Quality indicators in pediatric cardiopulmonary bypass: Current perspectives and future directions

Objective of this review

This review aims to provide a comprehensive analysis of quality indicators in pediatric cardiopulmonary bypass (CPB), highlighting their clinical relevance, evaluating current literature, and identifying avenues for standardization and technological innovation. By synthesizing available data, this review seeks to define best practices in pediatric perfusion and propose actionable recommendations for optimizing perioperative outcomes in congenital heart surgery (Figure 1).

Literature search strategy

A structured and comprehensive literature review was conducted using PubMed, Embase, and the Cochrane Library to identify relevant studies examining quality indicators in pediatric CPB. The search focused on indicators such as indexed oxygen delivery (DO₂i), mean arterial pressure (MAP), near-infrared spectroscopy (NIRS), activated clotting time (ACT), venous oxygen saturation (SvO₂/ScvO₂), lactate levels, hematocrit, cardioplegia and priming volume (Table 1). Search terms included combinations of: “pediatric cardiopulmonary bypass,” “quality indicators,” “oxygen delivery,” “mean arterial pressure,” “NIRS,” “venous oxygen saturation,” “lactate,” “hematocrit,” “cardioplegia,” “priming volume,” and “perfusion strategies.”

OPEN IN VIEWER

Table 1.

Quality indicators in the pediatric cardiopulmonary bypass.

