There are multiple potential contributors to brain injury in infants requiring neonatal surgery, beyond the direct impact of congenital anomalies, including: hypoxia and ischemia related to the underlying condition; vulnerability due to comorbid genetic anomalies; pain, sedation and analgesia; surgery and anesthesia; prolonged hospitalization; the altered family environment and impact on parental mental health, as well as parent-infant attachment (Govindaswamy et al., 2020; Diffin et al., 2016).

The literature regarding the impact of anesthesia on the neonatal brain is rapidly expanding. Whilst retrospective studies of infants have raised some concerns about anesthetic toxicity in the developing brain, prospective research has been somewhat reassuring. The GAS trial taught us that 40 min of general anesthesia for the repair of inguinal hernias in boys did not impact on measured 2- or 5-year neurodevelopmental outcomes, compared to those who received a spinal anesthetic (Davidson et al., 2016; McCann et al., 2019). A Canadian population-based study also showed no increased risk of adverse child development on school entry for children who had undergone surgery compared with sibling-matched pairs (O'Leary et al., 2019). However, that study intentionally excluded children with neurodevelopmental impairment. Of course, anesthesia and surgery occur for a few hours amongst what is often many weeks spent in the neonatal intensive care unit (NICU) for babies with congenital anomalies.

While clinical studies to date have largely focused on the impact of anesthesia on healthy infants, much pre-clinical research should lead us to interpret these findings with caution (Ben-Ari, 2002). Many anesthetic and sedative agents are GABA antagonists and have been shown to be neurotoxic in neonates (Andropoulos, 2018; Gascoigne et al., 2021). It also follows that the variability of development of GABA synapses may also further be altered by other neuronal disruptions such as prematurity. GABA is excitatory in the newborn brain, becoming inhibitory as chloride co-transporters mature, most importantly KCC2. Extensive animal research has identified the potentially toxic effects of a mismatch between GABA- and glutamate-mediated inhibition and excitation, which may occur with altered maturation of chloride co-transporters. Ben-Ari proposed that ‘the shift of GABA actions is a developmentally regulated function that signals the shift from a genetically determined programme to one that takes neuronal and environmental factors into account’.

Minimal human research exits regarding the impact of anesthesia on neonates who have already experienced some degree of brain compromise and in those requiring multiple operations over the first weeks of life. The extended peri-operative period and intensive care stay clearly exposes infants to ongoing risks beyond the surgery and anesthesia itself. For example, exposure to opiates both in utero and during preterm care has impacts on brain growth and later neurodevelopmental profile (Ranger et al., 2015; Zwicker et al., 2016; Steinbauer et al., 2021; Walhovd et al., 2007). Moreover, exposure to pain itself during the preterm course has an impact on thalamic development and longer-term neurodevelopment (Duerden et al., 2018; Schneider et al., 2018).

Functional deficits may also emerge through limits to prone positioning in the NICU following thoracic or abdominal surgery. The development of feeding disorders is also frequently experienced by neonates due to the balance of adverse oral experiences such as regular suctioning or severe gastro-esophageal reflux disease (Ferguson, 2017). The physical impact of impairment on normal nutritional intake for infants who are dependent on parenteral nutrition, or those with excessive caloric requirements due to underlying disease, also has the potential to impact both physical and brain development.

The risks to neonates undergoing surgery for congenital anomalies are demonstrated through imaging studies showing increased rates of brain injury and deviation from normal growth and development patterns. Moran et al. reported the post-operative MRI findings of 39 neonates with congenital diaphragmatic hernia, esophageal atresia or abdominal wall defects, who were subsequently assessed at 2 years of age (Moran et al., 2019). Compared with gender and gestational-age matched controls, infants requiring surgery had smaller brains on quantitative measures, were 10 times more likely to have white matter signal abnormalities and 6 times more likely to have delayed gyral maturation. The group who underwent surgery for congenital anomalies also had lower mean scores than controls in all three domains of the Bayley Scales at 2 years.

