FormalPara Key Points

The use of opioids before cord clamping for general anesthesia in a preterm Cesarean section is controversial, based on an evaluation of the maternal benefit/neonatal risk balance.

Low-dose remifentanil before cord clamping does not cause overt neonatal respiratory depression at birth and does not improve maternal or neonatal outcomes.

Low-dose remifentanil for general anesthesia before cord clamping may be considered if necessary.

1 Introduction

Although a Cesarean section (CS) is mostly performed using regional anesthesia, general anesthesia (GA) is required in emergency situations that endanger the life of the mother and/or her fetus, such as placenta abruption, eclampsia, placenta previa with hemorrhage, fetal heart rate abnormalities, or in the case of contraindication or failure of regional anesthesia [1]. In a recent US multi-center cohort, 18% of preterm CSs were performed under GA [2]. The conventional induction of GA during a CS is a modified rapid sequence induction, usually using an anesthetic agent and a rapid-acting muscle relaxant. After cord clamping, opioids are used, and anesthesia is deepened. The use of opioids before cord clamping is controversial, based on an evaluation of the maternal benefit/neonatal risk balance. Indeed, although opioids may reduce maternal sympathetic response, their transplacental passage can induce neonatal respiratory depression [3], and safety and potential long-term consequences are poorly described [4]. Remifentanil has a rapid onset of action (1–2 min), a short half-life (3–10 min), and a rapid offset of action [5, 6]. Remifentanil crosses the placental barrier with an umbilical venous/maternal arterial ratio from 0.73 to 0.88 [7, 8]. The use of remifentanil during a full-term CS has been evaluated, but never before 37 weeks of gestation [9]. We hypothesized that in GA for a CS before 37 weeks’ gestation, the use of low-dose remifentanil prior to cord clamping could reduce the use of adjunctive anesthetic agents, thereby improving the maternal hemodynamic status and neonatal adaptation of the preterm neonate. We aimed to assess the impact of low-dose remifentanil use during GA for a CS on preterm neonates at birth.

2 Methods

2.1 Study Design

This was a prospective, single-center, phase III, hospital-based, double-blind, randomized, placebo-controlled trial conducted in the tertiary maternity hospital at Rouen University Hospital, France, between September 2014 and June 2018. The study was approved by the Ethics Committee of Rouen, France (Comité de Protection des Personnes Nord-Ouest I, number CPP 01/015/2013) and registered at ClinicalTrials.gov (NCT02029898, EudraCT 2013-001850-83, primary investigator: Fabien Tourrel, date of registration, 8 January, 2014).

2.2 Patient Selection

The inclusion criteria were a maternal age more than 18 years, a single pregnancy with an indication for a CS under GA in the context of prematurity (24–37 weeks’ gestation), in the context of emergency or not (emergency grade 1 [emergency] to 4 [elective]) [10], and eligibility for social security. Non-inclusion criteria were a maternal disease requiring opioids during induction, severe pre-eclampsia, a 14-week delay between information and actual consent, patients under guardianship, and remifentanil contraindication. Pregnant women were informed during antepartum second trimester pre-anesthetic consultations, an information leaflet was provided to all potential participants, and written consent was obtained by the research team. Parturients were recruited when GA was decided, after screening for the inclusion criteria.

2.3 Randomization and Allocation of Treatment

After informed consent, patients were block randomized to remifentanil or placebo by computer generation. The allocation was blind to the participants, the care providers, and the investigator. Only the pharmacy unit was aware of the allocation.

2.4 Study Procedures

The anesthetic monitoring included non-invasive measurement of systemic pressure every 2 min and 30 s, as well as continuous electrocardiography and pulse oximetry. Hemodynamic parameters collected in the immediate preoperative period were the reference parameters.

The dose of remifentanil was a 0.5-μg/kg bolus followed by a continuous administration of 0.1 μg/kg/min until hysterotomy. For the remifentanil group, 1 mg of remifentanil was diluted into a volume of 20 mL with saline in a syringe labeled “study drug” (1 mL = 50 μg of remifentanil). For the control group, 20 mL of saline was drawn into an identically labeled syringe. Syringe preparation was performed by a nurse not involved in the patient’s care. The visual appearance of remifentanil and the placebo after dilution was identical.

