The effect of abnormal placentation on maternal serum fetal fraction of cell-free DNA

Introduction

Placenta accreta spectrum (PAS) refers to the range of abnormally adherent or invasive placenta, also known as placenta increta, placenta percreta, and placenta accreta [1]. The current concept of pathogenesis involves the preferential attachment of the blastocyst to an iatrogenic defect in the endometrium/decidua defect following uterine surgery, such as cesarean section. Early implantation to scar tissue facilitates abnormally deep trophoblastic invasion leading to interactions with the radial and arcuate arteries [2]. Even today, this age-old disease carries a heavy burden of complications with severe consequences, including maternal death. The most substantial risk factor for PAS is placenta previa in the index pregnancy after prior cesarean surgery [3, 4]. The incidence has markedly increased in recent years, parallel to the rising rate of cesarean deliveries [5].

Currently, women at risk for PAS are identified based on clinical risk factors mentioned above and imaging studies, ultrasound, and magnetic resonance imaging (MRI) [6], [7], [8]. However, their reported sensitivity and specificity of imaging vary substantially among studies, even when similar criteria are used, owing to interobserver variability [9]. Another challenge with imaging is its accuracy in predicting the depth of trophoblastic invasion. Of particular importance is to differentiate between abnormally adherent placenta (placenta accreta) and abnormally invasive placenta (placenta increta or percreta). The latter is associated with significantly higher mortality and morbidity [10, 11]. Determining the depth of villous invasiveness before delivery is essential to tailor individual care [12]. The current imaging modalities, both sonography [13] and MRI [14] fall short of accurate prediction of trophoblastic invasion.

The need for an accurate, rapid, and cost-effective diagnosis of PAS and its severity has led to a search for biomarkers (Reviewed in [15, 16]). One such biomarker is cell-free fetal DNA (cffDNA) which is widely used for non-invasive prenatal screening (NIPS) of common fetal aneuploidies. cffDNA derives from apoptotic trophoblastic cells, and the portion of the cff DNA of the total cell-free DNA in the maternal circulation is called the fetal fraction (FF) [17]. As the release of cffDNA is closely tied to placental morphogenesis, conditions that affect the placenta can directly impact the FF in maternal circulation [18]. An elevated FF is associated with placental disorders such as preeclampsia [19] and fetal growth restriction [20]. Dolin et al. [21] and Sekizawa et al. [22] reported increased levels of fetal fraction in cases of placenta previa (PP). Sekizawa et al. [22] also reported a high FF in two cases of PAS; however, they did not investigate the depth of invasion. Since concurrent placenta previa and PAS (placenta previa accreta) constitute two-thirds of the PAS cases, it is crucial to understand the interaction between placenta previa and PAS on the FF. In addition, we investigated the impact of the trophoblastic invasion depth on the FF.

Materials and methods

This is a retrospective case-control study approved by the Institutional Review Board (201,801,557). The eligible patients were selected from a cohort of 146 pregnant women managed at our institution for antenatally suspected PAS between January 2013 and December 2021. Only those women carrying a singleton and having NIPS before 20 weeks of gestation were included in the analyses. Those diagnosed with fetal anomalies and aneuploidies, delivered outside our institution, and lost to follow-up, and with missing FF results or histopathologic evaluation for PAS were excluded. The FF was determined by various commercial laboratories that perform NIPS. The patients’ source of reimbursement played a significant role in the choice of the commercial laboratory. Placental location was determined by sonography performed by maternal-fetal specialists. Transvaginal sonography was performed to evaluate the location of the placenta in relation to the internal cervical os. A patient with the leading edge of her placenta within 2 cm of the internal cervical os was labeled as having PP in our analyses. The diagnosis of PAS was established histologically according to the criteria proposed by the International Federation of Gynaecology and Obstetrics (FIGO). (1) Relevant demographic and clinical data were extracted from the patients’ medical records. The patients included in the study were divided into four groups: (1) normal placentation, (2) PAS), (3) PP, and (4) PAS + PP. Assuming a standard deviation of 3% for the population and using an α of 0.05 and β of 0.8., we estimated that we need at least six subjects in each group to detect a difference of 6% in means of cffDNA fraction. We used the Shapiro-Francia test to assess the distribution of our data for normality. One-way ANOVA and Kruskal-Wallis tests were used to compare parametric and non-parametric data.

