Lipman, Steve MD; Daniels, Kay MD; Cohen, Sheila E. MBChB, FRCA; Carvalho, Brendan MBBCh, FRCA
From the Departments of Anesthesia and Obstetrics & Gynecology, Stanford University School of Medicine, Stanford, California.
The authors thank Julie Arafeh, RN, MSN, and Andrea Puck, MS, RN, CNS, for their contributions to the OBSim program; Sara Snyder, RN, and Michelle Woodfall, RN, MS, CEN, CCRN, for helping conduct the drills in this study; Anne Marie Oakeson, MSN, RNC, RDMS, and Maurice Druzin, MD, for their support of multidisciplinary team training on our unit; and Brent Seibel, Department of Obstetrics & Gynecology, University of Florida College of Medicine, Jacksonville, for developing and sharing the abdominal overlay model used in these drills.
Portions of this data were reviewed during an oral presentation for the Best Paper Competition during the annual meeting of the Society for Obstetric Anesthesia and Perinatology, April 13–17, 2011, Lake Las Vegas, Nevada.
Corresponding author: Steve Lipman, MD, Department of Anesthesiology, MC5640, 300 Pasteur Drive, Stanford, CA 94305; e-mail: firstname.lastname@example.org.
Financial Disclosure The authors did not report any potential conflicts of interest.
In the event of maternal cardiopulmonary arrest, the American Heart Association recommends incision within 4 minutes and delivery of the fetus within 5 minutes through perimortem cesarean delivery if gestational size approximates 20 weeks or greater and vaginal delivery is not immediately possible.1,2
Despite consensus for early delivery during maternal arrest,1–6 there have been no studies to examine the optimal location for perimortem cesarean delivery. If the arrest takes place in the labor room, should perimortem cesarean delivery occur there or should the patient first be moved to the operating room?6 Clinical randomized controlled trials to determine the optimal location for perimortem cesarean delivery do not exist and are unlikely to ever be conducted as a result of ethical and logistic restraints.
The aim of this study was to compare the labor room and operating room settings for perimortem cesarean delivery during simulated maternal arrest drills on our labor ward. We hypothesized that moving to the operating room during perimortem cesarean delivery would delay incision and negatively affect maternal resuscitation.
MATERIALS AND METHODS
From May through September 2010, 15 teams were asked to participate in this randomized controlled study of perimortem cesarean delivery on the labor ward at Lucile Packard Children's Hospital in Stanford, California. The drills were part of an ongoing initiative called OBSim, an obstetric crisis team training program created and codirected by two of the authors (K.D., S.L.). OBSim is an integral part of the quality assurance and quality improvement process on our labor ward and is exempt from oversight per the institutional review board at Stanford University School of Medicine.
The obstetric teams participating in the series of perimortem cesarean delivery drills each consisted of four nurses, two obstetricians, one anesthesiologist, and neonatal staff. Before scenario start, all obstetric team participants were informed: 1) that the scenario was a perimortem cesarean delivery drill; 2) of their assigned setting (labor room or operating room) for perimortem cesarean delivery; and 3) that the time to delivery would be recorded. In addition, to identify and correct systems issues existing between labor ward and neonatal intensive care unit personnel, an actual neonatal team from our institution participated in the majority of cases. The neonatal teams did not know drills were being conducted and were the only participants who were naive to the scenario (per their request). A clinical nursing specialist from the trauma and emergency medicine services responded to all the drills when the adult code team was called. Support staff (unit secretaries and obstetric surgical technicians) participated in all the drills as per standard unit practice. The primary outcome measure was the time required to make skin incision in the labor room compared with the operating room. Secondary outcome measures included the times required to contact the neonatal and adult code teams, place the automatic external defibrillator, resume compressions after analysis, control the airway, and perform several other interventions.
The scenario involved a 36-year-old, healthy, gravida 2 para 1 woman presenting in active labor at 7-cm cervical dilation attempting a vaginal birth. Maternal monitors included noninvasive blood pressure cuff, heart rate monitor, and a pulse oximeter. Fetal monitors included external fetal heart rate monitor and tocograph. The parturient had no known drug allergies, a reassuring airway examination, an 18-gauge peripheral intravenous line, and an epidural catheter in situ with ongoing labor analgesia. The primary nurse was present before the start of each scenario in the labor room and was allowed to organize the room per her usual practice. She was also informed that the patient had just sustained a cardiopulmonary arrest from probable amniotic fluid embolus. To eliminate confounding that might have occurred as a result of delayed diagnosis among the primary nurses of each team, time zero was defined as the moment the primary nurse indicated she was ready to begin (by counting down “3, 2, 1, start”). Fetal heart rate, which had demonstrated moderate variability and a baseline of approximately 140 beats per minute before the maternal arrest, decreased to 60 beats per minute after the arrest but before the start of the scenario. The manikin used for the parturient was a Noelle birthing manikin.
