Abstract
Background
Most prior studies on anemia in pregnancy, preterm birth (PTB), and non-reassuring fetal status (NRFS) use only single-trimester hemoglobin concentrations and fail to capture gestational hemoglobin changes. This study aimed to assess the associations of pregnancy hemoglobin trajectories with these two adverse perinatal outcomes.
Methods
A retrospective cohort study was conducted using pregnancy records from Longquanyi District of Chengdu Maternity and Child Health Care Hospital (from July 2024 to June 2025) , involving 2,954 pregnant women with complete serial hemoglobin measurements. The outcome measures were PTB and NRFS. Latent growth mixture modeling was applied to identify distinct hemoglobin trajectories during pregnancy. Binary logistic regression was constructed to explore the association between hemoglobin trajectories and birth outcomes.
Results
The incidence of anemia during pregnancy was 29.8% in 2954 pregnant women. Three distinct hemoglobin trajectories were identified: Traj-1 (high-rapid decline), Traj-2 (moderate-gradual decline), and Traj-3 (low-delayed recovery). Compared with the Traj-2 group, both the Traj-1 (adjusted odds ratio (aOR) = 1.780, 95% confidence interval (CI) 1.144, 2.769) and Traj-3 (aOR = 2.051, 95% CI 1.084, 3.881) groups were significantly associated with PTB. Moreover, the Traj-3 (aOR = 2.580, 95% CI 1.223, 5.440) group exhibited an increased risk of NRFS.
Conclusion
This study suggests that specific longitudinal hemoglobin trajectories, particularly high-rapid decline and low-delayed recovery patterns, exhibited an increased risk of PTB and NRFS, highlighting the clinical value of dynamic monitoring over single-point assessment in the practice of maternal and child health care.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12884-026-09688-9.
Keywords: Anemia in pregnancy, Hemoglobin, Latent growth mixture modeling, Preterm birth, Non-reassuring fetal status
Introduction
Preterm birth (PTB) is defined as delivery before 37 weeks’ gestational age [1], which has short- and long-term negative impacts on the development of newborns [2, 3]. As a frequent adverse pregnancy outcome, PTB is regarded as a major global public health concern [4, 5]. The worldwide PTB rate was estimated at 9.9% (95% CI 9.1, 11.2) in 2020 [6], while over 7.0% of live births were preterm in mainland China in 2022 [7]. Accordingly, identifying PTB risk factors is critical to reducing preterm incidence and improving neonatal outcomes.
Anemia, a major pregnancy complication characterized by subnormal hemoglobin (Hb) or red blood cell levels, is significantly associated with PTB and other adverse birth outcomes [8–10]. In 2023, it is estimated that worldwide, 35.5% of pregnant women aged 15–49 years were anemic [11]. Given this high prevalence, understanding the relationship between maternal Hb and PTB is of clinical importance.
Hb is a key indicator of anemia [12]. Based on Hb concentrations, the World Health Organization (WHO) has established trimester-specific cutoff values for diagnosing anemia [13]. However, most prior studies rely on single-trimester Hb measurements and fail to explore nonlinear relationships between Hb and PTB or other adverse birth outcomes, leading to inconsistent findings across the existing literature. While some studies report that first-trimester maternal anemia raises PTB risk [14, 15], several other investigations find no significant association between PTB and anemia occurring in the second or third trimester [16–18]. However, additional studies have shown that second- or third-trimester anemia is associated with PTB [19, 20]. Furthermore, higher Hb levels during pregnancy do not always confer protection against PTB or other adverse birth outcomes [21]. Notably, recent evidence points to a U-shaped relationship between maternal Hb concentration and adverse birth outcomes, with elevated risks at both low and high Hb levels [22–25]. These findings suggest that a single measurement is insufficient to capture the dynamic physiological changes and associated risks throughout gestation.
To address this gap, it is essential to consider the entire pregnancy as a continuum and investigate dynamic changes in Hb trajectories across gestation in relation to PTB. Notably, studies have shown that non-reassuring fetal status (NRFS) due to intrauterine hypoxia may be influenced by Hb concentration during pregnancy [26, 27]. Given the paucity of studies exploring the association between maternal Hb and NRFS, we incorporated NRFS as a secondary outcome. Therefore, the present study aimed to identify distinct longitudinal trajectories of Hb and analyze their associations with PTB and NRFS.
Methods
Data sources
The data were extracted from the comprehensive Hospital Information System of Longquanyi District of Chengdu Maternity and Child Health Care Hospital. The system comprises two main modules: an outpatient workstation and an inpatient workstation, which systematically archive outpatient visit records and inpatient medical registration information, respectively. Hb and ferritin concentrations during pregnancy were obtained from the outpatient workstation, and the covariate data and birth outcomes were collected from the inpatient workstation. The two datasets were linked via a unique personal identification number.
Potential covariates were selected through literature reviews, encompassing sociodemographic characteristics, menstrual and reproductive history, maternal anthropometric characteristics, and gestational complications [28, 29]. Maternal sociodemographic characteristics included maternal age at delivery, employment status, marital status, and ethnicity. Menstrual and reproductive history included menarcheal age, menstrual duration, cycle length, gravidity, and parity. Maternal anthropometric characteristics included height, pre-pregnancy weight, and prenatal weight. Gestational complications included gestational diabetes mellitus (GDM), pregnancy-induced hypertension (PIH), and eclampsia gravidarum. Pre-pregnancy body mass index (BMI) was calculated as pre-pregnancy weight in kilograms divided by height in meters squared.
