Prevalence of Iron Deficiency and Iron Deficiency Anemia

Iron deficiency is the most common nutritional disorder worldwide. Severe or prolonged iron deficiency can cause iron deficiency anemia (IDA). The prevalence of IDA is sensitive to the age at testing and the diagnostic criteria used.

The hemoglobin concentration and hematocrit are the principal screening tests for detecting anemia. Hemoglobin can be measured quickly and accurately on a few drops of blood. ³ Data on infants aged 6–12 months are sparse. For infants aged 1–2 years and 3–5 years, most studies use cut-offs for serum hemoglobin (Hgb) of <110 g/L and <112 g/L, respectively. Typical cut-off values for females are <118 g/L for 12- 14-year-olds and <120 g/L for 15–39-year-olds. ^{4, 5}

These cut-off values were chosen by consensus or based on statistical analysis of the distribution of laboratory values in the population. Some experts argue that normal limits for Hgb and for iron studies should be based on analysis of the response to iron therapy, but efforts to define cut-off values in this manner have not yielded definitive results. ^{6 – 8}

Most cases of anemia are due to causes other than iron deficiency. When anemia is diagnosed, additional tests can determine whether iron deficiency is the cause. Centers for Disease Control and Prevention (CDC) analysts diagnose iron deficiency when two or more of the following tests are abnormal: free erythrocyte protoporphyrin (≥1.24 μmol/L red blood cells), transferrin saturation (<14% for 12–15 year-olds or <15% for 16–39 year-olds), and serum ferritin (<12 μg/L). ^{4, 9}

While the CDC criteria are arbitrary, they have been used consistently across several analyses of the National Health and Nutrition Examination Survey (NHANES) (1988-1994 and 1999-2000), making comparisons across time and between demographic groups possible. Table 1 shows that the prevalence of IDA in infants aged 1–2 years (2% to 3%) and in females aged 12–19 years (2%) did not change substantively between these time periods.

Table 1

Prevalence of iron deficiency anemia in selected populations - United States, National Health and Nutrition Examination surveys, 1988-1994 and 1999-2000*.

Not all studies use the CDC case definitions. The positive predictive value of a low hemoglobin for IDA varies with age and with the cut-off values used for case definition. Among children aged 12–35 months in NHANES III, the positive predictive value of Hgb concentration <110 g/L for iron deficiency was 29% (95% CI, 20–38%), and the sensitivity was 30% (95% CI, 20–40%). Changing the diagnostic cutoff point to Hgb <107 g/L resulted in a positive predictive value of 38% (95% CI, 24–52%) but lowered the sensitivity to 15% (95% CI, 7–22%). ⁸

Table 2 illustrates how the positive predictive value varies with age and with the cut-offs used to define IDA. In the Avon longitudinal study of pregnancy and childhood (ALSPAC), investigators developed criteria for the diagnosis of IDA based on the distributions of Hgb and ferritin levels in their own sample. By these (ALSPAC) criteria, 5% of infants 12 months or 18 months of age had a low Hgb value, and 10% or 12% of these infants, respectively, proved to have iron deficiency anemia. Using the World Health Organization (WHO) or Institute of Medicine (IOM) criteria, the apparent prevalence of anemia was between 17% and 18%, but the prevalence of IDA and the positive predictive value of a low Hgb value were much lower in infants 12 months of age than at 18 months of age.

Table 2

Percentage of infants with iron deficiency anemia at 12 and 18 months of age using different case definitions*.

The prevalence of iron deficiency anemia among pregnant women is uncertain. Data from NHANES II suggest that <2% of nonpregnant women aged 20–44 years had IDA in the late 1970s. ¹² The Pregnancy Nutritional Surveillance System (PNSS) (http://www.cdc.gov/pednss/publications/index.htm) has published annual rates of low Hgb or hematocrit (Hct) in a primarily low-income, pregnant U.S. program-based sample. These data indicate that the prevalence of anemia in the third trimester has not changed since the 1980s, but PNSS does not distinguish anemia related to iron deficiency from other causes. A surveillance program in Camden, New Jersey estimates that, in a low-income, mostly minority population, rates of IDA are 1.8% in the first trimester, 8.2% in the second trimester, and 27.4% in the third trimester. ¹³

Risk factors for Iron Deficiency Anemia in Different Groups

Iron deficiency without anemia is a precursor to IDA. Factors that cause iron deficiency include inadequate iron intake or absorption, or increased iron requirements due to growth or to loss of iron from bleeding. Most people who have iron deficiency never develop anemia. However, if iron deficiency is severe or prolonged, depletion of iron stores can cause inadequate hemoglobin production and anemia.

