Interventions to improve fetal heart rate patterns during open myelomeningocele repair

Introduction

In recent years, knowledge pertaining to neural tube defects has made great strides. Since the initial studies on their prevention, improvements in their detection and diagnosis have facilitated a new era in fetal therapy, significantly reducing the morbidity and mortality rates and improving the postnatal quality of life among patients with neural tube defects [1]. Myelomeningocele is responsible for orthopaedic, vesico-intestinal, and neurological sequelae, which are largely due to the high incidence of hydrocephalus associated with the Chiari II malformation [1]. It affects approximately 1 out of every 1000 to 2000 live births; approximately 80% of patients require surgery, and 46% experience complications in the first year of life [2]. Since the publication of the Management of Myelomeningocele Study (MOMS), fetal therapy centres around the world have broadened their research, and more centres have begun to perform open myelomeningocele repair [3]. Despite the increasing experience, there remain many challenges in the implementation of this highly complex procedure.
The frequent risk of fetal bradycardia is a major intraoperative concern; fetal heart rate (FHR) monitoring is essential for the identification of these changes and the rapid re-establishment of the haemodynamic state. As previously reported, the neurosurgical stage was shown to be a higher risk for fetal bradycardia, especially when nerve endings and the spinal cord were exposed [4]. Therefore, this study sought to evaluate FHR patterns after multifactorial changes to open fetal surgery and, in doing so, to contribute to improving the safety and quality of the procedure in addition to improving the understanding of fetal physiology.

Material and methods

A prospective, cross-sectional study was performed from July 2016 to July 2017. It included women who were pregnant with fetuses with myelomeningocele, who underwent open surgery. Patients were selected from the Centre for Fetal Medicine within the Department of Obstetrics of the Federal University of São Paulo (UNIFESP), the São Paulo Centre for Fetal Medicine, and Santa Joana Maternity Hospital in São Paulo, Brazil. The pregnant women who agreed to volunteer for the study signed an informed consent form, and this study was approved by the Research Ethics Committee of UNIFESP. The surgery was performed in operating rooms at UNIFESP’s São Paulo Hospital and at Santa Joana Maternity Hospital.
Fetal heart rate was evaluated at specific time points during surgery: before the mother received anaesthesia (1), after the mother received anaesthesia (2), neurosurgery (beginning of skin manipulation) (3a), during neurosurgery at exposure of nerve endings and the spinal cord (3b), during neurosurgery at synthesis (3c), and at the end of the surgery (4). FHR was obtained by means of pulsatile Doppler wave velocimetry, with measurements at the closure of the mitral valve at an interval of at least three consecutive beats. The FHR range considered normal in this study was 110 to 160 bpm (range, 24 to 27 weeks). Measurements were performed within 3-4 min of each stage of the surgery.
The ultrasound machine used in the operating room was a Logiq P5 (GE, Medical Systems, Milwaukee, WI, USA) with a multifrequency convex probe (2-5 MHz). The ultrasound probe was wrapped in a sterile laparoscopic cover, and gel was then placed directly on its surface. The transducer was positioned such that an adequate image could be obtained without hindering the surgery. All the images were saved on the device’s hard drive and printed in real time to be added to each patient’s medical record. The inclusion criteria of this study were a singleton pregnancy with a live fetus, myelomeningocele with a top margin between the first lumbar vertebra and the first sacral vertebra, evidence of brain stem herniation, gestational age between 24 and 27 weeks, and normal fetal karyotype. The exclusion criteria were fetuses with structural and/or chromosomal abnormalities, fetal kyphosis, increased risk of preterm labour (short uterine cervix < 25 mm, and/or prematurity in previous gestation), low-lying placenta, body mass index (BMI) > 35 kg/m², and hysterotomy of the anterior segment.
After a previous study by the authors, in which a reduction in fetal heart rate was observed during the neurosurgical stage, as were some instances of bradycardia, changes in the fetal surgery protocol were suggested to minimise the previous findings [4]. We decided against the use of intramuscular fentanyl for fetal anaesthesia because the surgical team initially experienced increases in cases of fetal bradycardia with no additional benefit to the surgical outcome or the well-being of the fetus. Temperature control became more rigorous. Using a laser thermometer, the temperature of the uterus while exposed to the external environment was controlled by a team member present outside the operating field. To prevent the uterine temperature from reaching levels below 30°C, the uterine surface was irrigated with a saline solution heated to 37°C. The temperature marked on the thermometer in the operating room itself was also monitored.
The team then wrapped the exposed uterus in a sterile plastic cover (Figure 1), leaving only the region of the hysterotomy exposed, to further minimise heat loss (Figure 2). Manual compression of the uterus (a subjective parameter) to display the operating field to the neurosurgeon has also become a focus in this procedure because it is associated with fetal chest compression, decreases in fetoplacental circulation, and fetal response to surgical stress.
In cases of maternal hypotension, ephedrine or metaraminol was administered if necessary, to preserve fetoplacental circulation. The plan for cases of fetal bradycardia included the infusion of atropine into the mother’s circulatory system at a dose of 0.02 mg/kg in addition to epinephrine at 1 µg/kg. All the data were processed in an Excel 2007 worksheet (Microsoft Corp., Redmond, WA, USA). We determined the mean ± standard deviation (SD) of the FHR of each stage and used analysis of variance (ANOVA) with repeated measures to assess differences between the surgical stages. To compare different stages of the fetal surgery, we used the mean difference with a 95% confidence interval (CI), and the differences were analysed using Tukey’s multiple comparison test.

