Prenatal ultrasound markers for predicting blood transfusion in placenta accreta spectrum

Abstract

Objective

Placenta accreta spectrum (PAS) is a leading cause of severe bleeding during childbirth and often requires complex surgery and blood transfusions. The aim of this study was to investigate whether prenatal ultrasound findings, particularly those reflecting vascular and cervical topography, can more accurately predict the need for transfusions.

Methods

Sixty-one surgically confirmed PAS cases from two tertiary centers were analyzed. The uterine spairing surgery and cesarean hysterectomy groups were compared in terms of antenatal ultrasonographic markers. Transfusion burden was first adjusted for preoperative hemoglobin level and cesarean hysterectomy status; then residual variability in the model was examined to assess the independent prognostic contribution of ultrasound markers.

Findings

The need for transfused erythrocyte suspension (RBC) was significantly higher in women undergoing cesarean hysterectomy compared to those undergoing uterine-sparing surgery (2.7 ± 1.7 vs. 1.4 ± 1.3 units, p = 0.003). After correction, uterovesical hypervascularity (β = 0.565, p = 0.034) and placental vessel extension to the cervix (β = 0.640, p = 0.011) remained independently associated with transfusion burden.

Conclusion

While classic findings such as loss of clear zone and placental lacunae support the diagnosis of PAS, vascular findings in the lower uterine segment and cervical region better reflect surgical difficulty and hemorrhagic burden. Focusing on these prognostic ultrasound markers can strengthen preoperative risk assessment, improve multidisciplinary surgical planning, and contribute to more reliable prediction of transfusion needs.

Introduction

Placenta accreta spectrum (PAS) disorders constitute a significant clinical problem in current obstetric practice and are associated with high maternal morbidity, particularly due to severe obstetric bleeding and the need for massive blood transfusions. The increase in PAS incidence parallels the rise in cesarean section rates globally [1–3], with a pooled prevalence of 0.17% [4]. This highlights the role of uterine scar remodeling in the pathophysiology of abnormal placental attachment [5].

Historically, PAS has been classified as accreta, increta, and percreta according to the histopathological depth of villous tissue into the uterine wall [6–8]. However, recent evidence challenges this approach and suggests that PAS is not a consequence of excessive trophoblastic invasion, but rather is associated with abnormal uterine healing, impaired decidualization, and loss of the normal uteroplacental separation plane [5, 9, 10]. Current data reveal that surgical difficulty and maternal outcomes are determined by the topographic extent of placental implantation, particularly its relationship to the cervix, lower uterine segment, parametrial tissues, and utero-vesical junction, rather than histological depth [11–15].

Accordingly, the focus of prenatal imaging has shifted from differentiating the histological subtypes of PAS to identifying ultrasonographic findings that can predict surgical difficulty, hemorrhagic risk, and the need for cesarean hysterectomy. Various ultrasound findings such as loss of clear zone, utero-vesical hypervascularization, placental lacunae, bridging vessels, and extension of placental vessels towards or into the cervix have been reported to be associated with increased surgical complexity and adverse maternal outcomes [16–18]. Among these findings, abnormal vascular patterns and signs reflecting cervical extension stand out in terms of predicting heavy bleeding and the need for more radical surgery [2, 19, 20].

Despite advances in prenatal diagnosis, predicting the transfusion burden remains challenging. Total blood transfusion requirements are influenced by numerous factors, including baseline maternal hemoglobin levels and the surgical approach used [21–23]. However, even considering these strong predictors, significant variability in transfusion needs persists, suggesting that some ultrasonographic markers may independently contribute to hemorrhagic risk.

In this context, the aim of this study is to evaluate the relationship between prenatal ultrasonographic findings and transfusion burden in patients diagnosed with placenta accreta spectrum disorder. By adjusting for the effects of preoperative hemoglobin levels and surgical approach, the goal was to identify ultrasonographic markers that independently predict transfusion needs. Furthermore, cases undergoing uterus-preserving surgery versus cesarean hysterectomy were compared in terms of clinical features, surgical outcomes, and neonatal data, focusing particularly on ultrasound findings reflecting placental topography and abnormal vascularization. With this approach, our study aims to contribute to preoperative risk stratification in PAS patients, support multidisciplinary surgical planning, and help predict transfusion requirements more accurately.

