Immediate Postoperative Evolution of Cerebral Doppler and Saturation Parameters After Cardiopulmonary Bypass in Infants With Congenital Heart Disease: A Prospective Cohort Study

Abstract

OBJECTIVES:

To describe temporal trends and instantaneous relationships between middle cerebral artery (MCA) cerebral blood flow velocity, cerebral oxygen saturation (Csat), cerebral fractional tissue oxygen extraction (cFTOE), and systemic arteriovenous oxygen difference (AVo₂) during the first 24 hours following congenital heart disease (CHD) surgery requiring cardiopulmonary bypass (CPB).

DESIGN:

Prospective observational study conducted from August 2021 to June 2022.

SETTING:

Single-center pediatric cardiac ICU in Canada.

PATIENTS:

Fifteen ventilated neonates and infants admitted to the PICU following CHD surgery requiring CPB.

INTERVENTIONS:

None.

MEASUREMENTS AND MAIN RESULTS:

Bilateral transcranial Doppler (TCD) assessments of the MCA were performed at seven predefined time points during the first 24 hours following CPB, with time 0 defined as the end of CPB. A total of 87 simultaneous measurements of systemic oxygen saturation, Csat, TCD-derived parameters, and AVo₂ were analyzed. The nadir in MCA velocities occurred at a mean of 9.3 hours (sd, 1.7 hr) post-CPB and coincided with the lowest Csat and the highest AVo₂ difference and cFTOE. Lower MCA mean velocities were associated with higher AVo₂ difference and increased cFTOE.

CONCLUSIONS:

In neonates and infants following CPB, MCA velocities reach its lowest point approximately 9 hours post-CPB, coinciding with the highest AVo₂ difference and previously published nadir of cardiac output. These findings could suggest a period of relative cerebral hypoperfusion. Future studies should evaluate whether targeted interventions guided by multimodal cerebral monitoring can mitigate the risk of acute and long-term brain injury in this vulnerable population.

KEYPOINTS.

Question: Are transcranial Doppler and near-infrared spectroscopy able to noninvasively detect low cardiac output syndrome in infants after congenital heart disease (CHD) repair?

Findings: In this prospective pilot study, infants with CHD after cardiopulmonary bypass (CPB), the nadir of cerebral saturations (CSat) and cerebral blood flow velocities (CBFv) occur 6–10 hours after CPB, similar than previously reported nadir of cardiac index. The nadir of CBFv and CSat occurred during the peak of arteriovenous oxygen difference.

Meaning: Monitoring CBFv and CSat after CPB can provide noninvasive information on cardiac output and promptly detect suboptimal cerebral perfusion.

Low cardiac output syndrome (LCOS) is a frequent complication observed in 25% of children after cardiac surgery (1), typically occurring 6–18 hours after the surgical intervention (2, 3). Early recognition and management of LCOS in post-cardiac surgery pediatric patients is essential to reduce mortality and morbidity (4). Invasive and continuous cardiac output (CO) monitoring is the gold standard to diagnose LCOS, however, is not routinely performed in children. Pediatric intensivists rely on clinical evaluation and indirect parameters of low CO such as heart rate, blood pressure, urine output, lactate, and arteriovenous oxygen difference (AVo₂) (5), which are insufficient to detect cerebral hypoperfusion. Technologies such as transcranial Doppler (TCD) and cerebral near-infrared spectroscopy (NIRS) (6) could help detect LCOS and brain hypoperfusion, and guide personalized brain-targeted care, potentially improving long-term neurologic outcomes.

TCD can measure cerebral blood flow velocities (CBFv) as a surrogate of cerebral blood flow (7, 8). TCD has been used during cardiopulmonary bypass (CPB) in pediatric patients (9–12). Due to impaired cerebrovascular autoregulation during CPB, arterial blood pressure and mean CBFv become correlated, suggesting that TCD can be used to adjust CPB pump flow to patient needs (13, 14). TCD may also aid in the detection of intraoperative cerebral emboli during CPB (15). Data on the evolution of CBFv in the postoperative cardiac setting in still lacking.

NIRS is employed not only as a cerebral oxygen monitor but also to assess adequacy of cerebral perfusion (12). To date, the observed correlations between NIRS, systemic hemodynamics, and pump flow have been limited (12). Cerebral saturation (CSat) trends measured by NIRS during cardiac surgery have been correlated with invasive central venous oxygen saturation using internal jugular vein blood gas oximetry (16). Low CSat in the immediate postoperative period has been associated with brain injury and neurodevelopmental impairment (17, 18).

