Materials and Methods
A retrospective cohort study included all pregnancies with a histopathologically confirmed diagnosis of placenta accreta, increta, or percreta treated in 1 of 3 tertiary teaching hospitals for Baylor College of Medicine, Houston, TX, between January 2000 and September 2013. This study was approved by the Baylor College of Medicine Institutional Review Board.
A formal multidisciplinary management program was introduced in 2011 for known and suspected cases of morbidly adherent placenta. We instituted a standardized protocol with a multidisciplinary strategy ( Figure ). Briefly, the protocol included admission at 33-34 weeks of gestation, planned preterm delivery by cesarean hysterectomy between 34 and 35 weeks of gestation, preoperative consultation, and prospective planning for maternal and neonatal care by the multidisciplinary team. All patients were admitted to the maternal-fetal medicine (MFM) service under the “percreta team” protocol and treated by an MFM on the team. The MFM physician coordinated all care and planning. The team had a nursing coordinator who ensured that all preadmission and admission checklist items were completed. All patients had a standardized consultation scheduled after admission with pulmonary critical care (a 24/7 service on our labor and delivery unit), urology, blood bank, anesthesiology, nursing, and neonatal intensive care unit services. Other specialized services such as interventional radiology or vascular surgery were consulted on an as-needed basis in an individualized fashion. In some cases, patients began contracting or started bleeding before all consultations were completed; in a few cases, emergency surgery was required. In these cases, the multidisciplinary approach was still followed systematically to the extent that it was safe.
Whenever possible, patients underwent combined spinal-epidural anesthesia for bilateral ureteric stent placement and cesarean delivery. This minimized fetal exposure to general anesthesia and facilitated postoperative pain management. Before induction of general anesthesia, the large-bore venous lines, an arterial line, and a central venous line were placed to optimize monitoring and ensure adequate vascular access before onset of hemorrhage. Patients were positioned in lithotomy with low-Allen stirrups to allow visualization of vaginal bleeding, if necessary, and for a third co-surgeon to have access to the surgical field. Abdominal entry was through a periumbilical midline abdominal incision, and the pregnant uterus was exteriorized gently to allow fundal or posterior classic hysterotomy. The placenta was allowed to remain in situ without attempt at removal. A modified radical hysterectomy technique that included ureterolysis was standardized and included extensive use of a bipolar cautery device (Ligasure; Covidien, Mansfield, MA).
Given the distended and enhanced deep pelvic collateral vasculature that is usually present when there is morbidly attached placentation (regardless of whether it is an accreta, increta or percreta, and the importance of reducing blood supply to the uterus in a systematic manner), we have developed what we refer to as a modified radical hysterectomy technique. This technique is designed to ensure wide margins from the friable uterine wall such that fragile and unsupported vessels and/or thinned out myometrium is not inadvertently disrupted by the placement of clamps or by attempts to suture or ligate pedicles against the uterus itself. The retroperitoneum is accessed lateral to the round ligaments, and the ureters and iliac vessels are exposed and identified as a first step in the hysterectomy. The uterus is then separated from its support structures, which leaves as wide a margin of broad ligament as is possible. The ovaries are preserved, but the tubes are removed. This technique allows us the ability to identify the feeder arteries early in the case and ligate (uterine) or temporarily clamp with a vascular clip (superior vesical) to reduce engorgement of the ballooned lower segment and bladder wall. Where necessary, a ureterolysis is carried out to protect the ureters and allow step-by-step devascularization of the lower segment. Even in cases of accreta and increta, in which the trophoblast has not invaded the bladder wall, there are still very distended and engorged blood vessels that traverse the interface between the bladder and uterine lower segment that can cause significant bleeding. By exposure of the lateral pelvic anatomy and performance of a ureterolysis in a systematic fashion, these vessels can be cauterized and closed down, and the bladder can be separated in a controlled manner with minimal blood loss.
