Chapter 8: PRENATAL DIAGNOSIS OF PLACENTA ACCRETA

Placenta accreta presents serious clinical problems that can lead to massive blood loss and even death. Torrential bleeding occurs when there is forceful separation of an abnormally implanted placenta from the uterine wall. Placenta accreta and its variations (placenta increta and percreta) account for 33% to 50% of all emergency peripartum hysterectomies.^{1,2, and 3} The major risk factor, the number of previous cesarean sections, is increasing at a rapid rate; in some hospitals in the United States, cesarean sections have reached over 40% of deliveries. At the University of Chicago, placenta accreta increased 10-fold in the years 1982 to 2002 and was present in 9.3% of women with placenta previa.⁴

Because management of placenta accreta requires substantial resources, such as the ability to transfuse large amounts of blood rapidly, it would be advantageous to make the diagnosis accurately before delivery. There are also new methods of prospective management that can markedly reduce blood loss and other morbidity. Considerable progress has been made in the use of ultrasound to predict its presence in all 3 trimesters. The exact role of magnetic resonance imaging (MRI) has not been defined, but it may be important in the future to evaluate patients in whom suspicion has been raised by ultrasound.

In placenta accreta, part or all of the decidua parietalis between the myometrium and placenta is missing, leaving the trophoblast to lie in direct contact with the myometrium. This lack of decidua means that there is no natural separation plane and the placenta is thus firmly adherent to the myometrium. In placenta increta, there is growth of the trophoblast into the myometrium, and in percreta all the way through it. In placenta percreta, the placenta may also invade the bladder, making removal of the uterus very difficult and often resulting in loss of some of the bladder. Percreta can also involve the broad ligament, cervix, or uterine artery. Usually, this lack of decidua involves only part of the placental myometrial surface, but even a limited area creates a problem if the placenta is forcefully separated from the myometrium. If complete separation of the placenta does not occur, leaving parts attached to the myometrium, the uterus does not contract well: it is uterine contraction that provides hemostasis after delivery.

Placenta accreta, increta, and percreta are known as placental attachment disorders (PADs). In the following discussion, the term placenta accreta refers to all 3 types of placental attachment disorders. The cause is not known because implantation is under the control of many molecular and cellular mechanisms.⁵ However, it seems that implantation in the lower uterine segment near the cervix is problematic because there is a higher incidence of placental attachment disorders in patients with a placenta previa, even without previous surgery.

Placenta previa by itself is a risk factor.^6,7 The risk of placenta accreta in patients with placenta previa, but no previous surgery, is increased over the baseline of 4 in 100,000 and ranges from 2% to 6.3%, depending upon whether maternal age is under or over 35 years.⁶ Wu et al found that the likelihood ratio of placenta previa for PAD was 51.42—by far the highest of any factor studied.⁴

Uterine surgery, such as a cesarean section myomectomy or reconstruction, is also a strong risk factor for placenta accreta in patients with placenta previa.^4,6,7 Silver et al found that in patients with placenta previa, the risk for accreta was 3, 11, 40, and 61 and 67 times the baseline for first, second, third, fourth, and fifth or more repeat cesarean deliveries.⁸ Clark et al found that one prior cesarean section in patients with placenta previa increased the risk to 24% and to 67% with 4 or more cesareans (and is increased even if the placenta does not lie over the uterine scar).⁹ However, the risk increases 5-fold if the placenta actually lies over the scar itself,⁶ again probably an additive factor of lower uterine segment location typical of placenta previa in addition to a surgically disturbed area of myometrium.