Administration of 0.01 mL/kg of the study drug followed by a continuous infusion of 0.002 mL/kg was performed. After the study drug bolus, 5 mg/kg of thiopental and 1.5 mg/kg of suxamethonium were administered. Study drug administration was stopped before the end of the infusion if there were clinical signs of poor tolerance (hypotension [mean blood pressure less than 45 mmHg], bradycardia [heart rate less than 45 beats per minute], or chest rigidity), or an anaphylactic reaction. The investigator could then unblind the drug. After cord clamping, the anesthetist decides whether to continue the remifentanil infusion or administer sufentanil.

2.5 Primary and Secondary Outcomes

The primary outcome measure was the rate of newborns with Apgar scores < 7 at 5 min. Secondary outcome measures were maternal haemodynamic parameters (systolic, diastolic, mean blood pressure, and heart rate), complications of anesthesia induction (difficult intubation defined as a failure after two attempts), use of adjuvant anesthetic agents, neonatal respiratory distress, umbilical cord pH, and lactate levels.

As done during routine care in our center, we collected the following data at corrected ages of 1 and 2 years: weight, height, and head circumference. Cerebral palsy was evaluated according to the European Cerebral Palsy Network definition [11]. Motor development and cognitive function were assessed by a routine score based on the Amiel-Tison and Denver developmental scales [12,13,14]. The psychosocial behavioral score was defined by items from the French version of the Strengths and Difficulties Questionnaire [15, 16].

2.6 Sample Size Calculation and Statistical Analysis

The sample size was based on an expected Apgar score < 7 at 5 min was 40% in the control group. A total of 60 mother-infant dyads (30 per group) was needed to detect a 28% decrease in the Apgar score (based on unpublished local data) with 80% power and a 5% significance level. The rate of neonates with an Apgar score < 7 at 5 min was estimated in each treatment group (remifentanil or placebo) and compared between the two groups using Pearson’s Chi-squared test on the intention-to-treat population, i.e., none of the neonates was excluded and the neonates were analyzed according to the randomization scheme. This unadjusted analysis was completed with a logistic regression adjusted on the gestational age at birth ([24–28], [28–32], [32–36]) and the induction time before cord clamping in minutes. Analysis of dichotomous endpoints other than the primary endpoint relied on the same methods, i.e., Pearson’s Chi-square test or Fisher’s exact test (if necessary) was used to compare observed proportions between groups (e.g., magnesium sulfate [yes/no], hypertension [yes/no], and the sex of the neonate). For quantitative endpoints (e.g., mask ventilation time, insufflation pressure), comparisons between treatment groups relied on the Student’s t-test or the Mann–Whitney’s non-parametric test as appropriate. All statistical tests used the two-sided 0.05 level as their significance threshold. For quantitative variables, means ± standard deviation or medians (first and third quartiles) are reported and for qualitative variables, the frequency and percentages are used.

3 Results

Between September 2014 and June 2018, 179 pregnant women who underwent a CS before 37 weeks’ gestation were screened for eligibility. Of these, 35 participants did not meet inclusion criteria, 33 were excluded because of an emergency status and 56 declined to participate, leaving 55 women to be randomized. After randomization, one participant no longer needed a CS, one participant withdrew consent before anesthesia, and one patient was determined to be under guardianship. Ultimately, 27 women in the remifentanil group and 25 women in the placebo group were analyzed. Each woman gave birth to a live newborn. Before discharge, two preterm newborns died in the remifentanil group and five died in the placebo group. At 1 year, eight participants dropped out in the remifentanil group, and three in the placebo group. At 2 years, six participants were lost to follow-up in each group (Fig. 1). We found no differences between the groups in terms of the general characteristics, for both the women and the newborns (Table 1).

