Placenta Previa and Accreta, Vasa Previa, Subchorionic Hemorrhage, and Abruptio Placentae

Definition and Epidemiology

Advances in the precision of sonographic diagnosis, particularly transvaginal ultrasound (TVUS) imaging, and increased understanding of the changing relationship between the placenta and the internal cervical os as pregnancy advances have rendered traditional definitions and classifications of placenta previa obsolete. A clinically significant abnormality of placental position occurs when the placenta covers the cervix completely or extends close enough to the cervix to cause bleeding when the cervix dilates or the lower uterine segment effaces. TVUS allows for localization of the placenta in relation to the internal cervical os with great precision. If the placenta is more than 2 cm from the cervix, there is no increased risk of bleeding. A definitive diagnosis of placenta previa should be avoided in asymptomatic patients before the third trimester, because many cases of placenta previa identified early in pregnancy resolve as pregnancy advances. The recent National Institute of Child Health and Human Development (NICHD) panel on fetal imaging recommended the replacement of all prior definitions with two terms: placenta previa and low-lying placenta . For pregnancies greater than 16 weeks, if the placental edge is 2 cm or more from the internal os, the placental location should be reported as normal. If the placental edge is less than 2 cm from the internal os but does not cover it, the placenta should be reported as low-lying and the distance reported. If the placental edge covers the internal os, it should be labeled as a placenta previa. Both low-lying placenta and placenta previa should have follow-up sonography at 32 weeks’ gestation. If the placenta remains low lying (<2 cm from the internal os) or continues to cover the internal os (placenta previa), follow-up TVUS at 36 weeks’ gestation is recommended. These recommendations apply only to asymptomatic women without bleeding.

Placenta previa affects about 1 (0.5%) of every 200 pregnancies at term. There is some evidence that the incidence of placenta previa is increasing. This increase may be related to the increasing rate of cesarean delivery observed in all developed countries. A single prior cesarean delivery or a prior pregnancy complicated by placenta previa increases the incidence of placenta previa in a subsequent pregnancy to as high as 5%, ^{, ,} and the risk increases further with a history of more than one prior cesarean delivery. Advanced maternal age increases the incidence of placenta previa to 2% after 35 years of age and 5% after age 40. Multiparity, prior suction curettage, and smoking are all associated with higher risks of placenta previa. The relative risks for these associated factors are summarized in Table 43.1 .

Pathogenesis

The underlying cause of placenta previa is unknown. There is a clear association between placental implantation in the lower uterine segment and prior endometrial damage and uterine scarring from curettage, surgical insult, prior placenta previa, or multiple prior pregnancies.

At least 90% of placentas identified as being low-lying or previa in early pregnancy ultimately resolve by the third trimester. The term placental migration is widely used to describe this phenomenon. The placenta clearly does not move, however; rather, it is likely that the placenta grows toward the better blood supply at the fundus (a process known as trophotropism), leaving the distal portions of the placenta, closer to the relatively poor blood supply of the lower segment, to regress and atrophy. As the uterus grows and expands to accommodate the developing fetus, there is differential growth of the lower uterine segment, and this may further increase the distance between the lower edge of the placenta and the cervix.

Bleeding from placenta previa may occur before labor as a result of development of the lower uterine segment and effacement of the cervix with advancing gestation. Prelabor uterine contractions may also produce bleeding, as may intercourse or injudicious vaginal examination. Once labor begins, significant bleeding will occur as the cervix dilates and the placenta is forced to separate from the underlying decidua.

Diagnosis

The classic history for placenta previa is that of painless third-trimester bleeding. Several small “herald bleeds” may occur in advance of major hemorrhage, but in up to 10% of cases there is no bleeding until the onset of labor. Some women experience pain secondary to uterine contractions. Bleeding may be provoked by labor, examination, or intercourse, but it usually has no identifiable precipitating cause. The patient is more likely to have a fetus with an abnormal lie because the placenta previa may prevent the fetus from establishing normal polarity. All women presenting with painless vaginal bleeding after 20 weeks’ gestation should be assumed to have a placenta previa until proven otherwise. Transabdominal ultrasound should be quickly utilized to screen for placenta previa. Unless the placenta is clearly fundal and the lower segment is clear, TVUS should then be performed. Transabdominal ultrasound has been shown to be inferior to TVUS for definitive placental localization. ^, Concerns regarding the potential for TVUS to provoke bleeding are unfounded, and several studies have confirmed the safety of a careful TVUS approach ^{, ,} ( Fig. 43.1 ). The placement of the transvaginal probe should be observed continuously on the ultrasound monitor during insertion, to avoid placing the probe into a potentially dilated cervix. If a transvaginal probe is unavailable, translabial imaging using a regular abdominal probe can produce excellent results, with better visualization of the relationship between the cervix and placenta than is obtained from transabdominal scanning. A digital or speculum examination to inspect the cervix for local causes of bleeding should not be performed until placenta previa has been excluded by ultrasonography.