Other drill team members were located together nearby, had no other patient care responsibilities, and responded immediately when called for help. The operating room was located approximately 35 meters from the labor room. A comprehensive review of maternal resuscitation in the obstetric setting was provided immediately before the scenario. This review involved a slide presentation and a cognitive aid poster, which were created from a pilot course on obstetric life support currently being conducted at our institution entitled OBLS.7
Group randomization allocation for the perimortem cesarean delivery setting was done using computer-generated random number allocation. All drills were videotaped and these were reviewed to confirm time intervals. Two stopwatches were used to measure the following time intervals during the scenario from time zero to: 1) start of chest compressions; 2) endotracheal intubation; 3) call for the neonatal team; 4) call for the adult code team; 5) code cart arrival in the room; 6) initiating uterine displacement; 7) placement of backboard under the patient; 8) placement of the defibrillator on patient; 9) resumption of compressions after rhythm analysis and shock; 10) skin incision; and 11) delivery. Of the 154 total interventions required over the course of the study, there were three omissions. Team 6, randomized to the labor room, did not use a backboard for chest compressions. Teams 10 and 11, both randomized to the operating room, called the neonatal intensive care unit team but not the adult code team for help. These omissions were deemed to have little effect on time to incision. The same scenario and manikin were used for all 15 drills.
We recorded demographic information from the participating personnel including the number of people on each team, their medical specialty, years of experience, whether they had received advanced cardiac life support training and were currently certified, if they had previously participated in simulator training, and whether they had ever rendered care during an actual maternal arrest.
We conducted 15 perimortem cesarean delivery drills to accommodate all 60 labor and delivery nurses at our institution with four nurses participating in each scenario. Previous timed OBSim drills for stat cesarean delivery (for uterine rupture) on our labor ward indicated a median (interquartile range) of 9 minutes 27 seconds (118 seconds) for relocating to the operating room, induction of anesthesia, intubation, prep, drape, gown, and skin incision.8 Based on those data, a sample size (n=7 per group) would provide 85% power to detect a 30% difference in the primary outcome of time to incision (α=0.05).
Demographic and outcome data are summarized with descriptive statistics. Results are expressed as the mean±standard deviation, median (interquartile range; range), and percentages, as appropriate. Outcome measures of interest between the two groups were compared using Student's t test for normally distributed variables and Mann–Whitney test for nonparametric comparisons. Normal distribution was determined by visual inspection QQ plots combined with Kolmogorov–Smirnov tests. Analyses were performed with SPSS 11.0 for Windows statistical package. P<.05 was considered statistically significant.
Data from 14 teams were analyzed for primary and secondary outcomes. Team 8 (randomized to the labor room) was excluded from analysis as a result of a protocol violation involving communicating with nonparticipating front desk personnel during the drill. Demographics of the remaining 14 teams are outlined in Table 1. There were no discrepancies greater than 3 seconds between the two timers using stopwatches. There were a total of eight discrepancies in all the drills. All eight discrepancies were resolved by studying the videotape; none of these eight involved incision time.
The median (interquartile range, range) from time zero until surgical incision was 4:25 (3:59–4:50, 3:27–7:04) minutes in teams randomized to the labor room and 7:53 (7:18–8:57, 4:49–12:45) minutes in teams randomized to the operating room (P=.004). Only two (29%) of the labor room teams and none of the operating room teams made incision within 4 minutes. Times required to reach specific resuscitation milestones are documented in Table 2 and demonstrate delays in the operating room group in a number of areas. The mean±standard deviation transfer time required from leaving the labor room to reaching the operating room door in the group assigned to the operating room was 50±13 seconds.
Vertical rather than Pfannenstiel incisions were used in six of seven (86%) in the labor room group compared with three of seven (43%) in the operating room group (P=.094). Seventy-one percent of teams (five of seven teams) in the labor room group used uterine tilt compared with 86% of teams (six of seven) in the operating room group. However, several of the operating room teams, which used uterine displacement initially before leaving the labor room, demonstrated delays in reinstating uterine tilt after transport to the operating room, making between-group comparisons difficult. Seventy-one percent of teams (five in each group) used backboards (P=1.0 between groups).
