Gestational stage-specific association of hemoglobin concentration with the risk of preterm birth and small for gestational age

Abstract

Aim: Accumulating evidence suggests that maternal hemoglobin concentration during pregnancy negatively influences perinatal outcomes, including preterm birth (PTB) and small for gestational age (SGA). This study aimed to determine gestational stage-specific reference ranges of hemoglobin in a recent Japanese population and clarify the impact of maternal peripheral hemoglobin concentration on pregnancy outcomes according to gestational stage.

Methods: Clinical data on hemoglobin concentration at four gestational stages were collected from the perinatal records of women who delivered at two perinatal centers between 2009 and 2018 (n=10,535). Statistical analyses were performed to evaluate the association of hemoglobin concentration with perinatal outcomes, focusing on gestational stage-specific characteristics.

Results: Fifty percentile values of hemoglobin among women in normal pregnancy at 8–12 weeks of gestation (WG), 18–26 WG, 28–32 WG, and 34–28 WG were 12.6, 11.4, 11.1, and 11.3 g/dl, respectively. The incidence of PTB was significantly higher among women with low hemoglobin concentrations than those with hemoglobin concentrations within the reference range. This association was more pronounced in the later stages of gestation. Women with higher hemoglobin concentrations in the third trimester had an increased risk of SGA.

Conclusions: The impact of hemoglobin concentration on the risk of PTB and SGA varied by gestational stage. The influence of gestational age on hemoglobin measurements should be considered for precise evaluation of the association of maternal hemoglobin concentration with pregnancy complications.

Introduction

During pregnancy, physiological changes occur in maternal blood to support healthy fetal growth. Plasma volume increases progressively to 150%, reaching a peak at approximately 34 weeks of gestation (WG).^1,2) However, because this plasma volume expansion surpasses the concurrent net gain in red blood cell volume, peripheral blood tests in pregnant women commonly show a reduced hemoglobin (Hb) concentration, blood cell count, and hematocrit value. This phenomenon is remarkable especially in the late second and early third trimesters.^2,3) This alteration in blood characteristics during pregnancy, referred to as physiological hemodilution, is presumed to be advantageous for efficient uteroplacental circulation to support the fetomaternal transfer of gases and nutrition.^2,3)

In addition to physiological hemodilution, iron deficiency anemia (IDA) is another factor that causes a reduction in Hb concentration during pregnancy. Iron demand increases in the mother due to increased red blood cell counts and continuous fetal- and placental-maternal transfer.^4,5) Therefore, an appropriate intervention with iron supplementation is necessary for IDA during pregnancy, especially in severe cases, although there is no established evidence that iron supplementation is effective for improving obstetrical outcomes.^6,7,8)

Hb concentration is commonly used as a representative parameter to assess anemia. The World Health Organization suggests an Hb concentration of <11 g/dl as a diagnostic criterion for anemia in pregnant women.⁹⁾ However, physiological hemodilution is a congruent adaptation to healthy pregnancy. In this context, it is clinically important to distinguish pathological anemia represented by IDA from sham Hb reduction that does not truly affect physiological pregnancy outcomes.^10,11) To this end, the threshold of Hb concentration for judging the need for medical intervention should be set according to hematological changes specific to each gestational stage.

The deterioration of hematologic status can lead to adverse pregnancy outcomes.^{11,12,13,14,15)} Previous studies have shown that abnormal maternal Hb concentrations, both low and high, are associated with obstetric complications, including preterm birth (PTB), stillbirth, low birth weight (LBW), infection, depression, and blood transfusion.^12,13,16) However, only a few studies have examined the influence of anemia on pregnancy outcomes with a focus on changes in Hb concentration according to gestational stage.^11,15)

The present study aimed to clarify the association of peripheral Hb concentrations at different stages of the gestation period with perinatal outcomes. By reviewing perinatal records from two perinatal centers, Hb concentrations routinely measured during perinatal visits were analyzed retrospectively. The incidence of adverse events in the following pregnancies, especially PTB and small for gestational age (SGA), was evaluated in relation to peripheral Hb concentrations at different gestational stages.

Materials and methods

This retrospective cohort study was conducted at two perinatal centers with tertiary care functions, the University of Tokyo Hospital and Juntendo University Hospital. The study was approved by the ethics committee of the two facilities (2019256NI). Informed consent was obtained from the hospitals’ websites in the form of opt-outs.

Routine protocol of prenatal management

Prenatal management was conducted following the protocol commonly implemented in the perinatal healthcare system in accordance with the recommendation of the Ministry of Health, Labour and Welfare in Japan. The intervals of prenatal visits were once every 4 weeks at <24 WG, once every 2 weeks from 24 to 36 WG, and once a week from 36 WG to delivery.

