Introduction

Anaemia is a state where the delivery of oxygen to the tissues is impaired because of a quantitative or qualitative deficiency of haemoglobin or red blood cells [1]. According to the World Health Organization (WHO), anaemia in pregnancy occurs when, at sea level, the haemoglobin is < 11 g/dL or the haematocrit is < 33%, regardless of gestation [2,3].

The United States Centers for Disease Control and Prevention’s (CDC) definition of anaemia in pregnancy differs from the WHO definition only in the second trimester where the cut off haemoglobin is < 10.5 g/dL and haematocrit is < 32% [4].

Despite these definitions, anaemia in pregnancy is not quite straight forward, because of the physiologic changes that occur during pregnancy, which also involve the haematologic system [5]. Some authorities maintain that additional variables of altitude, cigarette smoking and ethnicity may alter the definition of anaemia in individuals [6,7]. Others are of the opinion that altitude should not modify the definition [8] and that there is insufficient information to alter the definition of anaemia based on ethnicity [3].

These subtleties of defining anaemia in pregnancy are important as they have implications for comparisons between studies and between populations. Stephens et al. reported that approximately 38% of pregnant women worldwide are anaemic [9]. The estimated prevalence of anaemia in pregnancy differs widely between continents, being highest in Africa (55.8%) and Asia (41.6%), and lowest in Europe (18.7%) and North America (6.1%) [10]. In general, as pregnancy progresses, the prevalence of anaemia increases [11].

Anaemia in pregnancy is a major public health and economic problem worldwide and contributes to both maternal and fetal morbidity and mortality; anaemia of pregnancy can also have profound short-term and far-reaching sequelae for the newborn [12–14]. Anaemia, even in early pregnancy has been associated with adverse pregnancy outcome [15]. Clinical manifestations include fetal growth restriction, preterm delivery, low birth weight [16], impaired lactation, poor maternal/infant behavioural interactions, post partum depression and increased fetal and neonatal mortality [12,13]. Economic losses occur because iron deficiency anaemia has been associated with decreased work capability of adults and reduced cognitive function of children that may persist into adulthood; impaired motor development a manifestation of anaemia also adds to economic loss [9,17].

The risk factors/determinants/causes of anaemia can be found at multiple interacting levels [18].The immediate causes of anaemia can be considered to be decreased red blood cell/haemoglobin production and increased loss of red blood cells/haemoglobin, as a result of nutritional, infectious and genetic influences. Some of the important risk factors include deficiency of nutrients such as iron (apparently the most common risk factor), folate and vitamin B₁₂, infections such as human immunodeficiency virus (HIV), malaria and hook worms, and disorders in the structure or production of haemoglobin such as sickle cell disease and the thalassemias [12,13]. The aetiology of anaemia in pregnancy can be considered in the broader context of Balarajan and colleagues’ “Conceptual model of the determinants of anaemia” (see Fig 1), whereby alterations in the political economy, climate, ecology, cultural norms, etc. eventually culminate in decreased red blood cell/haemoglobin production and increased loss. Other risk factors in the conceptual model include teenage pregnancy, ‘low’ educational level, ‘poor’ socioeconomic status, a short inter pregnancy interval and high parity [18–20]. In aggregate, nutritional, infectious and genetic risk factors for anaemia are less common in high income than in low and middle income countries [21].

Impaired delivery of oxygen to tissues appears to be the central mechanism by which anaemia increases the risk of maternal organ (brain, heart, kidney) injury and mortality [22]. Oxygen delivery to the uterus (and fetus) may be reduced in pregnant women with anaemia because of this impaired tissue oxygen delivery [23].

The primary aims of this study were to determine the modifiable risk factors for anaemia in early pregnancy, and to compare the pregnancy outcomes between women with and without anaemia in early pregnancy.

Materials and Methods

The SCOPE (SCreening fOr Pregnancy Endpoints) study is an international, prospective, multicentre cohort study of 5 690 ‘low-risk’ nulliparous women with singleton pregnancies in four high income countries: New Zealand, Australia, England and Ireland (www.scopestudy.net/). The primary aim of the SCOPE study is to develop screening tests to predict pre-eclampsia, fetal growth restriction and spontaneous preterm births in a low risk population.

Details of the SCOPE study methods have been published in detail elsewhere [24] but in brief, recruitment of participants into the study started in Auckland, New Zealand in November 2004 and finished in Cork, Ireland in February 2011. The participants ranged in age from 16 to 45 years and were recruited into the study before 15 weeks’ gestation, through community midwives, general practitioners, hospital antenatal clinics, obstetricians and self referral. Women’s partners, along with their newborn infants, are also involved in the study. Women were excluded if they: 1) were considered to be at high risk of pre-eclampsia, fetal growth restriction or spontaneous preterm birth due to underlying medical conditions (chronic hypertension, diabetes, renal disease, systemic lupus erythematosus, anti-phospholipid syndrome, sickle cell disease, HIV), previous cervical knife cone biopsy, ≥ 3 previous terminations or ≥ 3 miscarriages, current ruptured membranes; 2) had a major uterine anomaly, a known major fetal anomaly or abnormal karyotype; or 3) received an intervention that could modify pregnancy outcome (e.g. aspirin therapy, cervical suture).

