A retrospective review of 10-year trends in general anesthesia for cesarean delivery at a university hospital: the impact of a newly launched team on obstetric anesthesia practice

Background The indications for general anesthesia (GA) in obstetric settings, which are determined in consideration of maternal and fetal outcome, could be affected by local patterns of clinical practice grounded in unique situations and circumstances that vary among medical institutions. Although the use of GA for cesarean delivery has become less common with more frequent adoption of neuraxial anesthesia, GA was previously chosen for pregnancy with placenta previa at our institution in case of unexpected massive hemorrhage. However, the situation has been gradually changing since formation of a team dedicated to obstetric anesthesia practice. Here, we report the results of a review of all cesarean deliveries performed under GA, and assess the impact of our newly launched team on trends in clinical obstetric anesthesia practice at our institution. Methods Our original database for obstetric GA during the period of 2010 to 2019 was analyzed. The medical records of all parturients who received GA for cesarean delivery were reviewed to collect detailed information. Interrupted time series analysis was used to evaluate the impact of the launch of our obstetric anesthesia team. Results As recently as 2014, more than 10% of cesarean deliveries were performed under GA, with placenta previa accounting for the main indication in elective and emergent cases. Our obstetric anesthesia team was formed in 2015 to serve as a communication bridge between the department of anesthesiology and the department of obstetrics. Since then, there has been a steady decline in the percentage of cesarean deliveries performed under GA, decreasing to a low of less than 5% in the latest 2 years. Interrupted time series analysis revealed a significant reduction in obstetric GA after 2015 (P = 0.04), which was associated with decreased use of GA for pregnancy with placenta previa. On the other hand, every year has seen a number of urgent cesarean deliveries requiring GA. Conclusions There has been a trend towards fewer obstetric GA since 2015. The optimized use of GA for cesarean delivery was made possible mainly through strengthened partnerships between anesthesiologists and obstetricians with the support of our obstetric anesthesia team.


(Continued from previous page)
Conclusions: There has been a trend towards fewer obstetric GA since 2015. The optimized use of GA for cesarean delivery was made possible mainly through strengthened partnerships between anesthesiologists and obstetricians with the support of our obstetric anesthesia team.
Keywords: Obstetric anesthesia, Cesarean delivery, Placenta previa, General anesthesia, Multi-disciplinary collaboration Background Ensuring safety for women receiving cesarean delivery remains a continuing challenge for anesthesiologists in general hospitals with an obstetric service. With sustained commitment to the mission of dealing with a variety of surgical procedures, the department of anesthesiology also plays a role in providing the highest standard of care during the labor and delivery process, which can be accompanied by potential risks for both the pregnant woman and her baby.
The choice of anesthetic technique for cesarean delivery should be determined with consideration of the degree of emergency in relation to maternal and fetal status and comorbidities, as well as of the difficulty or expected duration of procedures [1,2]. It is generally accepted that neuraxial anesthesia is the preferred choice over general anesthesia (GA), with advantages including avoidance of the potential complications with the maternal airway and neonatal exposure to anesthetic drugs used during induction and maintenance of GA [3,4]. Neuraxial anesthesia has the additional merit of allowing a mother to see her baby in the moments after birth, and reducing the need for neonatal respiratory support. On the other hand, however, administration of GA remains the more appropriate option under certain circumstances, especially when there is a perceived lack of time to apply neuraxial techniques [1,5]. In addition to emergent situations, GA may be adopted in women with a potentially life-threatening pregnancy-related condition such as placenta previa [2], but the indications for GA are complicated by the difficulty of identifying parturients at high risk for obstetric hemorrhage prior to cesarean delivery. Concerns with respect to encountering unexpected difficult maternal airway have been a deterrent to the frequent use of GA in obstetric settings [6].
As one of the largest and most comprehensive university hospitals in Japan, the University of Tokyo Hospital aims to offer a comprehensive obstetrics program for women in pregnancy with approximately 1000 deliveries per year. The growing need for perinatal care services for high-risk pregnancies, along with rising recognition of the anesthesiology subspecialty of obstetric anesthesia, prompted us, in 2015, to assemble an advanced team dedicated to clinical obstetric anesthesia practice. Although the team consists of staff anesthesiologists engaged in both obstetric anesthesia and a wide range of perioperative care, they have been in close partnership with multi-disciplinary professionals to contribute to better management for women before, during, and after childbirth.
Previously, we experienced a case of placenta accreta with intraoperative massive hemorrhage with a blood loss of almost 20,000 mL, which consequently led to the greater adoption of GA for women with placenta previa undergoing cesarean delivery at our institution. However, recent years have witnessed a progressive change in our obstetric anesthesia practice, especially since our obstetric anesthesia team was launched in 2015. The purpose of this study was to investigate the trends in cesarean deliveries carried out under GA, focusing on practices instituted by our obstetric anesthesia team.