Collection of data on maternal Hb

Blood tests, including Hb measurements, were routinely performed during prenatal visits at four gestational stages. Venous blood sampling was performed four times until term gestation: the first time at 8–12 WG (group A, the first trimester), the second time at 18–26 WG (group B, the second trimester), the third time at 28–32 WG (group C, early stage of the third trimester), and the fourth time at 34–38 WG (group D, late stage of the third trimester). In cases where additional data existed on Hb concentration apart from routine blood test results for some reason, laboratory data obtained at the time closest to 10 WG for group A, 22 WG for group B, 30 WG for group C, and 36 WG for group D were used for analysis. The total number of deliveries managed at the University of Tokyo and Juntendo University Hospitals between January 2009 and December 2018 was 10,837. Pregnant women analyzed in this study included those who were managed from early pregnancy to delivery at the two facilities where the study was conducted, as well as those who were managed at clinics until approximately 34 WG due to low risk. Hb concentrations measured at the clinics were not registered in the electronic medical records used in the present study. Therefore, pregnant women with missing data from some of the four time points were also included. Data from the records of women with deliveries before 22 WG, multiple gestations, and hematological disorders other than anemia were excluded. After exclusion, data of 10,535 women were included in the analysis. During the study period, iron tablets for anemia treatment were administered at the discretion of each physician. In some cases, prenatal management was conducted at different hospitals during certain periods of pregnancy, and Hb data during those periods were lacking. Consequently, data on Hb concentration were available for 3,441 women in group A, 4,442 in group B, 2,353 in group C, and 6,784 in group D (Figure 1). Incidences of PTB, SGA, hypertensive disorders of pregnancy (HDP), gestational diabetes mellitus (GDM), and placenta previa were obtained from perinatal records. HDP was diagnosed according to clinical criteria advocated by the Japan Society for the Study of Hypertension in Pregnancy.¹⁷⁾ GDM was diagnosed using IADPSG criteria.¹⁸⁾ SGA was defined as a birth weight <10th percentile for gestational age using a sex-specific Japanese neonatal anthropometric chart in 2000.¹⁹⁾

Figure 1.

Selection of study subjects.

The process of selecting pregnancy cases for the cross-sectional and longitudinal analyses is shown in the flowchart. After the initial step of exclusion according to the exclusion criteria, 10,535 cases remained. In the cross-sectional analysis, all cases for which Hb data were available during the corresponding gestational weeks were included; cases in which delivery ended before the time of Hb measurement were excluded. In the longitudinal analysis, cases for which Hb data from all four stages of the gestation period were available were included.

Statistical analysis

Data were analyzed using EZR statistical software (Jichi Medical University, Saitama, Japan).²⁰⁾ P<0.05 was considered statistically significant in all analyses. In the comparison of Hb concentrations between the control group and groups with pregnancy complications, univariate analysis and Student’s t-test were performed to clarify the association of Hb concentration with the incidence of pregnancy complications. Logistic regression analysis was performed to assess the association of Hb concentration with the incidence of PTB and SGA by gestational stage. Odds ratios (ORs) for developing PTB and SGA were evaluated among subgroups by Hb concentration. Multiple regression analysis using a bidirectional step-wise selection method was conducted to identify variables associated with the incidence of PTB and SGA. Along with clinical parameters potentially associated with the incidence of PTB and SGA, low Hb concentration (<0.5 g/dl) and high Hb concentration (≥11.5 g/dl) were included as variables in the analysis.

Results

Data selection

During the study period, 10,837 women delivered at the study facilities. Of these, 10,535 women were included in the study after excluding 302 who met the exclusion criteria. From perinatal records, two sets of Hb data were prepared for cross-sectional and longitudinal analyses; data subjected to the cross-sectional analysis included all Hb data available at each gestational stage (n=3,441 for group A, 4,442 for group B, 2,353 for group C, and 6,784 for group D). For the longitudinal analysis, only data from women whose Hb data at all four gestational stages were available were used (n=1,044) (Figure 1).

Results of cross-sectional analysis

1) Clinical characteristics by group according to gestational stage

Clinical characteristics of women in each group at different gestational stages are summarized in Table 1. In this analysis, Hb data were derived from the same population of 10,535 women for each group (Figure 1). There were no significant differences in clinical characteristics (i.e., maternal age at birth, pre-pregnancy body mass index, parity, and gestational age at delivery) among the four groups (Table 1).

Table 1. Clinical characteristics of pregnant women included in four gestational stage groups

	8–12 WG (Group A, n=3,441)	18–26 WG (Group B, n=4,442)	28–32 WG (Group C, n=2,353)	34–38 WG (Group D, n=6,784)
Maternal demographic characteristics
Maternal age at birth (years)	36 [16, 50]	35 [14, 50]	35 [14, 50]	35 [14, 50]
Pre-pregnancy BMI (kg/m2)	20.3 [13.8, 45.5]	20.3 [13.8, 41.6]	20.3 [14.3, 45.5]	20.2 [13.1, 41.6]
Parity
Nulliparity	2,311 (67.2%)	2,934 (66.1%)	1,516 (64.4%)	4,525 (66.7%)
Multiparity	1,130 (32.8%)	1,508 (33.9%)	837 (35.6%)	2,257 (33.3%)
Delivery outcomes
Gestational age at delivery (weeks)	39 [24, 42]	39 [25, 42]	38 [28, 42]	39 [34, 42]
Gestational age at delivery (days)	275 [169, 294]	275 [179, 294]	272 [201, 294]	276 [238, 294]
Birthweight (g)	3,024 [518, 4,852]	3,005 [493, 4,852]	2,919 [646, 4,852]	3,038 [1,245, 4,852]

Values are presented as median [minimum value, maximum value] or sample number (% to total samples in the group).

BMI, body mass index; SGA, small for gestational age; HDP, hypertensive disorders of pregnancy; GDM, gestational diabetes mellitus.

2) Distribution of Hb concentrations in normal pregnancy

In the cross-sectional analysis, a subgroup of women with normal pregnancies without any pregnancy complications (i.e., PTB, placenta previa, SGA, HDP, and GDM) was isolated from each group at four gestational stages. Hb concentrations at each gestational stage were evaluated using data from this subgroup. The distribution of Hb concentrations is summarized in Table 2, and differences by pregnancy period are shown in Figure 2. Hb concentrations in peripheral blood decreased from the first trimester (8–12 WG) to the early stage of the third trimester (28–32 WG), and slightly increased in the later stage of the third trimester.

