI 12.513.1 11.912.4 11.812.6 12.212.7 12.413.0 12.613.7 11.812.6 11.912.5 11.612.6 12.413.1 12.412.9 95%CI 12.512.9 12.413.1 12.213.0 12.412.9 12.613.2 13.014.2 11.912.7 12.012.5 11.812.7 12.613.4 12.613.2 p <0.01

0.282.54 0.304.12 0.836.00

0.662.30 0.905.62 1.044.08

<0.01

0.14

0.2027.06 0.3027.69

0.432.35

0.29

0.02

0.6014.61 0.8331.00

1.5921.76 0.321.03

0.34

0.14

95%CI

95%CI

p 0.25

p 0.56

Body fat percentage* Without excess 1.00 With excess 1.93 Age 1113.9 years 0.96 1416.9 years 1.90 1719.9 years 1.00 Sexual maturation stage (breasts/genitalia) Beginning of growth spurt 2.03 Peak of growth spurt 2.86 Growth deceleration 1.00 Sexual maturation stage (pubic hair) Beginning of growth spurt 3.37 Peak of growth spurt 3.35 Growth deceleration 1.00

0.204.30 0.273.42 0.605.96

0.421.39 0.885.33 0.993.93

<0.01

0.14

0.1724.48 0.2928.44

0.930.93

0.18

0.003

0.6916.59 0.7524.41

1.5921.54 0.310.99

0.46

0.09

CI: condence interval. OR: odds ratio. * Adjusted for age and sexual maturation stage (both breasts/genitalia and pubic hair).

e225

Also adjusted for the interaction of nutritional status with sexual maturation stage (breasts) (p = 0.006).

Author's personal copy

e226 U.V. Bagni et al.

Figure 1 Correlation of hemoglobin concentration with body mass index, and with body fat percentage Niteri, Rio de Janeiro, Brazil, 2008/2009.

during the peak of growth velocity, on the basis of the assessment of pubic hair development, boys presented higher chances of having anemia, a tendency also observed by Silva et al. [30] and Iuliano et al. [31]. In girls, the reduction of Hb concentration that begins with the peak of growth velocity according based on the assessment of breast development corroborates with the increase of vulnerability to the disease in puberty, as proven by other authors [3032]. The hypothesis that the occurrence of menarche as well as the onset of the menstrual cycles would be associated with a higher prevalence of anemia was not, however, conrmed in this study. It is believed that, because only 6% of the girls still had not experienced their rst menstruation, this association became weakened. In addition, the power of the sample to investigate the association between anemia and menarche was low, which impaired statistical signicance for this association even though the chance of anemia

among postmenarcheal girls was 2.4 times higher than among those who had not yet had their rst menstruation. The fact that overweight girls presented signicantly lower Hb concentrations in comparison with those who were not overweight, is in agreement with the results observed by Eftekhari et al. [47] among Iranian girls from 13 to 20 years old, in which the Hb concentrations reduced signicantly from 13.1 to 12.5 g/dL when the BMI classication changed from normal to overweight (BMI for age 95th percentile). An explanation for the relation between overweight and the impairment in iron status would be the increase of inammatory activity in the adipose tissue of obese individuals, which would lead to a higher production of hepcidin, a key hormone for the regulation of intestinal absorption and iron homeostasis [2224]. Inammatory cytokines, particularly interleukin-6, act directly

