Morbidly adherent placentation is defined as abnormal implantation of the placenta with chorionic villus invasion beyond the decidua basalis. Three degrees of morbidly adherent placenta have been described: placenta accreta (partial or complete absence of the decidua basalis with adherence of the placenta directly to the superficial myometrium); placenta increta (deep trophoblastic invasion into the myometrium but not through to the serosa); and placenta percreta (trophoblastic invasion through the myometrium into the uterine serosa and potentially beyond).1–4 The incidence of morbidly adherent placenta has progressively increased in the United States over the past three decades and parallels the increasing incidence of cesarean delivery.5–7
Improvements in surgical technique, utilization of a multidisciplinary team approach, and planned delivery at a high-volume referral center have all been shown to improve outcomes of pregnancies complicated by morbidly adherent placenta.8–10 However, the optimal timing of delivery in patients with morbidly adherent placenta remains a matter of debate. Some authors recommend planned delivery at 34–35 weeks of gestation,1 whereas others suggest that delivery may be deferred until 36 weeks of gestation in appropriately selected cases.11–13 In 2011, we established a multidisciplinary morbidly adherent placenta program with the goal of standardizing the clinical management of morbidly adherent placenta and improving outcomes. As part of our protocol, operative delivery is routinely scheduled at 34–35 weeks of gestation for all women in whom imaging suggests a high suspicion for morbidly adherent placenta.10 Over time, we have noted a relatively high incidence of patients in this cohort who did not reach their planned delivery date and who underwent urgent delivery, most commonly as a result of the development of sustained, symptomatic preterm contractions, vaginal bleeding, or both. At present, identification of women with morbidly adherent placenta at greatest risk for requiring urgent delivery is difficult, because risk factors have not been systematically defined. In addition, the effect of urgent compared with planned delivery of these patients on maternal and neonatal outcomes is unknown. The objectives of the current study were 1) to compare the outcomes of women and neonates from pregnancies complicated by morbidly adherent placenta who were delivered as scheduled at their planned date and time with the outcomes of those women and neonates who required urgent delivery, and 2) to delineate risk factors associated with the need for urgent delivery in patients with morbidly adherent placenta.
MATERIALS AND METHODS
This is a retrospective case–control study of all women diagnosed with morbidly adherent placenta who underwent either planned or urgent cesarean hysterectomy. After obtaining permission from the Baylor College of Medicine institutional review board (H-28609), a cohort that consisted of all patients between January 1, 2011, and February 30, 2017, was identified from clinical rosters maintained by the Maternal-Fetal Surgery Section at the Department of Obstetrics and Gynecology, Baylor College of Medicine–Texas Children's Hospital Pavilion for Women. Only women diagnosed with singleton pregnancies were included in this study. The details of our multidisciplinary approach have been previously described in detail.10 Morbidly adherent placenta was diagnosed based on standard findings of second- or third-trimester ultrasonography, magnetic resonance imaging, or both. These included multiple, irregular lacunae within the placenta, loss of the hypoechoic layer between the placental–uterine interface, thinning of the myometrium, and irregularity of the bladder wall.
