During pregnancy, maternal hemoglobin (Hb) concentration drops physiologically due to hemodilution. This effect reaches a maximum in the third trimester.1,2 Therefore, the World Health Organization set the cutoff for anemia in pregnancy to Hb <11 g/dL, instead of <12 g/dL in nonpregnant women.3 Still, anemia in pregnancy is common; in 2011, the Nutrition Impact Model Study estimated that 38% of pregnant women worldwide are anemic, with iron deficiency as the major cause (75%).4,5 Other causes of anemia are folate, vitamin B12 and vitamin A deficiencies, chronic inflammation, parasitic infections, and inherited disorders.5 Various studies reported on the consequences of anemia in pregnancy.5–15 It is thought that low maternal Hb concentration is a risk factor for adverse neonatal and maternal outcomes.5–15 A systematic review and meta-analysis reported a higher risk of preterm birth in case of maternal anemia in the first or second trimester,7 whereas a more recent meta-analysis showed also an increased risk of low birth weight.8

Apart from low Hb, also elevated Hb levels are associated with adverse perinatal outcome.4,15–17 As a result of poor plasma expansion, blood viscosity increases, leading to a reduction in blood flow and fetomaternal exchange of oxygen and nutrients in the placenta.15,16 High Hb concentrations are associated with pregnancy-induced hypertension and preeclampsia.15–17

Because low and elevated Hb levels seem to negatively influence pregnancy outcome, this may indicate a U-shaped optimum for Hb concentration in pregnancy.14,15,18 We hypothesize that in both anemic women and women with elevated Hb levels, there is a suboptimal oxygen supply to the placenta. This “placental hypoxemia” may lead to impaired fetal oxygenation, thus increasing the risk of fetal distress and possibly leading to impaired neonatal outcome. Furthermore, in case of anemia, maternal endurance during labor may be impaired, thus increasing the risk of assisted vaginal delivery or cesarean delivery (CD).

Until now, no systematic review has evaluated the relation among maternal Hb concentration, mode of delivery, and neonatal outcome. Individual studies reporting on the course of labor and short-term neonatal outcome in relation to maternal Hb show different results.10–12,19,20 Therefore, we aim to investigate the effect of maternal Hb in the second or third trimester of pregnancy on mode of delivery, Apgar score, umbilical cord pH, neonatal intensive care unit (NICU) admission, and perinatal mortality.

METHODS Data Sources

We systematically searched the electronic databases PubMed, EMBASE, and Cochrane for studies that reported on the relationship between maternal Hb and mode of delivery and/or neonatal outcome. The search terms included “h(a)emoglobin,” “h(a)ematocrit,” “mode of delivery,” “f(o)etal distress,” “pregnancy outcome,” “term birth,” and “childbirth.” This search was performed with the help of an information specialist. The study had to be available in the English or Dutch language. In addition, we manually reviewed the list of references of the identified articles and systematic reviews for additional eligible studies that were not identified in the initial search.

Inclusion and Exclusion Criteria

We included all studies that reported on both the maternal Hb concentration and at least 1 of the following outcome measures: mode of delivery, Apgar score, umbilical cord pH, NICU admission, fetal distress, or perinatal death. The study population had to consist of women with a singleton pregnancy, term birth, and with the intention of a spontaneous vaginal delivery. The maternal Hb concentration had to be measured during the second or third trimester of pregnancy, so it would reasonably reflect the peripartum Hb concentration.21 Exclusion criteria were planned CD, multiple pregnancies, or Hb concentration measured in the first trimester only.

Study Selection

Two independent investigators (J.S. and L.B.) screened all titles and abstracts of trials found in our search to determine if they met the inclusion criteria. Disagreements were discussed and consensus was reached. After eliminating noncompliant articles, the 2 investigators (J.S. and L.B.) analyzed the full text of the remaining studies to decide on eligibility for inclusion.

Data Extraction and Risk of Bias Assessment

Methodological quality was assessed from the following items: study type, number of subjects, risk of selection bias, including randomization and blinding (high or low), and description of inclusion and exclusion criteria (complete or incomplete). We used the GRADE instrument to provide an overall judgment of the study quality as described in the GRADE Handbook.22 Both reviewers evaluated the quality of eligible studies independently. Data were extracted from full texts, tables, and graphs. Data were entered into Microsoft Excel (Excel for Mac 2011; Microsoft Corporation, Redmond, WA), and the 2 reviewers double-checked accuracy.

