Epidural labor analgesia is a potential risk factor for increased blood loss within two hours after delivery in women with gestational hypertension: a retrospective cohort study

Abstract

Background

Postpartum hemorrhage remains the leading cause of maternal mortality during childbirth worldwide. Given that pregnant women with gestational hypertension are particularly vulnerable to blood volume reduction, there is currently insufficient evidence to determine whether epidural labor analgesia impacts these patients within the first 2 h postpartum.

Methods

A retrospective cohort study was conducted involving 1,903 term parturients with hypertensive disorders of pregnancy admitted to Guangdong Women and Children Hospital between January 2012 and December 2021. Participants were categorized into two groups based on receipt of epidural labor analgesia: the analgesia group (n = 884) and non-analgesia group (n = 1,019). Primary outcomes included early postpartum hemorrhage (PPH) defined as ≥ 300 mL blood loss within 2 h post-delivery and associated risk factors. Statistical analyses were performed using chi-square tests for categorical variables, Mann–Whitney U tests for continuous variables, and multivariate logistic regression to evaluate independent associations between epidural analgesia and PPH.

Results

Results showed significantly higher 2-h postpartum blood loss ≥ 300 mL in epidural analgesia vs non-analgesia groups among gestational hypertension patients (54.0% vs 46.0%, P = 0.003). Multivariate analysis identified epidural analgesia as an independent risk factor for postpartum hemorrhage (adjusted OR = 1.304, 95% CI:1.013-1.680, P = 0.039). Obstetric complications including placental adhesion (OR = 2.405) and macrosomia (OR = 2.644) also correlated with hemorrhage risk (all P < 0.05). Birth canal injury demonstrated a protective association (adjusted OR = 0.63, 0.49–0.81, P < 0.001).

Conclusion

Epidural labor analgesia may elevate the risk of early postpartum hemorrhage (≤ 2 h) in parturients with gestational hypertension. Clinical practice should prioritize intensive postpartum surveillance and individualized analgesic protocols to balance analgesic efficacy and hemorrhagic risk.

Background

Postpartum hemorrhage (PPH) remains a leading cause of maternal morbidity and mortality globally [1,2,3]. Early identification and intervention are critical. PPH is defined as ≥ 500 mL blood loss within 24 h of vaginal delivery, ≥ 1000 mL after cesarean section, or hypovolemic signs/symptoms regardless of volume [4]. The primary etiologies include uterine atony, birth canal lacerations, placental abnormalities, or coagulopathy, often occurring in combination [5]. Key risk factors include gestational hypertension, low maternal BMI, polyhydramnios, multiple gestation, macrosomia, multiparity, previous cesarean delivery, prolonged labor, and anesthetic use [6]. Approximately 75.8–80% of total 24-h blood loss occurs within the first 2 h postpartum [7, 8], making this period (or up to 4 h for high-risk cases) a critical window for PPH surveillance. Early hemorrhage strongly predicts subsequent 24-h blood loss [9], necessitating intensive monitoring during this timeframe.

Hypertensive disorders of pregnancy (HDP) are characterized by hypertension complicating pregnancy, affecting approximately 116 per 1,000,000 women of reproductive age globally [10]. Gestational hypertension represents a HDP subclass defined by new-onset hypertension ≥ 20 weeks' gestation without proteinuria or preeclamptic features, often asymptomatic or minimally symptomatic. While most affected women undergo vaginal delivery in the absence of obstetrical indications, 10–25% progress to preeclampsia [11]. Patho physiologically, gestational hypertension induces systemic arteriolar vasospasm, compromising uteroplacental perfusion and placentation. This ischemia–reperfusion injury during labor increases risk of uterine atony—a leading cause of postpartum hemorrhage (PPH)—due to impaired myometrial contractility from prolonged hypoxia [12].

Epidural labor analgesia is nationally covered under China's maternity insurance system, with our institution achieving an 85% epidural analgesia rate including hypertensive parturients. Both US and Chinese obstetric anesthesia guidelines [13, 14] recommend epidural analgesia for hypertensive parturients. Prior studies demonstrate no increased PPH risk [15] and even reduced hemorrhage risk [16] with epidural analgesia, but evidence specific to hypertensive parturients remains inconclusive. This study therefore focuses on hypertensive parturients, using 2-h postpartum blood loss as the primary outcome due to its critical role in PPH determination. Patients were stratified into ≥ 300 mL or < 300 mL groups, with epidural analgesia exposure evaluated as an independent variable. Additionally, established PPH risk factors [2, 6] including low BMI, polyhydramnios, multiple gestation, macrosomia, multiparity, prior cesarean, and labor duration were included as covariates.