Results

In a total of 37 open fetal myelomeningocele surgeries, the descriptive statistics tests showed that the mean maternal age was 32.6 years, mean gestational age at the time of surgery was 25.9 weeks, mean body mass index was 28.0 kg/m², and mean fetal weight was estimated at 830.9 g. The mean surgical time was 127.3 min. Delivery occurred after an average of 235.5 days of pregnancy. Mean fetal birth weight was 2225 g, and delivery occurred, on average, 56.1 days after surgery (Table 1). ANOVA with repeated measures was used to assess the differences between the following stages of surgery. The mean FHR before maternal anaesthesia (1), after maternal anaesthesia (2), during neurosurgery at the beginning of skin manipulation (3a), during neurosurgery at exposure of nerve endings and the spinal cord (3b), during neurosurgery at synthesis (3c), and at the end of the surgery was 138.6 ±9.1, 138.4 ±9.7, 132.8 ±9.1, 127.7 ±11.4, 131.4 ±8.1, and 132.7 ±8.5 bpm, respectively (p < 0.001) (Table 2).
We used Tukey’s multiple comparison test to show two-tailed comparisons (Table 3). Based on the 95% CI and 5% of level of significance, marked highlights were determined to be significant in the comparisons. There was a significant reduction in FHR between stages 2 and 3a and between stages 3a and 3b; it stayed at the same level until stage 4 (Figure 3). Unlike in previous protocols for fetal surgery, no episodes of fetal bradycardia were recorded in this series. ANOVA with repeated measures showed that mean values ​​for room temperature (in °C) during the different surgical stages (Table 4) were 23.2 (stage 1), 22.9 (stage 2), 22.9 (stage 3a), 22.9 (stage 3b), 22.9 (stage 3c), and 22.0 (stage 4). Therefore, the temperature of the room decreased only at the end of the surgery (between stages 3c and 4) (Figure 4). Mean uterine temperature (in °C) was 31.83 in stage 3a, 31.87 in stage 3b, and 31.26 in stage 3c (Table 4). There were no significant changes during the three neurosurgery stages (Figure 5).