Deliberate cystotomy and bladder excision was favored in cases with deep placental invasion over persistent attempts at bladder dissection. Intraoperative staged arterial embolization of the placental bed (after the cesarean delivery but before the hysterectomy) was performed for selected percreta cases that involved the lateral pelvic side walls. All patients received 500 mL albumin 5% before cesarean hysterectomy because acute volume expansion with colloid reduces intraoperative crystalloid requirement and facilitates hemodilution before hemorrhage. Early blood product replacement that used a massive transfusion protocol (transfusion of PRBCs and fresh frozen plasma in a 1:1 ratio) was encouraged, and electrolytes were measured frequently, particularly ionized calcium and potassium levels. During acute hemorrhage, we draw complete laboratory sets (arterial blood gas, K+, Ca++, H&H, prothrombin time, activated partial thromboplastin time, international normalized ratio, fibrinogen, platelets, D-dimer, glucose, Mg++) every 20 minutes. One jumbo dose (35-50 mL) of platelets was transfused after every 8 units of PRBCs. We tried to keep the platelet count ≥100,000/mL. Ureteric stents generally were put in with epidural anesthesia before the cesarean delivery. All patients received underbody and overbody forced air warming plus warmed intravenous infusions to achieve perioperative normothermia. Patients were given standardized immediate recovery in the intensive care unit after the operation. After referral for treatment by our team, all patients underwent ultrasound examination by a team member (W.L.) to confirm the diagnosis. Several parameters were investigated during ultrasound scans that included the type of placenta previa, loss of normal hypoechoic retroplacental zone, multiple vascular lacunae (irregular vascular spaces) within placenta, myometrial thickness <1 mm/loss of visualization of the myometrium, disruption of the placental-uterine wall interface and the presence of vessels that cross this area, blood vessels or placental tissue that bridge the uterine-placental margin, myometrial-bladder interface, or crossing uterine serosa. In cases with lateral or posterior placentation, magnetic resonance imaging was considered for more accurate evaluation. In cases with lateral or posterior placentation, magnetic resonance imaging was considered for more accurate evaluation. Before 2011, patients were treated on a case-by-case basis without a specific protocol and are included in the nonmultidisciplinary group. Since the introduction of the multidisciplinary program at our center, the number of the referred patients was increased markedly.
Medical records were reviewed, and data were collected that pertained to maternal death, EBL, anesthesia type and duration, operative time, use of blood products, vital signs during surgery, intraoperative and postoperative complications (ie, vascular, ureteral, and bowel injury), cystotomy, ureteric reimplantation, need for reexploratory laparotomy, postoperative intensive care unit complications, length of stay, and readmission after discharge. Data for patients who came to the hospital in an emergent state and those who were hospitalized but required emergency surgery for obstetric complications (labor, bleeding, fetal concerns) were also analyzed. Neonatal data were collected for birthweight and postdelivery complications (respiratory distress syndrome, sepsis, infection, necrotizing enterocolitis, intracranial bleeds, hypoxemic ischemic encephalopathy, or neonatal death).
Patient characteristics were compared with the use of descriptive statistics. The Mann-Whitney U test was used to compare continuous variables when appropriate. Categoric outcomes were compared with the use of χ 2 analysis and the Fisher exact test as appropriate. Multivariate logistic regression was performed to investigate the effect of several variables on the outcomes. A probability value of < .05 was considered significant. We used SPSS statistical software (version 17.0; SPSS Inc, Chicago, IL).
Results
Ninety cases of histopathologically confirmed placenta accreta, increta, or percreta were identified from 2000-2013. Of the 90 patients, 57 women (63%) were in the multidisciplinary group (March 2011 to September 2013), and 33 women (37%) were in nonmultidisciplinary group (January 2000 to March 2011). Maternal demographics are presented in Table 1 . There was no statistically significant difference in maternal characteristics or the number of previous cesarean deliveries between the 2 groups. The multidisciplinary group had a higher percentage of cases with placenta percreta ( P = .008) and placenta previa ( P = .047).
| Variable | Multidisciplinary group (n = 57) | Nonmultidisciplinary group (n = 33) | P value |
|---|---|---|---|
| Median maternal age, y (range) | 33 (24–45) | 33 (20–43) | .71 |
| Median body mass index at delivery, kg/m 2 (range) | 32.1 (17.7–53.7) | 28.9 (20.0–44.3) | .12 |
| Median gestational age at delivery, wk (range) | 34 (16–39) | 34 (19–40) | .50 |
| Median parity, n (range) | 3 (0–12) | 2 (1–8) | .66 |
| Median gravidity, n (range) | 4 (1–14) | 4 (2–11) | .46 |
| History of cesarean delivery, n (%) | 51 (93) | 28 (87) | .46 |
| Previous cesarean delivery, n (%) | .23 | ||
| 0 | 4 (7) | 4 (12) | |
| 1 | 12 (21) | 11 (33) | |
| 2 | 24 (42) | 13 (40) | |
| ≥3 | 17 (30) | 5 (15) | |
| Median time since last cesarean delivery, mo (range) | 41.5 (11–144) | 42 (12–216) | .31 |
Maternal outcomes of the 2 groups are shown in Table 2 . Patients in the multidisciplinary group had longer anesthesia time (because this included the time for stent and line placement in many cases), greater use of bipolar diathermy, and fewer attempts to remove the placenta manually ( P < .001 for all). There were significantly more cystotomies ( P = .008) in the multidisciplinary group (explained by deliberate per protocol cystotomy and excision of involved bladder dome rather than extensive dissection of bladder off the lower segment in percreta cases).