Methods

Inclusion and exclusion criteria

Diagnostic criteria for gestational hypertension

Gestational hypertension was defined as the onset of hypertension (systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg) after 20 weeks of gestation, with blood pressure normalization within 12 weeks postpartum; urine protein was negative ( −); diagnosis was confirmed postpartum [17].

Inclusion criteria

Eligible participants were:

• • Singleton term pregnant women (37–41 weeks’ gestation) meeting the gestational hypertension diagnostic criteria

• • Candidates for trial of vaginal delivery (TOVD) without contraindications for vaginal birth (no evidence of cephalopelvic disproportion, fetal malposition, or other cesarean indications)

• • Aged 18–49 years with American Society of Anesthesiologists (ASA) physical status classification II–III

• • No severe dysfunction of major organ systems (cardiac, hepatic, renal, pulmonary functions within normal limits)

• • Hemostatic competence: platelet count ≥ 80 × 10⁹/L and normal coagulation profile

• • Ability to tolerate labor analgesia and vaginal delivery process

Exclusion criteria

Participants were excluded if they:

• • Had preeclampsia (any severity), eclampsia, or chronic hypertension (with or without superimposed preeclampsia)

• • Had pre-existing circulatory/hematological disorders (e.g., congenital heart disease, coagulopathy)

• • Had contraindications to neuraxial anesthesia (e.g., spinal deformity, anticoagulant use)

• • Had active systemic infections (e.g., sepsis, meningitis)

• • Required operative vaginal delivery (forceps/vacuum extraction) or breech

Additionally excluded were cases of:

• • Preeclampsia/eclampsia spectrum disorders

• • Chronic hypertension in pregnancy (ICD-10: O10.0–O10.3)

• • Operative vaginal delivery (forceps/vacuum/breech extraction)

Methods of labor analgesia

Epidural labor analgesia was initiated after the first stage of labor entered the active stage (cervical opening of 3 cm). The parturients provided informed consent for anesthesia. The parturient was positioned on her left side, and the space between the lumbar vertebrae L2 and L3 for the median puncture was selected. Following a successful puncture, a 3–4-cm long steel wire epidural catheter was inserted. After successful epidural catheterization, a test dose was administered to determine the position of the epidural catheter, excluding intravascular or subarachnoid catheterization. The test dose consisted of 1% lidocaine 3 mL, followed by a 10 mL loading volume of 0.1% ropivacaine and 0.8 µg/mL of sufentanil, and was administered into the spinal area. After 30 min, a continuous epidural infusion (100 mL of 0.1% ropivacaine, and 0.3 µ g/mL sufentanil) was administered at a rate of 10 mL/h. The analgesic pump infusion was stopped 2 h after delivery of the fetus, and the epidural catheter was removed.

Case collection

We obtained the permission to extract electronic medical records from the Information and Medical Record Department of Guangdong Women and Children Hospital. By using the SQL query tool, we input the required fields and conditions for extraction into the tool and saved the results as an Excel spreadsheet. A total of 3,582 cases from January 1st, 2012, to December 30th, 2021, were extracted. For the quality—control of the screened data, we randomly selected cases and compared them with the original records, and also checked the diagnostic codes. Outliers were identified, and the corresponding cases were removed. Specifically, we excluded 291 cases with poor epidural labor analgesia outcomes, 656 cases without records of previous cesarean section, 554 cases without records of oxytocin use during labor, 2 cases of breech delivery assistance, 90 cases of forceps delivery, and 86 cases of vacuum extraction. Ultimately, 1,903 cases that met the criteria were included in the study. After storage, the data were encrypted and regularly backed up. The flowchart of data collection is presented in Fig. 1.

Fig. 1

Case screening flowchart. The figure depicts the flowchart for the case screening

Variables and grouping

Sample size calculation

Sample size was calculated using G-Power 3.1 (Faul et al., 2007) for a two-sample t-test (bilateral α = 0.05, power = 0.95). Based on pilot data (mean PPH: Group A 280 mL [SD 120], Group N 320 mL [SD 140]), the effect size (Cohen’s d = 0.31) required 494 participants/group (total n = 988). Anticipating 10% attrition, 1,903 participants were recruited (Group A = 884, Group N = 1,019).

