Placenta previa in in vitro fertilization and unassisted pregnancies–is there a difference in perinatal outcomes and placental histology?

Abstract

Introduction

In-vitro-fertilization (IVF) is an independent risk factor for placenta previa (PP). Our aim was to study this link by comparing the clinical characteristics and placental histology of pregnancies complicated by PP in IVF versus unassisted pregnancies.

Methods

A retrospective-cohort study of deliveries with PP between 2008 and 2021. Placental histology, obstetric and neonatal outcomes were compared between IVF and unassisted pregnancies. Included, were singleton deliveries complicated by PP at gestational weeks (GA) >24.

Results

A total of 182 pregnancies were included − 23 IVF pregnancies (IVF group) and 159 unassisted pregnancies (Control group). The control group was characterized by higher gravidity (p = .007) and parity (p < .001) and a trend of more past cesarean deliveries, whereas the IVF group- by a higher rate of nulliparity (p < .001) and diabetes mellitus (p = .04). The control group was characterized by a higher rate of placental weight below the 10th percentile (47.8 versus 13.9%, p = .001) and by a trend of a lower overall placental weight. No differences were noted in maternal and fetal vascular lesions.

Discussion

While PP in non-assisted pregnancies is probably associated with previous CDs, in IVF it is more “sporadic,” and may complicate any index pregnancy. A lower placental weight was more prevalent in the control group, supporting the concept that pregnancies complicated by PP following IVF can be attributed to initial abnormal location of placentation, rather than an underlying pathological uterine segment of implantation. Nevertheless, IVF and unassisted pregnancies entail similar perinatal outcomes in cases of PP.

Keywords:

• Placenta previa (PP)
• in vitro fertilization (IVF)
• placental histology
• placentation

Introduction

Placenta previa (PP) is defined as the presence of placental tissue over or adjacent to the cervical os. PP carries with it risks for serious maternal and adverse fetal outcomes, including severe bleeding, prematurity and increased perinatal morbidity and mortality. A complete PP is an absolute indication for cesarean delivery (CD) [1].

PP is more prevalent following in vitro fertilization (IVF) pregnancies as compared with unassisted pregnancies, with an increased risk of three to six-fold according to some authors [1,2]. The etiology for this increased risk in IVF pregnancies is unclear, and may be related to reproductive procedures performed or to subfertility characteristics which have led to IVF [3]. As women choose to postpone childbearing for many reasons, the requirement for IVF continues to increase [4]. Thus, while a growing proportion of pregnancies worldwide are conceived by IVF, possible iatrogenic side effects should be studied. As PP is of clinical importance, it is essential to diagnose it on time, as well as study the mechanisms and risk factors behind it, which could possibly help with its prevention [5].

The objective of our study was to compare the clinical characteristics and placental histology of pregnancies complicated with PP in IVF and unassisted pregnancies. We sought to shed light on the pathogenesis of this increased risk and to study whether the etiology is related to maternal risk factors or placental pathology, through an inspection of the placenta, which may point to an underlying placental process, or to changes which result from of intra uterine insults.

Material and methods

Patients

This was a retrospective cohort study. The medical records of all deliveries beyond 24 weeks of gestation (GA), at a single university-affiliated medical center, over the years 2008–2021, were reviewed. Included, were singleton gestations complicated by all types of PP (complete and marginal) for all of which, placental histopathological evaluations were available. Excluded from the study were pregnancies with lack of prenatal care, pregnancies in which data was missing regarding mode of conception and cases of intrauterine fetal demise. According to our departmental protocol, which is based on accepted guidelines, all asymptomatic pregnancies complicated by PP are delivered by CD at 36–38 GA and prior in cases of recurrent or significant bleeding [6].

The following data were collected from patients’ computerized files: maternal age, body mass index (BMI, kg/m²), gravidity and parity, history of previous CD, smoking, background diseases including diabetes mellitus (DM) and thrombophilia. Obstetric and neonatal outcomes in both groups were documented and included antepartum hemorrhage, need for blood transfusions, development of preeclampsia, GA at delivery and presence of placenta accreta. Neonatal outcomes included umbilical artery pH ≤7.1, need for blood transfusion or need for phototherapy, respiratory distress syndrome, meconium aspiration syndrome, diagnosis of sepsis, need for mechanical ventilation support, necrotizing enterocolitis, intraventricular hemorrhage, hypoxic ischemic encephalopathy, periventricular leukomalacia, seizures, hypoglycemia, hypothermia, and death [7].

