Prenatal diagnosis of bilobate placenta: incidence, risk factors and impact on pregnancy outcomes

Study design, participants

A prospective observational study took place between June 2018 and June 2022, at the Third Department of Obstetrics and Gynecology, School of Medicine, Faculty of Health Sciences, Aristotle University of Thessaloniki, Greece. The study was conducted according to the STROBE statement [7].

All pregnant women with a singleton pregnancy, undergoing both a routine first trimester scan at 11⁺⁰–13⁺⁶ and an anomaly scan at 20⁺⁰–24⁺⁰ gestational weeks were eligible to participate in the study. Exclusion criteria were missing data, miscarriage between the two scans, termination of pregnancy and vasa previa. All the scans were performed by maternal-fetal medicine consultants. Fetal scanning, including measurement of uterine artery pulsatility index (UtA PI), was performed transabdominally, according to the ISUOG guidelines [8]. Pregnancies diagnosed with a bilobate placenta (including those with succenturiate) were considered as the study group. Effort was taken for the postnatal confirmation of the finding. No false positive or false negative cases were recognized; however, confirmation was not always feasible, especially in cases of cesarean section. All pregnancies were dated according to the fetal crown-rump length at the first-trimester scan or the date of embryo-transfer in ART cases. A dedicated database (Astraia Software GmbH, Munich, Germany) was used for the storage of the patient data, including demographic and obstetric parameters.

It should be noted that the finding of a bilobate placenta did not affect the obstetric management as there currently is no guideline recommending modification of standard practice in such cases. The finding was mentioned in the ultrasound report with the only intention to increase awareness following a vaginal delivery for the complete removal of the placenta.

All the participants consented that their anonymized files may be used for research. The study protocol was approved by the Bioethics Committee of the School of Medicine, the Aristotle University of Thessaloniki, Greece (3.188/2-5-2018). Of note, no incentives were provided for participation in the study.

Measurements

With regards to the variables of interest, maternal medical and obstetric parameters were collected. Height and weight were measured during the first visit for the body mass index (BMI) calculation. Sociodemographic characteristics (age, parity, smoking, conception via ART) were collected through our database or directly during the recording of the maternal history.

The primary outcome under investigation was PE, defined as gestational hypertension accompanied by end-organ involvement [9]. Secondary outcomes were SGA, defined as birthweight below the 10th percentile [10], PTD, defined as the delivery between 22⁺⁰ and 36⁺⁶ weeks of gestation [11] and gestational age at delivery. Additionally, we studied all the available risk factors for possible correlation with the presence of bilobate placenta.

Statistical analysis

The continuous variables are represented as means (standard deviation), if normally distributed, otherwise as the medians (interquartile ranges). Categorical variables are expressed as numbers and percentages. Based on the type of statistical distribution, comparisons between groups were made using the parameterized Student’s T-test or nonparametric Mann–Whitney U test (continuous variables) and chi-square test (categorical variables).

Associations with the three main outcomes (PE, SGA and PTD), were assessed using multivariable logistic regression to adjust for the known and available confounders, while multivariable linear regression was conducted for gestational age at delivery. According to the most recent suggestions, a regression model forcing in all the existing factors is needed to calculate the unbiased prognostic effect of the new factor under investigation [12]. Thus, all the available variables that are known or expected to be confounders for PE, SGA and PTD according to published data were selected [13, 14]. Correlation analysis using Spearman’s rank correlation coefficient was performed among the continuous variables to assess multicollinearity.

A risk factor analysis was also conducted by employing a multivariable logistic regression model with the bilobate placenta as the dependent variable to consider possible confounders and offer more exact effect estimates; backward stepwise selection, based on p-value ≤0.1, was employed, in order to differentiate among the risk factors that had a significant prognostic value for bilobate placenta.

A complete-case analysis was employed in our study whereby cases with missing data were excluded from the analysis. This decision was based on the relatively small proportion of affected cases (approximately 2 % of the total population) and the absence of missing values in the study group, thereby not compromising the statistical efficiency by excluding any of our limited study cases [15]. A 5 % critical level of significance was used in every analysis. All analyses were conducted using R 2.15.1 (R foundation, Vienna, Austria).

