Skip to main content
NCBI home page
Maternal & Child Nutrition logoLink to Maternal & Child Nutrition
. 2017 May 3;14(1):e12460. doi: 10.1111/mcn.12460

Very low prevalence of iron deficiency among young French children: A national cross‐sectional hospital‐based survey

Anne‐Sylvia Sacri 1,2,3,✉, Serge Hercberg 4,5, Laurent Gouya 6, Corinne Levy 7,8, Alain Bocquet 9,10, Béatrice Blondel 1, Catherine Vincelet 11, Pascale Hebel 12, Isabelle Vinatier 13, Mariane de Montalembert 1,2, Henrique Barros 14, Yann Le Strat 15, Martin Chalumeau 1,2
PMCID: PMC6866040  PMID: 28466606

Abstract

Although iron deficiency (ID) is considered the most frequent micronutrient deficiency in industrialized countries and is associated with impaired neurodevelopment when occurring in early years, accurate recent estimations of its prevalence are lacking. Our objective was to estimate ID prevalence and associated sociodemographic markers in young children in France. The Saturn‐Inf national cross‐sectional hospital‐based survey recruited 3,831 French children <6 years old between 2008 and 2009 to assess lead poisoning prevalence and to establish a biobank. This secondary analysis measured serum ferritinemia (SF) in sera kept frozen at −80 °C for children with sufficient serum aliquots and C‐reactive protein <10 mg/L. For the 657 participating children (17% of the Saturn‐Inf study), the median age was 3.9 years (interquartile range: 2.2–5.1); 52% were boys. The median SF was 44 μg/L (interquartile range: 28–71). ID prevalence was 2.8% (95% confidence interval [1.7, 4.7]) and 3.2% (95% confidence interval [2.0, 5.1]) with an SF threshold of 10 and 12 μg/L, respectively. Low SF was significantly associated (p < .05) with mother being a migrant (32 vs. 45 μg/L for a mother born in France) or unemployed (37 vs. 50 μg/L for a mother employed).

In this first national cross‐sectional hospital‐based study in France, ID prevalence was much lower than that in other French and European studies performed in underprivileged populations but close to the lowest values observed in other population‐based studies in Europe.

Keywords: health policy, infant and child nutrition, infant iron status, infant milk formula, preschool children, socioeconomic factors

Abbreviations

CI

confidence interval

CRP

C‐reactive protein

ID

iron deficiency

n

number of observations

SF

serum ferritinemia

1. INTRODUCTION

Iron deficiency (ID) is considered the most frequent micronutrient deficiency worldwide, including in industrialized countries (World Health Organization, 2001). Young children are at high risk of ID because of their volemic mass expansion (Hercberg, Preziosi, & Galan, 2001; Wang, Zhan, Gong, & Lee, 2013; Zimmermann & Hurrell, 2007). Main ID risk factors in young children are poor perinatal iron stock (e.g., maternal ID and preterm birth), imbalance between requirements and dietary intake (e.g., delayed or inappropriate because of iron‐poor diversification), and parental low socioeconomic status (Baker & Greer, 2010; Domellof et al., 2014; Vincelet & Foucault, 2005). ID is associated with adverse short‐ and long‐term neurological sequelae in juvenile animal models (Beard, 2008) and young humans, as reported in several observational studies and randomized controlled trials (Lozoff, Jimenez, & Wolf, 1991; Moffatt, Longstaffe, Besant, & Dureski, 1994; Roncagliolo, Garrido, Walter, Peirano, & Lozoff, 1998; Wang et al., 2013).

In this context, ID is a target of various primary preventive strategies supported by medical societies and public health authorities, and some debate exists on the need for universal screening in young children (Baker & Greer, 2010; Domellof et al., 2014; EFSA, 2013; Ghisolfi et al., 2011; Hercberg et al., 2001). In France, the preventive strategy involves both parents' education oriented toward the consumption of naturally iron‐rich foods at food diversification, about 6 months of age, and consumption of iron‐fortified food (e.g., follow‐on formula; French ministry of health and welfare, 2005). Iron‐fortified follow‐on formulas (current mean iron content 0.94 mg/100 ml) have been recommended since the 1980s for all nonbreast‐fed infants 6 to 12 months old (French ministry of agriculture, 1978), and “growing‐up milk” (mean iron content 1.13 mg/100 ml) has been recommended since the 1990s for all nonbreast‐fed child 12 to 36 months old (Ghisolfi et al., 2011). In other industrialized countries, preventive strategies are based on parental education for adequate food diversification and iron‐fortified food as follow‐on formulas or cereals in the United States (Centers for Disease Control and Prevention, 1998; EFSA, 2013) or iron supplementation in Denmark (EFSA, 2013).

The evaluation of the effectiveness of these different strategies should be based on accurate estimations of ID prevalence in the general population and in at‐risk subgroups of young children. The US National Health and Nutrition Examination Survey performed between 1999 and 2002 provided general population estimates of ID prevalence based on a random sample of 672 children aged 1 to 3 years, for a mean prevalence of 9% (6.6% to 15.2% depending on ethnic origin and socioeconomic status; Baker & Greer, 2010). In European countries, ID prevalence in young children has been reported to range from 0% to 41% (all study design types; Bramhagen, Svahn, Hallstrom, & Axelsson, 2011; Cowin, Emond, & Emmett, 2001; Dube, Schwartz, Mueller, Kalhoff, & Kersting, 2010; Dura Trave & Diaz Velaz, 2002; Gompakis et al., 2007; Gunnarsson, Thorsdottir, & Palsson, 2004; Hay, Sandstad, Whitelaw, & Borch‐Iohnsen, 2004; Hopkins et al., 2007; Male et al., 2001; McCarthy et al., 2016; Michaelsen, Milman, & Samuelson, 1995; Tabone & Vincelet, 2000; Thane, Walmsley, Bates, Prentice, & Cole, 2000; Thorisdottir, Thorsdottir, & Palsson, 2011; Thorsdottir, Gunnarsson, Atladottir, Michaelsen, & Palsson, 2003; Tuthill et al., 2002; Vendt, Grunberg, Leedo, Tillmann, & Talvik, 2007; Vincelet & Foucault, 2005) or from 4% to 34% (random samples of participants; Cowin et al., 2001; Gompakis et al., 2007; Gunnarsson et al., 2004; Hopkins et al., 2007; McCarthy et al., 2016; Thane et al., 2000; Thorisdottir et al., 2011; Vendt et al., 2007). In France, the reported prevalence ranges from 33% for infants 10 months old to 40% for infants 16 to 18 months (Tabone & Vincelet, 2000; Vincelet & Foucault, 2005). Many of these results cannot be used to accurately evaluate the current prevalence of ID in European countries because they predate the implementation of prevention programs (Cowin et al., 2001; Hay et al., 2004; Hopkins et al., 2007; Male et al., 2001; Thane et al., 2000); they were exposed to strong selection bias because of single center recruitment (Dura Trave & Diaz Velaz, 2002; Gompakis et al., 2007; Tabone & Vincelet, 2000; Vincelet & Foucault, 2005), or they were based on small numbers of participants (Gunnarsson et al., 2004; Michaelsen et al., 1995).

