Spina bifida is the most serious neural tube defect compatible with life.1 Unprotected foetal neural tissue in the spine undergoes progressive neurodegeneration in the spinal cord, with impairment in brain development and hindbrain herniation or Chiari II malformation probably due to pressure disturbances resulting from continuous leakage of cerebrospinal fluid (CSF) through the defect. Foetal therapy, by closure of the defect, has proved to be useful in reversing hindbrain herniation and decreasing the need for postnatal CSF shunting.2
Although the first clinical trial of prenatal myelomeningocele surgery used an open surgery approach,2 other groups advocate minimally invasive surgery with the aim of reducing maternal morbidity and obstetric complications. Most studies of surgery for neural tube defects relate to perinatal and neonatal outcomes, and only two refer to anaesthetic management.3,4 One of them is a review of anaesthetic management during the peri-operative period during open surgery of neural tube defects.3 The other study is a retrospective review of 61 mothers who underwent percutaneous fetoscopic surgery.4
The aim of this study was to report on the anaesthetic management of women undergoing open or fetoscopic surgery for foetal neural tube defects. Specific objectives of the study are to compare the types and doses of anaesthetic drugs; the use of fluid therapy and vasoactive drugs; maternal haemodynamic changes, to describe maternal blood gas changes during insufflation of the amniotic cavity with carbon dioxide during fetoscopic surgery and to record maternal and foetal intra-operative and postoperative complications.
Materials and methods
Study design and setting
This was a retrospective cohort study carried out at the Vall d’Hebron University Hospital, Barcelona, between 2011 and 2016. Ethical approval for this study (PR (AMI)26/2016) was given by the Ethical Committee CEIC Vall d’Hebron Research Institute, Barcelona, Spain on 16 February 2016.
We recruited pregnant women undergoing intra-uterine foetal surgery for open neural tube defects. Inclusion criteria were maternal age 18 years or more, a singleton pregnancy, American Society of Anesthesiologists Class I and II, gestational age between 18 and 26 weeks, an isolated neural tube defect between T1 and S1 and cerebellum herniation. Exclusion criteria were BMI higher than 35 kg m−2, other foetal malformations, genetic or chromosomal abnormalities with poor prognosis, kyphosis (more than 30°) and maternal disorders contra-indicating surgery.
All women were assessed before surgery and gave informed consent for the procedure. Two units of packed red blood cell units were cross-matched. The night before surgery oral lorazepam (1 mg), oral omeprazole (20 mg), subcutaneous molecular low weight heparin (40 mg) and rectal indomethacin (50 mg) were given to all. Prior to the procedure, the dose of indomethacin (50 mg) was repeated, cefazolin (2 g, i.v.) was administered and deep venous thrombosis stockings were put on. Patients were positioned with left uterine displacement on the operating table. Temperature control was achieved by using a warming blanket, covering the exposed zones and administering warm fluids.
The two surgical techniques used were open and fetoscopic surgery. For open surgery, after a Pfannenstiel incision, the uterus was exteriorised and a hysterotomy was performed using a resorbable stapler.
For the fetoscopic procedure, after a Pfannenstiel incision, the uterus was exteriorised. A needle was inserted into the amniotic cavity under ultrasound visualisation to ensure its correct placement and to avoid puncturing the foetus, placenta or umbilical cord. Some of the amniotic fluid was then extracted and heated CO2 was insufflated into the amniotic cavity to improve the visualisation of amniotic contents, with the lowest possible partial pressure (6 to 9 mmHg). Two or three ports were used. The amniotic fluid was kept warm and sterile, and finally re-infused into the amniotic cavity after removing the remaining CO2.
Measures to reduce the risk of maternal or foetal CO2 embolism included ensuring the correct location of the needle by insertion under ultrasound guidance and aspirating amniotic fluid prior to the insufflation of CO2, and also by maintaining a low pressure of CO2 during surgery.
The shift from open surgery to fetoscopic surgery came as a result of gaining broad experience from fetoscopic surgery in a sheep model,5,6 and at the request of a woman who declined blood transfusion on religious grounds. Reduced blood loss was expected with the fetoscopic approach. The hospital's Ethics Committee allowed us to modify the technique and fetoscopic surgery became the technique of choice in our centre from March 2013.
A lumbar epidural catheter was inserted, and after the test dose, a single 0.25% bupivacaine bolus was administered to achieve a sensory block to the level of T10. No more was given until surgery was completed.
Anaesthesia was induced using the rapid sequence technique with midazolam [2 mg, intravenous (i.v.)], fentanyl (2 μg kg−1, i.v.), thiopental (3 to 5 mg kg−1, i.v.) up to December 2011 or propofol (2 mg kg−1, i.v.) after December 2011, vecuronium (0.1 mg kg−1, i.v.) or rocuronium (0.8 mg kg−1, i.v.) after July 2013.