Author's personal copy

Overweight and low hemoglobin levels in the hepatocytes, stimulating the expression of the hormone, which, when in contact with the bloodstream, binds to the cellular ferroportin, an iron exporter transmembrane protein. This binding induces the ferroportin internalization and degradation by the lysosomes, and with the loss of this transporter, iron export is blocked and the degraded iron is retained within the macrophage cell, favoring the reduction of circulating iron [2224]. In the enterocytes, the ferroportin internalization and degradation blocks the transfer of iron from the cytoplasm to the plasma transferrin [23], and such iron retention inhibits the apical uptake and the utilization of the dietary iron by intestinal cells [22], further increasing the effect over the concentration of the circulating mineral. Thus, even when the iron intake is adequate, individuals who were overweight and with excessive body fat have higher predisposition to develop iron deciency anemia, because there is lower utilization of the mineral in the body, as evidenced by several authors [1315,48,49]. Although, in this study, the change in nutritional status from not overweight to overweight did not increase signicantly the adolescents chance of developing anemia, the reduction in the Hb concentration observed, especially among girls, pointed at the important fact that the iron homeostasis may have been damaged among the individuals with higher a BMI. The impairment in iron status, whether the youngster present anemia or not, is cause for concern and can bring negative repercussions during adolescence, such as poor cognitive development, disruptions in learning and bad school performance [2,5,37,50,51]. A limitation of this study was that the sample was calculated to assess the prevalence of anemia, resulting in a reduction in the power of the study to investigate the association of anemia and hemoglobin with the exposure variables. However, despite this limitation, it was possible to show that overweight was associated with low hemoglobin concentration in girls, indicating that if the sample had been larger, stronger and signicant associations could have been detected with other variables, such as body fat percentage in boys, for which p value was borderline (Table 3). Other factor that can be related to the weak associations and correlations observed in this study concerning the impact of overweight in hemoglobin concentration and anemia prevalence was the reduced number of adolescents with obesity (4.5% of the total sample). The impairment of iron metabolism seems to be more evident when the level of overweight is higher, as noticed by Eftekhari et al. [47] in Iranian girls, in which the prevalence of e227 anemia between girls with weight between the 85th and the 95th percentiles of BMI for age (28.8%) was similar to the normal weight ones (27.8%), whereas those over the 95th percentile presented more anemia than the rest of the group (34.1%). Because the BMI has limitations in diagnosing obesity in teenagers (it does not adequately describe the variations of body composition during growth), in this study, we assessed the adolescents body composition by bioelectrical impedance, aiming to identify the effect of BF% excess in the prevalence of anemia and the concentration of Hb. However, no association or correlation was observed. A possible explanation would be that the Hb concentration, the parameter used in this study, is generally affected only at the nal stage of iron deciency [2], and thus, early and discrete alterations in the mineral homeostasis, which could already be occurring due to the excessive weight and BF, could not be observed in the adolescents assessed. However, given the difculty to collect, store, and process blood in epidemiological studies like this, we chose to assess the Hb concentration, the most commonly used marker when the objective is to evaluate anemia itself and not iron deciency, as recommended by the World Health Organization [2]. In future studies, we suggest that iron status, transferrin saturation and other early markers of iron deciency should also be assessed. The biochemical assessment of inammatory indicators, such as the C-reactive protein, can also help clarify the relation between excessive adiposity, inammation and iron deciency. In this study, there was an initial loss of 23.8% of eligible adolescents, which could represent a limitation in the results interpretation. However, it is believed that such losses have not had any inuence on the results, because upon comparison of the adolescents enrolled in the study with those eligible non-participants, no signicant difference was observed regarding the age mean (16.0 1.9 years old vs. 15.9 1.9 years old, p = 0.88) and the distribution by age group (p = 0.52), variables that are signicantly associated with anemia and sexual maturation. Although participants and non-participants had been heterogeneous regarding sex, a higher frequency of girls among the participants (60.3% vs. 43.3%), did not inuence the associations, as all the analyses were stratied by sex. We conclude that overweight had a negative inuence on the nutritional status of iron in the adolescents investigated, particularly among girls. Longitudinal studies may help understand the dynamics of iron homeostasis during puberty,