Material and method

This retrospective study was conducted between 2015 and 2024 at two tertiary care centers: Etlik Zübeyde Hanım Women’s Health Training and Research Hospital and Etlik City Hospital. The study was conducted in accordance with the Declaration of Helsinki after approval was granted by the hospital’s Medical Research Ethics Committee (Approval number: 30.10.2024–11.2024). Written informed consent was obtained from all participants for both participation and publication. This study represents a secondary analysis of an ongoing institutional cohort of patients with suspected placenta accreta spectrum, in which patient enrollment continued over the study period [17].

Between 2015 and 2024, a total of 158 patients were referred to our tertiary care centers with a prenatal suspicion of placenta accreta spectrum (PAS) based on ultrasound findings. All patients underwent a standardized prenatal ultrasound assessment, including both transabdominal and transvaginal ultrasound examinations, performed by an experienced maternal–fetal medicine specialist. Ultrasound evaluation focused on placental location, placental–myometrial interface, utero-vesical vascular patterns, and cervical involvement.

Following delivery, all cases were systematically assessed intraoperatively using a topographic and clinical approach. This assessment included evaluation of placental location relative to the cervix, lower uterine segment and parametrial tissues, the presence of abnormal hypervascularization, dense adhesions at the utero-vesical junction, and the feasibility of placental separation.

Of the initial 158 patients, 61 cases were confirmed as placenta accreta spectrum disorders based on a combination of intraoperative clinical findings, surgical observations, and supportive histopathological evaluation. These 61 patients constituted the final study population. The final cohort was divided according to the type of surgical management. Uterus-sparing surgical techniques were performed in 25 patients, while 36 patients underwent cesarean hysterectomy. The patient selection process is illustrated in Fig. 1.

Fig. 1.

Patient flowchart

Patients were excluded if they had a posterior placenta previa not covering the previous cesarean scar, fetuses with major anomalies, pregnancies without detectable intraplacental fetal vascular structures on ultrasound, missing clinical data, patients who declined participation, women who expressed no future desire for fertility preservation, as this could introduce bias in the selection of surgical approach, and cases in which intraoperative findings revealed placenta previa without evidence of PAS.

Data collection

Preoperative demographic and clinical data were retrospectively collected, including maternal age, body mass index (BMI), comorbidities, gravidity, parity, number of miscarriages, and number of previous cesarean sections. Prenatal ultrasonographic findings associated with placenta accreta spectrum (PAS) were recorded, together with prenatal classification according to the ultrasound staging system for PAS [24]; Intraoperative data included perioperative hemoglobin change, number of transfused erythrocyte suspension units, and intraoperative clinical grading of PAS based on standardized surgical assessment [25, 26]; Postoperative data comprised histopathological reports and neonatal outcomes, including gestational age at delivery, birth weight, and 5-minute Apgar scores.

The ultrasonographic markers evaluated in this study were loss of the clear zone, utero-vesical hypervascularity, presence of placental lacunae, and placental vessels extending toward or into the cervix. These markers were selected based on their established relevance for PAS assessment and their reported association with surgical complexity and maternal morbidity in the current literature [11, 24].

All patients underwent both transabdominal and transvaginal ultrasound examinations using a Voluson S10 ultrasound system (General Electric Medical Systems, Milwaukee, WI, USA). All examinations were performed by a single maternal–fetal medicine specialist with extensive experience in PAS imaging.

Given the long study period and the evolution of PAS imaging criteria over time, not all ultrasonographic markers were available in every case. Nevertheless, all recorded markers were included in the analyses to maximize the use of available data.

