Predicting the Risk of Cesarean Hysterectomy in the Management of Placenta Accreta Spectrum Disorders: A New Model Based on Clinical Findings and Ultrasonography

doi:10.21203/rs.3.rs-4800562/v1

Purpose

To develop a nomogram that can predict the probability of cesarean hysterectomy in patients for operation due to placenta accreta spectrum disorders(PASD).

Methods

Data from 520 patients who underwent surgery with a preliminary diagnosis of PASD at a tertiary center in eastern Turkey between 2013 and 2023 were recorded, and 302 were included in the study. A prediction model was developed using Penalized Maximum Likelihood Estimation (PMLE) regression analysis with clinical evaluation and ultrasonography findings as variables.

Results

The multivariate PMLE regression model included demographic data such as advanced age (aOR: 1.22, 95% CI: 1.08–1.44, p = 0.001), prior uterine surgeries (aOR: 3.18, 95% CI: 1.57–8.29, p = 0.001) were associated with an increased likelihood of cesarean hysterectomy. Regarding ultrasonographic findings, the multiple lacunaes (aOR: 48.53, 95% CI: 18.42–257.40, p < 0.001) and irregularity in the interface between the bladder and uterine serosa or signs of hypervascularization with Doppler Flow (aOR: 7.90, 95% CI: 2.66–35.12, p < 0.001) significantly increased the probability of hysterectomy. A retro-placental myometrial thickness of < 1mm showed borderline significance (aOR: 2.49, 95% CI: 0.89–8.27, p = 0.08). An anterior placental location (aOR: 9.60, 95% CI: 2.96–50.76, p < 0.001) increased the probability of hysterectomy, whereas advanced gestational age showed a negative correlation (OR: 0.78, 95% CI: 0.56–1.02, p = 0.07). The model's performance, evaluated using Harrell’s C-Index, was 0.974, and the Kappa Value was 0.819.

Conclusion

A nomogram predicting the probability of cesarean hysterectomy in the management of PASD patients has been developed, incorporating advanced age, gestational age, number of uterine surgeries, ultrasound findings, and placental location.

placenta accreta spectrum disorders

ultrasonography

cesarean hysterectomy

prediction model

The study presents a predictive model that incorporates clinical and ultrasonographic findings to enhance preoperative planning in the management of PASD. By emphasizing the importance of predictors such as advanced maternal age, prior uterine surgeries, and specific ultrasonographic findings, this study assists clinicians in making informed decisions regarding the type of surgical intervention before performing PASD surgeries. Additionally, the ability to predict the need for cesarean hysterectomy can reduce emergency interventions and associated complications, potentially decreasing maternal morbidity and mortality through better-managed, planned surgeries.

Placenta Accreta Spectrum Disorders (PASD) are increasingly common on a global scale and affect maternal morbidity and mortality significantly. For this reason, accurate diagnosis before and during birth is critical to distinguish adherent and invasive forms [1]. PASD describes an abnormal invasion of trophoblastic tissue extending beyond the decidua basalis layer into the uterine myometrium, uterine serosa, and beyond into the pelvic organs [2]. PASD is divided into three main categories (‘placenta accreta’, in which chorionic villi adhere to the myometrial surface but do not invade; ‘placenta increta,’ in which they invade the myometrium; and ‘placenta percreta,’ which is the most severe form in which it passes through the complete layer of the myometrium and spreads to the serosa and adjacent pelvic organs [3–4]. Increasing cesarean section rates also increase the risk of developing PASD (5). For PASD, aside from previous cesarean sections, ‘placenta previa’ also comes to the forefront as an essential risk factor [5–6]. PASD is considered one of the main indications for postpartum hysterectomy [7]. Although the method of birth used in these cases is a cesarean section, the procedure used and recommended in many cases [8] is a cesarean hysterectomy [9]. Postpartum hysterectomy performed under emergency conditions is riskier with high complication rates when compared to elective ones [10]. Therefore, the opinion has emerged that maternal mortality and morbidity rates can be decreased in this group by creating a risk assessment tool to determine the types of surgery in follow-ups starting from the early stages of pregnancy [11]. Despite the advances in imaging techniques and prenatal care, predicting the surgical intervention required for PASD remains a clinical challenge. Ultrasonography (USG), which is used frequently in the prenatal period (especially in the second and third trimesters), plays an essential role in determining the topography and depth of invasion of the placenta [12].