Quality indicator	Authors (published year)	Ranges and clinical impacts
DO2i	Dreher et al. (2023) 5	>340 mL/min/m2 in neonates	Maintaining DO2i > 340 mL/min/m2 in neonates and >400 mL/min/m2 in infants has been associated with a reduced incidence of AKI
		>400 mL/min/m2 in infants
	Zhang et al. (2022) 6	>353 mL/min/m2	DO2i of 353 mL/min/m2 is the lowest safe threshold in infants undergoing CPB. Below this threshold, there was a high probability of AKI.
	Hayward et al. (2022) 7	>350 mL/min/m2	Time spent on CPB with oxygen delivery <350 mL/min/m2 is independently associated with AKI in pediatric patients undergoing cardiac surgery
	Gao et al. (2022) 10	>258 mL/min/m2 in hypothermia	The critical DO2i for preventing AKI was 258 mL/min/m2 in the hypothermia phase and 281 mL/min/m2 during rewarming
		>281 mL/min/m2 for rewarming
(CPB flow)	Ramakrishnan et al. (2023) 13	200 mL/kg/min in neonates	The high-flow, high-hematocrit strategy is linked to a low incidence of AKI (∼3%) and neurological events (<1%), outperforming the conventional strategy
		150–175 mL/kg/min in older infants
MAP	Holmgaard F et al. (2019) 16	40–50 mmHg	Higher MAP induced by vasopressors was associated with lower mean regional cerebral oxygen saturation and an increased frequency of cerebral desaturation events
	Brady K et al. (2010) 17	Averaging 42 ± 7 mmHg	Hypotension impaired autoregulation, with individual MAP lower limits identified in 77% of cases, averaging 42 ± 7 mm Hg using a COx threshold of 0.4
SvO2	Ranucci et al. (2010) 2	ScvO2 <68% and lactate >3 mmol/L	Nadir ScvO2 <68% during CPB and peak lactate >3 mmol/L were strongly associated with major morbidity. The combined index had a positive predictive value of 89% for major complications
	Crowley et al. (2021) 21	ScvO2 <40% for >18 min	ScvO2 desaturation <40% for more than 18 min was associated with 100% sensitivity and 97.6% specificity for major adverse events. The area under the curve <40% also correlated with renal and hemodynamic complications
Lactate levels	Ranucci et al. (2010) 2	Lactate >3 mmol/L and ScvO2 <68%	The combination had a PPV of 89% for major morbidity. More accurate than either parameter alone. Suggests impaired oxygen delivery
	Kalyanaraman et al. (2008) 24	Lactime > 48 h	Lactime (duration of lactate >2 mmol/L) had the strongest predictive value for mortality. PPV: 60% overall, 75% in neonates. Stronger than initial or peak lactate alone
	Park et al. (2012) 3	ΔLactate (weaning → chest closure)	Significant intraoperative lactate rise associated with MAEs. AUC = 0.810. Best predictor among measured variables for critical complications
	Kanazawa et al. (2015) 27	ΔLactate after CPB	Post-CPB lactate increase independently predicted ICU stay and SAE risk. Used in regression analysis Emphasizes the value of post-CPB serial monitoring
	Toda et al. (2005) 5	Lactate >4 mmol/L in older children (bloodless prime)	In older children, elevated lactate was not associated with poor outcomes. Suggests bloodless CPB prime affects lactate independently of perfusion. Interpretation should be age- and context-specific
	Agrawal et al. (2012) 29	Persistent lactate >4.0 mmol/L	Postoperative persistent hyperlactatemia was associated with mortality, longer ventilation, and PICU stay. Also correlated with longer cross-clamp and CPB times. Reinforces prognostic utility of early PICU lactate trends
CPB fluid balance	Naik SK et al. (1991) 30	MUF: ∼10 min post-CPB to Hct 36%–42%	Reduced TBW rise (4.0% vs 11.1%), blood loss, and transfusion requirements; improved systolic/diastolic BP.
	Bierer J et al. (2019) 31	Not specified	MUF improves myocardial function and fluid balance; ZBUF reduces cytokines and improves pulmonary compliance
	Riegger L et al. (2002) 32	Albumin vs crystalloid prime	Albumin decreased fluid extravasation but increased PRBC transfusion; osmotic pressure was maintained post-CPB.
	Hassinger AB et al. (2014) 33	>5% fluid overload by POD1	Preceded AKI; associated with poor diuretic response, more inotrope use, and longer ICU/hospital stay
	Delpachitra M et al. (2017) 5	≥5% cumulative balance by day 2	Longer mechanical ventilation (211 vs 93 h), extended PICU/hospital stay; not directly linked to mortality
	Grist G et al. (2011) 35	Positive FBC at CPB end	Mortality rate 9.55% vs 4.65% in negative/neutral FBC; OR for mortality = 1.73
Hematocrit	Wypij D et al. (2008) 36	≥24%	Associated with higher psychomotor development scores and lower lactate levels. Below this threshold is linked to increased intraoperative fluid balance
	Newburger J et al. (2008) 37	25%–35%	No significant difference in major outcomes; both groups had below-normal development scores at 1 year
	Ramakrishnan K et al. (2023) 38	≥32%	High-flow, high-hematocrit strategy linked to lower incidence of AKI (∼3%) and neurological events (<1%)
	Fang W et al. (1997) 39	≤14% (general), ≤17% (high-risk)	Hematocrit ≤14% during CPB independently associated with increased mortality. Higher risk in high-risk patients at ≤17%
	Liu H et al. (2020) 40	Normal vs. high hematocrit	High preoperative hematocrit modified AKI risk during reoxygenation, suggesting individualized strategies are necessary
	Schill DM. (1990) 41	Priming with autologous blood	Maintained hematocrit levels with minimal drop on CPB; reduced need for transfusion and ultrafiltration
Cerebral NIRS	Ricci et al. (2015) 44	Cerebral rSO2 65%–75% during CPB	Cerebral NIRS remained 65%–75% on average, while somatic NIRS reached 80%–90%, suggesting cerebral-specific monitoring is crucial for detecting vulnerable perfusion regions
Temperature	Golab et al. (2000) 54	32°C vs. 28°C	Both temperatures facilitated CPB equally; mild hypothermia showed longer aPTT but no difference in blood loss
	Stocker C et al. (2011) 55	Mild (34°C) vs. Moderate (24°C) hypothermia	No significant difference in mechanical ventilation duration, ICU stay, blood loss, or inflammatory markers
	Xiong T et al. (2021) 56	Normothermic (≥36°C) vs. Hypothermic (<34°C)	Normothermic CPB is associated with lower lactate, creatinine, bleeding risk, and major adverse events
	Greely W et al. (1989) 15	18°C–32°C	CBF decreased under hypothermia and returned or exceeded baseline during rewarming, except after total circulatory arrest
	Browne L et al. (2021) 57	<37°C (arterial outlet); ≤10°C gradient	Limit arterial outlet temp to <37°C; maintain ≤10°C gradient during rewarming to prevent cerebral hyperthermia and gas emboli
	Engelman et al. (2015) 58	Normothermia rewarming to ∼36.5°C–37°C	Use of warming garment post-CPB significantly improved nasopharyngeal temp recovery; no device-related adverse events
Priming volume	Richmond et al. (2013) 60	Indexed CPB priming volume (mL/m2)	Indexed prime volume independently predicted peri-operative RBC transfusion; adjusted OR 1.006 per +1 mL/m2 after controlling for RACHS-1, surgeon, core temperature, and pre-op Hb. In infants, even small absolute prime changes materially shift risk. Lower prime associated with shorter LOS and lower mortality
	Mathieu et al. (2025) 61	Weight-indexed priming volume (mL/kg)	Patients below the median prime (∼37.9 mL/kg) had lower peak creatinine and CRP, shorter ICU stay and ventilation time; higher prime volumes correlated with need for renal replacement therapy and mediastinitis
	Boettcher et al. (2018) 63	Asanguineous (bloodless) prime with miniaturized circuits	Feasible across ≤15 kg using tiny oxygenators, short tubing. Static primes ∼75–280 mL by size; ability to avoid on-CPB transfusion depends strongly on weight (transfusion during CPB 53% <3 kg vs 14% >8 kg). Transfusion (especially on-CPB) independently associated with delayed extubation and longer ICU stay
	Kim et al. (2016) 64	Mini-volume priming (absolute mL reduction)	Adoption of “mini-volume” priming reduced prime from ∼322 mL → ∼126 mL and cut on-CPB transfusion volume (87 mL vs 227 mL). Larger priming volume (and higher RACHS-1) was a significant risk factor for post-op ECMO.
	Merkle et al. (2004) 62	Non-haemic prime technique in <6 kg	Circuit miniaturization (shortened/downsized lines, VAVD, exclude unused components) reduced prime ∼300 mL → ∼190 mL while maintaining safety (mean on-CPB Hct 22.5%; uneventful courses). Demonstrates technical pathway to smaller primes
	Dehaki et al. (2020) 65	Prime composition: Zero-balance ultrafiltration (Z-BUF) of priming blood	Z-BUF before initiation normalized electrolytes/osmolality and was associated with less postoperative bleeding/transfusion, faster extubation, and shorter ICU stay versus control. Supports prime optimization alongside volume reduction
Cardioplegia	Mylonas et al. (2017) 67	Blood vs. crystalloid (systematic review & meta-analysis)	Cardioplegia type: Blood vs. crystalloid (systematic review & meta-analysis). Across pediatric studies, no consistent advantage in clinical endpoints (ICU/hospital stay, 30-days mortality). Only cTnI at 4 h modestly lower with blood; otherwise biomarkers and outcomes comparable
	Zhou et al. (2021) 68	Warm blood vs. cold crystalloid (network meta-analysis)	Strategy ranking: Warm blood vs. cold crystalloid (network meta-analysis). Warm blood cardioplegia showed lower 24-h cTnI than cold crystalloid; HTK/Custodiol tended toward higher 24-h cTnI. No consistent differences in major clinical outcomes across strategies
	Mohammed et al. (2022) 72	del Nido vs. St. Thomas blood in neonates	Shorter CPB and cross-clamp times and fewer doses with del Nido; no difference in early mortality and most postoperative/discharge outcomes
	Ler et al. (2020) 73	del Nido vs. St. Thomas: Broader synthesis (adult + pediatric)	Del Nido associated with shorter CPB/ACC (notably in adults) and lower defibrillation rates in both adults and pediatrics; similar postoperative outcomes overall
	Fang et al. (2015) 66	Cardioplegia type: Blood vs. crystalloid	Blood cardioplegia associated with lower postoperative lactate; no significant differences in troponin I, ventilation duration, or ICU length of stay compared with crystalloid. Overall clinical outcomes similar