The key finding of this study is that moving patients to the operating room from the labor room during simulated perimortem cesarean delivery significantly delayed incision time and other critical tasks during maternal resuscitation. Teams randomized to perform perimortem cesarean delivery in the operating room required nearly twice as long to make incision as those in the labor room.
Whether to perform perimortem cesarean delivery in the labor room or move to the operating room is controversial.6,9 Proponents of moving to the operating room emphasize sterile conditions before incision, optimal lighting, familiar surroundings, access to necessary equipment, and surgical technicians to present instruments.9 However, sterility is a secondary concern in the face of an arrest, and ample lighting, oxygen delivery systems, and suction exist in the labor room. Patients in cardiac arrest possess no arterial blood pressure; therefore, multiple instruments, sutures, and electrocautery to control bleeding (as well as medications to induce anesthesia) are unnecessary. The only piece of equipment necessary for perimortem cesarean delivery is a scalpel.10 If the parturient regains spontaneous circulation postdelivery, uterine bleeding and infection prophylaxis can be addressed after transport to the operating room.11
The need for urgent delivery in the event of a maternal arrest is not controversial.1–6,10–13 The American Heart Association endorses this intervention and recommends activation of the emergency cesarean team at the onset of maternal arrest to facilitate initiation of perimortem cesarean delivery by 4 minutes if there is no return of spontaneous circulation.2 In one review of perimortem cesarean delivery (38 cases total), no woman deteriorated as a result of cesarean delivery.5 Two thirds of patients in whom vital signs were documented had immediate improvement of blood pressure, the return of spontaneous circulation, or both (“often in dramatic fashion”) after delivery.5 Other reports in the literature4–6,11,13 and media15 demonstrate rapid cesarean delivery may be the cornerstone in increasing maternal (and fetal) survival. In a report of 22 closed insurance claims of maternal cardiopulmonary arrest during regional analgesia in labor, the only woman who survived without permanent neurologic deficits underwent immediate perimortem cesarean delivery in her labor room; all the other patients were first transported to the operating room.6 Yet, many providers remain unfamiliar with both the recommendation for rapid delivery and other pregnancy-specific changes in resuscitation.8,16–18 Forty percent of obstetricians and anesthesiologists and 30% of emergency physicians in a tertiary care center were unaware of the recommendation for delivery within 5 minutes in the event of maternal arrest.16
Simulation is embedded in nearly every high-risk, high stakes domain (eg, NASA, the military, the aviation, and nuclear industries). Multiple state, national, and international medical organizations (eg, The California Maternal Quality Care Collaborative, The Joint Commission, and the Confidential Enquiries into Maternal and Child Health) specifically recommend simulation for obstetric emergencies.21–24 Our results support the use of these recommendations, in particular, that each institution develop and periodically “conduct drills … to train staff in local protocols and to identify and correct systems problems preventing optimal care.”21
Because cardiac arrests occur very infrequently in pregnant women, this series of simulated perimortem cesarean delivery drills allowed us to study questions that could not be addressed with routine clinical research. However, the drills were different than an actual event. No family members were represented, drill team members were stationed together nearby, the code team was on site for immediate response, the manikin facilitated transport (weighing only 36 kg), the delivery location was already determined, timing started only after the primary nurse knew an arrest had already occurred, and the drills were performed during normal working hours with staff who were not on duty (eliminating any competing patient care issues and mitigating any decrements in human performance associated with night call).25 Our teams had significant prior clinical experience with operating room transport on our compact labor ward and the majority (76 of 97 participants) had been involved in previous labor and delivery drills. Finally, all obstetric teams were instructed before start of the drill that the scenario was perimortem cesarean delivery and that they would be timed. This abundance of favorable conditions would never occur in reality, suggesting that the time intervals measured in this study were likely markedly shorter than could be expected during an actual event. Despite the advantages delineated here, only two teams in the labor room group and none in the operating room group were able to make the incision during the recommended 4 minutes from maternal arrest.
In conclusion, a randomized controlled study of actual perimortem cesarean delivery is both logistically and ethically prohibitive and will likely never occur. In this randomized controlled study of simulated perimortem cesarean delivery, the incision occurred approximately 3.5 minutes faster in the labor room than in the operating room group. Notably, delivery within 5 minutes was challenging in either location despite optimal conditions, raising concerns that the 5-minute recommendation for delivery in the event of maternal arrest is not feasible in most circumstances and should be considered a goal and not a “standard of care.” Our findings strongly suggest delivery would require more than 5 minutes in an actual arrest and support performing perimortem cesarean delivery in the labor room (if that is the site of maternal arrest) rather than relocating.