Methods
With the approval of our Institutional Review Board (#2203-(6)), we retrospectively analyzed our clinical database regarding the obstetric anesthetic care and management during 10 calendar years from January 1, 2010 through December 31, 2019. We reviewed the medical records of all parturients who received GA to determine the indications for cesarean delivery that required GA, whether elective, emergent, or urgent. The urgent cases in the current study were defined as corresponding to category 1 cesarean section described in the National Institute for Health and Clinical Excellence guidelines [7].
Interrupted time series analysis was used to evaluate how the launch of our obstetric anesthesia team impacted on our practice after 2015. Quarterly trends in the proportion of GA cases were analyzed to ensure an adequate number of time points both before and after the intervention. Residual autocorrelation was tested using the Durbin-Watson test. A P value < 0.05 was considered statistically significant. All statistical analyses were performed using SAS ver. 9.4 (SAS Institute, Cary, NC).

Trends overview
Overall, the annual number of cesarean deliveries was on the rise during the period 2010-2019 at our institution, in tandem with the increasing number of total deliveries (Fig. 1). The proportion of cesarean deliveries remained relatively high throughout the years covered by the study, ranging from 29.2 to 34.9% with a gradual increase over time (Table 1).

General anesthesia for cesarean delivery
As recently as 2014, more than approximately 10% of cesarean deliveries were conducted under GA, with an annual number of 27-39 cases, which peaked in 2011 at 39 cases (Table 1). Since 2015, however, there has been a steady decline in the number of cesarean deliveries requiring GA to 15 in 2019. As a percentage, administration of GA peaked at 14.5% in 2012, but has trended downward, deceasing to less than 5% in 2018 and 2019 (Fig. 2). Elective and emergent cesarean deliveries comprised the majority of all GA cases until recently (Table 1).
Interrupted time series analysis revealed that there was a significant decrease in the proportion of obstetric GA after the launch of our obstetric anesthesia team in 2015 (P = 0.04) (Fig. 3). The graphic provides a visual presentation of decreased use of obstetric GA in the postintervention period 2015-2019 (blue). All postintervention data points lie below the trend line extrapolated from the pre-intervention, 2010-2014, data (red). There was no significant change in slope from pre-to post-intervention (P = 0.74). The Durbin-Watson statistic showed that there was no evidence of autocorrelation (DW 1.64, P = 0.95).

General anesthesia for pregnancy with placenta previa
There has been an upward trend in the number of women with placental abnormalities at our institution, exceeding 20 per year (Supplementary Table 1). GA was chosen for pregnancy with placenta previa in the years 2010-2014, a time when placenta previa was the main indication for elective and emergent cesarean deliveries, with proportions ranging from 47.2 to 78.3% (Table 2). However, the annual number of parturients with placenta previa who received GA for elective or emergent cesarean delivery was on the decline ( Fig. 4), accounting for 28.6% of the elective and emergent cases in 2019 (Table 2). During the study period, the median blood loss in cesarean deliveries for placenta previa was 1920 mL with an interquartile range (IQR) of 1423 to 2896 mL, although there were 9 sporadic cases with a blood loss exceeding 4000 mL (Fig. 5). Despite these cases of massive hemorrhage, the use of obstetric GA has been on a downward trend in recent years.