Currently, there is limited data on CBF indices obtained via TCD and NIRS in the immediate postoperative period for infants undergoing cardiac surgery with CPB. We aimed to characterize the physiologic evolution and correlation of the middle cerebral artery (MCA) CBFv and bilateral CSat in this population.

Our long-term goal is to identify modifiable risk factors for poor neurologic outcome, investigate potential treatment approaches, and ultimately improve outcomes in neonates with congenital heart disease (CHD) undergoing cardiac surgery with CPB.

MATERIALS AND METHODS

Study Population

This was a prospective, single-center observational cohort study recruiting infants admitted to the PICU after congenital heart surgery between August 2021 and June 2022. The study population included infants 37 weeks old or older, corrected gestational age, and 12 months young or younger, sedated, ventilated, and admitted following cardiac surgery requiring CPB. Exclusion criteria were a history of intracranial bleed or stroke, or an expected death within 24 hours of admission. We included two participants with univentricular circulation as we aimed to evaluate trends over time, independent of the present or absence of intracardiac shunts. A participant (case 1) unexpectedly died a few hours after surgery. Although expected deaths within 24 hours were excluded, this participant’s death was not anticipated, so they were included in the study. “Time 0” refers to the moment when the participant was disconnected from the CPB intraoperatively. Patient demographic and clinical information were extracted from medical charts. Central venous saturation was sampled in the superior vena cava in 13 of 15 participants and in the right atrium in two of 15 participants that had no postoperative intracardiac shunt. Concurrent NIRS, TCD evaluations, and blood gas values were collected. Data collection was finished when patient reached 24 hours after the end of bypass or when patient was extubated, whichever came first. Critically ill extubated and nonsedated pediatric patients might exhibit different CBFv values compared with when they are intubated and sedated. This study was reviewed and approved by the McGill University Health Centre Research Ethics Board Pediatrics Panel on August 10, 2021 (MP-2022-7831, Use of near-infrared spectroscopy and transcranial Doppler in pediatric patients with low cardiac output syndrome after congenital heart surgery—a prospective cohort pilot study). Informed consent was obtained from their legal guardians before data collection. Procedures followed were in accordance with the Helsinki Declaration of 1975.

MCA Doppler

A single experienced operator (M.M.) conducted serial TCD of bilateral MCAs using a 1–4 MHz sector array probe with a Philips Sparq ultrasound system (Bothell, WA). The protocol followed adhered to expert consensus recommendations (7). At each time point, three pulse wave Doppler measurements of both MCAs were performed and repeated approximately every 4 hours from the time of admission up to 24 hours after CPB. An additional TCD measurement was performed if a clinical trigger was present, such as acute cerebral desaturation greater than 10%, acute mean arterial pressure drop of greater than 10 mm Hg, sustained tachycardia greater than 30 beats/min above baseline, etc (Supplemental Table 1, https://links.lww.com/CCX/B598). Data extracted from TCD included peak systolic velocity (PSV), end-diastolic velocity (EDV), and mean velocity (MV; Supplemental Fig. 1, https://links.lww.com/CCX/B598). The resistive index (RI) of the MCAs was calculated using the formula (PSV–EDV)/PSV. In the absence of diastolic flow, diastolic velocity was considered zero. The pulsatility index (PI) of the MCA was calculated using the formula (PSV–EDV)/MV (19). The sonographer was unblinded to blood gas results and CSat. Concurrent arterial and venous gas sampling were conducted, and the AVo₂ was calculated.

Cerebral and Peripheral Saturations

CSat for all participants were measured by NIRS using the INVOS 5100C cerebral/somatic oximeter (Medtronic, Minneapolis, MN) equipped with bilateral sensors. Right and left frontal CSat were collected at arrival in the PICU and concomitantly at each blood sampling and TCD. Right upper limb oxygen saturation (Spo₂), measured with a (Masimo Corporation, Irvine, CA) SET Pulse Oximeter, was recorded to calculate the corresponding cerebral fractional tissue oxygen extraction (cFTOE) with the formula: (Spo₂–CSat)/Spo₂ × 100 (20).