Eligible studies included randomized controlled trials, observational studies, meta-analyses, and systematic reviews. Studies focusing solely on adult populations or extracorporeal membrane oxygenation (ECMO) were excluded, as were case reports with fewer than five patients to maintain scientific robustness and relevance to pediatric cardiac surgery (Figure 2).

OPEN IN VIEWER

Figure 2.

Quality Indicators sheet in the Texas Children’s Hospital North Austin Campus.

Data extraction and analysis

Each eligible study was reviewed to extract key information, including study design, sample size, patient demographics, surgical details, reported outcomes, and their described relationship to CPB quality indicators. We organized and summarized the reported findings across studies to highlight commonly referenced thresholds and trends related to each perfusion parameter.

The extracted data were compiled to provide an overview of how different quality indicators have been applied in clinical settings and their potential associations with outcomes such as acute kidney injury, neurological events, ICU length of stay, and mortality. Particular attention was given to studies that reported clinically meaningful insights or proposed benchmarks. This synthesis aimed to support future efforts toward developing standardized CPB protocols in pediatric cardiac surgery.

Oxygen delivery rate

The oxygen delivery rate (DO₂) is a vital determinant of organ perfusion and metabolic function during cardiopulmonary bypass (CPB), calculated as:

However, more recent data show mixed results. A retrospective study of 479 pediatric patients found no significant association between DO₂-related metrics (nadir DO₂, mean DO₂, AUiDO₂) and AKI; however, lower DO₂ was strongly linked to elevated lactate levels, indicating metabolic stress.^8,9 Meanwhile, a prospective study of 413 low-weight infants showed that nadir DO₂ during mild hypothermia (32°C–34°C) and rewarming was independently associated with AKI but not during moderate hypothermia (26°C–32°C), suggesting temperature modulation may influence outcomes. ¹⁰

Studies from the Penn State Hershey Pediatric Cardiac Research Laboratory further emphasize that oxygen delivery, inflammatory responses, and gaseous microemboli all contribute to postoperative complications, including AKI. ¹¹ These findings support the ongoing exploration of individualized perfusion strategies and circuit design. Beyond renal injury, low DO₂ has also been linked to neurological injury. A prospective study found that lower nadir DO₂ correlated with elevated levels of glial fibrillary acidic protein, a biomarker of brain injury, during and after CPB—even after adjusting for surgical factors—highlighting its role in cerebral protection. ¹²

Additionally, a high-flow, high-hematocrit strategy has shown promising results in neonates and small infants, using flows of 200 mL/kg/min in neonates and 150–175 mL/kg/min in older infants, while maintaining hematocrit ≥32% during CPB. This approach has been linked to a low incidence of AKI (∼3%) and neurological events (<1%), outperforming conventional strategies. ¹³ Adjuncts such as aggressive ultrafiltration, pH/alpha-stat management, and limiting early postoperative milrinone and albumin have also contributed to improved outcomes.

Mean arterial pressure

Mean arterial pressure (MAP) is crucial in maintaining adequate systemic and organ perfusion during pediatric CPB. Striking the right balance is essential: insufficient MAP may compromise oxygen delivery, while excessive MAP can increase shear stress, contribute to endothelial injury, and trigger systemic inflammation. Despite its clinical significance, the optimal MAP target during pediatric CPB remains a subject of ongoing investigation, with recent studies advocating for individualized strategies tailored to patient-specific physiology.

