Key Abbreviations
Cesaean delivery scar defect CDSD
Magnetic resonance imaging MRI
Odds ratio OR
Placenta accreta PA
Postpartum hemorrhage PPH
Randomized controlled trial RCT
Trial of labor after cesarean TOLAC
Placenta accreta (PA) refers to different grades of morbid placental attachment to the uterine wall secondary to invasion of the trophoblast into the myometrium beyond the uteroplacental interface. The term placenta increta is used to describe deep myometrial invasion by trophoblast villi, and placenta percreta refers to villi perforating through the full thickness of the myometrium and uterine serosa with possible involvement of adjacent organs. The difference between placenta accreta, increta, and percreta is related to the extent of invasion, and the umbrella term disorders of invasive placentation is used to encompass all three ( Fig. 21-1 ).
PA has also been categorized as total, partial, or focal depending on the amount of placental tissue involved. This subclassification has been used less frequently because microscopic examination of the hysterectomy specimen is rarely complete, and attempts at manual removal often distort the placental anatomy. However, with increasing numbers of cases of PA being diagnosed antenatally, and with greater expertise in conservative therapeutic procedures, this classification might become more useful.
The first detailed description of a PA appeared within 20 years of major changes in the surgical techniques for, and more common use of, cesarean delivery (CD). It was not reported by the prominent pathologists of the eighteenth and nineteenth centuries, which may indicate the direct relationship between the increased use of CD and morbid placental adherence. Although PA remains a relatively uncommon obstetric pathology, it has become a complication that an obstetrician with a practice of average volume is likely to encounter more commonly, compared with the once or twice in a lifetime encounter of two decades ago. PA has also become a major cause of life-threatening obstetric complications worldwide.
When PA is present, the failure of the entire placenta to separate normally from the uterine wall after delivery is typically accompanied by severe postpartum hemorrhage (PPH; see Chapter 18 ). Attempts to remove the adherent tissue may provoke further bleeding and a cascade of ongoing hemorrhage, shock, and coagulation disorders that require complex clinical management ( Box 21-1 ). This chapter reviews the pathogenesis, epidemiology, antenatal diagnosis, and management of this obstetric disease.
- 1.
Placenta does not separate and is expelled spontaneously, fully, or partially after delivery of the baby.
- 2.
Attempt at removal of the placenta leads to brisk hemorrhage.
- 3.
Histopathology confirms myometrial fibers in apposition with trophoblasts at the suspected site of invasive placentation if partial or complete excision of the uterus has been performed.
Pathogenesis
Unlike many other placental disorders—such as hydatidiform moles, which have been known for centuries—PA was only recently described, in 1937 by Irving and Hertig. Before the nineteenth century, few successful attempts had been made at performing CDs. CD was a surgical procedure of last resort, and the operation was essentially performed to save the baby’s life. It is only when surgeons started to remove and subsequently to suture the uterus after delivery that the maternal death rate following CD started to improve. By the 1920s, the combination of surgical improvements by Kehrer (uterine closure), Pfannenstiel (suprapubic transverse entry), and Munro Kerr (lower segment uterine incision) and developments in anesthesia and microbiology had brought down the maternal mortality from 70% in the 1850s to less than 10%. The introduction of uterotonic agents, blood transfusions, and antibiotics just before and after the Second World War further reduced the complication rate of CD. These changes made the technique safer and allowed mothers not only to survive the surgical procedure but also to have one or more subsequent pregnancies (see Chapter 19 ).
Total or partial absence of decidua is the characteristic histologic feature of PA and is relatively clear cut in cases of implantation on a uterine scar. This results in the absence of the normal plane of cleavage above the decidua basalis, thus preventing placental separation after delivery. Myometrium does not actually heal by regenerating muscle fibers, rather it forms “foreign” substances, including collagen. The resulting fibrous tissue is weaker, less elastic, and more prone to injury than the original muscle. Myofiber disarray, tissue edema, inflammation, and elastosis have all been observed in uterine wound healing after surgery. Experiments in mice have indicated that differences in regenerative ability translate into histologic, proliferative, and functional differences in biomechanical properties of the scarred myometrium after CD.
Several concepts have been proposed to explain the abnormal placentation in PA including (1) a primary defect of trophoblast function; (2) a secondary decidua basalis defect as the result of a failure of normal decidualization; and, more recently, (3) abnormal vascularization and tissue oxygenation of the scar area. The strongest risk factor for placenta previa is a prior caesarean delivery, which indicates that a failure of decidualization in the area of a previous uterine scar can have an impact on both implantation and placentation. These findings suggest that the decidual defect following a uterine scar may have an adverse effect on early implantation by creating conditions for preferential attachment of the blastocyst to scar tissue and facilitating abnormally deep invasion of the extravillous trophoblast.