Statistics

We used descriptive statistics, including means (sds) and medians (interquartile ranges [IQRs]) for continuous variables, and counts (proportions) for categorical variables. We built random mixed-effects models with random slopes and intercepts to evaluate CSat, cFTOE, AVo₂, VM-MCA, RI-MCA, and PI-MCA. These models included the time since the end of CPB, measured in hours, as an adjustment variable. Temporal trends of CSat, cFTOE, RI, and PI were illustrated through graphs depicting the relationships between NIRS and TCD derived parameters. We conducted the statistical analyses using RStudio (Version 2023.09.1 + 494; Posit, PBC, Boston, MA).

RESULTS

Fifteen patients were included in the study of 37 infants admitted in our PICU after CPB between August 2021 and June 2022 (20 excluded due to absence of ultrasonographer and two declining participation; Fig. 1). CHD-type distribution is presented in Figure 1 and Supplemental Table 2 (https://links.lww.com/CCX/B598). With the exception of two participants, our cohort underwent biventricular repair. The average gestational age at birth was 37.9 weeks (± 1.9 wk). The median age at CPB was 75 days post-natal (IQR, 13–115 d), the median weight was 4900 g (IQR, 3500–5900 g). Among these individuals, 12 of 15 were males (80%).

Figure 1.

Study flowchart. Flow diagram with participant inclusion and exclusion. Ao = aorta, AVo₂ = arteriovenous oxygen difference, AVSD = atrioventricular septal defect, CHD = congenital heart disease, c-TGA = congenitally corrected transposition of the great arteries, d-TGA = dextro-transposition of the great arteries, NIRS = near-infrared spectroscopy, RV = right ventricle, Spo₂ = peripheral oxygen saturation, TA = tricuspid atresia, TAPVR = total anomalous pulmonary venous return, TCD = transcranial color-coded duplex, TOF = tetralogy of Fallot, VSD = ventricular septal defect.

All our patients were treated prophylactically with milrinone (21). There were seven time points available for analysis with a mean time of 4.7 (sd, ± 1.2), 6.9 (± 0.9), 9.3 (± 1.7), 11.7 (± 2.1), 14.4 (± 2.7), 16.3 (± 1.8), and 19 (± 0.8) hours after the end of CPB. From 15 included patients, one passed away at 8.6 hours after admission (missing data after time point 4, univentricular physiology participant), we lost data of one participant after time point 4 due to ultrasound storing failure, four participants were extubated after time point 5 (missing data for time points 6 and 7), and one patient was extubated after time point 6 (missing data time point 7). Our final dataset included 87 data points of Spo₂, CSat, CBFv, and AVo₂ parameters obtained simultaneously.

Stability of Blood Flow Determinants

In our cohort, mean arterial blood pressure values ranged from 54 ± 9 mm Hg at time 1 (lowest values) to 59 ± 9 mm Hg at time 7 (highest values). Mean Paco₂ values ranged from 36 ± 6 mm Hg at time 1 to 40 ± 5 mm Hg at time 7. Mean hemoglobin values ranged from 140 ± 17 g/L at time 1 to 141 ± 12 g/L at time 7 and mean rectal temperature values ranged from 35.8 ± 1.3°C to 36 ± 0.6°C. Overall, time trends of Paco₂, hemoglobin, and temperature were stable in our cohort, outlining that their contribution to fluctuations in CBF is expected to be minimal (Table 1).

TABLE 1.

Near-Infrared Spectroscopy, Mean Cerebral Artery Doppler, and Clinical Parameters by Transcranial Doppler Time Point

