High Risk Factors for Placenta Accreta Other than Pregnancy and Their Impact on Patient Prognosis.

Placenta accreta spectrum (PAS) disorders represent an obstetric condition where the placental trophoblast grows into the myometrium of the uterine wall. 1 As the placenta does not separate spontaneously after delivery, patients may experience greater volumes of blood loss, which may even be life-threatening. According to the International Federation of Gynecology and Obstetrics classification, PAS disorders may be categorized into three types: abnormally adherent placenta (accreta), abnormally invasive placenta (increta), and abnormally invasive placenta (percreta). 2 Placenta accreta is the mildest type and is defined as direct contact between the trophoblasts and myometrium without invasion. In placenta increta, trophoblasts grow into the myometrium; however, in placenta percreta (which is the most hazardous type), the trophoblasts grow through the myometrium and serosa and even protrude into surrounding organs, increasing the likelihood of lethal complications. 2 In this context, an increase in cesarean delivery rates in recent years has led to an increase in the incidence of PAS disorders. 3 In addition to cesarean delivery, any procedure damaging the endometrium may result in this condition. 1 Dilation and curettage (D&C), myomectomy, uterine artery embolization (UAE), and hysteroscopy are risk factors that cannot be overlooked. 1 Baldwin et al . 4 analyzed data from 854 primiparous women and found the relative risks for one, two, and three gynecologic procedures to be 1.5, 2.7, and 5.1, respectively. Maternal age, placenta previa, and the use of assisted reproductive technology are additional risk factors. 1 Recent studies suggest that smoking during pregnancy, uterine anomalies (such as endometriosis, uterine myomas, and adenomyosis), and hypertension may similarly increase the susceptibility to PAS disorders. 5 , 6 Ultrasound is the most convenient screening approach for this condition; it also offers favorable accuracy. 7 , 8 Findings related to PAS disorders have been identified and confirmed by recent studies. 9 , 10 Notably, many scoring systems use ultrasound data in combination with the patients’ history to predict the severity and prognosis of this condition. 11 – 13 These scoring systems place considerable emphasis on the number of previous cesarean sections and D&C procedures. Some of these systems can only be used to predict the risk in patients with a history of cesarean delivery. 13 This represents a disadvantage of these systems and limits their utility. Notably, a small group of patients with PAS disorders has no history of gestation. These patients are difficult to identify, as clinical studies in this population are lacking. As lower prenatal detection rates lessen the need for preoperative management, their prognosis remains unclear. Owing to a limited number of cases, the risk factors for PAS disorders without any history of pregnancy also remain unclear. This study aimed to determine the most pertinent factors responsible for PAS disorders in patients without any history of pregnancy and evaluate their prognostic implications.

Huixia Yang is the Editor-in-Chief of Maternal-Fetal Medicine . The article was subject to the journal’s standard procedures, with peer review handled independently of this editor and her research groups.

None.