The prevalence of IDA varies with age, sex, race, dietary intake, and socioeconomic factors. In the United States, the prevalence is higher among black and Mexican-Americans than among whites. Reliable estimates of rates of IDA in different subgroups are lacking, but good data on the prevalence of iron deficiency (with and without anemia) are available from NHANES These data indicate that age-, race-, and gender-specific prevalences of iron deficiency in the U.S. population did not change substantially between 1990 and 2000 (see Appendix Table 1). ¹⁰

As discussed below, other factors affect the risk of developing IDA in specific age and gender groups.

Risk factors among infants. The risk of iron deficiency anemia is high during the second year of life because of increased iron requirements related to rapid growth. ^{11, 14 – 16} Premature and low birth weight infants and infants with history of prolonged stay in the neonatal unit are at particularly high risk of developing iron deficiency anemia before 1 year of age. ¹⁷ Among term infants younger than 1 year, however, the prevalence of IDA is low, and Hgb and serum ferritin are uncorrelated. ^{7, 18}

Risk factors for developing IDA in the second year of life include the use of non-iron-fortified formula in the first year of life (without therapeutic iron supplementation); exclusive breastfeeding with no or erratic iron supplementation after 6 months of age; and the introduction of cow's milk before 1 year of age. ⁵ The prevalence of IDA increases between 12 and 18 months of age as these factors come into play.

At present, about 97% of formula sold in the United States is iron-fortified. ¹⁹ Randomized and nonrandomized controlled trials, observational studies, and time series studies have demonstrated substantial reductions in the incidence of iron deficiency and IDA in healthy infants fed iron-fortified formula, iron-fortified cereal, or breast milk with iron-fortified cereal added at 4–6 months, compared with infants fed cow's milk or unfortified formula.

U.S. data on the impact of race, ethnicity, and socioeconomic factors on the risk of developing IDA in infancy are surprisingly sparse. The Pediatric Nutrition Surveillance System (PedNSS) measures hemoglobin levels in a national sample of infants from families participating in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), but does not perform iron-related measures. In the 2003 PedNSS report, the most recent to be published, 16.2% of infants aged 6–11 months had Hgb < 110 g/L, and 15% of children aged 12–17 months had Hgb < 110 g/L. The overall prevalence of anemia in PedNSS children declined from 15.8% in 1994 to 12.8% in 2003. The highest prevalence of anemia was among black infants (19.0%). ²⁰ This survey establishes that black infants have a higher risk of anemia, but the proportion of cases that are related to iron deficiency is unknown.

In developing countries, and therefore among some groups of immigrants to North America, blood loss due to parasitic infection or malaria is a common cause of iron deficiency. ²¹ Native American infants and recent immigrants from Cuba are also at risk for IDA. A study of First Nations communities in Canada determined the prevalence of anemia (defined as Hgb < 110 g/L) among 9-month-old infants to be 31.9%, and estimated that the prevalence of IDA to be 5.6% to 10.8%, based on Hgb < 110 g/L and a low mean cell volume as proxy measures for IDA. ²² A 1998 Pan American Health Organization report estimated that IDA affects 40% to 50% of Cuban children aged 1–3 years. ²³

Risk factors among adolescent girls and adult women. Females of childbearing age require additional iron. Heavy menstrual blood loss (≥ 80 mL/mo) and pregnancy are associated with higher iron requirements. ^{5, 24}

Race, income, education, and other socioeconomic factors are associated with IDA in girls and women. In NHANES III, Mexican-American women aged 12–39 years were at higher risk of having IDA (6.2% ± 0.8%) than non-Hispanic white women of the same age (2.3%± 0.4%), a difference that was marked among poor women but small for women with higher household incomes (Table 3) ⁹ and which could not be accounted for by differences in dietary intake of iron. We did not find an analysis of risk factors among black women.

