Placenta accreta spectrum (PAS) disorders are severe maternal complications characterized by abnormal adherence of the placental trophoblast to the uterine myometrium. It can be classified as placenta accreta (PA), placenta increta (PI) and placenta percreta (PP) according to the invasion depth of the placenta, and the latter two types are regarded as invasive PAS (1). The prevalence of PAS ranges from 0.01% to 1.1% (2, 3). A study using a nationally representative sample in the United States showed that the prevalence of PAS was 0.29% among women who underwent cesarean delivery with live births, and the rate increased by 2.1% per quarter year from 2015 to 2017 (4). Patients with PAS present considerably higher risks of catastrophic hemorrhage, hysterectomy, organ damage, consumptive coagulopathy and maternal death (2, 4–12). The known risk factors for PAS include placenta previa, prior cesarean delivery, prior uterine surgery and assisted reproductive technology (ART) (4, 13, 14).

It remains unclear whether women with multiple gestation are at higher risk for PAS. Miller et al. (15) found that women with twins conferred a 2.5 times higher risk for PAS than those with singleton gestation. However, Matsuzaki et al. (4) suggested that PAS was less prevalent among multiple births than singletons (OR=0.73). In view of the above controversy, Guo et al. (16) pointed out that the differences may be attributed to the study population, data source and model design. In addition, it is worth noting that both studies only used the International Classification of Diseases (ICD) codes to extract the diagnosis and subtype information of PAS; however, Jotwani et al. (17) observed that the overall accuracy of the ICD-10 for PAS was rather low, which stressed the importance of medical records in confirming PAS diagnosis. Inpatient records can also provide vital information on ante-partum diagnosis and care scheme, as well as the clinical confirmation of PAS diagnoses (by intrapartum observation or pathological results), which added precision to PAS diagnoses and provided the whole picture of peripartum PAS diagnosis and management.

In terms of peripartum outcomes, Shamshirsaz et al. (18) found that PAS patients with twins had a significantly higher red blood cell (RBC) transfusion volume than those with singletons, but they only included PAS patients who underwent cesarean hysterectomy. Therefore, a clearer understanding of the peripartum outcomes of PAS patients with multiple and singleton gestations is warranted.

This study aimed to explore the association between multiple gestation and PAS as well as to compare the peripartum outcomes of PAS patients with singletons and multiple births in a cohort of Chinese patients. The results of this study may shed light on the screening, early detection and timely intervention of PAS patients while informing clinicians and the research community on the characteristics of PAS patients with different gestation modes.

Our research revealed that multiple gestation was independently associated with the risk of PAS. Descriptive analyses showed that PAS among multiple births had a lower rate of previous cesarean section and placenta previa than singleton. There were no significant differences in perioperative outcomes of PAS between singleton and multiple pregnancies.

Our study found that multiple pregnancies may be associated with an increased risk of PAS, which was similar to the findings of Miller et al. (15). Jauniaux et al. (25) theorized that the rising incidence of placental implantation disorders (PIDs), including PAS, could be attributed to an increase in iatrogenic factors such as cesarean section and ART. These iatrogenic procedures may have an impact on the endometrium’s functional integrity, either directly or indirectly. The application of ART in multiple pregnancies may result in disturbance of interactions between the blastocyst and the endometrium, which might be the underlying mechanism for the association of ART and PAS. After adjusting for ART, multiple gestation remained an independent risk factor for PAS. Miller et al. (15) pointed out that in multiple pregnancies, an increased placental surface area is more likely to be implanted over the prior cesarean section scar, incurring PAS. It is evident that the specific mechanism behind the association between multiple gestation and PAS needs further study.

This study found that the overall antenatal diagnosis rate of PAS was 53.69% but only 27.6% among multiple gestations. This might be due to the relatively low prevalence of common PAS screening factors, such as placenta previa and prior cesarean deliveries, in the multiple gestation PAS population, which was also reported in other studies (15, 18). Additionally, more than 75% of PAS in the multiple gestation population was the placenta accreta subtype (compared to 36.1% in the singleton population), which could increase the difficulty in prenatal detection since most ultrasound signs were significantly more likely to be observed in severe percreta cases than mild accreta cases (26). Antenatal detection of PAS as well as management by a multidisciplinary team can reduce severe maternal morbidity (SMM) (10, 27). Since missed antenatal detection can result in catastrophic maternal and fetal outcomes (5), clarifying whether multiple gestation pose an increased risk for PAS could have important clinical implications.