Comparison of the ultrasound features of the uterine cesarean scar with histologic findings has shown that large and deep myometrial defects are often associated with absence of reepithelialization of the scar area. These findings support the concept of a primary defect in PA, exposing the deep vasculature of the myometrium below the junctional zone to the migrating trophoblast. The loss of this normal plane of cleavage and the excessive vascular remodeling of the radial and arcuate arteries can explain the antenatal findings and the clinical consequences of PA. A recent study of the uterine circulation in women with a previous CD has shown that uterine artery resistance is increased, and the volume of uterine blood flow is decreased, as a fraction of maternal cardiac output compared with women with a previous vaginal birth. These data suggest a possible relationship between a poorly vascularized uterine scar area and an increase in the resistance to uterine blood flow with a secondary impact on reepithelialization of the scar area and subsequent decidualization. Overall, these data support the concept that abnormal decidualization and deep invasion of the placental bed by the trophoblast in PA are often secondary to the presence of a myometrial scar.
Fragments of myometrium are found in the products of conception in approximately one third of uterine curettages after miscarriage. The fact that it does not have a clear correlation with clinical PA in a subsequent pregnancy suggests that the trauma to the myometrium and the surface of the endometrial damage is often limited in a curettage procedure compared with that of a CD or other major uterine surgery. By contrast, severely deficient uterine scars with complete loss of large areas of the myometrium could explain rare reports of placenta percreta leading to uterine rupture in the first half of pregnancy. Although this is an extremely rare complication of placentation, the mechanism of uterine rupture due to a placenta percreta is likely to be similar to that of a tubal rupture in an ectopic placentation. These findings emphasize the pivotal role of the superficial myometrium in modulating normal uterine placentation.
Epidemiology
Any primary or secondary alterations of the uterine endometrial-myometrial integrity have been associated with the development of PA. With the rapid increase in the incidence of CD over the last several decades, this procedure has become associated with most cases of PA, whereas other factors are now responsible for a relatively small proportion of PA ( Box 21-2 ). CD is now the most commonly performed major operation around the world, with more than 1 million procedures performed each year in the United States alone (see Chapter 19 ). Not surprisingly, a substantial increase has been seen in the occurrence of PA over the last 50 years with as much as a tenfold rise in its incidence in most Western countries. Studies in the United States have indicated an overall incidence of PA of approximately 1 in 533 deliveries. A recent meta-analysis has shown a calculated summary odds ratio (OR) of 1.96 for PA after one CD. A recent case-control study has shown that, compared with primary intrapartum CD, primary prelabor CD significantly increased the risk of PA in a subsequent pregnancy in the presence of placenta previa. Overall, the data from epidemiologic studies consistently indicate that prior CD is the most important factor associated with subsequent PA and that the risk of PA increases with the number of prior CDs .
Primary Uterine Pathology
Major uterine anomalies
Adenomyosis
Submucous uterine fibroids
Myotonic dystrophy
Secondary Uterine Pathology
Cesarean delivery
Uterine curettage
Manual removal of the placenta
Cavity-entering myomectomy
Hysteroscopic surgery (endometrial resection)
In vitro fertilization procedures
Uterine artery embolization
Chemotherapy and radiotherapy
Placenta Previa With Accreta
Epidemiologic studies have also indicated that CD is associated with an increased risk of placenta previa (see Chapter 18 ) in subsequent pregnancies. A recent meta-analysis of five cohort and 11 case-control studies published between 1990 and 2011 revealed that after a CD, the calculated summary OR is 1.47 for placenta previa. The risk of previa is higher with an increasing number of prior CDs. Following a single CD, a 50% increase is seen in the risk of placenta previa in a subsequent singleton pregnancy. For women with two CDs, there is a twofold increase in the risk of placenta previa compared with women with a history of two vaginal deliveries.
PA complicates about 5% of pregnancies with placenta previa. The risk of PA in the setting of placenta previa increases with the number of prior CDs. Thus among women with placenta previa, 40% of those with two previous CDs and 61% of those with three prior CDs develop a PA. This risk is independent of other maternal characteristics such as parity, body mass index (BMI), tobacco use, and coexisting hypertension or diabetes. It has been recently estimated that if the CD rate continues to rise as it has in recent years, by 2020, there will be an additional 6236 placenta previas, 4504 PAs, and 130 maternal deaths annually.