Table 3

Prevalence of iron deficiency anemia in relation to poverty in Mexican-American and non-Hispanic white women aged 12–39 years.

Eating disorders are also associated with IDA. An analysis of NHANES III data on 9698 children aged 2–16 years found that overweight and obesity were associated with a higher risk of IDA; in a logistic regression model controlling for age, gender, ethnicity, poverty status, and parental education level, children who were overweight were 2.3 times as likely to be iron-deficient (2.3; 95% CI, 1.4–3.9, respectively) as were those who were not overweight. ²⁵ Adolescent girls who try to control their weight may inadvertently limit their iron intake. In Britain in the 1980s, the prevalence of IDA in adolescent girls was higher among girls who bought snacks at local shops instead of eating school lunches or bringing food from home. ¹⁷

Complications of Iron Deficiency Anemia

As early as the 1960s, researchers demonstrated that, in general, decreased hemoglobin alone does not have readily apparent adverse effects unless it is below 10 g/dL (100 g/L). ^{26 – 28} Persons with markedly reduced hemoglobin levels are at risk for cardiopulmonary and other complications. Screening is intended to find milder degrees of anemia before such complications have developed.

Infants and children. Several cross-sectional and case-control studies have demonstrated an association between IDA and psychomotor and cognitive abnormalities and poor school performance in children. ^{17, 29 – 35} For example, in a recent cross-sectional analysis of NHANES III data, 71% of iron-deficient children had below-average math scores, versus 49% of children who had normal iron status. ³⁴ Scores of tests on reading, block design, and digit span did not differ. After adjustment for age, gender, race, poverty status, caretaker education, and lead status, iron-deficient children were 2.4 times as likely to have low math scores (95% CI, 1.1–5.2; p=0.03). The effect was strongest among girls aged 12–16.

Several causal hypotheses have been proposed to explain this association. The oldest is that the brain functions poorly in IDA because of decreased oxygen delivery to tissues. According to this theory, correction of anemia could reverse the neurocognitive deficits seen in cross-sectional studies. An alternative hypothesis is that iron deficiency leads to increased absorption of lead, which can also cause brain damage.

Another alternative hypothesis is that, in the fetus, infant, and toddler, iron deficiency may cause abnormal metabolism of neurotransmitters or hypomyelination, leading to irreversible or very slowly reversible neurocognitive deficits. Evidence for this hypothesis comes primarily from animal studies. ^{36, 37} Investigators seeking supporting evidence in humans have measured auditory brainstem responses and visual evoked potentials in a cohort of Chilean children who were diagnosed to have IDA as infants. At the time of initial diagnosis at 6, 12, or 18 months of age, infants with IDA had slower transmission through the auditory brainstem pathway than healthy controls. Although IDA was diagnosed and treated early, at 4 years of age the children who had IDA as infants still had slower transmission than healthy infants. ³⁸

A recent critical review identified seven longitudinal studies in which low hemoglobin levels in early childhood were linked to poor cognitive development or school achievement in later childhood. ³⁵ (Two of these studies were available in 1996.) The older studies were small (range 20–41 anemic children) and the iron status of the anemic children was not clear.

One of the recent longitudinal studies using records from the WIC were linked to school records in Dade County, Florida. ³⁹ The outcome variable for the analysis was mild or moderate mental retardation on the basis of criteria used by the Florida Department of Education for special education placement. About 69% of the sample (n=3,771) were black, 23% were Hispanic, and 7% were white. After adjustment for birth weight, maternal education, sex, race-ethnicity, age of mother, and age of child, there was a significant association between Hgb level at entry into the WIC program and the probability of mental retardation at age 10 (odds ratio 1.28; 95% CI, 1.05–1.60).

The other recent study ⁴⁰ was a 10-year follow-up of a cohort of Costa Rican children, the subject of previous reports in infancy and at 5 years of age. In this cohort, 48 children who had severe iron deficiency in infancy were compared with 114 children who had good iron status in infancy. At ages 11–14 years, the children who had severe iron deficiency as infants still had worse scores on intelligence tests (101.8 ±2.0 vs. 104.6±1.3) and on a variety of tests of cognitive function, despite having similar Hgb levels in adolescence. Parents of children in the severe iron deficiency group were more likely to report behavior problems.

