Comparison of maternal and neonatal outcomes between general anesthesia and combined spinal-epidural anesthesia in cesarean delivery for pregnancy complicated with placenta previa

Abstract

Background

Placenta previa (PP) involves abnormal placental implantation in the lower uterine segment, partially or completely covering the internal cervical os, and is linked to severe maternal hemorrhage and fetal complications. The optimal anesthetic method between general anesthesia (GA) and combined spinal-epidural anesthesia (CSEA) for cesarean delivery (CD) with PP remains controversial.

Methods

We retrospectively analyzed 550 PP-complicated CD cases from the Longitudinal Placenta Previa Study (LoPPS) conducted in Southwest China between January 2018 and December 2024. Patients received either GA (n = 170) or CSEA (n = 380). Sociodemographic, obstetric, perioperative, and neonatal data were compared. Multivariate linear and logistic regression was employed to assess the association between anesthetic methods and other perioperative factors, and intraoperative blood loss, or neonatal asphyxia while adjusting for potential confounders.

Results

Patients undergoing GA had a higher age (32.4 vs 31.5 years, p = 0.020), higher body mass index (BMI) (26.5 vs 23.5 kg/m², p < 0.001), and greater parity (91.8% vs 78.4%, p < 0.001) compared to those under CSEA. Complete PP was more common in the GA group (47.1% complete PP vs 19.0%, p < 0.001). There were also more placenta accreta spectrum (PAS) (48.8% vs 15.8%, p < 0.001) and hysterectomy (12.9% vs 0.3%, p < 0.001) in the GA group. In accordance with the huge differences in anesthesia indications, the GA group experienced significantly greater intraoperative blood loss (1131.77 ± 77.29 mL vs. 707.50 ± 16.87 mL, p < 0.001), along with correspondingly higher rates of transfusion, including red cell suspension, plasma, and autologous blood (p < 0.001). The incidence of neonatal asphyxia was also significantly higher in the GA group (26.5% vs. 3.7%, p < 0.001), and the rate of preterm birth was notably higher (81.8% vs. 46.3%, p < 0.001). Among CSEA patients, increased blood loss was associated with placenta covering the uterine incision (HR = 58.49, p = 0.017), PAS type (HR = 29.02, p = 0.036), PP type (HR = 34.72, p = 0.048), and surgical duration (HR = 9.35, p < 0.001), while aortic balloon occlusion reduced blood loss (HR = –115.08, p = 0.009). In GA patients, similar risk factors were identified: placenta covering the incision (HR = 71.88, p = 0.015), PAS type (HR = 103.01, p = 0.042), PP type (HR = 106.16, p = 0.046), and surgical duration (HR = 13.83, p < 0.001). Aortic balloon occlusion remained protective in the GA group (HR = –300.01, p = 0.015), while GA (Exp(B) = 1.75, p = 0.002) and types of PAS are associated with increased risks of neonatal asphyxia.

Conclusion

CSEA is a safe option for selected cases of PP-related CD, particularly in the absence of PAS or in cases with milder forms of PP. Though the GA group exhibited greater intraoperative blood loss, due to the significant differences in the indications for anesthesia methods and the observational nature of current study, this should not be simply interpreted as a causative effect of GA on higher intraoperative blood loss. Thorough antenatal ultrasound assessment of placental status is critical. For complex cases, such as those involving severe PAS subtypes or complete PP, coordinated multidisciplinary perioperative management is essential.

Trial registration

ChiCTR2100052428, October 26^th, 2021.

Background

Placenta previa (PP) is a condition in which the placenta abnormally implants in the lower uterine segment, partially or completely covering the internal cervical os [1, 2]. It poses significant maternal morbidity and fetal risks, including life-threatening hemorrhage, need for hysterectomy, as well as placenta accreta spectrum(PAS), preterm birth, fetal growth restriction, and neonatal respiratory distress syndrome and is one of the leading causes of peripartum mortality [3–9]. The incidence of placenta previa varies globally but is estimated with an prevalence in Chinese pregnant women at 1.25% to 1.44% [10, 11].

Currently, the optimal anesthetic approach for cesarean delivery (CD) in patients with PP remains a topic of debate [12]. In general, cesarean delivery performed under combined spinal-epidural anesthesia (CSEA) reportedly exhibits a more stable hemodynamics and lower postoperative complications [13, 14]. In contrast, general anesthesia (GA) exhibits a rapid onset and can facilitate complex surgical procedures, particularly in managing massive hemorrhage during CD complicated by PP. However, GA presents more challenges in airway management and carries a risk of neonatal asphyxia due to the transplacental transfer of anesthetic agents [15].

Clinical data comparing maternal and neonatal outcomes between GA and CSEA in cesarean sections for PP complicated pregnancies are relatively lacking, leaving anesthesiologists without definitive guidance for decision-making. Therefore, we conducted this study to compare maternal and neonatal outcomes in CD of PP-complicated pregnancies under GA versus CSEA. Additionally, we also investigated the association between anesthetic methods, intraoperative blood loss, and neonatal asphyxia.

