Veno-arterial CO2 content gradient and veno-arterial CO2 to arterial-venous O2 content ratio for outcome prediction after pediatric cardiac surgery: a prospective study Authors: Vladimir L Cousin, Raphael Joye, Tomasz Nalecz, Tornike Sologashvili, Maurice Beghetti, Cyril Jaksic, Julie Wacker, Angelo Polito Published: 2025 ABSTRACT AbstractIntroductionCO2-derived variables, veno-arterial CO2 content gradient (ΔCCO2) and the ratio of ΔCCO2 with arterio-venous oxygen difference (AV-DO2) (ΔCCO2/AV-DO2), may have a potential role as indicators of low cardiac output and anaerobic metabolism, respectively. We sought to describe and evaluate the association of CO2-derived variables with patients’ outcomes in the post cardiopulmonary bypass (CPB) period in children.MethodsProspective, single-center, study enrolling children post-CPB with paired arterial and venous blood gases for determination of lactate, O2 extraction, ΔCCO2, and ΔCCO2/AV-DO2 at admission (H0), and at 6 (H6), 12 (H12) and 24 (H24) hours. Different clinical patterns were defined based on the presence of an anaerobic context or a hypoperfusion context, using both O2 and CO2-derived variables. The presence of anaerobic metabolism was defined with a lactate > 2 mmol/l and ΔCCO2/AV-DO2 > 1.8; the presence of hypoperfusion was defined with an O2 extraction > 30% and ΔCCO2 > 6 mL. The potential association of duration of amine support and mechanical ventilation was tested with CO2-derived variables and specific clinical patterns.ResultsA total of 51 patients with a median age of 36 (IQR 11–85) months were included. Median admission ΔCCO2 was 9.3 mL (IQR 5.6–11.4) with 72% above 6 mL. Median ΔCCO2/AV-DO2 was 2.1 (IQR 1.5–2.4) with 58% above 1.8. Admission ΔCCO2 showed a significant association with the duration of mechanical ventilation (R2 21.6, p value = 0.001) but not with the duration of vasoactive support. Neither H0 ΔCCO2 nor H0 ΔCCO2/AV-DO2 improved outcome prediction by a model including lactate and O2 extraction. Anaerobic metabolism context showed a significant association with prolonged vasoactive support [28.4 (CI 95% 12.2–44.6) p = 0.001] and mechanical ventilation duration [1.4 (95% CI 0.62–2.3) p = 0.003]. In hypoperfusion context, neither duration of vasoactive support nor mechanical ventilation appeared different in the subgroups analysis.ConclusionCO2-derived variables may improve outcome prediction after cardiac surgery in pediatric patients. Further evaluation in larger multicentered trials is necessary to improve its validation.Supplementary InformationThe online version contains supplementary material available at 10.1186/s40635-025-00834-9.Keywords: Cardiopulmonary bypass, Congenital heart disease, Postoperative care, Veno-arterial CO2 content gradient, Venous-arterial CO2 to arterial-venous O2 content difference, Oxygen extraction, Lactate Introduction Evaluation of oxygen (O2) metabolism is frequently used in the pediatric intensive care unit (PICU) to assess patients’ hemodynamic status [1–3]. Lactate level and O2 extraction are the main variables used in this setting [4]. However, both may experience limitations in clinical settings. Falsely elevated lactate can be caused by hyperglycemia or aminergic support with stimulation of beta-2 receptors while reduced clearance can be experienced in cases of renal or liver failure [5]. O2 extraction is affected by multiple variables, including cardiac output, arterial O2 saturation, hemoglobin level and O2 consumption [3]. Moreover, cases of falsely normal extraction have been reported, especially in the presence of sepsis [5, 6]. Carbon dioxide (CO2) metabolism has been proposed as an additional tool to assess hemodynamic status [6–8]. The main variables described are the veno-arterial CO2 content gradient (ΔCCO2) and the ratio of ΔCCO2 with arterio-venous oxygen difference (AV-DO2) (ΔCCO2/AV-DO2). Derived from the Fick equation, ΔCCO2 reflects the adequacy of the balance between the cardiac output and the CO2 production (VCO2), according to the formula ΔCCO2 = VCO2/cardiac output [6]. An increase in ΔCCO2 is then directly related to a decrease in cardiac output if the VCO2 is stable [9, 10]. Even in the case of anaerobic metabolism, a production of CO2 occurs through