The data for this study were collected based on the clear guidelines formulated by the China National PAS Working Group for data collection. This retrospective cohort study included a total of 1009 patients diagnosed with PAS disorders across 10 tertiary hospitals in China between January 1, 2018, and December 31, 2018. These hospitals included the Peking University First Hospital, Second Hospital of Hebei Medical University, First Affiliated Hospital of Zhengzhou University, Third Affiliated Hospital of Zhengzhou University, Nanjing Drum Tower Hospital, Xiangya Hospital Central South University, Third Affiliated Hospital of Guangzhou University, Maternal & Child Health Hospital of Guangxi Zhuang Autonomous Region, Tongji Hospital, and Hunan Provincial Maternal and Child Health Care Hospital. All patients were identified via the medical record systems of each hospital based on a definite diagnosis obtained during surgery or on histopathologic evaluation (as per the International Federation of Gynecology and Obstetrics recommendations). In cases without any pathology report, the type of PAS was determined based on the surgical findings recorded immediately after the operation. 2 Overall, 282, 530, and 197 patients had placenta accreta, placenta increta, and placenta percreta, respectively. Patients having no history of pregnancy or live births were included in the study. Cases with fetal malformations and those lacking basic information were excluded. Patients with ultrasound details (reports showing placenta privia or placenta accreta) and maternal-fetal outcomes were also recorded and analyzed. A total of 45 patients without any history of pregnancy were finally selected; data pertaining to their demographic characteristics, medical examinations, and treatments were obtained in detail. This study was approved by the ethics committee of the Peking University First Hospital (ID: 2019(232)). Written informed consent was obtained from all patients. At each center, all ultrasound reports were prepared by experienced radiologists who specialized in ultrasonography. As in previous studies, placenta accreta was considered mild PAS, and placenta increta and percreta were collectively considered severe PAS. 11 , 14 All the patients underwent routine ultrasound examinations at 11 to 13, 21 to 24, 30, 36, and 40 weeks (if applicable) at each hospital during clinical visits. Transvaginal and transabdominal ultrasound was performed by experienced radiologists. The following ultrasound findings were considered to be indicative of PAS disorders: placental lacunae, loss of hypoechoic space, abnormalities of the uterus-bladder interface, and color Doppler abnormalities including hypervascularity and bridging vessels. 15 Placenta previa was detected by transvaginal ultrasound after 28 weeks and was determined by the relationship between the lower edge of the placenta and internal os. 16 It was categorized into three types, namely, marginal, partial, and complete. 16 Certain procedures including myomectomy, hysteroscopy, and UAE may cause endometrial damage. 1 The clinical symptoms comprise antepartum hemorrhage, abdominal pain, anemia, and hematuria. All patients in this study received postpartum uterotonic treatment for hemostasis; none underwent the hysterectomy. Uterine tourniquets, sutures, arterial ligation, and uterine tamponade were reserved for use if and when necessary. The occurrence of intrapartum hemorrhage was recorded along with the volume of blood transfused. Maternal complications included postpartum hemorrhage, acute blood transfusion reactions, disseminated intravascular coagulation, shock, infection, anemia, and damage to other organs. Data pertaining to the fetus included the number of gestational weeks, Apgar scores from 1 minute (Apgar 1) to 5 minutes (Apgar 5), neonatal complications (including asphyxia, pneumonia, and infections), and the need for further treatment in the pediatric department. All data were obtained by searching the electronic medical record system. Continuous data have been presented as the mean ± standard deviation ( SD ) or median (range), after testing for normality (only age passed the normality test and was recorded as the mean ± SD ). Categorical data have been presented as the number ( n %). The Student t test and the Mann-Whitney U test were used to compare differences between continuous variables; the chi-squared test was used for comparison between groups of categorical data. Univariate analysis was used to identify the relationship between maternal age, endometrial damage, in vitro fertilization-embryo transfer (IVF-ET), placenta previa, and maternal-fetal outcomes (intraoperative bleeding, blood transfusion, maternal complications, maternal hospital stay in days, number of gestational weeks, Apgar score, neonatal complications, and admission to the ward during the neonatal period). Multivariate analysis included the variables demonstrating statistical significance on univariate analysis. Linear and logistic regression were used for identifying the relationship between two continuous and categorical data sets, respectively. A two-tailed P < 0.05 was considered statistically significant. The SPSS 24.0 software package (IBM Corp, Armonk, NY) was used for statistical analyses.