The present study was reviewed and approved by the Ethics Committee of Longquanyi District of Chengdu Maternity and Child Health Care Hospital (Approval number: LQFY-2026-009). Given the retrospective study design, all data were extracted from the Hospital Information System and subsequently de-identified, resulting in negligible risk to the participants. Informed consent was waived by the institutional ethics committee.
Study population
All enrolled pregnant women underwent their first prenatal visit at 6–13 weeks of gestation. In total, we have identified 3,242 pregnant women who delivered at the hospital between July 2024 and June 2025 from the Hospital Information System. Inclusion criteria for the study were: (1) singleton pregnancy; (2) availability of at least four Hb concentration records from different trimesters during pregnancy, and (3) delivery at or after 32 weeks of gestation. Exclusion criteria included incomplete birth outcome data. Finally, 2,954 participants remained for the final analysis (Fig. 1).
Fig. 1.
Flow chart of participants in the study
Exposure
Hb levels were assessed via routine blood tests conducted at various stages of pregnancy. To facilitate the description of Hb concentrations and anemia incidence across different gestational periods, the pregnancy period was categorized into six distinct intervals: the first trimester (T1, 6–12 weeks of gestation), the early second trimester (T2, 13–20 weeks), the late second trimester (T3, 21–28 weeks), the early third trimester (T4, 29–32 weeks), the late third trimester (T5, after 33 weeks), and the pre-delivery period (T6, within one week prior to delivery). If a pregnant woman has two or more sets of test data for a certain trimester, the minimum value should be taken to represent the most severe manifestation of maternal anemia [30, 31].
According to WHO, maternal anemia was defined as Hb < 110 g/L in the first and third trimesters, and < 105 g/L in the second trimester [13]. High hemoglobin was defined as Hb higher than 130 g/L [21].
Definition of outcome
Our primary clinical outcome was PTB, defined as delivery occurring before 37 weeks of gestation. The secondary outcome was NRFS, defined as progressive fetal hypoxia or acidemia stemming from insufficient fetal oxygenation, with clinical manifestations including abnormal fetal heart patterns, decreased fetal movement, fetal growth restriction, and meconium-stained amniotic fluid [27]. As described in Sect."Data sources", birth outcomes data were extracted from the Hospital Information System, with all clinical diagnoses adjudicated by attending clinicians according to individual maternal and fetal clinical findings.
Statistical analysis
Normality of continuous variables was tested using the Kolmogorov-Smirnov test. Normally distributed variables were expressed as mean (standard deviation, SD), while non-normal variables were expressed as median (interquartile range, IQR). Categorical variables were presented as frequencies (percentages). Intergroup comparisons of population characteristics across distinct Hb trajectory subgroups were performed using one-way ANOVA or the Kruskal-Wallis H test for continuous variables and the chi-square test for categorical variables.
Latent growth mixture modeling (LGMM) was applied to identify distinct trajectories of Hb concentration during pregnancy [32]. The LGMM divides the total sample into multiple latent classes, each comprising individuals with similar trajectories and characteristics. Hb levels during pregnancy follow different latent trajectories rather than a single homogeneous pattern. Analyzing only the average Hb trend of the overall population may obscure clinically meaningful subgroup differences and potentially mask associations with birth outcomes. Moreover, LGMM appropriately accommodates residual within-class individual variation. LGMM is therefore well-suited to address the limitations of average trend analysis.
We performed LGMM to identify potential Hb trajectories with all available measurements throughout pregnancy for each individual using the “lcmm” package in R [33]. With gestational age specified as the time variable and Hb concentration as the outcome, we constructed quadratic growth models ranging from 1 to 5 latent classes. After selecting the number of groups, the significance of linear, quadratic, and cubic polynomials was assessed to optimize trajectory shapes. The optimal model was selected according to the Akaike information criterion (AIC), the Bayesian information criterion (BIC), average posterior probability (AvePP), and percentage of each class. Lower AIC and BIC values, the percentage of each class > 5%, and AvePP > 0.7 indicate that the model is acceptable. After comprehensive consideration of model fit indices (Table S1) and the clinical interpretability of each subgroup, the 3 classes of trajectories were considered to be the best potential classes. Each participant was assigned to the trajectory group with the highest posterior class-membership probability.
Binary logistic regression models were constructed to evaluate the associations of Hb trajectories with PTB and NRFS, with trajectory 2 specified as the reference group and with adjustment for maternal age at delivery, employment status, menarcheal age, menstrual flow duration, cycle length, gravidity, parity, pre-pregnancy BMI, GDM, PIH, and eclampsia gravidarum.
To verify the robustness of the findings, sensitivity analysis was restricted to pregnant women with more than four Hb records throughout pregnancy. Statistical analysis was conducted using SPSS software (version 22.0) and R software (version 4.0.3), and a two-sided p value < 0.05 was considered statistically significant.
Results
Hb concentration and anemia prevalence across trimesters
In the first trimester, Hb concentration was the highest, with a prevalence rate of 41.4%. The incidence of anemia during pregnancy was the highest in the early third trimester, at 29.8%. Overall, the Hb concentrations followed a trend of an initial decrease and a subsequent increase, and anemia prevalence changed in the opposite pattern (Table 1).
Table 1.