buffering of bicarbonate [HCO3−] by acidic cations [H+], maintaining a certain level of VCO2, while the cardiac output is severely reduced, leading to an accumulation of CO2 in venous blood. Elevation of ΔCCO2 above 6 mL is related to worse outcomes in septic shock and other ICU admissions [4, 11]. The ratio of ΔCCO2/AV-DO2 has been less studied and acts as a surrogate for the respiratory quotient. It evaluates the balance between CO2 production and O2 consumption, both determined according to the Fick equation: the ΔCCO2 estimates VCO2 and the AV-DO2 estimates VO2. Under physiological conditions, VCO2 should not exceed VO2, and, therefore, a ratio > 1 should be encountered only in anaerobic situations when the anaerobic VCO2 exceeds the decreased VO2 [12]. In septic shock, elevated ΔCCO2/AV-DO2 with values ranging between 1.2 and 1.6 have been related to worse outcomes, even in cases of normal lactate levels [13, 14] Studies evaluating both variables and their use in clinical practice remain rare in the pediatric population, particularly for ΔCCO2/AV-DO2. Despite differences in study populations and outcomes definitions, ΔCCO2 has been associated with different outcomes after pediatric cardiac surgery, ranging from increased mortality, prolonged mechanical ventilation or worse composite outcomes including ECMO requirement or high vasopressor requirement [15–17]. Of particular significance, CO2-derived variables are not subject to the limitations associated with lactate and O2-extraction measurements [18]. Furthermore, integrating data from both O2 and CO2 variables could prove valuable in identifying distinct clinical patterns, according to the presence of anaerobic metabolism or a state of hypoperfusion, with potentially different trajectories in the PICU. This approach mirrors findings in septic shock patients, where these indices have been used to detect low cardiac output and inadequate oxygen delivery situations, with significantly different outcomes even in the presence of similar O2-derived variables [14, 19–22]. We hypothesized that the use of CO2-derived variables could improve hemodynamic assessment after surgery for congenital heart disease (CHD). Our study sought to evaluate the capacity of ΔCCO2 and ΔCCO2/AV-DO2 to predict outcomes after cardiac surgery in pediatric patients. Methods Study design and population This prospective study was conducted between November 2022 and December 2023 at the PICU of Geneva University Hospitals. The study, untitled “Cardiac output monitoring using CO2 metabolism in children after cardiac surgery” was registered and approved by the local ethic committee (Commission Cantonale d’Ethique et de la Recherche de l’Etat de Genève) on the 27th of September 2022, project number CCER 2022-00488. This study adheres to national regulations on medical research and international guidelines of the Helsinki Declaration. Informed consent was obtained from each subject or the subject’s legally authorized representative. The study population consisted of children (0–18 years) admitted after cardiopulmonary bypass (CPB) for CHD surgery. Patients were excluded if they did not have both a central venous catheter and an arterial catheter placed before the surgery, if they required extracorporeal membrane oxygenation or if the chest was left open upon leaving the operating room (N = 1). The primary outcome was the duration of vasoactive support. The secondary outcome was the duration of mechanical ventilation. We also sought to evaluate whether CO2-derived variables improve outcome prediction beyond O2-derived variables (lactate and O2 extraction) and, when integrated with them, enable refined definitions of hemodynamic clinical patterns. Data collection Demographic and clinical variables collected were: age, weight, cardiac defect, surgery risk evaluation using the Risk Adjustment for Congenital Heart Surgery (RACHS) score [23], CPB and aortic cross-clamp time. Data were prospectively collected at the following 4 time points: at PICU admission (H0) following cardiac surgery and after 6 h (H6), 12 h (H12) and 24 h (H24). At every time point, vasoactive support was reported using the Vaso-Inotropic Score (VIS) [24], and paired arterial and central venous blood gases