Materials and methods

Study design

The current study was derived from the Longitudinal Placenta Previa Study (LoPPS), a prospective pregnancy and birth cohort on PP based in Southwest China, the primary objective of which is to evaluate the respective influences of various genetic factors and environmental exposures on the onset and clinical outcomes of PP (Chinese Clinical Trial Registry ChiCTR2000041150, registered on December 19 th, 2020) [16]. The studied cases was retrospectively collected from Chengdu Women’s and Children’s Central Hospital (a municipal maternal and child health center and university-affiliated teaching hospital), and Yunyang County People’s Hospital (a tertiary hospitals at the county level). Figure 1 illustrates the study flowchart. Between January 2018 and December 2024, among approximately 30,000 deliveries at the participating institutions, 568 cases of PP were enrolled in the LoPPS cohort during the study period. After excluding 2 cases involving anesthesia methods other than GA and CSEA, 10 cases underwent anesthetic conversion from CSEA into GA prior to the delivery of neonates, and 6 case of vaginal delivery, a total of 550 patients were included in this study. Based on the type of anesthesia administered, the participants were categorized into GA or CSEA groups. We conducted a comparative analysis of sociodemographic and obstetric characteristics, maternal and neonatal perioperative outcomes, and anesthesia-related characteristics between both groups. Additionally, multivariable linear and logistic regression models were established to identify whether anesthesia methods were associated with intraoperative blood loss and neonatal asphyxia after adjusting potential confounding factors.

Fig. 1.

Study flowchart. Abbreviations: CS: Cesarean section. CSEA: Combined Spinal-Epidural Anesthesia. GA: general anesthesia. MRI: magnetic resonance imaging. PP: placenta previa

Patient selections

Inclusion criteria

1. Prenatal Diagnosis of PP: Identified through prepartum ultrasound examinations.

2. Patients who provided informed consent to be enrolled in the Longitudinal Placenta Previa Study (LoPPS) cohort.

Exclusion criteria:

1. Delivered vaginally.

2. Patients who received anesthesia other than GA or CSEA.

3. The planned CSEA was converted into GA prior to the delivery of neonates.

Diagnosis of PP

Previous studies have demonstrated the high sensitivity, accuracy, and safety in the prenatal diagnosis of PP and PAS under ultrasound examinations performed by experienced ultrasonographers [17, 18]. Given this, routine abdominal ultrasound was performed between 28 and 36 th week of gestation among patients with symptoms like vaginal bleeding or suspected with PP in her previous antepartum obstetric ultrasound examinations across participating medical institutions in this study. Based on standardized sonographic criteria, obstetric ultrasonographers identified suspected cases of placenta previa and provided such information to anesthesiologists for determining the appropriate anesthetic strategy prior to CD. The classification of PP was determined according to the distance between the placental edge and the internal cervical os as follows: (1) without PP, the placental edge was more than 2 cm away from the internal cervical os; (2) Marginal PP, the placental edge was located within 0–2 cm of the internal cervical os; (3) Partial or complete PP, the placenta partially or fully overlapped the internal cervical os [16]. Additionally, prior to both planned and emergency cesarean delivery, obstetric ultrasonographers conducted a final assessment to confirm the placental location relative to the internal cervical os. This evaluation also determined whether the placenta extended over the uterine incision site or was complicated by PAS, including its specific subtype [19].

Indication of CD for pregnancy complicated with PP

With the exception of a minority of marginal PP cases that may allow for vaginal delivery, CD is the standard approach for most pregnancies complicated by PP due to the high risk of massive hemorrhage and maternal–fetal complications [20]. In our institution, planned CD is generally scheduled at 36–37 weeks of gestation in stable patients, whereas emergency CD is reserved for cases with uncontrollable bleeding or fetal distress [21]. The primary indications for CD in pregnancies complicated by PP are as follows: Complete or Partial PP; Persistent Marginal PP at Term; Antepartum hemorrhage; PAS; Fetal Compromise; Multiple previous CD with PP.

Indications for peripartum hysterectomy

In line with previous publications, our study adopted similar indications for peripartum hysterectomy [9, 22]. Specifically, the procedure was indicated for cases with intraoperative blood loss exceeding 1500 mL accompanied by persistent active hemorrhage that could not be promptly controlled.

Preoperative preparation, method selection, and perioperative management of anesthesia for PP complicated CD

In this study, the preoperative anesthesia preparation, choice of anesthetic method, and perioperative anesthetic management of patients with PP were conducted in accordance with the Chinese Expert Consensus on Obstetric Anesthesia in 2023 and other international guidelines [23, 24]. The details are as follows:

Preoperative anesthesia preparation

In addition to the standard obstetric anesthesia preparation, the following key measures were implemented:

1. Identification of placental abnormalities: Special attention was given to cases with central placenta previa, placenta covering a previous CD scar, and PAS, particularly placenta percreta.

2. Assessment of preoperative circulatory function and severity of anemia, especially coagulation status.

3. Hemodynamic monitoring: Radial artery catheterization and internal jugular vein catheter placement were performed when appropriate. For patients at high risk of massive hemorrhage, prophylactic iliac artery balloon occlusion or other preventive measures were considered.

4. Preparation for massive hemorrhage management: For all cases of CD with PP, we ensure the availability of large intravenous access and the tools for massive resuscitation. Specifically, this includes the placement of dual intravenous access, preoperative cross-matched blood preparation, and the readiness of autologous blood salvage devices.

Choice of anesthetic method

1. CSEA: If both maternal and fetal conditions were stable, and the estimated blood loss was minimal, CSEA was preferred, with readiness for conversion to GA if needed.

2. GA: If the fetus was in distress requiring urgent delivery, or if the mother exhibited active bleeding, hypovolemic shock, significant coagulopathy, or disseminated intravascular coagulation (DIC), GA was the preferred approach.

Intraoperative anesthetic management

The induction and maintenance of GA followed standard CD protocols, with attention paid to intravascular volume resuscitation and hemodynamic stability. Key considerations included:

1. Close monitoring of vital parameters, including blood pressure, heart rate, volume status, coagulation indices, electrolytes, and acid–base balance.

