The Role of Anemia in Term and Preterm Pregnancies: Evidence from the Brazilian Multicenter Study on Preterm Birth (EMIP)

Camilla O. Figueira1, Helena M Gomide1, José P. Guida1, Tabata Z. Dias1, Giuliane J. Lajos1, Ricardo P. Tedesco2, Marcelo L. Nomura1, Patrícia M. Rehder1, José G. Cecatti1,3, Renato Passini Jr1, Fernanda G. Surita1, Maria Laura Costa1,3*, Brazilian Multicenter Study on Preterm Birth study group* 1Department of Obstetrics and Gynecology, University of Campinas, Campinas/SP, Brazil 2Department of Obstetrics and Gynecology, Jundiaí School of Medicine, Jundiaí, Brazil 3Centre for Studies in Reproductive Health of Campinas (CEMICAMP)

Physiopathology of preterm birth is not fully understood and is described as multifactorial, associated to social and environmental conditions that can be hardly identifiable. It is estimated that around 50% of these births remain without a defined cause after childbirth [2,3]. Risk factors for prematurity include maternal malnutrition, short interpregnancy interval (6 months or less), previous preterm birth, infections, smoking and anemia [1][2][3]. Also, multiple pregnancy and conditions that can affect fetal well-being such as preeclampsia, placental dysfunction and fetal growth restriction are responsible for increasing prematurity rates in the group of piPTB [4].
As preterm birth is highly associated with neonatal morbidity and mortality and also with the presence of anemia, this secondary analysis of a national multicentric study is intended to evaluate the prevalence of and factors associated with anemia among cases of preterm and term births and its association with poor maternal and perinatal outcomes.

Methods
The present study is a secondary analysis of the Brazilian Multicenter Study on Preterm Birth (EMIP). The study was reviewed and approved by the National Council for Ethics in Research (CONEP) and by the Institutional Review Board of each participating center (#0564.1.146.000-09). All women included in the study signed an individual informed consent after understanding the conditions of the study.
The study protocol and complete methods have been previously published elsewhere [1,11]. Briefly, the EMIP study was conducted in 20 referral obstetric hospitals from the three most populated Brazilian regions: Southeast, Northeast and South. It was a prospective cross-sectional study that screened 33,740 births from April 2011 to July 2012, enrolling all cases of preterm birth and their infants. Then, causes of preterm births were separately described. In addition, it enrolled an intentional sample of term births randomly identified immediately after a preterm one.
For the main study, sample size was calculated based on the Brazilian preterm birth prevalence at the time, reported as 6.5% in 2006 with allocation of at least 1054 subjects in each group to allow for all the planned analyses [2]. Information was collected from an interview with the women during admission for childbirth and through review of their medical records until 60 days if the neonate remained hospitalized [1,11].
For this secondary analysis, women with no information about anemia and hemoglobin levels in their medical records were excluded. The remaining women were divided in two groups, according to the presence of anemia, and analysis performed for preterm (before 37 weeks), term (37 weeks or more) and comparing both conditions. Anemia was defined as hemoglobin level below 11g/dL anytime during pregnancy, according to the WHO definition [4].
The prevalence of anemia was described in both term and preterm groups, and among different types of PTB (sPTB, pPROM and piPTB). We compared maternal and perinatal outcomes of preterm and term births with and without anemia, plus, only for women with anemia, term and preterm births.
The variables under study included sociodemographic characteristics (maternal age, urban or rural area, region of the country (Southeast, Northeast, South), skin color as a proxy for ethnicity, marital status, years of schooling, previous morbidity, previous obstetric history (other children under 5 years old, previous cesarean section, interpregnancy interval , previous preterm birth, previous newborn under 2500g), adequacy of prenatal care (evaluated through onset of prenatal care at first trimester or later, number or prenatal care visits, gestational weight gain, initial body mass index, use of alcohol or drugs). We also assessed the presence of urinary tract infection, vaginal bleeding, occurrence of fetal morbidity), route of childbirth, and neonatal outcomes as birth weight, 5 th minute Apgar score <7, fetal death, neonatal ventilatory support, neonatal morbidity, intraventricular hemorrhage and neonate's condition at discharge.

