The contribution of iron deficiency to the risk of peripartum transfusion: a retrospective case control study

A drop in hemoglobin (Hb) level is a physiologic consequence of pregnancy due to an expanded plasma volume. The maternal serum hemoglobin typically reaches the physiological nadir at 24–32 weeks gestation [1]. Iron requirements steadily rise in each trimester, reaching a peak of 7.5 mg/d in the 3rd trimester [1]. Overall, pregnancy has a net iron loss of approximately 740 mg [2]. In Canada, 23% of pregnant women are anemic (Hb < 110 g/L in 1st and 3rd trimesters, < 105 g/L in 2nd trimester) [3, 4] and an estimated 85% of these cases are attributable to iron deficiency [5]. In its early stages, iron deficiency can occur without anemia. Ferritin is the best marker of iron deficiency in pregnancy and a serum ferritin less than 30 μg/L has a sensitivity of 92% for iron deficiency with a positive predictive value of 83% [6]. Recent Canadian data showed that over 73% of pregnant women had ferritin levels < 30 μg/L at first obstetrical evaluation [7].

Maternal health is impacted by anemia during pregnancy. A recent analysis describes a 2.36 fold higher risk of maternal death in women with severe anemia (hemoglobin < 70 g/L) [8]. The risk of postpartum depression is also higher in women with iron deficiency anemia (IDA) [9]. Given iron’s critical role in DNA synthesis and cellular metabolism [10], it is not surprising that it also plays a crucial role in fetal development. In fact, maternal IDA increases the risk of prematurity [11, 12], fetal intrauterine growth restriction [11, 13], developmental delay [14] and perinatal mortality [12, 13]. The risks associated with anemia in pregnancy correlate best with hemoglobin levels tested in the first trimester, prior to significant plasma expansion [15].

In addition to these adverse outcomes, treating iron deficiency can also prevent red blood cell (RBC) transfusions for severe, symptomatic anemia. Avoiding RBC transfusions in pregnant women and women of childbearing age is key to preventing the development of red blood cell alloantibodies and hemolytic disease of the fetus and newborn in subsequent pregnancies [16], as one in fifteen recipients of RBC transfusions will develop an alloantibody [17]. Further, any transfusion carries the risk of identification errors, viral or bacterial infection, transfusion associated circulatory overload or other transfusion reactions [18].

Given the high prevalence of iron deficiency in pregnant women, our goal was to elucidate the relative contribution of untreated iron deficiency to peripartum RBC transfusions. We present a case control study aimed at identifying the prevalence of iron deficiency or anemia in pregnant women that were transfused in the peripartum period compared to controls.

A retrospective case-control study was carried out with cases defined as all obstetric patients receiving a peripartum RBC transfusion between January, 2014 and July, 2018 during admission to the obstetrical unit at Sunnybrook Health Sciences Centre, Toronto, Canada. High risk referrals comprise 20% of the approximately 4000 deliveries per year at our institution. Cases were identified through the blood bank information system (HCLL, Wellsky, USA). Patients were excluded if they had a clear alternative explanation for their anemia (e.g., chronic kidney disease, malignancy, bone marrow failure syndrome), had a pre-existing transfusion dependence, had differential hemoglobin targets for transfusion (e.g., sickle cell disease), or did not receive the majority of their prenatal care at our institution (defined as women whose care was transferred to our institution for delivery only). Controls were defined as obstetrical patients delivering at our institution that did not receive a transfusion at the time of delivery. A single control was selected as the next age-matched delivery following the transfused case.

Demographic and laboratory data were collected from patient charts retrospectively. Transfusion data were collected from the blood bank information system. Based on the literature and biologic plausibility, risk factors for and laboratory measures associated with iron deficiency, anemia and transfusion were established [19, 20]. Demographic variables collected included age, ethnicity, multiple gestation, social work involvement, history of heavy menstrual bleeding, history of chronic malabsorption (e.g. celiac disease, inflammatory bowel disease, bypass surgery), vegetarianism/veganism, hyperemesis gravidarum, and consumption of prenatal vitamins and iron supplements during pregnancy. Patients’ past medical history, including obstetrical and gynecologic history, and dietary restrictions were collected from their prenatal record. This record is standard and mandated for all pregnancies in the province of Ontario, Canada. Median household income was assigned based on geographic location of residence by postal code using publicly available records from Statistics Canada (Ottawa, Canada). Delivery outcomes and perinatal outcomes were also collected including incidence of primary postpartum hemorrhage (defined by > 1000 mL of blood loss within 24 h of delivery as estimated on their delivery record, or documentation of postpartum hemorrhage in the chart), maternal infection, fetal demise, prematurity (< 37 weeks gestational age), low birth weight (< 2500 g), maternal ICU admissions, hysterectomies and cesarean sections (planned and unplanned). We collected components of the complete blood count (hemoglobin, mean corpuscular volume (MCV), red blood cell distribution width (RDW)) and iron studies (ferritin, transferrin saturation (TSAT) and total iron binding capacity (TIBC)) in the year prior to pregnancy, during pregnancy and up to 6 months postpartum.

