Relationship between the iron status of pregnant women and their newborns

OBJECTIVE: To determine the relationship between iron nutritional status of pregnant women and their newborns using a combination of hematological and biochemical parameters for the diagnosis of iron defi ciency. METHODS: A cross-sectional study was conducted in Jundiaí, Southeastern Brazil, in 2000. Venous blood samples collected from 95 pregnant women and from their umbilical cord and used for the determination of complete blood count, serum iron, total iron-binding capacity, serum ferritin, zinc protoporphyrin, and transferrin saturation. Women were classifi ed into three groups: anemic, iron defi cient and non-iron defi cient. Statistical analysis included the Tukey-HSD test, Pearson’s correlation coeffi cient and multiple linear regression analysis. RESULTS: Among pregnant women, 19% were anemic (97.9% mildly anemic and 2.1% moderately anemic) and 30.5% were iron defi cient. No signifi cant difference was seen in mean values of any parameter studied between newborns in the three groups (p>0.05). Multiple linear regression analysis showed weak association between neonatal and maternal parameters. CONCLUSIONS: The iron nutritional status of pregnant women with iron defi ciency or mild anemia does not seem to have a signifi cant impact on the iron levels of their children.

Iron defi ciency anemia is the most frequent nutritional defi ciency in pregnancy, with an impact on maternal and fetal morbidity and mortality.It is regarded as the most important preventable cause of perinatal complications, such as premature delivery, intrauterine growth retardation and neonatal and perinatal death. 1,20 is still not clear whether iron defi ciency in pregnant women might lead to a defi cient iron status of their children.Many studies have supported the belief that iron transport from the mother to their fetus occurs independently of maternal iron levels, and that it might even induce defi ciency in the mother as a result of fetal "parasitism". 9,13,18However, later studies have questioned this belief and suggested that maternal iron deficiency can cause depletion of fetal iron stores. 3,8,14,16,17No consensus regarding this subject has been reached thus far.
The difficulty in establishing a precise diagnosis of the iron status of pregnant women represents a complicating factor in the understanding of the relationship between maternal and fetal iron levels.The physiological changes that occur during pregnancy (increased plasma volume and erythropoiesis) have a signifi cant impact on hematological and biochemical parameters available for the assessment of iron status.Hemoglobin concentration is the most used parameter to detect anemia in public health care services due to its low cost and available reference standards.Therefore, a combination of several parameters has been proposed in order to improve the diagnosis of iron defi ciency. 5,23e objective of the present study was to determine the relationship between the iron status of pregnant women and their newborns using a combination of several hematological and biochemical parameters for the diagnosis of iron defi ciency.

METHODS
A cross-sectional study was conducted including pregnant women/newborn pairs from two public hospitals in Jundiaí, Southeastern Brazil.All pregnant women had single children born at term (37 to 42 weeks) as determined by the method of Capurro et al, 4 performed by a trained researcher and confi rmed by medical examination.Women with systemic diseases (toxemia, hypertension, diabetes mellitus, and heart disease) were not included in the study.