The European Society of Pediatric Gastroenterology and Hepatology Nutrition and the European Food Security Authority have recently called for particular attention to ID prevention in infants and young children in Europe and for accurate estimations of ID prevalence in the general population and at‐risk subgroups of young children (Domellof et al., 2014; EFSA, 2013). Here, we aimed to estimate the ID prevalence in young children <6 years old in France, using for the first time a cross‐sectional national hospital‐based survey, and to study associated sociodemographic markers.

Key messages.

  • Iron deficiency (ID) prevalence was much lower (3%) in this first cross‐sectional national hospital‐based study of young children in France than that in other French and European studies performed in underprivileged populations.

  • The very low general ID prevalence and the strong associations with social and deprivation indicators we found strongly support prioritizing ID prevention strategies targeting young infants with low socioeconomic status.

  • The impact of the French national preventive strategy for ID in children, mainly based on the recommendation of iron‐fortified formulas in nonbreast‐fed children from age 6 months, should be explored.

2. METHODS

2.1. General methodology

This study is a secondary analysis of the biobank and databases of the Saturn‐Inf survey established by the French national health institute (Institut National de Veille Sanitaire; Etchevers et al., 2014). Both parents of included children gave their written consent after receiving information from the investigator. Approvals for the survey and the present ancillary analyses were obtained from local administrative and ethics authorities (CNIL no. 907,160, CPP IDF IX no. 08–022, and biobank no. DC‐2015‐2417, n° IRB 00003835). We used the Strengthening the Reporting of Observational Studies in Epidemiology guidelines to report this study (Table S1).

2.2. Participant selection

The Saturn‐Inf survey aimed at evaluating lead poisoning in France and was described in detail elsewhere (Etchevers et al., 2014). In summary, this cross‐sectional survey included 3,831 children aged 6 months to 6 years recruited in general pediatrics, pediatric surgery, or pediatric day hospital departments in 143 hospitals in mainland France and French overseas regions between September 2008 and April 2009. A two‐stage sampling design was used, the first being a random selection of hospitals, and the second being a selection of all hospitalised children with the following inclusion criteria: any health coverage, blood sampling planned during the hospitalisation, no severe chronical diseases (see detailed list, Lepoutre et al., 2008), no ongoing iron chelating treatment, and no ongoing diagnosis of lead poisoning. The hospital sampling frame was stratified by region and featured an oversampling of “high‐risk areas” of elevated blood lead to increase the precision of estimates of lead poisoning prevalence. The participation rate was 83% for the 273 sampled hospitals and 97% for the 3,949 sampled children.

All the 3,831 included children with the following criteria were eligible for this ancillary study: parents' written consent for secondary analyses, sufficient serum aliquots, lack of inflammatory syndrome defined as C‐reactive protein (CRP) <10 mg/L on frozen sera aliquot (because serum ferritinemia [SF] increases with inflammation), and available minimum set of sociodemographic data (i.e., at least age and gender).

2.3. Data collected

At inclusion in the Saturn‐Inf survey, the following sociodemographic data were collected by use of a standardized questionnaire during a face‐to‐face interview with pediatricians or nurses: age and sex of the child, country of birth of the mother (migrant or not), region of residence, underprivileged status, and each parent's professional status and educational. The family was considered underprivileged if its health coverage was fully funded by a public grant, which is linked to poverty in France (French National Institute of Statistics and Economics INSEE, 2013).

Serum was frozen at −80 °C and the following were measured immediately after samples were thawed in December 2014: SF (μg/l), measured by electro‐chemiluminescence immunoassay (COBAS‐Roche Diagnostics‐COBAS 6000 E601, measuring range 0.500–2000 μg/L; intra and interassay variation coefficients <1.1% and 5.7%, respectively), and CRP (mg/L), measured by immunoturbidimetric assay (COBAS‐Roche Diagnostics‐COBAS 6000 E501, measuring range: 0.3–350 mg/L, intra and interassay variation coefficients <1.4% and 1.8%, respectively).

2.4. Statistical analysis

We compared the general characteristics of the studied population to those of the Saturn‐Inf survey participants excluded from the present ancillary analysis and those of the 2010 French National Perinatal Survey, which is based on a representative sample of births (n = 15418) (Blondel, Lelong, Kermarrec, & Goffinet, 2012). Taking into account the Saturn‐Inf two‐stage sampling design (i.e., the cluster effects and the weights linked to the stratification), we described the distribution of SF and studied its crude then adjusted associations with sociodemographic characteristics thanks to the medians' 95% confidence interval (CI) and two multivariable linear regression models (after logarithmic transformation of SF, Bland & Altman, 1996) including variables selected based on a literature review. The first model included age, gender, and underprivileged status, which we considered as the most representative marker of poor sociodemographic status. The second included all available sociodemographic covariables (except parents' employment because of collinearity). Then, SF was dichotomized with a 10‐μg/L threshold to define ID, as suggested in the literature (Baker & Greer, 2010; Dallman, Siimes, & Stekel, 1980; de Montalembert et al., 2012). The prevalence of ID and its crude association with sociodemographic characteristics were studied by chi‐square and Fisher exact tests; no multivariable logistic regression model was used given the low number of ID cases. All analyses were performed again with a 12‐μg/L threshold to define ID, as also suggested in the literature (Baker & Greer, 2010; Cook, Lipschitz, Miles, & Finch, 1974; Dallman et al., 1980; World Health Organization/Centers for Disease Control and Prevention technical consultation on the assessment of iron status at the population level, 2007).

