Associations of characteristics of previous induced abortion with different grades of current placenta accreta spectrum disorders

Abstract Objective Induced abortion could increase the risk of placenta accreta spectrum (PAS) in the next pregnancy. We aimed to explore the associations between characteristics of previous induced abortion and grades of current PAS. Methods A retrospective case-control study was performed in eligible pregnant women with PAS between January 2014 and June 2022. Data collected included demographics, obstetric characteristics, and information on previous induced abortion. Results The study included 211 pregnant women, with 51 and 160 in the invasive (placenta increta or percreta) and adherence (placenta creta) PAS groups, respectively. The risk of invasive PAS was 14.3-fold higher in patients with abnormal vaginal bleeding after abortion (odds ratio = 14.3, 95% confidence interval 5.6–36.4, p < .01) than those without abnormal vaginal bleeding and approximately 5.8-fold higher in patients with the last induced abortion ≥5 years ago than those within 5 years (odds ratio = 5.8, 95% confidence interval 2.2–15.2, p < .01). The risk of invasive PAS was 13.4-fold higher in patients with placenta attached to uterine cornu than patients with the placenta attached to uterine wall (odds ratio = 17.5, 95% confidence interval 5.5–55.5, p < .01). The number of previous induced abortions, hospital grades, and gestational age at abortion were not different between two groups. Conclusion In pregnant women with a history of induced abortion, abnormal vaginal bleeding after induced abortion and prolonged duration after the last induced abortion increased the risk for invasive PAS in the current pregnancy. The number of previous induced abortions and gestational age at abortion had no relation to the grades of PAS.


Introduction
Placenta accreta spectrum (PAS) disorders refer to the abnormal attachment of the placenta to the uterine myometrium.Based on the invasion depth of the placental villi into the myometrium, PAS has been categorized into three grades, 1) placenta creta or adherenta (PA) when the chorionic villi simply adhere to the myometrium surface with no invasion; (2) placenta increta (PI) when the chorionic villi invade up to the external layer of the myometrium; and (3) placenta percreta (PP) when the chorionic villi penetrate through the myometrium and reach surrounding structures [1].PA is commonly considered as an adherent grade of PAS and PI and PP are considered as invasive grades of PAS [2].PAS can cause various short-and long-term adverse maternal and neonatal outcomes [3].The depth of placental invasion into the myometrium is associated with the severity of outcomes, with PI/PP having higher risks for postpartum hemorrhage, hysterectomy, and preterm birth compared with PA [4][5][6], which require special attention and management [7].
Uterine curettage, as a basic technique for induced abortion, could result in an injury to the endometrium [2,8,9].In China, there have been 13 million abortions every year, and the rate of repetitive abortion has been reported to be as high as 55.9%.The average annual rate of induced abortion was approximately 29% among married women between 1984 and 2011 [10], higher than those in many developed countries [11].Given the large number of induced abortions in China, there is an increasing population of women at their childbearing age with potential endometrium injury.The endometrium injury could lead to PAS.Although most studies have demonstrated a close association between the previous induced abortion and the PAS in the current pregnancy, the exact relationship between these two was rarely reported [12,13].More studies are required in order to help clinicians to make early correct diagnosis of PI/PP and apply appropriate managements.
In the present study, we hypothesized that the characteristics of previous induced abortion could be associated with different PAS grades in the current pregnancy.We investigated the correlation between the characteristics of previous induced abortion and the PAS grades in the current pregnancy, with the purpose to provide clinical evidence for easy identification of patients at high risk for PI/PP.

Study design and participant selection
This was a retrospective case-control study that included women who were diagnosed with PAS in the current pregnancy and with a history of induced abortion at the Affiliated Hospital of North Sichuan Medical College, China, between January 2014 and June 2022.The study protocol was approved by the hospital ethics committee (No 2023ER070-1).Informed consent was waived due to the retrospective design of the study.
Clinical examination was recommended as the most important criteria to define and categorize PAS disorders [1].Therefore, we included women who had a history of induced abortion and who were diagnosed with PAS in the current pregnancy based on clinical or histological examinations [14].In women undergoing hysterectomy, PI/PP was diagnosed histologically when there was placental villi invading the myometrium or reaching the uterine serosa.In women without hysterectomy, the PAS grade was diagnosed based on International Federation of Gynecology and Obstetrics (FIGO) classification of the clinical diagnosis of PAS [15], or with magnetic resonance imaging (MRI) when necessary.At the postpartum, PI/PP was diagnosed clinically when there was no (partial or total) cleavage plane palpable between the uterus and the placenta during manual exploration.Gentle cord traction resulted in the uterus being pulled inwards without separation of the placenta, with significant amounts of hypervascularity or placental tissue observed to invade through the uterine serosa or other pelvic tissue/organs.PA was clinically diagnosed when the placenta could not detach spontaneously from the uterus but could be removed successfully with a gentle attempt and no placental tissue was observed to invade through the uterus surface, with no or minimal neovascularity.
The exclusion criteria were pregnant women with any of the following, 1) history of cesarean section; 2) history of hysteroscopic surgery; 3) history of myomectomy; 4) history of embryo transplantation; (5) twin pregnancies; 6) incomplete clinical data.