were collected with the following data: pH,O2 saturation, partial pressure in O2, partial pressure in CO2, hemoglobin level (g/dL), temperature, inspired O2 fraction, lactate (mmol/L). O2- and CO2-derived variables were then calculated. Blood gases were performed using a Radiometer ABL-800 (Radiometer, Copenhagen, Denmark). O2-derived variables were calculated as follows: O2 extraction = arterial O2 saturation − venous O2 saturation. Other O2-derived perfusion markers were directly available on the blood gases: arterial lactate (mmol/L) and superior vena cava oxygen saturation [SVO2 (%)]. CO2-derived variables were calculated according to Douglas’ formula [25] (detailed in supplementary materials): ΔCCO2/AV-DO2 calculation: Calculation for oxygen content in arterial (CaO2) or venous (CvO2) blood were determined as follows: Definition of anaerobic metabolism and hypoperfusion subgroups Two distinct hemodynamic clinical patterns were defined based on O2− and CO₂- derived variables, as incorporating CO₂ parameters alongside lactate and O2-extraction may refine outcome predictions in the PICU. To assess anaerobic metabolism, we categorized patients using a combination of lactate levels and the ΔCCO₂/AV-DO₂. Since lactate reflects tissue hypoxia and impaired oxidative metabolism, and ΔCCO₂/AV-DO₂ serves as an index of respiratory quotient changes in low-flow states, their integration might enhance the identification of anaerobic metabolism. Four subgroups were defined: Group 1: lactate ≤ 2 mmol/L and ΔCCO₂/AV-DO₂ ≤ 1.8 (normal metabolism) Group 1: lactate ≤ 2 mmol/L and ΔCCO₂/AV-DO₂ ≤ 1.8 (normal metabolism) Group 2: lactate > 2 mmol/L and ΔCCO₂/AV-DO₂ ≤ 1.8 (isolated hyperlactatemia without anaerobic CO2 production) Group 2: lactate > 2 mmol/L and ΔCCO₂/AV-DO₂ ≤ 1.8 (isolated hyperlactatemia without anaerobic CO2 production) Group 3: lactate ≤ 2 mmol/L and ΔCCO₂/AV-DO₂ > 1.8 (CO₂ anaerobic production despite normal lactate) Group 3: lactate ≤ 2 mmol/L and ΔCCO₂/AV-DO₂ > 1.8 (CO₂ anaerobic production despite normal lactate) Group 4: lactate > 2 mmol/L and ΔCCO₂/AV-DO₂ > 1.8 (both metabolic acidosis and anaerobic CO2 production, suggesting potential severe anaerobic metabolism) Group 4: lactate > 2 mmol/L and ΔCCO₂/AV-DO₂ > 1.8 (both metabolic acidosis and anaerobic CO2 production, suggesting potential severe anaerobic metabolism) To evaluate hypoperfusion, we combined O2 extraction with ΔCCO₂, given that O2 extraction reflects the balance between O2 delivery and consumption, while ΔCCO₂ increases in the presence of stagnant or low-flow states. Group 1: O2 extraction ≤ 30% and ΔCCO₂ ≤ 6 mL (normal perfusion) Group 1: O2 extraction ≤ 30% and ΔCCO₂ ≤ 6 mL (normal perfusion) Group 2: O2 extraction > 30% and ΔCCO₂ ≤ 6 mL (high oxygen demand with preserved CO₂ clearance) Group 2: O2 extraction > 30% and ΔCCO₂ ≤ 6 mL (high oxygen demand with preserved CO₂ clearance) Group 3: O2 extraction ≤ 30% and ΔCCO₂ > 6 mL (low oxygen extraction with CO₂ stagnation, suggesting circulatory inefficiency) Group 3: O2 extraction ≤ 30% and ΔCCO₂ > 6 mL (low oxygen extraction with CO₂ stagnation, suggesting circulatory inefficiency) Group 4: O2 extraction > 30% and ΔCCO₂ > 6 mL (both high oxygen demand and impaired CO₂ clearance, indicating significant hypoperfusion). Group 4: O2 extraction > 30% and ΔCCO₂ > 6 mL (both high oxygen demand and impaired CO₂ clearance, indicating significant hypoperfusion). Statistical analysis Categorical variables were expressed as proportions (%) and continuous variables as median and interquartile range (IQR) due to the non-normal distribution of variables. Elevated ΔCCO2 and ΔCCO2/AV-DO2 were defined according to published cutoffs: >6 mL for ΔCCO2 and >1.8 for ΔCCO2/AV-DO2 [7, 12, 26, 27]. Associations between ΔCCO2, ΔCCO2/AV-DO2; anaerobic or hypoperfusion subgroups at admission and clinical PICU outcomes were evaluated using linear regression. In order to assess the extent to which each of the different parameters (ΔCCO2 gradient, ΔCCO2/AV-DO2 or a ratio of ΔCCO2 or ΔCCO2/AV-DO2 between H0 and H6 or H24) improves the explained proportion of variance of different outcomes on top of what both lactate and O2-extraction already explain, multiple linear regressions were conducted. For each