In all analyses, parents' educational level was classified in four categories according to the maximum level obtained: secondary school (including no education, primary school, and general secondary school), professional education (without high school), high school (designating both professional and general education), and higher learning. Parents' professional status was redefined in two categories, employed or not (including housewives, students, and pensioners). Age was used as a continuous variable in multivariable analysis after testing the lack of deviance to linearity; results are presented with age in three categories by periods of variation in iron metabolism and requirements (<2, 2 to less than 3, and ≥3 years old). Regions of residence were categorized (n = 9) by the first level of the nomenclature of territorial units for statistics of the European Union.

All analyses were performed also directly on the population of children selected from Saturn‐Inf without applying the two‐stage sampling design (results presented in Supporting Information). Missing data were not imputed. The analyses involved use of Stata/SE 13.1 (StataCorp LP Statistics/Data Analysis, Texas 77845, USA).

3. RESULTS

3.1. Participants

Among the 3,831 children included in the Saturn‐Inf study, data for 3,174 (83%) were excluded because of lack of parents' written consent (n = 1), insufficient serum aliquots (n = 2,586, 68%), lack of data on age and/or gender (n = 15, 0.4%), or presence of an inflammatory syndrome (n = 572, 15%, Figure S1). The current analysis is based on data for 657 (17%) children.

Considering the two‐stage sample design, the mean age of children was 3.7 years (standard error: 0.13; median: 3.9, interquartile range: 2.2–5.1), 52% were boys, 14% had a migrant mother, 47% had an unemployed mother, and 21% a family with an underprivileged status (Tables 1 and S2). Except for father's employment, participant characteristics significantly differed from those of participants included from the Saturn‐Inf survey and those of the 2010 French National Perinatal Survey participants (Table S3; Blondel et al., 2012).

Table 1.

Sociodemographic characteristics and association with serum ferritinemia (SF; taking into account the Saturn‐Inf two‐stage sampling design)

Characteristics Proportion (%) SF (μg/L) Iron deficiency (SF <10 μg/L)
Median 95% CIa Model 1b Model 2b

ID

(%)

No ID

(%)

 OR 95% CIc
n = 657 β 95% CId β 95% CId
Age (years)e
0.8–2 22 43.1 28.9, 49.6 Ref. Ref. 3 97 Ref. Ref.
2–3 17 40.6 37.8, 45.4 −0.14, 0.21 −0.13, 0.23 5 95 1.59 0.55, 4.63
3–7 61 44.3 38.0, 47.7 0.003, 0.31 0.07, 0.48 2 98 0.57 0.20, 1.66
Gender
Female 48 48.0 39.0, 55.5 Ref. Ref. 2 98 Ref. Ref.
Male 52 38.9 35.1, 44.6 −0.29, −0.10 −0.27, −0.11 4 96 2.17 1.00, 4.70
Mother's country of birth
France 86 44.9 40.6, 49.5 Ref. 2 98 Ref. Ref.
Other 14 32.1 25.9, 34.0 −0.37, 0.13 7 93 3.12 1.06, 9.18
Underprivileged status
No 79 44.2 38.4, 49.6 Ref. Ref. 2 98 Ref. Ref.
Yes 21 43.1 34.5, 50.5 −0.37, 0.12 −0.25, 0.28 6 94 3.83 1.39, 10.53
Mother's employment
Employed 53 49.5 44.9, 55.5 1 99 Ref. Ref.
Unemployed 47 36.6 33.3, 44.6 4 96 3.16 0.93, 10.75
Father's employment
Employed 88 44.9 38.6, 49.5 3 97 Ref. Ref.
Unemployed 12 43.0 33.0, 61.3 4 96 1.69 0.48, 6.01
Mother's educational level
Higher learning 32 47.2 44.9, 49.6 Ref. 2 98 Ref. Ref.
High school 22 57.0 39.4, 71.3 −0.04, 0.47 2 98 1.17 0.23, 5.92
Professional 31 36.6 33.0, 49.5 −0.36, 0.13 2 98 1.25 0.32, 4.81
Secondary school 15 32.1 21.8, 46.8 −0.78, −0.10 8 92 4.48 1.26, 15.98
Father's educational levelf
Higher learning 32 40.2 34.9, 47.2 Ref. 0 100
High school 15 45.0 35.1, 54.4 −0.38, 0.09 6 94
Professional 42 52.3 40.7, 58.5 −0.10, 0.26 2 98
Secondary school 10 39.9 28.8, 57.1 −0.16, 0.54 8 92
Regions
Paris region 17 34.0 34.0, 34.0 Ref. 2 98 Ref. Ref.
Paris basin 24 47.2 40.7, 60.3 0.05, 0.51 3 97 2.05 0.19, 22.32
North France 2 35.1 35.1, 35.1 −0.23, 0.25 4 96 2.25 0.16, 31.65
East France 10 41.8 37.3, 43.6 −0.0005, 0.53 2 98 1.08 0.06, 19.55
West France 14 39.0 36.6, 47.7 0.08, 0.38 1 99 0.49 0.03, 7.73
Southwest France 6 55.5 49.5, 55.5 0.28, 0.60 6 94 3.64 0.37, 35.46
Center‐east France 2 50.6 50.6, 50.6 0.35, 0.73 2 98 1.43 0.09, 23.59
Mediterranean France 10 48.9 41.8, 49.6 0.01, 0.78 5 95 3.20 0.22, 46.03
Overseas regions 15 41.7 38.6, 43.7 −0.12, 0.83 2 98 1.35 0.11, 0.17
Open in a new tab

Note. ID = iron deficiency; CI = confidence interval; n = number of observations; OR = odds ratio; SF = serum ferritinemia.

a

95% CI of medians taking into account the two‐stage sampling design.

b

Multivariable linear model after a logarithmic transformation of the dependent variable (SF).

c

95% CI of OR.

d

95% CI of β coefficients of a multiple linear regression model. Analysis of age as a continuous variable (no deviance to linearity) yielded similar results.

e

Analysis of age as a continuous variable (no deviance to linearity) yielded similar results.

f

OR not calculable given presence of cells with a zero value.