Surgical procedure

All surgical procedures were performed by a dedicated multidisciplinary team led by a senior surgeon with more than 20 years of experience in the management of placenta accreta spectrum (PAS). Perioperative surgical findings were prospectively documented using a standardized clinical grading system, as previously described [25, 26].

All procedures were initiated under spinal anesthesia until fetal delivery, after which general anesthesia was administered for the remainder of the operation. A vertical midline abdominal incision was routinely performed, and the uterus together with the adnexa was exteriorized to allow optimal surgical exposure.

The choice of surgical approach was determined intraoperatively based on topographic and clinical findings, rather than on histopathological classification. Key factors guiding surgical decision-making included placental extension toward or into the cervix, involvement of the parametrial tissues, the degree and distribution of abnormal hypervascularization, and the presence of dense adhesions at the utero-vesical interface.

In cases deemed suitable for conservative management, uterus-sparing surgical techniques were employed. These included segmental uterine resection and repair, placement of interrupted transverse sutures at the placental bed, uterine artery ligation, and/or insertion of an intrauterine Bakri balloon (CooperSurgical Inc., Trumbull, CT, USA), either alone or in combination.

In uterus-sparing procedures, the fetus was delivered through a transverse incision in the unaffected upper uterine segment, avoiding areas of suspected placental invasion. The placenta was subsequently managed with controlled cord traction, followed by application of conservative techniques or localized resection when limited invasion was identified.

In cases requiring cesarean hysterectomy, fetal delivery was achieved via a vertical fundal uterine incision, after which the umbilical cord was clamped and transected, leaving the placenta in situ. The cesarean hysterectomy was then completed without attempting placental separation.

Statistical analysis

All statistical analyses were performed using SPSS version 29.0 (IBM Corp., Armonk, NY, USA). Continuous variables were assessed for normality using the Shapiro–Wilk test. Normally distributed variables were presented as mean ± standard deviation (SD), while non-normally distributed variables were reported as median with interquartile range (IQR; 25th–75th percentile). Categorical variables were expressed as frequencies and percentages. Comparisons between surgical management groups (uterus-sparing surgery vs. cesarean hysterectomy) were performed using the independent samples t-test or Mann–Whitney U test for continuous variables, and the Chi-square or Fisher’s exact test for categorical variables, as appropriate.

To investigate predictors of transfusion burden, linear regression analysis was conducted using the total number of transfused blood units as the dependent variable. Given the strong clinical influence of baseline anemia and surgical approach, preoperative hemoglobin level and cesarean hysterectomy status were included as covariates in the initial model. Residuals from this model were subsequently extracted and used as the dependent variable to evaluate the independent contribution of individual ultrasonographic markers. Ultrasonographic variables with a p-value < 0.05 in univariate analyses were entered into a multivariable regression model. Additionally, estimated marginal means of total transfused blood units were calculated to illustrate adjusted transfusion requirements according to preoperative hemoglobin levels and surgical management, facilitating clinical interpretation of the regression models. All tests were two-sided, and a p-value < 0.05 was considered statistically significant.

Results

Maternal age and body mass index did not differ significantly. However, patients who underwent cesarean hysterectomy had a higher median gravidity, parity, and number of previous cesarean sections compared with those managed with uterus-sparing techniques. Preoperative and postoperative hemoglobin levels were similar between groups. Despite this, the total number of transfused blood units was significantly higher in the cesarean hysterectomy group compared with the uterus-sparing group (mean ± SD: 2.7 ± 1.7 vs. 1.4 ± 1.3 units, p = 0.003). Neonatal outcomes, including gestational age at delivery, birthweight, and Apgar scores, did not differ significantly between groups (Table 1).

Table 1.