Patients are generally admitted several days before planned cesarean hysterectomy at 34–35 weeks of gestation to stabilize existing comorbid medical conditions, administer corticosteroids for fetal lung maturation, allow necessary evaluation by multiple team members from different disciplines, and ensure comprehensive preoperative preparation per our previously published protocol.10 Women with vaginal bleeding or signs of labor before scheduled 33-week admission are admitted at the onset of these signs or symptoms and are monitored as inpatients. Timing of delivery in these women is based on clinical assessment of the patient's condition. After admission by the maternal–fetal medicine service, other team members including nursing, obstetric anesthesiology, pulmonary–critical care, urology, and neonatology team members are consulted. Largely depending on the results of imaging studies, additional services are consulted on an “as-needed” basis. These services include gynecologic oncology interventional radiology and vascular surgery. House officers assist in this process with careful supervision. A dedicated morbidly adherent placenta team coordinator ensures that our standardized preoperative checklists are appropriately completed.14 In brief, the surgical approach utilized to manage the operative delivery of patients with morbidly adherent placenta includes cystoscopy under anesthesia with placement of bilateral ureteral catheters anchored to a 20-French indwelling urethral balloon catheter, cesarean delivery with fundal hysterotomy remote from the site of placentation followed by hysterotomy closure and modified radical hysterectomy. The placenta is left in situ with no attempt at manual removal. Since its initiation in January 2011, the morbidly adherent placenta program and the surgical approach as outlined previously have remained standardized. Furthermore, there has been no turnover in the three core maternal–fetal medicine surgeons (A.A.S., M.A.B., and K.A.F.). A number of incremental improvements in process and technique have been implemented over time and include the use of a hemostatic hysterotomy incision created using T-clamps or a linear stapler and increased use of a posterior uterine approach.15,16 Whenever possible, regional anesthesia is used before delivery to avoid neonatal depression and to facilitate postoperative pain management with conversion to general endotracheal anesthesia for the hysterectomy. Because this conversion is planned in advance, induction of general anesthesia and intubation do not take a significant amount of additional time. However, cases in which a patient is actively bleeding frequently start with general anesthesia to avoid any undue delay in delivery. All deliveries were performed in the main operating room and ureteral stents were placed in all women before surgery.
Urologic surgeons assisted with bladder resection and repair when there was extensive trophoblastic invasion into the bladder. The bladder was repaired with a first layer of continuous chromic suture in the mucosa, a second layer of continuous absorbable polyglycolic acid-based suture in the detrusor muscle bolstered by a third layer of interrupted suture (absorbable polyglycolic) closure of the paravesical fat and bladder serosa followed by omental interposition between the bladder repair and vaginal cuff suture lines. A 20-French indwelling balloon catheter was left to continuous drainage for minimum 4 weeks after repair.
For purposes of analyses, enrolled participants were categorized into two groups: women who underwent cesarean hysterectomy at their planned date and those who underwent earlier urgent cesarean hysterectomy as a result of significant vaginal bleeding, uterine contractions, or both. In some instances, short-term tocolysis with magnesium sulfate is considered in cases of marked fetal immaturity to allow administration of corticosteroids for fetal benefit. Data regarding maternal demographics, gynecologic and obstetric history, antenatal course including presence of symptomatic uterine contractions, vaginal bleeding, gestational age at delivery, and pathology findings were collected contemporaneously from patient charts. We note that during this timeframe, an additional 14 patients were diagnosed antenatally with morbidly adherent placenta who were found not to have this condition at the time of cesarean delivery and did not undergo hysterectomy. All patients included in this study underwent hysterectomy and had histologic confirmation of morbidly adherent placenta.
The primary outcome was defined as composite maternal morbidity. This was defined as the presence of any of the following variables: intraoperative transfusion of eight or greater red blood cell units, development of peri- or postoperative coagulopathy, ureteric injury, bowel injury, unintentional cystotomy, vesicovaginal fistula, acute kidney injury, postoperative ventilation for 4 hours or greater, occurrence of deep vein thrombosis or pulmonary embolism, reoperation or readmission for a complication, infection, or maternal death. Coagulopathy was defined as the presence of any of the following: platelet count less than 70,000/microliter, international normalized ratio greater than 1.2, or fibrinogen less than 200 mg/dL. Given the frequency of multiple unit transfusion in these patients, 8 units of packed red blood cell transfusion was utilized rather than the more standard 4-unit definition of severe morbidity to enhance discrimination between normally and abnormally morbid outcomes. Secondary maternal outcome variables included estimated blood loss, total units of blood products transfused, volume of crystalloid transfused, massive transfusion protocol activation, need for interventional radiology, need for intentional cystotomy, type of anesthesia, total duration of the surgery, skin-to-skin surgery duration, and length of hospital stay in days. Composite neonatal morbidity was defined as presence of any of the following variables: respiratory distress syndrome, intraventricular hemorrhage, necrotizing enterocolitis, and 5-minute Apgar score less than 7. Other neonatal outcome variables included neonatal birth weight, need for neonatal intensive care unit admission, length of stay in the neonatal intensive care unit, and neonatal death.