Results

Figure 1 is the flowchart to depict selection of the patients for our analysis. Of the 146 patients in our cohort, 53 underwent NIPS and were delivered at our hospital. Of the remainder, 70 had no genetic screening or testing, 19 had either biochemical genetic screening or prenatal diagnosis, and 4 were delivered at another hospital and lost to follow-up. Of the 53 patients who underwent NIPS, we excluded 10 who lacked a documented FF and 12 due to other criteria. The final analyses were performed on 31 patients: 8 with normal placentation, 10 with PP, 6 with PAS, and 7 with PP + PAS. Among the groups, there were no significant differences in baseline demographic and clinical characteristics except the median number of prior uterine surgeries (Table 1). Also, the groups did not significantly differ in their median FF (Figure 2).

Figure 1:

Flow diagram of patient selection. The unfilled boxes depict the selected patients for the study. n, number; wks, weeks; PP, placenta previa; PAS, placenta accreta spectrum.

Figure 2:

Fetal fraction and placentation. PAS, placenta accreta spectrum; PP, placenta previa.

Of the 13 patients with PAS with or without PP, 7 had abnormally adherent placenta, and 6 had abnormally invasive placenta. When stratified according to the depth of placental invasion, i.e., no PAS, abnormally adherent, and abnormally invasive placenta, the FF did not discriminate any group (Figure 3).

Figure 3:

Fetal fraction and the depth of placental invasion. PAS, placenta accreta spectrum.

Discussion

FF determined before 20 weeks of gestation is not predictive of PAS in patients with or without PP. These findings are consistent two previous studies that did not find any significant relation between means or median FF and abnormal placentation, either PAS or PP. However, these studies did not control for the concurrent PAS and PP. We demonstrated that the fetal fraction does not differ between patients with placenta previa accreta or PAS alone.

To date, there has been no publications regarding the trophoblastic invasion depth on FF except a case report [22] of a patient with a high FF and the histologically confirmed invasive placenta. Here, we also demonstrated that FF does not correlate with the depth of trophoblast invasion among patients with PAS to be of clinical significance.

There was a significant difference between the median number of surgeries between the groups. This is potentially due to the fact that the patients were selected from a cohort of patients referred for PAS evaluation. Patients with greater than or equal to four prior cesareans were recommended to be referred for evaluation regardless of other ultrasound findings.

The major strength of our study was the number of PAS cases analyzed, which is almost twice as many as in the previous reports [23, 24]. Despite its relatively large size, our study may not be powered to detect small differences which are unlikely to be of clinical significance. Yet, a meta-analysis based on individual patient data from this and other studies [21], [22], [23], [24] would improve the power. Another strength is limiting our analysis to those patients who underwent NIPS before 20 weeks’ gestation, as often performed. This approach should minimize the effect of gestational age on the FF.

The major limitation is the retrospective design. The FF data were unavailable in 19% (10/53) of patients who had undergone NIPS at other institutions and then transferred to our care. Our population’s screening rate with cffDNA was only 35% (51/146). Failure to analyze the cffDNA of the entire cohort may have confounded our results. Nevertheless, this study represents the real-world clinical setting where patients chose other options over NIPS.

From a biological perspective, the lack of a relation between FF and PAS is not entirely surprising. Histologic studies by Kim et al. demonstrated no significant difference in apoptotic rates and proliferative index between abnormally adherent or invasive placentas compared to normally implanted placentas [25]. This is consistent with the current concept that abnormally deep trophoblastic invasion is the result of endometrium/decidua defect rather than excessive trophoblastic proliferation [2].

Although cffDNA is a valuable tool for screening common fetal aneuploidies, FF does not aid the clinical management of PAS. Imaging and clinical risk-based approaches remain the mainstay of identifying patients at risk for PAS, and the quest for biomarkers should continue.