The relationship between MAP and cerebral oxygenation has been evaluated in pediatric patients, revealing a modest positive correlation between MAP and regional cerebral oxygen saturation (rSO₂) in children (r = 0.27). However, this association was negligible in neonates (r = 0.06) and infants (r = 0.05), likely reflecting immature cerebral autoregulation. ¹⁴ Furthermore, temperature management significantly affects cerebral blood flow (CBF) during CPB. CBF declines across all age groups under hypothermic conditions (18°C–32°C). In moderate hypothermia (25°C–32°C), CBF typically recovers during rewarming and may exceed baseline post-bypass. Conversely, after deep hypothermic circulatory arrest at 18°C for an average of 32 ± 12 min, CBF fails to return to baseline levels. Notably, a correlation between MAP and CBF is observed only under deep hypothermia, not at moderate temperatures. ¹⁵ These findings support using individualized MAP targets based on age, temperature, and surgical context.

The use of vasopressors to regulate MAP has also been studied. A randomized trial comparing high MAP (70–80 mm Hg) to low MAP (40–50 mm Hg) during CPB found that vasopressor-induced MAP elevation was associated with lower mean rSO₂ and more frequent cerebral desaturation events. ¹⁶ This suggests that pharmacologically increasing MAP does not necessarily enhance cerebral perfusion and may even be harmful.

To optimize MAP management, real-time monitoring of cerebral autoregulation using the cerebral oximetry index (COx) has emerged as a promising tool. A pilot study identified individual lower limits of MAP in 77% of pediatric patients using a COx threshold of 0.4, with an average autoregulation threshold of 42 ± 7 mm Hg. ¹⁷ While this approach allows for personalized MAP targets, it is not yet widely adopted in clinical practice and requires further validation. Additionally, elevated or fluctuating MAP has been linked to impaired cerebrovascular pressure autoregulation (CPA), especially in neonates and infants. High end-tidal CO₂ levels (≥40 mm Hg) further increase the risk of disturbed CPA, indicating that both pressure and ventilation parameters must be considered together. These findings emphasize the complexity of cerebral perfusion regulation and reinforce the need for patient-specific MAP targets during pediatric CPB. ¹⁸

Venous oxygen saturation

Venous oxygen saturation (SvO₂) is a critical parameter for evaluating the balance between oxygen delivery and consumption during CPB. Continuous SvO₂ monitoring offers real-time insight into systemic oxygenation status, allowing for timely perfusion adjustments to optimize outcomes. Among in-line monitoring systems, the CDI 500/550 has demonstrated superior accuracy and consistency in measuring both SvO₂ and hematocrit compared to other devices like the Gish StatSat and Bentley Oxysat, especially across varying blood flow rates, temperatures, and hematocrit levels, making it a reliable tool in pediatric CPB management. ¹⁹

Central venous oxygen saturation (ScvO₂), a surrogate for mixed venous oxygenation, has also gained prominence as a dynamic indicator of oxygen extraction during and after CPB. Studies have shown that a nadir ScvO₂ below 68% during bypass correlates with elevated peak lactate levels and is significantly associated with major morbidity and mortality in pediatric cardiac surgery patients. Notably, the combination of ScvO₂ <68% and lactate >3 mmol/L yields a positive predictive value of 89% for major morbidity. ²

For continuous and precise ScvO₂ monitoring, devices such as the Pediasat catheter have been validated in pediatric populations, demonstrating minimal bias and acceptable percentage errors across intraoperative and postoperative phases. ² Similarly, modified pediatric ScvO₂ catheters have shown a strong correlation with co-oximetry readings (r = 0.81) and accurate trending within physiological ranges in critically ill children. ²⁰ Importantly, prolonged periods of low ScvO₂ have been linked to poor outcomes. In one prospective study, central venous saturations <40% for more than 18 min were associated with major adverse events, with 100% sensitivity and 97.6% specificity. Furthermore, the ScvO₂ area under the curve <40% correlated significantly with clinical markers such as reduced creatinine clearance, increased inotrope use, and longer ICU and hospital stays. ²¹

Despite the widespread use of SvO₂ monitoring, its ability to predict cerebral oxygenation remains limited. Studies have shown that SvO₂ does not consistently correlate with jugular bulb venous oxygen saturation (SjO₂), which more directly reflects cerebral oxygen extraction. In one analysis, the correlation between SvO₂ and SjO₂ was poor (r² = 0.14), and significant cerebral desaturation (SjO₂ < 50%) occurred in 50% of cases following rewarming, despite seemingly adequate SvO₂ values. ²² This discrepancy underscores the limitations of relying solely on SvO₂ for cerebral perfusion assessment and highlights the importance of complementary cerebral monitoring strategies. Near-infrared spectroscopy (NIRS) has emerged as a non-invasive alternative for assessing regional cerebral oxygenation. However, comparisons between ScvO₂ and NIRS have revealed only moderate agreement, with wide limits of agreement (e.g., a mean difference of more than 10% between ScvO₂ and brain or somatic NIRS). ²³ While NIRS may not serve as a precise surrogate for global perfusion, it provides valuable insight into regional oxygenation, particularly for vulnerable organs like the brain and kidneys in neonates and infants.