The patients were classified based on the type of PAS disorder. Among the 45 patients (4.46% of 1009), 22 (48.9%) had placenta accreta, 21 (46.7%) had placenta increta, and 2 (4.4%) had placenta percreta. Patients with placenta increta and percreta were included in the severe PAS group for the purposes of statistical analysis. Three of 45 patients who delivered by spontaneous labor experienced 200 to 500 mL of blood loss; they were finally diagnosed with placenta accreta. Among them, 42 underwent cesarean delivery for medical reasons; 21 had a suspected PAS disorder, and 9 had placenta previa. The other patients underwent cesarean section for twin pregnancies, fetal distress, oligohydramnios, breech/transverse fetal positions, or severe preeclampsia. The blood loss related to vaginal deliveries was not considered. A comparison of the demographic and clinical characteristics of patients with mild and severe PAS disorders is presented in Tables 1 and 2 . None of these patients had a history of myomectomy or UAE. One patient underwent hysteroscopy and the other patient had surgery using the uterine manipulator; they were considered to have endometrial damage. Patients with mild and severe PAS disorders demonstrated no differences in terms of their medical history and clinical symptoms. Although there was a statistically significant difference in the number of days they were hospitalized ( P = 0.040), the average duration did not significantly differ between each type (10 vs . 8 vs . 9 days). The diagnosis of placenta previa showed a strong association ( P = 0.003) with the incidence of PAS disorders. Demographic and clinical characteristics in PAS patients without pregnancy history. Data are presented as n (%), mean ± SD , and median (range). *Compare between mild PAS and severe PAS. † Student t test. ‡ Chi-squared test. § Mann–Whitney U test. ∥ Calculated in 19 patients. ¶ Calculated in 18 patients. ** Calculated in 20 patients. —: Not applicable; IVF-ET: In vitro fertilization-embryo transfer; PAS: Placenta accreta spectrum; SD : Standard deviation. Multivariate analysis of demographic and clinical characteristics in PAS patients without pregnancy history between mild and severe PAS. *Logistic regression. CI : Confidence interval; IVF-ET: In vitro fertilization-embryo transfer; OR : Odds ratio; PAS: Placenta accreta spectrum. The total ultrasound recognition rate for PAS was not high. Data pertaining to ultrasound examination were available for 39 of 45 patients; 5 (12.8%) of these 39 patients had been preoperatively diagnosed with PAS disorders by ultrasound. In relation to maternal prognosis, intraoperative bleeding and blood transfusion data were available for 41 of 45 patients, and the rate ( P = 0.004) and amount ( P = 0.006) of blood transfusion differed between PAS types. Although the 1-minute and 5-minute Apgar scores showed significant differences, there was no significant difference in the rate of neonatal asphyxia ( P > 0.050). Maternal age, endometrial damage, and history of IVF-ET or placenta previa were considered prognostic factors; these have been discussed below. The results of univariate and multivariate analyses of maternal prognosis are shown in Tables 3 and 4 . Intraoperative bleeding, blood transfusion volumes, maternal complications, and duration of hospital stay in days were used as indices to demonstrate the significance of various prognostic factors. On univariate analysis, older age was found to be related to more complications ( P = 0.011), and IVF-ET caused more intraoperative bleeding ( P = 0.007; 1257.1 vs . 722.2 mL); however, none of these factors demonstrated an impact on multivariate analysis ( P > 0.050). Univariate analysis of PAS patients without pregnancy history with maternal prognosis. *Linear regression. † Mann-Whitney U test. ‡ Chi-squared test. IVF-ET: In vitro fertilization-embryo transfer; PAS: Placenta accreta spectrum. Multivariate analysis of PAS patients without pregnancy history with maternal prognosis. *Linear regression. † Logistic regression. CI : Confidence interval; IVF-ET: In vitro fertilization-embryo transfer; OR : Odds ratio; PAS: Placenta accreta spectrum. The results of univariate and multivariate analyses of neonatal prognosis are shown in Tables 5 and 6 . The number of gestational weeks, Apgar scores, fetal complications, and receipt of further treatment in the pediatric ward were used as prognostic indices. The findings on both univariate and multivariate analyses indicated that endometrial damage led to lower Apgar scores in the newborn and susceptibility to neonatal asphyxia (average Apgar 1 score, 7.8). Placenta previa was found to be associated with a shorter gestational period ( P = 0.047; 36.9 vs . 37.3 weeks); IVF-ET was only associated with lower neonatal hospitalization rates ( P = 0.044; 25.8% vs . 0.0%) on univariate analysis. None of these factors demonstrated any influence on multivariate analysis. Univariate analysis of PAS patients without pregnancy history with neonatal prognosis. *Logistic regression. † Mann-Whitney U test. ‡ Chi-squared test. IVF-ET: In vitro fertilization-embryo transfer; PAS: Placenta accreta spectrum. Multivariate analysis of PAS patients without pregnancy history with neonatal prognosis. *Linear regression. † Logistic regression. CI : Confidence interval; IVF-ET: In vitro fertilization-embryo transfer; OR : Odds ratio; PAS: Placenta accreta spectrum.

Our study showed that in patients without a history of pregnancy, placenta previa was strongly associated with severe PAS disorders; this was similar to the observation in patients with a history of a prior pregnancy. Endometrial damage was associated with lower Apgar scores; particular care should therefore be taken during clinical management of the mother and baby. Finally, MRI should be used more proactively in cases where the patient provides a related gynecologic history and has a suspicious ultrasound finding.