Cesarean Delivery Scar Defects
With the increasing use of transvaginal ultrasound, a uterine cesarean delivery defect (CDSD), or “niche,” is a condition that has been described recently. Much of the focus has been on identifying ultrasound criteria of CDSD during pregnancy that may indicate a higher risk of scar dehiscence and/or uterine rupture during a trial of labor after caesarean (TOLAC) in subsequent pregnancies. Yet ultrasound imaging, including three-dimensional (3-D) imaging and sonohysterography, have been increasingly used to investigate uterine scars in nonpregnant women ( Fig. 21-2 ). CDSDs have been seen in between 20% and 65% of women undergoing transvaginal ultrasound with a previous CD. A CDSD is described as a tethering of the endometrium that can serve as a reservoir for blood and fluid, and it can be associated with clinical gynecologic symptoms such as intermenstrual and postmenstrual spotting and dysmenorrhea. A CDSD may range from a small defect of the superficial myometrium to a larger defect with a direct communication between the endometrial cavity and the visceral serosa ( Fig. 21-3 ). Studies have demonstrated that a history of multiple CDs and uterine retroflexion are significantly associated with larger CDSDs, and larger CDSDs are more likely to be associated with symptoms such as intermenstrual spotting and pelvic pain.
A recent retrospective cohort study has shown that a clinically evident uterine scar dehiscence in a previous pregnancy is a potential risk factor for preterm delivery, low birthweight, and peripartum hysterectomy in the following pregnancy. Interestingly, in this study, previous uterine scar dehiscence did not increase the risk of uterine rupture, PA, or adverse perinatal outcomes, such as low Apgar scores at 5 minutes and perinatal mortality. A recent series of 14 women (20 pregnancies) with a prior uterine rupture and 30 women (40 pregnancies) with a prior uterine dehiscence has shown that these women can have excellent outcomes in subsequent pregnancies if managed in a standardized manner, including CD before the onset of labor or immediately at the onset of spontaneous preterm labor.
Theoretically, uterine scar surgical repair could prevent recurrent cesarean scar pregnancies and could also prevent PA ( Fig. 21-4 ). A recent systematic review of studies reporting on hysteroscopic and laparoscopic CDSD resection has found that abnormal uterine bleeding improved in the vast majority (i.e., 87% to 100%) of patients after these interventions. However, the methodologic quality of the reviewed papers was considered to be moderate to poor, and therefore data were insufficient to make firm conclusions. Until surgical interventions are evaluated in a prospective trial, the benefits and efficacy of surgical repair when a CDSD is observed should be considered unknown. A recent small retrospective study found that the use of monofilament suture for hysterotomy closure at the time of CD significantly reduced the chance of having placenta previa in the subsequent pregnancy. However, these data have not been replicated, and there is a need for properly designed, quality randomized controlled trials (RCTs) with larger sample sizes to evaluate the impact of different suture materials on uterine scar defects and risk of uterine rupture during TOLAC and on the development of placenta previa/accreta in subsequent pregnancies.
Prenatal Diagnosis
Before the development of newer imaging techniques, it was not possible prenatally to assess the depth of placental invasion or, in many cases, to even assess the existence of a PA. Consequently, many publications on PA report what can be considered a mixture of the conditions of accreta, increta, and percreta as well as antenatally suspected and pathologically diagnosed PA, which makes it difficult to assess the results of management for each individual pathologic entity and clinical situation. At present, however, prenatal diagnosis of PA has become a useful tool for a better understanding of the pathophysiology of an abnormally adherent placenta and its appropriate management.
Ultrasound Imaging
Ultrasound has become the primary screening tool for women at risk of PA. Gray-scale ultrasound features suggestive of PA include the loss of the myometrial interface or retroplacental clear space, reduced myometrial thickness, and chaotic intraplacental blood flow and intraplacental lacunae ( Figs. 21-5 to 21-8 ) . Placenta increta may be suggested by an examination of the border between the bladder and myometrium, which is normally echogenic and smooth. In placenta percreta, the placenta often is seen bulging into the bladder. In rare cases, frank placental invasion into the bladder, which appears as exophytic masses, can be seen.
Much attention has been focused on the presence of “intraplacental lacunae” as a marker of PA. These sonolucent spaces contain slow-moving maternal blood on gray-scale imaging and have been described as intraplacental “lakes.” The mere presence of some lacunae, however, does not signify the presence of a PA. When they involve a small area of the placenta or are found in an area of low villous tissue density, such as in the center of the cotyledons or under the chorionic plate, these lakes may be normal variants and have no clinical significance. Conversely, when a PA is present, the lacunae are extensive and create a “moth-eaten” appearance of the placenta (see Figs. 21-5 and 21-6 ).