Nomograms are graphical representations of statistical prediction models used to simplify complex calculations by employing some variables. They are developed to calculate the probability of a particular outcome for individual patients [13]. These models, which can be employed in the clinical field, help clinicians make risk assessments and treatment or surgical intervention decisions.

The present study aimed to develop a prediction model by performing regression analysis with various clinical and ultrasonographic findings to facilitate preoperative planning and determine the most accurate type of surgery for PASD patients.

Data Collection

A total of 520 patients who underwent surgery with a preliminary diagnosis of PASD at Gaziantep University Faculty of Medicine, Department of Gynecology and Obstetrics between 2013 and 2023 were evaluated, and 302 suitable patients were included. All patients included in the present study were assessed regarding placenta previa totalis, and patients who did not meet this criterion were excluded. Figure 1 summarizes the study's inclusion and exclusion criteria. In the study, 195 patients underwent cesarean hysterectomy, and 107 patients underwent conservative surgery. Those who underwent hysterectomy because of bleeding during conservative surgery and those who underwent emergent surgery were not included in the study. Before the study, approval was received from the Gaziantep University Faculty of Medicine Ethics Committee with Decision No: 2023/44 on 01.02.2023. The cases for whom conservative approaches such as placental bed suturing, uterine or hypogastric artery ligation, anterior segmental resection, and uterine balloon tamponade were applied because of PASD, and those for whom postpartum hysterectomy was performed were examined through the Hospital Information Management System (HIMS) retrospectively. All patients included in the study were informed about the risk of hysterectomy, and written consent was obtained.

Investigated Clinical Characteristics

The maternal age, gravida, parity, abortion, prior uterine surgeries (cesarean section and myomectomy history), weeks of pregnancy, ultrasonographic findings, pre-and post-operative hemoglobin values, perioperative blood product replacement, surgery performed, complications developing during the surgery, follow-up in the postoperative period and hospitalization times and postoperative pathology results of the patients were recorded.

The findings given below, which increased the risk of PASD [14–15] in the second and third trimester USG and Doppler Examinations during the antenatal period, were included in the study.

USG 1: Lack or absence of normal hypoechoic retroplacental zone

USG 2: Multiple lacunaes in the placenta

USG 3: Irregularity in the line between the bladder and uterine serosa and findings of hypervascularization with Doppler Flow

USG 4: Myometrium thickness in the placenta location < 1 mm

These findings were included in the modeling developed by the researchers as parameters used in the preliminary diagnosis of PAS disorders in the USG reports of the patients in the present study.

Management of Patients

In the our clinic, patients with a pre-diagnosis of PASD are treated at 34–36 months, in line with the literature data. Birth is planned during the gestational age, and antenatal corticosteroids are administered routinely before the surgery. After the 34th week of pregnancy, patients are generally hospitalized and monitored or followed up with frequent checks (every three days) until the surgery is planned. For patients who are followed up, blood bank stocks are checked daily for emergencies, a bed is always available for the risky pregnancy ward, Fibrinogen (1 g) in a vial is made ready to be used in the surgery, and Bakri Baloon® is available in the obstetrics and gynecology operating room for uterine balloon tamponade. For cases with suspected PASD before the surgery, five units of Erythrocyte Suspension (ES) and Fresh Frozen Plasma (FFP) are ready to be used in emergency conditions at the blood center.

Patients pre-diagnosed with PASD but have severe vaginal bleeding before the 36th week of pregnancy or are referred from external centers because of vaginal bleeding undergo surgery under emergency conditions. However, PASD patients who underwent surgery under emergency conditions were excluded from the study so as not to disrupt the data standardization and the Nomogram Model quality.