Transcranial Doppler	1	2	3	4	5	6	7
Number of subjects	15	15	15	13	13	9	7
Hours post-bypass	4.7 (1.2)	6.9 (0.9)	9.3 (1.7)	11.7 (2.1)	14.4 (2.7)	16.3 (1.8)	19 (0.8)
Hemoglobin (g/L)	140 (17)	144 (19)	149 (17)	151 (13)	146 (13)	142 (14)	141 (12)
Oxygen saturation (%)	96 (7)	94 (8)	96 (7)	96 (6)	96 (6)	95 (8)	98 (2)
pH	7.41 (0.08)	7.42 (0.09)	7.38 (0.06)	7.42 (0.07)	7.40 (0.06)	7.38 (0.06)	7.36 (0.05)
Paco2 (mm Hg)	36 (6)	35 (6)	37 (4)	35 (4)	36 (3)	39 (5)	40 (5)
Lactate (mmol/L)	2.4 (1.6)	2 (1.2)	2 (1.4)	1.6 (0.9)	1.34 (0.6)	1.3 (0.6)	1.6 (0.6)
Arteriovenous oxygen difference (%)	0.34 (0.1)	0.38 (0.1)	0.37 (0.1)	0.35 (0.1)	0.33 (0.1)	0.31 (0.07)	0.26 (0.14)
Temperature (ºC)	35.8 (1.3)	36 (1.1)	36.1 (1.1)	36.2 (1.2)	36.2 (1.1)	36.3 (1.3)	36 (0.6)
Invasive systolic blood pressure	77 (15)	78 (17)	72 (13)	72 (8)	72 (9)	73 (11)	78 (12)
Invasive mean arterial pressure	54 (9)	59 (11)	57 (11)	57 (6)	56 (7)	55 (7)	59 (9)
Invasive diastolic blood pressure	51 (12)	52 (11)	48 (9)	48 (5)	47 (6)	45 (5)	48 (7)
State Behavioral Scale	–2.3 (1)	–1.9 (1.3)	–1.9 (1.3)	–1.8 (1)	–1.5 (1.1)	–1 (1.2)	–0.5 (1)
Right CSat (%)	59 (11)	57 (9)	57 (11)	60 (8)	63 (6)	64 (8)	65 (8)
Left CSat (%)	62 (12)	59 (13)	60 (25)	60 (10)	64 (8)	66 (8)	62 (8)
Right FTOE (%)	38 (12)	39 (13)	41 (13)	37 (12)	34 (9)	32 (13)	33 (9)
Left FTOE (%)	35 (12)	37 (16)	37 (15)	36 (12)	33 (10)	28 (12)	35 (8)
Right PSV (cm/s)	68 (28)	65 (23)	65 (22)	65 (25)	73 (29)	84 (33)	92 (35)
Left PSV (cm/s)	70 (31)	68 (24)	66 (25)	69 (28)	77 (32)	89 (34)	99 (44)
Right MV (cm/s)	35 (14)	34 (11)	34 (12)	35 (12)	38 (14)	46 (20)	50 (23)
Left MV (cm/s)	35 (15)	35 (12)	34 (12)	37 (15)	41 (17)	48 (20)	54 (27)
Right EDV (cm/s)	18 (8)	18 (6)	18 (9)	20 (7)	24 (14)	27 (14)	28 (16)
Left EDV (cm/s)	18 (9)	18 (7)	18 (7)	21 (9)	23 (10)	28 (13)	32 (18)
Right RI-MCA	0.73 (0.06)	0.71 (0.07)	0.71 (0.09)	0.68 (0.07)	0.67 (0.09)	0.68 (0.06)	0,71 (0.09)
Left RI-MCA	0.74 (0.07)	0.72 (0.06)	0.71 (0.07)	0.69 (0.05)	0.69 (0.04)	0.69 (0.06)	0.70 (0.08)
Right PI-MCA	1.44 (0.26)	1.37 (0.26)	1.42 (0.48)	1.28 (0.24)	1.27 (0.14)	1.27 (0.23)	1.37 (0.36)
Left PI-MCA	1.48 (0.25)	1.42 (0.23)	1.39 (0.28)	1.28 (0.21)	1.3 (0.15)	1.31 (0.23)	1.32 (0.3)

CSat = cerebral saturation, EDV = end-diastolic velocity, FTOE = fractional tissue oxygen extraction, MV = mean velocity, PI-MCA = pulsatility index of mean cerebral artery, PSV = peak systolic velocity, RI-MCA = resistive index of mean cerebral artery.

Values are expressed as mean (sd).

The Effect of Time

The postsurgical time trends of right and left MCA CBFv, PI, RI, and CSat values are summarized in Table 1. For the right side, the mean CSat at time point 1 (4.7 hr post-bypass) was 59% ± 11%, reaching its lowest value at 9.3 hours post-bypass, then gradually increasing to a final average of 65% ± 8% at time point 7 (19 hr post-bypass). Right MCA CBFv followed a similar pattern, with PSV starting at 68 ± 28 cm/s, MV at 35 ± 14 cm/s, and EDV at 18 ± 8 cm/s. The lowest values occurred between 6.9 and 9.3 hours post-bypass, followed by a rise to PSV of 92 ± 35 cm/s, MV of 50 ± 23 cm/s, and EDV of 28 ± 18 cm/s. Right MCA RI was 0.73 ± 0.06 and PI was 1.48 ± 0.25 at time point 1 with no significant change at time point 7 (RI-MCA 0.71 ± 0.06 and PI-MCA 1.37 ± 0.36 with respective p values of 0.20 and 0.49). Right FTOE was 38% ± 12% at time point 1, peaked at 9.3 hours post-CPB and gradually decreased to 33% ± 9% at time point 7. Similar trends and values were observed in the left MCA. AVo₂ difference peaked, exceeding 0.35, between 6.9 and 9.3 hours post-bypass. Graphical representations of these time trends are shown in Figure 2 and Supplemental Figures 2–4 in Appendix 1 (https://links.lww.com/CCX/B598).