parameter, the R2 of both models with and without the parameter of interest was compared. A small increase in R2 was interpreted as a poor improvement in explaining the outcome. All models included both changes in lactate and O2-extraction as predictors and the outcomes were the duration of vasopressor support and the duration of mechanical ventilation. Standard errors were estimated using the Huber–White (HC3) robust method. P value < 0.05 was considered statistically significant and statistical analysis was performed using R version 4.2.2 and Stat software v.18 (STATA, College Station, Texas USA). Results Patients A total of 51 patients were included. Patient characteristics and clinical outcomes are summarized in Table 1. The median duration of PICU stay was 2 days (IQR 2–3). No patient required emergent chest reopening or ECMO and there were no in-hospital deaths during the study period. Table 1Characteristics of patients at baselineVariableMedian (IQR) or N (%)Age (months)36 (11–85)Weight (kg)13.7 (9.9–28)RACHS categories 15 (9.8) 212 (23.5) 320 (39.2) 46 (11.8) 61 (1.9)Not classified7 (13.8)Duration of CPB (min)70 (52–139)Duration of aortic cross-clamp (min)48 (32–90)Admission lactate (mmol/L)1.6 (1–2.1)Admission O2 extraction (%)29 (25–33)Admission VIS11 (5–22.5)Aminergic support (hours)17 (5–24.5)Mechanical ventilation (days)0 (0–1)PICU stay duration (days)2 (2–3)CBP cardiopulmonary bypass, IQR interquartile range, PICU pediatric intensive care unit, RACHS risk adjustment for congenital heart surgery Characteristics of patients at baseline CBP cardiopulmonary bypass, IQR interquartile range, PICU pediatric intensive care unit, RACHS risk adjustment for congenital heart surgery CO2 and O2-derived variables At PICU admission, median ΔCCO2 was 9.3 mL (IQR 5.6–11.4) and median ΔCCO2/AV-DO2 was 2.1 (IQR 1.5–2.4). The evolution of CO2 and O2-derived variables over the first 24 h is depicted in Fig. 1. A gradual increase in ΔCCO2 and ΔCCO2/AV-DO2 was observed during this period, whereas lactate levels and O2 extraction tended to decrease after H12. At admission, 36 out of 50 patients (72%) had a ΔCCO2 > 6 mL and 29 out of 50 patients (58%) had a ΔCCO2/AV-DO2 > 1.8. The proportion of patients with elevated CO2-derived variables over the first 24 h after admission is detailed in Supplementary Fig. 1. Evolution of O2 and CO2 variables over the first day of PICU stay post cardiopulmonary bypass. Evolution of O2-derived variables [lactate (mmol/L) and O2 extraction (%)] as well as CO2-derived variables (ΔCCO2 (mL) and ΔCCO2/AV-DO2) in the first 24 h of PICU after the cardiac surgery. Time points are as follows: admission H0, then at 6 h (H6), 12 h (H12) and 24 h (H24) of admission Admission ΔCCO2, ΔCCO2/AV-DO2 and association with clinical outcomes No significant association was identified between ΔCCO2 and the primary outcome. Univariate analysis showed a significant association between admission ΔCCO2 and duration of mechanical ventilation (R2 21.6, p value = 0.001) (Table 2). None of the outcomes were significantly associated with ΔCCO2/AV-DO2 at PICU admission (Table 2). Elevated ΔCCO2 (>6 mL) and ΔCCO2/AV-DO2 (>1.8) failed to significantly predict patients’ outcomes; only VA-CCO2/AV-O2 ratio > 1.8 approached significance for duration of mechanical ventilation (Supplementary Table 1). Table 2CO2-derived variables and their association with clinical outcomesVariableCoefficient (95% CI)R2P valueΔCCO2 H0 Vasoactive support (h)0.93 (−0.64–2.5)2.80.24 Mechanical ventilation (d)0.14 (0.06–0.21)21.60.001ΔCCO2/AV-DO2 H0 Vasoactive support (h)2.87 (−6.1–11.8)0.870.52 Mechanical ventilation (d)0.43 (−0.36–0.89)6.60.07Anaerobic subgroup Vasoactive support (h)*6.3 (0.96–11.6)10.30.02 Post hoc subgroup I vs IV28.4 (12.2–44.6)25.10.001 Mechanical ventilation (d)*0.38 (0.09–0.66)12.90.01 Post hoc subgroup I vs IV1.4 (0.62–2.3)18.10.003Hypoperfusion subgroup Vasoactive support (h)*#1.01 (−4.6–6.6)0.20.72 Mechanical ventilation (d)*#0.31 (0.02–0.6)8.90.03Association of CO2-derived variables at admission (H0) with clinical outcomes of interest and association of hemodynamic subgroups with clinical outcomes of interest*Subgroup 1 is defined as reference#None of the post hoc subgroups comparison was significantly