Placenta praevia and placenta accreta occur when the placenta is located wholly or partly in the lower uterine segment and is morbidly adherent to it. Placenta accreta is defined as an abnormally deep attachment of the placenta, through the endometrium and into the myometrium. There are three forms of placenta accreta, distinguishable by the depth of penetration: placenta accreta describes the partial or complete absence of decidua, with adherence of the placenta directly to the superficial myometrium; placenta increta is an invasion into, but not all the way through, the myometrium; and placenta percreta describes penetration through the full thickness of the myometrium and into the serosa. This is relatively rare but potentially life-threatening, as it often results in complications in the peripartum period such as severe haemorrhage, a possible need for caesarean hysterectomy and even severe injuries to pelvic organs.
The aim nowadays is to manage placenta accreta conservatively not only to avoid massive pelvic bleeding and massive transfusions, but also to preserve fertility. The improvement of imaging investigations facilitates a diagnosis, allowing management plans to be discussed with other clinical professionals (e.g. interventional radiologists and anaesthetists), as well as the patient and her family.
The association between placenta accreta, placenta praevia and caesarean section (CS) has become more striking in recent years, and the incidence of placenta accreta (including increta and percreta) may have risen 10-fold in the past 50 years.1 Recent reports suggest that the incidence ranges from 1 in 110 to 1 in 2500 deliveries,2 and this incidence is expected to rise as the number of CSs increases.
The present study was conducted over a period of 30 months. The aims of the study were as follows: to determine the incidence of confirmed placenta accreta at Dubai Hospital, United Arab Emirates, and to compare this with the literature; to examine the influence of an antenatal diagnosis in determining approaches to management; to assess the effect of subsequent separation of the placenta during CS on maternal outcomes; to discuss new challenges in management; and to institute new guidelines in our department for the management of future patients with the same pathology, including undertaking caesarean hysterectomy in patients who do not wish to have further pregnancies, or leaving the placenta in situ in those who wish to preserve their fertility.
Materials and methods
A retrospective and descriptive study was conducted of all pregnant women at or beyond 24 weeks of gestation who were diagnosed with placenta praevia accreta, either antenatally or during CS, between 1 June 2009 and 31 December 2011. Patients were identified from medical files, record books and electronic medical record databases at Dubai Hospital.
The medical notes were examined and checked against the database for the following: gestational age at diagnosis and delivery; risk factors; imaging findings; operative interventions and any complications; estimated blood loss; blood unit transfusion; and findings confirmed by the pathology department in cases where hysterectomy was carried out. Admission to the intensive care unit (ICU) and the mean hospitalization period were analysed. The literature was reviewed and our results compared with those of international reports.
During the study period of 30 months, there were 9577 deliveries at Dubai Hospital, of which 2966 deliveries (31%) were by CS. Eleven cases of placenta praevia accreta were diagnosed. The estimated incidence of the condition, based on the results of this study, is 1 in 909 deliveries. During this study period, no vaginal deliveries were followed by post-partum diagnosis of placenta accreta, nor were there any cases of CS with placenta accreta only.
The mean maternal age was 33.5 years (range 25–37 years). The mean gestational age at delivery was 35.7 weeks (range 24–37 weeks). Patients with placenta praevia accreta were pauciparous, with an average parity of three. This does not correspond to the literature, with multiparity considered a risk factor for this condition.
All patients with placenta praevia accreta had undergone at least one prior CS (Table 1).
|Case number||Age||Parity||Number of previous CSs||Number of miscarriages||Other risk factors|
ERPOC, evacuation of retained products of conception.
An ultrasound examination is performed routinely in all pregnant women attending the hospital. Screening for placenta accreta is carried out only once an antenatal diagnosis of placenta praevia has been made, in view of the association between placenta praevia and the diagnosis of placenta accreta. In most cases, the diagnosis of placenta praevia accreta was made before 30 weeks of gestation. Table 2 shows the level of concordance between ultrasound results and final diagnosis. Although the sample was small, it is clear nonetheless that the sensitivity of ultrasound was poor, compared with that reported in the literature. Only 4 out of 11 patients with placenta praevia accreta (36.4%) had corresponding ultrasound findings.
MRI, magnetic resonance imaging.
Seven out of 11 patients were misdiagnosed by ultrasound (63.6%), and hysterectomy was necessary in six of these patients. In two cases, the pathology indicated the presence of placenta accreta increta, and in one case placenta accreta percreta. Magnetic resonance imaging (MRI) was requested in five cases. In one case this was cancelled as the patient did not tolerate the examination. Diagnoses based on ultrasound were confirmed in two of the remaining four patients; in the other two patients these diagnoses were wrongly excluded. Both of these patients underwent hysterectomy for placenta accreta with intractable post-partum haemorrhage (PPH). With regards to procedure, six patients had elective CSs whereas five underwent emergency CSs.
All CSs were carried out in the presence of a consultant. Four units of blood were cross-matched as per protocol for any patient undergoing a CS. As the diagnosis of placenta praevia accreta was not made in 7 out of 11 cases, an attempt was made to separate the placenta after delivery in all patients. In two cases, part of the placenta was left in situ; however, one of these patients subsequently returned to the operating theatre for secondary PPH and required a hysterectomy due to massive bleeding with profound hypotension. Another patient underwent a second laparotomy after collapsing, with 1 l of haemoperitoneum resulting from intraperitoneal bleeding. Ligation of the internal iliac artery was performed to stop the bleeding.