Author's personal copy

e228 pointing with more certainty the periods of greatest risks of developing this mineral deciency. Facing the impressive increase of overweight prevalence among adolescents, the development of studies to identify cutoff points for overweight and excessive body fat that dene the risk of iron deciency anemia in this age group is recommended, so that actions and programs focused on youngsters can include anemia among their priorities. U.V. Bagni et al.
[11] Urbano MRD, Vitalle MSS, Juliano Y, Amncio OMS. Iron, copper and zinc in adolescents during pubertal growth spurt. J Pediatr (Rio J) 2002;78(4):32734. [12] Vaskonen T. Dietary minerals and modication of cardiovascular risk factors. J Nutr Biochemistry 2003;14:492506. [13] Zimmermann MB, Zeder C, Muthayya S, Winichagoon P, Chaouki N, Aeberli I, et al. Adiposity in women and children from transition countries predicts decreased iron absorption, iron deciency and a reduced response to iron fortication. Int J Obes (Lond) 2008;32(7):1098104. [14] Menzie CM, Yanoff LB, Denkinger B, McHugh T, Sebring NG, Calis KA, et al. Obesity-related hypoferremia is not explained by differences in reported intake of heme and nonheme iron or intake of dietary factors that can affect iron absorption. J Am Diet Assoc 2008;108(1):1458. [15] Aeberli I, Hurrell RF, Zimmermann MB. Overweight children have higher circulating hepcidin concentrations and lower iron status but have dietary iron intakes and bioavailability comparable with normal weight children. Int J Obes (Lond) 2009;33(10):11117. [16] Lecube A, Carrera A, Losada E, Hernndez C, Cim R, Mesa J. Iron deciency in obese postmenopausal women. Obesity 2006;14:172430. [17] Micozzi MS, Albanes D, Stevens RG. Relation of body size and composition to clinical biochemical and hematologic indices in US men and women. Am J Clin Nutr 1989;50(6):127681. [18] Yanoff LB, Menzie CM, Denkinger B, Sebring NG, McHugh T, Remaley AT, et al. Inammation and iron deciency in the hypoferremia of obesity. Int J Obes (Lond) 2007;31(9):14129. [19] Pinhas-Hamiel O, Neweld RS, Koren I, Agmon A, Lilos P, Phillip M. Greater prevalence of iron deciency in overweight and obese children and adolescents. Int J Obes Relat Metab Disord 2003;27(3):4168. [20] Moayeri H, Bidad K, Zadhoush S, Gholami N, Anari S. Increasing prevalence of iron deciency in overweight and obese children and adolescents (Tehran Adolescent Obesity Study). Eur J Pediatr 2006;165(11):8134. [21] Nead KG, Halterman JS, Kaczorowski JM, Auinger P, Weitzman M. Overweight children and adolescents: a risk group for iron deciency. Pediatrics 2004;114(1):1048. [22] Mena NP, Esparza A, Tapia V, Valds P, N nez MT. Hepcidin inhibits apical iron uptake in intestinal cells. Am J Physiol Gastrointest Liver Physiol 2008;1:G1928. [23] Ganz T, Nemeth E. Iron imports. IV. Hepcidin and regulation of body iron metabolism. Am J Physiol Gastrointest Liver Physiol 2006;290:199203. [24] Fleming RE. Iron and inammation: cross-talk between pathways regulating hepcidin. J Mol Med 2008;86(5):4914. [25] Brasil. Ministrio da Sade. Pesquisa Nacional de a PNDS 2006 Demograa e Sade da Mulher e da Crianc (National Demographic and Health of Women and Child PNDS 2006). Final report. Braslia: Ministrio da Sade, 2008. [26] Santos EB, Amancio OMS, Oliva CAG. Estado nutricional, ferro, cobre e zinco em escolares de favela da cidade de So Paulo (Nutritional status, iron, copper and zinc in the slum school in So Paulo). Rev Assoc Med Bras 2007;53:3238. [27] Grillo LP, Klitzke CH, Campos IC, Mezadri T. Nutritional risks of low family income students residents in the coast of Santa Catarina, Brazil. Texto Contexto Enferm 2005;14(Esp.):1723. [28] Tsuyuoka R, Bailey JW, Guimares AMAN, Gurgel RQ, Cuevas LE. Anemia and intestinal parasitic infections in primary

Conicts of interest
There are no conicts of interest associated with the publication of this work, and that were not omitted information about any afliations or nancial agreements between the authors and companies or individuals who might have interest in the subject matter in the article.