Dating of the unassisted pregnancies was accomplished utilizing the day of last menstrual period and was altered if needed according to crown rump length (crl) measurement according to accepted criteria [8].

Small for gestational age was defined as birth weight ≤10th percentile for GA, according to updated local growth scales [9]. Preeclampsia, DM and thrombophilia were diagnosed according to the current American College of Obstetricians and Gynecologists criteria [10–13]. Placenta accreta was defined as an abnormal placental attachment to the uterine wall due to an invasion of the trophoblasts beyond the uteroplacental interface [14].

For the purpose of the study, maternal demographics, pregnancy outcomes and placental histopathological lesions were compared between IVF and unassisted pregnancies.

Approval for the study was obtained from the Edith Wolfson Institutional Review Board Committee, approval number # 0282-20-WOMC, dated November 1st, 2020.

Placental examination

According to our departmental protocol, all placentas of complicated pregnancies including preeclampsia, diabetes, PP and more are sent to histopathological evaluation. Placental pathology examinations were performed using our standard protocol, by a single pathologist (author L.S). At least five tissue samples from each placenta were embedded in paraffin blocks for histo-pathological assessment, which was classified according to the “Amsterdam criteria” [15,16] into four groups, as previously reported by our group [17]: maternal vascular malperfusion lesions, fetal vascular malperfusion lesions, and inflammatory lesions–either chorioamnionitis (defined as any maternal or fetal inflammatory response) or villitis of unknown cause. Composite maternal vascular malperfusion lesions was defined as the presence of one or more maternal vascular supply abnormalities and composite fetal vascular malperfusion lesions as the presence of one or more fetal vascular supply abnormalities. Placental weight percentile was defined according to population-based nomograms [18].

Statistical analysis

Data were analyzed with Epi Info, version 7.0 (Centers for Disease Control and Prevention, Atlanta, GA). Continuous variables were calculated as mean ± SD or median (range) as appropriate and compared using the student’s t-test . Categorical variables were calculated as number (percentage) and compared with chi-squared or Fisher’s exact test as appropriate. All tests were two tailed, and the threshold for statistical significance was defined as p < .05.

To calculate the sample size required, we assumed a higher rate of maternal vascular malperfusion lesions which would be found in pregnancies complicated by PP following IVF. As the maternal vascular malperfusion lesion rate among PP pregnancies was previously reported to be 40% [19], we assumed this rate constitutes an average between a 50% rate among IVF patients as compared with 20% in unassisted pregnancies. We assumed a 1:10 ratio between the groups, thus we calculated 17 IVF cases and 169 unassisted pregnancies as sufficient, with 80% power and an alpha of 0.05.

Results

During the study period, 245 deliveries with a diagnosis of PP took place at our institution, for 182 of which placental pathology reports were available, and were included in the study. A total of 23 pregnancies were following IVF, and 159 following unassisted pregnancies.

Patient demographics are presented in Table 1. IVF pregnancies were of lower gravidity and parity compared to unassisted pregnancies (p = .007, p < .001, respectively). Consequentially, the rate of nulliparous patients was higher in the IVF group (47.8 vs. 17.0%, p < .001). In addition, there were more cases of DM in IVF pregnancies (21.7 vs. 8.1%, p = .04). There were no differences between the groups with regards to maternal age, BMI, smoking and a diagnosis of thrombophilia. There was a higher rate of CDs among the control group, but this did not reach statistical significance (38.9 vs. 21.7%, p = .10).

Table 1. Demographic characteristics of IVF and control groups.

	IVF group n = 23	Control group n = 159	p
Age, (years), mean ± SD	34.4 ± 4.8	32.6 ± 4.8	.10
BMI (kg/m^2), mean ± SD	24.6 ± 4.9	23.6 ± 4.3	.37
Gravidity, median (range)	2 (1-6)	3 (1-12)	.007
Parity, median (range)	0 (0-2)	2 (0-7)	<.001
Nulliparous, n (%)	11 (47.8)	27 (17.0)	<.001
Previous CD, n (%)	5 (21.7)	62 (38.9)	.10
Previous previa, n (%)	1 (4.3)	5 (3.1)	.56
DM, n (%)	5 (21.7)	13 (8.1)	.04
Smokers, n (%)	3 (13.0)	35/158(22.1)	.41
Thrombophilia, n (%)	1 (4.3)	5 (3.1)	.56

IVF: in vitro fertilization; BMI: body mass index, pre-gestational; CD: cesarean delivery; DM: diabetes mellitus, pre-gestational or gestational.