2. Ensuring adequate venous and arterial access to allow for rapid fluid and blood product administration.

3. Preventing acute kidney injury by optimizing fluid balance and renal perfusion.

4. DIC prevention and management: For patients suspected of developing DIC, prophylactic administration of low-dose heparin, along with coagulation factor and platelet supplementation, was cautiously considered while awaiting laboratory confirmation.

Statistical analysis

Statistical analyses in this study were conducted using the Statistical Product and Service Solutions (SPSS) software, version 27.0 (IBM Corp., Armonk, NY, USA) and Prism version 9.5.1 (GraphPad, San Diego, CA, USA). Categorical variables were analyzed using the Chi-square test or Fisher’s exact test, with results presented as frequencies and percentages. The continuous variables were expressed as the mean ± standard deviation or median with 25 percentile and 75 percentile, and were compared using either Student’s t test or the Mann–Whitney U test, when appropriate. Subsequently, we performed multivariate linear regression analysis to evaluate the factors associated with intraoperative blood loss. It was conducted separately within the CSEA and GA groups to avoid potential confounding effect arising from the different indications for CSEA and GA. Moreover, the multivariate logistic regression analysis to investigate the sociodemographic and perioperative variables that correlates with neonatal asphyxia. These variables included maternal age, gestational age at pregnancy termination, mode of conception, history of prior cesarean delivery and induced abortion, hepatitis B infection status, history of recurrent pregnancy loss, placental location, presence and classification of placenta accreta spectrum disorders, anesthetic methods, and surgery duration, etc. All statistical tests were two-tailed, with a significance threshold set at p < 0.05.

Result

Sociodemographic and obstetric characteristics of participants

As presented in Table 1, women undergoing GA had a significantly higher mean age than those receiving CSEA (32.4 ± 4.8 vs 31.5 ± 4.1 years, p = 0.020). Pre-pregnancy BMI was also greater in the GA group (26.5 vs 23.5 kg/m², p < 0.001), while gestational weight gain was modestly lower (12.0 ± 3.5 vs 13.3 ± 4.1 kg, p = 0.024). GA group had markedly higher parity and gravidity, with 91.8% of GA patients were multiparous compared to 78.4% in CSEA (p < 0.001), and 62.9% had been pregnant at least twice before versus 31.1% in CSEA (p < 0.001). Correspondingly, a history of multiple abortions was more common in the GA group (52.4% in GA vs 36.8% in CSEA underwent more than 2 prior abortions, p = 0.002). Significantly higher proportion of patients in the GA cohort had a history of prior CD, with two-thirds (≈66%) had one previous CD (vs 28.4% in CSEA, p < 0.001), and 13.5% had at least 2 prior CD (vs 1.3% in CSEA). They were also more likely to have undergone a previous emergency CD (11.8% vs 5.8%, p = 0.015). Consistent with this, uterine scarring was much more frequent in GA patients (61.8% vs 26.6%, p < 0.001). The placenta was often abnormally located in GA cases, with nearly half of GA patients (49.7%) had a placenta completely covering the uterine incision (cesarean scar) compared to only 4.2% in CSEA, with any placental coverage of the scar (partial or complete) in 47.6% vs 8.7% of cases (p < 0.001). Lastly, concurrent miscarriage was more often noted in the GA group (4.1% vs 0.8%, p = 0.009), whereas other comorbidities (e.g. syphilis, hepatitis, hypertension) did not differ significantly between groups.

Table 1.

Sociodemographic and obstetric profile of PP complicated pregnancies undergoing GA or CSEA

	CSE (n = 380)	GA (n = 170)	P value
Age (years)	31.50 ± 4.10	32.42 ± 4.79	0.020a
BMI before pregnancy (kg/m2)	23.52 ± 3.84	25.12 ± 6.05	< 0.001a
Weight gain during pregnancy (kg)	13.27 ± 4.09	11.99 ± 3.53	0.024a
Parity			< 0.001b
1	82 (21.6%)	14 (8.2%)
≥ 2	298 (78.4%)	156 (91.8%)
Gravity			< 0.001b
0	92 (24.2%)	10 (5.8%)
1	170 (44.7%)	53 (31.2%)
≥ 2	118 (31.1%)	107 (62.9%)
Previous abortion			0.002b
0	140 (36.8%)	42 (24.7%)
1	100 (26.3%)	39 (22.9%)
≥ 2	140 (36.8%)	89 (52.4%)
Previous CD			< 0.001b
0	272 (71.6%)	57 (33.5%)
1	103 (27.1%)	90 (52.9%)
2	5 (1.3%)	23 (13.5%)
Previous emergency CD	22 (5.8%)	20 (11.8%)	0.015b
Previous pelvic surgery
0
1
≥ 2
Previous hysteroscopic examinations or surgeries (times)			0.196c
0	332 (87.4%)	156 (91.8%)
1	39 (10.3%)	13 (7.7%)
2	6 (1.6%)	1 (0.6%)
3	3 (0.8%)	0
History of blood transfusion	8 (2.1%)	5 (2.9%)	0.788b
Scarred uterine	101 (26.6%)	105 (61.8%)	< 0.001b
Placenta covering uterine incision			< 0.001b
Partial	17 (4.5%)	18 (13.5%)
Complete	16 (4.2%)	63 (49.7%)
Comorbidity
Concurrent miscarriage	3 (0.8%)	7 (4.1%)	0.009c
Syphilis	8 (2.1%)	2 (1.2%)	0.128c
Type B hepatitis	21 (5.5%)	15 (8.8%)	0.173b
GBS infection	2 (0.5%)	0	0.406c
Anemia			0.112b
None	239 (62.9%)	97 (57.1%)
Iron deficiency anemia	114 (30.0%)	61 (35.9%)
Thalassemia	22 (5.8%)	6 (3.5%)
Anemia due to other causes	5 (1.3%)	6 (3.5%)
GDM	57 (15.0%)	36 (21.2%)	0.166b
Hypertensive disease			0.597c
None	366 (96.3%)	163 (95.9%)
Gestational Hypertension	8 (2.1%)	5 (2.9%)
Mild Preeclampsia	3 (0.8%)	2 (1.2%)
Severe Preeclampsia	3 (0.8%)	0
Hypothyroidism	26 (6.8%)	16 (9.4%)	0.465b
ICP	20 (5.3%)	7 (4.1%)	0.427b
Intrauterine infection	11 (2.9%)	4 (2.4%)	0.718c
Previous myomectomy	1 (0.3%)	2 (1.2%)	0.176c