References
[1] World Health Organization. Worldwide prevalence of anaemia 19932005: WHO global database on anaemia. Geneva: WHO; 2008. [2] World Health Organization. Iron deciency anemia: assessment, prevention and control. A guide for programme managers. Document WHO/NHD/01.3. Geneva: WHO; 2001. [3] World Health Organization. Vitamin and mineral requirements in human nutrition: report of a joint FAO/WHO expert consultation. 2nd ed. Bangkok, Thailand: WHO; 2004. [4] Dallman PR, Yip R, Oski FA. Iron deciency and related nutritional anemias. In: Nathan OF, editor. Hematology of infancy and childhood. 4th ed. Philadelphia: WB Saunders; 1993. p. 41350. [5] Halterman JS, Kaczorowski JM, Aligne CA, Auinger P, Szilagyi PG. Iron deciency and cognitive achievement among school-aged children and adolescents in the United States. Pediatrics 2001;107:13816. [6] World Health Organization/Food and Agriculture Organization of the United Nations. Guidelines on food fortication with micronutrients. Geneva: WHO; 2006. [7] Estima CCP, Salles-Costa R, Sichieri R, Pereira RA, Veiga GV. Meal consumption patterns and anthropometric measures in adolescents from a low socioeconomic neighborhood in the metropolitan area of Rio de Janeiro, Brazil. Appetite 2009;52:7359. [8] Levy RB, Castro IRR, Cardoso LO, Tavares LF, Sardinha LMV, Gomes FS, et al. Food consumption and eating behavior among Brazilian adolescents: National Adolescent Schoolbased Health Survey (PeNSE), 2009. Cinc Sade Coletiva 2010;15(Suppl. 2):308597. [9] Toral N, Slater B, Cintra IP, Fisberg M. Adolescent eating behavior regarding fruit and vegetable. Rev Nutr 2006;19(3):33140. [10] Borges CQ, Silva RCR, Assis AMO, Pinto EJ, Fiaccone RL, Pinheiro SMC. Factors associated with anemia in children and adolescents in public schools in Salvador, Bahia State, Brazil. Cad Sade Pblica 2009;25(4):87788.

Author's personal copy

Overweight and low hemoglobin levels e229

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38] [39]