There were no differences between the IVF and the control groups with regards to obstetric and neonatal outcomes, as depicted in Table 2 including GA at delivery, incidence of placenta accreta, antepartum hemorrhage and need for blood products. There was a trend toward a higher incidence of preeclampsia in the IVF group compared with the control group (p = .07). Neonatal outcome was similar between the groups including birth weights, Apgar scores, cord pH and adverse composite neonatal index.

Table 2. Obstetric and neonatal outcome in the IVF and control groups.

	IVF group n = 23	Control group n = 159	p
Bleeding during pregnancy, n (%)	14 (60.8)	90 (56.6)	.69
Betamethasone treatment during pregnancy, n (%)	6 (26.0)	58 (36.4)	.32
Preeclampsia, n (%)	2 (8.7)	2 (1.2)	.07
Gestational age, (weeks), mean ± SD	36.0 ± 2.8	35.9 ± 2.8	.81
Premature delivery, n (%)	11 (47.8)	76 (47.8)	>.99
Placenta accrete, n (%)	0	14 (8.8)	.22
Maternal blood transfusion, n (%)	3 (13.0)	47 (29.5)	.13
Maternal hemoglobin post CD, (g/dL), mean ± SD	9.7 ± 1.2	9.2 ± 1.5	.10
Birth weight, (grams), means ± SD	2731 ± 662	2607 ± 624	.37
Small for gestational age, n (%)	2 (8.7)	19 (11.9)	>.99
Five-minute Apgar score <7, n (%)	1 (4.3)	5/158 (3.1)	.56
Cord pH <7.1, n (%)	0/20	3/147 (2.0)	>.99
Adverse composite neonatal index, n (%)	9 (39.1)	61 (38.3)	.94

IVF: in vitro fertilization, premature delivery–under 37 weeks, adverse composite neonatal index was defined as at least one of the following: phototherapy, respiratory distress, sepsis, blood transfusion, cerebral complications, necrotizing enterocolitis (NEC), neonatal intensive care unit (NICU) admission or neonatal death.

Table 3 describes the placental findings in both groups. The placentas of the control group were characterized by a higher rate of placental weights below the 10th percentile (47.8 vs. 13.0%, respectively, p = .001). There were no differences between the groups in the incidence of maternal or fetal vascular malperfusion lesions, or inflammatory lesions.

Table 3. Placental findings in the IVF and control groups.

	IVF group n = 23	Control group n = 159	p
Placental weight, (grams), mean ± SD	496 ± 135	450 ± 118	.09
Placental weight <10th percentile, n (%)	3 (13.0)	76 (47.8)	.001
Macroscopic abruption, n (%)	2 (8.7)	12/157 (7.6)	.69
Maternal vascular malperfusion (MVM) lesions
Composite MVM lesions, n (%)	6 (26.0)	57 (35.8)	.35
Marginal/retroplacental hematoma, n (%)	0	13 (8.1)	.37
Vascular lesions	2 (8.7)	11 (6.9)	.67
Villous changes	5 (21.7)	41 (25.7)	.67
Fetal vascular malperfusion (FVM) lesions
Composite FVM lesions, n (%)	2 (8.7)	27 (16.9)	.54
Vascular lesions	1 (4.3)	19 (11.9)	.47
Villous changes	1 (4.3)	11 (6.9)	>.99
Inflammatory lesions
VUE, n (%)	1 (4.3)	7 (4.4)	>.99
Chorioamnionitis, n (%)	0	1 (0.6)	>.99

IVF: in vitro fertilization; VUE: villitis of unknown etiology, chorioamnionitis–maternal and/or fetal inflammatory responses.

Maternal vascular malperfusion lesions – marginal and retro-placental hematoma; vascular changes–acute atherosis and mural hypertrophy; villous changes–increased syncytial knots, villous agglutination, increased intervillous fibrin deposition, distal villous hypoplasia, and parenchymal infarcts.

Fetal vascular malperfusion lesions – Vascular lesions–thrombosis of the chorionic plate and stem villous vessels; villous changes–villous stromal-vascular karyorrhexis and avascular villi.