Table 2. Hb concentrations at different gestational stages in normal pregnancy

	Hb (g/dl)
	Minimum	2.5percentile	7percentile	50percentile	93percentile	97.5percentile	Maximum
Group A (8–12 WG) (n=2,726)	7.7	10.6	11.1	12.6	13.9	14.3	15.7
Group B (18–26 WG) (n=3,484)	5	9.6	10.1	11.4	12.7	13.1	14.4
Group C (28–32 WG) (n=1,605)	7.5	9.3	9.7	11.1	12.5	12.9	13.9
Group D (34–28 WG) (n=5,498)	6.4	9.3	9.9	11.3	12.7	13.2	14.8

Figure 2.

Maternal Hb concentrations in normal pregnancy by gestational stage.

In the cross-sectional analysis, maternal Hb concentrations in normal pregnancy cases were obtained for each of the four stages of gestation: 8–12 weeks of gestation (WG) (Group A), 18–26 WG (Group B), 8–32 WG (Group C), and 34–38 WG (Group D). The thick lines inside the boxes indicate median values, and the upper and lower ends of the boxes represent the range of the 75th and 25th percentiles. The whiskers indicate the smallest and largest values within the 1.5 interquartile range of the box. The circles indicate outliers.

3) Association of Hb concentration with pregnancy complications by gestational stage

Hb concentrations were compared between groups with and without pregnancy complications by gestational stage. Women with PTB had a lower Hb concentration than those with normal pregnancies at 34–38 WG (group D, 10.3 g/dl vs. 11.3 g/dl, P<0.001). Women with SGA had a higher Hb concentration than those with normal pregnancies at 18–26 WG (group B, 11.5 g/dl vs. 11.4 g/dl, P=0.006), 28–32 WG (group C, 11.5 g/dl vs. 11.1 g/dl, P<0.001), and 34–38 WG (group D, 11.5 g/dl vs. 11.3 g/dl, P<0.001). Women with HDP and GDM had a higher Hb concentration than those with normal pregnancies at 8–12 WG (group A), 18–26 WG (group B), and 28–32 WG (group C). Women with placenta previa had a significantly lower Hb concentration than those with normal pregnancies at 34–38 WG (group D) (Table 3).

Table 3. Hb levels at different stage of gestation among the groups with pregnancy complications

	Hb (g/dl)
	Normal	PTB	P	SGA	P	HDP	P	GDM	P	Previa	P
Group A (8–12 WG)	12.6 (0.9) n=2,726	12.5 (1.1) n=233	0.221	12.5 (1.1) n=262	0.551	13.0 (1.1) n=89	<0.001	12.8 (0.9) n=217	<0.001	12.4 (1.2) n=28	0.443
Group B (18–26 WG)	11.4 (0.9) n=3,484	11.3 (1.0) n=308	0.184	11.5 (1.0) n=364	0.006	11.7 (1.1) n=111	<0.001	11.6 (0.9) n=288	<0.001	11.2 (1.1) n=41	0.086
Group C (28–32 WG)	11.1 (0.9) n=1,605	11.1 (1.1) n=283	0.583	11.5 (1.0) n=254	<0.001	11.4 (1.2) n=80	0.003	11.4 (1.0) n=205	<0.001	11.0 (1.0) n=69	0.349
Group D (34–38 WG)	11.3 (1.0) n=5,498	10.3 (1.5) n=301	<0.001	11.5 (1.1) n=565	<0.001	11.2 (1.5) n=153	0.499	11.4 (1.1) n=366	0.152	10.0 (1.0) n=78	<0.001

Hb concentration data was divided into four groups based on the gestational age at blood sampling: 8–12 weeks (Group A), 18–26 weeks (Group B), 28–32 weeks (Group C), and 34–38 weeks (Group D). The means of Hb concentrations (the upper line) and the sample numbers (the lower line) are presented in each of the groups with pregnancy complications. Hb values are expressed as mean (SEM).

4) Risks for PTB and SGA by Hb concentration

Based on the observations above, the association of Hb concentration with the risk for disease development was further investigated, focusing on PTB and SGA. The population subjected to the cross-sectional analysis was divided into subgroups according to Hb concentration. In group A (8–12 WG), subjects were divided into six subgroups (Hb <9.5 g/dl, 9.5≤ Hb <10.5 g/dl, 10.5≤ Hb <11.5 g/dl, 11.5≤ Hb <12.5 g/dl, 12.5≤ Hb <13.5 g/dl, and Hb ≥13.5), and ORs were determined with the 11.5≤ Hb <12.5 g/dl subgroup as the reference (Tables 4 and 5). In groups B, C, and D, ORs were determined for five subgroups (Hb <9.5 g/dl, 9.5≤ Hb <10.5 g/dl, 10.5≤ Hb <11.5 g/dl, 11.5≤ Hb <12.5 g/dl, and Hb ≥12.5) with the 10.5≤ Hb <11.5 g/dl subgroup as the reference, as the distribution of Hb concentrations shifted to a lower range in the second to third trimesters (corresponding to groups B, C, and D) (Tables 4 and 5).