A cuffed endotracheal tube was inserted. Anaesthesia was maintained with sevoflurane [medium alveolar concentration (MAC) 1.5 to 2%], remifentanil infusion (0.05 to 0.2 μg kg−1 min−1) and vecuronium or rocuronium, and the dose was adjusted according to the train-of-four (TOF), aiming at TOF less than 0.3 to ensure a sufficient degree of abdominal wall relaxation to facilitate the surgical procedure.
Mechanical ventilation in volume-controlled mode was used. Initially, the ventilatory settings were as follows: tidal volume 8 ml kg−1, respiratory frequency 12 min−1, positive end-expiratory pressure (PEEP) 5 cmH2O and inspiratory-to-expiratory time ratio 1 : 2. Subsequently, the settings were adjusted according to the arterial blood gas measurements. Neostigmine (until July 2013) or sugammadex (after July 2013) were used for reversal of neuromuscular blockade.
Drugs for foetal anaesthesia included a combination of vecuronium (0.2 mg kg−1) or rocuronium (0.8 mg kg−1), fentanyl (15 μg kg−1) and atropine (20 μg kg−1) by intramuscular injection into the foetal gluteus muscle immediately before surgery.
Maternal intra-operative monitoring
Two i.v. cannulae were inserted. A 20-gauge cannula was inserted in a radial artery to monitor arterial pressure and take blood samples before, during and after surgery for arterial blood gas analysis. A central venous 7-French bilumen catheter was inserted in the basilic or cephalic veins to monitor central venous pressure (CVP).
Additional maternal intra-operative monitoring included end-tidal CO2 (ETCO2), bispectral index (BIS), neuromuscular relaxation by train-of-four, oesophageal temperature and hourly urine output. After March 2014, the Vigileo system™ (Vigileo; Edwards Lifesciences, Irvine, CA, USA) was used to estimate cardiac output index (CI), stroke volume variation (SVV) and systemic vascular resistance index (SVRI).
SBP, DBP, mean (MBP) arterial pressure and heart rate (HR) were measured at specific times during surgery in both open and fetoscopic procedures:
- Open surgery: beginning of anaesthesia (1), beginning of surgery (2), uterine exteriorisation (3), hysterotomy (4), beginning of foetal surgery (5), end of foetal surgery (6), uterine return to abdominal cavity (7), end of surgery (8), end of anaesthesia (9).
- Fetoscopy: beginning of anaesthesia (1), beginning of surgery (2), uterine exteriorisation (3), beginning of fetoscopy (4), foetal surgery (5), end of fetoscopy (6), uterine return to abdominal cavity (7), end of surgery (8), end of anaesthesia (9).
Fluid therapy and management of maternal hypotension
In 2011, the fluid therapy protocol was relatively liberal, ranging from 6.7 to 7.5 ml kg−1 h−1, and was initially intended to prevent and treat hypotension. After March 2014, and following the introduction of goal-directed fluid therapy (GDFT), fluid administration was based on haemodynamic variables SVV, CI, SVRI and MBP. Crystalloid fluid (Ringer solution; Grifols Laboratory, Barcelona, Spain) and colloid fluid (hydroxyethyl starch; Voluven, Fresenius Kabi, Spain) were used.
Maternal hypotension was defined as a reduction of SBP below 20% of the baseline value or MBP under 65 mmHg. For the first 10 women, ephedrine was the preferred vasoconstrictor. Subsequently, the choice of drug depended on the basal HR: phenylephrine was given if the HR was 65 bpm or more, and ephedrine if HR was less than 65 bpm. The dose of ephedrine was 5 to 10 mg per bolus. Phenylephrine was administered as a continuous infusion at an initial rate of 25 μg min−1 and then titrated according to effect.
Foetal monitoring and uterine relaxation
Foetal heart rate (FHR) was monitored by ultrasound. In the first 25 cases, only episodes of bradycardia (FHR under 110 bpm) were registered, but in subsequent cases, the value of the FHR was registered at specific times of surgery.
Uterine relaxation was assessed by the surgeon. Additional volatile anaesthetic (MAC = 2), nitroglycerine bolus (50 to 100 μg, i.v.) or continuous perfusion (0.5 to 1 μg kg−1 min−1) were administered if uterine relaxation was insufficient.