Discussion

The development of the placenta is a complex process which is susceptible to epigenetic changes including environmental factors and assisted reproductive technologies (ART), mainly IVF [20]. Bearing in mind a possible link of IVF to primary placental anomalies of implantation, and that IVF is likely to become increasingly common as women delay childbearing, physicians are required to monitor these conditions and be familiarized with the patho-mechanism behind their occurrence. In our study the prevalence of PP was 3.7% which is in concordance with its known prevalence [21].

The exact mechanism by which IVF increases the risk for PP is uncertain [22,23]. Some reports have attributed maternal risk factors to the higher occurrence. Faiz et al. have demonstrated an increased risk with advanced maternal age, parity, prior CD and abortions [24], while Shobeiri et al. attributed the increased risk of PP to maternal smoking regardless to the need of ART [25]. Our findings indicate that the explanation behind the increased risk of PP in ART treatments lies on abnormal placentation or specifically–an abnormally located placentation process. Taken together, these hypotheses can indicate both ART and maternal risk factors may act in a synergic manner and bring rise to the development of PP.

In a previous large scale sibling cohort [23], a comparison of the incidence of PP between consecutive pregnancies among 1,349 women with both unassisted and IVF pregnancies found a 3-fold higher risk associated with IVF, which caused the authors to conclude a direct relation between reproductive technologies and the occurrence of PP. In addition, Tsutsumi et al. studied pregnancy outcomes of infertile patients and compared those who required ART and those who eventually conceived without assistance. They found a higher prevalence of PP among the former and concluded that ART is an independent risk factor for the development of PP [26].

As PP remains a major obstetrical concern with ART as an obvious risk factor, we sought to further examine the etiologies of PP and the link between IVF and its occurrence [2]. To our knowledge, our study considers a novel aspect of the investigation between ART and PP with respect to the additional utilization of placental pathology. As it has been well established that the placenta remains “the diary of intrauterine life,” it remains an intriguing field of study and we feel it has shed some light on the link between ART and PP [16].

In our study, we found a lower incidence of small placentas among the unassisted pregnancies compared to IVF pregnancies. This may imply that the mechanism of the increased incidence of PP among IVF-pregnancies may relate to an abnormally located placentation process, whereas unassisted pregnancies were more likely associated with a defective uterine segment and parity, associated with smaller placentas. This hypothesis is further supported by the fact that there were no differences in neonatal outcomes including birth weight.

Our groups were similar with regards to demographics. The number of previous CDs between the groups was similar but this could be attributed to a small sample size. In addition, there were no differences between the groups with regards to the incidence of placenta accreta. It can be postulated that if the groups were larger, a significantly higher incidence of the previous CDs would have been witnessed among unassisted pregnancies, and accordingly, a higher incidence of placenta accrete.

The strength of this study lies in the fact that we were able to include placental histopathology to shed more light on the pathogenesis for PP and better understand the link between this important and often life-threatening condition and ART. In addition, we excluded twin gestations and focus solely on singleton gestations, as there are reports of an association between multiple gestations and PP [23,27]. Finally, our cohort originates from a single institutional registry, so that elaborate coding and access to patient files was available. Placental examinations were performed by a single pathologist (author L.S.), unaware of mode of conception.

Our study is not without limitations. Our major limitation was our small sample size in the IVF group. Since each case of the IVF group had to fulfill 3 conditions to be included- being IVF patients who gave birth at our center and for whom we had their placental histopathological reports available, this made a unique study group a consequentially- made a relatively small sample size. And despite having a power calculation, larger study groups would have possibly allowed for the demonstration of additional differences in outcomes, including previous CD and placenta accreta. Moreover, this limitation prevented us from matching against possible cofounders.

Secondly, we did not have information regarding the specific reproductive technology utilized in our IVF-group as most patients did not undergo IVF at our institution, and consequently cannot shed light on the role of specific aspects of ART and the occurrence of PP.

In conclusion, our study might suggest that PP following IVF most probably results from a sporadic abnormally placed placentation processs. Our findings are best reaffirmed with a larger study group, also accounting for specific ART treatment characteristics.

Ethical approval

Approval for the study was obtained from the Edith Wolfson Institutional Review Board Committee, approval number # 0282-20-WOMC, dated November 1st, 2020.

Acknowledgment

No funding was required.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author, [EB], upon reasonable request.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.