Table 4. Risk of developing preterm birth according to Hb levels at different stage of gestation

Hb (g/dl)	Group A (8–12 WG)		Hb (g/dl)	Group B (18–26WG)		Group C (28–32WG)		Group D (34–38WG)
	PTB/total cases n=233/3,441	PTB/cases without other complications n=160/2,886		PTB/total cases n=308/4,442	PTB/cases without other complications n=205/3,689	PTB/total cases n=283/2,353	PTB/cases without other complications n=178/1,783	PTB/total cases n=301/6,784	PTB/cases without other complications n=193/5,691
	OR (95% CI)			OR (95% CI)		OR (95% CI)		OR (95% CI)
Hb<9.5	4.51 (1.19–17.0) n=3/12	4.56 (0.92–22.5) n=2/9	Hb<9.5	3.59 (2.01–6.4) n=16/92	3.85 (1.93–7.67) n=11/72	2.01 (1.14–3.55) n=17/90	2.25 (1.17–4.33) n=13/74	8.08 (5.69–11.5) n=64/285	9.19 (5.92–14.3) n=40/216
9.5≤Hb<10.5	2.15 (0.94–4.94) n=7/51	1.26 (0.38–4.21) n=3/41	9.5≤Hb<10.5	1.51 (1.03–2.22) n=40/491	1.72 (1.11–2.69) n=31/415	1.44 (1.03–2.01) n=65/455	1.43 (0.95–2.15) n=44/370	2.39 (1.76–3.24) n=88/1,115	2.71 (1.85–3.95) n=60/957
10.5≤Hb<11.5	1.04 (0.64–1.7) n=23/321	0.86 (0.47–1.58) n=14/273	10.5≤Hb<11.5	1.0 n=93/1,678	1.0 n=64/1,430	1.0 n=96/924	1.0 n=62/717	1.0 n=87/2,514	1.0 n=52/2,155
11.5≤Hb<12.5	1.0 n=73/1,060	1.0 n=53/898	11.5≤Hb<12.5	1.43 (1.08–1.88) n=128/1,655	1.44 (1.03–2.01) n=87/1,376	1.18 (0.86–1.62) n=81/673	1.22 (0.83–1.80) n=51/493	0.59 (0.41–0.86) n=42/2,029	0.68 (0.43–1.08) n=28/1,692
12.5≤Hb<13.5	0.90 (0.65–1.25) n=86/1,373	0.91 (0.62–1.32) n=63/1,168	12.5≤Hb	1.07 (0.70–1.62) n=31/526	0.67 (0.36–1.25) n=12/396	1.11 (0.69–1.78) n=24/211	0.70 (0.33–1.5) n=8/129	0.68 (0.42–1.11) n=20/841	0.80 (0.43–1.48) n=13/671
13.5≤Hb	0.95 (0.64–1.41) n=41/624	0.84 (0.52–1.38) n=25/497

The left row in each gestational stage group is odds ratio determined using total cases of cross-sectional analyses. The right row in each group is odds ratio calculated by excluding the data of the cases with complications other than PTB (SGA, HDP, GDM and placenta previa). P-values were determined in comparison with reference groups (1.5≤Hb<12.5 in Group A, and 1.5≤Hb<12.5 in the other groups) The odds ratios with statistical significance (P<0.05) are shown in bold font.

Table 5. Relation between SGA and maternal hemoglobin levels according to the gestational period

Hb (g/dl)	Group A (8–12 WG)		Hb (g/dl)	Group B (18–26WG)		Group C (28–32WG)		Group D (34–38WG)
	SGA/total cases n=262/3,441	SGA/cases without other complications n=208/2,934		SGA/total cases n=364/4,442	SGA/cases without other complications n=282/3,766	SGA/total cases n=254/2,353	SGA/cases without other complications n=190/1,795	SGA/total cases n=565/6,784	SGA/cases without other complications n=480/5,978
	OR (95% CI)			OR (95% CI)		OR (95% CI)		OR (95% CI)
Hb<9.5	2.41 (0.52–11.2) n=2/12	3.76 (0.77–18.5) n=2/9	Hb<9.5	1.27 (0.60–2.69) n=8/92	1.18 (0.46–3.00) n=5/66	0.54 (0.21–1.36) n=5/90	0.60 (0.21–1.71) n=4/65	0.81 (0.49–1.36) n=17/285	0.83 (0.44–1.55) n=11/187
9.5≤Hb<10.5	1.61 (0.67–3.88) n=6/51	1.04 (0.31–3.46) n=3/41	9.5≤Hb<10.5	1.28 (0.89–1.85) n=43/491	1.35 (0.91–2.02) n=36/420	0.72 (0.47–1.08) n=33/455	0.68 (0.42–1.10) n=24/350	0.74 (0.55–1.00) n=61/1,115	0.66 (0.47–0.93) n=45/942
10.5≤Hb<11.5	1.24 (0.80–1.93) n=30/321	1.27 (0.79–2.06) n=25/284	10.5≤Hb<11.5	1.0 n=117/1,678	1.0 n=95/1,461	1.0 n=91/924	1.0 n=71/726	1.0 n=182/2,514	1.0 n=159/2,262
11.5≤Hb<12.5	1.0 n=81/1,060	1.0 n=64/909	11.5≤Hb<12.5	1.18 (0.92–1.53) n=135/1,655	1.15 (0.86–1.53) n=103/1,392	1.34 (0.98–1.83) n=86/673	1.35 (0.95–1.93) n=65/507	1.39 (1.13–1.72) n=199/2,029	1.38 (1.10–1.73) n=174/1,838
12.5≤Hb<13.5	0.85 (0.62–1.16) n=90/1,373	0.91 (0.64–1.28) n=76/1,181	12.5≤Hb	1.75 (1.26–2.42) n=61/526	1.61 (1.10–2.35) n=43/427	2.07 (1.37–3.12) n=39/211	1.98 (1.21–3.22) n=26/147	1.85 (1.43–2.38) n=106/841	1.83 (1.39–2.40) n=91/749
13.5≤Hb	1.12 (0.78–1.61) n=53/624	1.06 (0.70–1.61) n=38/510

The left row in each gestational stage group is odds ratio determined using total cases of cross-sectional analyses. The right row in each group is odds ratio calculated by excluding the data of the cases with complications other than SGA (PTB, HDP, GDM and placenta previa). P-values were determined in comparison with reference groups (1.5≤Hb<12.5 in Group A, and 1.5≤Hb<12.5 in the other groups) The odds ratios with statistical significance (P<0.05) are shown in bold font.