In the unusual setting of significant late-pregnancy bleeding where ultrasound is not available and the diagnosis is not clear, there is still a place for the “double-setup” examination. The patient is taken to the operating room and preparations are made for a cesarean delivery. A vaginal examination is then performed, beginning in the vaginal fornices and avoiding placing the fingers directly in the cervix. If a placenta previa is detected, cesarean delivery is performed. If no placenta previa is found, a search for other causes of third-trimester bleeding ensues.

Management

Any woman with vaginal bleeding after 20 weeks’ gestation should be assessed on a labor and delivery unit. The primary focus should be on hemodynamic assessment of the mother and assessment of fetal well-being. Vital signs are obtained, and electronic fetal monitoring is initiated. One or two large-gauge intravenous lines should be placed, and maternal blood should be sent for determination of the hematocrit and type and screen. For substantial bleeding episodes, 2 to 4 units of blood should be cross-matched. Obstetric units might consider the use of an “obstetric hemorrhage protocol” to facilitate access to the resources of the hospital blood bank for this and any other obstetric hemorrhage ( Box 43.1 ). Rh immune globulin is administered, when appropriate, to Rh-negative, nonimmunized women.

Once the patient is stable and fetal condition has been assessed, the definitive cause of the bleeding can be addressed. If the diagnosis is clearly placenta previa and the patient is at or beyond 36 weeks’ gestation, delivery is appropriate. If bleeding is excessive or continues, or if there are concerns about the condition of the fetus, the patient should be delivered regardless of gestational age. In all other cases, management may be conservative and has been shown to be safe, ^, with prolongation of pregnancy by an average of 4 weeks after the initial bleeding episode. The closer the gestational age to 36 weeks, the less likely it is that a significant prolongation of pregnancy will be gained. Betamethasone to enhance fetal lung maturation should be administered to patients who are at less than 34 weeks’ gestation if expectant management is planned.

There is controversy regarding the role of tocolytics in the setting of hemorrhage from placenta previa. Both β-mimetics and magnesium sulfate have been used in this setting and appear to be associated with significant prolongation of pregnancy without adverse effects.

After the initial presentation with bleeding, patients should remain in the hospital until they have been free of bleeding for at least 48 hours. Some may then be considered for home management. Several studies have addressed the issue of safety of outpatient management in a controlled setting at home. ^, With the exception of one report of an increase in perinatal mortality and morbidity and earlier gestational age at birth, it appears to be a safe approach. Patients selected for home management should be asymptomatic with regard to bleeding and abdominal pain, should be able to remain at home with limited activity, and should have adequate support as well as adequate access to transport to a nearby hospital if bleeding recurs. A second significant bleeding episode usually results in readmission until delivery.

Several strategies have been proposed to reduce the risk of hemorrhage in women with a known placenta previa. Bed rest, reduced activity, and avoidance of intercourse are commonly mandated and seem logical, although there is no conclusive evidence to support these measures. Cervical cerclage was evaluated in two small prospective studies ^, without clear evidence of benefit and is not recommended.

All women whose placenta lies within 2 cm of the cervix, as documented by a late third-trimester TVUS scan, should be delivered by cesarean section. An asymptomatic woman whose placenta lies more than 2 cm from the cervical os can be allowed to labor safely. The presence of a low-lying placenta, even if it does not cause intrapartum bleeding, increases the risk of postpartum hemorrhage because of lower uterine segment atony.