Fig. 1
figure 1

Enrollment, randomization, and follow-up of the participants

Full size image
Table 1 Patient baseline characteristics
Full size table

3.1 Primary Outcome

The rate of newborns with an Apgar score < 7 at 5 min was not significantly different between the remifentanil (9/27, 33%) and the placebo (11/25, 44%) groups (p = 0.49, adjusted odds ratio 0.66, 95% confidence interval 0.20–2.18).

3.2 Secondary Outcomes

3.2.1 Maternal Outcomes

Hemodynamically, the anesthesia procedure was well tolerated in both groups. None of the participants required hemodynamic support with a vasopressor. Only one woman received propofol instead of thiopental to induce anesthesia (remifentanil group). The duration of anesthesia and the cumulative treatment dose until cord clamping did not differ between the studied groups. Induction complications were difficult intubations in two participants in the placebo group. Only one participant (placebo group) required mask ventilation, which the clinician attributed to chest rigidity (Table 2).

Table 2 Maternal outcomes
Full size table

3.2.2 Neonatal Outcomes

The blood cord gases were also similar for the two groups. While the use of mask ventilation was similar, the duration of ventilation was longer in the remifentanil group (394 vs 211 s in the placebo group, p = 0.01). Similarly, in the case of intubation, the duration of mechanical ventilation was longer in the remifentanil group (29 h vs 17 h in the placebo group), while surfactant administration was similar. Chest compression (4/25, 16%) and epinephrine administration (2/25, 8%) were performed only in the placebo group. Respiratory distress was distributed equally (77.8% in the remifentanil group and 92% in the placebo group, p = 0.25). We did not find more chest rigidity in the remifentanil group (28.6% vs 17.4% in the placebo group, p = 0.48). Before discharge, two neonates had died in the remifentanil group and five in the control group (7.4% vs 20%, p = 0.24) (Table 3).

Table 3 Neonatal outcomes
Full size table

3.2.3 1-Year and 2-Year Follow-Up

Eleven participants were lost to follow-up at 1 year (eight in the remifentanil group and three in the placebo group), and 12 additional participants at 2 years (six in each group). We did not find any differences in terms of motor, cognitive, behavior, sensory, sleeping, or feeding scores between the groups (Table 4).

Table 4 1-Year and 2-year follow-up
Full size table

4 Discussion

A low dose of remifentanil during GA before cord clamping in preterm CS did not result in worsening of 5-min Apgar scores compared to placebo, though an increased duration of face mask ventilation at birth was observed in the remifentanil group. Interestingly, we did not find evidence of more chest rigidity after remifentanil exposure. No statistical difference was noted in maternal hemodynamic parameters. Pediatric follow-ups at 1 and 2 years of age did not show any difference in neurosensory outcomes between the remifentanil and placebo groups.

4.1 Strengths and Limitations

The main strength of our trial is the randomized, double-blind, controlled design for an emergency procedure. Another strength is that the control group intervention was saline and not another anesthetic agent that could bias the interpretation of the impact of remifentanil.

Our results should be interpreted keeping in mind the limitations of our study. Our study was a single-center investigation that reflects our current local practices. We approached the number of participants to be included without reaching this. Difficulties relating to inclusion were mainly owing to the emergency nature of the CS, despite the information delivered during a pre-anesthetic consultation. The high rate of loss to follow-up at 1 and 2 years of age, and the small number of inclusions clearly limit the interpretation of these results. It should be noted that this study was not designed to show a difference in the long-term neurological outcome, but in neonatal cardiorespiratory adaptation at birth, and that the data at 1 and 2 years were described because they were routinely collected in our center.