All continuous variables were tested for normality using descriptive statistics for skewness and kurtosis, visual evaluation of histograms, and the Kolmogorov Smirnov test. Continuous data are reported as mean with SD or as median with interquartile range as appropriate. Categorical data are reported as proportions and percentages. When appropriate, relevant continuous variables were converted to categories. Comparisons between planned and urgent groups were made using Student t test, Mann-Whitney U, test and χ2 test or Fisher exact test, as appropriate. Given that maternal outcomes in our cohort improved over time after the introduction of our multidisciplinary approach,17 we adjusted comparisons of the maternal outcome variables for the epoch in which they were recorded using the Kruskal-Wallis test and logistic regression. Multivariate logistic regression was used to identify those predictors of urgent delivery and to adjust for potential confounders in a subset of 58 patients whose cervical length measurements (as per American Institute of Ultrasound in Medicine guidelines18) were available. The number of previous cesarean deliveries, history of preterm birth, cervical length less than 25 mm, gravidity three or greater, maternal age 35 years or older, body mass index (calculated as weight (kg)/[height (m)]2) 35 or greater, and gestational age at the time of cervical length measurement were considered in the model. We also used Kaplan-Meier statistics to create a survival curve for delivery and its significance was determined by the log-rank test for factor urgent delivery. We used SPSS 17.0 and a P value of <.05 was considered statistically significant.
One hundred thirty patients met the inclusion criteria. All patients, urgent and planned, had the antenatal diagnosis of morbidly adherent placenta and postdelivery pathology confirmation. Demographics and baseline characteristics of these participants are summarized in Table 1. Seventy (53.8%; 95% CI 45.3–62.4) women underwent planned delivery at 34–35 weeks of gestation, and 60 (46.2%; 95% CI 37.6–54.7) required urgent delivery before 34 weeks of gestation. Of those participants who underwent urgent delivery, 26 (43.3%; 95% CI 30.8–55.9), nine (15%; 95% CI 6.0–24.0), and 11 (18.3%; 95% CI 8.5–28.1) had been admitted to the hospital for vaginal bleeding, premature symptomatic uterine contraction, and both, respectively. Although an attempt is generally made to temporize delivery of these patients on the antepartum unit until 34–35 weeks of gestation, in these women, the severity or persistence of these clinical complications necessitated urgent delivery. The remaining 14 (23.3%; 95% CI 12.6–34.0) patients in the urgent group were admitted before 34 weeks of gestation per protocol but then required urgent and unscheduled intervention as a result of a change in clinical condition. Reasons for urgent delivery were vaginal bleeding in 18 (30%; 95% CI 18.4–41.6), premature symptomatic uterine contraction in 18 (30%; 95% CI 18.4–41.6), and both in 23 (38.3%; 95% CI 26.0–50.6) patients. One patient (1.7%; 95% CI 0.0–4.9) underwent urgent delivery as a result of severe abdominal pain (Fig. 1). The median time from decision to surgery in the urgent group was 3 (interquartile range 2.5–3.5) hours. Figure 2 shows the percentage of women delivered in each group before a given gestational age. None of the patients in the urgent delivery group were diagnosed with ruptured membranes, preeclampsia, or a nonreassuring fetal heart rate tracing. No patient required an emergency delivery in less than 1 hour. Median parity was higher in the urgent group than in the planned group (three [interquartile range 2–4] versus one [interquartile range 2–4], P=.02).