Data Analysis

The systematic review was conducted using the PRISMA guidelines and checklist (2009).23 A meta-analysis could not be performed because the included articles show large heterogeneity in study population and study methods. Therefore, we described the results and displayed the evidence in relation to the quality of each study.

RESULTS Data Search

After removal of duplicates, 940 studies published before January 2021 were found (Fig. 1). The studies were screened for eligibility by title and abstract, and 14 articles were found eligible for full-text assessment. All references were screened, and we found another 44 articles that were screened by title and abstract, of which 23 articles were eligible for full-text assessment. In total, we performed a full-text assessment on 37 articles (Fig. 1). Of these 37 articles, 23 were excluded because of the following reasons: the article was not available in English or Dutch (n = 2); the articles were reviews or meta-analyses (n = 5); the maternal Hb concentrations were only measured in the first trimester or it was not stated in the article when the Hb concentrations were measured (n = 6); no relevant outcome measures were concerned (n = 9); and breech presentations were also included (n = 1). A total of 14 articles, including a total of 422,180 women, met the inclusion criteria. Among the included articles were 6 prospective cohort studies, 2 case-control studies, and 6 retrospective cohort studies (Table 1). The study characteristics and quality assessment are displayed in Table 2, and the outcomes are shown in Tables 3A–F.

FIG. 1:

Results from the literature search and the different steps in the selection process of eligible articles.

TABLE 1 - Number of Included Studies and Study Type Per Outcome Measure Outcome Measure Available Evidence Mode of delivery 1 prospective cohort studies, 5 retrospective cohort studies Apgar score 4 prospective cohort studies, 1 prospective case-control study, 4 retrospective cohort studies, 1 retrospective case-control study Fetal distress 1 retrospective cohort study NICU admission 2 retrospective cohort studies Perinatal mortality 1 prospective cohort study, 4 retrospective cohort studies Umbilical cord pH 1 retrospective cohort study
TABLE 2 - Characteristics and Quality of Included Studies Author Year of Publication Country of Origin Study Design Hb Cutoffs for Anemia Timing Hb Measurement Sample Size Description In/Exclusion Blinding and Randomization Risk on Selection Bias Mode of Delivery Apgar Score NICU Admission Perinatal Mortality Fetal Distress Umbilical Cord pHa Missing Results Reported Cointerventions Described Confounding Variables Described GRADE Van Bogaert24 2006 South Africa Retrospective cohort study <10.00 g/dL Second trimester 3214 Incomplete NA High Yes No No No No No No Iron and folic acid No Low Orlandini et al12 2016 Italy Retrospective cohort study <11.00 g/dL Between AD 35 + 0 and 36 + 6 wk 1131 Complete NA High Yes Yes No No No No No Multivitamin Yes Low Aimakhu and Olayemi25 2003 Nigeria Prospective cohort study <9.67 g/dL Every antenatal visit 633 Incomplete NA Moderate Yes Yes No Yes No No No Iron Yes Low Fareh et al26 2009 United Arab Emirates Retrospective case-control study <11.00 g/dL Around AD 30 wk 200 Complete NA High No Yes No No No No No No Yes Low Zhang et al27 2009 China Prospective cohort study <10.00 g/dL Every trimester 164,667 Incomplete NA Moderate No No No Yes No No No Iron and folic acid Yes Low Sekhavat et al28 2011 Iran Prospective cohort study <10.00 g/dL First stage of labor 18 Incomplete NA Moderate No Yes No No No No No No No Low Lone et al29 2004 Pakistan Prospective cohort study <11.00 g/dL In labor 629 Incomplete NA Moderate No Yes No Yes No No Yes No Yes Low Drukker et al19 2015 Israel Retrospective cohort study <11.00 g/dL In labor 7566 Complete NA High Yes Yes Yes No No No No No Yes Low Hwang et al20 2010 Republic of Korea Retrospective cohort study <10.00 g/dL Third trimester 356 Incomplete NA High Yes Yes Yes Yes No No No Iron Yes Low Little et al30 2004 United Kingdom Prospective cohort study <11.00 g/dL Second trimester 144,209 Incomplete NA High No No No Yes No No No No Yes Low Lee et al31 2006 Korea Prospective cohort study <10.8 g/dL Second trimester 248 Incomplete NA Moderate No Yes No No No No No No No Low Xiong et al32 2003 China Retrospective cohort study <10.00 g/dL AD 32 wk 15943 Incomplete NA High No No No Yes No No No No Yes Low Lelic et al33 2014 Bosnia and Herzegovina Prospective case-control study <10.50 g/dL Second trimester 100 Complete NA High No Yes No No No No No No No Low Bullens et al34 2019 The Netherlands Retrospective cohort study <11.0 g/dL <2 wk prior labor 9144 Complete NA Moderate Yes Yes No No Yes Yes No No Yes Low