The cesarean section hysterectomy or conservative approach to be applied to patients is performed by considering the recommendations of published guidelines in our clinic [16–17]. Severe placental invasion to neighboring organs and increased vascularization are suggested in selected cases in which percreta findings in the visceral peritoneum and parametrium are involved. A direct hysterectomy is performed following the delivery of the fetus. If the macroscopic degree of invasion is high in these cases, hypogastric artery ligation can be applied to reduce bleeding before cesarean hysterectomy or when bleeding continues afterward. However, perioperative hypogastric artery ligation is not performed routinely in our clinic. Hypogastric Artery Ligation is used for cases if bleeding continues after a hysterectomy [18].

In our clinic, conservative surgery is performed according to the conservative approach in the literature [19]. In general, following the delivery of the fetus, hemostasis is achieved with clamps, the bladder is released from the front of the uterus by sharp scissors or blunt dissection with fingers [20], and then, if needed, anterior segmental resection or posterior bed suturing is performed.

Data analysis

The categorical data were presented as frequency (n) and percentage (%). The Chi-Square Test was used to evaluate the relationship between categorical variables. The Shapiro-Wilk Test was used to assess the distribution properties of numerical variables. Median and Interquartile Range (IQR: 25%-75% range) were reported for numerical data that did not show a normal distribution, and the Mann-Whitney U Test was used to compare two independent groups. A p-value of < 0.05 was considered significant for the statistical significance of the difference between groups. Model-building processes are described in detail in the Statistical Modeling section.

The sample size was kept sufficiently large in the study, and the number of explanatory variables used in developing the clinical model was chosen conservatively. Each explanatory variable (degree of freedom, DF) used in the model included at least ten events (patients) to provide reliable results and increase the generalizability of the model by preventing overfitting (Outcome/DF > 10) [21]. The present study included 302 patients (197 underwent hysterectomy). Eight candidate explanatory variables were used in the model (df = 8). In this context, a sufficient number of events was provided for each explanatory variable of the model.

Statistical Modeling

The relationship of each variable with cesarean hysterectomy was evaluated in univariate analyses, multivariate analyses were used, and the variables that were significant or close to significance as a result of the analyses were selected for modeling. This study examined the relationship between the result and candidate explanatory variables using the Penalized Maximum Probability Estimation (PMLE) method. Instead of traditional logistic regression, PMLE maximizes the penalized log probability and adjusts the maximum log probability of the model by adding a penalty factor. Standard logistic regression may reduce the model's predictive power and increase variance in small sample sizes or datasets with many variables. PMLE solves this problem by adding a penalizing term and making the model more stable. Standard logistic regression is more likely to overfit with lower prediction accuracy on new observations. For this reason, the PMLE method gives better results in studies with fewer patients and when there is no trial and validation dataset [21].

According to the results of the regression analysis used with the PMLE Method, a nomogram was developed, including some variables (e.g., age, number of cesarean sections, and previous uterine surgeries (cesarean section, myomectomy), USG findings (USG 1, USG 2, USG 3, USG 4) and anterior localized placenta). Statistical tests and validation methods were used to evaluate the internal performance and fitness of the model. Harrell’s C-Index measured the discriminative ability of the model, and the Kappa Value evaluated the classification accuracy of the model. The McNemar Test was used to compare the false positive and negative rates. The Hosmer-Lemeshow Test evaluated the fit of the model. Also, the 10-fold cross-validation method was used to measure the generalization and performance of the model more reliably.

The R Software version 4.4.1 ( http://www.r-project.org ) was used to perform all statistical analyses and create the charts.

The demographic and clinical characteristics of the patients who underwent cesarean hysterectomy or conservative surgery because of PASD were examined in detail in the present study. As a result of the multivariate analyses, the risk factors of these patients were identified, and important findings were obtained in light of these factors. Clinical parameters and demographic data of the patients who underwent surgery for PASD are summarized in Table 1, which includes various demographic and clinical characteristics of the patients, such as age, gestational age, previous uterine surgeries, placental localization, and USG findings.