Urgent cesarean delivery
Unlike the substantial decrease in elective and emergent cases, we have experienced a somewhat variable but non-declining number of urgent cesarean deliveries each year (Fig. 6). Overall, non-reassuring fetal status was responsible for many of urgent cases ( Table 3). The decision to delivery interval (DDI) differed depending on the case, with a median value of 19 min (IQR 15-25 min) in the years 2010-2019 ( Table 4). The mean neonatal umbilical arterial pH (UApH) was approximately 7.20, albeit with the exception of 2010, 2011, and 2014, whereas the median 1-and 5-min Apgar scores varied throughout the study period (Supplementary Table 2). The relationship between DDI and UApH is illustrated in Fig. 7.

Neuraxial anesthesia for pregnancy with placenta previa
Recently, we have adopted the use of neuraxial anesthesia over GA for pregnancy with placenta previa. A total of 75 women with placenta previa without suspected placenta accreta (63 elective and 12 emergent) received neuraxial anesthesia for cesarean delivery in 2016-2018. The use of combined spinal-epidural anesthesia comprised the majority of these cases (88.0%).
Internal iliac artery balloon catheters were placed in 6 cases prior to elective cesarean delivery. During the same period, GA was administered in 13 women with placenta previa (9 elective and 4 emergent). The 9 parturients scheduled for elective cesarean delivery had preoperative placement of internal iliac artery balloon catheters. Total hysterectomy was additionally Table 2 Indications for elective and emergent cesarean deliveries performed under general anesthesia   2010  2011  2012  2013  2014  2015  2016  2017  2018  2019   Total cases  24  36  34  29  23  18  13  12  12  7 Abnormal Maternal factors included threatened premature labor, failure of labor process, hypertensive disorders of pregnancy (HDP), syndrome of hemolysis, elevated liver enzymes, and low platelets (HELLP), impaired placental function, intrauterine infection, and other maternal comorbidities Fetal factors included diaphragmatic hernia, umbilical cord problems, prolapse of fetal extremity, and other fetal comorbidities

Mortality and morbidity
We observed no anesthesia-related mortality during the entire study period. There was one case of amniotic fluid embolism associated with disseminated intravascular coagulation, which was followed by intraoperative massive hemorrhage. This case required emergent use of extracorporeal membrane oxygenation during the surgery. No case of difficult intubation was identified in which more than three attempts at laryngoscopy were made. We recorded 19 cases where the lowest oxygen saturation (SpO 2 ) was below 90% following the rapid sequence induction of GA (Supplementary Table 3), but the total duration of desaturation was less than 1 min in every case. During the 10-year period, the parturients who received GA for cesarean delivery, including the one with amniotic fluid embolism, were discharged without any anesthesia-related morbidity. Anesthetic agents and muscle relaxants used for induction Table 5 represents trends in sedative-hypnotic agents and muscle relaxants that were used at our institution for induction of GA for cesarean delivery. Although the use of thiopental and succinylcholine accounted for a great majority in 2010, the following years saw a significant shift towards the use of propofol and rocuronium. The combination of these agents has been common in recent years, especially since 2016. The use of vecuronium was occasionally seen until 2013, but it has rarely been used in the past 6 years.

Main findings
Our review of obstetric anesthesia practice from 2010 to 2019 at our institution revealed a noticeable decline in the number of cesarean deliveries performed under GA, particularly in the recent 5 years. The reduction in obstetric GA was accompanied by a decrease in the number of parturients with placenta previa who received GA for cesarean delivery. On the other hand, however, urgent GA use for cesarean delivery was inevitable in some unplanned situations, indicating its importance throughout the study years.