The primary objective was to determine the proportion of patients with iron deficiency in pregnancy (ferritin < 30 μg/L, TSAT < 20% or last documented MCV < 80 fL when previously normal) or anemia in pregnancy (Hb < 110 g/L in 1st or 3rd trimester or < 105 g/L in 2nd trimester) in cases and controls.

The secondary objectives were to determine the rate of suboptimal care as defined by any of the following quality indicators: 1) Unaddressed anemia in pregnancy (defined by anemia (as above) without either the measurement of a ferritin or transferrin saturation and total iron binding capacity, or without iron treatment documented in the antepartum record; and absence of a known red cell disorder as documented in laboratory information system or antenatal record. The standard of care in Canada is to offer hemoglobin electrophoresis or high performance liquid chromatography to at-risk women [21]); 2) Identified but uncorrected iron deficiency in pregnancy: clear iron deficiency (ferritin < 30 μg/L) without documentation of commencing oral iron replacement at a dose of at least 30 mg of elemental iron or administration of intravenous iron; 3) Identified but uncorrected iron deficiency anemia in pregnancy: clear iron deficiency anemia (ferritin < 30 μg/L and anemia (defined above)) without documentation of commencing oral iron replacement at a dose of at least 30 mg of elemental iron or administration of intravenous iron; 4) Mild anemia (Hb 100–109 g/L in 1st and 3rd trimester, Hb 100–104 g/L in 2nd trimester) and moderate anemia (Hb 90–99 g/L) in pregnancy; and 5) Hb < 90 g/L postpartum with uncorrected iron deficiency (defined above) at last measurement in the current pregnancy.

Descriptive statistics were summarized using medians and interquartile ranges for continuous variables and proportions for categorical variables. To identify statistically significant predictive factors associated with peripartum transfusion, univariate logistic regression was performed with case/control status as the dependent variable. We included predefined risk factors for iron deficiency, anemia and transfusion which were selected based on biologic plausibility and sufficient data. Statistically significant variables at a threshold of alpha 0.05 were included in a multivariate logistic regression model to quantify predictive factors associated with transfusion at delivery. Maternal and fetal outcomes were excluded from our multivariate analysis in order to focus on preexisting risk factors for transfusion. Odds ratio (OR) and the 95% confidence interval (CI) were calculated for each predictive factor. All statistical analyses were performed using Statistical Analysis Software (SAS version 9.4; Cary, NC).

This study was approved by the Sunnybrook Health Sciences Centre Research Ethics Board, #221–2018; this approval allowed access to the data used for this research.

To our knowledge this is the first study to assess the burden of iron deficiency in women transfused in the peripartum period. We find that transfused women were more than 2-times more likely to have iron deficiency or anemia in pregnancy. Women treated with intravenous iron in pregnancy (OR 3.8), women who delivered by unplanned cesarean section (OR 2.8), and women being followed by a social worker (OR 4.1) were also found to have increased odds of peripartum transfusion. In addition, 77% of women received only 1–2 units of red blood cells. These findings suggest that there is an opportunity to identify and treat iron deficiency before and during pregnancy to mitigate the risk of transfusion and that coordination between physicians and social workers may be a useful strategy to identify and reach women at higher risk.