Immediately after placenta removal, all patients experienced severe bleeding with hypotension. In all cases, every effort was made to arrest the bleeding before resorting to emergency hysterectomy [including the use of oxytocin (Syntocinon®, Sandoz, Sandoz International GmbH, Holzkirchen, Germany), prostaglandin, uterine sutures, bed stitches and Bakri™ Balloon (Cook Medical, Limerick, Ireland)], and this may account for the significantly larger volume of blood lost and blood transfused in this group (Table 3). There were two cases of bladder injury during dissection (see Table 3). There were no bowel injuries, despite adhesions and post-partum infection.
|Case number||CS||Bed stitches||Bakri Balloon||Lynch technique||Uterine artery ligation||Internal iliac artery ligation||Hysterectomy||Pathology|
Blood loss leads to marked circulatory collapse, vasoconstriction and oliguria, and may even be life-threatening. Mean blood loss in the patients in this study was estimated at 2.9 l (range 1.5–5.0 l). An analysis of blood units received during and after surgery suggests that this may be an underestimate (Figure 1).
The mean number of blood units transfused was 8.1 (range 5.0–15.0 units). An average of 8.5 units of other blood products were required (range 2.0–15.0 units). All patients with placenta accreta or placenta praevia with PPH were admitted to the ICU for between 24 and 72 hours. The mean length of hospital stay after delivery was 7.1 days.
As numbers of CSs continue to rise, both within Dubai Hospital and worldwide, the problem of placenta praevia accreta will become more common. At Dubai Hospital, 1 in 909 deliveries is affected. This is within the range of figures reported internationally, which varies from 1 in 1102 to 1 in 2510 deliveries affected.3 The association between placenta praevia and prior CS delivery with placenta accreta and the risk of requiring a hysterectomy is well documented.4 Miller et al.3 found that the risk of placenta accreta ranged from 2% in women < 35 years of age with no history of CS to almost 39% in women who had undergone two or more previous CSs and who had an anterior or central placenta praevia.
Greyscale ultrasonography is now an established first-line investigation for suspected placental invasion of the myometrium and has clear diagnostic criteria: placental lacunae and loss of retroplacental clear space.5 It has reduced the morbidity and mortality from placenta accreta by helping to establish the diagnosis and detect complications antepartum. Comstock5 found that, at 15–20 weeks of gestation, the presence of lacunae in the placenta is the most predictive sonographic sign of placenta accreta, with a sensitivity of 79% and a positive predictive value of 92%. MRI has been shown to be beneficial in some cases when ultrasound findings are equivocal or not diagnostic.6 The use of MRI has a variable sensitivity ranging from 38% to 88%, and specificity of up to 100% when it is applied as a secondary diagnostic tool. In this study, errors were made on both ultrasound and MRI; the direct consequences were a high rate of morbidity (hysterectomies, bladder injuries and massive blood transfusions). In future, antenatal diagnosis should be improved by training obstetricians and radiologists to approach the diagnosis according to clear, established criteria, which will benefit those patients who wish to avoid hysterectomy in order to preserve their fertility.
Owing to a lack of antenatal diagnosis, all patients experienced massive bleeding, leading to emergency peripartum hysterectomy. Hsu et al.7 reported an estimated mean blood loss of 3.8 l, as compared with 2.9 l in this study. The mean quantity of whole blood transfused in the study by Hsu et al. was 15 units,7 as compared with 8.1 units for our patients. Wong et al.8 compared one group of patients who did not receive an antenatal diagnosis with another group given an antenatal diagnosis of placenta accreta. They concluded that women without an antenatal diagnosis of placenta accreta experienced significantly more haemorrhaging (a mean total blood loss of 3.6 l vs. 1.4 l in the group given an antenatal diagnosis; P = 0.003).8 Despite the limitations of studying a small number of conservatively managed cases, these authors concluded that a clear relationship had been shown between antenatal diagnosis of placenta accreta and non-separation of the placenta, with accompanying improved maternal outcome. According to Chan et al.,9 by leaving the morbidly adherent placenta in situ, with or without pelvic arterial embolization, they succeeded in preserving the uterus in three patients. Kayem et al.10 reported on a much larger number of patients with placenta accreta, managed using a similar conservative approach. In this study, the uterus was preserved in 80% of patients, and the number of complications was low compared with historical attempts to remove the placenta.10 In 2009, at a 33 national days meeting of the French National College of Gynaecologists and Obstetricians (CNGOF), the findings of a large multicentre study comprising 167 patients from 25 university hospitals, who had conservative management for placenta accreta, were presented.11 The results showed that the uterus was preserved in 78.4% of these patients, with maternal morbidity reported in 6% of cases. Of the patients who had conservative management for placenta accreta, 28.2% later became pregnant without the aid of medication. The main drawback of conservative management of placenta accreta is that a long period of follow-up is required for those patients who experience a delay in the spontaneous resorption of the placenta, which may take up to 17 weeks on average (range 4–45 weeks). For this reason, patients should be fully advised of the options available for the management of placenta accreta during the antenatal period. Our department at Dubai Hospital is working towards this goal in order to decrease the rate of morbidity, despite well-known risks of conservative management, e.g. blood transfusion (required in 40% of patients), infection (can affect 28% of patients) and recurrence of placenta accreta in future pregnancies (can affect 30% of patients).
Placenta accreta is a relatively rare but potentially catastrophic obstetric complication. It is likely that its incidence will continue to increase because of increasing numbers of repeat CS deliveries. Antenatal diagnosis is an invaluable aid in perinatal management, as it allows the clinician to anticipate and recognize complications that might not otherwise be expected. Conservative management is one option that should be discussed with the patient, with a view to decreasing massive maternal bleeding and conserving fertility should the patient wish to do so.