Abbreviation: BMI Body mass index, PP Placenta Previa, GA General Anesthesia, CSEA Combined Spinal and Epidural Anesthesia, CD Cesarean Delivery, GDM Gestational Diabetes Mellitus, ICP Intrahepatic Cholestasis of Pregnancy, SD Standard Deviation

^aMean ± SD, Student t test

^bNumber(percentage), Chi square test

^cNumber(percentage), Fisher exact test

Maternal perioperative outcomes

The maternal perioperative outcomes of patients underwent different anesthesia methods were listed in Table 2. The GA group had more severe PP, with nearly half (47.1%) had complete PP, compared to 19.0% in the CSEA group. Conversely, marginal PP was more common with CSEA (28.7% vs 4.1%) (p < 0.001). PAS disorders were also markedly more frequent in GA group. In particular, placenta increta and percreta occurred in a combined 48.8% of GA cases, versus 15.8% in CSEA, while cases with no accreta were far fewer in GA (17.7% vs 45.0%) (p < 0.001). The placental attachment site differed as well. GA cases were much more likely to have a posteriorly attached placenta (posterior uterine wall in 47.1% of GA vs 0.5% of CSEA, whereas in CSEA the placenta was usually on the anterior uterine wall (71.1% vs 45.9% in GA) (p < 0.001). Peripartum hysterectomy was significantly more common in the GA group (p < 0.001), with 12.9% of GA patients required a hysterectomy (10% total, 2.9% subtotal) compared to virtually none in CSEA (only 1 case of cornual resection, accounting for 0.3% of). To manage anticipated hemorrhage, preoperative abdominal aortic balloon occlusion was employed in the majority of GA patients (71.2%) but in almost none of the CSEA group (0.3%, p < 0.001). An intrauterine balloon tamponade for hemorrhage control was also used slightly more often in GA (67.1% vs 56.3%, p = 0.045). Unplanned reoperations were rare but occurred only in the GA group (1.8% vs 0%, p = 0.029). Postoperative hypoproteinemia was noted in 14.1% of GA patients, significantly higher than 1.6% in CSEA (p < 0.001). There were also four cases of intraoperative bladder injury in the GA group (2.4%) versus one case (0.3%) in CSEA (p = 0.034). The duration of postoperative antibiotic therapy was longer in GA patients (mean 76.8 vs 50.4 h, p < 0.001). ICU admissions were much more frequent in the GA group, with 15.3% required ICU care postoperatively, compared to only 0.5% of the CSEA group (p < 0.001). Other complications (ileus, fever, and poor incisional wound healing) were not significantly different between groups.

Table 2.

Maternal perioperative outcomes of PP complicated pregnancies undergoing GA or CSEA

	CSE (n = 380)	GA (n = 170)	P value
Type of PP			< 0.001a
Low-lying	3 (0.8%)	0
Marginal	109 (28.7%)	7 (4.1%)
Partial	13 (3.4%)	3 (1.8%)
Complete	72 (19.0%)	105 (47.1%)
Types of PAS			< 0.001b
None	171 (45.0%)	30 (17.7%)
Accreta	149 (39.2%)	57 (33.5%)
Increta	30 (7.9%)	33 (19.4%)
Percreta	30 (7.9%)	50 (29.4%)
Marginal umbilical cord insertion	19 (5.0%)	8 (4.7%)	0.773b
Location of placental attachment			< 0.001a
Frontier uterine wall	270 (71.1%)	78 (45.9%)
Lateral uterine wall	27 (7.1%)	10 (5.9%)
Posterior uterine wall	2 (0.5%)	80 (47.1%)
Uterine Fundus	2 (0.5%)	0
Cesarean Scar Pregnancy	1 (0.3%)	3 (1.8%)	0.095a
Emergency CD	81 (21.3%)	47 (27.6%)	0.104b
Hysterectomy			< 0.001a
None	379 (99.7%)	148 (87.1%)
Cornual	1 (0.3%)	0 (0%)
Subtotal	0 (0%)	5 (2.9%)
Complete	0 (0%)	17 (10%)
Preoperative abdominal aortic balloon occlusion	1 (0.3%)	121 (71.2%)	< 0.001a
Placement of Intrauterine Balloon	214 (56.3%)	114 (67.1%)	0.045b
Unplanned reoperation	0	3 (1.8%)	0.029a
Hypoxemia	1 (0.3%)	2 (1.2%)	0.227a
Hypoproteinemia	6 (1.6%)	24 (14.1%)	< 0.001b
Ileus	2 (0.5%)	2 (1.2%)	0.591a
Fever	66 (17.4%)	41 (24.3%)	0.060b
Maternal death	0	0	–
Poor surgical wound healing	1 (0.3%)	0	1.000a
Injury of bladder	1 (0.3%)	4 (2.4%)	0.034a
Duration of postoperative antibiotic administration (hour)	48.00 (24.00,72.00)	72.00 (48.00, 81.00)	< 0.001d
ICU administration	2 (0.5%)	26 (15.3%)	< 0.001a