A lower Hb concentration was associated with a higher OR for PTB in all groups. This association was unchanged even when women with other complications (HDP, GDM, SGA, and placenta previa) were excluded from the analysis (Table 4).

A higher Hb concentration was associated with a higher incidence of SGA at 28–32 WG and 34–38 WG. On the other hand, the association between Hb concentration and SGA at 8–12 WG and 18–26 WG was “U-shaped”; higher ORs were obtained for the lower and higher Hb subgroups, although a significantly high OR was obtained for the Hb ≤12.5 subgroup at 18–26 WG. This trend was consistent with the results of the analysis in which women with other complications (HDP, GDM, PTB, and placenta previa) were excluded (Table 5).

5) Multiple regression analysis on the association of Hb concentration with PTB and SGA incidence

We further investigated the association of Hb concentration with the incidence of PTB and SGA by multiple regression analysis. Based on the findings from the subgroup analyses (Tables 4 and 5), Hb concentrations of “<10.5 g/dl” and “≤11.5 g/dl” were included as variables representing low and high Hb concentrations, respectively. In addition to variables related to Hb concentration, clinical factors (maternal age, pre-pregnancy body mass index, parity, and pregnancy complications with GDM, HDP, and placenta previa), which potentially affect the development of PTB and SGA, were included as variables (Tables 6 and 7).

Table 6. Multiple regression analysis; Association of low Hb concentrations at different stages of gestation with preterm birth

	Group A (8–12 WG)	Group B (18–26WG)	Group C (28–32WG)	Group D (34–38WG)
	PTB/total cases n=233/3,441	PTB/total cases n=308/4,442	PTB/total cases n=283/2,353	PTB/total cases n=301/6,784
	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
Maternal age	0.98 (0.95–1.02)	0.97 (0.94–1.00)	0.98 (0.95–1.01)	1.01 (0.98–1.03)
BMI	0.97 (0.93–1.02)	0.97 (0.93–1.01)	0.96 (0.92–1.00)	0.95 (0.91–0.99)
Parity	0.77 (0.56–1.05)	0.80 (0.60–1.05)	0.85 (0.64–1.15)	0.84 (0.64–1.10)
GDM	0.77 (0.39–1.51)	0.78 (0.43–1.41)	0.60 (0.33–1.09)	1.03 (0.59–1.79)
HDP	4.07 (2.25–7.36)	3.6 (2.09–6.20)	3.76 (2.16–6.55)	5.19 (3.21–8.40)
Placenta previa	9.59 (4.04–22.8)	12.3 (6.09–24.8)	6.77 (3.91–11.7)	8.25 (4.82–14.1)
Hb<10.5	2.36 (1.06–5.23)	1.36 (0.95–1.96)	1.37 (0.99–1.88)	4.01 (3.08–5.22)

The odds ratios with statistical significance (P<0.05) are shown in bold font. BMI, body mass index; HDP, hypertensive disorders of pregnancy; GDM, gestational diabetes mellitus.

Table 7. Multiple regression analysis; Association of low and high Hb concentrations at different stages of gestation with SGA

	Group A (8–12 WG)	Group B (18–26WG)	Group C (28–32WG)	Group D (34–38WG)
	SGA/total cases n=262/3,441	SGA/total cases n=364/4,442	SGA/total cases n=254/2,353	SGA/total cases n=565/6,784
	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
Maternal age	1.01 (0.98–1.05)	1.00 (0.98–1.03)	1.01 (0.98–1.04)	1.01 (0.99–1.03)
BMI	1.09 (1.03–1.15)	1.09 (1.04–1.14)	1.07 (1.02–1.12)	1.07 (1.03–1.11)
Parity	0.99 (0.72–1.35)	0.85 (0.66–1.10)	0.93 (0.69–1.26)	0.90 (0.73–1.10)
GDM	0.96 (0.51–1.82)	0.70 (0.39–1.25)	0.33 (0.15–0.72)	0.80 (0.51–1.27)
HDP	2.86 (1.44–5.68)	2.81 (1.58–4.99)	1.96 (1.01–3.80)	1.80 (1.06–3.07)
Placenta previa	1.22 (0.28–5.31)	1.19 (0.36–3.96)	0.80 (0.31–2.04)	1.03 (0.41–2.60)
Hb<10.5	1.63 (0.68–3.92)	NA	NA	NA
11.5≤Hb	NA	1.30 (1.02–1.65)	1.86 (1.40–2.48)	1.62 (1.33–1.97)

The odds ratios with statistical significance (P<0.05) are shown in bold font. BMI, body mass index; HDP, hypertensive disorders of pregnancy; GDM, gestational diabetes mellitus; SGA, small for gestational age; NA, not analyzed.

A low Hb concentration (Hb <10.5) was identified as a significant variable associated with PTB at 8–12 WG and 34–38 WG (OR 2.36; 95% CI 1.06–5.23 at 8–12 WG and OR 4.01; 95% CI 3.08–5.22 at 34–36 WG). HDP and placenta previa were significantly associated with an increased incidence of PTB regardless of gestational stage. No remarkable associations were observed between PTB and other clinical parameters at any gestational stage (Table 6).

A high Hb concentration (Hb ≥11.5) was a significant variable associated with an increased incidence of SGA at 18–26 WG (OR 1.30; 95% CI 1.02–1.65), 28–32 WG (OR 1.86; 95% CI 1.40–2.48), and 34–38 WG (OR 1.62; 95% CI 1.33–1.97). A low Hb concentration (Hb <10.5) was not a significant variable associated with SGA at 8–12 WG. Maternal BMI and HDP were identified as high-risk factors for SGA regardless of gestational stage (Table 7).