Comparison of patients with uterus spairing surgery and Cesarean-Hysterectomy in terms of demographic, surgical, maternal and neonatal outcomes

	Uterus Spairing Surgery (n = 25)	Cesarean-Hysterectomy (n = 36)	P value
Demographics, mean ± SD/median (25,75 centile)
Maternal age in years	32.9 ± 5.19	32.4 ± 4.52	0.705
Gravida	4.0 (3.0, 4.0)	5.0 (3.0, 6.0)	0.015
Parity	2.0 (1.0, 3.0)	3.0 (2.0, 3.2)	0.008
Previous Caesarean section	1.0 (1.0, 2.0)	2.0 (2.0, 3.0)	0.005
Maternal body mass index at booking in Kg/m2	28.5 ± 4.40	29.5 ± 3.63	0.316
Surgery, mean ± SD
Preoperative Hemoglobin level	11.4 ± 1.18	11.2 ± 1.46	0.650
Postoperative Hemoglobin level	10.0 ± 1.43	9.9 ± 1.28	0.807
Total blood transfusion	1.4 ± 1.33	2.7 ± 1.70	0.003
Postoperative Neonatal and Maternal Findings, mean ± SD/median (25,75 centile)
Gestational week at birth	35.0 (34.0, 35.0)	34.0 (34.0, 35.0)	0.297
Birthweight in grams	2335 ± 455	2329 ± 496	0.967
APGAR score at 1th min.	7.0 (7.0, 8.0)	7.0 (7.0, 8.0)	0.217
APGAR score at 5th min.	9.0 (8.0, 10.0)	9.0 (8.0, 9.0)	0.616

Data are presented as mean ± standard deviation (SD) for normally distributed variables, and as median (25th–75th percentile) for non-normally distributed variables. Categorical variables are shown as number and percentage (n, %). P values were calculated using independent samples t-test, Mann–Whitney U test, or Chi-square test as appropriate. Values in bold indicate statistical significance (p < 0.05)

Loss of the clear zone and bridging vessels were almost universally observed in patients undergoing cesarean hysterectomy. Utero-vesical hypervascularity and placental vessels extending toward or into the cervix were significantly more frequent in the hysterectomy group, while placental lacunae were also more commonly detected (Table 2).

Table 2.

Comparison of ultrasonographic findings between uterus spairing surgery and Cesarean-Hysterectomy cases in placenta accreta spectrum (PAS) disorders

Variables, n (%)	Uterus Spairing Surgery (n = 25)	Cesarean-Hysterectomy (n = 36)	P value
Loss of clear zone	18/25 (72.0)	36/36 (100.0)	< 0.001
Utero-Vesical Hypervascularity	12/25 (48.0)	31/36 (86.1)	0.001
Placental vessels extending into the cervix	4/25 (16.0)	19/36 (58.0)	0.004
Presence of lacunae	5/20 (20.0)	23/36 (63.9)	0.012
Bridging vessels	14/25 (56.0)	33/36 (91.7)	0.001
Retroplacental myometrium thickness < 1 mm	1/22 (4.5)	21/34 (61.8)	< 0.001

Data are presented as n (%) for categorical variables and median (25th–75th percentile) for continuous variables. P-values calculated using Chi-square test or Mann–Whitney U test as appropriate. Values in bold indicate statistical significance (p < 0.05)

To account for the strong baseline effects of surgical approach and anemia, total transfusion requirement was first adjusted for cesarean hysterectomy status and preoperative hemoglobin level. Residuals from this model were subsequently analyzed to evaluate the independent contribution of ultrasonographic markers (Table 3). In univariate analyses, utero-vesical hypervascularity and placental vessels extending into the cervix were significantly associated with increased transfusion requirement. In the multivariable model, both markers remained independently associated with transfusion burden after adjustment (β = 0.565, p = 0.034 and β = 0.640, p = 0.011, respectively).

Table 3.