3.2. SF and association with sociodemographic factors

Median SF was 44 μg/L (interquartile range: 28–71, mean [standard deviation]: 64 [63]; Figure 1). Low SF was significantly associated (p < .05) with mother being a migrant (32 vs. 45 μg/L for a mother born in France) or unemployed (37 vs. 50 μg/L for a mother employed). SF was not significantly associated with participants' age (as a linear variable [no deviance to linearity] or in three categories in years [≥0.8 and <2, ≥2 and <3, and ≥3 and <7]) or gender, underprivileged family status, father's employment, parents' educational level, or region of residence (Table 1).

Figure 1.

Figure 1

Open in a new tab

Serum ferritinemia distribution. Dotted line corresponds to the serum ferritinemia threshold at 10 μg/L. Serum ferritinemia scale: Serum ferritinemia is presented in μg/L, values spaced every 10 μg/L from 0 to 300

In a first multivariable linear regression model including age, gender, and underprivileged status, a significantly lower SF was associated with male gender (Table 1). In a second model including age, gender, mother's country of birth, underprivileged status, parents' educational level, and region of residence, a lower SF was associated with only male gender (Table 1).

Analyses that did not consider the Saturn‐Inf sampling design are presented in Table S4.

3.3. ID prevalence and sociodemographic markers

ID prevalence was 2.8% (95% CI [1.7, 4.7]) and 3.2% (95% CI [2.0, 5.1]) with an SF threshold of 10 and 12 μg/L, respectively (Tables 1 and S2). With a 10‐μg/L threshold, ID was significantly associated with male gender (4%), the mother being a migrant (7%), underprivileged family status (6%), and low maternal educational level (8% vs. 2% for mothers having a secondary school level versus a higher learning level). ID was not associated with age (as a linear variable [no deviance to linearity] or in three categories), parents' professional status or region of residence (Table 1). ID defined with a 12‐μg/L threshold was associated with the same markers as ID defined with a 10‐μg/L threshold and also mother's unemployment and low father educational level (Table S2).

Analyses that did not consider the Saturn‐Inf sampling design are presented in Tables S4 and S5.

4. DISCUSSION

In this first national cross‐sectional hospital‐based study in France, we found a very low prevalence of ID in children from 6 months to 6 years old: 2.8% (95% CI [1.7, 4.7]) and 3.2% (95% CI [2.0, 5.1]) depending on the SF threshold (10 or 12 μg/L, respectively). The large sample size, 657 children, allowed for producing precise estimates. We found no significant geographical variations in ID prevalence. This very low prevalence is an unexpected finding given the two most recent estimates of ID prevalence in French children of 33% and 40% issued from studies performed in 1998 and 2002 (Tabone & Vincelet, 2000; Vincelet & Foucault, 2005) but is within the lower range of prevalences reported recently in young children in Europe (0% to 41%) (Bramhagen et al., 2011; Cowin et al., 2001; Dube et al., 2010; Dura Trave & Diaz Velaz, 2002; Gompakis et al., 2007; Gunnarsson et al., 2004; Hay et al., 2004; Hopkins et al., 2007; Male et al., 2001; McCarthy et al., 2016; Michaelsen et al., 1995; Tabone & Vincelet, 2000; Thane et al., 2000; Thorisdottir et al., 2011; Thorsdottir et al., 2003; Tuthill et al., 2002; Vendt et al., 2007; Vincelet & Foucault, 2005).

Variability in ID prevalence in children can be related to variability in ID definitions, as various ones were used in the literature: SF alone (Bramhagen et al., 2011; Cowin et al., 2001; Dube et al., 2010; Dura Trave & Diaz Velaz, 2002; Gompakis et al., 2007; Gunnarsson et al., 2004; Hay et al., 2004; Hopkins et al., 2007; Male et al., 2001; McCarthy et al., 2016; Tabone & Vincelet, 2000; Thane et al., 2000; Thorisdottir et al., 2011; Thorsdottir et al., 2003; Tuthill et al., 2002; Vendt et al., 2007; Vincelet & Foucault, 2005) or combined with other indicators (red‐cell mean corpuscular volume, transferrin saturation; and/or serum transferrin receptor; Baker & Greer, 2010; Dura Trave & Diaz Velaz, 2002; Male et al., 2001; McCarthy et al., 2016; Michaelsen et al., 1995; Thorisdottir et al., 2011). In the four studies that used both SF alone or combined with the same blood sample, ID prevalence was higher with SF used alone versus in a combined marker: 5.8% versus 1.4% (Thorisdottir et al., 2011), 15.6% versus 7.2% (Male et al., 2001), 3.2% versus 2.1% (Dura Trave & Diaz Velaz, 2002), and 4.6% versus 2.6% (McCarthy et al., 2016). Thus, our results cannot be explained by using SF as a definition of ID. ID definitions based on SF may also vary depending on the thresholds used and the exclusion of participants with inflammatory status, because SF increases with inflammation (Zimmermann & Hurrell, 2007). However, among studies evaluating ID prevalence in young children without inflammatory syndrome in industrialized countries, the range of estimations, 1.8% to 20% for SF threshold of 10 μg/L (Baker & Greer, 2010; McCarthy et al., 2016; Thane et al., 2000) and 4% to 41% for threshold 12 μg/L (Gunnarsson et al., 2004; McCarthy et al., 2016; Thane et al., 2000; Vendt et al., 2007), are close to those of studies not accounting for inflammation (1.1% to 34% for SF threshold of 10 μg/L (Gompakis et al., 2007; Male et al., 2001; Tuthill et al., 2002) and 4% to 40% for threshold of 12 μg/l (Cowin et al., 2001; Dube et al., 2010; Tabone & Vincelet, 2000; Thorisdottir et al., 2011; Vincelet & Foucault, 2005).