Prior placenta accreta is also a risk factor. In a group of patients with presumed placenta accreta due to difficulty removing the placenta, 16% of cases had previous events of placenta accreta.¹⁰ Prior curettage and history of abortion have been mentioned in many publications as risk factors, but in a case controlled series neither of these were actually significant.⁴ Gielchinsky et al found dilatation and curettage (D and C) to be a factor in the less severe cases only; in the more severe cases leading to hysterectomy, increased parity, anterior low placenta, and repeated cesarean sections were the significant risk factors.¹⁰

For some reason maternal age also increases risk in cases of placenta previa. Maternal age over 35 can double or triple baseline risk.⁶

Although most patients with placenta accreta will have a placenta previa, placenta accreta occasionally can be found in patients without placenta previa¹¹ and in patients without previous surgery, the latter often in the posterior or fundal areas. In the series of Gielchinsky et al, 20% were primiparous women without known risk factors.¹⁰

Before the beginning of a scan, ask all patients about previous uterine surgery. Then look at the location of the placenta to determine if it could be over the scar.

Placenta accreta probably starts at the moment of implantation. It has been found at the time of D and C in the first trimester, as early as 7 weeks, most frequently in patients who have had previous uterine surgery.^{12,13,14, and 15}

Low implantation of the sac upon a previous cesarean section scar is the typical first trimester appearance of placenta accreta in such patients. The usual location of the normal gestational sac is in the fundus (Figure 8-1A); only occasionally does a normal sac implant low in the uterus (Figure 8-1B). In a series of pathology proven cases of placenta accreta in which there had been early scans, the sac was found to be implanted in the lower anterior uterine segment in all patients who had had a previous cesarean section (6 of 6) (Figures 8-2 and 8-3).¹⁶ Although the mean thickness of the lower uterine segment in patients with a previous cesarean section is as little as 7 mm,¹⁷ in the reported early accretas there was almost no anterior uterine wall and the wall was seen to dramatically thin just above the sac in most cases. Five of the eight pregnancies in 2 publications on early accretas continued into the third trimester.^16,18 Abnormal vascular lacunae have also been seen in the first trimester.¹⁹ Recently, a mass seen between the uterus and bladder at 2 weeks in one patient and in 4 months in another after a first trimester D and C was found to be a PAD. Upon repeat D and C there was marked bleeding, leading to total abdominal hysterectomies in both. Pathology showed the remains of placenta accreta in one and percreta in the second with degenerating placental tissue that had perforated the area of the cesarean section scar.^20,21 The conclusion is that a mass in the anterior lower uterine segment myometrium after a D and C could be placenta accreta in a patient with a persistent positive human chorionic gonadotropin (HCG) and previous surgery.

A pregnancy can also develop in the actual scar, reaching it through a presumed fistula. These are termed "scar pregnancies" and do not involve the endometrial cavity. They are not included in the present discussion.

Placenta percreta can be suspected in an early pregnancy when there is free fluid in the abdomen in a situation in which the patient is at risk for accreta²² or even without risk factors.²³ Nine to sixteen weeks is the time during which perforation and subsequent rupture seems most frequent,^{24,25, and 26} but ruptures at 20 and 25 weeks have also been reported.^27,28 When an at–risk patient presents with fluid in the abdomen, the most likely diagnosis is perforation (Figure 8-4). Severe abdominal pain in a patient with a previous cesarean section should also suggest this diagnosis.

There are a number of ultrasound findings that have been thought to predict placenta accreta over the years. The most sensitive sign is the presence of linear irregular placental spaces with turbulent flow, usually known as lacunae. The nature of these is not known—sometimes the rate per minute suggests fetal origin, but more often the rate suggests maternal origin.

