Results
During the study period, 263 caesarean sections were performed. General anaesthesia was used in 47 (17.9%) cases and neuraxial anaesthesia was used in 214 (81.3%) cases. Two parturients received both general and spinal anaesthesia. Since the reasons for using both techniques were not clear, these parturients were excluded from the analysis. In the neuraxial anaesthesia group, simple epidural anaesthesia was performed in 24 (11%) cases and spinal anaesthesia with or without epidural anaesthesia in the remaining 190 (89%) cases. The characteristics of the study patients are shown in Table 1. The general anaesthesia group included more parturients with NYHA class III–IV disease, mWHO class IV disease, systemic ventricle EF <50%, and emergency caesarean section. In the past, parturients more frequently received general anaesthesia at delivery. More recently, parturients frequently received neuraxial anaesthesia at delivery. At delivery, gestational age was lower in the general anaesthesia group than in the neuraxial anaesthesia group. Indications for caesarean section differed significantly between the 2 groups. The amount of missing data was expressed in Table S1.
Postpartum Outcomes
Postpartum outcomes are shown in Table 2. There were no deaths or respiratory events. Cardiovascular events were more common in the general anaesthesia group (14.9%) than in the neuraxial anaesthesia group (7.9%), but this difference was not significantly different (P =0.16). Heart failure occurred in parturients with Eisenmenger syndrome (n=1), transposition of the great arteries (TGA) repaired with the Jatene procedure (n=2), TGA functionally repaired with the Mustard procedure (n=1), repaired ventricular septal defect (VSD) (n=2), bicuspid aortic valve (BAV) with aortic regurgitation (n=1), and unrepaired Epstein anomaly (n=1). Pulmonary hypertension occurred in parturients with Eisenmenger syndrome (n=3), repaired atrioventricular septal defect (n=1), and repaired VSD (n=1). Arrhythmia occurred in parturients with Fontan circulation (n=2), repaired tetralogy of Fallot (n=2), BAV with aortic dilation (n=1), coarctation of the aorta (n=1), and Kawasaki disease with coronary aneurysm (n=1). Heart failure occurred more frequently in parturients with higher NYHA class or mWHO class. Three parturients in the general anaesthesia group required postpartum transfusion for intraperitoneal bleeding after administration of anticoagulants because of a mechanical valve.
Results from mixed-effects logistic regression models with a random intercept for each mWHO class for maternal cardiovascular risk are shown in Table 3. Compared with neuraxial anaesthesia, general anaesthesia was not significantly associated with composite postpartum cardiovascular events (odds ratio [OR], 1.00; 95% confidence interval [CI], 0.30–3.29).
Neonatal Outcomes
Neonatal outcomes are shown in Table S2. Three newborns died of the following causes: intraventricular haemorrhage probably due to maternal anticoagulation with the mechanical valve during pregnancy (n=1), after surgery to correct neonatal CHD (n=1); immediately after birth due to severe neonatal CHD (n=1). The general anaesthesia group had a significantly higher percentage of neonatal intubations (53.2% vs. 7.0%, P <0.001), Apgar scores <7 at 1 or 5 minutes (57.4% vs. 2.8%, P <0.001) and composite neonatal outcome (63.4% vs. 8.9%, P<0.001). Mixed-effects logistic regression models with a random intercept for each mWHO class for maternal cardiovascular risk revealed that general anaesthesia is significantly associated with a higher rate of composite neonatal outcomes (OR, 13.3; 95% CI, 5.52–32.0) (Table 4).
Sensitivity analysis
In the sensitivity analysis with date of delivery as a confounding factor, maternal outcomes remained not significantly associated with general anaesthesia (OR, 0.76; 95% CI, 0.19–3.00). Results were also unchanged in the sensitivity analysis where emergency caesarean section status was changed to gestational weeks at delivery as a potential confounding factor in the model (OR, 0.51; 95% CI, 0.24–1.11). Regarding neonatal outcomes, in the sensitivity analysis with date of delivery as a confounding factor, composite neonatal outcomes remained associated with general anaesthesia (OR, 8.84; 95% CI, 3.54–22.1). Results were also unchanged in the sensitivity analysis where emergency caesarean section status was changed to type of indication for caesarean section as a potential confounding factor in the model (OR, 12.8; 95% CI, 5.49–30.0). In the propensity score analysis using IPTW, composite postpartum cardiovascular events remained not significantly associated with general anaesthesia (OR, 0.51; 95% CI, 0.12–2.20) and composite neonatal outcomes remained significantly associated with general anaesthesia (OR, 4.56; 95% CI, 1.48–14.0). The c-index for the logistic regression model that was used to calculate propensity scores was 0.90.