Abbreviation: PP placenta previa, GA general anesthesia, CSEA Combined spinal and epidural anesthesia, CD Cesarean Delivery, PP placenta previa, SD Standard Deviation

^aNumber(percentage), Fisher exact test

^bNumber(percentage), Chi square test

^cMean ± SD, Student t test

^dMedian (25 th percentile, 75 th percentile), Mann–Whitney U test

Fetal characteristic and neonatal outcomes

We subsequently investigated the fetal characteristic and neonatal outcomes (Table 3). The mean gestational age at delivery was 35.6 ± 2.0 weeks for GA vs 36.7 ± 1.7 weeks for CSEA (p = 0.001). Consequently, preterm birth was far more common with GA (81.8% of cases) than with CSEA (46.3%, p < 0.001). There were also differences in how the pregnancies were conceived, with the vast majority of GA cases were naturally conceived (97.1% vs 87.1%), whereas CSEA cases included a higher proportion of IVF conceptions (12.9% vs 2.9% in GA, p = 0.001). Twin or multiple gestations and fetal growth restriction rates were similar between groups (no significant differences). Neonatal outcomes were notably worse in the GA group. Neonatal asphyxia was significantly more frequent in GA-delivered infants: 26.5% experienced asphyxia (mostly mild), compared to only 3.7% of neonates in the CSEA group (p < 0.001). There was a single neonatal death in the GA group (0.6%) and none in CSEA (not statistically significant). GA neonates were also smaller on average, with a lower mean birth weight (2631 ± 486 g vs 2881 ± 524 g, p < 0.001) and were shorter in length at birth (p < 0.001). There was no significant difference in congenital anomalies between the groups. Overall, the earlier gestational age and associated complications under GA translated into higher rates of preterm, low-weight babies with neonatal respiratory depression compared to the CSEA group.

Table 3.

Fetal characteristics and neonatal outcomes of PP complicated pregnancies undergoing GA or CSEA

	CSE (n = 380)	GA (n = 170)	P value
Gestational age at delivery (weeks)	36.72 ± 1.74	35.58 ± 1.98	0.001a
Twin or multiple pregnancy	15 (3.9%)	4 (2.4%)	0.344b
FGR	17 (4.5%)	5 (2.9%)	0.592c
Mode of conception			0.001b
Natural conception	331 (87.1%)	165 (97.1%)
IVF	49 (12.9%)	5 (2.9%)
Fetal death	0	1 (0.6%)	1.000b
Fetal malpresentation	63 (16.6%)	36 (21.2%)	0.194c
Single umbilical artery	3 (0.8%)	1 (0.6%)	0.773b
Preterm birth	176 (46.3%)	139 (81.8%)	< 0.001c
Nuchal cord			0.058b
None	258 (67.9%)	132 (77.6%)
Single	121 (31.8%)	38 (22.4%)
Double	1 (0.3%)	0
Neonatal anomaly	16 (4.2%)	5 (3.9%)	0.473c
Neonatal asphyxia			< 0.001b
None	366 (96.3%)	125 (74.2%)
Mild	14 (3.7%)	43 (25.3%)
Severe	0 (0%)	2 (1.2%)
Neonatal death	0 (0%)	1 (0.6%)	1.000b
PROM	31 (8.2%)	7 (4.1%)	0.193c
Oligohydramnios	9 (2.4%)	8 (4.7%)	0.341c
Polyhydramnios	7 (1.8%)	3 (1.8%)	0.637b
Neonatal body length (cm)	48.26 ± 2.829	47.38 ± 2.65	< 0.001a
Neonatal body weight (g)	2888.29 ± 502.906	2631.43 ± 485.77	< 0.001a

Abbreviation: PP placenta previa, CSEA, Combined spinal and epidural anesthesia, FGR Fetal Growth Restriction, GA general anesthesia, GBS Group B Streptococcus, PROM Premature Rupture of Membranes, SD Standard Deviation

^aMean ± SD, Student t test

^bNumber(percentage), Fisher exact test

^cNumber(percentage), Chi square test

Anesthesia related characteristics

Then we analyzed the anesthesia related characteristics of both groups (Table 4). Operative times and blood loss were substantially greater in the GA group. The duration of anesthesia was prolonged under GA, averaging 100.5 ± 5.5 min vs 69.2 ± 2.4 min for CSEA cases (p < 0.001). Similarly, surgery lasted significantly longer with GA (78.6 ± 4.6 vs 49.4 ± 1.0 min, p < 0.001). There was also higher intraoperative blood loss, with an estimated mean hemorrhage volume of 1132 ± 77 mL in GA group, about 60% more than in the CSEA group (708 ± 17 mL, p < 0.001). In line with larger fluid shifts, urine output during surgery was also higher under GA (270 ± 279 vs 135 ± 87 mL, p < 0.001). Their hemodynamic management also differed, with vasopressor support was needed more often with CSEA (to counter spinal-induced hypotension), reflected by a significantly higher use of other vasoactive agents in the CSEA group (28.2%) compared to GA (14.1%, p < 0.001).

Table 4.