Data collections and group assignments
Medical records were reviewed to collect data, including demographics, gestational age, gravidity, parity, obstetrics features (mode of delivery, and ultrasound or MRI records for placental location-main body of placenta attachment site, blood loss, transfusion volume, hospital stay, and puerperal infection), and information on the previous induced abortion (numbers of induced abortion, gestational age at abortion, abnormal vaginal bleeding, and hospital grades).Abnormal vaginal bleeding was defined as vaginal blood loss after induced abortion exceeding individual previous menstrual blood volume and/or the duration of bleeding exceeding 15 days [16].
Due to the small number of patients with PP, we combined patients with PI and PP into one group (invasive group).Pregnant women with PA were included in the adherence group.

Statistical analysis
SPSS 22.0 for Windows (SPSS Inc., Chicago, IL, USA) was used for statistical analysis.Continuous variables are presented as median (interquartile range), while categorical variables are presented as numbers and percentages.Normal distribution of continuous variables was determined using the Kolmogorov-Smirnov test.Group comparisons were done with categorical variables using the Chi-square test, and for non-normally distributed continuous variables using the Mann-Whitney U test.Multivariate logistic regression analysis was conducted to determine the potential risk factors of PP or PI.The odds ratio (OR) and 95% confidence intervals (95% CI) were reported.A P-value <.05 was considered statistically significant.

Results
Of the 16,325 women who give birth in our hospital, 211 (1.3%) women were diagnosed with PAS.In these 211 women, 3 women received hysterectomy and their PAS grades were diagnosed histologically.The PAS grades in 26 women were diagnosed based on the FIGO classification and MRI results.The PAS grades in the rest of women were diagnosed clinically based on the FIGO classification.Finally, 160 women had PA (adherence group) and 51 had PI/PP (invasive group).The patient enrollment flowchart is shown in Figure 1.
The maternal demographic characteristics and peripartum outcomes were compared between the two groups (Table 1).Gestational age at delivery, maternal age, and parity did not differ significantly between the adherence and invasive groups.The percentage of women with gravidity !4 times was significantly higher in the invasive group.Estimated blood loss, blood transfusion volume, and length of hospital stay, and the percentages of women with puerperal infection, pelvic arterial embolization, and late postpartum hemorrhage were higher in the invasive group than the adherence group.There was a higher percentage of women with the placenta attached to the uterine cornu in the invasive group than in the adherence group.The percentage of women with placenta previa did not differ between the adherence and invasive groups.
The information on the previous induced abortions was comparable between the two groups (Table 2).The percentage of abnormal vaginal bleeding after curettage was higher in the invasive group than the adherence group (62.8% vs. 13.7%,p < 0.01).The number of induced abortions, hospital grades, gestational age at abortion, and ages when undergoing the first induced abortion were not different between the adherence and invasive groups.
Multivariate logistic regression analysis was performed to show the relationship between different previous abortion characteristics with PI/PP.The results demonstrated that the risk of invasive PAS was 14.3-fold higher in patients with abnormal vaginal bleeding (OR ¼ 14.3, 95% CI 5.6-36.4,p < 0.01).The risk of invasive PAS was approximately 5.8-fold higher in patients with a duration !5 years between the last abortion and current pregnancy compared with those with a duration <5 years (OR ¼ 5.8, 95% CI 2.2-15.2,p < 0.01).
In addition, the placental position in current pregnancy also increased the risk of invasive PAS.The risk of invasive PAS was 17.4-fold higher in patients with the placenta partially or fully attached to the cornu compared with patients with the placenta attached to the wall (OR ¼ 17.5, 95% CI 5.5-55.5, p < 0.01) (Table 3).