Composite maternal morbidity for enrolled participants is presented in Table 2. Such morbidity was present in 34 (56.7%; 95% CI 43.2–69.4) of the urgent and 26 (37.1%; 95% CI 25.8–48.5) of the planned delivery groups (P=.03). This difference remains statistically significant even when adjusted for the epoch in which the delivery occurred (to eliminate any confounding effect of team improvement over time) (P=.02). When adjusting the data for the depth of placental invasion documented in the pathology report, the higher composite morbidity seen in the urgent delivery group appears to be primarily related to a greater depth of placental invasion (placenta increta or percreta) in this group (Appendix 1, available online at http://links.lww.com/AOG/B54). These observations also suggest that the depth of abnormal placental invasion may predispose to an earlier onset of bleeding and contractions in women with morbidly adherent placenta. Maternal intraoperative transfusion requirements are presented in Table 3. Although intraoperative visual estimated blood loss did not significantly differ between the two groups, fewer units of both red blood cells and fresh frozen plasma were transfused in the planned delivery group (red blood cells, median [interquartile range] 3 [0–8] versus 1 [0–4], P=.02; fresh frozen plasma, median 1 [0–2] versus 0 [0–0], P=.001) (Table 3).
There was one maternal death in the urgent delivery group. This occurred in a patient at 33 weeks of gestation with extensive placenta percreta who was delivered as a result of regular symptomatic contractions without antenatal bleeding. During surgery, this patient experienced massive retroperitoneal hemorrhage from a previously unknown renal artery aneurysm that ruptured intraoperatively, further complicated by coagulopathy, ventricular tachycardia, and subsequent cardiac arrest. There was one case of delayed cecal perforation associated with ileus (Ogilvie syndrome) that required colostomy and a delayed reanastomosis in the planned delivery group.
Table 4 details intraoperative and postpartum outcomes. General anesthesia at the beginning of surgery was more common in the urgent compared with the planned delivery group (49 [81.7%; 95% CI 71.9–91.5] versus 41 [58.6%; 95% CI 47.0–70.1], P<.01) (Table 4).
Interventional radiology service was consulted for five women (two in urgent and three in planned delivery groups). In each case, the invasive placental mass extended to the pelvic sidewalls, and access to the uterine vascular bundles was not technically possible before embolization. Embolization resulted in sufficient reduction in placental perfusion pressure to significantly shrink the placental mass and allow surgical access to the uterine vascular bundles.
Composite neonatal morbidity was present in 45 (78.9%; 95% CI 68.4–89.5) of the neonates in the urgent group and 27 (45.8%; 95% CI 33.1–58.5) of those in the planned group (P<.01). As expected from the group definitions, the urgent delivery cohort neonates had a lower gestational age (median 32 [interquartile range 30–33] vs 34 [interquartile range 34–35], P<.01). Similarly, urgent delivery neonates weighed less (mean±SD [kg] 1.4±0.6 vs 2.1±0.4, P<.01), had a higher neonatal intensive care admission rate (54 [98.2%; 95% CI 94.7–100.0] versus 53 [88.3%; 95% CI 80.2–96.5], P=.03) and required a longer length of stay (median [days] 26 [interquartile range 14–43] vs 9 [interquartile range 4–15], P<.01). There was no intraventricular hemorrhage or necrotizing enterocolitis in either group. However, low Apgar score and respiratory distress syndrome were higher in the urgent compared with the planned delivery group (5-minute Apgar score less than 7, 34 [59.6%; 95% CI 46.9–72.4] versus 14 [23.3%; 95% CI 12.6–34.0], P<.01; respiratory distress syndrome, 34 [61.8%; 95% CI 49.0–74.7] vs 16 [27.1%; 95% CI 15.8–38.5], P<.01). There were three neonatal deaths before discharge in the urgent and none in the planned delivery groups. Reasons for neonatal deaths before discharge were severe hypoplastic left heart syndrome with comfort care after birth, prematurity and omphalocele with pulmonary hypoplasia at 25 weeks of gestation, and previable gestational age at 20 weeks of gestation.
Cervical length was measured in 58 women between 18 and 23 weeks of gestation (33 women in the urgent and 25 in the planned delivery groups). According to our logistic regression analysis, history of two or more prior cesarean deliveries was the only independent predictor of urgent delivery (odds ratio [OR] 4.9, 95% CI 1.5–16.0; adjusted OR 11.4, 95% CI 1.8–71.1; P=.01). The predictive role of cervical length less than 25 mm and a history of prior preterm birth was not confirmed by our logistic model (cervical length less than 25 mm: OR 4.3, 95% CI 0.5–39.2, adjusted OR 6.2, 95% CI 0.5–86.0, P=.17; history of prior preterm birth: OR 3.2, 95% CI 0.8–13.1, adjusted OR 8.1, 95% CI 1.0–65.9, P=.05).