Discussion

Maintaining the haemodynamic balance of the fetus during open surgery is complex and depends on multiple factors such as maternal anaesthesia, fetal anaesthesia, fetal response to surgical stress, uterine manipulation by fetal chest compression, possible decreases in fetoplacental circulation, and possible heat loss [4]. It is evident that the mother’s haemodynamic state influences fetal circulation. Van de Velde et al. reported that the use of vasopressors in cases of maternal hypotension during fetal procedures may be necessary [5]. Therefore, the surgical team’s monitoring should be as vigilant as possible.
As demonstrated by Boat et al., the administration of desflurane along with intravenous anaesthetics is beneficial in reducing bradycardia, left ventricular dysfunction, and the need for fetal resuscitation during open surgery when compared with that of only high doses of desflurane. The protocol included propofol (150-250 mg/kg/min) and remifentanil (0.2-0.5 mg/kg/min) during the beginning of procedure, and then administration of desflurane (1-1.5 MAC) for uterine relaxation. At the time the uterus was exposed, propofol infusion rates were decreased (50-75 mg/kg/min) and the concentration of desflurane was gradually increased (2-2.5 MAC) until appropriate uterine relaxation was provided [6]. This finding was also described by Nguyen et al., who reported fetal bradycardia during surgery after the inhalation of high doses of anaesthetics [7]. In our patients, a 3% sevoflurane concentration was elevated to 5% when the uterus was removed from the abdominal cavity, and in this series, no similar episodes were observed.
It is known that the fetal response to stress and pain involves the activation of the hypothalamic–pituitary–adrenal axis and the release of stress-related hormones, such as noradrenaline, cortisol, beta-endorphin, and corticotropin [8]. As Anand et al. describe, this theory is limited, and often these stress responses cannot be interpreted as a painful sensation and do not present cortical involvement [9]. The viability of an intact system to transmit pain from the peripheral receptor to the cerebral cortex must be complete for the fetus to experience pain. The development of these peripheral receptors begins at the 17th week of gestation and is complete by the 20th week. This mechanism makes the sensation of pain possible; however, the serotonin-releasing downstream inhibitory system does not develop completely until after birth [10, 11].
Although the MOMS trial protocol [3] consisted of the use of intramuscular fentanyl in the fetus, in our experience in recent years, we have observed an increase in the incidence of fetal bradycardia associated with the use of this drug, and atropine is much more commonly used for vagolysis. This finding differs from those reported by Richick et al., who found that fentanyl and vecuronium increased FHR and decreased cardiac output [12]. Bellieni et al. also believed that because anaesthetic levels are lower in fetal circulation than in maternal circulation, analgesics and anaesthesia should be administered directly to the fetus [13]. Volatile anaesthetics are, in fact, known to pass through the placenta. This is due to the properties of these drugs, which include low-molecular-weight, high-fat solubility, and a non-ionic nature. Although it is believed that the use of fentanyl may reduce the fetal response to surgical stress, the confirmation of this effect and a full understanding of fetal physiology still require further studies [14].
Adequate control of body temperature during surgical procedures is essential [15]. It is known that hypothermia increases cardiac morbidity, increases oxygen consumption, is associated with coagulation disorders, and is also related to a high incidence of operative wound infection [16]. Mann et al. also note their concern over cases of maternal hypothermia even after fetal surgery [17]. In our practice, the patient is covered with upper-body blankets to minimise heat loss. The operating room remains at a stable temperature during critical periods, and after the uterus is removed from the abdominal cavity, it is protected with a plastic cover, further minimising heat loss and thus protecting maternal–fetal haemodynamics. The continued use of uterine irrigation with heated saline during neurosurgery is an important step for maintaining a safe thermal balance. We observed that the uterine surface temperature did not change significantly during the three neurosurgical stages, and the room temperature decreased only during the final stage of the surgery after the uterus had already been returned to the abdominal cavity. Although a reduction in the fetal heart rate was observed during the neurosurgical stage (when the nerve endings and spinal cord are exposed), there were no episodes of fetal bradycardia, which previously had been more frequent.

Conclusions

Although several factors are involved in the pathogenesis of fetal cardiovascular disorders, we believe that these changes allow for better haemodynamic control of the fetus and, in doing so, increase the safety of the procedure. Of course, further studies are needed to improve our understanding of fetal behaviour during open surgery as well as that of the physiological aspects involved in this complex scenario. Maybe the assessment of FHR variability would be important in this context.

Acknowledgments

We would like to thank the Department of Obstetrics of the Federal University of São Paulo (UNIFESP) and the Association for the Improvement of Higher Education Personnel (CAPES) for enabling this study. We would also like to thank Santa Joana Maternity Hospital for supporting our team and the growth of open fetal surgery in Brazil.

Conflict of interest

The authors declare no conflict of interest.