Discussion

Our multicentre survey provides a comprehensive look at the perioperative neurological care landscape for neonates undergoing cardiac surgery. Despite well-documented evidence that this population experiences substantial rates of perioperative brain injury, we identified significant variability in neuromonitoring and neuroimaging practices across respondent Pediatric Cardiac Intensive Care Society centres and a discordance between perceived utility and clinical adoption. ^{Reference Peyvandi, Xu and Barkovich3,Reference Stegeman, Feldmann and Claessens5,Reference Andropoulos, Hunter and Nelson6,Reference Petit, Rome and Wernovsky17} Several key practical themes emerged: (1) a preference for preoperative over postoperative imaging, with heavy reliance on head ultrasound; (2) neuromonitoring was more consistently implemented postoperatively, with near-infrared spectroscopy as the predominant modality; (3) a lack of correlation between specialised neurocritical care teams and increased diagnostic utilisation; and (4) a perceptual mismatch regarding the necessity and sufficiency of these tools among physicians, advanced practice providers, and nursing staff. Collectively, these findings highlight an implementation gap and the urgent need for multidisciplinary guidelines to standardise neurological care in this vulnerable population.

We found a clear preference for routine neuroimaging before rather than after cardiopulmonary bypass (87% vs. 32%). Brain MRI was protocolised more commonly preoperatively than postoperatively, but HUS remained the predominant modality in both periods, despite its well-documented limited sensitivity for detecting white matter injury and stroke. ^{Reference Rios, Welty and Gunn18} Our findings mirror previous survey studies within the United States and Europe, which report wide variability in preoperative MRI use, performed in 17–72% of centres, and near universal use of screening HUS. ^{Reference Feldmann, Hagmann and de Vries19–Reference Belfi, Vega and Aguar21} Similarly, we found that postoperative MRI was largely reserved for symptomatic neonates or those requiring mechanical circulatory support. ^{Reference Feldmann, Hagmann and de Vries19,Reference Belfi, Vega and Aguar21} This predominantly reactive rather than proactive approach to postoperative imaging is concerning, as brain MRI detects new postoperative injuries in up to 50% of neonates with CHD, many of which are clinically silent. ^{Reference Stegeman, Feldmann and Claessens5,Reference Andropoulos, Hunter and Nelson6,Reference Alablani, Chan and Beishon22} Reliance on head ultrasound or symptom-based imaging triggers, therefore, likely underestimates the true burden of postoperative brain injury. Further, perioperative brain injury has been linked with slowed postoperative brain growth and substantial neurodevelopmental impairment. ^{Reference Marino, Lipkin and Newburger23–Reference Dimitropoulos, McQuillen and Sethi26} Early identification of injury may therefore inform neuroprotective strategies and facilitate early neurodevelopmental referral and family counselling.

Perioperative neuromonitoring was widely utilised among centres in our survey, with near-universal use of near-infrared spectroscopy, consistent with prior studies of paediatric cardiac centres. ^{Reference Feldmann, Hagmann and de Vries19–Reference Belfi, Vega and Aguar21,Reference Hoskote, Tume and Trieschmann27} Near-infrared spectroscopy is widely regarded as a noninvasive, readily available neuromonitoring tool that provides continuous real-time feedback and may facilitate early detection of acute neurologic events. ^{Reference Nagdyman, Fleck and Schubert28,Reference Loomba, Rausa and Sheikholeslami29} While near-infrared spectroscopy serves as a valuable trend-monitoring tool, evidence remains limited regarding its correlation with neuroimaging abnormalities or long-term neurodevelopmental outcomes. ^{Reference Hoskote, Tume and Trieschmann27,Reference Pardo, Carrasco and Wintermark30} Our findings demonstrate a substantial increase in electroencephalography utilisation compared with historical cohorts, with 50% of centres now reporting postoperative electroencephalography use, compared with 17–20% in prior studies. ^{Reference Feldmann, Hagmann and de Vries19,Reference Leon, Levy and Hu20} This shift likely reflects a growing recognition of the high incidence and clinical impact of postoperative seizures and an evolving alignment with recommendations published by the American Clinical Neurophysiology Society. ^{Reference Andropoulos, Mizrahi and Hrachovy9–Reference Clancy, Sharif and Ichord11,Reference Feng, Lin and Zhang31,Reference Wusthoff, Numis and Pressler32} Importantly, our findings regarding clinician perceptions suggest that electroencephalography was highly valued across all provider roles, regardless of whether it was routinely incorporated into institutional practice. This suggests that electroencephalography is broadly perceived as a high-utility tool in the postoperative care of these patients, supporting its inclusion in future consensus guidelines and underscoring the need for formal cost-benefit analyses to inform its optimal implementation.