| Variable | Multidisciplinary group (n = 57) | Nonmultidisciplinary group (n = 33) | P value |
|---|---|---|---|
| Median estimated blood loss, L (range) | 2.1 (0.5–18) | 3 (0.8–14) | .025 |
| Median packed red blood cell transfusion units, n (range) | 4 (0–24) | 4.5 (1–25) | .114 |
| Packed red blood cell transfusion of ≥ 4 units, n (%) | 37 (65) | 26 (79) | .166 |
| Median hemoglobin decrease, mg/dL (range) | 1.1 (–4.6 to 5.5) | 1 (–3 to 5.1) | .760 |
| Median crystalloid transfusion, mL (range) | 4300 (1000–16,200) | 5250 (2000–17,000) | .166 |
| Median length of hospital stay, d (range) | 4 (2–12) | 4 (2–14) | .523 |
| Use of bipolar diathermy device, n (%) | 40 (70) | 0 | < .001 |
| No attempt to remove placenta, n (%) | 45 (80) | 7 (22) | < .001 |
| Median anesthesia time, min (range) | 287 (74–608) | 180 (62–398) | < .001 |
| General anesthesia after epidural, n (%) | 25 (44) | 7 (22) | .057 |
| Median neonatal birthweight, g (range) | 2400 (800–3900) | 2300 (300–3900) | .460 |
| Cystotomy and bladder repair, n (%) | 17 (30) | 2 (6) | .008 |
| Bowel injury, n (%) | 1 (2) | 1 (3) | .999 |
| Ureteral injury, n (%) | 1 (2) | 2 (6) | .550 |
| Reoperation, n (%) | 3 (5) | 1 (3) | .999 |
| Readmission, n (%) | 2 (3) | 1 (3) | .999 |
Overall, patients in the multidisciplinary group had significantly lower EBLs ( P = .025) when compared with that seen in the nonmultidisciplinary group, with a median of 2.1 L (range, 0.5–18 L) vs 3 L (range, 0.8–14 L). The median of PRBC units transfused was not statistically different (multidisciplinary group, 4 units [range, 0–24 units]; nonmultidisciplinary group, 4.5 units [range, 1–25 units]; P = .114). Postoperative hospital length of stay was no different ( P = .52). Table 3 shows a comparison of the multidisciplinary and nonmultidisciplinary groups with respect to their antenatal diagnosis. The number of patients with antenatal diagnosis was significantly higher ( P < .001) in the multidisciplinary group compared with nonmultidisciplinary group (44 [79%] vs12 [37%], respectively).
| Variable | Multidisciplinary group (n = 43) | Nonmultidisciplinary group (n = 12) | P value |
|---|---|---|---|
| Median estimated blood loss, L (range) | 2 (0.5–18) | 2.3 (1.2–5.5) | .236 |
| Median packed red blood cell transfusion units, n (range) | 3 (0–24) | 4.5 (1–13) | .536 |
| ≥4 packed red blood cell transfusion units, n (%) | 27 (61) | 8 (67) | .999 |
| Median hemoglobin decrease, mg/dL (range) | 1.1 (–4.6 to 5.5) | 0.2 (–1.3 to 5.1) | .507 |
| Median crystalloid transfusion, mL (range) | 4450 (1000–16,200) | 3900 (2800–10,000) | .753 |
| Median length of hospital stay, d (range) | 4 (2–12) | 4.5 (3–14) | .793 |
| Use of bipolar diathermy device, n (%) | 35 (79) | 0 (0) | < .001 |
| No attempt to remove placenta, n (%) | 42 (98) | 6 (50) | < .001 |
| Median anesthesia time, min (range) | 300 (74–608) | 185 (116–316) | .036 |
| Median neonatal birthweight, kg (range) | 2.3 (1.5–3.5) | 2.6 (0.5–3.3) | .787 |
| Cystotomy and bladder repair, n (%) | 16 (36) | 1 (8) | .080 |
| Bowel injury, n (%) | 1 (2) | 0 | .999 |
| Ureteral injury, n (%) | 1 (2) | 1 (8) | .386 |
| Reoperation, n (%) | 1 (2) | 1 (8) | .386 |
| Readmission, n (%) | 0 | 1 (8) | .214 |
Table 4 shows the maternal outcomes specifically in cases of deeper placental invasion (increta or percreta). EBL was significantly lower in the multidisciplinary group than in the nonmultidisciplinary group (2.1 L [range, 0.9–18 L] vs 3.5 L [range, 0.8–14 L]; P = .031). There was not a statistically significant trend to fewer transfused units of PRBCs (median, 3.5 units per case [range, 0–24 units] vs 6 units per case [range, 2–18 units]; P = .058), respectively. There was no significant difference in maternal outcomes between the multidisciplinary and nonmultidisciplinary groups when only accreta cases were compared (excluding increta and percreta). The length of hospital stay in the 2 groups was not different.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