According to the multivariate PMLE regression analysis, independent variables associated with cesarean hysterectomy were identified. Variables such as USG findings and anterior localization of the placenta were associated with cesarean hysterectomy significantly. In particular, the USG 2 finding was among the strongest predictors among USG findings, and the USG 3 finding was very significant. However, the findings of USG 1 and USG 4 showed a less significant relationship. Also, advanced gestational age (aOR = 0.78, 95% CI: 0.56–1.02, p = 0.07) decreased the risk of cesarean hysterectomy. The findings are summarized in Table 2.

The present study enabled the development of a nomogram that included variables such as age, gestational age, and ultrasonography findings and can predict the possibility of cesarean hysterectomy (Fig. 2). This nomogram makes it possible to predict the risk of cesarean hysterectomy more accurately and supports surgical decision-making in clinical practice. Figure 3 is the importance chart showing the importance levels of the variables in the nomogram. It shows which variables have more weight in the model. This chart also shows the relative contributions and order of importance of the variables in the model, emphasizing which factors clinicians should pay more attention to. As a result of the analyses, the USG 2 finding was effective when deciding on a hysterectomy.

Model Performance and Internal Validation

The 10-fold cross-validation method was used to evaluate the predictive accuracy and generalizability of the model. According to the results, the Root Mean Square Error (RMSE) of the model was 0.2660, the R-Square (Rsquared) value was 0.6690 and the Mean Absolute Error (MAE) value was 0.1349. These results showed that the magnitude of the model’s prediction errors was, on average, 0.266, explaining 66.9% of the total variation in the dependent variable, and the average magnitude of the prediction errors was 0.1349. These findings showed that the model's generalization ability was good, and its predictions were close to the actual values. The discrimination of the model was measured with Harrell’s C-Index, and the correct classification performance of the model was found to be 0.974, which showed that the model had very good discrimination. The accuracy of the model was evaluated with the Kappa Value, which measures the fit between the predicted classes and the actual classes, and the Kappa Value was calculated as 0.819, indicating a good fit. Also, the McNemar Test was used to test the agreement of the model’s predictions and actual values. The results (χ2 = 0, p-value = 1) showed no significant differences between the classes predicted by the model and the actual classes. The overall model fit was evaluated with the Hosmer-Lemeshow Test (χ2 = 4.7828, Degrees of Freedom = 8, p-value = 0.7805) and confirmed that the model provided a good fit to the data.

The Log-Loss Value of the model was 0.2001, which indicated how close the predicted probabilities were to the actual classes. A low Log-Loss value indicates that the model’s probability estimates are highly accurate.

These results showed that our logistic regression model was highly effective in predicting the probability of hysterectomy. Internal validation of the model was performed using 1000-Bootstrap Replication. The calibration chart shows the consistency between the model’s observed and predicted results. The model’s calibration chart (Fig. 4) shows how well the predictions match the actual results and how reliable the model is.

The demographic and clinical data of patients underwent cesarean hysterectomy or conservative surgery for PASD were analyzed to identify risk factors. A nomogram incorporating age, gestational age, and ultrasonography findings was developed to predict the likelihood of cesarean hysterectomy.

Studies report that advanced maternal age is associated with PASD and increases the need for cesarean hysterectomy [22–23]. Similarly, in multivariate analyses in the present study, advanced maternal age was found to be a factor in increasing the probability of hysterectomy (aOR = 1.22, 95% CI: 01.08–1.44, p = 0.001).

Conservative surgery or cesarean hysterectomy is recommended to be planned between 34 and 36 weeks in patients over 28 weeks [24]. Still, the timing of the surgery might vary depending on the patients' maternal and fetal indications. The patient group in our modeling was between 27 and 38 weeks of gestation and underwent elective surgery. Early gestational age was a finding that relatively increased the possibility of cesarean hysterectomy in the patients included in the present study (p = 007).

An increase in the number of previous cesarean sections [25] and a history of myomectomy [26] are important risk factors for PASD requiring peripartum hysterectomy [27]. In parallel with these findings being frequently reported in the literature data, patients with a history of cesarean section and myomectomy were included as individuals who had previous uterine surgery. In the present study, an increased number of prior uterine surgeries was evaluated as an essential parameter that increased the possibility of cesarean hysterectomy (aOR = 3.18, 95% CI: 1.57–8.29, p = 0.001).