Vascular Spaces

In a prospective analysis of all patients with previous uterine surgery presenting for obstetrical ultrasound over a 12-year period, most patients who developed placenta accreta had ultrasound findings by the time of their screening or booking exam at 18 weeks.¹¹ The most sensitive marker consisted of vascular lacunae, which are long and thin (Figure 8-5A) echolucent spaces, rather than round (Figure 8-5B), and show blood flow within (Figures 8-6 and 8-7). Placental lacunae or vascular spaces as a finding in placenta accreta have been described in several publications.^{29,30, and 31} Finburg and Williams graded placentas by the number of abnormal vascular spaces and found that the more lacunae there were, the more likely the patient had placenta accreta. This was recently corroborated by Yang et al who graded 51 at–risk cases using a method similar to Finberg and William's.³² Sensitivity for placenta increta or percreta for grade 2 (4 to 6 irregular lacunae) or above was 100%, specificity 97.2%, positive predictive value (PPV) 93.8%, and negative predictive value 100%. Of their cases requiring cesarean hysterectomy all had lacunae, with 2 at grade 1 and all the rest grades 2 to 4. The more lacunae present, the more frequent and the larger were transfusions. If lacunae were present, 23% developed disseminated intravascular coagulation (DIC) and 46% needed intensive care unit (ICU) admission. Finberg and Williams also showed that placenta accreta can occasionally be present with no lacunae, but when there were 4 or more lacunae, placenta accreta was present in 100% of cases.

Clear Zone

Another frequently mentioned sign is thinning or absence of the clear zone between the placenta and myometrium. This clear or echolucent zone between the placenta and uterine wall is thought to represent the decidua parietalis (Figure 8-8). However, it is frequently absent anteriorly in the normal patient,³³ although rarely in the fundal or posterior areas (Figure 8-9). Thus, absence of the clear zone in an anterior placenta previa can be normal. Actually, the absence of this clear zone as the only finding is the most frequent cause of false–positive diagnoses.^11,29 In one large study the PPV of the clear zone was 6% and its sensitivity was 7%.¹¹

The clear zone has not been evaluated in posterior placenta accreta. Since McGahan et al found that the clear space was always seen posteriorly, its absence may be much more significant in a posterior placenta than in an anterior one.

Uterine-Bladder Serosa

Disruption of the usual white line between the bladder and uterus, termed uterine-bladder serosa, has been cited as a very reliable sign of placenta accreta (Figure 8-10). Its disruption does not automatically mean that the placenta has invaded this wall, but is due to enlarged vessels in these spaces. It is an infrequent sign but has been quite specific for a PAD in 2 large studies.^11,29

Diagnosis in the Patient Without Uterine Scarring

Placenta percreta can occur rarely in patients who have had no previous surgery. A spontaneous rupture through the fundus due to placenta percreta has been reported.²⁷ All the major papers have evaluated patients who have had cesarean sections, most of whom also have placenta previa. Does the ultrasound appearance of placentas not implanted over a lower uterine segment or a surgical scar differ from that described in most cases of placenta accreta with a placenta previa? There is as yet no available data.

Diagnosis of Placenta Percreta

It appears that the more lacunae that are present the more likely there is placenta increta or percreta. Yang et al found that all patients with a grade 2 or above had a PAD, and all but 1 of these had placenta increta or percreta. Invasion can occur into the cervix and broad ligaments as well. Ultrasound detection of this has not been reported.

In a blinded study, the addition of color Doppler to gray-scale sonography did not increase diagnostic accuracy.³⁴ Although it is not necessary for the diagnosis, if available it is useful for evaluation of lacunae that may not be typical in appearance, allowing detection of mixed or arterial blood flow within them. There are several papers that speak about the utility of color Doppler studies, but none of them showed increased accuracy over gray-scale diagnosis.^35,36 Abnormal vessels linking the placenta to the bladder were seen in all cases of placenta accreta by Jarparaj et al.³⁵

The place for MRI has not yet been completely defined, but with improved scan times and sequences it will probably be used to determine the extent of invasion after suspicion has been raised by ultrasound. It is doubtful that it will be used as a primary screening tool in the near future. Problems at the present relate primarily to the thinness of the lower uterine segment, to the distinction between the myometrial wall and the placenta, to the lack of pathologic proof, and to the paucity of case controlled and blinded studies. One blinded study showed that no additional information was gained with MRI except in a case with a posterior placenta.³⁴ Gadolinium contrast material will increase the definition of the placenta from the uterine wall,³⁷ but it has not been used in the United States in fetuses because of several animal model reports showing mental retardation. Gadolinium is used in pediatric radiology, however.