1. American Heart Association, Part 10.8: cardiac arrest associated with pregnancy. Circulation 2005;112:IV-150–3.
2. American Heart Association, Part 12.3: cardiac arrest associated with pregnancy. Circulation 2010;122:S833–8.
3. Marx GF. Cardiopulmonary resuscitation of late-pregnant women. Anesthesiology 1982;56:156.
4. Katz VL, Dotters DJ, Droegemueller W. Perimortem cesarean delivery. Obstet Gynecol 1986;68:571–6.
5. Katz V, Balderston K, DeFreest M. Perimortem cesarean delivery: were our assumptions correct? Am J Obstet Gynecol 2005;192:1916–20; discussion 1920–1.
6. Lofsky A. Doctors company reviews maternal arrests cases. Anesthesia Patient Safety Foundation (APSF) Newsletter. 2007. p. 28–30.
7. Lipman S, Daniels KI, Arafeh J, Halamek LP. The case for OBLS: a simulation-based obstetric life support program. Semin Perinatol 2011;35:74–9.
8. Lipman S, Daniels KI, Carvelho B, Arafeh J, Harney K, Puck A, et al.. Deficits in the provision of cardiopulmonary resuscitation during simulated obstetric crises. Am J Obstet Gynecol 2010;203:179.e1–5.
10. Warraich Q, Esen U. For perimortem caesarean section, the surgical knife is the most important instrument. BJOG 2010;117:768.
11. McDonnell NJ. Cardiopulmonary arrest in pregnancy: two case reports of successful outcomes in association with perimortem caesarean delivery. Br J Anaesth 2009;103:406–9.
12. Dijkman A, Huisman CM, Smith M, Schutte JM, Zwart JJ, van Roosmalen JJ, et al.. Cardiac arrest in pregnancy: increasing use of perimortem caesarean section due to emergency skills training? BJOG 2010;117:282–7.
13. Warraich Q, Esen U. Perimortem caesarean section. J Obstet Gynaecol 2009;29:690–3.
14. Yeomans E, Gilstrap L 3rd. Physiologic changes in pregnancy and their impact on critical care. Crit Care Med 2005;33(suppl):S256–8.
16. Cohen SE, Andes LC, Carvalho B. Assessment of knowledge regarding cardiopulmonary resuscitation of pregnant women. Int J Obstet Anesth 2008;17:20–5.
17. Einav S, Matot I, Berkenstadt H, Bromiker R, Weiniger CF. A survey of labour ward clinicians' knowledge of maternal cardiac arrest and resuscitation. Int J Obstet Anesth 2008;17:238–42.
18. Leighton BL. Why obstetric anesthesiologists get sued. Anesthesiology 2009;110:8–9.
19. Ristagno G, Tang W, Chang YT, Jorgenson DB, Russell JK, Huang L, et al.. The quality of chest compressions during cardiopulmonary resuscitation overrides importance of timing of defibrillation. Chest 2007;132:70–5.
20. Hazinski MF, Nadkarni VM, Hickey RW, O'Connor R, Becker LB, Zaritsky A. Major changes in the 2005 AHA Guidelines for CPR and ECC: reaching the tipping point for change. Circulation 2005;112(suppl):IV206–11.
21. The Joint Commission. Preventing maternal death. Sentinel Event Alert, 2010, Vol. 44.
22. The Joint Commission. Preventing infant death and injury during delivery. Sentinel Event Alert, 2004. Vol. 30.
23. Shields L, Veile JC. Postpartum hemorrhage: lessons learned from other states. In: CMQCC Obstetric Hemorrhage Toolkit, Obstetric Hemorrhage Care Guidelines and Compendium of Best Practices Reviewed by CADPH-MCAH: 12/21/09, California Maternal Quality Care Collaborative (CMQCC).
24. Al-Foudari KE, Catling S. Saving mothers' lives: lessons for anaesthetists. (CEMACH 2003–5). Crit Care Pain 2010;10:81–7.
25. Smith Coggins R, Rosekind MR, Buccino KR, Dinges DF, Moser RP. Rotating shiftwork schedules: can we enhance physician adaptation to night shifts? Acad Emerg Med 1997;4:951–61.