INTRODUCTION
The sample size (N=130) was estimated using the t-test table to compare means of continuous variables 7 at α=1% and β=5%, and considering 30% for the prevalence of anemia in pregnant women and 10% to compensate for possible losses.There were fi rst studied 135 pregnant women: 23 had hemolytic blood or insuffi cient blood volume; 15 women had hemoglobinopathies and thalassemia, determined by hemoglobin electrophoresis at pH 8.6 and measurement of fetal hemoglobin; and two had postpartum complications.Therefore, the fi nal sample comprised 95 pregnant women/newborn pairs.Venous blood samples were collected by a trained researcher from pregnant women during labor and from their umbilical cord immediately after clamping into Vacutainer tubes containing EDTAK 3 as anticoagulant and into tubes without anticoagulant (Becton Dickinson Vacutainer System, USA).The following hematological and biochemical parameters were determined: 1) complete blood count using the Cell Dyn 3000 automated counter (Abbott, USA), 2) serum iron determined by a colorimetric method using the Synchron CX system (Beckman, USA), 3) total iron-binding capacity determined by a turbidimetric method using the Cobas Mira Plus system (Roche, USA), 4) serum ferritin measured by chemiluminescence in an Immulite apparatus (Immulite, DPC, USA), and 5) zinc protoporphyrin (ZPP) measured with a Protofluor Z hematofluorometer (Helena Laboratories, Beaumont, TX).Transferrin saturation was estimated based on the ratio between serum iron and total iron-binding capacity.
Pregnant women were divided into three groups: 1) anemic, 2) iron defi cient, and 3) non-iron defi cient.A combination of hematological and biochemical parameters (hemoglobin, ferritin, transferrin saturation and ZPP) was used for the classifi cation of the three groups according to the model proposed by Cook & Finch. 5Pregnant women showing low hemoglobin levels (<11 g/dL) and with at least one of the other parameters altered (ferritin <10 μg/dL or transferrin saturation <16% or ZPP >60 μmol/mol heme) were considered to be anemic.Pregnant women with normal hemoglobin values (≥11 g/dL) and with at least two of the other parameters altered were classifi ed as iron defi cient, and those with a maximum of one altered parameter (except for hemoglobin) were considered to be non-iron defi cient.
The EpiInfo and Statistica software programs were used for statistical analysis.Differences in mean parameters between the women and their newborns in the three groups were determined by two-factor analysis of variance, with "group" being one factor and "pregnant women and newborn" the other factor.Multiple comparison analyses were performed by the Tukey-HSD test.The assumption of normality was tested by the Kolmogorov-Smirnov test, with the variables ferritin, ZPP and serum iron not showing a normal distribution and therefore being converted into a logarithmic scale before statistical analysis.Pearson's correlation coeffi cient was used for correlation analysis.Multiple linear regression analysis was applied to determine whether newborn parameters could be explained by maternal parameters.A p-value <0.05 was considered to be signifi cant for all tests.
The research protocol was approved by the Ethics Committee of the Faculdade de Saúde Pública, Universidade de São Paulo, and the women signed an informed consent form agreeing to participate.

RESULTS
Nineteen percent (19%) of the pregnant women studied were anemic, 30.5% were iron defi cient and 50.5% were non-iron defi cient.It should be emphasized that among anemic women, 33.3% did not take any iron supplements during pregnancy, while 66.7% did so during at least one trimester of gestation.Among iron defi cient pregnant women, 14.3% did not use any iron supplements and 85.7% took supplements during at least one trimester.In contrast, only 8.3% of noniron defi cient pregnant women did not take any iron supplements, while 91.7% used supplements during at least one trimester.
Table 1 shows the characteristics of pregnant women and their newborns in the three groups studied, showing no signifi cant differences in the variables studied.Comparison of the variables between anemic, irondeficient and non-iron deficient pregnant women showed a gradual decline in both hematological and biochemical parameters according to the course of iron defi ciency (except for RDW and ZPP which tended to increase).Furthermore, no signifi cant differences (p>0.05) in ferritin levels or transferrin saturation were seen between anemic and iron-defi cient women (Figure).With respect to parameters determined in The distribution of the hematological and biochemical parameters of maternal and umbilical cord blood is shown in Figure .Comparison of these parameters in each group revealed signifi cantly higher values for umbilical cord blood than for maternal blood (p<0.05),except for red cell distribution width (RDW) and ZPP.
Higher RDW and ZPP values in cord blood were only seen in the non-iron defi cient group (p<0.05).

Figure.
Boxplots of hematological parameters in pregnant women and newborns according to the iron status of the women (anemic, iron defi cient and non-iron-defi cient).Southeastern Brazil, 2000.
Multiple linear regression analysis showed that the parameters seen in newborn children (hemoglobin, ferritin, ZPP and transferrin saturation) could not be explained as a function of the same maternal parameters.As shown in Table 3, only the model considering newborn transferrin saturation as the dependent variable had a statistically significant association with other maternal parameters; however, the low r 2 value obtained suggests that this association was probably a random event.