Figure 2.

Bilateral mean cerebral artery (MCA) Doppler trends over first 24 hr. Time 0 is the end of bypass. Trends over time for right and left MCA. EDV = end-diastolic velocity, MV = mean velocity, PSV = peak systolic velocity.

The mixed-effect model revealed a significant association between time in hours after the end of CPB and TCD velocities (PSV, MV, EDV), left CSat, and AVo₂ as detailed in Supplemental Table 3 (https://links.lww.com/CCX/B598). For each hour after the end of CPB, right PSV increased by 1.43 cm/s (95% CI, 0.86–2 cm/s), left PSV increased by 1.68 cm/s (95% CI, 1.05–2.31 cm/s), right MV increased by 0.81 cm/s (95% CI, 0.46–1.17 cm/s), left MV increased by 1.05 cm/s (95% CI, 0.69–1.41 cm/s), right EDV increased by 0.61 cm/s (95% CI, 0.34–0.89 cm/s), and left EVD increased by 0.72 cm/s (95% CI, 0.46–0.99 cm/s). Left CSat increased by 0.54% every hour after the end of CPB (95% CI, 0.46–1.17), and right CSat did not show significance, but had a similar evolution, increasing over time by 0.47% per hour (95% CI, 0.265–0.678).

There was a significant association between AVo₂ and MV, AVo₂ and cFTOE, while accounting for the timing of evaluation since the end of CPB, as detailed in Table 2. The analysis indicated that, on average, an increase of 0.1 units in the AVo₂ difference was associated with a decrease of 5.8 cm/s of right MCA MV (95% CI, –8 to –3,5), a decrease of 5.4 cm/s of the left MCA MV (95% CI, –8 to –2.8), an increase of 3.4% (95% CI, 2–4.9%) for right cFTOE, and an increase of 4% for the left cFTOE (95% CI, 2.2–5.8). The mixed-effect models for CSat and Doppler indices and cFTOE and Doppler indices, also shown in Table 2, are less significant and clinically less relevant.

TABLE 2.

Random Mixed-Effect Models: Association Between Near-Infrared Spectroscopy, Doppler Parameters, and Arteriovenous Oxygen Difference

Outcome	Predictor	Coefficient (95% CI)	p
Right MV	AVo2	–5.8 (–8 to –3.5)	< 0.001
Left MV		–5.4 (–8 to –2.8)	< 0.001
Right FTOE	AVo2	3.4 (2–4.9)	< 0.001
Left FTOE		4 (2.2–5.8)	< 0.001
Right CSat	Right RI (*)	–1.4 (–3.3 to 0.6)	0.172
	Right PI (*)	–0.5 (–1.1 to 0.01)	0.059
Left CSat	Left RI	–0.6 (–1.1 to –0.12)	0.016
	Left PI	–0.7 (–1.5 to –0.02)	0.047
Right FTOE	Right RI	1.2 (–9.1 to 32.5)	0.27
	Right PI	0.5 (0.04–1)	0.037
Left FTOE	Left RI	0.7 (0.2–1.2)	0.008
	Left PI	0.6 (–0.1 to 1.5)	0.11

AVo₂ = arteriovenous oxygen difference, CSat = cerebral saturation, FTOE = fractional tissue oxygen extraction, MV = mean velocity, PI = pulsatility index, RI = resistive index

*RI and PI are expressed as 0.1 unit of increment or decrease.

In this table, the variable “Time” represents the time in hours since the end of bypass. Bolded values are statistically significant.