The American College of Obstetricians and Gynecologists and the Society for Maternal-Fetal Medicine have included morbidly adherent placenta as an indication for planned late preterm delivery.19 We and others have previously shown that managing morbidly adherent placenta using a multidisciplinary team approach results in improved maternal outcomes.8,10,17 In this study, we tested our hypothesis that women who undergo cesarean hysterectomy for morbidly adherent placenta urgently are at increased risk for composite morbidity when compared with those who undergo this surgery as a scheduled procedure.
When we analyzed the relative contribution of each individual component of the composite maternal morbidity variable, two outcomes were more common in the urgent delivery group: coagulopathy and postoperative ventilation for 4 hours or greater. Nearly 70% of patients in the urgent delivery group underwent cesarean hysterectomy earlier than scheduled as a result of bleeding (either alone or in conjunction with symptomatic premature uterine contractions). The increased rate of bleeding before surgery might contribute to the higher rate of coagulopathy seen in this group. Despite the fact that there was not a significantly different estimated intraoperative blood loss between the two groups, more red blood cell units were transfused in the urgent delivery group; this difference is likely explained by the fact that many patients in the urgent delivery group had already lost an appreciable amount of blood before surgery, although it is also possible that the need for urgent delivery itself led to more aggressive transfusion. A greater volume and higher rate of blood transfusion is a known contributing factor to prolonged ventilation, and this may explain this outcome in the urgent delivery group.
We found that a history of two or more prior cesarean deliveries is an independent predictor for urgent delivery. However, we were unable to show any association between cervical length or a history of preterm birth and the need for urgent delivery in our population of patients with morbidly adherent placenta. These negative findings must be viewed with caution, because our sample size was underpowered to confirm a finding of no difference. The major finding of this study was a confirmation of increased morbidity in women undergoing urgent compared with planned cesarean delivery for morbidly adherent placenta. Our data suggest that a history of two or more prior cesarean deliveries in this specific population may be added to the list of known risk factors for preterm labor in the general population (for example, short cervical length or a history of preterm birth) in considering timing of delivery.
Almost half of our patients with morbidly adherent placenta required an urgent, unscheduled surgery, despite our general plan to deliver at 34–35 weeks of gestation. Given these data, it may be reasonable to individualize the scheduled delivery time in selected patients with identifiable risk factors and to plan delivery for 33–34 weeks of gestation. This is an important consideration given the logistics of the multidisciplinary management required for these patients. This becomes especially important in centers that lack round-the-clock availability of all necessary multidisciplinary team members, although the wisdom of plan management of these patients in such a center is questionable. Our data demonstrate that in a specialized, high-volume morbidly adherent placenta referral center, maternal and neonatal outcomes are generally good even with the need for unplanned, urgent delivery. However, such cases are associated with increased morbidity. Accordingly, we have changed our approach and currently admit women with known major risk factors for preterm labor and birth in the general population (previous preterm birth, short cervix between 18 and 23 weeks of gestation) and those with two or more prior cesarean deliveries, now demonstrated to be a risk factor in this specific population, before 33 weeks of gestation for standard preoperative evaluation and preparation for delivery at 33 weeks of gestation.
This study has a number of strengths including a large sample size of antenatally suspected and pathologically confirmed patients with morbidly adherent placenta, all of whom were managed in a single center by the same team using the same surgical methodology and preoperative, intraoperative, and postoperative management protocols. In addition, trained physician researchers contemporaneously abstracted the clinical data with few missing data points. Limitations of this study include those inherent with any retrospective analyses. We were not able to generalize the negative results of our logistic regression model as a result of small counts for the number of included contributing variables in preterm labor.