Variability in ID prevalence can also be related to differences in population studied, notably participants' age and socioeconomic status. The mean (standard error) age of participants in the present study was 3.7 (0.13) years, which is older than the age in the other published studies involving young children (Baker & Greer, 2010; Bramhagen et al., 2011; Cowin et al., 2001; Dube et al., 2010; Dura Trave & Diaz Velaz, 2002; Gunnarsson, Thorsdottir, Palsson, & Gretarsson, 2007; Hay et al., 2004; Hopkins et al., 2007; Male et al., 2001; McCarthy et al., 2016; Michaelsen et al., 1995; Tabone & Vincelet, 2000; Thorisdottir et al., 2011; Tuthill et al., 2002; Vendt et al., 2007; Vincelet & Foucault, 2005). However, prevalence did not differ among the age ranges in our study, so age variation cannot explain our results. Nor could selection bias related to socioeconomic status explain our results. Indeed, with the oversampling strategy of children at risk of lead poisoning, 21% of our population was underprivileged, which is a much higher rate than in the two former French studies on ID deficiency (7%, Vincelet & Foucault, 2005 and 19%, Tabone & Vincelet, 2000) and in the French National Perinatal Survey (13%; Blondel et al., 2012). This study involved children hospitalized in public general pediatrics units or surgery units which could explain the high rate of underprivileged markers found. If we consider only population‐based studies performed in industrialized countries and excluding participants with inflammatory status, reported ID prevalence was higher than in the present study, from 9.2% to 20% (Baker & Greer, 2010; Thane et al., 2000) and 14% to 31% (Gunnarsson et al., 2004; Thane et al., 2000; Vendt et al., 2007) with a threshold of 10 and 12 μg/L, respectively.

The very low prevalence we found as compared to other European studies may be due to national variations in iron intake (Baker & Greer, 2010; EFSA, 2013; Eussen, Alles, Uijterschout, Brus, & van der Horst‐Graat, 2015). Indeed, the French ID prevention strategy relies on the recommendation for iron‐fortified follow‐on formula and growing‐up milk for nonbreast‐fed children from age 6 to 36 months (Ghisolfi et al., 2011). In 2005 and 2013, cross‐sectional national studies have shown that iron intakes among French infants and toddlers were above the recommended intakes of 7 mg/day until the age of 18 and 24 months old but slightly lower (around 6.5 mg/day) after (Bocquet & Vidailhet, 2015; Fantino & Gourmet, 2008). In the 2013 study, the consumption of iron‐fortified formulas among nonbreast‐fed infants and toddlers was gradually declining with age: 89% from 6 to 7 months, 84% from 8 to 11 months, 60% from 12 to 17 months, 57% from 18 to 23 months, 32% from 24 to 29 months, and 24% from 30 to 36 months (Bocquet & Vidailhet, 2015). The positive effect of follow‐on formulas on iron status has been shown in randomized controlled studies (Daly et al., 1996; Fuchs et al., 1993; Gill, Vincent, & Segal, 1997), with a persistent effect several months after stopping consumption (Daly et al., 1996). In this study, involving children without chronic disease hospitalized in pediatric medical and surgical departments, we cannot anticipate the direction and the force of a potential selection bias on the usual consumption of iron‐fortified formulas compared to nonhospitalized children. Thus, the very low ID prevalence found in this study may be related to adequate iron intakes in young children in France, notably due to the high coverage of follow‐on formula.

The socioeconomic markers of ID in children we identified were directly or indirectly related to poor economic or educational level of the family: immigration, underprivileged status, parents' unemployment, and low educational level (Baker & Greer, 2010; Siu, 2015; Tabone & Vincelet, 2000; Vincelet & Foucault, 2005). These results confirm the statements of expert societies that underlie the need for prevention policies for ID targeting infants and children with inadequate or at risk of inadequate iron status (i.e., low socioeconomic status; Baker & Greer, 2010; Domellof et al., 2014; EFSA, 2013). We found male gender an independent marker associated with low SF, as in other studies, which also reported its association with elevated of soluble transferrin receptors (a marker of ID) and anemia (Choi, Pai, Im, & Kim, 1999; Domellof et al., 2002; Male et al., 2001; Sherriff, Emond, Hawkins, & Golding, 1999). Several hypotheses have been raised to explain this association (e.g., greater growth rate than females), but none rely on a robust rationale (Choi et al., 1999; Domellof et al., 2002; Male et al., 2001; Sherriff et al., 1999).

Our study has several limitations. First, despite a two‐stage probability sample design, the initial sample significantly differed from that in the 2010 French National Perinatal survey (Blondel et al., 2012; Table S3), which may have modified the results of our main outcome, ID prevalence. Second, the attrition rate was almost 83%, mainly because of a factor that is probably independent of ID prevalence in children (insufficient serum aliquots). Third, serum samples were stored for 5 to 6 years at −80 °C, conditions that could have modified the SF. There is no data in the literature about the freezing protocol of SF at this temperature and during this lapse of time. Preliminary results from the birth cohort Generation XXI in Porto, Portugal, in which SF was measured twice in the same aliquot, at the beginning of the survey and after a freezing protocol at −80 °C for 3 to 4 years, showed highly preserved values (Barros H, personal communication 2016). Reassuring data exist for the stability of CRP after long periods of conservation at −80 °C (Nilsson et al., 2005). The distribution of SF in this study showed several high values despite the exclusion of participants with inflammation (CRP >10 mg/L). Because CRP is an acute‐phase inflammation marker, some participants might have been at the end of an inflammatory phase with their CRP already returned to normal and an SF still elevated (Zimmermann & Hurrell, 2007), which could have decreased the prevalence of ID.