Hemoglobin concentrations across pregnancy trimesters
| Trimester | n | Hb | (n,%) | |
|---|---|---|---|---|
| High hemoglobin | Maternal anemia | |||
| T1 - first trimester | 2292 | 128 (122–135) | 948 (41.4%) | 104 (4.5%) |
| T2 - early second trimester | 2479 | 116 (110–122) | 125 (5.0%) | 271 (10.9%) |
| T3 - late second trimester | 2920 | 114 (108–119) | 83 (2.8%) | 449 (15.4%) |
| T4 - early third trimester | 2878 | 115 (108–121) | 131 (4.6%) | 859 (29.8%) |
| T5 - late third trimester | 2518 | 119 (111–126) | 342 (13.6%) | 471 (18.7%) |
| T6 - prebirth | 2912 | 119 (111–126) | 376 (12.9%) | 575 (19.7%) |
Continuous variables were presented as mean (SD) or median (IQR) according to their distribution, and categorical variables were shown as frequencies (percentage)
Characteristics of participants across different Hb trajectories
Baseline characteristics across different Hb trajectories were presented in Table 2. The median age of all participants was 29 years, and the median pre-pregnancy BMI was 21.5 kg/m². The prevalence of overweight and obesity was 28.3% in Traj-1 and 18.5% in Traj-3, while the rate of underweight was 16.1% in Traj-3 and 11.2% in Traj-1. Overall, 97.7% of participants were married, and 97.9% were of Han ethnicity. There were significant differences in parity, height, pre-pregnancy weight, pre-pregnancy BMI, prenatal weight, PIH, and eclampsia gravidarum. Basic characteristics were similar between the final analytic sample and those excluded due to missing data (Table S2).
Table 2.
Baseline characteristics at preconception enrollment according to maternal hemoglobin trajectories
| Variable | All women (n = 2954) |
Traj-1 (n = 827) |
Traj-2 (n = 1878) |
Traj-3 (n = 249) |
H/χ2 Value | p-Value |
|---|---|---|---|---|---|---|
| Maternal age at delivery (years) | 29 (27–32) | 29 (27–32) | 29 (27–32) | 29 (27–33) | 1.480 | 0.477 |
| Employment status | 5.086 | 0.079 | ||||
| Employed | 1315 (44.5%) | 341 (41.2%) | 862 (45.9%) | 112 (45.0%) | ||
| Unemployed | 1639 (55.5%) | 486 (58.8%) | 1016 (54.1%) | 137 (55.0%) | ||
| Menarcheal age (years) | 13 (12–14) | 13 (12–14) | 13 (12–14) | 13 (12–14) | 2.251 | 0.324 |
| Menstrual duration (days) | 0.353 | 0.838 | ||||
| ≤ 3 | 435 (14.7%) | 118 (14.3%) | 282 (15.0%) | 35 (14.1%) | ||
| > 3 | 2519 (85.3%) | 709 (85.7%) | 1596 (85.0%) | 214 (85.9%) | ||
| Menstrual cycle (days) | 3.607 | 0.165 | ||||
| ≤ 30 | 2304 (78.0%) | 626 (75.7%) | 1480 (78.8%) | 198 (79.5%) | ||
| > 30 | 650 (22.0%) | 201 (24.3%) | 398 (21.2%) | 51 (20.5%) | ||
| Gravidity | 2 (1–3) | 2 (1–3) | 2 (1–3) | 2 (1–3) | 1.981 | 0.371 |
| Parity | 1 (1–2) | 1 (1–2) | 1 (1–2) | 1 (1–2) | 7.412 | 0.025 |
| Height (cm) | 158 (155–162) | 159 (156–163) | 158 (155–162) | 158 (155–162) | 7.837 | 0.020 |
| Pre-pregnancy weight (kg) | 54 (49–60) | 55 (50–62) | 54 (49–60) | 53 (48–58) | 24.164 | <0.001 |
| Pre-pregnancy BMI (kg/m2) | 21.5 (19.7–23.7) | 21.8 (20.0-24.2) | 21.4 (19.7–23.6) | 20.9 (19.3–23.0) | 18.082 | < 0.001 |
| < 18.5 | 368 (12.5%) | 93 (11.2%) | 235 (12.5%) | 40 (16.1%) | 17.982 | 0.006 |
| 18.5–23.9 | 1889 (63.9%) | 500 (60.5%) | 1226 (65.3%) | 163 (65.5%) | ||
| 24.0-27.9 | 533 (18.0%) | 177 (21.4%) | 321 (17.1%) | 35 (14.1%) | ||
| ≥ 28.0 | 164 (5.6%) | 57 (6.9%) | 96 (5.1%) | 11 (4.4%) | ||
| Prenatal weight (kg) | 68 (62–74) | 69 (63–75) | 68 (62–74) | 66 (61–72) | 28.685 | <0.001 |
| Gestational diabetes | 2.849 | 0.241 | ||||
| Yes | 690 (23.4%) | 205 (24.8%) | 436 (23.2%) | 49 (19.7%) | ||
| No | 2264 (76.6%) | 622 (75.2%) | 1442 (76.8%) | 200 (80.3%) | ||
| Pregnancy-induced hypertension | 6.971 | 0.031 | ||||
| Yes | 153 (5.2%) | 49 (5.9%) | 84 (4.5%) | 20 (8.0%) | ||
| No | 2801 (94.8%) | 778 (94.1%) | 1794 (95.5%) | 229 (92.0%) | ||
| Eclampsia gravidarum | 10.528 | 0.005 | ||||
| Yes | 65 (2.2%) | 21 (2.5%) | 32 (1.7%) | 12 (4.8%) | ||
| No | 2889 (97.8%) | 806 (97.5%) | 1846 (98.3%) | 237 (95.2%) | ||
Continuous variables were presented as mean (SD) or median (IQR) according to their distribution, and categorical variables were shown as frequencies (percentage). p values of continuous variables were calculated with one-way ANOVA or the Kruskal-Wallis H test, and chi-square tests for categorical variables
Abbreviations: SD standard deviation, IQR interquartile range, BMI body mass index
Ferritin levels declined throughout gestation with a prominent drop in the second trimester (Table 3). Ferritin concentrations differed significantly across the three trajectory groups at all gestational trimesters. Traj-1 consistently exhibited lower ferritin than the other two groups across pregnancy, and Traj-2 had significantly lower ferritin than Traj-3 in the third trimester.