It is difficult to prove that the relationship between anemia and developmental abnormalities in longitudinal studies is causal. Many other factors associated with abnormal neurocognitive development are also associated with iron deficiency. These include nutritional factors, such as intake of iodine, zinc, and other micronutrients; environmental factors (e.g., exposure to lead); prematurity and low birth weight; caretaker characteristics (e.g., maternal education, household income); and other socioeconomic factors. ^{35, 41} In all cross-sectional studies, iron-deficient children and their families differed in nutritional status, income, education, and other factors from the comparison groups. ³⁵ Most longitudinal studies did not include enough children to control for all environmental variables that could be associated with iron deficiency and with the outcomes. ³⁵

Socioeconomic factors are so strongly associated with cognitive outcomes, and so highly inter-correlated, that the ability of statistical adjustment to eliminate confounding is uncertain. In the Dade County study, for example, maternal education was a powerful predictor of mental retardation after adjustment for other risk factors. Compared with maternal education greater than 12 years, the adjusted odds associated with only 12 years of maternal education and less than 12 years were 8.32 (95% CI, 1.12–62.0) and 11.9 (95% CI, 1.63–88.1), respectively. In the Costa Rican study, maternal IQ and education were strongly associated with children's IQ and with cognitive abnormalities.

Screening is most likely to influence neurodevelopmental outcomes if it is done at an age when IDA is present and development is still normal. ¹⁴ Investigators from the AVON longitudinal study of pregnancy and childhood sought to identify the best age for screening by examining the relationship between serum Hgb and developmental outcomes, measured at age 18 months. ^{11, 14} Delayed development by age 18 months was associated with anemia at 8 months of age. However, most abnormalities that would lead to a diagnosis of iron deficiency without anemia resolved spontaneously by 12 or 18 months of age.

Pregnancy. Numerous observational studies have reported an association between severe to moderate anemia (hemoglobin <9–10 g/dL) and poor pregnancy outcome. ² However, the relationship between maternal iron deficiency or IDA during pregnancy and birth outcome is not well understood. Older studies, including three large, population-based studies, evaluated the relationship between Hgb or Hct and low birth weight or premature birth without assessing the iron status of the mother. Recent cohort studies ⁴² and reviews, ^{13, 43} including a critical review of studies published between 1966 and 1999, ⁴⁴ emphasize that the relationship of maternal Hgb to birth weight is U-shaped—that is, low and high Hgb values are markers for poor birth outcomes. In white women, maternal hemoglobin values of 105–125 g/L were associated with the lowest rate of LBW. For black women, the rate of low birth weight was lowest for maternal hemoglobin values of 85–95 g/L, but this estimate is based on data that are now over 25 years old. In the first trimester, IDA is associated with a greater than two-fold increase in the risk of preterm delivery. In the third trimester, however, lower Hgb and Hct levels are not associated with higher rates of low birth weight or preterm delivery.

Maternal IDA might have other complications. One prospective, longitudinal human study found an association between low umbilical cord serum ferritin concentrations and poor performance on mental and psychomotor tests at 5 years of age. ⁴⁵ Low postpartum Hgb or Hct levels may be associated with postpartum depression. ⁴⁶

Postpartum maternal IDA may also be associated with developmental delay in children. A controlled trial of iron therapy in young, South African mothers with IDA, published in 2005, compared non-anemic mothers with anemic mothers administered either placebo (25 mg ascorbic acid and 10 μg folate) or daily iron treatment (125 mg FeSO4) plus ascorbate and folate). ⁴⁷ All mothers had full-term, normal birth weight infants (n = 81) and were enrolled in the study at 6–8 weeks postpartum. At baseline, anemic mothers tended to be less responsive to, and more controlling of, their infants than non-anemic mothers. Infants of anemic mothers were delayed at 10 weeks in hand-eye movement and overall development. Infants whose mothers were anemic in the early postpartum period scored worse on developmental tests at 10 weeks and 9 months of age. At 9 months, anemic mothers in the placebo group were significantly more negative toward their babies, engaged less in goal setting, and were less responsive than non-anemic mothers in the control group.