For tocolytic purposes, at the end of the foetal surgery and before uterine return to abdominal cavity, an atosiban (Tractocile) bolus (6.75 mg) followed by a continuous perfusion (3 mg h−1) was administered. At the same time, for tocolytic and analgesic purposes, local anaesthetic was continuously infused (bupivacaine 0.1%, 5 to 7 ml h−1) through the epidural catheter.
The aim of postoperative management was good pain control, tocolysis and prevention of maternal pulmonary oedema.
An epidural catheter was used to administer bupivacaine 0.1% for 48 to 72 h. i.v. paracetamol (2 g every 8 h) was also administered, and additional i.v. morphine (bolus of 0.05 mg kg−1, followed by a continuous infusion 1 to 2 mg h−1) was administered if required.
Special attention was paid to the detection and treatment of early signs of pulmonary oedema.
A continuous infusion of atosiban was maintained during the early days and oral nifedipine (20 mg every 8 h) was added if necessary.
Subcutaneous molecular low-weight heparin (40 mg) was administered to prevent deep vein thrombosis.
All data were collected and entered anonymously into an electronic database for further analysis.
Categorical data were reported as frequency and percentage. Differences between the types of surgery were estimated by χ2 or Fisher's test, as appropriate. Continuous variables were given as median [range]. The Mann–Whitney test was used to assess differences between groups.
Mean and confidence intervals at each surgical timepoint were used to assess the longitudinal evolution of haemodynamic variables. A linear regression model with random effects for repeated measurements was used to assess changes from the initial value. The model was adjusted with an interaction to assess if the evolution was different according to the type of surgery, and the prediction was represented in a graph.
The statistical software Stata 13.1 (StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP.) was used for all data analysis. Statistical significance level was set at P value less than 0.05.
Twenty-nine foetuses with neural tube defects (21 cases of myelomeningocele and eight of myelocele) between January 2011 and December 2016 were operated on. Seven (24.1%) women underwent open surgery and 22 (75.9%) fetoscopic surgery of whom 12 (54.5%) had the patch & glue technique and 10 (45.4%) had the skin closure technique.
Most patients were nulliparous, under 30 years of age and had no co-existing disease. Maternal characteristics, gestational age at surgery, operative time and hospital stay were similar in both groups (Table 1).
Anaesthetic drugs during the intra-operative period
All patients had a combined epidural and general anaesthetic technique. Intra-operative drugs are summarised in Table 2.
After the first two cases of open surgery, in which thiopental was used for induction, the anaesthetic protocol was modified and propofol was used in the subsequent 27 cases. Likewise, vecuronium was replaced by rocuronium after the first 10 cases, and neostigmine by sugammadex, owing to a change in the protocol. No statistical differences were found in the dose of opioids or neuromuscular blocking agents. No patient required morphine during the postoperative period. The median dose of sevoflurane was significantly lower in fetoscopic surgery than in open surgery (P = 0.001). In both groups, the maximum dose of sevoflurane occurred during the period when the uterus was exteriorised. Nitroglycerine was required in fewer cases in fetoscopic surgery compared with open surgery (P = 0.001), although no differences in dose were found in those cases wherein the drug was needed (P = 0.693).
Fluid therapy and vasoconstrictor drugs during the intra-operative period
Fluid therapy and vasoconstrictor drugs are summarised in Table 3. Colloid volumes were lower in fetoscopic surgery than in open surgery (P = 0.036), and no significant differences were found in the volume of crystalloids between both groups.
No intra-operative or postoperative blood transfusion was required. One case with a fetoscopic approach required a blood transfusion before surgery owing to a low haemoglobin concentration (8.2 g dl−1). No significant differences were found in haemoglobin prior to surgery or haematocrit between groups, but the postoperative haemoglobin and haematocrit were significantly lower in the open surgery group (P = 0.004 and P = 0.005, respectively).
The change in haemoglobin concentration was not associated with the surgery time (r = -0.003, P = 0.987), maximum MAC of sevoflurane (r = -0.200, P = 0.308), need for a second intra-operative tocolytic drug (r = -0.087, P = 0.869) or gestational age at the time of surgery (r = 0.221, P = 0.258).
Vasoconstrictor drugs were administered in all cases. No statistical differences were found in the dose of ephedrine or phenylephrine between groups. The maximum dose of ephedrine was given during uterine exteriorisation, and the highest dose of phenylephrine was given at the beginning of foetal surgery (Fig. 1).
Use of vasoconstrictor drugs was not associated with the surgery duration (r = -0.394, P = 0.231 and r = 0.182, P = 0.455, respectively), maximum MAC of sevoflurane (r = 0.133, P = 0.696 and r = 0.358, P = 0.133, respectively), bleeding (r = -0.286, P = 0.394 and r = 0.006, P = 0.980, respectively), need for intra-operative nitroglycerine (r = 0.316, P = 0.684 and r = 0.866, P = 0.333, respectively) or use of postoperative tocolytic drugs (r = -0.273, P = 0.417 and r = 0.284, P = 0.239, respectively).