The incidence of PAS disorders is increasing with the rise in cesarean section deliveries. However, certain patients with PAS disorders have no history of cesarean section or even pregnancy. Because PAS disorders represent a serious condition, the risk factors and prognosis of these patients should also be considered. In our study, placenta previa was found to be strongly associated with severe PAS disorders. Many studies have shown that apart from previous cesarean delivery scars, placenta previa is an independent risk factor for the development of PAS disorders. Placenta previa is also associated with greater volumes of blood loss. In this context, Mulla et al . 17 reported that women with PAS disorders who have coexistent placenta previa experience greater blood loss compared with those without the condition (3500 vs . 1200 mL). The findings from our study concurred with those of previous studies in that it confirmed the notable influence of placenta previa in patients without a history of pregnancy. However, it failed to demonstrate its significance in terms of the extent of hemorrhage and blood transfusion requirements. This may be attributed to the small number of included cases. The findings demonstrated a relatively low rate of diagnosis in patients with PAS disorders who had no history of pregnancy. Notably, the Society of Obstetricians and Gynecologists of Canada has recently recommended ultrasound screening for women with high-risk pregnancies who have a history of uterine surgery (including multiple cesarean section); however, women without previous cesarean section have not been considered. 18 None of the patients with mild PAS disorders were diagnosed with the condition before delivery; only five (27.8%) patients with severe PAS disorders were identified. Berkley et al . 7 found the sensitivity and specificity of ultrasound detection to range from 67% to 97% and 50% to 98%, respectively, in seven studies; this was considerably higher than the rates observed in our cohort. We speculate that in cases where the placenta was positioned on the anterior uterine wall, the proximity between the cesarean section scar and placenta may have led to unsatisfactory ultrasound detection rates. In cases where the placenta was positioned on the posterior wall, it was also difficult to identify the condition using ultrasound. Magnetic resonance imaging (MRI) offers higher sensitivity than ultrasound for identification of the extent of placenta accreta. 19 MRI may therefore be helpful in women with placenta previa who have no history of pregnancy but do have a history of gynecological surgery. Interestingly, a strong relationship was observed between endometrial damage and Apgar scores in our 45 cases. Following evaluation of the pathophysiology of PAS, Jauniaux et al . 20 suggested that the depth of villous penetration into the myometrium is likely to be related with the extent of the deciduomyometrial damage. Microscopic damage secondary to D&C and endometritis is therefore more likely to lead to mild PAS. Surgical scars that cause deficiencies in local vascular remodeling may lead to severe PAS. It is reasonable to conclude that poor vascularization and tissue oxygenation in the area of endometrial damage may be associated with failure of local re-epithelialization and decidualization. These changes may impact implantation, placentation, and placental development. 20 Insufficient placental function may ultimately cause acute or chronic fetal distress and asphyxia, leading to lower Apgar scores after birth. This may explain why endometrial damage affected the Apgar score in our cohort (although it did not lead to premature birth). Based on our findings, we recommend detailed evaluation of the history of previous gynecological surgeries (including possible use of the uterine manipulator), as two of the four patients with endometrial damage had a history of such surgeries (and not hysteroscopy). Our study is the first to provide insights into the characteristics of patients with PAS disorders who have no history of pregnancy. Clinical studies in this area are lacking. As in patients with PAS disorders who have a history of cesarean section, placenta previa contributed significantly to their diagnosis. The lack of distinction on ultrasound was a serious impediment in these cases. In cases with an obscure gynecologic history or indistinct ultrasound findings, MRI may be the preferred option for obtaining a definite diagnosis. Endometrial damage was associated with lower Apgar scores in this cohort; particular care should therefore be taken in the clinic, and adequate preparations should be made for the newborn. There are some limitations to our study. It had a retrospective design, which limited its effectiveness compared with that of a prospective study. The sample size of 45 was small, and the numbers in each group may have reduced the accuracy of statistical analysis. It is possible that some patients failed to provide correct information regarding their pregnancy history owing to personal reasons; in this context, sex-selective abortions are not permitted in China, and at-home abortions place women at high risk of uterine damage 21 – 23 ; the impact of abortion cannot therefore be ignored. Larger prospective studies are warranted in the future.

None.

The authors thank the clinicians from all 10 tertiary care centers in China who notified them of cases and completed data collection forms.

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Huixia Yang and Jie Yan designed the study; Xinrui Yang wrote the manuscript; Xinrui Yang and Weiran Zheng performed statistical analyses and analyzed the data; Huixia Yang and Jie Yan coordinated the study over the entire time. All authors approved the final manuscript.