GA, gestational age; NA, not applicable.

TABLE 3 - Outcome of Included Studies Author Study Design Association Comment 3a Mode of delivery  Van Bogaert24 Retrospective cohort study Negative The prevalence of booking anemia (±GA 24 wk) in primigravidas and multigravidas with a CD was significantly higher compared with spontaneous vaginal deliveries (P = 0.002).  Orlandini et al12 Retrospective cohort study Negative Anemic women showed a significantly higher rate in emergency CD (P = 0.006).  Aimakhu and  Olayemi25 Prospective cohort study Unclear 37.5% of moderate anemic patients delivered by CD compared with 22.2% of the nonanemic patients, there was no statistical test performed.  Drukker et al19 Retrospective cohort study Negative Rates of CD were significantly higher among anemic women if they were (grand) multiparas. Anemia was identified as a significant independent risk factor of CD (P < 0.001).  Hwang et al20 Retrospective cohort study Negative Anemic women showed a significantly higher rate of CD for fetal distress compared with nonanemic women (P < 0.001).  Bullens et al34 Retrospective cohort study Negative There was a unique significant contribution of Hb for the prediction of IVD for any reason (P = 0.00), IVD for fetal distress (P = 0.05), CD for any reason (P = 0.01), and CD for nonprogressive labor (P = 0.01). 3b Apgar score  Orlandini et al12 Retrospective cohort study No No differences were observed in terms of Apgar scores at 5 minutes between anemic women and nonanemic women.  Aimakhu and Olayemi25 Prospective cohort study Moderate Neonates of nonanemic mothers had a higher Apgar score at 1 minute (P < 0.05) compared with anemic mothers. There was no significant difference in Apgar score at 5 minutes between the anemic and nonanemic groups.  Fareh et al26 Retrospective case-control study No There was no significant difference in Apgar score between the case group and control group.  Sekhavat et al28 Prospective cohort study Negative The risk of low Apgar score was significantly increased in women with anemia. However, the given P value given is P = 0.8, which indicates no significant association. Therefore, we interpret no significant difference.  Lone et al29 Prospective cohort study Negative Multivariate analysis showed that the risk of an Apgar score <5 at 1 minute was 1.8 times higher for anemic women compared with nonanemic women.  Drukker et al19 Retrospective cohort study Negative Anemia was identified as a significant independent risk factor of Apgar score <7 at 5 minutes (P < 0.001).  Hwang et al20 Retrospective cohort study No There was no significant difference in Apgar scores found.  Lee et al31 Prospective cohort study Negative Anemic women had significantly lower Apgar scores at 1 and 5 minutes compared with nonanemic women (P < 0.05).  Lelic et al33 Prospective case-control study No There was no significant difference in Apgar score between the case group and control group.  Bullens et al34 Retrospective cohort No There was no relation between Hb level and low Apgar score. 3c Fetal distress  Bullens et al34 Retrospective cohort study No Hb levels did not have a significant unique contribution to the likelihood of fetal distress. 3d Neonatal intensive care unit admission  Drukker et al19 Retrospective cohort study Moderate Anemia was identified as a significant independent risk factor for NICU admission (P = 0.018). When anemia was distributed by severity (mild, moderate/severe), there was no significant difference found between study groups in NICU admission. Retrospective cohort study No There was no significant difference in NICU admission between groups. 3e Perinatal mortality  Aimakhu and  Olayemi25 Prospective cohort study Unclear 97.4% of the nonanemic patients had live births compared with 75% of the moderately anemic patients and 100% of mild anemic patients. There were no statistical tests performed.  Zhang et al27 Prospective cohort study Moderate No association among maternal anemia, intrapartum stillbirth, and neonatal mortality was detected. In fact, a trend toward a slightly increased risk of early neonatal death in relation to elevated Hb levels in the second and third trimester was found (hazards ratio, 1.1; 95% CI, 1.0–1.2).  Lone et al29 Prospective cohort study No The adjusted relative risk of perinatal mortality was 3.2 times higher among anemic women; this was not significant (95% CI, 0.7–14.6).  Little et al30 Prospective cohort study No A U-shaped pattern was found with lowest recorded Hb concentration (in most cases between GA 26 and 28 wk) for early neonatal mortality; however, when adjusted for prematurity, the relationship of early neonatal mortality with lowest Hb largely disappeared and was no longer significant.  Xiong et al32 Retrospective cohort study No Anemia in the third trimester was not associated with poor birth outcomes. When divided into different degrees of anemia, there were no significant differences found in the frequency of perinatal mortality. 3f Umbilical cord pH  Bullens et al34 Retrospective cohort study No Before correction for confounders (age, parity, and birth weight), there was a significant correlation between Hb levels and umbilical cord pH (≤7.05 of >7.05). After correction, Hb level did not contribute to the prediction of low umbilical cord pH ≤7.05.
Mode of Delivery