different CO2-derived variables and their association with clinical outcomes Association of CO2-derived variables at admission (H0) with clinical outcomes of interest and association of hemodynamic subgroups with clinical outcomes of interest *Subgroup 1 is defined as reference #None of the post hoc subgroups comparison was significantly different Neither the duration of vasoactive support or mechanical ventilation prediction was significantly improved by the addition of CO2-derived variables to a model compelling lactate and O2 extraction (Supplementary Table 2). Similarly, ΔCCO2/AV-DO2 did not significantly improve the prediction model when including lactate and the O2 extraction for outcome prediction (Supplementary Table 2). In a similar manner, the dynamic evolution of ΔCCO2 and ΔCCO2/AV-DO2 at H6 or H24 was also tested with a model including lactate and O2 extraction. Neither the evolution at H6 or H24 of ΔCCO2 or ΔCCO2/AV-DO2 significantly improved outcome prediction, as depicted in Supplementary Table 2. Impact of admission hypoperfusion or anaerobic metabolism patterns on outcome Patients were categorized into anaerobic patterns as follows: Subgroup I included 17/51 (33.3%), Subgroup II 4/51 (7.8%), Subgroup III 20/51 (39.2%), and Subgroup IV 10/51 (19.7%). In these anaerobic subgroups, the VIS at admission was 10 (IQR 5–14) for Subgroup I, 19.75 (IQR 13–29) for Subgroup II, 10.5 (IQR 3.75–16.5) for Subgroup III, and 20 (IQR 6–27.5) for Subgroup IV. Clinical outcomes—duration of vasoactive support and mechanical ventilation—differed significantly across the anaerobic subgroups. Post hoc analysis revealed significant differences between Subgroups I and IV for all outcomes (Fig. 2A, B). Clinical outcome according to hypoperfusion and anaerobic clinical patterns. Anaerobic subgroups comparison for duration of amine support (A) and mechanical ventilation (B). Group I, lactate ≤ 2 mmol/L and ΔCCO2/AV-DO2 ≤ 1.8; group II, lactate > 2 mmol/L and ΔCCO2/AV-DO2 ≤ 1.8; group III, lactate ≤ 2 mmol/L and ΔCCO2/AV-DO2 > 1.8; and group IV lactate > 2 mmol/L and ΔCCO2/AV-DO2 > 1.8. *p < 0.05 for subgroup 1 vs. 4. Hypoperfusion subgroups comparison for duration of amine support (C), mechanical ventilation support (D). Group I, extraction ≤ 30% and ΔCCO2 ≤ 6 mL; group II, extraction > 30% and ΔCCO2 ≤ 6 mL; group III, extraction ≤ 30% and ΔCCO2 > 6 mL; group IV, extraction > 30% and ΔCCO2 gradient > 6 mL For hypoperfusion pattern, the distribution was as follows: Subgroup I 13/51 (25.5%), Subgroup II 1/51 (2%), Subgroup III 21/51 (41.2%), and Subgroup IV 16/51 (31.3%). VIS at admission in these hypoperfusion subgroups was 10 (IQR 5–22) for Subgroup I, 17 (IQR 17–17) for Subgroup II, 10 (IQR 5–22) for Subgroup III, and 14 (IQR 7.25–28.25) for Subgroup IV. In contrast to the anaerobic subgroups, no significant differences were observed among the hypoperfusion subgroups for the duration of amine support or mechanical ventilation (Fig. 2C, D). Discussion This prospective study evaluated the utility of CO2-derived variables, specifically ΔCCO2 and ΔCCO2/AV-DO2, as predictors of clinical outcomes in pediatric patients following CPB for CHD surgery. This study provides novel insights into the potential role of CO2-derived variables in post-CPB hemodynamic assessment and their integration with O2-derived markers to define distinct hemodynamic clinical patterns. Our findings indicate that ΔCCO2 and ΔCCO2/AV-DO2 were persistently elevated above established adult thresholds (ΔCCO2 > 6 mL and ΔCCO2/AV-DO2 > 1.8) during the first 24 h post-CPB, even in patients with uneventful recoveries. At PICU admission, 72% of patients exhibited ΔCCO2 > 6 mL, and 58% had ΔCCO2/AV-DO2 > 1.8, suggesting that these cutoffs, derived primarily from adult and septic shock populations, may not be applicable to pediatric post-CPB settings [4, 7, 8]. Elevated CO2-derived variables have been reported in prior pediatric studies post-CPB, with ΔPCO2 (a proxy for ΔCCO2) frequently exceeding 6 mmHg [15–17, 28–30]. For instance, Akamatsu et al. noted that 90% of patients had ΔPCO2 > 6 mmHg despite low complication rates [28]. Similarly, Taina et al. reported ΔCCO2/AV-DO2 values between 1.8 and 2 in the post-CPB period [30]. These