Cerebral Saturation and Fractional Tissue Oxygen Extraction Are Associated with Anterior Cerebral Artery Doppler Parameters in Neonates with Congenital Heart Defects Open Access

After birth, neonates experience a drop in pulmonary vascular resistance (PVR) and a rise in systemic vascular resistance (SVR) due to placental disconnection [1]. In neonates with congenital heart defects (CHDs), this transition may disrupt oxygen transport and end-organ perfusion due to shunts and altered systemic/pulmonary blood flow, potentially reducing cerebral blood flow (CBF) and increasing the risk of brain injury and long-term neurodevelopmental issues [1‒4].

Bedside tools like near-infrared spectroscopy (NIRS) and ultrasonography (US) help assess cerebrovascular status in neonates at risk [5‒8]. NIRS, placed on the forehead, non-invasively measures venous-weighted cerebral oxygen saturation (CSat) at the frontal cortex and is used in neonatal cardiac surgeries to estimate cerebral oxygen consumption and perfusion [9‒11]. CSat is influenced by tissue perfusion, systemic oxygenation, local oxygen extraction, metabolism, and hemoglobin levels, which are affected by cardiac function and vascular distribution [4]. Given CHD-related variations in arterial oxygen content, cerebral fractional tissue oxygen extraction (cFTOE), calculated as ([preductal oxygen saturation {SpO2} – CSat]/preductal SpO2), provides insight into oxygen utilization. However, data on CSat and cFTOE as cerebrovascular metrics in CHD neonates during postnatal adaptation are limited.

Cerebral US with Doppler assesses blood flow velocities, including peak systolic (PSV) and end-diastolic (EDV) velocities, influenced by vascular resistance, blood viscosity, and cardiac output. Parameters like resistive index (RI) and pulsatility index (PI) of the anterior cerebral artery (ACA) evaluate vascular resistance and flow characteristics in the frontal cortex [8]. These markers have been shown to correlate with cerebral perfusion pressure and are recognized indicators of increased intracranial pressure in newborns. In neonatal hypoxic-ischemic encephalopathy, elevated RI may also signal brain injury, particularly during the hyperemic reperfusion phase. Finally, diastolic steal has been associated with increased RI and PI [2]. This study aimed to examine correlations between CSat and cFTOE (via NIRS) and RI and PI of the ACA (via Doppler US) in neonates with different CHD types during their first week of life. We hypothesize that CHD variations influence CBF, enabling investigation into consistent relationships between CBF and these bedside parameters, potentially offering stronger evidence that these modalities fluctuate in a consistent when vascular resistances changes.

This is a prospective observational cohort study in neonates with CHD born at 35 weeks gestational age or more and admitted to a single-center neonatal intensive care unit within their first 7 days of post-natal life from November 2019 to December 2021. Neonates either had a prenatal or a postnatal diagnosis of CHD. Neonates were recruited as early as possible after admission. We excluded those with a significant syndrome with more than one organ anomalies, including cerebral, pulmonary, airway, or intra-abdominal malformations. Demographic and clinical information was extracted from the medical chart. When available, blood gas and hemoglobin levels were collected on the same day as the NIRS and US measurements. In instances where multiple blood samples were collected within a single day, the sample temporally closest to the measurements was selected. We included values from any type of gas sample available, whether it was obtained through arterial or capillary sampling (ABL 800/835 Flex Analyzers). If a complete blood count was performed on the same day as the NIRS and US measurements, we prioritized the hemoglobin value from the complete blood count. If not, we used the hemoglobin level from our gas analysis. Written informed consent was obtained from the parent/legal guardians. This study was approved by the Pediatric Research Ethics Committee of McGill University Health Centre Research Institute, Approval No. 2019-3823.

CSat was recorded every 5 s using the INVOSTM 5100C cerebral oximeter (Medtronic), with sensors placed on the forehead for continuous monitoring until day 7, discharge, or cardiac intervention. An average of the CSat values at the exact minute of Doppler US acquisition of the ACA was extracted, along with pre-ductal saturation (Masimo SET Pulse oximetry) for cerebral FTOE calculation.