Table 3.
Ferritin concentration during different pregnancy trimesters
| Variable | n | All women | Traj-1 | Traj-2 | Traj-3 | H Value | p-Value |
|---|---|---|---|---|---|---|---|
| Ferritin Concentration (µg/L) | |||||||
| T1-first trimester | 2227 | 48.6 (27.7–79.0) | 38.5 (22.5–60.4) | 52.7 (31.8–83.9) | 53.2 (19.0-85.9) | 65.181 | <0.001 |
| T2/T3-second trimester | 2733 | 13.3 (8.3–22.8) | 9.7 (6.9–15.1) | 15.0 (9.3–24.8) | 16.9 (9.8–29.9) | 175.312 | <0.001 |
| T4/T5-third trimester | 2522 | 11.5 (8.1–18.6) | 8.4 (6.6–11.2) | 13.1 (9.1–20.6) | 18.2 (11.3–29.2) | 390.425 | <0.001 |
Continuous variables were presented as mean (SD) or median (IQR) according to their distribution
p values of continuous variables were calculated with one-way ANOVA or the Kruskal-Wallis H test
Birth outcomes were shown in Table 4. The median gestational age of the 2954 pregnant women was 39.1 weeks. The overall incidences of PTB and NRFS were 3.4% and 1.8%, respectively, and there were significant differences among the three groups. Maternal and neonatal health outcomes not analysed in this study are shown in Table S3.
Table 4.
Comparison of birth outcomes among maternal hemoglobin trajectories
| Variable | All women (n = 2954) |
Traj-1 (n = 827) |
Traj-2 (n = 1878) |
Traj-3 (n = 249) |
H/χ2 Value | p-Value |
|---|---|---|---|---|---|---|
| Gestational age at delivery (week) | 39.1 (38.7–39.9) | 39.3 (38.9–39.9) | 39.1 (38.7–39.9) | 39.0 (38.4–39.6) | 20.595 | <0.001 |
| < 37 | 100 (3.4%) | 38 (4.6%) | 48 (2.6%) | 14 (5.6%) | 11.460 | 0.003 |
| ≥ 37 | 2854 (96.6%) | 789 (95.4%) | 1830 (97.4%) | 235 (94.4%) | ||
| Non-reassuring fetal status | 8.136 | 0.017 | ||||
| Yes | 52 (1.8%) | 14 (1.7%) | 28 (1.5%) | 10 (4.0%) | ||
| No | 2902 (98.2%) | 813 (98.3%) | 1850 (98.5%) | 239 (96.0%) |
Continuous variables were presented as mean (SD) or median (IQR) according to their distribution, and categorical variables were shown as frequencies (percentage)
p values of continuous variables were calculated with one-way ANOVA or the Kruskal-Wallis H test, and chi-square tests for categorical variables
Longitudinal Hb trajectories
In this study, three dynamic trajectories were identified based on 16,892 Hb records during pregnancy. Trajectory 1 (Traj-1), named high-rapid decline, had the highest initial Hb levels of about 132 g/L, rapidly decreasing to about 110 g/L. Traj-2, named moderate-gradual decline, had the higher initial Hb level of about 128 g/L, slowly decreasing to the lowest point of about 113 g/L and then rising to 121 g/L. Traj-3, named low-delayed recovery, had the lowest initial Hb level of about 119 g/L, slowly decreasing to the lowest point of about 114 g/L and then rising to 129 g/L. All three trajectory groups exhibited a consistent pattern of an initial decrease followed by an upward trend (Fig. 2). The number and proportion in the three trajectories were 827 (28.0%), 1878 (63.6%), and 249 (8.4%), respectively.
Fig. 2.
Maternal hemoglobin trajectory groups across pregnancy. Three trajectories of hemoglobin dynamics were identified based on initial levels and subsequent changes: (1) high-rapid decline, (2) moderate-gradual decline, and (3) low-delayed recovery
Association of Hb trajectories with PTB and NRFS
Compared with the reference Traj-2 group, after adjustment for maternal age at delivery, employment status, menarcheal age, menstrual flow duration, cycle length, gravidity, parity, pre-pregnancy BMI, GDM, PIH, and eclampsia gravidarum, both the Traj-1 group (aOR = 1.780, 95% CI 1.144, 2.769) and Traj-3 group (aOR = 2.051, 95% CI 1.084, 3.881) were significantly associated with PTB. Traj-3 group also showed a significant association with NRFS (aOR = 2.580, 95% CI: 1.223, 5.440) (Table 5).
Table 5.