Longitudinal analysis

The association of Hb concentration with the incidence of PTB and SGA was further investigated using a longitudinal approach. As mentioned above, this analysis was performed only in women (n=1,044) for whom Hb data from all four gestational stages were available, with no missing data. In other words, those included in the longitudinal analysis delivered their babies after 34 WG. Changes in Hb concentration during gestation were compared between the control, PTB, and SGA groups. Consistent with the results of the cross-sectional analysis, the PTB group had a lower Hb concentration than the control group. Moreover, the gap widened with the progression of pregnancy, showing significant differences for groups B, C, and D (i.e., periods corresponding to the second trimester and later) (Table 8 and Figure 3). The trend in differences between SGA and control groups was similar to that observed in the cross-sectional analysis. Hb concentrations were significantly higher at 28–32 WG (group C) and 34–38 WG (group D). Hb concentrations at 8–12 WG and 18–26 WG were comparable between SGA and non-SGA groups (Table 8 and Figure 4).

Table 8. Hb levels at different stages of gestation in the groups with or without PTB and SGA

	Hb (g/dl) n=1,044
	PTB group n=106	Non-PTB group n=938	P	SGA group n=90	Non-SGA group n=954	P
Group A (8–12 WG)	12.4 (1.1)	12.6 (1.0)	0.23	12.5 (1.1)	12.6 (1.0)	0.80
Group B (18–26 WG)	11.2 (1.1)	11.4 (1.0)	0.03	11.5 (1.0)	11.4 (1.0)	0.4
Group C (28–32 WG)	11.0 (1.0)	11.2 (1.0)	0.03	11.5 (1.0)	11.1 (1.0)	<0.001
Group D (34–38 WG)	10.1 (1.4)	11.2 (1.1)	<0.001	11.4 (1.1)	11.0 (1.2)	<0.001

In the longitudinal analysis, the mean (SD) of the Hb concentrations in the groups with and without PTB and SGA determined are presented on the Table. P-values were determined in comparison between with and without the disease. The P-values with statistical significance (P<0.05) are shown in bold font.

Figure 3.

Maternal Hb concentrations in groups with and without preterm birth by gestational stage.

In the longitudinal analysis, Hb concentrations were compared between groups with and without preterm birth (PTB). The group with PTB (n=106) is presented as blank circles connected by solid lines, and the group without PTB (n=938) as blank triangles connected by dotted lines. The error bars represent the standard deviation at each time point.

Figure 4.

Maternal Hb concentrations in groups with and without small for gestational age by gestational stage.

In the longitudinal analysis, Hb concentrations were compared between groups with and without small for gestational age (SGA). The group with SGA (n=90) is presented as blank circles connected by solid lines, and the group without SGA (n=954) as blank triangles connected by dotted lines. The error bars represent the standard deviation at each time point.

Discussion

This study demonstrated that Hb concentration in pregnant women is associated with the risk of developing PTB and SGA. PTB occurred more frequently in women with low Hb concentrations, and the impact of a low Hb concentration on the development of PTB was more pronounced in the later stages of pregnancy, especially the third trimester. The association of Hb concentration with the incidence of SGA differed from that with the incidence of PTB; the risk of developing SGA was higher among women with high Hb concentrations in the third trimester, although the association between a low Hb concentration and SGA was not remarkable.

In the present study, we determined the reference range of Hb concentration specific to each of the four stages of the gestation period in uncomplicated pregnancy. Hb concentrations decreased progressively from the first trimester to the early third trimester among women with normal pregnancies without major pregnancy complications and slightly increased in the late third trimester. These changes in Hb concentration are consistent with physiological hemodilution caused by the expansion of plasma volume,^1,2,3) which is a well-known phenomenon characteristic of healthy pregnancy and is part of the hematological adaptation favorable for efficient uteroplacental circulation. Disturbance of this process is linked to pathological pregnancy, such as HDP and fetal growth restriction.²⁾

The incidence of PTB was higher in women with low Hb concentrations, and this trend was especially evident in women with very low Hb concentrations (Hb <9.5 g/dl). These results are consistent with previous reports on the effect of maternal anemia on perinatal outcomes. Maternal anemia has been identified as a risk factor for PTB; however, since the timing of Hb measurement during gestation was not consistent in previous studies, the gestational stage-specific effect of anemia has not been fully clarified. In the present study, the cross-sectional analysis revealed the association of low Hb concentration with PTB consistently across all groups (i.e., at four different gestational stages). In addition, the multiple regression analysis revealed that a low Hb concentration was a risk factor associated with PTB, especially in the first and the late third trimesters, suggesting that a low Hb concentration is a robust indicator of PTB risk. More importantly, ORs in the low Hb subgroups were markedly higher at 34–38 WG, suggesting that a low Hb concentration is more strongly associated with PTB occurring at 34 WG or later (generally referred to as ‘late PTB’) than with PTB occurring at earlier stages of gestation (Table 4). Furthermore, the longitudinal analysis revealed that the difference in Hb concentration between the normal pregnancy and PTB groups increased during gestation (Figure 3). Since only women for whom Hb data at 34–38 WG were available were selected, those who delivered earlier than 34 WG were not included in the longitudinal analysis. Therefore, it is accurate to interpret that a low Hb concentration in the preceding gestation period is associated with late PTB.