Daily transfontanellar cranial US with Doppler ACA sampling was performed using the Philips Epiq7c system. Imaging started at enrollment and continued until day 7, discharge, or cardiac intervention. Pulse-wave Doppler was obtained from coronal and sagittal planes with an insonation angle <20°. Doppler data included PSV, EDV, VTI, and MV, with RI ([PSV-EDV]/PSV) and PI ([PSV-EDV]/MV) calculated (online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000546675). Data were anonymized and masked to CHD and NIRS values.

We used descriptive statistics to summarize the data. Mixed-effects models assessed associations between CSat or cerebral FTOE and RI/PI-ACA, adjusting for time since birth, pCO₂, and hemoglobin levels. Analyses were performed using RStudio (v2023.12.1-402).

Among 135 neonates with CHD admitted to our neonatal intensive care unit during the study period, 34 neonates were enrolled in our study (Fig. 1). The final dataset includes 142 concomitant data measurements of NIRS and US Doppler. Among the enrolled neonates, 62% (n = 21) were male, the mean gestational age at birth was 38.5 ± 1.4 weeks, and the mean birth weight of 3.2 ± 0.6 kg. The types of CHD included are presented in Figure 1. Most (79%, n = 27) neonates were inborn and had a prenatal diagnosis. Most (74%, n = 25) were initiated on prostaglandins. Among the 34 patients who underwent an intervention during their first hospitalization, 70% (n = 24) required a surgical procedure or intervention at a mean of 5 (SD 1) days of life. Five neonates died from: septic shock (n = 1), bradycardia (n = 1), cardiac arrest at home (n = 1), and multiorgan failure (n = 2). Additional demographic and clinical information are shown in Table 1.

The postnatal trends of CSat, cFTOE, RI-ACA, and PI-ACA are outlined in Table 2. CSat on the first day was at a mean of 72 ± 10%, gradually declining to an average of 63 ± 9% by day 7. cFTOE increased during the same period, starting at an average of 23 ± 7 on day 1 and reaching 32 ± 10 on day 7. The RI-ACA went from an average of 0.77 ± 0.13 on day 1 to an average of 0.85 ± 0.13 on day 6. A drop can be noticed on the day 7 to an average of 0.81 ± 0.11, although this value represented data from only 9% (13 neonates out of 34 neonates enrolled still being monitored). PI-ACA followed a similar trend with an overall increase over time from 1.40 ± 0.31 to 1.73 ± 0.41 on day 6 and a drop on day 7 at 1.57 ± 0.36. The PSV of the ACA progressively increased over time, while the EDV remained stable. CSat decreased over time, while the cerebral FTOE increased, mirroring the trends observed in RI and PI of the ACA (see Fig. 2). The majority of blood gas analyses were performed using capillary sample (n = 66, 63%). The pCO₂ level decreased from 45 ± 9 mm Hg on day 1 to 40 ± 7 mm Hg on day 5, subsequently rising to 46 ± 5 mm Hg by day 7. Concurrently, hemoglobin experienced a slight decline from 169 ± 28 g/L on day 1 to 164 ± 22 g/L by day 7.

The mixed-effect models demonstrated a significant association between CSat/cFTOE and the corresponding RI and PI, while accounting for the timing of evaluation (Table 3). The analysis indicated that, on average, a 2.3% decrease in CSat (95% confidence interval [CI]: −4.4; −0.3%) corresponded to a 0.1-point increase in RI-ACA. Further, a 3.0-point elevation in cerebral FTOE (95% CI: 0.9; 5.2) corresponded to a 0.1-point increase in RI-ACA. Similarly, a 0.9% reduction in CSat (95% CI: −1.5; −0.3%) corresponded to a 0.1-point increase in PI-ACA. As well, a 1.1-point rise in cerebral FTOE (95% CI: 0.4; 1.7) corresponded to a 0.1-point increase in PI-ACA. Adjusted random mixed-effects models were constructed to account for confounders, including pCO₂ and hemoglobin levels at the time of the ultrasound (online suppl. Table 1). These adjustments did not alter the strength of the association. No association was observed between lactate levels and cerebral NIRS markers in this cohort. Figure 2 presents the NIRS and Doppler parameters over time in hours since birth, while online supplementary Figure 2 illustrates these parameters by day of life. These figures demonstrate that, as time progresses, CSat decreases, cFTOE increases, and both RI and PI increase. Online supplementary Figure 3 depicts the unadjusted linear regression between CSat/cFTOE and RI/PI, showing that higher CSat is associated with lower RI and PI, while lower cFTOE corresponds to lower RI and PI. These unadjusted analyses highlight a consistent directionality of correlation between NIRS and Doppler US flow markers. However, they do not account for the inherent within-participant correlation over time.