In conclusion, our findings suggest urgent delivery of women with morbidly adherent placenta is associated with worse outcomes than women with planned delivery even when managed by a dedicated multidisciplinary team. This worse outcome is primarily the result of increased rates of maternal coagulopathy and the need for more aggressive blood transfusion when the surgery is precipitated by vaginal bleeding as well as increased neonatal morbidity. Although the current national standards recommend scheduled cesarean delivery at 34–35 weeks of gestation for most women, in patients with two or more prior cesarean deliveries, it may be reasonable to individualize scheduled cesarean hysterectomy at 33 weeks of gestation.
1. Publications Committee, Society for Maternal-Fetal Medicine, Belfort MA. Placenta accreta. Am J Obstet Gynecol 2010;203:430–9.
2. Belfort MA. Indicated preterm birth for placenta accreta. Semin Perinatol 2011;35:252–6.
3. Silver RM. Abnormal placentation: placenta previa, vasa previa, and placenta accreta. Obstet Gynecol 2015;126:654–68.
4. Hull A, Resnick R. Placenta previa, placenta accreta, abruptio placentae, and vasa previa. Philadelphia (PA): Maternal-Fetal Medicine Principles and Practices Saunders Elsevier; 2009. p. 727.
5. Silver RM, Landon MB, Rouse DJ, Leveno KJ, Spong CY, Thom EA, et al. Maternal morbidity associated with multiple repeat cesarean deliveries. Obstet Gynecol 2006;107:1226–32.
6. Clark SL, Koonings PP, Phelan JP. Placenta previa/accreta and prior cesarean section. Obstet Gynecol 1985;66:89–92.
7. Bailit JL, Grobman WA, Rice MM, Reddy UM, Wapner RJ, Varner MW, et al. Morbidly adherent placenta treatments and outcomes. Obstet Gynecol 2015;125:683–9.
8. Eller AG, Porter TF, Soisson P, Silver RM. Optimal management strategies for placenta accreta. BJOG 2009;116:648–54.
9. Robinson BK, Grobman WA. Effectiveness of timing strategies for delivery of individuals with placenta previa and accreta. Obstet Gynecol 2010;116:835–42.
10. Shamshirsaz AA, Fox KA, Salmanian B, Diaz-Arrastia CR, Lee W, Baker BW, et al. Maternal morbidity in patients with morbidly adherent placenta treated with and without a standardized multidisciplinary approach. Am J Obstet Gynecol 2015;212:218.e1–9.
11. Oyelese Y, Smulian JC. Placenta previa, placenta accreta, and vasa previa. Obstet Gynecol 2006;107:927–41.
12. Upson K, Silver RM, Greene R, Lutomski J, Holt VL. Placenta accreta and maternal morbidity in the Republic of Ireland, 2005-2010. J Matern Fetal Neonatal Med 2014;27:24–9.
13. Rac MW, Wells CE, Twickler DM, Moschos E, McIntire DD, Dashe JS. Placenta accreta and vaginal bleeding according to gestational age at delivery. Obstet Gynecol 2015;125:808–13.
14. Fox KA, Shamshirsaz AA, Belfort MA. Center of Excellence for morbidly adherent placenta. In: Silver RM, editor. Placenta accreta spectrum. Boca Raton (FL): CRC Press; 2017. p. 113–21.
15. Belfort MA, Shamshiraz AA, Fox K. Minimizing blood loss at cesarean-hysterectomy for placenta previa percreta. Am J Obstet Gynecol 2017;216:78.e1–2.
16. Belfort MA, Shamshirsaz AA, Fox KA. A technique to positively identify the vaginal fornices during complicated postpartum hysterectomy. Am J Obstet Gynecol 2017;217:222.e1–3.
17. Shamshirsaz AA, Fox KA, Erfani H, Clark SL, Salmanian B, Baker BW, et al. Multidisciplinary team learning in the management of the morbidly adherent placenta: outcome improvements over time. Am J Obstet Gynecol 2017;216:612.e1–5.
18. American Institute of Ultrasound in Medicine. AIUM practice guideline for the performance of obstetric ultrasound examinations. J Ultrasound Med 2010;29:157–66.
19. Medically indicated late-preterm and early-term deliveries. Committee Opinion No. 560. American College of Obstetricians and Gynecologists. Obstet Gynecol 2013;121:908–10.