We identified 6 articles reporting on maternal Hb concentration and mode of delivery.12,19,20,24,25,34 Van Bogaert24 performed a retrospective cohort study in a rural hospital in South Africa. Of the included 3214 patients, 2707 patients had a spontaneous vaginal delivery and a total of 507 patients had a secondary CD. The prevalence of anemia in patients with a CD was significantly higher compared with spontaneous deliveries (odds ratio [OR], 0.55; 95% confidence interval [CI], 0.37–0.80; P = 0.002). Unfortunately, they did not describe the reason to perform a CD (eg, fetal distress or nonprogressive labor). Orlandini et al12 performed a retrospective cohort study and included 1131 women with uncomplicated pregnancies. There were 2 groups, group A (n = 156) with Hb concentrations <11 g/dL (mild anemia) and group B with Hb ≥11.1 g/dL. Anemic women showed a higher rate of emergency CD than nonanemic women (P = 0.006). Aimakhu and Olayemi25 performed a prospective cohort study in a University College Hospital in Nigeria. With a finger prick, the packed cell volume (PCV) was measured every antenatal visit until delivery. Hb can be estimated by dividing the PVC by 3,35 a PVC below 30% was considered anemic. There were more CD in the moderate anemic group (37.5%), compared with the nonanemic and mildly anemic group (respectively, 22.2% and 21.4%). There were no statistical tests performed because of the small sample size of moderately anemic women (n = 24). Drukker et al19 conducted a large retrospective cohort study containing 75,660 women; the Hb values were determined on the day of labor. Maternal anemia was significantly associated with higher rates of CD in multiparous women (2.6% vs 2.1% for multiparas [P = 0.039] and 3.2% vs 2.0% for grand multiparas [defined as >5 childbirths] [P < 0.001]). There was no significant difference among nulliparous women. In addition, they showed that an increase in Hb levels of 1.0 g/dL was associated with a reduction of 8.3% in CD rate (OR, 0.92; 95% CI, 0.88–0.95; P < 0.001). Two stepwise backward logistic regression models were performed to evaluate the independent effect of anemia on CD rate. Both models identified anemia as a significant independent risk of CD (OR, 1.30; 95% CI, 1.13–1.49; P < 0.001; and OR, 1.56; 95% CI, 1.23–1.97; P < 0.001). Hwang et al20 performed a retrospective cohort study as well, including 3560 women of whom 377 had anemia. In the anemic group, there were higher rates of CD for fetal distress compared with the nonanemic group. Also in the multivariate analysis, CD for fetal distress was independently associated with anemia (OR, 1.5; 95% CI, 1.2–1.7; P < 0.001). Bullens et al34 conducted a large retrospective cohort study in a tertiary hospital in the Netherlands. They included 9144 women where the Hb concentration was determined within 2 weeks before labor. The mean Hb concentration was 12.2 g/dL (±1.2, 7.7–16.4). Intrapartum Hb concentration was low in 12.7%, normal in 60.4%, and high in 26.9%. No women had severe anemia.

Hemoglobin levels were significantly different in the groups where women had an instrumental vaginal (IVD) for fetal distress (P < 0.001), IVD for nonprogressive labor (P < 0.001), or a CD for nonprogressive labor (P < 0.001), compared with the women where these specific assisted-delivery measures were not undertaken.