In conclusion, in this first national cross‐sectional hospital‐based study in France, ID prevalence was much lower (2.8 to 3.2% depending on the SF threshold) than that from other French and European studies of underprivileged populations but close to the low values observed in other population‐based studies in Europe. Further studies are needed to explore the relation between this very low ID prevalence and the French national preventive strategy for ID in children that is mainly based on the recommendation of iron‐fortified formulas in nonbreast‐fed children from age 6 months. The very low general ID prevalence and the strong associations with social and deprivation indicators we found strongly support prioritizing ID prevention strategies targeting young infants with low socioeconomic status.

CONFLICTS OF INTEREST

The funders mentioned in “sources of support” had no role in the study design, data collection and analysis, the decision to publish, or the preparation of the manuscript. The authors have no patents, products in development, or marketed products to declare.

CONTRIBUTIONS

AB, BB, MC, MdM, LG, PH, SH, CL, ASS, IV, and CV designed and conducted research; HB and YLS provided essential materials; ASS, MC, and YLS performed statistical analysis; ASS and MC wrote the manuscript; ASS and MC have primary responsibility for final content. All authors read, revised critically, and approved the final manuscript.

Supporting information

Table S1. Strengthening the Reporting of Observational Studies in Epidemiology –STROBE Statement — checklist of items that should be included in reports of observational studies

Table S2. Participant characteristics, association between serum ferritinemia (SF) and socio‐demographic factors for iron deficiency threshold 12 μg/l (taking into account the Saturn‐Inf two‐stage sampling design)

Table S3. Characteristics of participants in the 2010 French National Perinatal survey, the total Saturn‐Inf study, and excluded and excluded Saturn‐Inf participants

Table S4. Sociodemographic characteristics of participants and association with serum ferritinemia (SF) (without taking into account Saturn‐Inf sampling design)

Table S5. Participant characteristics, association between serum ferritinemia (SF) and sociodemographic factors for iron deficiency (ID) threshold 12 μg/l (without taking into account the Saturn‐Inf sampling design)

Figure S1. Flow chart of participants in the study

Click here for additional data file. (90.1KB, docx)

Sacri A‐S, Hercberg S, Gouya L, et al. Very low prevalence of iron deficiency among young French children: A national cross‐sectional hospital‐based survey. Matern Child Nutr. 2018;14:e12460 10.1111/mcn.12460