GDFT protocol was applied after the first 20 cases, and a significant reduction in the volume of colloids given was observed in this group (median 1.26 [0 to 1.90] ml kg−1 h−1 compared with the non-GDFT group median 2.54 [1.65 to 2.80] ml kg−1 h−1, P = 0.002). No differences were observed in crystalloid volume (6.89 [6.70 to 7.45] ml kg−1 h−1 in the GDFT group vs. 7.32 [6.70 to 7.50] ml kg−1 h−1 in the non-GDFT group, P = 0.564) or dose of phenylephrine (754 [310 to 900] μg, in the GDFT group, vs. 666 [600 to 1000] μg, in the non-GDFT group, P = 0.156) between both groups.
Maternal haemodynamic changes during the intra-operative period
Figures 2 and 3 show the evolution of maternal SBP, DBP, MBP and HR during surgery for both techniques. No significant differences were found in baseline SBP (P = 0.251), DBP (P = 0.146), MBP (P = 0.120) and HR (P = 0.759) between open and fetoscopic techniques. Lowest median SBP and MBP reached during surgery were 89 [70 to 117] mmHg and 62 [46 to 73] mmHg respectively, and were significantly lower during open surgery than fetoscopic surgery (P = 0.009 and P = 0.018, respectively). No significant differences were found in the lowest values of DBP (P = 0.108), where the median was 48 [34 to 58] mmHg. The lowest median HR was 65 [51 to 81] bpm and was significantly lower in the fetoscopic group (P = 0.007). No significant differences were found in the highest HR (P = 0.646), where the median was 90 [72 to 139] bpm. Median CVP in the open surgery group was 10 [7 to 19] and 8 [6 to 9] mmHg in the fetoscopic group (P = 0.142).
In the 10 patients in whom the Vigileo system was used for advanced haemodynamic monitoring the initial values of CI, SVRI, SVV were 2.6 [2.2 to 3.13] l min−1 m−2, 1450 [1130 to 2320] dynes-s-cm−5 m−2 and 8%, respectively. No significant differences were found during the intra-operative period in any of these variables (P = 0.566, P = 0.731 and P = 0.466, respectively).
Maternal blood gas changes during CO2 insufflation into the amniotic cavity during fetoscopy
In patients undergoing fetoscopic surgery, arterial gas measurements were made before, during and after CO2 insufflation into the amniotic cavity (Table 4) and no significant differences were found.
Maternal and foetal complications
Three cases (10.3%) of postoperative acute pulmonary oedema were identified, one case in the open surgery group (14.3%) and two in the foetoscopic group (9.1%; P = 0.694). No significant association between acute pulmonary oedema and fluid administration was observed (P = 0.552) nor with crystalloid volume (P = 0.508), colloid volume (P = 1.0), sevoflurane dose (P = 0.654), use of nitroglycerine (P = 0.999) or number of postoperative tocolytic drugs (P = 0.847). The three cases of pulmonary oedema were diagnosed 2 days after surgery; all three were suspected after a reduction in oxygen saturation and confirmed with chest radiograph, but patients were asymptomatic. All three resolved with diuretics (furosemide) and oxygen and none of the cases required mechanical ventilation or additional treatment.
Pregnant women who underwent open surgery required more postoperative tocolytic drugs (three drugs in all seven cases) than those who underwent fetoscopic surgery (two drugs in 21 cases and three drugs in one case) (P < 0.001). None of the women required morphine for postoperative pain management.
No cases of foetal bradycardia were recorded.
The main finding of this study is that open surgery is associated with more maternal haemodynamic changes and higher doses of halogenated anaesthetic and tocolytics agents than fetoscopic surgery. Also, advances in haemodynamic monitoring and GDFT protocols have brought about changes in the peri-operative treatment of pregnant women.
Our protocol considered that a MAC equal to or greater than 1.5 was necessary to achieve adequate uterine relaxation, although the maximum MAC of sevoflurane was significantly higher in open surgery, probably because it is more invasive.3,7,8 The use of volatile agents in combination with tocolytic agents such as atosiban allows a reduction in the dose (MAC < 1) necessary for optimal uterine relaxation with the percutaneous fetoscopic approach.4 In our series, this dose would probably be insufficient. The highest concentration of sevoflurane coincided with the period of uterine exteriorisation, a circumstance that does not occur in the percutaneous technique. In addition, most women having open surgery required an additional tocolytic drug, nitroglycerine, as reported by other authors.3
Mean blood pressure was stable during surgery. This was achieved through pharmacological anticipation of the stages of surgery and with continuous monitoring of mother and foetus. A reduction in blood pressure was observed in both groups during uterine exteriorisation. This reduction was more significant in the open surgery group, as previously reported.4,9,10 A combination of several factors that coincide at this moment of the procedure would explain this finding: the need for intense uterine relaxation, surgical manipulation and a higher heat loss. Consumption of vasoconstrictor drugs was also higher during this period of surgery.