Results
The EMIP study identified 4150 preterm births (≤37 weeks gestation) during data collection and selected a sample of 1156 term births as a control group. For our current secondary analysis, we excluded 434 (10.46%) and 108 (9.42%) cases from preterm and term groups respectively, due to lack of information on the hemoglobin level for identifying anemia. Among all preterm births, 1278 women (34.39%) had spontaneous preterm birth (sPTB), 1066 (28.69%) preterm premature rupture of membranes (pPROM) and 1372 women (36.92%) had a provider-initiated preterm birth (piPTB) (Figure 1).
The overall prevalence of anemia in preterm births was 33.23% while in term births was 27.74% (p=0.01).
Frequency of anemia was significantly higher among the pPROM group (36.3%, p=0.029). Among spontaneous preterm births the frequency of anemia was 32.79% and in the piPTB group it was 31.27%. Table 1 shows the maternal characteristics according to gestational age at birth and the presence of anemia. Age ≤19 years (p=0.047), non-white skin color (p=0.001), living in the Northeast region (p=0.001), ≤8 schooling years (p=0.003), history of morbidity (p=0.003) and having children under 5 years old (p=0.001), were significantly associated with anemia and preterm birth. Among the group of term childbirth, a significant association with anemia was found for women living in urban area (p=0.001) and with non-white skin color (p=0.015). Furthermore, comparing only the women with anemia, between preterm and term pregnancies, living in the Southeast region (p=0.026), ≤8 schooling years (p=0.001), ≤12 months interpregnancy interval (p=0.021), previous preterm birth (p=0.001) and previous low birth weight (p=0.001) were associated with preterm deliveries.
Gestational characteristics associated with preterm birth and the presence of anemia includes late onset of prenatal care (p=0.001), <6 prenatal care visits (p=0.007), being pre-pregnancy underweight or normal weight (p=0.001), use of drugs before or during pregnancy (p=0.037), urinary tract infection (p=0.001) and vaginal bleeding (p=0.001), as shown in Table 2     bleeding (p=0.001) and cesarean section (p=0.0026) were associated with preterm. Table 3 shows perinatal outcomes according to gestational age at birth and the presence of anemia. In the preterm group, the presence of anemia was significantly associated with the need of ventilatory support (p=0.003), sepsis (p=0.006) and endocrine dysfunction (hypoglycemia) (p=0.001). This last one was the only association found between term delivery and anemia (p=0.012). In the analysis of the anemia group, newborns classified as small for gestational age were more frequent in the preterm (25.41%) than term (1.74%) births. Furthermore, association of anemia and prematurity was found for: 5 th minute Apgar score <7 (p=0.001), fetal death (p=0.002), endotracheal intubation at birth (p=0.001), use of surfactant (p=0.001), ventilatory support (p=0.001), sepsis (p=0.001), respiratory distress (p=0.001),  In the multivariate analysis (Table 4), for the term group, housing area was the only significant association found.
For the preterm group, the factors associated to anemia include: urinary tract infection in sPTB and piPTB groups (p<0.001), neonatal morbidity (p=0.001), adequacy of prenatal care visits (p=0.009) and children under 5 years old in pPROM group.