Cesarean delivery for placenta previa should be performed by the most experienced team available because of the substantial risk of intraoperative hemorrhage. In most instances, a lower uterine segment incision is appropriate. If the placenta is anterior, it is necessary to clamp the umbilical cord immediately to prevent excessive blood loss caused by disruption of the placenta during entry. A vertical incision is also reasonable in such cases and may be preferable if the fetus is premature or if a transverse lie exists. Postpartum hemorrhage may occur from the placental implantation site secondary to atony and may require the use of additional pharmacologic agents to control blood loss, such as methylergonovine maleate (Methergine); carboprost tromethamine (Hemabate), also known as 15-methyl prostaglandin F _2α ; and high-dose oxytocin, used either singly or in combination. The B-Lynch suture or local suturing of the placental bed may be needed to control bleeding; the Bakri Postpartum Balloon (Cook Medical, Bloomington, IN) may also be useful in the control of postpartum hemorrhage following cesarean section for placenta previa. In rare cases of refractory hemorrhage, hysterectomy may be required.

Among women known to have a placenta previa who do not require very early delivery, elective delivery should be performed before significant bleeding has occurred. It is reasonable to plan on delivery at or just after 36 weeks’ gestation, because there is little fetal advantage after that time when weighed against the risk of a sudden and possibly excessive bleeding episode. Amniocentesis to confirm lung maturity before delivery is not needed.

The selection of anesthesia to be used for cesarean delivery in cases of placenta previa should be decided by the obstetrician and anesthesiologist involved with the delivery, in concert with the patient. In the United Kingdom, regional anesthesia was preferred by most obstetric anesthesiologists in a survey and was used in 60% of cases in a retrospective series. Regional anesthesia is associated with lower operative blood loss and less need for transfusion than general anesthesia, ^, probably because many inhaled anesthetics cause uterine relaxation.

Placenta Accreta Spectrum

Placenta accreta occurs when trophoblast is attached to the myometrium without intervening decidua; that is, invasion beyond the normal boundary established by the Nitabuch fibrinoid layer. If invasion extends into the myometrium, the term placenta increta is used; placental invasion beyond the uterine serosa (at times involving the bladder or other pelvic organs and vessels) is termed placenta percreta. In this chapter, placenta accreta will be used to refer to the entire spectrum unless otherwise specified. The term morbidly adherent placenta is also used to describe placenta accreta. Histologic examples of normal placental implantation and placenta accreta are shown in Fig. 43.2 .

The most important risk factors for placenta accreta are placenta previa and prior cesarean delivery, especially multiple cesareans. ^, Indeed, the risk exceeds 60% in women with previa and three or more cesareans. Prevalence appears to be similar in women with these risk factors who are undergoing second-trimester pregnancy termination. Other risk factors include other uterine surgeries (e.g., endometrial ablation, myomectomy, uterine curettage), pelvic radiation, in vitro fertilization, short interpregnancy interval, increased maternal age, smoking, and parity.

The diagnosis of placental invasion of the myometrium can often be made by ultrasound, with a reported sensitivity and specificity for the diagnosis of approximately 80% to 90%. However, most published studies are prone to bias and may overestimate the accuracy of ultrasound. In a study wherein “experts” were blinded to clinical status, ultrasound had a sensitivity of 55%, specificity of 88%, and accuracy of 65%. Also, there was considerable variation among experienced clinicians in the prediction of accreta, underscoring the fact that features of accreta are often present in normal placentas.

The ultrasound appearance of a normal placental implantation site is shown in Fig. 43.3A . Normal attachment is characterized by a homogeneous appearance of the placenta and a hypoechoic boundary between the placenta and the bladder that represents the myometrium and the normal retroplacental myometrial vasculature. The bladder wall is intact throughout. In contrast, placental accreta is associated with loss of the normal hypoechoic boundary, and there are usually intraplacental sonolucent spaces adjacent to the involved uterine wall ( Fig. 43.3B ). Color flow and power Doppler sonography can also facilitate the diagnosis. Turbulent flow in placental lacunar spaces and increased vascularity in the placenta-uterus interface are all characteristic of accreta. The same sonographic features may be identified as early as the first trimester ( Fig. 43.4 ). First-trimester diagnosis is particularly accurate in the setting of cesarean scar ectopic pregnancy. All women with clinical risk factors for accreta should be referred for obstetric sonography in a unit with appropriate expertise in the diagnosis of accreta.