REFERENCES

  1. Baker, R. D. , & Greer, F. R. (2010). Diagnosis and prevention of iron deficiency and iron‐deficiency anemia in infants and young children (0–3 years of age). Pediatrics, 126, 1040–1050. [DOI] [PubMed] [Google Scholar]
  2. Beard, J. (2008). Why iron deficiency is important in infant development. The Journal of Nutrition, 138, 2534–2536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bland, J. M. , & Altman, D. G. (1996). Transformations, means, and confidence intervals. BMJ, 312, 1079. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Blondel, B. , Lelong, N. , Kermarrec, M. , & Goffinet, F. (2012). Trends in perinatal health in France from 1995 to 2010. Results from the French National Perinatal Surveys. J Gynecol Obstet Biol Reprod (Paris), 41, e1–e15. [DOI] [PubMed] [Google Scholar]
  5. Bocquet, A. , & Vidailhet, M. (2015). Nutri‐Bebe 2013 study part 2. How do French mothers feed their young children? Archives de Pédiatrie, 22, 10s17–10s19. [DOI] [PubMed] [Google Scholar]
  6. Bramhagen, A. , Svahn, J. , Hallstrom, I. , & Axelsson, I. (2011). Factors influencing iron nutrition among one‐year‐old healthy children in Sweden. Journal of Clinical Nursing, 20, 1887–1894. [DOI] [PubMed] [Google Scholar]
  7. Centers for Disease Control and Prevention . (1998). Recommendations to prevent and control iron deficiency in the United States. MMWR ‐ Recommendations and Reports, 47, 1–29. [PubMed] [Google Scholar]
  8. Choi, J. W. , Pai, S. H. , Im, M. W. , & Kim, S. K. (1999). Change in transferrin receptor concentrations with age. Clinical Chemistry, 45, 1562–1563. [PubMed] [Google Scholar]
  9. Cook, J. , Lipschitz, D. , Miles, L. , & Finch, C. (1974). Serum ferritin as a measure of iron stores in normal subjects. The American Journal of Clinical Nutrition, 27, 681–687. [DOI] [PubMed] [Google Scholar]
  10. Cowin, I. , Emond, A. , & Emmett, P. (2001). Association between composition of the diet and haemoglobin and ferritin levels in 18‐month‐old children. European Journal of Clinical Nutrition, 55, 278–286. [DOI] [PubMed] [Google Scholar]
  11. Dallman, P. , Siimes, M. , & Stekel, A. (1980). Iron deficiency in infancy and childhood. The American Journal of Clinical Nutrition, 33, 86–118. [DOI] [PubMed] [Google Scholar]
  12. Daly, A. , MacDonald, A. , Aukett, A. , Williams, J. , Wolf, A. , Davidson, J. , & Booth, I. W. (1996). Prevention of anaemia in inner city toddlers by an iron supplemented cows' milk formula. Archives of Disease in Childhood, 75, 9–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Domellof, M. , Lonnerdal, B. , Dewey, K. G. , Cohen, R. J. , Rivera, L. L. , & Hernell, O. (2002). Sex differences in iron status during infancy. Pediatrics, 110, 545–552. [DOI] [PubMed] [Google Scholar]
  14. Domellof, M. , Braegger, C. , Campoy, C. , Colomb, V. , Decsi, T. , Fewtrell, M. , … van Goudoever, J. (2014). Iron requirements of infants and toddlers. Journal of Pediatric Gastroenterology and Nutrition, 58, 119–129. [DOI] [PubMed] [Google Scholar]
  15. Dube, K. , Schwartz, J. , Mueller, M. J. , Kalhoff, H. , & Kersting, M. (2010). Iron intake and iron status in breastfed infants during the first year of life. Clinical Nutrition, 29, 773–778. [DOI] [PubMed] [Google Scholar]
  16. Dura Trave, T. , & Diaz Velaz, L. (2002). Prevalence of iron deficiency in healthy 12‐month‐old infants. Anales Españoles de Pediatría, 57, 209–214. [PubMed] [Google Scholar]
  17. EFSA . (2013). European Food Security Authority. Panel on Dietetic Products, Nutrition and Allergies (NDA). Scientific opinion on nutrient requirements and dietary intakes of infants and young children in the European Union. European Food Security Authority (EFSA) Journal [serial online], 11, 3408–3511. [Google Scholar]
  18. Etchevers, A. , Bretin, P. , Lecoffre, C. , Bidondo, M. L. , Le Strat, Y. , Glorennec, P. , & Le Tertre, A. (2014). Blood lead levels and risk factors in young children in France, 2008–2009. International Journal of Hygiene and Environmental Health, 217, 528–537. [DOI] [PubMed] [Google Scholar]
  19. Eussen, S. , Alles, M. , Uijterschout, L. , Brus, F. , & van der Horst‐Graat, J. (2015). Iron intake and status of children aged 6–36 months in Europe: A systematic review. Annals of Nutrition & Metabolism, 66, 80–92. [DOI] [PubMed] [Google Scholar]
  20. Fantino, M. , & Gourmet, E. (2008). Nutrient intakes in 2005 by non‐breast fed French children of less than 36 months. Archives de Pédiatrie, 15, 446–455. [DOI] [PubMed] [Google Scholar]
  21. French ministry of agriculture . (1978). Decree of 30 March 1978 laying down provisions relating to dietetic milk foods. Paris, France: French Government General Secretariat Journal officiel de la République Française C.F.R. [Google Scholar]
  22. French ministry of health and welfare . (2005). National Health Nutrition Plan. [Nutrition guide from birth to 3 years‐old]. Retrieved 2015
  23. French National Institute of Statistics and Economics INSEE . (2013). [Universal Health coverage and Complementary Universal Health coverage—Definitions]. Retrieved 10 may 2016, from http://www.insee.fr/fr/themes/document.asp?reg_id=7&ref_id=21042&page=dossiers_etudes/publications_elec/zoom_CMU-C/CMU-C_definitions.htm
  24. Fuchs, G. J. , Farris, R. P. , DeWier, M. , Hutchinson, S. W. , Warrier, R. , Doucet, H. , & Suskind, R. M. (1993). Iron status and intake of older infants fed formula vs cow milk with cereal. The American Journal of Clinical Nutrition, 58, 343–348. [DOI] [PubMed] [Google Scholar]
  25. Ghisolfi, J. , Vidailhet, M. , Fantino, M. , Bocquet, A. , Bresson, J. L. , Briend, A. , … Turck, D. (2011). Cows' milk or growing‐up milk: What should we recommend for children between 1 and 3 years of age? Archives de Pédiatrie, 18, 355–358. [DOI] [PubMed] [Google Scholar]
  26. Gill, D. G. , Vincent, S. , & Segal, D. S. (1997). Follow‐on formula in the prevention of iron deficiency: A multicentre study. Acta Paediatrica, 86, 683–689. [DOI] [PubMed] [Google Scholar]
  27. Gompakis, N. , Economou, M. , Tsantali, C. , Kouloulias, V. , Keramida, M. , & Athanasiou‐Metaxa, M. (2007). The effect of dietary habits and socioeconomic status on the prevalence of iron deficiency in children of northern Greece. Acta Haematologica, 117, 200–204. [DOI] [PubMed] [Google Scholar]
  28. Gunnarsson, B. S. , Thorsdottir, I. , & Palsson, G. (2004). Iron status in 2‐year‐old Icelandic children and associations with dietary intake and growth. European Journal of Clinical Nutrition, 58, 901–906. [DOI] [PubMed] [Google Scholar]
  29. Gunnarsson, B. S. , Thorsdottir, I. , Palsson, G. , & Gretarsson, S. J. (2007). Iron status at 1 and 6 years versus developmental scores at 6 years in a well‐nourished affluent population. Acta Paediatrica, 96, 391–395. [DOI] [PubMed] [Google Scholar]
  30. Hay, G. , Sandstad, B. , Whitelaw, A. , & Borch‐Iohnsen, B. (2004). Iron status in a group of Norwegian children aged 6–24 months. Acta Paediatrica, 93, 592–598. [PubMed] [Google Scholar]
  31. Hercberg, S. , Preziosi, P. , & Galan, P. (2001). Iron deficiency in Europe. Public Health Nutrition, 4, 537–545. [DOI] [PubMed] [Google Scholar]