school students in Aracaju, Sergipe, Brazil. Cad Sade [40] Onis M, Onyango AW, Borghi E, Siyam A, Nishida C, Siekmann Pblica 1999;15(2):41321. J. Development of a WHO growth reference for school-aged Franc a MGG. Prevalncia de anemia ferropriva em adoleschildren and adolescents. Bull WHO 2007;85:6607. centes da Vila Princesa lixo em Porto Velho RO [41] Houtkooper LB, Going SB, Lohman TG, Roche AF, Loan MV. (Prevalence of iron deciency anemia in adolescents from Biolectrical impedance estimation of fat body mass index in the Princess Town dump in Porto Velho - RO). MSc children and youth: a cross validation study. J Appl Physiol Thesis, University of Braslia, 2006. 1992;72(1):36673. Silva FC, Vitalle MSS, Quaglia EC, Braga JAP, Medeiros EHGR. [42] Williams DP, Going SB, Lohman TG, Harsha DW, Srinivasan Anemia proportion according to pubertal stage using two SR, Webber LS, et al. Body fatness and risk for elevated diagnostic criteria. Rev Nutr 2007;20(3):297306. blood pressure, total cholesterol, and serum lipoprotein Iuliano BA, Frutuoso MFP, Gambardella AMD. Anemia among ratios in children and adolescents. Am J Public Health adolescents according to sexual maturation. Rev Nutr 1992;82:35863. 2004;17(1):3743. [43] Tanner JM. Growth at adolescence. 2nd ed. Oxford: BlackFujimori E, Szarfarc S, Oliveira IM. Prevalence of iron well Scientic Publication; 1962. deciency anemia in female adolescents in Taboo da o sexual: auto avaliac o do adolescente [44] Saito MI. Maturac Serra, SP. Brazil. Rev Latinoam Enfermagem 1996;4(3): [Sexual maturation: self assessment of the adolescent]. 4963. Pediatrics 1984;6:1115. Frutuoso MFP, Vigantzky VA, Gambardelella AMD. [45] Ferreira MU, Silva-Nunes M, Bertolino CN, Malafronte RS, Hemoglobin serum levels in adolescents according to Muniz PT, Cardoso MA. Anemia and iron deciency in sexual maturation stage. Rev Nutr 2003;16:15562. school children, adolescents, and adults: a communityMariath AB, Henn R, Matos CH, Lacerda LLV, Grillo LP. Prevabased study in rural Amazonia. Am J Public Health lence of anemia and hemoglobin serum levels in adolescents 2007;97(2):2379. according to sexual maturation stage. Rev Bras Epidemiol [46] Rezende EG, Santos MA, Lamounier JA, Galvo MAM, Leite 2006;9(4):45461. MC. Iron deciency and anemia in students from the rural Norton RC, Figueiredo RC, Diamante R, Goulart EM, Mota rea in Novo Cruzeiro (Minas Gerais) Brazil. Rev Med Minas J, Viana MB, et al. Prevalence of anemia among schoolGerais 2009;19(2):1038. children from Rio Acima (State of Minas Gerais, Brazil): [47] Eftekhari MH, Mozaffari-Khosravi H, Shidfar F. The relationuse of the standardized prevalence method and evaluaship between BMI and iron status in iron-decient adolestion of iron deciency. Braz J Med Biol Res 1996;29(12): cent Iranian girls. Public Health Nutr 2009;12:237781. 161724. [48] Del Giudice EM, Santoro N, Amato A, Brienza C, Calabr P, Miglioranza LHS, Matsuo T, Caballero-Crdoba GM, Goulart Wiegerinck ET, et al. Hepcidin in obese children as a potenEM, Mota JA, Viana MB, et al. Anemia prevalence in children tial mediator of the association between obesity and iron and adolescents from educational centers in the outskirts deciency. J Clin Endocrinol Metab 2009;94:51027. of Londrina, PR, Brazil. Rev Nutr 2002;15(2):14953. [49] Tussing-Humphreys LM, Liang H, Nemeth E, Freels S, Duarte LFM, Santos Jnior A, Taddei JAAC. Erythrocyte Braunschweig CA. Excess adiposity, inammation, and count, hemoglobin, hematocrit and ferritin values in adoiron-deciency in female adolescents. J Am Diet Assoc lescents of a public school from So Paulo, Brazil. Rev 2009;109:297302. Paulista Pediatr 2006;24(3):2448. [50] Pollitt E, Leibel RL. Iron deciency and behavior. J Pediatr Low S, Chin MC, Deurenberg-Yap M. Review on epidemic of 1976;88:37281. obesity. Ann Acad Med Singapore 2009;38(1):579. [51] Youdim MBH, Ben-Shachar D, Yehuda S. Putative biological Instituto Brasileiro de Geograa e Estatstica. Pesquisa de mechanisms of the effect of iron deciency on brain and Orc amentos Familiares 20082009. Antropometria e estado behavior. Am J Clin Nutr 1989;50:60717. nutricional de crianc as, adolescentes e adultos no Brasil [Household Budget Survey 20082009. Anthropometry and nutritional status of children, adolescents and adults in Brazil]. Rio de Janeiro: Instituto Brasileiro de Geograa ewww.sciencedirect.com Available online at Estatstica; 2010.