The association between maternal hemoglobin trajectories and birth outcomes
| Variable | Traj-1 | Traj-2 | Traj-3 | |
|---|---|---|---|---|
| Preterm Birth | Model1 | 1.836 (1.190, 2.833) | 1 | 2.271 (1.233, 4.183) |
| Model2 | 1.801 (1.163, 2.791) | 1 | 2.289 (1.237, 4.236) | |
| Model3 | 1.780 (1.144, 2.769) | 1 | 2.051 (1.084, 3.881) | |
| Non-reassuring Fetal Status | Model1 | 1.138 (0.596, 2.173) | 1 | 2.764 (1.326, 5.762) |
| Model2 | 1.160 (0.605, 2.226) | 1 | 2.691 (1.285, 5.636) | |
| Model3 | 1.149 (0.599, 2.206) | 1 | 2.580 (1.223, 5.440) |
Traj-1: high-rapid decline; Traj-2: moderate-gradual decline; Traj-3: low-delayed recovery Model 1, crude. Model 2, adjusted for maternal age at delivery, employment status, menarcheal age, menstrual flow duration, cycle length, gravidity, parity, and pre-pregnancy BMI;Model 3, Model 2 plus additionally adjusted for gestational complications such as pregnancy-induced hypertension, gestational diabetes mellitus, and eclampsia gravidarum
Sensitivity analysis
In the sensitivity analysis, used 15,913 records of 2,707 women with more than four Hb records in the whole pregnancy, the results were robust and consistent with the main findings. Compared with the reference Traj-2 group, after adjustment for the same confounding factors, both the Traj-1 (aOR = 1.957, 95% CI 1.215, 3.154) and Traj-3 (aOR = 2.299, 95% CI 1.179, 4.484) groups were significantly associated with PTB (Table S4).
Discussion
In this retrospective cohort study, we identified three distinct nonlinear longitudinal Hb trajectories during pregnancy and found that Hb trajectories were significantly associated with PTB and NRFS. Compared to the moderate-gradual decline trajectory (Traj-2), the high-rapid decline (Traj-1) and low-delayed recovery (Traj-3) trajectories had higher risk of PTB, and the Traj-3 had a higher risk of NRFS. Sensitivity analysis yielded results consistent with the main analyses in both the direction and magnitude of associations, confirming its robustness. Women with a rapid Hb decline throughout pregnancy and no notable recovery in late gestation had a higher risk of PTB. Notably, pregnant women who had low Hb in early gestation remained at increased risk of PTB and NRFS, even when their Hb gradually rose to relatively high levels later in pregnancy. Our findings highlighted the value of dynamic serial monitoring of Hb throughout pregnancy, in contrast to relying solely on a single measurement.
Accumulating evidence supports the general association between abnormal Hb dynamics and adverse pregnancy outcomes. Deborah A et al. indicated that pregnant women with persistent anemia or a declining Hb trajectory had a higher risk of PTB, at 15.8% and 14.8%, respectively [31]. Similarly, Zhang et al reported that, compared to the stable trajectory, the descending Hb trajectories, but not ascending ones, were associated with significantly increased risks [34]. It should be noted, however, that these prior studies primarily assessed monotonic Hb changes (e.g., persistent anemia or declining trends), whereas our study identified non-linear latent trajectories. Therefore, rather than directly aligning our trajectory groups with previously defined "declining" or "ascending" categories, we interpret our findings as associations between distinct non-linear Hb trajectory patterns and adverse outcomes.
During normal pregnancy, maternal plasma volume expands more than red blood cell mass, leading to a relative decrease in Hb concentration, a phenomenon known as physiological hemodilution. This process reduces blood viscosity and optimizes uteroplacental circulation perfusion [35]. The Traj-1 group with 28.0% participants is characterized by Hb levels rapidly declining from a relatively high baseline to a low level and then remaining stable. This pattern may reflect reduced oxygen-carrying capacity in the third trimester, leading to insufficient oxygen supply to the placenta and uterine myometrium. Hypoxia in the placenta is associated with vascular remodeling, hypertension, metabolic changes, oxidative stress, mitochondrial dysfunction, and endoplasmic reticular stress [36], which directly or indirectly precipitate PTB [37, 38]. Other studies showed that low Hb may cause chronic hypoxia, inducing a stress response with increased placental corticotropin releasing hormone production. If this occurs too early, it may induce PTB [21, 39].
Conversely, the Traj-3 group with 8.4% participants features a rapid Hb rise during the third trimester, reaching 129 g/L before delivery. We interpret this pattern as reflecting an inadequate or delayed physiological response. The rationale is as follows. First, the incidence of first-trimester anemia in Traj-3 was 19.5%, and these women had significantly higher third-trimester serum ferritin levels than the other two groups. Elevated ferritin in late pregnancy may reflect inflammation rather than true iron repletion [40, 41] and has been associated with increased PTB risk [42]. Second, according to Philip J. Steer's research [43], elevated Hb (> 120 g/L) may indicate inadequate plasma volume expansion, which is associated with reduced placental blood flow, fetal growth restriction, PTB, and an increased risk of pre-eclampsia. In this study, the Traj-3 group had significantly higher rates of eclampsia (4.8% vs. 2.5% and 1.7%) compared to the other two groups, supporting this mechanism. Third, a large body of studies has confirmed that early-pregnancy anemia is closely related to PTB [14, 15]. The high prevalence of early anemia in Traj-3 suggests that the increased risk of PTB may remain irreversible even if iron status improves to a relatively high level in late pregnancy.Additionally, the Traj-3 group was significantly associated with NRFS. The mechanism may be that Hb concentration is low in the first trimester, but due to reduced normal plasma expansion and hemoconcentration in the third trimester, the Hb concentration rises rapidly, leading to reduced placental-fetal perfusion and consequently causing fetal distress[26, 44].