Hypothermic CPB further complicates the interpretation of oxygenation metrics. In pediatric patients undergoing surgery with hypothermic CPB, cerebral rSO₂ consistently decreased at all major operative stages, whereas SjO₂ increased and SvO₂ remained relatively unchanged. ²⁴ These findings suggest that temperature-induced changes in cerebral perfusion and oxygen metabolism must be considered when interpreting oxygenation markers. BIS monitoring has also been shown to decrease during hypothermia, and both rSO₂ and BIS correlate positively with cerebral perfusion pressure (CPP), reinforcing the need for multimodal neuromonitoring. ²⁵

Lactate levels

Lactate is a metabolic byproduct of anaerobic respiration and serves as a dynamic biomarker for tissue hypoxia and impaired oxygen delivery during CPB. Under normal conditions, lactate levels during CPB are expected to remain below 2 mmol/L. However, persistent or rising lactate concentrations have been consistently associated with metabolic acidosis, poor perfusion, increased morbidity, and prolonged ICU stay in pediatric cardiac surgery patients.^2,24

The combination of low central venous oxygen saturation (ScvO₂ < 68%) and elevated lactate (>3 mmol/L) during CPB has demonstrated strong predictive value for significant postoperative morbidity, with a positive predictive value of 89% for adverse outcomes. ² While peak lactate alone offers some prognostic information, recent evidence suggests that the duration of hyperlactatemia, termed lactime, is a more reliable predictor of mortality. In one study, a lactime >48 h had a positive predictive value of 60% across all pediatric age groups and 75% among neonates, compared to only 15% for peak lactate >6 mmol/L. ²⁴

Intraoperative lactate trends provide additional clinical insight. Notably, the increase in lactate between CPB weaning and chest closure was significantly associated with major adverse events (MAEs), including hemodynamic instability, ECMO support, renal dysfunction, or death. ²⁶ This rise in lactate (ΔLactate) demonstrated a strong predictive value for MAEs (AUC: 0.810), reinforcing its role as a real-time intraoperative risk marker.

Furthermore, retrospective analysis of 459 pediatric cases confirmed that a post-CPB lactate rise was independently associated with both prolonged ICU stay and increased incidence of serious adverse events (SAEs). Multivariate regression showed a significant correlation between ΔLactate and poor outcomes, suggesting that serial lactate measurements—especially after CPB—may offer a practical and actionable index of postoperative risk. ²⁷ Factors influencing lactate levels include patient age, inflammatory responses, and the composition of CPB prime solutions. Recent findings suggest that lactate accumulation in older children, particularly those receiving a bloodless prime, may not reflect the same level of metabolic stress or oxygen debt seen in younger children undergoing CPB with a blood prime. In one study, 32% of older children had postoperative lactate levels >4 mmol/L despite uncomplicated clinical courses, while only 6% of younger children exhibited similar elevations (p = 0.0002). Multivariate analysis identified the priming solution as an independent predictor of elevated lactate, highlighting the need for age- and context-specific interpretation of lactate trends. ²⁸

Consistent and timely lactate monitoring during and after CPB remains a valuable adjunct to perfusion assessment. Serial point-of-care lactate measurements, both intraoperatively and during the early postoperative period, may help identify patients at risk for adverse outcomes. In a prospective study, persistent lactate elevation >4.0 mmol/L in the postoperative period was significantly associated with mortality, prolonged mechanical ventilation, and extended PICU stays. Peak lactate levels correlated considerably with longer aortic cross-clamp time and CPB duration. ²⁹ These results support the integration of lactate trends, rather than isolated values, into routine hemodynamic monitoring to optimize postoperative care and enable early intervention in pediatric cardiac surgery patients.

CPB fluid balance

Intraoperative CPB fluid management in pediatric cardiac surgery significantly influences postoperative outcomes. Modified ultrafiltration (MUF) after CPB has been shown to effectively reduce hemodilution, total body water (TBW), blood loss, and transfusion requirements while improving arterial blood pressure and postoperative hemodynamics. In one randomized study, MUF significantly lowered TBW rise (median 4.0% vs 11.1%, p = 0.0001) and decreased blood transfusion needs compared to controls. ³⁰ Various ultrafiltration techniques have demonstrated physiological benefits, including conventional ultrafiltration (CUF), MUF, and zero-balance ultrafiltration (ZBUF). MUF improves myocardial function and reduces fluid overload, while ZBUF is associated with decreased inflammatory cytokines and better pulmonary compliance. However, evidence on whether these techniques reduce ventilation time or ICU stay remains mixed due to heterogeneity in protocols and patient populations. ³¹

The composition of the CPB prime solution also affects perioperative fluid dynamics. A prospective study comparing albumin versus crystalloid priming showed that albumin priming attenuated fluid accumulation and maintained higher colloid osmotic pressure, though it also increased the need for blood transfusion. ³² These findings underscore the importance of balancing vascular volume preservation with transfusion risk.

Postoperative fluid overload, particularly when it exceeds 5% of body weight by the end of the first postoperative day, is associated with adverse outcomes, including prolonged mechanical ventilation, increased ICU stay, and acute kidney injury (AKI). In a multicenter analysis, fluid overload not only preceded AKI but also correlated with poor diuretic response and higher inotrope requirements. ³³ Similarly, early postoperative fluid accumulation ≥5% was associated with longer ventilation (median 211 vs 93 h) and longer PICU and hospital stays, even though it was not directly associated with mortality. ³⁴

Notably, positive fluid balance during CPB has correlated with increased mortality. In a study of 1540 pediatric patients, those with positive fluid balance at the end of CPB had a significantly higher mortality rate (9.55%) compared to those with neutral or negative balances (4.65%) (p < 0.05). The adjusted odds ratio for mortality was 1.73, reinforcing that intraoperative fluid accumulation may contribute to poor outcomes and should be closely monitored. ³⁵ The target CPB fluid balance should ideally be ≤5% at CPB termination to reduce the risk of fluid-related complications. Aiming for a net-neutral to slightly negative fluid balance (∼0 to −5% of body weight) helps prevent excessive hemodilution and tissue edema while maintaining hemodynamic stability. Optimized ultrafiltration (CUF/MUF) strategies should be employed to reach the goal hematocrit without removing excessive volume.