Anesthesia related characteristics of PP complicated pregnancies undergoing GA or CSEA

	CSE (n = 380)	GA (n = 170)	P value
Duration of anesthesia	69.16 ± 2.38	100.49 ± 5.46	< 0.001a
Surgical duration (min)	49.41 ± 1.00	78.6 ± 4.610	< 0.001a
Estimated volume of intraoperative bleeding (mL)	707.50 ± 16.87	1131.77 ± 77.29	< 0.001a
Intraoperative urine output (mL)	134.62 ± 87.10	269.98 ± 278.90	< 0.001a
Administration of vasoactive agents
Ephedrine	6 (1.6%)	3 (1.8%)	0.826b
Metaraminol	0	1 (0.6%)	0.372b
Norepinephrine	1 (0.3%)	4 (2.4%)	0.124b
Others	107 (28.2%)	24 (14.1%)	< 0.001c
Preoperative Hb	115.90 ± 12.70	112.82 ± 12.41	0.008a
Postoperative Hb (on 3rd postoperative day)	100.94 ± 15.01	96.34 ± 16.18	0.002a
Blood transfusion
Red cell suspension	0.00 (0.00, 0.00)	0.00 (0.00, 1.50)	< 0.001d
Plasma	0.00 (0.00, 0.00)	0.00 (0.00, 0.00)	< 0.001d
Autologous blood transfusion	0.00 (0.00, 0.00)	0.00 (0.00, 203.00)	< 0.001d

Abbreviation: CSEA Combined spinal and epidural anesthesia, PP Placenta Previa, GA General Anesthesia, Hb hemoglobin, SD Standard Deviation

^aMean ± SD, Student t test

^bNumber(percentage), Fisher exact test

^cNumber(percentage), Chi square test

^dMedian (25 th percentile, 75 th percentile), Mann–Whitney U test

Although preoperative hemoglobin was only slightly lower in the GA group (112.8 ± 12.4 vs 115.9 ± 12.7 g/L, p = 0.008), the drop in hemoglobin post-surgery was more pronounced. By the third postoperative day, mean hemoglobin in GA patients was 96.3 g/L, significantly below the CSEA mean of 100.9 g/L (p = 0.002). On average they received nearly one unit of packed red blood cells per patient (0.95 units), versus virtually no transfusion in the CSEA group (0.13 units, p < 0.001). Plasma transfusion was likewise higher under GA (about 90 mL vs 8 mL in CSEA, p < 0.001), and cell-salvaged autologous blood use was also greater under GA (88 mL vs 16 mL, p < 0.001).

Correlation analysis of sociodemographic and perioperative features with estimated intraoperative blood loss stratified by anesthesia methods

Table 5 presents a stratified multivariate logistic regression analysis examining the associations between sociodemographic and perioperative variables and estimated peripartum blood loss among patients underwent CSEA or GA. The model included maternal age, gestational age at delivery, mode of conception (natural vs. IVF), number of previous CDs, number of abortions, hepatitis B status, history of recurrent miscarriage, placenta covering the uterine incision, PAS, type of PP, use of preoperative abdominal aortic balloon occlusion, and surgical duration. Among patients who underwent CSEA, factors significantly associated with increased blood loss included placenta covering the uterine incision (adjusted HR = 58.49, p = 0.017), type of PAS (adjusted HR = 29.02, p = 0.036), type of PP (adjusted HR = 34.72, p = 0.048), and surgical duration (adjusted HR = 9.35, p < 0.001). Conversely, preoperative aortic balloon occlusion was associated with reduced blood loss (adjusted HR = –115.08, p = 0.009). Among GA patients, significant predictors of increased blood loss included placenta covering the uterine incision (adjusted HR = 71.88, p = 0.015), type of PAS (adjusted HR = 103.01, p = 0.042), type of PP (adjusted HR = 106.16, p = 0.046), and surgical duration (adjusted HR = 13.83, p < 0.001); while the application of aortic balloon occlusion is a protective factor for intrapartum blood loss within GA group(adjusted HR = –300.01, p = 0.015).

Table 5.

Multivariate Logistic Regression Analysis on the Associations Between Sociodemographic and Perioperative Factors and Estimated Peripartum Blood Loss Stratified by Anesthesia Type (CSEA vs GA)

Variables	Adjusted HR	95%CI		P value
		Lower limit	Upper limit
Patients underwent CSEA
Maternal age (years)	4.46	−6.02	14.93	0.403
Gestational age at delivery (weeks)	16.24	−13.53	46.00	0.283
Mode of conception (natural conception or IVF)	41.69	−95.69	179.06	0.550
Times of previous CD	46.70	−43.61	137.01	0.309
Times of abortion	19.92	−6.19	46.03	0.134
Type B hepatitis	−97.45	−281.09	86.20	0.296
History of recurrent miscarriage	−16.78	−195.01	161.45	0.853
Placenta covering uterine incision	58.49	12.85	129.82	0.017
Type of PAS (Accreta/Increta/Percreta)	29.02	12.93	94.88	0.036
Type of PP (none/marginal/partial/complete)	34.72	13.50	132.94	0.048
Preoperative temporary occlusion of abdominal aorta using balloon catheterization	−115.08	−28.92	−201.23	0.009
Surgical duration(min)	9.35	7.16	11.53	< 0.001
Patients underwent GA
Maternal age (years)	−5.71	−30.47	19.06	0.649
Gestational age at delivery (weeks)	−12.05	−91.63	67.53	0.765
Mode of conception (IVF)	−500.77	−1284.98	283.45	0.208
Times of previous CD	114.56	−42.78	271.91	0.152
Times of abortion	−11.30	−95.46	72.87	0.791
Type B hepatitis	−49.34	−428.51	329.84	0.797
History of recurrent miscarriage	72.23	−1104.23	1248.68	0.903
Placenta covering uterine incision	71. 88	39.88	125.51	0.015
Type of PAS (Accreta/Increta/Percreta)	103.01	76.05	255.11	0. 042
Type of PP (none/marginal/partial/complete)	106.16	75.68	328.01	0.046
Preoperative temporary occlusion of abdominal aorta using balloon catheterization	−300.01	−63.61	−663.62	0. 015
Surgical duration (min)	13.83	11.12	16.54	< 0.001