4.2 Interpretation

General anesthesia for a CS remains challenging, as an acceptable depth of anesthesia needs to be obtained for the parturient with a limited risk for neonates [3]. Traditionally, opioids have been used only after cord clamping to avoid neonatal respiratory depression. It remains a matter of debate, and the use of opioids before cord clamping without any neonatal side effects could be a feasible alternative. For parturients, the addition of opioids potentially improves the depth of anesthesia [17] and prevents hypertension at intubation, especially in pre-eclampsia [18, 19]. We did not find an impact of remifentanil on maternal hemodynamic parameters. The absence of a preventive effect on the pressor response to laryngoscopy, compared with other studies that showed a pressor preventive effect of remifentanil [17], may be explained by the low dose used to prevent potential adverse neonatal respiratory effects, and the non-inclusion of parturients with pre-eclampsia. We chose not to include patients with severe pre-eclampsia because there was strong evidence that remifentanil controlled post-intubation blood pressure compared with placebo [19,20,21]. Since 2020, the Société Française d’Anesthésie et de Réanimation has recommended that patients with severe pre-eclampsia undergoing GA receive an injection of morphine during the induction to limit the hemodynamic consequences of tracheal intubation [22]. At the time the protocol was written, these practices were already in place in the department, and it did not seem ethical for a patient with severe pre-eclampsia not to receive an opioid if she was in the placebo group. We hypothesize that the antihypertensive beneficial effect of remifentanil is more readily demonstrated in patients with pre-eclampsia, who exhibit higher blood pressure. Finally, a low dose of remifentanil did not change the anesthesia procedure and appears to be safe for the parturient. Two randomized controlled trials with 40 and 50 women found that remifentanil for GA for a term CS was as safe as saline and dexmedetomidine, respectively [8, 23]. This reassuring evidence is also supported by a meta-analysis about induction opioids for a CS [24].

Given the high placental transfer of remifentanil, respiratory depression and chest rigidity are concerns at birth. In a randomized controlled trial with 40 parturients, two term newborns exposed in utero to remifentanil required naloxone and were admitted to the neonatal intensive care unit for observation [8]. Draisci et al. demonstrated that newborns in the remifentanil group had lower 1-min and 5-min Apgar scores, but still above 8 at 5 min [25]. A recent meta-analysis did not show any evidence of a deleterious impact of remifentanil on cardiorespiratory adaptation [24]. To our knowledge, there have been no studies to date showing a deleterious effect of remifentanil on newborns at birth [17, 26,27,28,29].

At birth, neonatal outcomes were broadly similar between the remifentanil and placebo groups. In the case of intubation, the duration of mechanical ventilation was longer in the remifentanil group, while surfactant administration was similar. Given the pharmacokinetic properties of remifentanil, it seems unlikely that the increase in ventilation time was related to remifentanil exposure. Antenatal remifentanil exposure did not change the cord pH and lactate levels, suggesting no adverse impact on fetus vitality.

Interestingly, we did not find that remifentanil had any impact on the occurrence of chest rigidity. Furthermore, we noted chest rigidity in 4/25 patients in the placebo group (16%). This misdiagnosis could be partly explained by the difficulty in differentiating chest rigidity from other causes of neonatal respiratory distress. However, the participants in the remifentanil group had a longer duration of mask ventilation (394 ± 274 s vs 211 ± 171 s in the control group, p = 0.01), without other respiratory morbidities. The absence of a difference in neonatal adaptation is of paramount importance as GA was most often indicated in vulnerable critical situations at a high risk of asphyxia at birth, neonatal mortality, morbidity, or long-term neurodisability [30].

The results of the follow-up at 1 and 2 years did not show any difference in terms of growth and neurosensory outcomes. However, this study cannot conclude because of the high rate of lost to follow-up at 1 and 2 years and the modest number of inclusions. Nevertheless, in a murine model with intracortical injections of ibotenate, remifentanil exerted a beneficial effect against excitotoxicity associated with a reduction in the brain lesion and prevention of some behavioral impairments [31]. In humans, this potential neuroprotective effect could be of interest, as GA is performed in emergency situations that increase the risk of neurodevelopmental sequels in preterm infants, but a larger multi-center randomized study is needed.

5 Conclusions

A low dose of remifentanil before cord clamping during GA in preterm CS does not cause overt neonatal respiratory depression at birth. Further studies are needed to assess the clinical effect of remifentanil on long-term neurodevelopmental outcomes in infants.