Stolwijk et al. performed MRI brain scans post-operatively in 101 infants, of whom 32 were born prematurely (Stolwijk et al., 2017). MRI abnormalities were found in 75% pre-term and 58% full-term infants. In the pre-term group, most abnormalities were parenchymal and included white matter lesions, punctate cerebellar lesions, thalamic infarction and periventricular hemorrhagic infarction. Prematurity and the type of congenital anomaly were most predictive of brain injury.

Congenital anomalies have long been known to be an independent risk factor for neonatal morbidity in infants born prematurely (Linhart et al., 2000). In an outcome study of 5- and 8-year-olds with abdominal wall defects, infants born before 37 weeks gestation performed worse on every test compared with term-born infants at both 5 and 8 years (Burnett et al., 2018). In infants with congenital diaphragmatic hernia, lower gestational age was an independent risk factor for impaired language and motor development at 2 years but had an ongoing impact on full-scale intelligence quotient (FSIQ) and processing speed even at 8 years (Gunn-Charlton et al., 2019). For infants born with esophageal atresia, the acute impact of comorbid prematurity is substantial, including pre-operative respiratory compromise, delayed repair and gastrostomy formation, increased risk of surgical complications, oral aversion and feeding difficulties, length of stay as well as potential comorbid lesions including cardiac anomalies (Hunt et al., 2016). In 32.2-year-olds with esophageal atresia, infants with a Bayley score more than 1 standard deviation below the test normative mean at 2 years were, on average, 3 weeks more premature, 1 kg lighter at birth and had more than twice the hospital length of stay, compared with infants with a 2-year Bayley score in the normal range (Gunn et al., 2016b). However, by school-age the impact of prematurity was overcome by other contributing illness-related and environmental factors (Burnett et al., 2021).

If surgery itself (as well as the condition leading to surgery) is a risk to term-born neonates, then the additive impact of prematurity is potentially cumulative. Hunt et al. reported the 8-year-old outcomes of three cohorts of extremely preterm infants who underwent a surgical procedure during their NICU admission (Hunt et al., 2018). They found that rates of neurosensory disability were more than 4-fold higher in children who had undergone surgery (33%) compared with those who didn’t (10%). Surgery in those infants would most likely have been for conditions such as necrotizing enterocolitis or a patent ductus arteriosus ligation rather than for management of a congenital anomaly per se.

Walsh et al. specifically excluded congenital anomalies from a group of preterm infants less than 30 weeks gestation included in a study comparing MRI at term-equivalent age and 2-year outcome between those who did and did not receive surgery and anesthesia (Walsh et al., 2021). A relationship was found between exposure to surgery (with a dose effect of time) and relative reductions in white matter and deep gray matter volume. At 2 years, surgical exposure was associated with lower cognitive and motor scores and there remained a dose effect of surgical time, even when controlling for the MRI findings, suggesting that more complex microstructural factors are at play beyond brain growth. Gano et al. also described adverse neurodevelopmental outcomes in preterm infants born before 33 weeks who underwent two or more surgeries during the preterm period, with 20-point lower FSIQ at school age compared with children had not undergone surgery (Gano et al., 2015).

Most major neonatal surgery is time-critical such that pre-operative MRI is very challenging to obtain. Those researchers who have managed to obtain such images are subject to extreme bias as the more complex or unwell infants are least likely to have been able to access the MRI scanner prior to surgery. Fetal imaging would be the obvious solution to exploring brain growth prior to non-cardiac surgery but has not yet been widely explored to date.

Many studies using both ultrasound and MRI have now shown reductions in brain growth during the second and third trimester in fetuses with CHD, the variance from normal fetal brains of which becomes more pronounced through pregnancy (Limperopoulos et al., 2010; Ortinau et al., 2019). Early neonatal MRI confirms that brain abnormalities include delayed myelination patterns, impaired white matter microstructural development, abnormal brain metabolism, abnormal cortical folding and reductions in brain volumes (Beca et al., 2013; Ortinau et al., 2013; Ortinau et al., 2012; Miller et al., 2007; Licht et al., 2009). These findings are not specific to the cardiac diagnosis but are seen in infants with right and left heart obstruction, suggesting the impact is related more closely to chronic cerebral hypoxia rather than altered cerebral blood flow per se. This potential link has been confirmed by MRI studies linking cerebral oxygenation with impaired growth in fetuses with CHD (Sun et al., 2015). Gestational age and chronicity of hypoxia affect the fetal response, which most commonly includes cerebral vasodilation (Ortinau and Shimony, 2020).