Abbreviations: CSEA Combined spinal and epidural anesthesia, CD Cesarean Delivery, GA General Anesthesia, IVF In Vitro Fertilization, PAS Placenta Accreta Spectrum, PP Placenta Previa

Correlation analysis of sociodemographic and perioperative features with neonatal asphyxia

Correlation analysis of sociodemographic and perioperative features with neonatal asphyxia was conducted using multivariate logistic regression analysis (Table 6). Several factors were found significantly correlated with the severity of neonatal asphyxia. Gestational age at delivery was a significant predictor, earlier delivery was associated with higher odds of asphyxia (p = 0.040). PAS emerged as a strong risk factor for neonatal asphyxia. Even after adjustment, placenta accreta was associated with increased odds and severity of neonatal asphyxia (p = 0.024), and the risk became progressively higher with placenta increta (p = 0.011) and percreta (p = 0.003). The mode of anesthesia also showed an independent association with neonatal outcomes, with CD under GA carried higher odds of neonatal asphyxia compared to those under CSEA (p = 0.002). Maternal demographic factors (age, conception method) were not significantly correlated with neonatal asphyxia risk.

Table 6.

Multivariate logistic regression analysis of sociodemographic, perioperative features and neonatal asphyxia

Variables		95%CI		P value
	Exp (B)	Lower limit	Upper limit
Maternal age (years)	1.27	0.66	3.02	0.395
BMI before pregnancy (kg/m2)	1.17	0.21	1.81	0.763
Gestational age at delivery (weeks)	−0.63	−1.06	0.00	0.040
Mode of conception (Natural conception or IVF)	1.05	0.70	2.20	0.523
Times of previous CD	1.24	0.62	4.04	0.290
Times of abortion	1.20	0.73	3.02	0.843
Type B hepatitis	1.13	0.91	2.27	0.070
History of recurrent miscarriage	1.04	0.24	3.16	0.711
Placenta covering uterine incision	1.55	0.91	3.10	0.088
Type of PAS
Accreta	2.07	1.71	3.13	0.024
Increta	2.91	1.94	4.71	0.011
Percreta	4.42	1.77	7.91	0.003
Type of PP
Low-lying	1.13	0.79	2.09	0.993
Marginal	1.41	0.71	2.31	0.741
Partial	1.58	0.83	3.13	0.534
Complete	1.73	0.87	4.70	0.131
Mode of anesthesia (CSEA or GA)	1.75	1.06	2.74	0.002

Abbreviations: CSEA Combined Spinal and Epidural Anesthesia, CD Cesarean Delivery, GA General Anesthesia, IVF in vitro fertilization, PAS Placenta Accreta Spectrum, PP Placenta Previa

Discussion

PP is a pregnancy complication that significantly impacts maternal and fetal health due to the potential massive hemorrhage. Risk factors for PP include previous CD, uterine surgeries, advanced maternal age, and IVF [8]. The increasing rate of CD has greatly contributed to the rise in PP and its associated complications [25, 26]. In recent decades, we have gradually developed better understandings about the maternal and fetal outcomes and indications for various anesthetic methods for CD [27–30]. However, the choice of anesthesia for CD complicated by PP remains a subject of ongoing debate [12, 31–33].

CSEA provides both effective neural blockade and muscle relaxation, with minimal impact on the fetus. It also offers postoperative analgesia, making it the preferred anesthetic method for CS in obstetric clinical practice. However, anesthetics using CSEA generally have a prolonged duration of effects and are prone to causing vasodilation, thus the sympathetic blockade induced by CSEA can lead to severe hypotension. In cases of massive hemorrhage, this can exacerbate hemodynamic instability. Furthermore, coagulopathy is often present when massive hemorrhage occurs, increasing the risk of hematoma formation at the puncture site [13–15]. On the other hand, GA has a rapid onset and a lower risk of hypotension, making it suitable for patients with conditions such as lumbar spine disease, infections, severe anxiety, or neurological disorders, where CSEA is not recommended. However, GA may increase the risk of fetal exposure to anesthetic agents and can induce maternal vomiting, which may lead to aspiration and complicate airway management [24, 27, 28]. Probably because of these factors, the Chinese Expert Consensus on Obstetric Anesthesia in 2023 does not recommend GA as the preferred anesthetic method for CS [23].

Two or three decades ago, many anesthetists hold the view that GA is the optimal anesthetic methods for CD complicated with PP [34, 35], while the subsequent studies comparing CSEA and GA’s maternal and neonatal outcomes provided accumulating evidence supporting CSEA’s safety over GA [31, 36]. Several studies focusing on Chinese obstetric populations complicated with PP have emerged in recent years [23, 31]. One previous study, which enrolled approximately 60 cases from Southwest China, yielded findings on maternal and neonatal outcomes for GA versus CSEA that were similar to ours [37]. However, that study did not control for potential confounding factors. It is important to note that, according to the current Chinese domestic guideline, GA is recommended in situations where the fetus is in distress or the mother is at high risk of massive hemorrhage [23]. Consequently, comparisons of perioperative outcomes between GA and CSEA are inherently subject to significant bias due to the huge difference of indications.