Figure 3 shows the graphical relationship between AVo₂ and MV, as well as between AVo₂ and FTOE. Lower MCA MV correlated with a higher AVo₂ difference, with extremes observed at an AVo₂ close to 0.5 when MV was around 20 cm/s, and an AVo₂ of 0.2 when MV exceeded 80 cm/s. A right MV of 25.1 cm/s corresponded to an AVo₂ value of 0.35 on the locally estimated scatterplot smoothing curve, while an MV of 23 cm/s corresponded to an AVo₂ value of 0.40. Similarly, a left MV of 25.3 cm/s corresponded to an AVo₂ value of 0.35, and an MV of 17.7 cm/s corresponded to an AVo₂ value of 0.40. The graph shows three distinct zones: zone 1, where AVo₂ is directly proportional to the decrease in MV from 30 to 20 cm/s; zone 2, where MV ranges between 30 and 45 cm/s and AVo₂ stabilizes around 0.35–0.37; and zone 3, where further increases in MV (above 45 cm/s) result in a steady decline in AVo₂. Regarding the AVo₂/FTOE relationship, higher cFTOE corresponded with higher AVo₂ values. There was a stable region around an FTOE of 40, where AVo₂ remained around 0.35.

Figure 3.

Correlations arteriovenous oxygen difference (AVo₂)/mean velocity (MV) and AVo₂/fractional tissue oxygen extraction (FTOE). The top outline the relationship between mean cerebral artery (MCA) MV and AVo₂ for the right and left MCA. The bottom graphs represent the relationship between cerebral FTOE (cFTOE) and AVo₂ difference bilaterally. A right MV of 25.1 cm/s corresponded to an AVo₂ value of 0.35 on the locally estimated scatterplot smoothing curve, while an MV of 23 cm/s corresponded to an AVo₂ value of 0.40. Similarly, a left MV of 25.3 cm/s corresponded to an AVo₂ value of 0.35, and an MV of 17.7 cm/s corresponded to an AVo₂ value of 0.40.

DISCUSSION

Our study analyzed seven time points and 87 data points (including of Spo₂, CSat, CBFv, and AVo₂) collected from 15 infants with CHD during the first 24 hours following CPB, a critical period of vulnerability regarding cerebral perfusion. Overall, there was a parallel evolution of CBFv and CSat with correlation of AVo₂ and cFTOE and cerebral mean flow velocity.

Cerebral Saturation and Cerebral Blood Flow Velocities

During the first night after CPB, we observed that bilateral MCA CBFv, including PSV, MV, and EDV, followed a parallel pattern, decreasing from the time of PICU admission and reaching their lowest values between 6.9 and 9.3 hours post-CPB, followed by a steady increase from 9.3 to 24 hours. Interestingly, these findings compare to Wernovsky et al (2) results in neonates and infants after an arterial switch operation. Their team examined the hemodynamic profile, including cardiac index, pulmonary vascular resistance, and systemic vascular resistance, in 122 participants during the first postoperative night. In that study, the cardiac index decreased by 32.1% ± 15.4%, while pulmonary and systemic vascular resistance increased. A cardiac index of less than 2.0 L/min/m² was observed in 24% of the patients, with the nadir typically occurring 9–12 hours post-surgery. Similarly, in our cohort, we noted the nadir of bilateral CBFv, CSat, and the peak of both cFTOE and AVo₂ differences between 6.9 and 9.3 hours post-bypass. This suggests that the period of suboptimal cerebral perfusion likely occurs between 6 and 10 hours after CPB marking a critical window for monitoring and intervention. During this time window, cerebral hemodynamics could serve as a potential surrogate for CO. This information can guide brain-based goal-directed therapies, potentially optimize outcomes by mitigating neurologic risks. Further research is needed to explore the potential benefits of maneuvers such as vasopressors/inotropes titration to improve CO and enhance cerebral perfusion, adjust Paco₂ targets, and reduce mean airway pressure to improve cerebral venous drainage. The higher PSV observed at time points 6 and 7 (16.3 and 19 hr post-CPB) could be a combination of higher CO and hyperemia. Further research is needed to understand the risk for acute brain injury in this scenario (e.g., reperfusion injury) and the possible need for intensivist to readjust hemodynamic therapies in this context.