Our study analyzed 142 data points from 34 neonates with various CHD subtypes, revealing a significant association between CSat, cFTOE, and ACA RI/PI during the first week of life. Higher RI/PI values indicate increased resistance to flow velocities. A decrease in CSat with increased cFTOE suggests greater oxygen extraction, likely due to reduced blood flow, impaired oxygen delivery, or heightened metabolic demand. This association remained significant after adjusting for pCO₂ and hemoglobin, reinforcing CSat as a marker of cerebral perfusion and FTOE as an indicator of oxygen consumption. These findings align with physiological expectations and may aid in monitoring cerebral perfusion in CHD neonates during intensive care. Further research is needed to assess if interventions targeting low CSat or high FTOE can reduce brain injury risk.

NIRS has emerged as a potential tool for monitoring cerebral venous-weighted oxygen saturation as an indirect reflection of CBF [3, 4]. Our study observed a decline in CSat and a rise in cerebral FTOE throughout the first week in neonates with CHD. These findings align with prior research by Lynch et al. [3], who also reported a significant negative trend of CSat and a positive trend of cFTOE over time in CHD. Although the precise cause of the reduction in CSat remains unclear, it may partly be attributed to an altered adaptation in the context of shunts and the underlying abnormal cardiac structures, as well as an increase in brain metabolic demand over time [1]. Transitional physiology can significantly affect vascular resistance and blood flow, particularly with shunts or cardiac anomalies. Despite this variability, our findings show that NIRS and ultrasound maintained consistent directional trends, highlighting their complementary value in monitoring cerebral hemodynamics. In a recent cohort of neonates with CHD, we observed increased ACA RI and PI associated with retrograde holodiastolic flow in the descending aorta, likely due to ductal steal [5]. This suggests that altered flow dynamics during transition can impact cerebral perfusion, yet NIRS and ultrasound remain reliable monitoring tools. Although we did not measure parameters of vascular autoregulation, our findings may be further explained by inherent alterations in cerebral autoregulation and exposure to systemic agents affecting vascular tone (e.g., prostaglandins). In the current analysis, pre-ductal oxygen saturation and hemoglobin levels remained stable throughout the first week of life. As such, increasing trend in cerebral FTOE may indicate increased tissue oxygen extraction, possibly secondary to increased metabolic demand and/or decreased perfusion by steal effect or poor systemic cardiac output.

CBF velocities adapt to changes in vascular resistance, blood viscosity, and systemic ventricular contraction. Diastolic steal or rising vascular resistance can alter velocities, triggering feedback to maintain CBF. Indices like RI and PI describe these changes, with Camfferman et al. [6] noting RI’s minimal inter-observer variability and angle independence. In preterm infants, higher ACA RI correlates with lower CSat, similar to our findings [7]. Zhang et al. [8] also reported an inverse relationship between CSat and RI/PI in neonates undergoing VSD repair. Our data link CSat with ACA Doppler flow, reflecting frontal cortex CBF, with results remaining consistent after adjusting for pCO₂ and Hgb despite varied cardiac conditions. These insights underscore the potential of CSat/cFTOE as bedside, noninvasive, and continuous monitoring tools essential for understanding hemodynamics, facilitating early detection, prompt management, and guiding surgical decisions to improve neurological outcomes in this vulnerable population [9, 10].