Principal findings

The main findings of this study were: 1) the incidence of prenatally diagnosed bilobate placenta in singleton pregnancies is 2 %, 2) bilobate placenta is associated with an increased prevalence of abnormal cord insertion and increased mean UtA PI z-score, 3) a statistically significant association between bilobate placenta and PE was identified, while no association was found with SGA and PTD and 4) conception via ART and history of 1st trimester miscarriage are the main risk factors for bilobate placenta.

Interpretation of the findings

Based on the findings of the present study, prenatally diagnosed bilobate placenta involves about 1 in 50 singleton pregnancies and these women have an almost two-fold increased risk for PE, compared to the control group. Given that bilobate placentas are typically regarded as benign anatomical variations, this finding is intriguing and implies that bilobate placenta and PE may share the same pathophysiological procedure of abnormal placentation. Our findings are consistent with a clinicopathological analysis of 130 abnormally shaped placentas, including 48 succenturiate and 12 bilobed, which found an association between the abnormal placental shape and insufficient maternal-fetal perfusion [16]. A systematic review by Schiffer et al. showed that pregnancies complicated by placental syndromes, such as PE and SGA have higher placental thickness and greater presence of placental lakes and calcifications compared to the uncomplicated ones, however, bilobate placenta was not examined individually [17]. This is further supported by the significant association noted between the existence of bilobate placenta and a higher mean UtA PI z-score in our study, indicating a potential indirect mechanism where bilobate placentas may cause PE through abnormally high UtA PI; abnormal UtA PI is a well-studied predictor of PE [18]. Additionally, we found that bilobate placenta was linked to a higher prevalence of abnormal cord insertion (9.8 vs. 27.1 %). Given that the most recently published relevant meta-analyses [19, 20] demonstrated that both marginal and velamentous cord insertions have been linked to an increased risk of PE, this could be another pathophysiological mechanism explaining the association between bilobate placenta and PE. Of note, abnormal cord insertion has been previously found to be associated with higher mean UtA PI as well [21].

With regards to SGA, no association with bilobate placenta was identified; this is in accordance with the findings by Suzuki et al. who analyzed succenturiate placentas and found no statistically significant difference in the incidence of SGA compared to the controls [3]. We additionally investigated PTD, which was also not associated with the bilobate placenta. Of note, bilobate placenta was associated with a 3 days’ decrease in the mean gestational age at delivery; this is of minor clinical importance, since prematurity rates were not affected.

Following a risk factor analysis, conception via ART, history of previous miscarriage in the first trimester and advancing maternal age were identified as independent risk factors for bilobate placenta. These findings are in accordance with those by Suzuki et al., who identified maternal age above 35 years and the use of in vitro fertilization as possible risk factors [3]. The underlying increasing arterial endothelial damage brought on by aging may be the cause of placental anatomical variations like the bilobate placenta in older women [22, 23]. A systematic review assessed the association between ART and bilobate or succenturiate placenta and found an increased rate of succenturiate lobes in ART vs. non-ART pregnancies (OR: 6.97) [24]. The most plausible theory explaining why ART is an important risk factor, is based on the in vitro formation of the chorion that could affect the formation of the placenta and predispose to morphological variations such as bilobate placenta [25].

To our knowledge, this is the first study investigating the risk factors for bilobate placenta, as well as its associations with pregnancy outcomes. We consider the design and methodology of the present study as its major strength; excluding cases of vasa previa eliminated the possible bias of its coexistence with the bilobate placenta and by including only prenatally diagnosed cases, our study is rendered clinically relevant. The main weakness of the study stems from the limited study sample and statistical power. Thus, our results should be interpreted with caution until larger population studies are conducted to enable the extraction of safer and more robust conclusions. Furthermore, the inclusion of succenturiate placenta cases in the study group may have affected the results. Although conducting a complete-case analysis may be perceived as a weakness of our study, it is worth noting that the proportion of cases with missing values was limited (about 2 %) and no study case had missing values. Additionally, the data were missing due to no reporting of some values or due to loss in follow-up, considering both as missing at random; according to the literature complete-case analysis does not introduce any bias [15]. We should also state that statistical significance level was not adjusted for multiple testing, thus p-value is of no confirmatory value. Finally, the substantial inter-observer and intra-observer variability in the sonographic assessment of the placenta should not be disregarded.