USG is a standard imaging method used to determine placental localization and topography in pregnancy follow-up, which can be applied with high accuracy and specificity in diagnosing PASD in experienced centers [28]. In the present study, there were four USG findings associated with cesarean hysterectomy.It was shown that lacunar areas in the placenta increased invasion and adhesion in PASD patients [15]. Increasing the depth of invasion also increased the possibility of cesarean hysterectomy [23]. The present study showed that the presence of lacunar areas in the placenta (USG 2), which is one of the findings that can be detected in ultrasonography, is among the strongest predictors for cesarean hysterectomy (aOR = 48.53, 95% CI: 18.42–257.40, p < 0.001). This shows that the presence of lacunar areas is an important determinant in diagnosing PASD and predicting the risk of cesarean hysterectomy.Also, irregularity in the line between the bladder and uterine serosa and hypervascularization with Doppler Flow (USG 3) is an important finding that increases placental adhesion [14]. It was also found in the present study that these irregularities and hypervascularization, which are indicators of PASD, increased the possibility of cesarean hysterectomy (aOR = 7.90, 95% CI: 2.66–35.12, p < 0.001).

The presence or absence of hypoechoicity in the retroplacental field (USG 1) (p = 0.03) and the myometrium thickness in the placenta location field is less than 1 mm (USG 4), which showed a statistically significant difference in demographic data (p < 0.001) in the patients included in the present study who underwent cesarean hysterectomy and conservative surgery were reported to increase placental adhesion in previous studies [14]. These two findings that increase adhesion were not considered significant determinants in our PMLE regression analysis. Although they were found to be significant in univariate analyses, the fact that these variables were not significant in regression analysis shows that they need to be more sufficient as independent predictors in predicting the risk of hysterectomy.Wright et al. showed that the anterior location of the placenta on USG was strongly associated with placenta percreta [29]. One of the variables included in the present study (having the placenta in an anterior location) increases the probability of cesarean hysterectomy significantly when compared to having the placenta in a posterior location (aOR = 9.60, 95% CI: 2.96–50.76, p < 0.001).

Nomograms are important and effective tools for predicting disease risks in clinical practice. Many studies have been conducted to predict PASD and evaluate its clinical outcomes. In a previous study, a nomogram was developed to emphasize that factors such as the number of prior cesarean deliveries, vaginal bleeding during pregnancy, and placenta previa were important in determining the risk of PASD [30]. Also, Chen et al. developed a nomogram to be used in clinical practice to predict the amount of postpartum bleeding in women undergoing uterine surgery [31]. What makes the present study different from these previous studies is that the focus of the nomogram developed was created to predict cesarean hysterectomy. The present study is the first in the literature to develop a nomogram to predict the possibility of cesarean hysterectomy and is the study conducted with the largest patient group.

The study was conducted retrospectively, and the data were collected from a single center. The model was created using a single dataset in this study, and a separate set was not used for validation, which caused a risk of overfitting and made it difficult to fully evaluate the model's generalization ability. In the management of PASD, the most widely accepted approach is cesarean hysterectomy, and conservative surgery must be applied to an appropriately selected patient group. This nomogram emphasizes the importance of choosing the right surgery and aims to make it easier to decide on the most appropriate type of surgery for the patient by predicting signs of invasion before the surgery. We believe that making a preoperative decision on conservative surgery or cesarean hysterectomy in PASD patients will play essential roles in decreasing intrapartum and postpartum complications. This study will also shed light on future studies and contribute to the literature in this field.

A nomogram that can predict the probability of cesarean hysterectomy in managing PASD patients has been developed. The nomogram includes advanced age, gestational age, prior uterine surgeries, ultrasound findings, and placental location. Validation tests showed that the nomogram yielded good results in predicting cesarean hysterectomy and could be used in clinical practice.

Acknowledgments:

The authors have no acknowledgments to declare for this study.

Disclosure:

The authors declare no Conflict of Interest for this article.