Maternal
Impact of the launch of our obstetric anesthesia team on clinical practice In 2011 the Perinatal Center was officially established, and two years later, in 2013, one of our staff anesthesiologists took up his new duties with obstetric anesthesia.
The year 2015 saw the launch of our obstetric anesthesia team, which served as a catalyst that provided team members with wide opportunities to discuss practices with obstetricians. Their successful attempts at identifying women with pregnancy-related risk factors prior to cesarean delivery led to a rethinking of the indications for GA in obstetric practice. There was a shift towards using GA only for parturients with placenta accreta, who have high risk of intraoperative massive hemorrhage, prolonged operating times, and the need for application of extended surgical procedures. This was largely responsible for the reduction in GA administration for pregnancy with placenta previa. The overall use of GA in obstetric practice has been declining in concert with the increasing adoption of neuraxial anesthesia for cesarean delivery [12,13]. However, there has been controversy and interinstitutional variation regarding the preferred option of anesthetic technique for cesarean delivery complicated by placenta previa [14,15]. As part of optimizing the indications for GA in an effort to avoid GA-related complications, our current clinical practice encourages the use of neuraxial techniques in parturients without any abnormally invasive placentation. This trend, which has been driven by our obstetric anesthesia team, is in accordance with the Fig. 7 The relationship between decision to delivery interval and neonatal umbilical arterial pH. There were 59 urgent cesarean cases in the study years 2010 to 2019, including two cases of twin deliveries. Decision to delivery interval (DDI) was not determined in two cases, and neonatal umbilical arterial pH (UApH) was not recorded in four cases. Analysis of a total of 55 cases suggested a slight inverse correlation between DDI and UApH evidence that neuraxial anesthesia can be used for cesarean delivery even in the presence of placenta previa [16][17][18]. Unlike those days when GA was virtually the only choice for pregnancy with placenta previa, we now administer GA only in parturients with high risk of hemorrhage, who were prenatally diagnosed as having placenta accreta with a high degree of certainty. This is related in part to the finding that the amount of intraoperative blood loss was larger in cases using GA. Our new strategy provides a major contribution to optimizing the selection of cesarean deliveries that deserve the use of GA.

Activities of our obstetric anesthesia team
Since its formation in 2015, our obstetric anesthesia team has acted as a communication bridge between the department of anesthesiology and the department of obstetrics. The team consists of a team leader and several team members, who take turns staying in the obstetric ward on weekdays, assisting obstetricians in performing epidural anesthesia for labor and delivery. Once a week, the team members participate in a multi-disciplinary conference with obstetricians and neonatologists where the group shares detailed information on parturients with comorbidities. They discuss the individual's current physical status, the preferred method of delivery and its timing, and the appropriate choice of anesthetic technique in terms of maternal and neonatal outcome. This has made it possible to keep the members informed of the most updated medical condition of high-risk parturients.
In close cooperation with the department of obstetrics, the team leader has been engaged in making clinical assessments of all women who undergo elective cesarean delivery. Obstetricians are encouraged to consult him about parturients scheduled for vaginal delivery with maternal or fetal complications, should they need support from our obstetric anesthesia team. Through discussion with the parturient and obstetricians in outpatient settings, the leader makes clear in advance what type of anesthetic technique to choose for scheduled cesarean delivery. This newly established system has benefited us even when the parturient should undergo an emergency cesarean delivery, as it has made it easier to quickly grasp her medical condition and problems.

Decision to delivery interval
At any hospital that has an obstetric service, there is a constant need for being prepared for obstetric emergencies. When there is an immediate threat to the life of the mother or baby, a 30-min DDI has been suggested as the time frame within which delivery should be accomplished [19]. Although the recommended DDI is widely accepted as a pragmatic rule [20], whether it improves maternal and fetal outcomes remains unclear [21,22]. Additionally, the feasibility of the 30-min rule in urgent settings has been questioned in some studies [23,24]. During the study period, we have achieved a DDI of less than 30 min in most urgent cesarean deliveries, despite the difficulty in predicting when they would occur. This suggests the practicability of the recommended DDI at a university hospital, where the department of anesthesiology is expected to provide anesthesia for every surgical patient all day and night. It must be remembered that our success in maintaining the optimum DDI has also been owed to obstetricians, neonatologists, nurses, midwives, and all medical staff members, including those in the operating room, who work in cooperation with anesthesia providers in an endeavor to keep prepared for emergency or urgency for the sake of maternal and fetal safety.
Theoretically, a shorter DDI may save the distressed fetus from being exposed to intrauterine hypoxia [22]. In our analysis, the value of UApH varied markedly regardless of DDI, probably because each case had different reasons and situations for requiring an urgent cesarean delivery. Even in urgency, however, every possible effort must be made to ensure the earliest possible delivery, given that DDI was associated, though less directly, with the severity of fetal acidosis.