The overall prenatal MRI use was low in this study, especially in the PAS patients with twin gestation. It was controversial regarding whether MRI evaluation could facilitate more prenatal detection of PAS. A recent systematic review comparing the additional clinical value of MRI found that both ultrasound and MRI showed comparable accuracy in the antenatal diagnosis of PAS and showed that the routine employment of MRI with high expense should not be recommended (28). Recent guideline (20) by Chinese experts also suggested that MRI should be used as an adjunct tool for patients who have already been screened on US and suspected to have PAS, rather than as a routine screening tool. However, whether additional MRI evaluation could increase the prenatal diagnostic accuracy of PAS in multiple gestation patients warrants further investigation.

The perioperative events of PAS were not significantly different between multiple and singleton gestation in this study. However, a multicenter study by Shamshirsaz et al. (18) showed that the PAS patients among multiple gestation received more units of RBC transfusion during and within 24 hours of surgery. Miller et al. (15) also found that multiple births with PAS were associated with a significantly higher rate of SMM (blood transfusion, hysterectomy, amniotic fluid embolism, disseminated intravascular coagulation, etc.) than singletons. In our study, the multiple gestation group had a significantly lower prevalence of invasive PAS, which could decrease SMM but it also had lower prenatal diagnosis rate which could cause more SMM. Therefore, whether twin gestation posed an additional risk of SMM in the PAS patients may need further investigation.

Our study’s strengths included the utilization of a multiyear database from a tertiary teaching hospital. The PAS patients in this study were pathologically or clinically confirmed instead of solely based on ICD codes, thus ensuring diagnostic accuracy for the included patients. Since the included patients were all managed in a single center by the same team, using a consistent surgical strategy with standardized pre, intra- and postoperative management protocols, the outcomes of this population were less likely to be confounded by the variations among caregivers.

There were also several limitations in the study. First, there could be some unmeasured biases that were inherent in retrospective studies. Future registered cohorts with prospective designs and data collection plans may provide higher-level evidence on the association. Second, the technology and healthcare quality evolved over a six-year observation time, although we did adjust for the year of delivery in the multivariate models. Third, the hysterectomy rate among PAS patients can be relatively low in our center, which could affect the generalizability of the results on patient outcomes. Finally, this was a single-center study despite its relatively large sample size. Multicenter studies could be designed to validate the current association and explore whether the results of the current study can be generalized.

In conclusion, multiple gestation was associated with a higher risk of PAS. In the multiple births PAS group, traditional risk factors (placenta previa and previous cesarean deliveries) were identified less frequently, which posed challenges to prompt diagnosis of the disease. With the release of the three-child policy in China, the increase in pregnancies with previous cesarean section may lead to an increased incidence of PAS, which stresses the need to further understand this severe maternal complication. In addition, the development and increased utilization of ART technology can result in an increase in multiple gestations, which was the main focus of the current study. Further studies are warranted to explore the association between PAS and multiple births, whether multiple gestation is related to peripartum outcomes, and the mechanisms behind this phenomenon.

The datasets used in this study are available from the corresponding author upon reasonable request.

The studies involving human participants were reviewed and approved by Ethics Committee of Peking University First Hospital. The ethics committee waived the requirement of written informed consent for participation.

ZG and XH collected and analyzed the data, prepared tables and figures, and drafted the paper; JM designed the research and revised the paper; WZ and HY provided clinical supervision. All authors have read and approved the final manuscript.

This study was supported by the Strategic Collaborative Research Program of the Ferring Institute of Reproductive Medicine (Grant No. FIRMA 181104) and the National Key Technologies R&D program of China (Grant No. 2016YFC1000303).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.