However, standardizing fluid balance tracking during CPB remains challenging due to differences in measurement practices and data capture methods across institutions. Real-time monitoring using automated perfusion data systems and integration into electronic health records could improve consistency and enable better benchmarking. Databases such as PediPerform could support threshold development and quality improvement, although implementation cost is a barrier.

Hematocrit

Hematocrit represents the proportion of red blood cells in the total blood volume and plays a vital role in oxygen-carrying capacity during CPB. Maintaining an optimal hematocrit level is essential to ensure sufficient oxygen delivery while avoiding the complications associated with both excessive hemodilution and elevated blood viscosity. Generally, a target hematocrit range of 25%–40% during CPB is recommended, balancing perfusion efficacy and rheological safety.

Recent studies have emphasized the importance of maintaining hematocrit levels at or above 24% during low-flow CPB. In the combined Boston hematocrit trials, hematocrit levels below 23.5% were associated with poorer psychomotor development at 1 year of age and higher intraoperative lactate levels, while outcomes plateaued beyond 24%. ³⁶ Similarly, a randomized comparison of hematocrit 25% versus 35% during hypothermic CPB in infants found no major differences in overall outcomes, though the lower hematocrit group had worse intraoperative fluid balance and cerebral oxygen saturation. Both groups scored below normative developmental values at 1 year, indicating that hematocrit is only one of many neurodevelopmental determinants. ³⁷

A more aggressive approach using a high-flow, high-hematocrit (≥32%) CPB strategy has been associated with better outcomes, including reduced postoperative acute kidney injury rates and neurological complications. This approach incorporates adjunct techniques such as ultrafiltration and tailored pharmacologic support, contributing to favorable fluid balance and reduced transfusion needs. ³⁸ In adult cardiac surgery, a hematocrit threshold ≤14% during CPB was independently associated with increased mortality, and high-risk patients had a higher mortality risk when hematocrit dropped below 17%, underscoring the critical threshold values even beyond the pediatric population. ³⁹

Hematocrit also modifies the risk relationship between perfusate reoxygenation and AKI, particularly in pediatric patients with Tetralogy of Fallot. Among patients with elevated preoperative hematocrit, excessive oxygenation during CPB rewarming was associated with significantly higher AKI risk, suggesting the need for hematocrit-adjusted perfusion strategies. ⁴⁰

Efforts to preserve hematocrit during CPB have included patient-specific strategies such as autologous priming. In one report, replacing the standard crystalloid prime with the patient’s own blood reduced hematocrit drop by approximately 4%, allowing CPB without additional transfusion or ultrafiltration. ⁴¹ Furthermore, pediatric-specific blood conservation guidelines have been published, advocating for strategic perfusion management and minimizing blood loss, although exact hematocrit targets continue to vary. ⁴²

Near-infrared spectroscopy

NIRS is a noninvasive modality increasingly used in pediatric cardiac surgery to monitor regional tissue oxygenation and guide perfusion during CPB. ⁴³ Beyond basic monitoring, cerebral NIRS—particularly the cytochrome aa3 (CytOx) signal—has shown promise as a predictor of neurological injury. Experimental studies link CytOx reductions with decreased cerebral energy metabolites and histologic brain damage, while clinical findings associate lower intraoperative CytOx levels with neuropsychological deficits. ⁴⁴

During CPB, cerebral NIRS values are typically lower than somatic sites, such as renal or peripheral tissues, due to regional differences in oxygen consumption and autoregulatory capacity, especially in neonates and infants. One study in 25 infants showed cerebral NIRS averaged 65–75%, while somatic sites often exceeded 80–90% during CPB. ⁴⁴ These disparities emphasize the importance of brain-specific monitoring to detect subtle perfusion vulnerabilities.

NIRS has also demonstrated sensitivity to cerebral oxygenation changes that are not always reflected in standard hemodynamic metrics. For example, cerebral rSO₂ levels may increase during cooling despite stable MAP, hematocrit, or pCO₂. ⁴⁵ In deep hypothermic circulatory arrest (DHCA) models, NIRS identified significant deoxygenation through decreased oxyhemoglobin and increased deoxyhemoglobin, with partial recovery upon reperfusion, closely paralleling cerebral blood flow changes. ⁴⁶ NIRS is especially useful in distinguishing between cyanotic and non-cyanotic congenital heart disease oxygenation patterns. While both show a drop in cerebral-oxygenated hemoglobin at CPB initiation, cyanotic patients exhibit a more pronounced and sustained decline in deoxygenated and total hemoglobin (p < 0.001, p < 0.01), ⁴⁷ suggesting increased cerebral vulnerability and the need for individualized perfusion strategies.