The most common sonographic finding associated with PA is the loss of myometrial interface with enlargement of the underlying uterine vasculature. The addition of color or power Doppler evaluation has been valuable in improving the screening capacity of ultrasound for PA. Doppler features that suggest PA include chaotic intraplacental blood flow, the presence of increased blood flow in the retroplacental space, and aberrant vessels crossing between placental surfaces (see Figs. 21-7 and 21-8 ). In at-risk women, gray-scale ultrasound is moderately sensitive (70% to 90%), although this sensitivity may be improved by color flow mapping. A recent meta-analysis of 22 studies that included a total of 3641 pregnancies at risk of invasive placentation indicated that the overall sensitivity of ultrasound in diagnosing PA antenatally is around 90%. Quality assessment of the studies showed that the study quality was generally high. The negative predictive value (NPV) of other recent studies ranged between 95% and 100% ( Box 21-3 ). However, not all studies have yielded sensitivities and specificities that were quite as high. For example, a recent prospective study has shown that the sensitivity, specificity, positive predictive value (PPV), and NPV were 53.5%, 88.0%, 82.1%, and 64.8%, respectively. Moreover, it should be noted that most of these values come from studies of high-risk women, and the predictive values among lower risk women (e.g., those without a placenta previa) would be expected to be lower.
Presence of Lacunae
Sensitivity: 77.8% (95% CI, 70.7-83.6)
Specificity: 96.5% (95% CI, 95.6-97.1)
Loss of Hypoechoic Space Between Myometrium and Placenta
Sensitivity: 63.9% (95% CI, 55.1-71.9)
Specificity: 97.3% (95% CI, 96.6-97.9)
Color Doppler Imaging
Sensitivity: 91.2% (95% CI, 87.2-96.7)
Specificity: 91.9% (95% CI, 88.8-94.2)
Ultrasound Overall
Sensitivity: 90.8% (95% CI, 87.0-93.5)
Specificity: 96.9% (95% CI, 96.2-97.4)
Positive predictive value: 74.8% (95% CI, 70.2-78.8)
Negative predictive value: 99.0% (95% CI, 98.6-99.3)
CI, confidence interval.
Disruption of the normal appearance of continuous color flow that results in a gap in myometrial blood flow can also be seen in cases of PA. This gap represents the site of placental invasion into the myometrium and can be diagnosed as early as the first trimester of pregnancy. Within this context, it is possible that the suspicion for PA may even be heightened (see Fig. 21-5 ) or lessened ( Fig. 21-9 ) at the time of an ultrasound examination at 11 to 14 weeks of gestation.
In cases of a placenta in the lower segment, a transvaginal scan will help evaluate the uteroplacental interface near the internal os of the uterine cervix, and it may assist in assessing the degree of placental invasion. For example, the degree of excess vascularity of the lower uterine segment assessed by ultrasound correlates with disease severity. A recent study using logistic regression modeling has shown that the use of formal mathematical modeling to predict PA may have a better PPV than qualitative assessment of ultrasound alone. Therefore integrating PA screening into the 11- to 14-week and/or midgestation obstetric ultrasound examination of women with a previous cesarean scar who present with a low anterior placenta can be helpful in the management of PA.
Magnetic Resonance Imaging
Magnetic resonance imaging (MRI) has recently been introduced as a tool that can be used to evaluate the possibility of PA. Controversy exists as to its usefulness, although some authors have suggested that MRI is better than ultrasound in defining areas of abnormal placentation and assessing the depth of myometrial invasion, particularly in cases of posterior placentae ( Fig. 21-10 ). A recent meta-analysis of 18 studies that involved 1010 pregnancies at risk for invasive placentation has indicated that the sensitivity and specificity of MRI in diagnosing PA antenatally were 94.4% and 84.0%, respectively. MRI was also useful in assessing both the depth and topography of placental invasion ( Table 21-1 ). Focal interruption of the myometrium and the presence of dark intraplacental bands on T2-weighted sequences showed the best sensitivity (92.0%), whereas tenting of the bladder and uterine bulging had the best specificity (98.6%). The prevalence of PA in the cohort under review was nearly 75%, which suggests that women included in these studies were highly selected and at very high risk. Therefore, the diagnostic accuracy of MRI in a more general population is difficult to ascertain, and it has not yet been shown that the addition of MRI to ultrasound results in improved clinical outcomes.