Association between prenatal ultrasound markers and transfusion requirement after adjustment for hysterectomy and preoperative hemoglobin levels in placenta accreta spectrum

	Univariete Lineer Regression			Multivariete Lineer Regression
	OR	%95 CI	P value	OR	%95 CI	P value
Previous Caesarean section count	0.038	(−0.222) – (0.299)	0.767
Loss of clear zone	−0.449	(−1.250) – (0.354)	0.268
Utero-Vesical Hypervascularity	0.670	(0.060) – (1.150)	0.030	0.565	(0.044) – (1.080)	0.034
Placental vessels extending into the cervix	0.671	(0.168) – (1.170)	0.010	0.640	(0.151) – (1.130)	0.011
Presence of lacunae	0.098	(−0.419) – (0.617)	0.704
Retroplacental myometrium thickness < 1 mm	0.194	(−0.357) – (0.745)	0.484

Linear regression analysis was used to identify ultrasound markers associated with transfusion requirement in placenta accreta spectrum. In the first step, total number of transfused blood units was regressed on hysterectomy status and preoperative hemoglobin level to account for their strong baseline effects. Model residuals were subsequently extracted and used as the dependent variable in univariate linear regression analyses assessing individual ultrasound markers. Ultrasound variables showing statistical significance in univariate analysis were entered into a multivariable linear regression model. Regression coefficients are presented with 95% confidence intervals. Values in bold indicate statistical significance (p < 0.05)

Estimated marginal means analysis further illustrated the combined impact of baseline hemoglobin levels and surgical management on transfusion requirement (Table 4). Across all hemoglobin strata, cesarean hysterectomy was consistently associated with a higher expected number of transfused units compared with uterus-sparing surgery. While higher preoperative hemoglobin levels were associated with reduced transfusion needs overall, the difference between surgical approaches persisted even at higher hemoglobin levels.

Table 4.

Estimated marginal means of total transfused blood units according to preoperative hemoglobin levels and Cesarean hysterectomy status in placenta accreta spectrum

				95% Confidence Interval
Preoperative Hb Level	Cesarean-Hysterectomy	Marginal Mean	SE	Lower	Upper
9.93⁻	0	1.92	0.369	1.180	2.66
	1	3.11	0.314	2.483	3.74
11.27µ	0	1.51	0.304	0.901	2.12
	1	2.70	0.253	2.196	3.21
12.61⁺	0	1.10	0.354	0.391	1.81
	1	2.29	0.326	1.640	2.95

⁻ mean − 1SD, ^µ mean, ⁺ mean + 1SD

Discussion

In this study, we evaluated the relationship between prenatal ultrasonographic findings and surgical complexity, transfusion burden, and the need for cesarean hysterectomy in women with placenta accreta spectrum (PAS), using a topographic and clinically based approach rather than histopathological subclassification. Our findings support the growing literature arguing that the primary aim of prenatal imaging in PAS should be to identify features that can predict surgical difficulty and hemorrhagic risk, rather than to differentiate histological subtypes. For many years,

PAS was considered a spectrum defined by the depth of villus invasion penetration into the myometrium. However, numerous studies published in recent years have shown that histological invasion depth is not reliably associated with surgical difficulty, blood loss, or maternal morbidity. Instead, it is increasingly understood that PAS is a pathology resulting from uterine scar remodeling and loss of the normal placental separation plan, rather than true trophoblast invasion [27, 28].

Jauniaux et al. demonstrated that in most cases previously classified as “percreta,” the villus tissue did not penetrate the entire uterine wall and did not invade adjacent organs. The underlying mechanism of abnormal placental adhesion has been shown to be dense fibrinoid deposition at the utero-placental interface and disruption of the Nitabuch membrane [9, 16]. These findings support the idea that the distinction between accreta and percreta has limited clinical value and that a topographic and vascular criteria-based approach is more appropriate for surgical planning.

According to this conceptual change, our study showed that vascular abnormalities and cervical involvement were more strongly linked to surgical difficulty and the need for transfusions than traditional ultrasound findings that indicated myometrial invasion. Even after controlling for preoperative hemoglobin level and surgical technique, uterovesical hypervascularization and placental vessel extension toward or into the cervix were found to be independent predictors of transfusion burden. These results are in line with earlier research showing a strong correlation between intraoperative bleeding and surgical difficulty and aberrant vascular patterns in the lower uterine segment and cervico-isthmic region [19, 29].