Current trends and future concerns
Throughout the study period, difficulty with intubation was not common among our 267 parturients receiving GA. The minimum SpO 2 was lower than 90% in 19 cases without prolonged oxygen desaturation after the induction of GA. This was partly because we are also committed to medical residency education, and the first attempt at laryngoscopy was made by a resident anesthesiologist.
Ensuring opportunities for training with obstetric GA is still a matter of concern. Some studies warn that trainee anesthesiologists may have little or no clinical experience with GA administration for cesarean delivery, as GA in obstetric practice has been largely replaced by the widespread use of neuraxial techniques [13,25]. The observed trend towards fewer cesarean deliveries conducted under GA may foreshadow an eventual lack of exposure to obstetric GA during residency training. The situation could be even more serious, considering that anesthesia providers at our institution are continually expected to deal with unpredictable cases that necessitate urgent administration of GA. Awareness of the necessity for maintaining obstetric airway management skills should be raised.

Anesthetic agents and muscle relaxants used for induction
Rapid-sequence induction using thiopental and succinylcholine was the previous standard for induction of GA for cesarean delivery [26]. More recently, propofol and rocuronium, which are commonly used for induction of GA in non-obstetric settings, have mostly replaced them, serving as the agents of choice for obstetric GA. Succinylcholine, a depolarizing neuromuscular blocking agent, has the fastest onset and shortest duration, and it remained the preferred choice for difficult maternal airway management, despite its fatal side effects including hyperkalemia and malignant hyperthermia [26]. However, the situation has changed since rocuronium, a nondepolarizing neuromuscular blocking agent, and sugammadex became available for clinical use. A high dose (1.0-1.2 mg/kg) of rocuronium creates excellent intubating conditions in 60 s [27], and 16 mg/kg sugammadex can provide a rapid reversal of profound neuromuscular blockade [28]. Manufacturing approval for sugammadex was granted in Japan in 2010, and there has been a notable shift towards the use of rocuronium since the year 2011. Given recent trends, our prediction is that the combination of propofol and rocuronium will continue to be the standard, replacing thiopental-succinylcholine induction.

Limitations
Some limitations of our study should be mentioned. First, it is to be noted that there have been changes and improvements in anesthetic techniques and drugs that have created a paradigm shift over the years in the way obstetric anesthesia is provided. Our analysis highlights the maturing process where delivery of better perinatal care, including optimization of obstetric GA administration, was enabled through the contributions of our obstetric anesthesia team. Even then, we recognize that this achievement was driven in part by a growing body of evidence on obstetric anesthesia practice. Second, providing detailed information on comfort and recovery of parturients was difficult in the current study. The database used for our analysis deals mainly with obstetric anesthetic care and management in the operating room. Further studies will be needed to assess the effect of our renewed practice on obstetric women with respect to post-operative pain management, feasibility of early ambulation, and hospital length of stay.

Conclusion
In summary, our retrospective analysis of cesarean deliveries performed under GA over the past 10 years has revealed reduced GA administration in obstetric women, which was associated with a decrease in the number of parturients with placenta previa receiving GA. In a successful attempt to identify high-risk parturients, our obstetric anesthesia team has made possible the optimized use of GA for cesarean delivery, contributing to a reduction in use of GA in obstetric settings. Continuing efforts are required to further strengthen the relationship with multi-disciplinary team members, so that we can take the initiative to play the role of peripartum physicians to ensure better obstetric management for women in pregnancy.