Discussion
The global prevalence of PAS was estimated to be approximately 0.2%, with the adherent type of PAS slightly more than the invasive type of PAS [17].However, the prevalence of PAS secondary to various causes was reported to be 2.4% among pregnant women in China [18].In our study, the incidence of PAS in pregnant women with previous induced    abortion was 1.3%.Considering that our hospital is a tertiary obstetric critical care referral center, the actual incidence of PAS in pregnant women with previous induced abortion might be lower in the general population.
There have been inconclusive findings on the relationship between the depth of placenta accreta and the number of abortions.Tantbirojn et al. reported that a history of two or more curettages increased the chance of PAS (22% and 0% in PAS and non-PAS groups, respectively, p ¼ 0.05) [12].Peng et al. constructed a multivariate logistic regression model to predict the severity of PAS disorders and found a correlation between an increased PP incidence and a higher number of previous abortions [13].Another recently published study in Japan did not find an association between the number of induced abortions and increased risk of PAS [19].In the present study, we excluded women with a previous history of cesarean section and other intrauterine operations, such as hysteroscopic surgery and embryo transfer.This allowed us to focus on the influence of induced abortions on the depth of PAS and reach a less biased conclusion.We included women with placenta previa in our study, since placenta previa could be resulted from the endometrial injury during the induced abortion.The inter-group analysis did not show a statistically significant difference in the incidence of placenta previa between two groups.The multivariate regression analysis also did not show a strong relationship between placenta previa and grades of PAS.However, considering the lower uterine segment is a common site for PAS, we still recommend careful evaluations of this uterine area by ultrasound or MRI during the prenatal evaluations in pregnant women with placenta previa.In addition, we analyzed the detailed information from the previous induced abortion, such as gestational age at abortion, hospital grades and ages at first abortion.Our results did not show an increased risk of PI/PP with increased number of previous induced abortions.
Our results showed the risk of invasive PAS was 14.3-fold higher in patients with postoperative abnormal vaginal bleeding.Postoperative abnormal vaginal bleeding may be associated with uterine injury and infection which both increase the risk of PI/PP.It suggested that, in pregnant women with a history of induced abortion, abnormal vaginal bleeding after abortion should be carefully screened to identify patients at a high risk for invasive PAS in the current pregnancy.These high-risk patients should receive further ultrasound examination or MRI.
In our study, the risk of PI/PP in patients with a duration ! 5 years between the last abortion and current pregnancy was approximately 5.8-fold higher than in patients with a duration < 5 years.We speculated that endometrial injury secondary to curettage could result in uterine scar tissues that might develop and extend with the time goes on.Consequently, the next pregnancy in women with prolonged duration between last abortion and current pregnancy would be more likely to have PI or even PP.
In our study, the risk of PI/PP in patients with placenta attached to cornu was 17.4-fold higher than that in patients with placenta attached to wall.To our knowledge, there were very few studies on the correlation between placenta location and placenta accrete.In a case report study, cornual pregnancy was reported to be accompanied by PA [20].Guan et al. also described a patient who had a previous cornual pregnancy with PA and developed an acquired uterine arteriovenous fistula [21].Rathod et al. reported a case of cornual pregnancy with PP in a woman with her first pregnancy [22].Our results suggest that current placental attachment to the uterine cornu was a risk factor for PI/PP among women with a history of induced abortion.Therefore, obstetricians should rule out the possibility of PI/PP in pregnant women with a previous induced abortion and current placental attachment to the cornu.
The limitations of our study include its retrospective design in a single hospital.The information on the previous induced abortion and pregnancy were self-reported by the pregnant women, which could have led to inaccurate data collection and potential bias in our analysis.Therefore, large-scale, prospective, and multi-center studies are required to confirm our findings.

Conclusions
In conclusion, in pregnant women with a history of induced abortion, abnormal vaginal bleeding after induced abortion, long duration after the last induced abortion, and current placental attachment to the uterine cornu were risk factors for PI/PP in the current pregnancy.The number of previous induced abortions and gestational age at abortion had no relation to the grades of PAS.Therefore, careful prenatal evaluation of placental location and detailed abortion history should be obtained in the high-risk patients.

Table 1 .
Comparisons of demographic characteristics and peripartum outcomes between the adherence and invasive groups.
a Chi-square test.b Mann-Whitney U-test.

Table 2 .
Previous abortion information between two groups.
b Mann-Whitney U-test.

Table 3 .
Multivariate logistic regression analysis of risk factors for PI/PP patients.