Contrary to our findings, a recent meta-analysis reported that maternal anemia in the first trimester, but not in the second or third trimester, increased the risk of premature birth.¹⁴⁾ This discrepancy might be attributed to differences in race, public health, and the medical environment among the study populations. Although several factors leading to PTB have been identified, including multiple pregnancies, infections, genetic factors, and chronic conditions such as diabetes, most PTBs occur without a clear causative factor.²¹⁾ The underlying mechanism causing PTB in patients with low Hb concentrations, as observed in this study, needs to be addressed in future studies.

Presumably, maternal Hb concentration directly affects oxygen supply to the fetus and influences fetal development. Supporting this, previous meta-analyses demonstrated an association of maternal anemia with an elevated risk for LBW.^2,16) From a statistical standpoint, LBW is not the best outcome to examine the effects of maternal Hb concentration on fetal growth, as pregnancies with LBW could overlap with those with PTB, which also increases with maternal anemia. In this context, we used SGA as a parameter to explore the association between maternal Hb concentration and fetal growth. Consequently, the association between Hb concertation and SGA varied by gestational stage. In the analysis of Hb concentrations in the first and second trimesters, a U-shaped trend showing an increased risk of SGA was observed for both subgroups with high and low Hb concentrations, although the ORs for these groups were not statistically significant (Table 5). In contrast, subgroups with higher Hb concentrations (11.5≤ Hb <12.5 g/dl and 12.5≤ Hb) showed an elevated incidence of SGA. The increasing influence of low Hb concentration on the risk of SGA with the progression of gestation was also observed in the longitudinal analysis. Consistent with our findings, some studies reported that maternal Hb concentration had a “U-shaped” influence, rather than a linear influence, on the risk of adverse events during pregnancy.^11,13) Excessively high and low maternal Hb concentrations are linked with poor outcomes in some perinatal complications. Those studies also reported that low maternal Hb concentration in the first trimester and high maternal Hb concentration in the third trimester are closely related to adverse perinatal outcomes. Indeed, in the present study, SGA was a representative adverse event that required a distinctive interpretation of Hb concentrations according to gestational stage.

Congenital disease in the fetus, multiple gestations, and placental dysfunction are major causes of SGA. Since congenital disease in the fetus and multiple gestations were excluded from the present analyses, placental dysfunction is presumed to be the dominant cause of SGA observed among pregnant women in this study. However, the mechanisms leading to SGA in mothers with low and high Hb concentrations were not clear from the present observations. As a possible explanation, low Hb concentration in the first trimester may be associated with SGA because suboptimal maternal nutritional status could trigger poor placental development. Indeed, a recent study provided evidence for positive associations between the mother’s nutritional score and biomarkers for placental function and placental volume in the first trimester.²²⁾ As for the association of high Hb concentration in the third trimester with SGA, deterioration of placental function might lead to insufficient plasma volume expansion.²⁾ Placental dysfunction, a severe subtype of HDP, is the central pathology of preeclampsia.²³⁾ Fetal growth restriction is a clinical symptom that meets the criteria for preeclampsia diagnosis for categorizing the HDP phenotype.¹⁷⁾ In this context, both the analyses of the SGA group and HDP group included women with preeclampsia and fetal growth restriction (Table 3). Furthermore, the association of maternal Hb concentration with the incidence of SGA was also examined in women with SGA without HDP, GDM, PTB, and placenta previa. However, the risk of SGA was significantly higher in the low Hb subgroup in the first trimester and the high Hb subgroup in the third trimester. In addition, women with HDP had higher Hb concentrations from the first trimester to the early stage of the third trimester, and this pattern differed from that observed in those with SGA (Table 5). The multiple regression analysis confirmed that a high Hb concentration is a significant and independent variable associated with an increased risk of SGA in the second and third trimesters (Table 7). Taken together, the pattern of changes in maternal Hb concentration observed in pregnancies with SGA may reflect a phenomenon related to placental dysfunction leading to insufficient fetal growth; this phenomenon can occur independently from the development of preeclampsia.

Iron deficiency is a major cause of anemia in pregnancy. While iron supplementation is the primary approach in the treatment of maternal anemia during pregnancy, evidence is not fully established regarding its efficacy for improving perinatal outcomes. A previous meta-analysis assessing the effects of daily oral iron supplements for pregnant women⁷⁾ concluded that, although maternal anemia was reduced, the positive effect of daily oral iron supplements on perinatal outcomes was unclear. Therefore, future efforts are necessary to identify and reduce the risk of adverse outcomes in pregnant women requiring iron supplementation. In the present study, Hb concentration was associated with an increased risk of pregnancy complications, and this association was influenced by the gestational stage at which the measurement of Hb concentration was performed. In particular, physiological hemodilution in the second and third trimesters makes it difficult to judge the significance of the reduction in Hb concentration. Given that the present and previous studies^11,12,14,16) consistently demonstrated an association between severe anemia in early pregnancy and an increased risk of PTB and SGA, checking Hb concentrations and initiating iron supplementation in the first trimester, when hemodilution is less remarkable, may be a promising approach.

This study has several limitations. First, we examined the risk of adverse pregnancy outcomes by focusing only on Hb concentration, whereas maternal anemia should ideally be assessed by multiple parameters including MCV, serum ferritin, and serum iron, in addition to Hb concentration. However, Hb was the only item for which data were available in most women receiving prenatal care at the study facilities. Second, the data analyzed in this study were obtained from two facilities located in the central area of Tokyo, Japan. Therefore, the observations in this study were based on a relatively homogeneous population with relatively small variations in racial, social, and economic backgrounds.