At our institution, 70% of pregnant women had a ferritin checked and only 65% with documented iron deficiency were prescribed oral iron supplementation. The American College of Obstetricians and Gynecologists (ACOG) [22], Centers for Disease Control (CDC) [23], and World Health Organization (WHO) [4] all recommend screening asymptomatic pregnant women for IDA using serum hemoglobin and ferritin levels, and universal iron supplementation during pregnancy with a dose of 30–60 mg/day of elemental iron [23, 24]. In anemic women, they suggest increasing the dose to 60–120 mg/day of elemental iron [23]. A consensus statement from the Network for the Advancement of Patient Blood Management, Haemostasis and Thrombosis (NATA) suggests daily oral iron (30-60 mg) supplementation in areas with a high prevalence of anemia in pregnancy and screening for iron deficiency in all non-anemic women at risk of ID [25]. The U.S. Preventative Task Force cites a lack of evidence to support routine screening or routine iron supplementation in pregnancy [26]. Perhaps as a result of this ambiguity, less than one third of American obstetrician-gynecologists have been shown to provide routine iron supplementation to their pregnant patients [27]. The NATA consensus statement also supports intravenous iron consideration in women with ID and Hb < 80 g/L or in newly diagnosed IDA after 34 weeks gestation [25]. In our study patients, however, we noted a delay in the initiation of intravenous iron in pregnancy: three-quarters of intravenous iron was given within 3 weeks of delivery, leaving little time for erythropoiesis to correct the red cell mass. The reasons for this delay are likely multifactorial, including a lack of physician comfort with prescribing intravenous iron in pregnancy, difficulty accessing the drug, and challenges in facilitating drug infusion. When possible, we recommend earlier consideration of intravenous iron in women not responding to oral iron, not tolerating oral iron or with moderate to severe anemia to ensure sufficient time to correct the anemia before delivery.

Despite a publicly funded healthcare system in Ontario, oral and intravenous iron are not universally funded as general list products and as such coverage is provided out-of-pocket, through third-party private insurance plans or via exceptional access public funding. However, this does not address the “working poor” who cannot pay out-of-pocket, do not have private insurance and also do not meet criteria for exceptional access public funding. This results in health inequity between high and low income families and likely contributes to the higher incidence of untreated iron deficiency in those of lower socioeconomic status [28]. Moreover, these women are less likely to eat a diet rich in iron as many of these foods are particularly expensive (e.g. red meat). Since children born to iron deficient mothers have higher rates of developmental delay [14] and impaired self-regulation [29], there is a compounding disadvantage of being born into a low-income household. Further, these babies may represent a more medically complex population since in our study, babies born to transfused mothers had higher rates of prematurity and lower birth weights. We also found that Caucasian women had lower peripartum transfusion rates which is in keeping with prior literature [30]. Multiple factors including diet, socioeconomic status and genetics may be at play since Black women have lower hemoglobin levels [31] and non-Caucasians have higher rates of hemoglobinopathies [32, 33]. Further research is required to fully understand this phenomenon and the maternal and fetal benefits of removing barriers to IV iron in pregnancy.

The peripartum transfusion rate at our institution (0.9%) was comparable to that in other high resource settings [19, 20]. Interestingly, 44% of transfused women received just one unit of red blood cells, suggesting that many of the transfusions given were potentially avoidable with diligent prevention of iron deficiency anemia prior to delivery. Avoiding blood transfusions in women of childbearing age is essential to preventing complications such as alloantibody formation and hemolytic disease of the fetus and newborn in future pregnancies. Postpartum transfusion also has other important implications for women: those transfused 1 to 2 units of blood after obstetrical hemorrhage have longer hospital stays, higher ICU admission rates and hospital readmission rates and lower breastfeeding rates [34]. Further, when postpartum women are transfused in the absence of hemorrhage, their risk of severe morbidity with subsequent pregnancies is over four fold [35].

Next steps include clarifying the factors that contribute to higher peripartum transfusion rates in non-Caucasian women and women with complex social situations and quality improvement initiatives to target earlier and more intensive iron replacement in pregnant women. Additional studies need to be performed to understand the proportion of transfusions in these subgroups of patients that can be prevented with oral and intravenous iron. Prospective studies are needed to determine if aggressive iron management in pregnancy will reduce the proportion of women transfused and impact important maternal and fetal outcomes.

In summary, iron deficiency in pregnancy is common, under-recognized and undertreated. It has implications for mothers and babies and predisposes women to blood transfusions which may have severe consequences for future pregnancies. Women being followed by a social worker appear particularly vulnerable and warrant more attention paid to iron replacement and we hypothesize that social and financial barriers may be impairing access to treatment. We advocate for early recognition of iron deficiency by screening all pregnant women with a first trimester ferritin, diligent use of early oral iron replacement, and improved access to intravenous iron replacement in pregnancy for all women irrespective of their financial resources.