The severity and type of PP and PAS play crucial roles in determining the anesthetic method for cesarean deliveries. In line with previous studies, our findings indicate that types of PP and PAS are significantly associated with a higher risk of massive hemorrhage within both CSEA and GA groups [33]. In contrast, cases with less extensive placental coverage, such as marginal or partial PP, may be managed safely with CSEA, allowing for faster maternal recovery and potentially fewer complications. These findings underscore the importance of individualized anesthetic planning based on PP and PAS types, and prenatal ultrasound examination on the placenta in case of vaginal bleeding [18]. The prenatal ultrasound can therefore be used as one of the investigations to decide who can safely be treated with CSEA instead of GA. It also highlighted the essential role of multidisciplinary team approach for optimizing outcomes in high-risk obstetric cases [38].

Furthermore, our multivariate logistic regression analysis also demonstrates that GA is significantly associated with neonatal asphyxia. In our obstetric clinical practice, we generally convert patients at high risk of major bleeding from CSEA to GA only after fetal delivery, in order to avoid neonatal asphyxia caused by anesthetic agents used in GA. A meta-analysis of 46 randomized trials involving 3,689 women comparing general, spinal, epidural, and combined spinal-epidural anesthesia for cesarean delivery found that spinal and epidural anesthesia were associated with higher Apgar scores, and epidural anesthesia was linked to better umbilical venous pH [39]. In addition, our findings revealed that though PP types were not significantly associated with the odds or severity of neonatal asphyxia, the concurrence and types of PAS were significant predicators of neonatal asphyxia. Given that a study has shown that pregnant women with PP at 28 weeks of gestation were still diagnosed with previa at 36 weeks [16], we recommend that pregnant women with signs of PP or those already diagnosed with PP be closely monitored during the second and third trimesters. These patients should be promptly referred to maternal and child health centers or comprehensive hospitals with better neonatal resuscitation capabilities, ensuring timely intervention and improved outcomes for both the mother and newborn. Considering the crucial role that GA plays in CD of pregnancies with severe PP or PAS (with a high risk of massive hemorrhage), a multidisciplinary team involving neonatologists experienced in neonatal resuscitation is essential for these cases [27, 38]. However, given the substantial differences in the indications for CSEA and GA, and considering that our results demonstrated significant associations between all three types of PAS and neonatal asphyxia, the poorer neonatal outcome in the GA group may also be related to greater blood loss as a result of more complete previa, percreta and hysterectomy cases as revealed in the prepartum placental ultrasound findings rather than due to the GA itself.

Our study’s strengths include a large sample size, detailed obstetric, perioperative and neonatal profiles, and a relatively comprehensive analysis of maternal and neonatal outcomes, including correlations with neonatal asphyxia and estimated peripartum blood loss. However, the single-center, retrospective design may limit the generalizability of our findings and the likelihood of confounding by different indications of CSEA and GA should also be acknowledged. A prospective, multicenter randomized controlled trial(RCT) is warranted to further our study.

Conclusion

CSEA can be used safely for some cases of PP complicated CD, especially those without PAS or having a less severe type of PP. Active prepartum ultrasound examination of placenta is essential. In complicated cases, like those with severe subtypes of PAS or complete PP, multidisciplinary perioperative cooperation is necessary. Although the GA group demonstrated higher intraoperative blood loss, this should not be simply explained as a causal role of GA on higher intraoperative blood loss, given the substantial differences in anesthesia indications and the observational design of the present study. A large-scale, multicenter RCT on this issue is warranted to clarify their association.

Abbreviations

CI

Confidence Interval

CD

Cesarean Delivery

CSEA

Combined Spinal-Epidural Anesthesia

DIC

Disseminated Intravascular Coagulation

GA

General Anesthesia

HR

Hazard Ratio

ICU

Intensive Care Unit

IVF

In Vitro Fertilization

LoPPS

Longitudinal Placenta Previa Study

MRI

Magnetic Resonance Imaging

PAS

Placenta Accreta Spectrum

PP

Placenta Previa

SPSS

Statistical Product and Service Solutions

Authors’ contributions

Xin Li, Tianjiao Liu, and Chunmei Liu conceptualized the current study. Tianjiao Liu, Xin Li, Chunmei Liu drafted the manuscript. Chunmei Liu, Xinyu Xiao, Xin Li and Yangyang Wang collected the data. Tianjiao Liu, Xin Li, Xinyu Xiao, Zhi Chen, and Chunmei Liu analyzed the data. Chunmei Liu, Xin Li, Tianjiao Liu, and Zhi Chen supervised this project.

Funding

This study was financially supported by the Yingcai Scheme of CWCCH (No. YC2023004), the Japan–China Sasakawa Medical Fellowship (No. 4408), the Chengdu Science and Technology Program (2021-YF05-00627-SN), the Chengdu Key Clinical Specialty Project, and the Tianfu Jincheng Laboratory Foundation (TFJC2023010001).

Data availability

The data analyzed in this study can be obtained from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study adhered to the principles of the Declaration of Helsinki and received ethical approval from the Institutional Review Board of Chengdu Women’s and Children’s Central Hospital (Approval No. 2022208). Written informed consent was obtained from all participants prior to enrollment. To ensure patient confidentiality, all identifiable information was removed during data extraction from the hospital information system.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.