Major strengths include the prospective design, repeated measures, use of advanced trajectory modeling, and adjustment for a comprehensive set of potential confounders. By applying the LGMM method, we categorized the dynamic changes in Hb in pregnant women, which may help implement targeted care and early interventions for high-risk groups in clinical practice. Furthermore, the results of multiple subgroup and sensitivity analyses were consistent with the primary analysis, further reinforcing the current study's conclusions. However, our study is not without its limitations. First, our study population was recruited from a single district-level maternal and child health institution in Chengdu, limiting the generalizability of the findings to broader populations and other geographical settings. Second, despite adjustments, residual confounding by unmeasured factors such as detailed dietary intake, underlying hemoglobinopathies, or subclinical infection cannot be ruled out. Thus, future studies should aim to address these limitations to validate our findings and explore their generalizability. Additionally, the biological mechanisms underlying the observed associations remain speculative, and future studies are needed to explore causal pathways.
Conclusions
In conclusion, this study shows that specific hemoglobin trajectories during pregnancy are significantly associated with adverse perinatal outcomes. Both a high-rapid decline and a low–delayed recovery pattern are linked to preterm birth, while the latter is also associated with non-reassuring fetal status. Monitoring the dynamic pattern of Hb change, rather than isolated values, offers a more nuanced approach to identifying at-risk pregnancies and paves the way for more personalized and timely antenatal interventions.
Supplementary Information
Acknowledgments
Institutional review board statement
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Longquanyi District of Chengdu Maternity and Child Health Care Hospital (Approval number: LQFY-2026-009, dated 26 February 2026).
Abbreviations
- PTB
Preterm birth
- Hb
Hemoglobin
- WHO
World Health Organization
- NRFS
Non-reassuring fetal status
- GDM
Gestational diabetes mellitus
- PIH
Pregnancy-induced hypertension
- BMI
Body mass index
- SD
Standard deviation
- IQR
Interquartile range
- LGMM
Latent growth mixture modeling
- AIC
Akaike information criterion
- BIC
Bayesian information criterion
- AvePP
Average posterior probability
Authors' contributions
Conceptualization, Y.Z. and X.H.; methodology, H.N.; software, X.Y. and H.N.; validation, H.N. and X.Y.; formal analysis, X.Y.; investigation, W.Y. and J.T.; resources, W.Y. and J.T.; data curation, X.Y.; writing—original draft preparation, X.Y.; writing—review and editing, H.N.; visualization, X. Y.; supervision, Y.Z. and X.H.; project administration, Y.Z. and X.H.; funding acquisition, Y.Z. and X.H. All authors have read and agreed to the published version of the manuscript.
Funding
This research was supported by the Health System Scientific Research Fund of Longquanyi District, Chengdu, Sichuan Province (No. WJKY2026107) and the Chengdu Medical Research Project, Sichuan Province (No. 2026297).
Data availability
The data supporting the findings of this study are available from the corresponding authors upon reasonable request.
Declarations
Ethics approval and consent to participate
This retrospective study used routinely collected clinical data, which have already been desensitized. According to relevant local regulations, the requirement for informed consent was waived by the Ethics Committee of Longquanyi District of Chengdu Maternity and Child Health Care Hospital.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Xiaohong Yu, Hongli Nie, Xuexue Huang and Yongmei Zhao contributed equally to this work.
Contributor Information
Xuexue Huang, Email: 315764437@qq.com.
Yongmei Zhao, Email: 1558690278@qq.com.
References
- 1.Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet Lond Engl Engl. 2008;371:75–84. 10.1016/S0140-6736(08)60074-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Lapidaire W, Proano A, Blumenberg C, Loret de Mola C, Delgado CA, Del Castillo D, et al. Effect of preterm birth on growth and blood pressure in adulthood in the Pelotas 1993 cohort. Int J Epidemiol Engl. 2023;52:1870–7. 10.1093/ije/dyad084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Villar J, Restrepo-Mendez MC, McGready R, Barros FC, Victora CG, Munim S, et al. Association Between Preterm-Birth Phenotypes and Differential Morbidity, Growth, and Neurodevelopment at Age 2 Years: Results From the INTERBIO-21st Newborn Study. JAMA Pediatr United States. 2021;175:483–93. 10.1001/jamapediatrics.2020.6087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Bradley E, Blencowe H, Moller A-B, Okwaraji YB, Sadler F, Gruending A, et al. Born too soon: global epidemiology of preterm birth and drivers for change. Reprod Health Engl. 2025;22:105. 10.1186/s12978-025-02033-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ashorn P, Ashorn U, Muthiani Y, Aboubaker S, Askari S, Bahl R, et al. Small vulnerable newborns-big potential for impact. Lancet Lond Engl Engl. 2023;401:1692–706. 10.1016/S0140-6736(23)00354-9. [DOI] [PubMed] [Google Scholar]
