Prevalence of Iron Deficiency and Iron Deficiency Anemia in <scp>US</scp> Pregnant Individuals, 2003–2023

Iron deficiency with and without anemia are common in pregnancy and associated with significant maternal and perinatal morbidity and mortality despite widespread availability of effective treatment [1]. Although the majority of literature focuses on the negative effects of iron deficiency anemia, iron deficiency without anemia is also deleterious for maternal and neonatal health. For the mother, iron deficiency without anemia is associated with increased preterm birth, low birth weight and small for gestational age infants, and development of iron deficiency anemia [2]. For the child, non-anemic iron deficiency early in pregnancy is associated with low iron status at birth and worse language and motor outcomes at 2 years of age [3]. Screening historically has focused on anemia with the American College of Obstetrics and Gynecology (ACOG) recommending complete blood count screening in the first trimester and again between 24 and 28 weeks gestation [4]. The United States Preventive Services Task Force (USPSTF) concluded that there is insufficient evidence to support screening for iron deficiency with or without anemia in asymptomatic pregnant persons [5]. These recommendations are at odds with those of the International Federation of Obstetrics and Gynecology who advise dedicated iron assessment as early as possible in pregnancy and treatment of iron deficiency with or without anemia [6]. We previously reported that almost 40% of 12–21 year old females are iron deficient [7] and hypothesize a higher prevalence in pregnant persons given accumulation of menstruating years and blood loss and the increased iron demands of pregnancy. To assess this hypothesis, we examined iron deficiency prevalence among pregnant persons in the National Health and Nutritional Examination Survey (NHANES) dataset between 2003 and 2023 to inform future screening practices. NHANES is a complex survey administered through the United States Centers for Disease Control and Prevention in two-year cycles with interview, examination, and biospecimen components designed to be nationally representative of a non-institutionalized civilian population. It receives ethics approval from the National Center for Health Statistics review board. We identified pregnant persons, defined as those with a positive urine pregnancy during NHANES examination, from all years in which ferritin was measured (2003–2010 and 2015–2023). Gestational age data were not available. We identified the proportion with iron deficiency (defined as ferritin < 30 μg/L in primary analysis [4], 15 μg/L [World Health Organization (WHO) definition] and 50 μg/L [physiologically supported threshold in non-pregnant females [8]] when stated elsewhere as sensitivity analyses). We then described the prevalence of anemia (defined as hemoglobin < 11 g/dL per WHO definition) and iron deficiency anemia (hemoglobin of 11 g/dL and ferritin < 30 μg/L). Finally, we used quasibinomial generalized linear models to generate independent adjusted odds ratios to assess associations among age, body mass index, race/ethnicity, income-to-poverty ratio, and education level with having iron deficiency with or without anemia. Unless otherwise specified, counts are unweighted while proportions and statistical analyses are weighted using the survey package in R version 4.4.1. Of 953 NHANES participants who had a positive pregnancy test, 726 (corresponding to a weighted sample of 2.3 million pregnant people) had complete data (Table 1). The overall prevalence of iron deficiency was 52.9% (95% CI 47.6%–58.1%); in sensitivity analyses prevalence ranged from 26.9% (95th CI 22.5%–31.3%) for ferritin < 15 μg/L to 73.0% (95th CI 67.9%–78.1%) for < 50 μg/L. Prevalence was similar between the two time periods (2003–2010 and 2015–2023). The prevalence of anemia from any cause was 9.6% (95% CI 6.5–12.6), while 6.9% (95% CI 4.4–9.5) of participants had iron deficiency anemia. Of patients with iron deficiency, 86.9% (95% CI 82.2–91.5) had hemoglobin > 11 g/dL, meaning they would be missed with hemoglobin screening alone. In adjusted analyses, no demographic characteristics were independently associated with iron deficiency (Table 2). Only income to poverty ratio < 150% was associated with iron deficiency anemia (OR 2.78, 95% CI 1.25–6.16, p = 0.012). Among pregnant persons in the United States between 2003 and 2023, iron deficiency affected 53% and iron deficiency anemia affected 7%, with variation depending on the ferritin threshold used. Income to poverty ratio was associated with iron deficiency anemia. It is important to note that the prevalence of iron deficiency cited in the latest USPSTF recommendations was much lower at 18% (~30% in the third trimester), using NHANES data from 1999 to 2006 and an overly restrictive iron deficiency definition of total body iron < 0 mg/kg [5, 7]. The prevalence in this study of the general population was similar to that seen in a Canadian study (52.8%) that used the same ferritin cutoff to define iron deficiency but described the prevalence of iron deficiency only in those patients who were screened by their providers (59.4% of pregnant patients in the study had iron screening) [9]. The primary limitation of this study is lack of gestational age data which may lead to underestimation of prevalence as iron deficiency and iron deficiency anemia both increase over the course of pregnancy with prevalence rates of iron deficiency as high as 84% by 33 weeks gestation in a high resource setting [10]. Prevalence may also be underestimated as iron deficiency was defined using ferritin, an acute phase reactant, and therefore may miss affected individuals with underlying inflammation. Given the higher than previously reported prevalence of iron deficiency without anemia, current ACOG guidelines recommending the use of complete blood count for screening will miss many individuals with iron deficiency without anemia. These patients are at risk for the negative consequences of iron deficiency without anemia, as well as progression to iron deficiency anemia and its associated adverse outcomes. Lack of a recommendation to repeat screening closer to term, when the prevalence is highest, also likely leads to underdiagnosis and ultimately undertreatment. As likelihood of iron deficiency screening in pregnancy is associated with socioeconomic status [11], lack of universal iron specific screening recommendations is also likely to perpetuate existing health disparities. Iron deficiency with and without anemia in pregnancy is incredibly prevalent and has been shown to be harmful to mothers and their babies despite effective available therapies. Data on ideal timing and frequency of screening, as well as optimal diagnostic thresholds is lacking. Further study is needed to identify best practices for screening and associated outcomes after screening implementation. The authors have nothing to report. NHANES receives ethics approval from the National Center for Health Statistics review board. The authors declare no conflicts of interest. All data are freely accessible online at https://wwwn.cdc.gov/nchs/nhanes/default.aspx.