It is likely that hypoxic injury in the third trimester has a similar effect on brain development to that seen in preterm infant brains. It is therefore not surprising that the pattern of injury seen in term newborns with CHD is remarkably similar that of term-corrected preterm brains. In particular, MRI studies have shown that periventricular punctuate white matter injury is very common in both populations, although the more extensive patterns of periventricular leukomalacia seen historically, and still occasionally contemporaneously, in preterm infants, are seen less commonly in CHD (Beca et al., 2013; Ortinau and Neil, 2015; Guo et al., 2019). Intriguingly, the long-term motor development trajectory varies somewhat between the two groups, with lower rates of cerebral palsy in CHD, except where later cerebrovascular accidents have occurred. However, the pattern of attentional problems into school age is very similar between the groups (Claessens et al., 2018; O'Shea et al., 2013). The differences are most likely impacted by chronicity of illness and subsequent brain impacts, especially low cardiac output syndrome and stroke.

Given the impaired fetal pattern of brain development in CHD, it follows that the brains of newborns born with CHD are more immature than their healthy counterparts (Beca et al., 2013; Licht et al., 2009). Furthermore, the brains of neonates showing evidence of immaturity relative to their gestational age on pre-operative scans, have higher rates of injury detected on post-operative MRI scans and, perhaps more importantly, are more likely to exhibit subsequent developmental delay (Beca et al., 2013; Andropoulos et al., 2014).

As with non-cardiac congenital anomalies, neonates with CHD are more likely than those without CHD to be born preterm and/or SGA and preterm infants are more likely to be born with CHD than those who reach full-term. Lower gestational age is a strong predictor of brain injury (encompassing a range of patterns of anomalies on neuroimaging) and neurodevelopmental impairment in CHD, even in early term-born infants. A study of 971 consecutive neonates with critical CHD was stratified by gestational age and adjusted for severity of lesion, against a reference group of neonates born at 39–40 weeks gestation. The risk of cerebral injury increased 2-fold in the 37–39 week gestation group; 8-fold in the 34–37 week gestation group and 25-fold in infants less than 34 weeks gestation at birth (Costello et al., 2010). In the Hearts and Minds Study of babies over 36 weeks at birth with critical CHD, on multivariable analysis of 2-year outcomes, the lowest developmental quotients on the Bayley Scales related to intrinsic patient factors including reduced gestational age at birth, single ventricle physiology with pulmonary obstruction and the need for repeat surgery (Gunn et al., 2016a). Ortinau et al. demonstrated that CHD does not increase the risk of intraventricular hemorrhage over that of gestation-associated risk (Ortinau et al., 2018), but Bell et al. showed that preterm neonates cared for in a cardiac intensive care unit are 3.2 times more likely to display acute neurological events (including cerebral hemorrhage or white matter injury) compared with older patients (Bell et al., 2019).

There has been much debate regarding optimal timing of surgery for infants born prematurely with critical CHD (Curzon et al., 2008; Hickey et al., 2012; Chu et al., 2017). Intentionally delayed surgery, beyond the best-practice timing for a given cardiac lesion, allows for organ growth and maturity, potential nutritional benefits, fall in pulmonary vascular resistance, improvement in immune competence and safe operative scheduling. Conversely, the risks of delaying surgery include the secondary effects of unrepaired CHD including hypoxemia, increased pulmonary blood flow, systemic pulmonary artery pressures, ductal steal, intensive care morbidity and the overall cost of care. Studies have shown that elective delayed surgery in preterm infants increases the risk of mortality, white matter injury on MRI and secondary morbidities such as infection, apnea, and necrotizing enterocolitis, whereas intentional delays reduce the risk of intracerebral hemorrhage, renal dysfunction and coagulopathy. These difficult decisions regarding competing clinical interests will continue to be made at an institutional level, weighing up risk and benefit for individual patients. However, it is clear than in most centers, a collaborative approach in care between cardiac intensive care, cardiology and neonatology is likely to lead to the best outcomes for preterm infants with CHD (Krishnamurthy et al., 2013).