Grouping criteria

Participants were stratified into two groups based on intrapartum analgesia administration:

• • Analgesia Group (Group A): Receiving neuraxial labor analgesia (epidural or combined spinal-epidural)

• • Non-Analgesia Group (Group N): No labor analgesia or intravenous opioid use only

Dependent variable

Postpartum Hemorrhage (PPH) within 2 Hours

Measurement: Gravimetric method (standardized for obstetric blood loss quantification)

Procedure: Weigh dry blood-collecting materials (sterile gauze pads, chux pads) pre-delivery

Post-delivery: Weigh saturated materials using a calibrated scale (precision: ±1g) Calculate blood loss (mL) using formula: blood loss (mL) = [wet weight of blood dressing after delivery of the fetus (g) - dry weight of dressing before blood collection (g)] /1.05 (blood specific gravity g/mL). The maternal classification was based on blood loss of 300 mL or more within 2 h of birth. Unit: Milliliters (mL), rounded to the nearest whole number

Independent variables

Demographic & obstetric characteristics

• Age (years, calculated as years at delivery)

• Pre-pregnancy weight (kg, measured at first prenatal visit)

• Body mass index (BMI, kg/m²; calculated as weight/height.²)

• Gestational age (completed weeks, confirmed by early ultrasound dating)

• Parity (nulliparous/primiparous/multiparous [≥ 2 prior deliveries])

• Previous cesarean delivery (yes/no, with indication documented)

• Oxytocin induction/augmentation (yes/no, defined as ≥ 2 mU/min intravenous oxytocin)

• Epidural analgesia (yes/no, Neuraxial technique)

Intrapartum variables

Labor stage (first/second/third stage duration, minutes)

Fetal position (occiput anterior/occiput posterior/breech/other malposition)

Newborn Apgar score (1-min and 5-min scores, 0–10 scale)

Pathological conditions

Placental adhesion (morphological confirmation: accreta/increta/percreta)

Macrosomia (birth weight ≥ 4000 g, confirmed by delivery room measurement)

Polyhydramnios (amniotic fluid index ≥ 25 cm via ultrasound)

Birth canal injury (perineal laceration: modified O'Connell classification; vaginal/uterine trauma).

Database construction

Data curation & software

De-identified clinical data were systematically curated and imported into IBM SPSS Statistics 25.0 (IBM Corp., Armonk, NY, USA) using a double-entry validation protocol. A password-protected database was established with locked variable definitions to ensure consistency.

Quantitative variables included maternal age (appropriate:20–34, advanced:35–39, super-advanced: ≥ 40), pre-pregnancy BMI (1: < 18.5,2:18.5–23.9,3: ≥ 24.0), gestational age (completed weeks:1:37⁺⁰–37⁺⁶, 2:38⁺⁰–38⁺⁶, 3:39⁺⁰–39⁺⁶, 4: ≥ 40⁺⁰), newborn Apgar Score: 0–10 scale (1/5/10 min), Labor stage(minutes).Categorical variables (0/1/2coding), macrosomia: No (0),Yes (1);parity:nulliparous (0),primiparous (1),multiparous (≥ 2);previous cesarean delivery; oxytocin Induction; fetal position: Cephalic (LOA/ROA) (1)- Non-cephalic (2);placental adhesion: No (0),Yes (1);polyhydramnios: No (0),Yes (1);birth canal injury: No (0),Yes (1);epidural analgesia No (0),Yes (1); PPH: < 300 mL (0) ≥ 300 mL (1).Logic checks ensured consistency (e.g., previous cesarean → parity ≥ 1). Gestational age validated by early ultrasound, fetal position confirmed by obstetricians.

Quality control: Double entry by two researchers (κ = 0.98 for categorical variables).

Statistical analysis

IBM SPSS Statistics 25.0 (IBM Corp., Armonk, NY, USA) was used for primary data analysis. R version 4.2.2 (R Core Team, Vienna, Austria) was employed for forest plot visualization.

Data distribution & univariate analysis

Quantitative variables

Normality Assessment: Shapiro–Wilk test (α = 0.05 threshold for non-normality). Normal Distribution: Reported as mean ± standard deviation (SD); analyzed using independent samples t-test. Non-Normal Distribution: Reported as median (interquartile range, IQR); analyzed using Mann–Whitney U test.

Categorical variables

Reported as frequency (n) and proportion (%), Analyzed using: Pearson chi-square test (expected cell counts ≥ 5), Fisher’s exact test (expected cell counts < 5).

Multivariate analysis (Logistic Regression)

Model specification

Dependent Variable: Postpartum hemorrhage (PPH) ≥ 300 mL (binary: 0 = No, 1 = Yes).