In our study, six neonates had minor infarcts outside the ACA territory, and eight had minor subdural hematomas on magnetic resonance imaging. While the timing of these lesions relative to Doppler NIRS measurements is unclear, we observed an association between ACA Doppler (deep vascular flow) and NIRS (superficial flow) despite these findings. Clinicians should interpret cerebral perfusion markers cautiously, especially without advanced imaging. As our study focused only on the pre-intervention period, further research during and after surgical interventions is needed to explore dynamic changes in cerebral oxygenation and Doppler markers.

Our prospective study provided a rich dataset across various CHD subtypes, with blinded data extraction for NIRS-US Doppler evaluations and standardized imaging protocols. While we did not assess inter-rater variability in this study, our group previously reported good to excellent intraclass correlation coefficients (>0.75) for ACA Doppler parameters using the same methodology [5]. Several limitations may affect generalizability. As a single-center study with a small sample, extrapolation to broader CHD populations is limited. The study design did not allow for evaluating associations between specific NIRS or Doppler thresholds and clinical outcomes. Future larger studies with defined intervention protocols are thus needed. Although bedside staff were not blinded to NIRS values, this is unlikely to have impacted the relationship between NIRS and ACA Doppler. Data for hemoglobin and pCO₂ were incomplete (available for 101 and 104 of 142 encounters, respectively), potentially limiting analysis robustness. Since only 37% of pCO₂ samples were arterial, standardizing sampling methods in future studies is recommended. While heart rate and blood pressure may influence NIRS Doppler relationships, we prioritized pCO₂ and hemoglobin as the most biologically relevant markers to minimize confounding. Inconsistencies in ultrasound acquisitions over the 7-day period, due to admission, discharge, or procedures, also posed a challenge. Additionally, the small sample size prevented assessment of findings based on CHD severity or brain injury risk. To address some variability, we applied random mixed-effects models to account for individual trends. Finally, both NIRS and Doppler offer indirect CBF assessments; direct measures would require magnetic resonance imaging or catheterization, which were not conducted in this study [11].

Our study revealed an association between CSat/cFTOE measurements and RI-ACA/PI-ACA indices, obtained simultaneously. Doppler signs indicating increased CBF resistance in the ACA correlated with decreased CSat and increased cerebral FTOE captured on the forehead using NIRS. Future research should assess whether a multimodal bedside approach to monitoring cerebrovascular hemodynamics can facilitate early detection of cerebral hypoperfusion and prevent brain injury, as well as adverse neurodevelopmental outcomes in this vulnerable population.

We would like to acknowledge the contributions of all the families who agreed to participate in this study.

The Institutional Review Board of MUHC-RI (Montreal University Health Center Research Institute) approved this study, Approval No. REB#2019-3823. Written informed consent was obtained from the parent/legal guardian of participants prior to the study.

Medtronic provided NIRS monitors and sensors for the study through a peer-reviewed competition grant. Medtronic was not involved in elaborating the protocol, acquiring the data, or interpreting the results.

Medtronic provided NIRS monitors and sensors for a study. Just-for-Kids Foundation provided the funds to obtain the echocardiography machine and the TomTEC software for analysis. The Department of Pediatrics of McGill University provided funds to support this research. Division of Neonatology of the Montreal Children’s Hospital provided funds to support this research.

P. Kanaprach conceptualized and designed the study, collected the data, extracted the data from the hospital records, analyzed the data, drafted the manuscript, and adjusted the manuscript according to the comments of the co-authors. C. Michel-Macias, M. Mazzarello, E. Rampakakis, S.S. Moore, P. Wutthigate, J. Simoneau, and D. Villegas collected the data, critically appraised the analysis, and revised the manuscript. P. Wintermark, S.D. Shemie, M. Brossard-Racine, G. Bertolizio, and A. Dancea critically appraised the data analysis and reviewed and revised the manuscript. G. Altit conceptualized and designed the study, supervised data collection, critically appraised the data analysis, and wrote and critically reviewed the manuscript for important intellectual content. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

The data that support the findings of this study are not publicly available due to confidentiality but are available from the corresponding author upon reasonable request.