In recent years, there has been a special emphasis on the significance of transvaginal ultrasonography in identifying these high-risk vascular features. Transvaginal evaluation during the third trimester is crucial for predicting complicated surgery and significant blood loss, as Jauniaux et al. showed [30]. Similarly, Aryananda et al. reported that intracervical hypervascularity is strongly associated with the need for radical surgery and increased maternal morbidity [19].

Additionally, our research shows a noteworthy difference between prognostic and diagnostic ultrasound results in PAS. Although placental lacunae and clear zone loss are frequently used to diagnose PAS, corrected analyses have not shown that these symptoms independently predict transfusion burden. This suggests that these findings reflect abnormal placental adhesion but do not provide sufficient information about the degree of vascular recruitment or surgical risk [22].

In contrast, vascular findings, particularly those involving the uterovesical junction and cervical region, better reflect the hemodynamic consequences of scar dehiscence and abnormal placental implantation, and ultimately show a stronger association with hemorrhagic risk. Numerous factors, such as baseline maternal hemoglobin level, surgical technique, and center experience, influence the overall blood transfusion requirement in PAS. Higher transfusion needs at all hemoglobin levels were also linked to cesarean hysterectomy in our cohort; this finding is in line with earlier research [11, 20].

However, even after adjusting for these strong determinants, significant variability in transfusion requirements persisted. The fact that some of this residual variation could be explained by ultrasonographic vascular markers suggests that prenatal imaging may provide additional prognostic information to clinical variables. Estimated marginal mean analyses show that the surgical approach modifies the transfusion risk at each hemoglobin level, highlighting the importance of accurate preoperative risk stratification.

Taken together, all these findings support the idea that prenatal ultrasonography in PAS should focus on findings that can predict surgical difficulty, hemorrhagic risk, and the need for radical surgery, rather than attempting to differentiate histological subtypes. This approach is consistent with current recommendations advocating for individualized surgical planning based on placental topography, vascular anatomy, and center resources [31].

Our study’s small sample size and retrospective design are significant limitations. Additionally, a certain amount of heterogeneity may have resulted from modifications to surgical techniques and ultrasound criteria during the lengthy study period. However, the validity of our results is strengthened by the use of clinically significant correction strategies, standardized intraoperative assessment, and imaging carried out by a single, experienced ultrasonographer.

Conclusion

Prenatal ultrasound vascular markers, particularly cervical and utero-vesical hypervascularization, are independently associated with transfusion burden in PAS. A prognostic imaging approach focusing on placental topography and abnormal vascular recruitment may improve surgical readiness and transfusion anticipation, even in women with normal or high preoperative hemoglobin levels.

Abbreviations

PAS

Placenta Accreta Spectrum

MRI

Magnetic Resonance Imaging

ACOG

American College of Obstetricians and Gynecologists

BMI

Body Mass Index

OR

Odds Ratio

AUC

Area Under the Curve

CI

Confidence Interval

SE

Standard Error

Authors’ contributions

A.KINDAN: Manuscript writing. C.O.U: Data analysis. A.KURT: Manuscript editing, Methodology. M.L.D: Data curation, Validation. S.S: Data analysis, Resources. A.G.Y: Data curation, Resources. K.Y.Y: Manuscript editing, Methodology. Ş.Ç: Investigation, Conceptualization. A.T.Ç: Protocol development, Conceptualization. Y.E.Ü: Protocol management, Supervision.

Funding

This research received no specific grant from any funding agency.

Data availability

Due to hospital policies, patient data and study materials cannot be shared. However, the data are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Institutional Ethics Committee of Etlik Zübeyde Hanım Women’s Health Training and Research Hospital (Protocol ID: 30/10/2024-11) and written informed consent was obtained from all participants. The veri were anonymized and handled in compliance with ethical guidelines to ensure participant confidentiality.

Consent for publication

Not applicable. This study does not contain any individual person’s veri in any form (including any individual details, images, or videos) that would require consent for publication.

Competing interests

The authors declare no competing interests.