Discussion
The overall prevalence of anemia in our study was 31.70%, with 27.74% in term deliveries and significantly higher, 33.23%, among preterm deliveries. Anemia is a public health problem worldwide, especially in low and middle-income countries [11][12][13] and it is associated with worse maternal and perinatal outcomes, including increased maternal and perinatal death, preterm birth, cesarean section, preeclampsia and low birth weight [14][15][16]. WHO's national estimates report anemia prevalence of 41.8% [4], higher than our results. Furthermore, our study presented higher prevalence of anemia then reported in a study in Lagos, Nigeria in 2020, that was 20.0% [17], higher than a multicenter Chinese study that found an overall prevalence of anemia of 23.5% [18], higher than a study conducted in Jerusalem with 10.5% of anemia [19] and higher than a large retrospective study in British Columbia that found anemia in 12.8% of its records [20]. The difference in findings throughout studies may represent the population analyzed as having anemia is more prevalent in low-income countries and in regions with poor antenatal care [21]. A Brazilian study with adolescents showed an overall prevalence of anemia of 41% [22]. This higher prevalence may be due to the additional risk factors related to pregnancy in adolescents as higher need of iron intake [23] and nutritional deficiencies [24]. Also, our study was conducted in referral obstetric hospitals in the three most populous Brazilian regions, so the prevalence of anemia could have been higher as the most challenging and severe cases are treated in these hospitals.
Our results also found association with maternal age and the presence of anemia in preterm deliveries, as it was more frequent in pregnant adolescents. This is in agreement with the Chinese study that found higher prevalence of anemia in women under 20 [17]. In addition, positive association for preterm birth and anemia was evidenced in non-white women, living in the Northeast region, schooling years less than 8, presence of children below 5 years old, onset of prenatal care at second or third trimester and less than six prenatal care visits. These findings are consistent with literature as anemia is associated with poor access to antenatal care and low socio-economic condition [5,17,18]. In our country, non-white skin color and the Northeast region are associated with lower income regions and therefore associated with anemia [25].
Regular antenatal care is an important intervention that may avoid many maternal adverse complications, by the identification of risk factors, early diagnosis and surveillance, including anemia and its consequences, as deviations from the usual can be early detected and treated. Late onset of prenatal care was a finding associated with anemia in both term and preterm deliveries. It may be related to the fact that iron supplementation wasn't long enough to recover at least the expected dilutional anemia and a chronic history of malnutrition and self-care. As a universal recommendation, prophylactic iron should be prescribed during gestation throughout postpartum [4].
Anemic women had high cesarean section rates in our study (especially among preterm deliveries) and such procedure can worsen hemoglobin level leading to other complications. Also, in the preterm group, fetal growth restriction was more frequent, and anemia could be partially responsible for the restriction. Previous preterm birth and previous low birth weight newborn were also more frequent in the preterm group. Moreover, anemia increases the risk for postpartum hemorrhage, worsening an already deficient hemoglobin level [10].
Considering factors independently associated to anemia, living in urban area was associated to anemia in term pregnancies. This could most likely be a consequence of the great majority of the population being urban, but also could be worsen by nutritional habits in urban centers where availability of food with low nutritional quality is higher. Differently, a study in China found that rural population is at greater risk for anemia [18], however studies correlate anemia with family income and schooling years, being poor education and low economic conditions associated to anemia [5,18,26]. Urinary tract infection was associated with anemia in sPTB and piPTB. It might have been the trigger of spontaneous preterm birth and could deteriorate maternal clinical condition, as it is an already known risk factor for prematurity, independently the gestational age of the infection [27]. In the sPTB, non-white skin color was also associated with prematurity, probably due to less access to health care. Having other child below 5 years old, any neonatal morbidity and inadequate numbers of antenatal care visits were associated with anemia and pPROM, suggesting that antenatal care is quite important to follow the women and identify alterations that can lead to complications. Studies suggest screening for infections, nutritional adequacy, iron supplementation and a periodic antenatal visit to check maternal health and early detection of possible pathologic conditions [20,28].
It is important to point out that this study has limitations as it was based on a secondary analysis of a major multicentric study. As a consequence, some variables were difficult to standardize and it was hard sometimes to define the cases based on past medical records. The main study (EMIP) was a multicenter cross-sectional study that was conducted in 20 referral obstetric hospitals and was the largest Brazilian study in prematurity so far [1]. One of its findings was that anemia is a risk factor for prematurity, what was an important point for further analysis.
Prematurity is enhancing worldwide and is a complex situation as the goals in order to reduce its rates are difficult to reach [29,30]. WHO estimates almost 15 million preterm deliveries every year (11.1% of all births) increasing costs in public and private healthcare [4]. In our study, prematurity was divided into three etiologies: premature rupture of membranes, spontaneous premature birth and provider-initiated preterm birth. The frequency of anemia in each group was 36.3%, 32.79% and 31.27% respectively. In all the three situations, anemia might be related to the cause of prematurity, however could also be worsen as a consequence of reduced time for iron supplementation. Overall, anemia was more prevalent among preterm birth, what is in agreement with the study in India [5] and the retrospective cohort studied in British Columbia that found an Odds Ratio of 2.44 for preterm birth and unspecified anemia severity [20].
Besides, prematurity costs millions to healthcare and is strongly associated with neonatal morbidity and mortality [20,31,32] with rates that have not decreased despite the knowledge and efforts to reduce its prevalence [30]. The improvement of maternal and neonatal intensive care is a situation that tends to persist and enables women with several morbidities to control the underlying condition and pursue pregnancy and neonates delivered in a very preterm gestational age to survive, sustaining high rates of preterm birth, especially medically indicated.
Our study highlighted the high prevalence of anemia and prematurity in our setting and could bring light to further studies considering the association of anemia and preterm birth and the importance of additional studies to evaluate the topic and improve maternal healthcare aiming the reduction in neonatal morbidity and mortality rates due to prematurity.