Funding:

Our study did not receive any funding

Morlando M, Collins S. (2020) Placenta Accreta Spectrum Disorders: Challenges, Risks, and Management Strategies. Int J Womens Health. Nov 10;12:1033-1045. https://doi.org/10.2147/ijwh.s224191
Piñas Carrillo A, Chandraharan E. (2019) Placenta accreta spectrum: Risk factors, diagnosis and management with special reference to the Triple P procedure. Womens Health (Lond). Jan-Dec;15:1745506519878081. https://doi.org/10.1177/1745506519878081
Jauniaux E, Chantraine F, Silver RM et al. (2018) FIGO Placenta Accreta Diagnosis and Management Expert Consensus Panel. FIGO consensus guidelines on placenta accreta spectrum disorders: Epidemiology. Int J Gynaecol Obstet;140:265–273. https://doi.org/10.1002/ijgo.12407
Parra-Herran C, Djordjevic B (2016) Histopathology of placenta creta: Chorionic villi intrusion into myometrial vascular spaces and extravillous trophoblast proliferation are frequent and specific findings with implications on diagnosis and pathogenesis. Int J Gynecol Pathol. 35:497–508. https://doi.org/10.1097/pgp.0000000000000250
Bowman ZS, Eller AG, Bardsley TR et al (2014) Risk factors for placenta accreta: a large prospective cohort. Am J Perinatol. Oct;31(9):799-804. https://doi.org/10.1055/s-0033-1361833
Pegu B, Thiagaraju C, Nayak D et al (2021) Placenta accreta spectrum-a catastrophic situation in obstetrics. Obstet Gynecol Sci. May;64(3):239-247. https://doi.org/10.5468/ogs.20345
Garmi G, Salim R (2012) Epidemiology, etiology, diagnosis, and management of placenta accreta. Obstet Gynecol Int. 2012;2012:873929.Epub May 7. https://doi.org/10.1155/2012/873929
American College of Obstetricians and Gynecologists; Society for Maternal-Fetal Medicine (2018) Obstetric Care Consensus No. 7: Placenta Accreta Spectrum. Obstet Gynecol. Dec;132(6):e259-e275. https://doi.org/10.1097/aog.0000000000002983
Einerson BD, Gilner JB, Zuckerwise LC (2023) Placenta Accreta Spectrum. Obstet Gynecol. Jul 1;142(1):31-50. https://doi.org/10.1097/aog.0000000000005229
Flores-Mendoza H, Chandran AR, et al (2022) Outcomes in emergency versus electively scheduled cases of placenta accreta spectrum disorder managed by cesarean-hysterectomy within a multidisciplinary care team. Int J Gynaecol Obstet. Nov;159(2):404-411. https://doi.org/10.1002/ijgo.14121
Yang T, Li N, Qiao C et al (2019) Development of a Novel Nomogram for Predicting Placenta Accreta in Patients With Scarred Uterus: A Retrospective Cohort Study. Front Med (Lausanne). Dec 17;6:289. https://doi.org/10.3389/fmed.2019.00289
Yu FNY, Leung KY (2021) Antenatal diagnosis of placenta accreta spectrum (PAS) disorders. Best Pract Res Clin Obstet Gynaecol. Apr;72:13-24. https://doi.org/10.1016/j.bpobgyn.2020.06.010
Steyerberg, E (2009) W. Clinical Prediction Models: A Practical Approach to Development, Validation, and Updating, 1st edn. New York, NY: Springer. 317–319. http://dx.doi.org/10.1007/978-0-387-77244-8
Silver RM, Fox KA, Barton JR et al (2015) Center of excellence for placenta accreta. Am J Obstet Gynecol. May;212(5):561-8. https://doi.org/10.1016/j.ajog.2014.11.018
Zhang J, Li H, Feng D et al (2023) Ultrasound scoring system for prenatal diagnosis of placenta accreta spectrum. BMC Pregnancy Childbirth. Aug 7;23(1):569. https://doi.org/10.1186/s12884-023-05886-x
Sentilhes L, Kayem G, Chandraharan E et al (2018) FIGO Placenta Accreta Diagnosis and Management Expert Consensus Panel. FIGO consensus guidelines on placenta accreta spectrum disorders: Conservative management. Int J Gynaecol Obstet. Mar;140(3):291-298. https://doi.org/10.1002/ijgo.12410