DISCUSSION
Iron deficiency is the most common nutritional defi ciency in pregnancy and has an important impact on maternal and fetal morbidity and mortality.In the present study, practically half of the pregnant women studied had at any stage iron defi ciency even if anemia had not been manifested.The prevalence of iron defi ciency was very high, especially given that this condition can be prevented by dietary guidance in combination with the use of iron supplements during pregnancy.The present results show that the prevalence of anemia and iron defi ciency were markedly lower when the mother took iron supplements during at least one trimester of gestation.
Various studies have been carried out to determine the relationship between anemia and iron defi ciency in pregnant women and their newborns, but the results are inconsistent and inconclusive.It has been currently discussed whether maternal iron nutritional status has an impact or not on the iron status of their children at birth and throughout childhood. 10,11,13,14,21,22,24 is known that the fetus receives iron indirectly from the maternal circulation through a rapid and unidirectional process.Maternal iron is transported to the placenta bound to a protein, maternal transferrin, which does not cross the placenta, transferring this mineral element to placental receptors.Iron then binds to a protein with the same structure, fetal transferrin, and is transported through the bloodstream to fetal tissues. 15n general, iron-related hematological and biological parameters are markedly higher in umbilical cord than maternal blood. 15,19The present study corroborates this observation for most of the parameters studied.However, ZPP values would be expected to be higher in maternal than cord blood, since this parameter shows an inversely proportional relationship to body iron levels.Yet ZPP levels were signifi cantly higher in newborns, whereas normal levels were found in the pregnant women.One possible explanation is the fact that, according to some authors, 2,12 this parameter is invariably increased in newborns irrespective of their iron levels.In the present study, all parameters analyzed were statistically similar in children born to anemic, iron defi cient and non-iron defi cient pregnant women.This result suggests that maternal iron nutritional status at the end of pregnancy does not have an impact on iron levels of the newborn.Although not signifi cant, the biochemical parameters analyzed showed a discrete and gradual increase as the iron nutritional status of pregnant women improved.,18 In a case-control study carried out in Jordan, Kilbride et al 11 investigated the relationship between maternal anemia and iron status of newborns at birth (umbilical cord) and at 3-4, 6, 9 and 12 months of age.The authors found that iron content in cord blood was similar in anemic and control pregnant women, with no signifi cant difference in hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration or plasma ferritin.However, the rate of anemia was higher in children born to anemic pregnant women than in those born to non-anemic pregnant women at 3-4 (6% vs. 1%), 6 (43% vs. 28%), 9 (68% vs. 48%) and 12 months of age (81% vs. 65%).According to other investigators, maternal anemia is an important risk factor that can trigger anemia during infancy, even if no differences in hematological or biochemical parameters are seen in newborns at birth.
It should be emphasized that no case of severe anemia was identified in the present study, with anemia being classifi ed based on criteria of the World Health Organization, 6 which considers individuals with hemoglobin concentrations lower than 7 g/dL to be severely anemic.In the present study, 97.9% of the pregnant women had hemoglobin levels higher than 9 g/dL.In view of this fact, one may suppose that in cases of severely anemic pregnant women their newborn's iron levels at birth might be signifi cantly compromised.
In the present study, weak correlations were found between the hematological parameters of newborns and their mothers.Several investigators have determined the correlation between some of these parameters; however, the results vary from study to study probably because of the low specifi city of most parameters, which can be altered by different conditions, including sub-clinical infections.Rios et al, 18 Kelly et al 9 and Puolakka et al 17 did not fi nd any correlation between maternal and umbilical cord ferritin levels.Lao et al 13 reported a correlation between maternal mean corpuscular volume and mean corpuscular hemoglobin and cord serum iron (r=0.24 and r=0.25, respectively).
The multivariate regression analysis confi rmed the weak correlation between newborn and maternal parameters and suggested that the hematological and biochemical parameters of children born to mildly or moderately anemic pregnant women cannot be predicted on the basis of maternal parameters.These results suggest that the fetus is able to maintain adequate iron levels even when their mother has iron defi ciency or mild anemia.However, this behavior might be altered in cases of severely anemic pregnant women.In addition, anemia or iron defi ciency during pregnancy has been suggested to have an impact on the nutritional status of children only after the fi rst months of life, even when no signifi cant compromise is seen at birth.
Therefore, studies investigating the different stages of anemia are needed, especially moderate and severe disease, with follow-up of children from birth to at least six months of age.

Table 1 .
Age and obstetric characteristics of pregnant women and birth weight of newborns (mean and SD) according the iron status of the women (anemic, iron defi cient and non-iron defi cient).Southeastern Brazil, 2000.

Table 2 .
Pearson's correlation coeffi cients (r) between iron parameters of pregnant women and their respective newborns.Southeastern Brazil, 2000.

Table 3 .
Associations between iron parameters in pregnant women and their newborns assessed by multiple linear regression analysis.Southeastern Brazil, 2000.