At 14–16 weeks’ and 19–21 weeks’ gestation, participants were interviewed and examined by a research midwife, and underwent blood and urine tests. An ultrasound scan was performed at 19–21 weeks. Participants were followed prospectively, with pregnancy outcome data and baby measurements collected by research midwives.

Ethical approval was obtained from local ethics committees [New Zealand AKX/02/00/364, Australia REC 1712/5/2008, London and Manchester 06/MRE01/98 and Cork ECM5(10)05/02/08] and all women provided written informed consent. For the secondary analysis of the SCOPE study dataset, presented in this paper, ethics approval was sought from the Human Research Ethics Committee (Medical) of the University of the Witwatersrand M130966.

Prior to analysis, each participant’s measured booking haemoglobin (obtained typically in the first trimester) was adjusted to take into account cigarette smoking during three months pre-pregnancy up to and including the first study visit at 14–16 weeks (data on smoking was available for this time period). The adjustment was made according to the WHO criteria [3,6] (see Table 1). Because smoking has effects which can persist via epigenetic changes [25], and haematological indices can take years to return to normal after smoking cessation [26], an adjustment to the measured haemoglobin was made considering the entire period for which data on smoking was available. The adjusted haemoglobin was then used to classify participants as having anaemia or not according to the WHO definition of anaemia in pregnancy. No adjustment to the haemoglobin for altitude was made because all the participating SCOPE centres were below 1 000m above sea level.

We investigated the following risk factors for their association with anaemia: maternal age, country, marital status, ethnicity, schooling, paid work, body mass index, maternal socioeconomic index, fruit consumption, vegetable consumption, folate intake, iron and mineral intake, alcohol, psychological scales, paternal age and socioeconomic index. We investigated the following outcomes for their association with anaemia: small for gestational age, preterm delivery, mode of delivery, low birth weight and APGAR score. The variables were examined as presented in Table 2.

Results

Fig 2 depicts the inclusion of participants in the final analysis; 5 690 participants were recruited into the SCOPE study at 14–16 weeks (STROBE Statement: S1 Table). Forty-eight participants (0.8%) were lost to follow-up and 14 (0.2%) were ineligible after recruitment. Nineteen (0.3%) of the 5 628 remaining participants did not have a booking haemoglobin and were excluded, resulting in a final study population of 5 609 participants at 14–16 weeks. Without adjustment to the measured haemoglobin, 103 (1.8%) of the participants were anaemic; 125 (2.2%) were found to be anaemic after adjusting for cigarette smoking.

Table 2 shows the differences between participants with and without anaemia. The median haemoglobin for those without anaemia was 12.8 g/dL (IQR 12.3–13.4); for those with anaemia it was 10.7 g/dL (IQR 10.4–10.8). Factors that were significantly associated with having anaemia were ethnic origin, reporting folate intake before pregnancy and no iron or mineral intake in the first trimester.

In the final model, which was of good fit, the variables that were independently associated with anaemia in early pregnancy were country, ethnic origin and having a marital partner (Table 3). Assuming that the 19 participants without a booking haemoglobin were anaemic lead to having paid work and reporting folate intake prior to pregnancy being protective of anaemia in early pregnancy.

Birth outcomes were similar for anaemic and non-anaemic women (Table 4). The pregnancy outcome findings suggested a trend towards a higher risk of adverse pregnancy outcomes for some, but not all outcomes.

Although 76.0% of participants did not have data on serum ferritin, 12 (0.88%) participants with data on serum ferritin were found to have iron deficiency anaemia (defined as serum ferritin < 12μg/L and haemoglobin < 11g/dL [15]). The median gestational age at delivery was similar for the anaemic and non-anaemic women: 38.9 and 40.0 weeks, respectively (see Fig 3), logrank test: p > 0.05. Participants with moderate anaemia had a mean and median gestational age at delivery of 39.8 and 40.1 weeks respectively while the corresponding values for mildly anaemic participants were 39.3 and 39.9 weeks. There was no correlation between the adjusted haemoglobin and birth weight, correlation coefficient 0.0017, p-value = 0.8972.

Discussion

In this contemporary, large multicenter cohort of nulliparous women in their first ongoing pregnancy we found a very low prevalence of anaemia. The 2.2% prevalence of anaemia in the SCOPE cohort differs sharply with the 22% prevalence reported from high-income regions in recent literature [9]. One possible explanation for this is that SCOPE participants were selected to be ‘low risk’ and were all nulliparous. It is well recognised that co-morbidities such as high parity and short birth interval can affect anaemic status [18]. Furthermore, despite the multicenter nature of the cohort, the ethnicity of SCOPE participants was homogeneous with 89.9% of European ancestry. We did not adjust our findings for ethnic specific variations in haemoglobin concentration because of the low numbers of non-Caucasian participants and because there is sparse data on how to adjust the haemoglobin for ethnicity [7], this may also partially explain our findings. Mandatory folic acid supplementation does not explain the lower incidence of anaemia in this cohort because at the time of patient recruitment none of the participating countries had mandatory folic acid supplementation programs [29].