Allen L, Jauniaux E, Hobson S et al (2018) FIGO Placenta Accreta Diagnosis and Management Expert Consensus Panel. FIGO consensus guidelines on placenta accreta spectrum disorders: Nonconservative surgical management. Int J Gynaecol Obstet. 2018 Mar;140(3):281-290. https://doi.org/10.1002/ijgo.12409
Sucu S, Özcan HÇ, Karuserci ÖK et al (2021). Is there a role of prophylactic bilateral internal iliac artery ligation on reducing the bleeding during cesarean hysterectomy in patients with placenta percreta? A retrospective cohort study. Ginekol Pol. 92(2):137-142. https://doi.org/10.5603/gp.a2020.0145
Shabana A, Fawzy M, Refaie W (2015). Conservative management of placenta percreta: a stepwise approach. Arch Gynecol Obstet. May;291(5):993-8. https://doi.org/10.1007/s00404-014-3496-x
Çağlayan Özcan H, Gurol Uğur M, Sucu S et al (2020) Blunt dissection technique with finger and vessel skeletonization in the posterior vesical wall for abnormally invasive placenta previa. J Matern Fetal Neonatal Med. Jul;33(14):2441-2444. https://doi.org/10.1080/14767058.2018.1554043
Harrell, F. E., et al. (2015). Regression Modeling Strategies. Springer Series in Statistics.
Kasraeian, M., Hashemi, A., Hessami, K. et al (2021) A 5-year experience on perinatal outcome of placenta accreta spectrum disorder managed by cesarean hysterectomy in southern Iranian women. BMC Women's Health 21, 243. https://doi.org/10.1186/s12905-021-01389-z
Zeng C, Yang M, Ding Y et al (2018) Placenta accreta spectrum disorder trends in the context of the universal two-child policy in China and the risk of hysterectomy. Int J Gynaecol Obstet. 140:312–8. https://doi.org/10.1002/ijgo.12418
Nguyen, P.N., Vuong, A.D.B. & Pham et al (2024) Neonatal outcomes in the surgical management of placenta accreta spectrum disorders: a retrospective single-center observational study from 468 Vietnamese pregnancies beyond 28 weeks of gestation. BMC Pregnancy Childbirth 24, 228 https://doi.org/10.1186/s12884-024-06349-7
Silver RM, Landon MB, Rouse DJ et al (2006) National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. Maternal morbidity associated with multiple repeat cesarean deliveries. Obstet Gynecol. Jun;107(6):1226-32. https://doi.org/10.1097/01.aog.0000219750.79480.84
Gyamfi-Bannerman C, Gilbert S, Landon MB et al (2012) Risk of uterine rupture and placenta accreta with prior uterine surgery outside of the lower segment. Obstet Gynecol. 120:1332–7. https://doi.org/10.1097/aog.0b013e318273695b
Noël I, Ghesquiere L, Guerby P et al (2024). Clinical Risk Factors for Placenta Accreta or Placenta Percreta: A Case-Control Study. J Obstet Gynaecol Can. Apr;46(4):102294. https://doi.org/10.1016/j.jogc.2023.102294
Ari SA, Suner A, Senkaya AR et al (2023) A prospective cohort study: can advanced ultrasonography replace magnetic resonance imaging in the diagnosis of placental adhesion disorders? J Perinat Med. Jan 6;51(5):675-681. https://doi.org/10.1515/jpm-2022-0407
Collins SL, Alemdar B, van Beekhuizen HJ et al (2019) International Society for Abnormally Invasive Placenta (IS-AIP). Evidence-based guidelines for the management of abnormally invasive placenta: recommendations from the International Society for Abnormally Invasive Placenta. Am J Obstet Gynecol. Jun;220(6):511-526. ü https://doi.org/10.1016/j.ajog.2019.02.054
Yang T, Li N, Qiao C et al (2019) Development of a Novel Nomogram for Predicting Placenta Accreta in Patients With Scarred Uterus: A Retrospective Cohort Study. Front Med (Lausanne). Dec 17;6:289. https://doi.org/10.3389/fmed.2019.00289
Chen B, Zhang L, Wang D et al (2020) Nomogram to predict postpartum hemorrhage in cesarean delivery for women with scarred uterus: A retrospective cohort study in China. J Obstet Gynaecol Res. Sep;46(9):1772-1782. https://doi.org/10.1111/jog.14354