CSat has been described to correlate with cerebral oxygen consumption (22) and variably correlate with CO in preterm neonates and pediatric cardiac population (23–25), but not in term neonates during the transitional period (26). A retrospective study of 91 single ventricle infants below 6 months of age, a CSat of less than 57% at 6 hours post-CPB had 91% sensitivity and 72% specificity to detect LCOS (18). In postoperative sedated patients, cerebral oxygen consumption should remain stable in the absence of seizures, fever, new parenchymal brain disease, brain death, stable hemoglobin, and Co₂ levels. The observed decreasing trend in CSat during the first 9.6 hours and in conjunction with decreasing trend of CBFv is likely associated with a lower CO related to myocardial dysfunction following CPB.

From Cerebral Doppler to Cerebral Blood Flow

During the 1990s, TCD began to be used with the aim of understanding CBFv, to provide information on cerebral perfusion, and correlating these measures with neurologic outcomes (11, 27, 28). Current scientific evidence is variable on the association between changes in CBF and MCA CBFv (29, 30). TCD now plays an essential role in multimodal neuromonitoring during CPB, allowing monitoring of CBF during circulatory arrest or low-flow states (9, 10). Additionally, both high and low CBFv during CPB have been shown to have diagnostic value for predicting postoperative cognitive dysfunction in adult patients, potentially helping to minimize postsurgical delirium (14, 31, 32). Although the behavior of CBFv after CPB is not well-studied, the same principles may apply. In our cohort, after ensuring stability of central body temperature, MAP, Co₂, pH, and hemoglobin, we extrapolated that changes in CBFv reflect acute changes in CO, potentially with or without a component of cerebral vasodilation. We believe that the coincidence of the nadir in CBFv by TCD in the time window of low CO suggests a reduction in CBF in a population with often impaired cerebral autoregulation (33). The observed combination of increased FTOE and decreased NIRS during the first 6 to 10 hours post-CPB may further suggest a decrease in CO rather than cerebral vasodilation. Requiring further research, the identification of low brain perfusion due to low CO could trigger the intensivist to titrate hemodynamic resuscitation. Similarly, hemodynamic and ventilation therapies could be adjusted in case of evolution to high CO with cerebral hyperemia.

Cardio-Cerebral Coupling

Control of CBF is vital in managing critically ill cardiac patients. Although the link between CO and CBF is theoretically sound, consistent evidence is lacking (29). Our cohort’s findings—showing the timing of peak AVo₂, peak cFTOE, and nadir CBFv and CSat coinciding with a known nadir in CO around 12 hours post-CPB (2)—suggest potential cardio-cerebral coupling. In addition, studies have shown that approximately 15% of children with CHD have impaired autoregulation post-surgically, making them more vulnerable to fluctuations in CO and at greater risk of cerebral hypoperfusion during perioperative periods (34).

Limitations

Several limitations may affect the generalizability of our findings. As a single-center study with a small sample size, extrapolation is limited. Two participants with univentricular physiology were included in this cohort. While their physiology differs from those with biventricular repairs, their inclusion was intended to reflect the diversity of the postoperative CHD population in the cardiac ICU. Nonetheless, this may introduce variability in CSat values and should be considered when interpreting the results. Importantly, the use of cFTOE in our analysis helps mitigate this variability by accounting for differences in systemic oxygen delivery and providing a more physiologically surrogate measure of cerebral oxygen extraction. The observational nature meant our single ultrasonographer was not blinded to CSat, which is routinely used in the PICU. Additionally, patient 1 had a premature death, and patient 13 lost data after time 5, potentially affecting the analysis. Regarding study’s strengths, our data were prospectively and rigorously acquired using a standardized protocol by one ultrasonographer. To our knowledge, this is the first study to provide new data on CSat and CBFv monitoring after CPB in infants before the arrival of robotic TCD.

CONCLUSIONS

Our cohort of CHD patients monitored post-CPB revealed a parallel evolution of CBFv and CSat with correlation of AVo₂ with MV. CSat and CBFv have an earlier nadir (6–10 hr) compared with the previously published evolution of cardiac index in infants after CPB (9–12 hr). In the absence of direct CO monitoring in critically ill children, noninvasive cerebral oxygenation and blood flow monitoring could be an alternative method to measure LCOS. Further studies utilizing advanced bedside monitors and more sophisticated multimodal neuromonitoring platforms are essential to deepen our understanding of CBF dynamics following CPB and to guide brain-focused resuscitation strategies.

ACKNOWLEDGMENTS

We thank the cardiac intensive care nursing team at the Montreal Children’s Hospital for their disposition and collaboration.

Supplementary Material

cc9-8-e1364-s001.pdf ^{(979.5KB, pdf)}