In addition to cerebral monitoring, renal NIRS serves as a valuable tool for the early detection of AKI. In a study of 59 infants, 48% developed AKI, and these patients had significantly lower renal NIRS values intraoperatively and up to 48 h postoperatively (p < 0.05). ⁴⁸ Among AKI patients, 11% required renal replacement, and 7% died, while no deaths occurred in the non-AKI group. Renal NIRS may detect perfusion deficits earlier than conventional biomarkers, aiding in perioperative renal protection.

Despite its growing use, interpreting NIRS values remains challenging. Factors such as probe placement, hemoglobin levels, and inflammation can confound readings. ⁴⁹ Near-infrared spectroscopy (NIRS) for somatic monitoring is most reliable in small children and becomes less dependable as body size and tissue depth increase. In a cohort of children with heart disease, renal NIRS correlated closely with renal venous saturations in those ≤10 kg (r ˜ 0.82) but showed no meaningful correlation in patients >10 kg, implicating increased subcutaneous/perirenal tissue thickness as a key limitation. ⁵⁰ Contemporary neonatal guidance concurs that somatic NIRS can validly trend capillary-venous oxygenation in superficial tissues but is constrained by optode spacing, penetration depth, device-to-device variability, and site-specific artifacts—factors that are particularly favorable in neonates/infants and progressively less so with growth. ⁵¹ As a result, experts recommend using somatic NIRS primarily as a trend monitor, interpreted alongside systemic perfusion indices (e.g., DO₂, lactate, ScvO₂) and, when feasible, combining cerebral and somatic sites to improve detection of low-flow states. ⁵² In older or higher-BMI children, practical mitigations include optimizing probe placement (posterolateral flank), securing consistent contact, and prioritizing multimodal perfusion monitoring rather than relying on a single NIRS value. Experts, therefore, recommend incorporating NIRS into a multimodal brain monitoring strategy—using it alongside tools like transcranial Doppler and EEG. A low rSO₂ should be seen as a warning signal rather than a definitive marker of hypoxia and must be evaluated within the broader clinical context. ⁵³

Temperature management

Temperature management during pediatric CPB is critical to optimizing metabolic control and protecting vital organ function. Traditionally, moderate hypothermia (28°C) has been widely accepted to reduce oxygen consumption and provide cerebral and myocardial protection. However, recent findings suggest that mild hypothermia (32°C–34°C) or even normothermic CPB (36°C–37°C) may be equally effective in many cases without the risks associated with deeper hypothermia, such as coagulation abnormalities or prolonged recovery times.^54,55

A randomized study comparing mild hypothermia (32°C) and moderate hypothermia (28°C) in infants undergoing ventricular septal defect repair found no significant difference in clinical outcomes—including blood loss and coagulation profile—between the two groups, though the 32°C group had slightly prolonged activated partial thromboplastin time (aPTT). ⁵⁴ Another randomized study comparing 24°C and 34°C bypass temperatures found no statistically significant differences in mechanical ventilation duration (median 22 vs 14 h, p = .14), ICU stay (43 vs 29 h, p = .79), or inflammatory markers, supporting the notion that milder hypothermia may be adequate for many pediatric patients. ⁵⁵ A 2021 meta-analysis of 13 studies involving 837 pediatric patients showed that normothermic CPB (NCPB) was associated with significantly better outcomes in simple congenital heart surgery. These included reduced risk of postoperative bleeding (OR 0.11, 95% CI: 0.01–0.89), lower lactate levels 2–4 h post-CPB (MD –0.60 mmol/L), lower serum creatinine at 24 and 48 h (MD –2.73 and −2.08 μmol/L, respectively), and shorter CPB time (MD –19.10 min), as well as fewer major adverse events (OR 0.37, 95% CI: 0.15–0.93). ⁵⁶

Despite these findings, deeper hypothermia (18°C–22°C) is still warranted in cases requiring total circulatory arrest, as it offers superior neuroprotection. Cerebral blood flow (CBF) was shown to decrease under hypothermic conditions in all temperature groups (18°C–32°C), but recovered during rewarming in moderate hypothermia groups. In contrast, CBF did not return to baseline in patients undergoing total circulatory arrest at 18°C, suggesting impaired cerebral recovery. ¹⁵

Temperature measurement accuracy remains crucial. Surveys of clinical perfusion practices across Great Britain and Ireland revealed significant variability in monitoring techniques. Nasopharyngeal temperature was most commonly used (52%), but not all centers adhered to best practices such as maintaining rewarming gradients <10°C or avoiding temperatures >37°C. ⁵⁷ Guidelines developed by major thoracic and perfusion societies recommend using oxygenator arterial outlet blood temperature as the surrogate for cerebral perfusate temperature during CPB and warn against exceeding a 10°C gradient during cooling or rewarming phases to avoid cerebral injury from microemboli or gas outgassing. Moreover, arterial outlet blood temperature should not exceed 37°C to prevent hyperthermia-related cerebral damage. ⁵⁸

Innovative strategies for rewarming are also being explored. In a randomized trial of infants undergoing hypothermic CPB, use of the ALLON 2001 thermoregulation system (a warming garment) resulted in significantly higher nasopharyngeal temperatures 20 and 40 min after separation from CPB compared to routine thermal care (36.5°C vs 35.0°C at 20 min, p = 0.004; 36.98°C vs 35.30°C at 40 min, p = 0.034), without any adverse effects, supporting its efficacy in preventing postoperative after-drop. ⁵⁹

Priming volume

Minimizing cardiopulmonary bypass (CPB) priming volume is a key aspect of pediatric perfusion strategy because prime size directly affects the level of hemodilution, exposure to donor blood, and downstream inflammatory and organ-specific complications. Indexed prime volume (mL/m²) is a particularly useful metric in children of varying sizes and is associated with clinically relevant outcomes throughout the perioperative period.