Table 1: Clinical Parameters and Demographic Data of the Patients Who Underwent Surgery for PASD

Demographic Data

Cesarean Hysterectomy

(n, %)

195 (64.5%)

Conservative Surgery

(n, %)

107 (35.5%)

P

Maternal Age

(Median [25-75%])

34 [31-36]

30 [26-33]

0.0001

Gravida

(Median [25-75%])

4 [3-5]

4 [3-4]

0.007

Parity

(Median [25-75%])

3 [2-3]

2 [2-3]

0.0001

Prior Uterine Surgeries

(Median [25-75%])

2 [2-3]

2 [2-2]

0.0001

Gestational Weeks

(Median [25-75%])

35 [34-36]

0.26

Hgb Preoperative

(Median [25-75%])

10.6 [9.9-11.3]

10.6 [9.8-11.2]

0.48

Hgb Postoperative

(Median [25-75%])

9.8 [9.1-10.1]

9.6 [9.2-10]

0.65

Bladder Injury

(n, %)

34(%17.44)

7 (%6.54)

0.01

USG Findings:

USG 1 (n, %)

USG 2 (n, %)

USG 3 (n, %)

USG 4 (n, %)

172 (%88.21)

169 (%86.67)

180 (%92.31)

159 (%81.54)

84 (%78.50)

9 (%8.41)

57 (%53.27)

47 (%43.93)

0.03

<0.0001

Placental Location:

Anterior (n, %)

Posterior (n, %)

177 (%90.77)

18 (%9.23)

63 (%58.88)

44 (%41.12)

<0.0001

Pathological Specimen:

Percreta (n, %)

Increta (n, %)

Acreta (n, %)

162 (%83.07)

15 (%7.69)

18 (%9.23)

25 (%23.36)

34 (%31.77)

48 (%44.85)

0.0001

Hgb: Hemoglobin

USG 1: Lack or absence of normal hypoechoic retroplacental zone

USG 2: Multiple lacunaes in the placenta

USG 3: Irregularity in the line between the bladder and uterine serosa and findings of hypervascularization with Doppler Flow

USG 4: Myometrium thickness in the placenta location < 1 mm

Table 2: Multivariate Model of Patients Who Underwent Caesarean Hysterectomy For PASD Compared to Those Who Received Conservative Surgery

Variables

BETA

OR

%95 CI

P

Maternal Age

0.20

1.22

1.08-1.44

0.001

Prior Uterine Surgeries

1.15

3.18

1.57-8.29

0.001

Gestational Weeks

-0.24

0.78

0.56-1.02

0.07

USG Findings:

USG 1

0.73

2.09

0.61-9.56

0.25

USG 2

3.88

48.53

18.42-257.40

<0.001

USG 3

2.06

7.90

2.66-35.12

<0.001

USG 4

0.91

2.49

0.89-8.27

0.08

Placental Location

Posterior

2.26

9.60

2.96-50.76

<0.001

USG 1: Lack or absence of normal hypoechoic retroplacental zone

USG 2: Multiple lacunaes in the placenta

USG 3: Irregularity in the line between the bladder and uterine serosa and findings of hypervascularization with Doppler Flow

USG 4: Myometrium thickness in the placenta location < 1 mm

Predicting the Risk of Cesarean Hysterectomy in the Management of Placenta Accreta Spectrum Disorders: A New Model Based on Clinical Findings and Ultrasonography

Status:

Version 1

Abstract

Purpose

Methods

Results

Conclusion

Figures

What does this study add to the clinical work?

INTRODUCTION

MATERIALS AND METHODS

Data Collection

Investigated Clinical Characteristics

Management of Patients

Data analysis

Statistical Modeling

RESULTS

Model Performance and Internal Validation

DISCUSSION

CONCLUSION

Declarations

References

Tables

Status:

Version 1