In conclusion, the present study revealed an association of maternal Hb concentration with the risk of pregnancy complications, especially PTB and SGA, among Japanese women receiving prenatal care in the metropolitan area of Tokyo. A low Hb concentration was consistently associated with the risk of PTB throughout the gestation period. The risk of SGA was associated with a high Hb concentration in the third trimester. These findings can serve as a basis for setting criteria for Hb concentrations that require therapeutic intervention for anemia in social and medical settings in Japan.

Acknowledgments

The study was conducted with financial support from Nippon Shinyaku Co., Ltd.

Conflict of interest

The authors have no conflicts of interest to disclose.

References

• 1.   Lund  CJ,  Donovan  JC. Blood volume during pregnancy. Significance of plasma and red cell volumes. Am J Obstet Gynecol. 1967; 98: 394–403.
• 2.   de Haas  S,  Ghossein-Doha  C,  van Kuijk  SM,  van Drongelen  J,  Spaanderman  ME. Physiological adaptation of maternal plasma volume during pregnancy: a systematic review and meta-analysis. Ultrasound Obstet Gynecol. 2017; 49: 177–187.
• 3.   Aguree  S,  Gernand  AD. Plasma volume expansion across healthy pregnancy: a systematic review and meta-analysis of longitudinal studies. BMC Pregnancy Childbirth. 2019; 19: 508.
• 4.   Rodger  M,  Sheppard  D,  Gándara  E,  Tinmouth  A. Haematological problems in obstetrics. Best Pract Res Clin Obstet Gynaecol. 2015; 29: 671–684.
• 5.   Raut  AK,  Hiwale  KM. Iron deficiency anemia in pregnancy. Cureus. 2022; 14: e28918.
• 6.   Haider  BA,  Olofin  I,  Wang  M,  Spiegelman  D,  Ezzati  M,  Fawzi  WW. Anaemia, prenatal iron use, and risk of adverse pregnancy outcomes: systematic review and meta-analysis. BMJ. 2013; 346: f3443.
• 7.   Peña-Rosas  JP,  De-Regil  LM,  Garcia-Casal  MN,  Dowswell  T. Daily oral iron supplementation during pregnancy. Cochrane Database Syst Rev. 2015; 2015: Cd004736.
• 8.   Athe  R,  Dwivedi  R,  Pati  S,  Mazumder  A,  Banset  U. Meta-analysis approach on iron fortification and its effect on pregnancy and its outcome through randomized, controlled trials. J Family Med Prim Care. 2020; 9: 513–519.
• 9.   World Health  O. Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity. Geneva: World Health Organization; 2011. Contract No.: WHO/NMH/NHD/MNM/11.1.
• 10.   Shinar  S,  Shapira  U,  Maslovitz  S. Redefining normal hemoglobin and anemia in singleton and twin pregnancies. Int J Gynaecol Obstet. 2018; 142: 42–47.
• 11.   Young  MF,  Oaks  BM,  Tandon  S,  Martorell  R,  Dewey  KG,  Wendt  AS. Maternal hemoglobin concentrations across pregnancy and maternal and child health: a systematic review and meta-analysis. Ann N Y Acad Sci. 2019; 1450: 47–68.
• 12.   Scanlon  KS,  Yip  R,  Schieve  LA,  Cogswell  ME. High and low hemoglobin levels during pregnancy: differential risks for preterm birth and small for gestational age. Obstet Gynecol. 2000; 96: 741–748.
• 13.   Dewey  KG,  Oaks  BM. U-shaped curve for risk associated with maternal hemoglobin, iron status, or iron supplementation. Am J Clin Nutr. 2017; 106: 1694S–1702S.
• 14.   Rahmati  S,  Azami  M,  Badfar  G,  Parizad  N,  Sayehmiri  K. The relationship between maternal anemia during pregnancy with preterm birth: a systematic review and meta-analysis. J Matern Fetal Neonatal Med. 2020; 33: 2679–2689.
• 15.   Xu  S,  Wang  W,  Li  Q, et al. Association of Maternal Longitudinal Hemoglobin with Small for Gestational Age during Pregnancy: A Prospective Cohort Study. Nutrients. 2022; 14: 1403.
• 16.   Jung  J,  Rahman  MM,  Rahman  MS, et al. Effects of hemoglobin levels during pregnancy on adverse maternal and infant outcomes: a systematic review and meta-analysis. Ann N Y Acad Sci. 2019; 1450: 69–82.
• 17.   Makino  S,  Takeda  J,  Takeda  S,  Watanabe  K,  Matsubara  K,  Nakamoto  O, et al. New definition and classification of “Hypertensive Disorders of Pregnancy (HDP). Hypertens Res Pregnancy. 2019; 7: 1–5.
• 18.   Metzger  BE,  Gabbe  SG,  Persson  B, et al. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care. 2010; 33: 676–682.
• 19.   Itabashi  K,  Miura  F,  Uehara  R,  Nakamura  Y. Japanese neonatal anthropometric charts. Pediatr Int. 2014; 56: 702–708.
• 20.   Kanda  Y. Investigation of the freely available easy-to-use software ‘EZR’ for medical statistics. Bone Marrow Transplant. 2013; 48: 452–458.
• 21.   Vogel  JP,  Chawanpaiboon  S,  Moller  AB,  Watananirun  K,  Bonet  M,  Lumbiganon  P. The global epidemiology of preterm birth. Best Pract Res Clin Obstet Gynaecol. 2018; 52: 3–12.
• 22.   Parisi  F,  Savasi  VM,  di Bartolo  I,  Mandia  L,  Cetin  I. Associations between first trimester maternal nutritional score, early markers of placental function, and pregnancy outcome. Nutrients. 2020; 12: 1799.
• 23.   Jung  E,  Romero  R,  Yeo  L, et al. The etiology of preeclampsia. Am J Obstet Gynecol. 2022; 226: S844–S866.