  32. Hopkins, D. , Emmett, P. , Steer, C. , Rogers, I. , Noble, S. , & Emond, A. (2007). Infant feeding in the second 6 months of life related to iron status: An observational study. Archives of Disease in Childhood, 92, 850–854. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lepoutre, A. , Lecoffre, C. , Fréry, N. , Le Tertre, A. , Pascal, M. , De Crouy‐Chanel, P. , … Gaillard, C. (2008). Saturn‐Inf 2008–2009. National prevalence survey about lead poisoning and infectious diseases seroprevalences in children from 6 months to 6 years. Protocol. Paris, France: French national health institute Departement Health Environment. [Google Scholar]
  34. Lozoff, B. , Jimenez, E. , & Wolf, A. W. (1991). Long‐term developmental outcome of infants with iron deficiency. The New England Journal of Medicine, 325, 687–694. [DOI] [PubMed] [Google Scholar]
  35. Male, C. , Persson, L. A. , Freeman, V. , Guerra, A. , van't Hof, M. A. , & Haschke, F. (2001). Prevalence of iron deficiency in 12‐mo‐old infants from 11 European areas and influence of dietary factors on iron status (Euro‐Growth study). Acta Paediatrica, 90, 492–498. [DOI] [PubMed] [Google Scholar]
  36. McCarthy, E. K. , Ni Chaoimh, C. , O'B Hourihane, J. , Kenny, L. C. , Irvine, A. D. , Murray, D. M. , & Kiely, M. (2016). Iron intakes and status of 2‐year‐old children in the Cork BASELINE Birth Cohort Study. Maternal & Child Nutrition. 10.1111/mcn.12320 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Michaelsen, K. F. , Milman, N. , & Samuelson, G. (1995). A longitudinal study of iron status in healthy Danish infants: Effects of early iron status, growth velocity and dietary factors. Acta Paediatrica, 84, 1035–1044. [DOI] [PubMed] [Google Scholar]
  38. Moffatt, M. E. , Longstaffe, S. , Besant, J. , & Dureski, C. (1994). Prevention of iron deficiency and psychomotor decline in high‐risk infants through use of iron‐fortified infant formula: A randomized clinical trial. The Journal of Pediatrics, 125, 527–534. [DOI] [PubMed] [Google Scholar]
  39. de Montalembert, M. , Bresson, J. L. , Brouzes, C. , Ruemmele, F. M. , Puy, H. , & Beaumont, C. (2012). Diagnosis of hypochromic microcytic anemia in children. Archives de Pédiatrie, 19, 295–304. [DOI] [PubMed] [Google Scholar]
  40. Nilsson, T. K. , Boman, K. , Jansson, J. H. , Thogersen, A. M. , Berggren, M. , Broberg, A. , & Granlund, A. (2005). Comparison of soluble thrombomodulin, von Willebrand factor, tPA/PAI‐1 complex, and high‐sensitivity CRP concentrations in serum, EDTA plasma, citrated plasma, and acidified citrated plasma (Stabilyte) stored at −70 degrees C for 8‐11 years. Thrombosis Research, 116, 249–254. [DOI] [PubMed] [Google Scholar]
  41. Roncagliolo, M. , Garrido, M. , Walter, T. , Peirano, P. , & Lozoff, B. (1998). Evidence of altered central nervous system development in infants with iron deficiency anemia at 6 mo: delayed maturation of auditory brainstem responses. The American Journal of Clinical Nutrition, 68, 683–690. [DOI] [PubMed] [Google Scholar]
  42. Sherriff, A. , Emond, A. , Hawkins, N. , & Golding, J. (1999). Haemoglobin and ferritin concentrations in children aged 12 and 18 months. ALSPAC Children in Focus Study Team. Archives of Disease in Childhood, 80, 153–157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Siu, A. L. (2015). Screening for iron deficiency anemia in young children: USPSTF recommendation statement. Pediatrics, 136, 746–752. [DOI] [PubMed] [Google Scholar]
  44. Tabone, M. D. , & Vincelet, C. (2000). Socioeconomic status and child health: The experience of the Paris Child Health Checkup Center. Archives de Pédiatrie, 7, 1274–1283. [DOI] [PubMed] [Google Scholar]
  45. Thane, C. W. , Walmsley, C. M. , Bates, C. J. , Prentice, A. , & Cole, T. J. (2000). Risk factors for poor iron status in British toddlers: Further analysis of data from the National Diet and Nutrition Survey of children aged 1.5–4.5 years. Public Health Nutrition, 3, 433–440. [DOI] [PubMed] [Google Scholar]
  46. Thorisdottir, A. V. , Thorsdottir, I. , & Palsson, G. I. (2011). Nutrition and iron status of 1‐year olds following a revision in infant dietary recommendations. Anemia, 2011, 986303. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Thorsdottir, I. , Gunnarsson, B. S. , Atladottir, H. , Michaelsen, K. F. , & Palsson, G. (2003). Iron status at 12 months of age—Effects of body size, growth and diet in a population with high birth weight. European Journal of Clinical Nutrition, 57, 505–513. [DOI] [PubMed] [Google Scholar]
  48. Tuthill, D. P. , Cosgrove, M. , Dunstan, F. , Stuart, M. L. , Wells, J. C. , & Davies, D. P. (2002). Randomized double‐blind controlled trial on the effects on iron status in the first year between a no added iron and standard infant formula received for three months. Acta Paediatrica, 91, 119–124. [DOI] [PubMed] [Google Scholar]
  49. Vendt, N. , Grunberg, H. , Leedo, S. , Tillmann, V. , & Talvik, T. (2007). Prevalence and causes of iron deficiency anemias in infants aged 9 to 12 months in Estonia. Medicina (Kaunas, Lithuania), 43, 947–952. [PubMed] [Google Scholar]
  50. Vincelet, C. , & Foucault, C. (2005). Measuring iron levels relative to the type of milk consumed within a population of 16 to 18 month old French infants. Santé Publique, 17, 339–346. [DOI] [PubMed] [Google Scholar]
  51. Wang, B. , Zhan, S. , Gong, T. , & Lee, L. (2013). Iron therapy for improving psychomotor development and cognitive function in children under the age of three with iron deficiency anaemia. Cochrane Database of Systematic Reviews, 6, CD001444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. World Health Organization . (2001). Iron deficiency anaemia. Assessment, prevention and control. A guide for programme managers. Retrieved 2016, from http://www.who.int/nutrition/publications/en/ida_assessment_prevention_control.pdf
  53. World Health Organization/Centers for Disease Control and Prevention technical consultation on the assessment of iron status at the population level . (2007). Assessing the iron status of populations. Geneva, Switzerland: World Health Organization (WHO) Press. [Google Scholar]
  54. Zimmermann, M. B. , & Hurrell, R. F. (2007). Nutritional iron deficiency. Lancet, 370, 511–520. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S1. Strengthening the Reporting of Observational Studies in Epidemiology –STROBE Statement — checklist of items that should be included in reports of observational studies

Table S2. Participant characteristics, association between serum ferritinemia (SF) and socio‐demographic factors for iron deficiency threshold 12 μg/l (taking into account the Saturn‐Inf two‐stage sampling design)

Table S3. Characteristics of participants in the 2010 French National Perinatal survey, the total Saturn‐Inf study, and excluded and excluded Saturn‐Inf participants

Table S4. Sociodemographic characteristics of participants and association with serum ferritinemia (SF) (without taking into account Saturn‐Inf sampling design)

Table S5. Participant characteristics, association between serum ferritinemia (SF) and sociodemographic factors for iron deficiency (ID) threshold 12 μg/l (without taking into account the Saturn‐Inf sampling design)

Figure S1. Flow chart of participants in the study

Click here for additional data file. (90.1KB, docx)

Articles from Maternal & Child Nutrition are provided here courtesy of Wiley

RESOURCES