Large observational cohorts demonstrate that higher indexed prime volumes independently predict perioperative red blood cell transfusion after pediatric CPB—even after adjustment for surgical complexity (RACHS-1), core temperature, preoperative hemoglobin, and surgeon. ⁶⁰ Beyond transfusion exposure, greater prime volumes have been associated with higher postoperative inflammatory markers, worse renal parameters, and longer durations of mechanical ventilation and ICU stay, underscoring the systemic impact of hemodilution and circuit burden. ⁶¹

Pragmatic approaches to prime minimization span circuit design, conduct of perfusion, and blood conservation. Miniaturized, low-surface-area circuits with shortened tubing and vacuum-assisted venous drainage can safely reduce absolute prime (e.g., from ∼300 mL to ∼190 mL) while maintaining adequate venous return and cardiotomy suction. ⁶² Asanguineous (bloodless) priming with miniaturized CPB circuits is feasible and clinically effective. Avoidance of transfusion on CPB was achieved in ∼47% of patients <3 kg and ∼86% of those >8 kg, though many heavier patients still required transfusion later during hospitalization. ⁶³ These findings highlight the strong interaction between body size, priming volume, and transfusion risk. In very low–body-weight neonates, “mini-volume” priming strategies have been associated with fewer transfusions and reduced need for postoperative ECMO support, suggesting potential benefits that extend beyond the operating room. ⁶⁴

Prime composition also matters. Zero-balance ultrafiltration (Z-BUF) of priming blood helps normalize electrolytes and acid–base status before initiation of bypass and has been linked to shorter ventilation times, reduced ICU length of stay, and fewer electrolyte-related complications. ⁶⁵ Taken together, these data support a pragmatic, multi-component program: (1) standardize use of pediatric-specific oxygenators and shortened tubing, (2) employ vacuum-assisted venous drainage judiciously to permit lower venous reservoir volumes without compromising safety, (3) preferentially use asanguineous prime when physiologically tolerable, (4) apply Z-BUF prior to initiation and modified ultrafiltration after weaning, and (5) track indexed prime volume as a key quality indicator with routine feedback to the team.

Type of cardioplegia

Choice of cardioplegia formulation and delivery strategy directly influences myocardial metabolic recovery, reperfusion stability, and whole-body perfusion physiology during pediatric CPB. Contemporary evidence in children shows a consistent metabolic signal favoring blood-based solutions, whereas most hard clinical endpoints remain similar across solutions when modern techniques and standardized protocols are applied. In a meta-analysis focused on pediatric surgery, blood cardioplegia was associated with lower postoperative lactate at early (0–6 h) and later (∼48 h) time points, without differences in aortic cross-clamp or CPB times, troponin/CK-MB release, ventilation duration, ICU length of stay, or mortality. ⁶⁶ A second pediatric meta-analysis reached largely concordant conclusions, reporting improved metabolic indices with blood cardioplegia but no consistent differences in major outcomes. ⁶⁷ Complementing these pairwise syntheses, a pediatric network meta-analysis ranked warm blood cardioplegia as achieving the lowest 24-h troponin-I, while HTK/Custodiol tended to yield higher troponin-I; nonetheless, between-strategy differences in clinical outcomes were not consistent across studies. ⁶⁸ Methodologically, the pediatric evidence base still contains many small, single-center randomized trials with surrogate endpoints and important heterogeneity in dosing, temperature, and redosing intervals, as highlighted by a systematic review of randomized pediatric studies. ⁶⁹

Solution-specific dosing schemas and redosing intervals also carry system-level implications for hemodilution and perfusion. Conventional cold blood cardioplegia is typically delivered as an induction dose with maintenance every 15–20 min at low temperatures; intermittent warm blood dosing is another common variant. In contrast, del Nido (1:4 blood:crystalloid, lidocaine-containing) is designed as a single-dose strategy (often ∼20 mL/kg) that provides 60–90 min of electrical arrest at moderate hypothermia, thereby reducing the number of interruptions and total cardioplegia volume administered. Across pediatric series and controlled experiences, del Nido generally results in fewer doses, lower total cardioplegia volume, quicker return of spontaneous rhythm, and lower vasoactive-inotropic requirements versus conventional blood cardioplegia, while cross-clamp/CPB times and major postoperative outcomes are typically comparable when care pathways are standardized.^70–73 For example, in neonates and infants, retrospective and comparative series reported similar early outcomes between del Nido and conventional strategies despite higher lesion complexity in some del Nido cohorts, with significantly fewer redoses and lower administered cardioplegia volume in the del Nido arm.^71,72 These observations align with broader meta-analytic comparisons of del Nido versus St Thomas solutions across adult and pediatric populations, which commonly note lower cardioplegia volume and fewer redoses with del Nido and small reductions in cross-clamp or CPB times, without consistent differences in major adverse events. ⁷³