In the SCOPE cohort, not having a marital partner was associated with higher odds of having anaemia in early pregnancy. This is not surprising because there is evidence that involvement of fathers during pregnancy is associated with diminished negative maternal behaviours and better neonatal outcomes [30]. In addition, not having a partner suggests that the pregnancy was unintended and therefore women did not take steps to optimize their health prior to pregnancy. In general, marriage protects pregnancy [31].

In the adjusted analysis, United Kingdom participants had a higher odds ratio of anaemia in early pregnancy compared to the other countries. It is difficult to untangle the disparate potential contributions of political economy, ecology, geography and climate, all of which are found within the conceptual framework of anaemia’s determinants.

Previous studies mainly from low and middle income countries have shown an association between low education [18,20], the Edinburgh postnatal depression score (the depression being linked to folic acid deficiency [32,33]) and teenage pregnancy [19], but we did not find this in our study.

In our study, from a statistical significance viewpoint, anaemia was not associated with adverse pregnancy outcomes. However, adverse pregnancy outcomes tended to be more common in those with anaemia than in those without. Low birth weight and preterm delivery were similar between pregnant women with and without anaemia in early pregnancy. This is at odds with findings from a recent comprehensive systematic review and meta-analysis [15]. The low prevalence of anaemia in this study (with small numbers of relevant pregnancy outcomes for anaemic participants), due to the deliberate recruitment of ‘low risk’ women, could possibly explain the absence of an effect of anaemia on these adverse pregnancy outcomes.

Contrary to the finding that anaemia prevalence is consistently higher in those of lower socioeconomic status and in those with low body weight [34], in this study the prevalence of anaemia was similar across paternal and maternal socioeconomic groupings and body weight.

Although confirmation of iron deficiency in pregnancy is difficult [13], iron deficiency anaemia is reportedly the most common cause of anaemia in pregnancy. Relatively easy access to iron in fortified cereals and other food products (important sources of iron in industrialized countries) [35], irrespective of socioeconomic status, may partly explain this lack of association.

Although there were few participants with moderate anaemia, they paradoxically delivered later, on average, than participants with mild anaemia (of lesser severity). In fact, none of the participants with moderate anaemia had pre-term labour. This finding contrasts with the u-shaped relationship described in literature (where pre-term birth is more common at both very high and very low maternal haemoglobin concentrations and is uncommon at normal haemogloin levels) [8]. Our finding could be spurious given the small number of participants; however the finding is biologically plausible because paradoxical results have been found in transfusion studies where individuals with more severe anaemia do better [36,37]. We are by no means suggesting that women be made to have moderate anaemia as moderate anaemia seems to be associated with a longer pregnancy compared to mild anaemia, but the finding is worth noting because it can generate hypothesis about underlying biological mechanisms and reveal potential therapeutic targets.

The strengths of this study include its large multi-country prospective cohort design with excellent follow-up where outcome data were available for approximately 99% of the participants. Inclusion of parental infant trios and the availability of a large number of clinical variables further strengthened the study. Stringent real time data monitoring helped to ensure the quality of the data.

The primary study was designed to develop predictive biomarkers for three late pregnancy conditions, and not specifically to answer the question posed in the study reported in this paper.

‘Healthy’ nulliparous women with singleton pregnancies recruited into the SCOPE study are not representative of the general pregnant population. The primary study was conducted in high income countries, thus risk factors for anaemia in pregnancy such as malaria and hook worm infection (which are more common in low and middle income countries) are unlikely to be identified as significant causes of anaemia. In addition, women with HIV and sickle cell disease, which are known risk factors for anaemia, were excluded from the primary study.

Cigarette smoking was evaluated by self report as is usual in clinical practice, however underreporting of smoking is possible as cotinine levels—a sensitive marker of smoking tobacco—were not measured. Nevertheless, in pregnancy, self-reported tobacco use has been found to be a valid marker of tobacco exposure [38].

Conclusion

In this low risk healthy pregnant population we found a low anaemia rate. The absence of a marital partner was a non-modifiable factor, albeit one which may reflect a variety of confounding factors, that should be considered for addition to the conceptual framework of anaemia’s determinants. Although not statistically significant, clinically, a trend towards a higher risk of adverse pregnancy outcomes was observed in women that were anaemic in early pregnancy.

Supporting Information

Acknowledgments

This study was initially carried out as a research report by Gwinyai Masukume in partial fulfillment of the requirements of the MSc in the field of Epidemiology and Biostatistics at the University of the Witwatersrand School of Public Health. We are grateful to the participants who volunteered for the SCOPE study, to make pregnancy and child birth even safer. We thank the SCOPE Country Project Managers R. Taylor, University of Auckland, D. Healy, University of Adelaide, A. Briley, King’s College London, and N. Murphy and E. Snapes, University College Cork, and the database and statistical support provided by E. H. Y. Chan, University of Auckland. We acknowledge and thank the SCOPE Principal Investigators, Professors Robyn North, London, Lesley McCowan, Auckland, Gustaaf Dekker, Adelaide, Lucilla Poston, London and James Walker, Leeds.