To our knowledge, this is the first study to quantify the presence of neurocritical care consultation services in paediatric heart centres (42%). Neurocritical care consultants are increasingly recognised as essential stakeholders in the comprehensive care of critically ill children. ^{Reference LaRovere, Murphy and Horak33} This is particularly important in the cardiac ICU, where patients are at high risk for neurological injury due to complex physiology and operative interventions. In our survey, the presence of a neurocritical care team did not increase neuroimaging or neuromonitoring utilisation, and while noteworthy, the trend toward higher electroencephalography use in centres with neurocritical care teams did not reach statistical significance. While our limited sample size may have obscured more significant findings, the data suggests that specialised expertise alone cannot drive clinical adoption without the support of formal institutional protocols. Nonetheless, as neurocritical care teams expand, they are well-positioned to play a central role in developing guidelines, implementing targeted interventions, and integrating neurocritical care into postoperative management.

A key finding was the divergence between the perceived utility of perioperative neurological practices and their actual clinical utilisation, alongside significant perceptual variation across clinical roles. While brain MRI was widely recognised as valuable, this perception was notably stronger among nurses than among physicians or advanced practice providers. This trend may reflect the nursing role’s focus on neurodevelopmental advocacy and the need for prognostic clarity during family counselling. However, this high perceived value did not consistently translate into clinical practice, particularly within postoperative neuroimaging protocols. This disconnect suggests that implementation barriers are not driven solely by clinician beliefs, but rather by a combination of structural, logistical, and evidence-based gaps. Specifically, clinicians may be hesitant to navigate the logistical hurdles of neonatal transport and the structural costs of MRI without definitive data proving these diagnostics directly alter immediate management or improve long-term outcomes.

Furthermore, a significant discrepancy emerged regarding the beliefs about the institutional adequacy of neurological practices. Physicians were more likely than advanced practice providers and nurses to report their centre practices as sufficient. These conflicting perceptions likely stem from differences in training, proximity to bedside care, and clinical workflow. Addressing these gaps will require interdisciplinary dialogue, targeted education, and the establishment of consensus guidelines to align multidisciplinary perceptions with evidence-based practice and standardised care.

While this study provides novel information about the landscape of neurological practices within the paediatric cardiac ICU, several limitations should be considered. Participation in the Pediatric Cardiac Intensive Care Society is voluntary, and the self-selected nature of the sample may have preferentially captured centres with a greater interest in neurological care. The modest individual response rate of 13.8%, the high response rate by physicians relative to other provider roles and the predominance of academic centres in the United States may limit generalisability to smaller or international programmes. In addition, reported institutional practices were not independently verified and may therefore be subject to misclassification bias. Practices were based on self-reports by medical directors and therefore may not reflect actual clinical utilisation of neuroimaging and neuromonitoring at each institution. Future studies should validate reported institutional practiced using objective data from sources such as chart review and registry-based analysis. An additional limitation was that the terms “useful” and “sufficient” were not standardised within the survey, which may have resulted in variable interpretation across respondents and influenced perception-based analyses.

Importantly, the perceptions captured in this study should be considered hypothesis-generating, particularly among advanced practice providers and nursing providers, who were under-represented relative to physicians. Future studies should aim to improve multidisciplinary engagement, incorporate standardised terminology, and evaluate the relationship between neuromonitoring and neuroimaging practices and patient-centred outcomes. Broader recruitment efforts targeting nursing and APP subgroups within the Pediatric Cardiac Intensive Care Society, as well as collaboration with external professional organisations representing non-physician providers, may further enhance participation and improve the representativeness of future studies.