Maternal HR was lower in fetoscopic surgery than in open surgery. We think that this could be secondary to phenylephrine11 that reduces HR, and was the most used drug in fetoscopic surgery, but no episodes of maternal bradycardia were reported. No elevation of maternal MBP was observed at the time of phenylephrine perfusion. Our goal was to keep maternal and foetal haemodynamic stability, with MBP above 65 mmHg, trying to minimise the side effects of phenylephrine related to HR and cardiac output.12
Mean crystalloid volume administered in our study was 1500 ml, lower than the 2000 ml recommended in the literature.13 The significant reduction in fluid volume probably followed the introduction of a GDFT protocol. In a retrospective study with 61 pregnant women who underwent percutaneous fetoscopic surgery, the administered fluid volume was lower when GDFT was applied.4 However, there is insufficient evidence to claim that in these patients, restrictive fluid therapy improves peri-operative outcomes.3
In the open surgery group, significant lower postsurgery concentrations of haemoglobin and haematocrit were found. These differences could be due to the more invasive hysterotomy required for open surgery, and higher uterine relaxation that increases the risk of bleeding. However, no blood transfusions were required in any of the groups, in common with previous reports.4,9,10,14
In the 22 cases of fetoscopic surgery, heated CO2 was insufflated into the amniotic cavity to improve the visualisation of amniotic contents, with the lowest possible partial pressure (6 to 9 mmHg). This technique has been previously described in a sheep model by our group.5,6,15 Kohl et al.16 observed that the use of CO2 pressure between 7 and 15 mmHg is not associated with foetal acidosis or macroscopic or microscopic neurological damage. The same authors performed 37 procedures in humans and concluded that this technique is well tolerated for both mother and foetus, and improves surgical conditions.17 In our study, we preferred to heat any fluid or gas that was to be introduced in the uterine cavity to avoid potential foetal acidosis, hypoxaemia or bradycardia. No episodes of bradycardia were reported during FHR monitoring. No signs of hypercapnia or other blood gas changes were observed in the mother during fetoscopy.
The prevalence of acute pulmonary oedema was 10%, similar to other studies.2,14 Gestational physiologic changes, tocolytic administration and liberal fluid therapy are the main causes of acute pulmonary oedema. In a prospective study with 13 pregnant women undergoing foetal surgery, a catheter was placed in the pulmonary artery. The authors concluded that there was an increase in extravascular lung water and pulmonary vascular permeability with stable cardiopulmonary function, indicating an increased risk of maternal pulmonary oedema.18
Foetal anaesthesia achieved its goal, as no episodes of foetal bradycardia or foetal movements were observed in our study.
One of the concerns in foetal surgery is the effect of volatile agents on the developing foetal brain. The real effect of anaesthetic exposure on foetal brain remains unknown. Some studies in animal models have reported neuronal damage.19 There are several retrospective studies on anaesthetic exposure in infants under 4 years of age, suggesting adverse neurodevelopment in learning and language ability particularly following prolonged or repeated exposure. With regard to foetal anaesthetic exposure, there are no retrospective studies to guide practice.20
The main limitation of our study is the small sample size. However, the number of published studies of anaesthetic management in surgery for neural tube defect is also small.3,4,7 Another limitation is the retrospective nature of the study, and also that during the study period the anaesthetic protocol and use of drugs was changed, making comparisons of haemodynamic management, doses of anaesthetic drugs and monitoring variables difficult. The Vigileo system has been used in pregnant women by other investigators,21,22 although we acknowledge that it has not been validated in for this group of patients. In our study, it was rather used as a trend monitor for each individual patient.
Fetoscopic surgery for neural tube defects seems to be associated with less maternal haemodynamic changes than open repair. This is probably due to a lower dose of halogenated and other tocolytic agents. The use of vasocontrictor drugs was related to the uterine exposure, but not to the surgical technique.
CO2 insufflation during fetoscopy did not affect maternal blood gas analysis greatly and no episodes of foetal bradycardia were reported.
Acknowledgements relating to this article
Assistance with the study: none.
Financial support and sponsorship: none.
Conflicts of interest: none.