Independent Variables:

Demographics: Age (years), BMI (kg/m²), Labor stage(min)

Obstetric history: Previous cesarean (yes/no), parity (0/1/ ≥ 2)

Intrapartum factors: Gestational age (completed weeks), fetal position (cephalic/non-cephalic), oxytocin use (yes/no), epidural analgesia (yes/no)

Output reporting

Effect Estimates: Odds ratio (OR) with 95% confidence interval (CI), Model Fit: Hosmer–Lemeshow test (p > 0.05 indicates good fit), Statistical Significance: P < 0.05 (two-tailed, α = 0.05).

Results

Bivariate analyses (Table 1) of 1,903 participants (16.9% PPH ≥ 300 mL) identified five significant associations: multiparity (37.3% vs. 62.7% in primiparas; χ² = 5.08,P = 0.024), epidural analgesia (54.0% vs. 46.0%; χ² = 8.60, P = 0.003), placental adhesion (9.3% vs. 4.4%; χ² = 12.32, P < 0.001), macrosomia (5.0% vs. 2.2%; χ² = 6.77,P = 0.009), and birth canal injury (43.2% vs. 56.8%; χ² = 16.80, p < 0.001). Prolonged first (390 vs. 335 min; z =  − 2.61, P = 0.009) and shorter third (5 vs. 6 min; z =  − 2.30, P = 0.022) stages correlated with PPH. No significant associations were found for age, BMI, previous cesarean, fetal position, oxytocin use, polyhydramnios, precipitate labor, or Apgar scores (all P ≥ 0.125).

Table 1 Bivariate Predictors of 2-Hour Postpartum Hemorrhage (≥ 300 mL) in 1,903 Parturients

Multivariate logistic regression (Table 2 and Fig. 2) in this observational cohort (n = 1,903) identified three independent risk factors for 2-h PPH (≥ 300 mL): epidural analgesia (adjusted OR = 1.30, 95%CI: 1.01–1.68, P = 0.039), macrosomia (adjusted OR = 2.64, 1.38–4.88, P = 0.002), and placental adhesion (adjusted OR = 2.41, 1.50–3.77, P < 0.001). Birth canal injury demonstrated a protective association (adjusted OR = 0.63, 0.49–0.81, P < 0.001), potentially confounded by shorter labor duration. No significant effects were observed for age, BMI, or oxytocin use (all P ≥ 0.076). The model (Hosmer–Lemeshow p = 0.514) explained 17.9% of PPH variance (Nagelkerke R² = 0.179) and included 16.9% PPH cases. These findings highlight the utility of epidural use, fetal size, and placentation status in PPH risk prediction, while challenging the traditional assumption of birth canal injury as a bleeding risk. All effect estimates were reported with 95% CI, and no significant multicollinearity was detected (tolerance ≥ 0.85).

Table 2 Multivariate logistic regression for 2-h postpartum hemorrhage (n = 1,903)

Fig. 2

Forest plots. The figure depicts the Forest plot of the results of the multifactor logistic regression analysis of the impact on the postpartum hemorrhage within 2 h after delivery

Discussion

Main findings

In this single-center retrospective cohort study (n = 1,903), epidural labor analgesia was identified as an independent risk factor for 2-h postpartum hemorrhage (PPH ≥ 300 mL) in pregnant women with gestational hypertension (adjusted OR = 1.30, 95%CI: 1.01–1.68, P = 0.039). This association remained significant after adjusting for confounders (macrosomia, placental adhesion) in multivariate logistic regression (Table 2).

Comparison with existing literature

This study's findings demonstrate significant discrepancies with previous non-pregnant hypertensive parturient cohorts. Existing evidence indicates no significant association between epidural labor analgesia and postpartum hemorrhage (PPH) in non-pregnant hypertensive populations [18, 19], with analgesic effects potentially reducing bleeding risk via hemodynamic stabilization mechanisms [20]. Based on the unique pathophysiological characteristics of hypertensive disorders of pregnancy (HDP) [21], we propose the following mechanistic explanations:

• 1. Uteroplacental perfusion dysregulation: HDP-induced vasospasm of uterine spiral arteries reduces placental perfusion, impairing myometrial cell energy metabolism and compromising postpartum uterine contractility;

• 2. Vascular-coagulation cascade disruption: Vascular endothelial injury triggers inflammatory mediator release, establishing an ischemia-inflammation-thrombosis vicious cycle that exacerbates uteroplacental ischemia and coagulation dysfunction;

• 3. Uterine contractility modulation abnormalities: HDP may disrupt contraction regulation through dual mechanisms: ① downregulating oxytocin receptor expression/affinity; ② inhibiting prostaglandin synthesis pathways, both leading to diminished uterine contractile efficacy.