- 6.Ohuma EO, Moller A-B, Bradley E, Chakwera S, Hussain-Alkhateeb L, Lewin A, et al. National, regional, and global estimates of preterm birth in 2020, with trends from 2010: a systematic analysis. Lancet Lond Engl Engl. 2023;402:1261–71. 10.1016/S0140-6736(23)00878-4. [DOI] [PubMed] [Google Scholar]
- 7. Xie Y, Liang J, Mu Y, Chen P, Wang Y, Liu Z, et al. Prevalence estimates of small vulnerable newborns in China (2012-22): a modelling study. Lancet Glob Health Engl. 2026;14:e539–51. 10.1016/S2214-109X(25)00498-X. [DOI] [PubMed] [Google Scholar]
- 8.Benavente MT, Geifman N, Bath S, Ahankari A. Maternal Anemia and Offspring Outcomes in India: A Scoping Review. Anemia. 2025;2025:2850956. 10.1155/anem/2850956. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Benson AE, Lo JO, Caughey AB. Iron Deficiency and Iron Deficiency Anemia During Pregnancy-Opportunities to Optimize Perinatal Health and Health Equity. JAMA Netw Open. 2024;7:e2429151. 10.1001/jamanetworkopen.2024.29151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tsamantioti E, Alfven T, Hossin MZ, Razaz N. Maternal anaemia and risk of neonatal and infant mortality in low- and middle-income countries: a secondary analysis of 45 national datasets. BMJ Glob Health. 2025;10. 10.1136/bmjgh-2023-014654. [DOI] [PMC free article] [PubMed]
- 11.WHO global anaemia estimates. key findings, 2025. Geneva: World Health Organization; 2025. [Google Scholar]
- 12.Means RT. Iron Deficiency and Iron Deficiency Anemia: Implications and Impact in Pregnancy, Fetal Development, and Early Childhood Parameters. Nutrients Switz. 2020;12. 10.3390/nu12020447. [DOI] [PMC free article] [PubMed]
- 13.Guideline on haemoglobin cutoffs to define anaemia in individuals and populations. Geneva: World Health Organization. 2024 [cited 2025 Dec 4]. http://www.ncbi.nlm.nih.gov/books/NBK602198/. Accessed 4 Dec 2025. [PubMed]
- 14.Shi H, Chen L, Wang Y, Sun M, Guo Y, Ma S, et al. Severity of Anemia During Pregnancy and Adverse Maternal and Fetal Outcomes. JAMA Netw Open. 2022;5:e2147046. 10.1001/jamanetworkopen.2021.47046. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chen Y, Zhong T, Song X, Zhang S, Sun M, Liu X, et al. Maternal anaemia during early pregnancy and the risk of neonatal outcomes: a prospective cohort study in Central China. BMJ Paediatr Open. 2024;8:e001931. 10.1136/bmjpo-2023-001931. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.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 Engl. 2020;33:2679–89. 10.1080/14767058.2018.1555811. [DOI] [PubMed] [Google Scholar]
- 17.Wu L, Sun R, Liu Y, Liu Z, Chen H, Shen S, et al. High hemoglobin level is a risk factor for maternal and fetal outcomes of pregnancy in Chinese women: A retrospective cohort study. BMC Pregnancy Childbirth. 2022;22:290. 10.1186/s12884-022-04636-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Pathirathna ML, Haruna M, Sasaki S, Yonezawa K, Usui Y, Hagiwara Y. Maternal Hemoglobin Concentrations in Early and Late Pregnancy and Adverse Birth Outcomes: A Nationwide Cohort Study in Sri Lanka. J Nutr. 2025;155:3938–48. 10.1016/j.tjnut.2025.06.029. [DOI] [PubMed] [Google Scholar]
- 19.Go H, Hashimoto K, Kyozuka H, Maeda H, Nishigori H, Sato A, et al. Maternal hemoglobin levels and neonatal outcomes: the Japan Environment and Children’s Study. J Matern Fetal Neonatal Med Engl. 2022;35:10472–80. 10.1080/14767058.2022.2130237. [DOI] [PubMed] [Google Scholar]
- 20.Barut A, Mohamud DO. The association of maternal anaemia with adverse maternal and foetal outcomes in Somali women: a prospective study. BMC Womens Health. 2023;23:193. 10.1186/s12905-023-02382-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Burden CA, Smith GC, Sovio U, Clayton GL, Fraser A. Maternal hemoglobin levels and adverse pregnancy outcomes: individual patient data analysis from 2 prospective UK pregnancy cohorts. Am J Clin Nutr. 2023;117:616–24. 10.1016/j.ajcnut.2022.10.011. [DOI] [PubMed] [Google Scholar]
- 22.Tang L, Luo Y, Sheng Y, Lai T, Song W, Yang X, et al. Hemoglobin concentrations in early pregnancy and their associations with adverse pregnancy outcomes in Western China: a population-based study. BMC Pregnancy Childbirth. 2024;24:761. 10.1186/s12884-024-06968-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Olaniyan H, Weyand AC. Optimal haemoglobin concentration in pregnancy. Lancet Haematol. 2023;10:e700–2. 10.1016/S2352-3026(23)00206-5. [DOI] [PubMed] [Google Scholar]
- 24.Lian Y, Lv Y, Qiao Y, He T. Relationships between hemoglobin levels at admission and adverse maternal and perinatal outcomes in patients with preeclampsia. PLoS ONE. 2025;20:e0335079. 10.1371/journal.pone.0335079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Young MF, Oaks BM, Rogers HP, Tandon S, Martorell R, Dewey KG, et al. Maternal low and high hemoglobin concentrations and associations with adverse maternal and infant health outcomes: an updated global systematic review and meta-analysis. BMC Pregnancy Childbirth. 2023;23:264. 10.1186/s12884-023-05489-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Yip R. Significance of an abnormally low or high hemoglobin concentration during pregnancy: special consideration of iron nutrition123. Am J Clin Nutr. 2000;72. 10.1093/ajcn/72.1.272S. :272S-279S. [DOI] [PubMed]