Magnetic resonance imaging (MRI) has also been used to confirm the diagnosis or to better delineate the presence or extent of accreta. ^, MRI has been reported to have high accuracy for the diagnosis of placenta accreta. However, studies are prone to even more bias than those for ultrasound, and it is uncertain that MRI provides accuracy beyond that achieved by ultrasound alone. It is speculated to be most useful in the presence of a posterior placenta and in the assessment of deep myometrial, parametrial, and bladder involvement. The use of gadolinium may improve the ability of MRI to diagnose placenta accreta. However, a large Canadian study noted an increased risk of a broad set of rheumatologic, inflammatory, or infiltrative skin conditions and for stillbirth or neonatal death after exposure to gadolinium-enhanced MRI during pregnancy. The same study found no risk of any adverse outcome after exposure to MRI without gadolinium during pregnancy, even in the first trimester.

The preferred and generally recommended treatment for placenta accreta is a cesarean delivery and hysterectomy. It is important, if at all possible, to make the diagnosis before delivery, because intraoperative hemorrhage can be massive, and placenta accreta has been reported to be the most common indication for emergency peripartum hysterectomy. ^, There is now ample evidence that predelivery diagnosis of placenta accreta, combined with elective delivery under optimal circumstances and care provided by a multidisciplinary team, markedly reduces maternal morbidity and intraoperative blood loss. The timing of elective delivery (34 to 37 weeks’ gestation) has been somewhat controversial. We plan for delivery at 34 to 35 weeks after maternal corticosteroid administration. With this approach, our neonatal outcomes have been favorable. Delivery is performed without amniotic fluid confirmation of fetal lung maturity. This strategy is supported by a decision analysis comparing various gestational ages with and without amniocentesis before delivery.

In addition, delivery should be accomplished in a center with appropriate multidisciplinary expertise and experience in the care of placenta accreta. Outcomes are improved and the chances of complications diminished when accretas are managed at regional referral centers or “centers of excellence.”

In an effort to diminish blood loss, it is recommended that delivery be accomplished through a fundal incision (avoiding the placenta) followed by clamping of the cord. The placenta is allowed to remain in situ while the surgeons proceed to a total abdominal hysterectomy. This may require very complex surgical technique and planning, and a pelvic surgeon capable of wide resection of the lower uterine segment and parametrial areas should be available, as should ample transfusion capability.

Although published reports are not extensive, it has been suggested that balloon occlusion of the aorta or internal iliac vessels may help to prevent excessive blood loss during resection of the lower uterine segment. This involves preoperative placement of balloon-tipped catheters retrograde through the femoral arteries immediately before surgery. The catheters are guided under fluoroscopic direction into the internal iliac arteries and inflated during the dissection. However, the value and safety of this approach have been challenged since reports of no proven benefit and embolic complications were published. ^, The utility of balloon-tipped catheters remains unclear because the timing of deployment during surgery has been highly variable and inconsistent. A recent comprehensive literature review concluded that there is a need for larger studies or, preferably, a randomized trial to provide adequate answers to this controversial issue.

On the surface, conservative management seems like an attractive option because it avoids a morbid surgery and may lead to preserved fertility. However, safety remains uncertain. It is appealing if there is suspicion of a small focal accreta, if there is a fundal location after a myomectomy or classic cesarean delivery, or with a posteriorly implanted placenta. Several case series note good outcomes after leaving the uterus and placenta in situ (delivery via fundal hysterotomy). Most women are treated with uterine artery embolization postoperatively. Sentilhes and colleagues reported on a cohort of 167 women with accreta in whom conservative management was planned before delivery. Eighteen required a primary hysterectomy for intraoperative bleeding, and 18 underwent a delayed hysterectomy. Severe morbidity occurred in 10 women, and there was one death due to complications of methotrexate therapy. Due to risks as well as lack of proven efficacy, methotrexate is unequivocally not advised for conservative management of accreta. Follow-up was available for 35 women, of whom 14 desired a subsequent pregnancy. Of the 15 pregnancies achieved, there were six first-trimester losses, four preterm infants, and two recurrent placenta accretas. Although these outcomes are acceptable to some women, it is impossible to be sure that they all had true accretas because there were no specimens to assess histologically. Accordingly, the risk in “true” accretas may be considerably higher than that reported in these studies. At present, conservative management should be undertaken with extreme caution and appropriate counseling regarding risks.