The absolute difference in mean Hb level was 0.1–0.5 g/dL, and the effect size was small (range Cohen d, 0.01–0.03). Logistic regression was performed to correct for the confounders age, parity, and birth weight. There was a unique significant contribution of Hb for the prediction of IVD for any reason (OR, 1.10; 95% CI, 1.04–1.17; P < 0.05), IVD for fetal distress (OR, 1.10; 95% CI, 1.00–1.21; P = 0.05), CD for any reason (OR, 0.91; 95% CI, 0.84–0.98; P = 0.01), and CD for nonprogressive labor (OR, 0.89; 95% CI, 0.81–0.97; P = 0.01).

Apgar Score

Ten articles reporting on Apgar score were included.12,19,20,25,26,28,29,31,33,34 Orlandini et al12 did not observe any differences in 5-minute Apgar score between the anemic and nonanemic group. The study by Aimakhu and Olayemi25 reported a significantly higher mean 1-minute Apgar scores in the nonanemic group compared with the moderate anemic group. The mean 1-minute Apgar score was 7.9 in the nonanemic group (n = 567), 7.8 in the mild anemia group (n = 42), and 6.4 in the moderate anemia group (n = 24) (P < 0.05). However, the mean 5-minute Apgar was not significantly different; the mean Apgar scores for nonanemic, mild, and moderate anemia were, respectively 9.5, 9.6, and 8.6. Fareh et al26 performed a retrospective case-control study. Records of 100 consecutive anemic mothers who received antenatal care and had a vaginal delivery in the hospital were reviewed. Within 1 week of the delivery, a nonanemic patient was enrolled for inclusion in the control group. There were no statistically significant differences in baseline characteristics between case and control groups. Apgar scores at 1 and 5 minutes after birth were not different between the 2 groups. Sekhavat et al28 performed a prospective cohort study in Iran. A total of 1842 patients fulfilled inclusion criteria of whom 328 patients had anemia and 598 patients had high Hb concentrations. The authors state that the risk of low Apgar score was significantly increased in women with anemia; however, in the results table, a P value of 0.8 is mentioned. It is also unclear whether the 1- or 5-minute Apgar score is considered in their study. Lone et al29 performed a prospective cohort study and included 629 women. The univariate analysis showed that the risk of a 1-minute Apgar score <5 and 5-minute Apgar score <7 was 2.1 (95% CI, 1.2–3.7) and 1.7 (95% CI, 1.0–3.1), respectively. The multivariate analysis showed that the risk of a low 1-minute Apgar score was 1.8 times higher for anemic women compared with nonanemic women (95% CI, 1.2–3.7). The study of Drukker et al19 observed a significantly higher incidence of a 5-minute Apgar score <7 (P < 0.001). In the multivariate logistic stepwise regression model, anemia was an independent risk factor for 5-minute Apgar score <7 (OR, 2.21; 95% CI, 1.84–2.64; P < 0.001). When a multivariate logistic stepwise regression was performed by the degree of anemia, it showed that women with moderate or severe anemia had significantly increased risks for a low 5-minute Apgar score, compared with women with mild or no anemia (OR, 2.98; 95% CI, 2.20–4.03; P < 0.001).

Hwang et al20 performed a retrospective cohort study. They did not find any significant differences among study groups for 5-minute Apgar score <7. Lee et al31 performed a prospective cohort study with 248 healthy pregnant women. The Hb concentration was measured at 24 to 28 weeks of gestation, and the women were divided into 3 groups: Hb <10.8 g/dL (anemia), Hb 10.8–11.9 g/dL (normal), and Hb ≥12.0 g/dL (high). Newborn infants from anemic mothers had significantly lower Apgar scores at 1 and 5 minutes than the normal and high Hb groups (P < 0.05). They also found a significantly positive correlation between maternal Hb concentration and Apgar scores at 1 minute (r = 0.231, P < 0.05) and at 5 minutes (r = 0.201, P < 0.05). Lelic et al33 performed a prospective case-control study with 2 groups, each consisting of 50 women with healthy term pregnancies. The control group contained women with neither signs of anemia nor any other pregnancy disorder that could affect pregnancy outcomes. The Apgar scores at 1 and 5 minutes in both groups were similar. Th