- 27.Gravett C, Eckert LO, Gravett MG, Dudley DJ, Stringer EM, Mujobu TBM, et al. Non-reassuring fetal status: Case definition & guidelines for data collection, analysis, and presentation of immunization safety data. Vaccine Neth. 2016;34:6084–92. 10.1016/j.vaccine.2016.03.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Peng Z, Si S, Cheng H, Zhou H, Chi P, Mo M, et al. The Associations of Maternal Hemoglobin Concentration in Different Time Points and Its Changes during Pregnancy with Birth Weight Outcomes. Nutrients. 2022;14. 10.3390/nu14122542. [DOI] [PMC free article] [PubMed]
- 29.Ohuma EO, Jabin N, Young MF, Epie T, Martorell R, Pena-Rosas JP, et al. Association between maternal haemoglobin concentrations and maternal and neonatal outcomes: the prospective, observational, multinational, INTERBIO-21st fetal study. Lancet Haematol. 2023;10:e756–66. 10.1016/S2352-3026(23)00170-9. [DOI] [PubMed] [Google Scholar]
- 30.Ringshaw JE, Hendrikse CJ, Wedderburn CJ, Bradford LE, Williams SR, Nyakonda CN, et al. Persistent impact of antenatal maternal anaemia on child brain structure at 6–7 years of age: a South African child health study. BMC Med. 2025;23:94. 10.1186/s12916-024-03838-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Randall DA, Patterson JA, Gallimore F, Morris JM, Simpson JM, McGee TM, et al. Haemoglobin trajectories during pregnancy and associated outcomes using pooled maternity and hospitalization data from two tertiary hospitals. Vox Sang. 2019;114:842–52. 10.1111/vox.12839. [DOI] [PubMed] [Google Scholar]
- 32.Nagin DS, Jones BL, Elmer J. Recent Advances in Group-Based Trajectory Modeling for Clinical Research. Annu Rev Clin Psychol. 2024;20:285–305. 10.1146/annurev-clinpsy-081122-012416. [DOI] [PubMed] [Google Scholar]
- 33.Nguena Nguefack HL, Page MG, Katz J, Choiniere M, Vanasse A, Dorais M, et al. Trajectory Modelling Techniques Useful to Epidemiological Research: A Comparative Narrative Review of Approaches. Clin Epidemiol. 2020;12:1205–22. 10.2147/CLEP.S265287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Zhang P, Shen Y, Zou G, Xiang X. Associations of hemoglobin trajectories and time in target hemoglobin range with adverse pregnancy outcomes: a real-world pregnancy records-based study. Front Glob Womens Health. 2025;6:1682957. 10.3389/fgwh.2025.1682957. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Costantine MM. Physiologic and pharmacokinetic changes in pregnancy. Front Pharmacol. 2014;5:65. 10.3389/fphar.2014.00065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.van Patot MCT, Ebensperger G, Gassmann M, Llanos AJ. The hypoxic placenta. High Alt Med Biol. 2012;13:176–84. 10.1089/ham.2012.1046. [DOI] [PubMed] [Google Scholar]
- 37.Lien Y-C, Simmons RA. Deciphering the Molecular Dialogue: Mitochondria, Epigenetics, and Extracellular Vesicles in Placental Function and Pregnancy Complications. Compr Physiol United States. 2025;15:e70040. 10.1002/cph4.70040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Lis N, Lamnisos D, Bograkou-Tzanetakou A, Hadjimbei E, Tzanetakou IP. Preterm Birth and Its Association with Maternal Diet, and Placental and Neonatal Telomere Length. Nutrients Switz. 2023;15. 10.3390/nu15234975. [DOI] [PMC free article] [PubMed]
- 39.Ochedalski T, Lachowicz A. Maternal and fetal hypothalamo-pituitary-adrenal axis: different response depends upon the mode of parturition. Neuro Endocrinol Lett Swed. 2004;25:278–82. [PubMed] [Google Scholar]
- 40.Arosio P, Levi S. Ferritin, iron homeostasis, and oxidative damage1, 2 1Guest Editor: Mario Comporti 2This article is part of a series of reviews on Iron and Cellular Redox Status. The full list of papers may be found on the homepage of the journal. Free Radic Biol Med. 2002;33:457–63. 10.1016/S0891-5849(02)00842-0. [DOI] [PubMed] [Google Scholar]
- 41.Yuan X, Gao Y, Zhou W, Long W, Liu J, Wang H, et al. Effect of the inflammatory response on serum indices of iron status in late pregnancy. J Trace Elem Med Biol. 2020;61:126516. 10.1016/j.jtemb.2020.126516. [DOI] [PubMed] [Google Scholar]
- 42. Zeng Y, He J, He G. The impact of increased serum ferritin levels in late pregnancy on pregnancy outcomes: a multi-center cohort study in China. BMC Pregnancy Childbirth Engl. 2026;26. 10.1186/s12884-026-08910-y. [DOI] [PMC free article] [PubMed]
- 43.Steer PJ. Maternal hemoglobin concentration and birth weight. Am J Clin Nutr. 2000;71. 10.1093/ajcn/71.5.1285s. :1285S-7S. [DOI] [PubMed]
- 44. Gaillard R, Eilers PHC, Yassine S, Hofman A, Steegers EAP, Jaddoe VWV. Risk factors and consequences of maternal anaemia and elevated haemoglobin levels during pregnancy: a population-based prospective cohort study. Paediatr Perinat Epidemiol Engl. 2014;28:213–26. 10.1111/ppe.12112. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data supporting the findings of this study are available from the corresponding authors upon reasonable request.