Previous investigations have identified birth canal trauma [22, 23]and maternal high body mass index (BMI) [24] as risk factors for postpartum hemorrhage (PPH). Conversely, low BMI confers significant protection against PPH, particularly reducing severe bleeding risk [25]. A landmark study employing multivariable regression models to control for labor patterns and confounding variables demonstrated that oxytocin induction does not independently predict PPH or estimated blood loss [26]. Classically, precipitous labor increases bleeding through birth canal injury; however, this study's bivariate and multivariate analyses revealed no association between birth canal trauma and 2-h postpartum blood loss.

This discrepancy may relate to obstetric team interventions during precipitous labor, and importantly, reflects the study's exclusive focus on 2-h bleeding metrics—clinicians must remain alert to delayed hemorrhage within 24 h. Variable comparisons with prior literature confirm consistency in non-analgesia factors: all PPH associations align with historical findings except for epidural labor analgesia.

Determination of 2-hour postpartum hemorrhage threshold (≥ 300 mL)

In different regions, postpartum hemorrhage may be measured by methods such as visual estimation and gauze weighing. However, the amount of blood loss measured by these methods is often lower than the actual amount. More accurate measurement methods, such as quantitative blood loss (QBL) [27] and calculated blood loss (CBL) via pre/postpartum hematocrit [28], are difficult to implement clinically in developing countries. A secondary outcome of a UK cohort study mentioned that the median postpartum blood loss after spontaneous vaginal delivery measured by QBL was 300 mL [29]. A Chinese study in low-risk primiparas showed that the median blood loss in normal vaginal delivery was approximately 300 mL (control group: < 300 mL; case group: ≥ 500 mL). The 300 mL threshold was determined based on data distribution characteristics to sensitively identify early hemorrhage risk in low-risk primiparas [30]. In 2014, Say et al. emphasized that rapid blood loss (e.g., > 300 mL within 1 h postpartum) is a critical indicator for predicting postpartum hemorrhage, particularly in resource-limited settings [3]. A 2015 review in BMC Pregnancy and Childbirth highlighted that most clinical blood loss assessment methods are underestimated, recommending a diagnostic threshold of 300 mL to improve detection [31].

This study used gauze weighing (1 g = 1 mL) to calculate blood loss, which may underestimate actual hemorrhage. Based on the above evidence, a 300 mL threshold was adopted. The first 2 h postpartum is a critical observation window for immediate postpartum hemorrhage. A lower threshold reduces missed diagnosis risk and enables early intervention. Thus, 300 mL was established as the 2-h postpartum hemorrhage threshold.

Research limitations

• 1. Inherent retrospective design constraints:

While multivariate modeling controlled for measured confounders, residual confounding from unmeasured factors (e.g., epidural analgesic dose, maternal psychological stress) cannot be excluded.

• 2. Sample selection bias:

Exclusion of parturients with systolic blood pressure ≥160 mmHg and/or diastolic blood pressure ≥105 mmHg may limit generalizability to severe preeclampsia cases, requiring validation in this subgroup.

• 3. Blood loss quantification method:

Gauze weighing, the primary method in Chinese clinical practice, introduces systematic underestimation (20–40% discrepancy) compared to reference standards (e.g., quantitative blood loss [QBL] [27]).

• 4. Outcome limitations:

  • ◦ Focus restricted to 2-h postpartum hemorrhage (PPH), without evaluating 24-h cumulative blood loss or long-term morbidities (e.g., anemia, transfusion requirements).

  • ◦ No longitudinal analysis of hemoglobin trajectory pre/post-PPH.

• 5. Threshold validation:

The 300 mL threshold lacks guideline endorsement and may lead to overdiagnosis in low-risk populations. Clinical decision-making should incorporate multidimensional assessment (e.g., hemodynamic parameters, coagulation profile).

Conclusions

This single-center retrospective cohort study (n = 1903) with stringent inclusion criteria identified a potential independent association between neuraxial analgesia and early postpartum hemorrhage (EPPH, defined as ≥ 300 mL blood loss within 2 h of delivery) in parturients with uncomplicated gestational hypertension (UGH).Future research should prioritize rigorous prospective investigations to validate these findings, while exploring the clinical feasibility of personalized analgesic protocols. Such optimization strategies may include: 1) opioid dosage titration, 2) analgesic duration optimization, and 3) pharmacological modulation to mitigate the inhibitory impact of neuraxial analgesia on uterine contractility – thereby ameliorating the trade—off between analgesic efficacy and hemorrhagic risk in this obstetric population.