consistent findings across studies raise questions about the appropriateness of adult-derived CO2 thresholds in children, particularly in the context of CPB. Several factors may contribute to elevated CO2-derived variables post-CPB. The post-CPB period is characterized by increased CO2 production (VCO2) due to repayment of oxygen debt incurred during surgery and active metabolic recovery [31]. Paediatric patients may also exhibit higher baseline VCO2 compared to adults, potentially necessitating age-specific thresholds [32, 33]. Additionally, microcirculatory dysfunction, a known feature of the post-CPB state, may impair CO2 clearance, leading to elevated venous CO2 levels despite adequate cardiac output [34, 35]. This is supported by the hypothesis that CO2 diffusion from poorly perfused tissues increases ΔCCO2, even in the absence of overt hypoperfusion [27, 34]. These physiological nuances underscore the need to establish pediatric-specific normative values for CO2-derived variables post-CPB to enhance their clinical utility. Regarding outcome prediction, univariate analysis revealed a significant association between admission ΔCCO2 and prolonged mechanical ventilation, consistent with findings by Gong et al. [15]. However, no significant association was observed between ΔCCO2 or ΔCCO2/AV-DO2 and the duration of vasoactive support, nor did ΔCCO2/AV-DO2 independently predict the primary outcome. Moreover, the addition of CO2-derived variables to models incorporating lactate and oxygen extraction did not significantly enhance outcome prediction. Discrepancies in the predictive value of ΔCCO2 have been noted in prior studies. Rhodes et al. and Insom et al. reported associations between elevated ΔCCO2 and adverse outcomes, including mortality, particularly in neonates and patients with univentricular physiology [16, 17]. In contrast, Akamatsu et al. and Taina et al. found no such associations, potentially due to differences in patient populations or outcome definitions [28, 30]. These inconsistencies highlight the need for standardized approaches to population selection and cutoff determination in future studies. A key contribution of this study is the exploration of hemodynamic clinical patterns defined by combining CO2- and O2-derived variables. Patients categorized into anaerobic metabolism subgroups based on lactate and ΔCCO2/AV-DO2 demonstrated significant differences in clinical outcomes. Notably, Subgroup IV (high lactate and high ΔCCO2/AV-DO2, indicative of severe anaerobic metabolism) exhibited prolonged vasoactive support and mechanical ventilation compared to Subgroup I (normal metabolism). This suggests that integrating ΔCCO2/AV-DO2 with lactate may improve the identification of patients at risk for adverse outcomes, even when lactate levels are normal. In contrast, hypoperfusion subgroups, defined by oxygen extraction and ΔCCO2, did not show significant outcome differences, indicating that anaerobic metabolism may be more clinically relevant in this setting. These findings align with adult septic shock studies, where combining ΔCCO2 with lactate or oxygen extraction has refined patient stratification and risk assessment [14, 19, 36]. Castanuela-Sanchez et al. similarly reported that combining ΔCCO2 and lactate predicted multiple organ failure post-CPB, supporting the value of multimodal monitoring [37]. However, incorporation of those items in an algorithm-based resuscitation protocol did not change significantly patients’ outcomes in comparison to standard care, underlining the interest in those variables and the need for further studies to specify the role of combining CO2- and O2-derived variables [38]. The ability of CO2-derived variables to complement oxygen-derived markers is particularly promising given the limitations of lactate and oxygen extraction. Lactate levels can be confounded by hyperglycemia, adrenergic support, or impaired clearance, while oxygen extraction may appear falsely normal in conditions like sepsis [5, 6]. CO2-derived variables, less susceptible to these confounders, may provide a more robust assessment of hemodynamic status when used in conjunction with traditional markers. To our knowledge, this is the first study to propose anaerobic and hypoperfusion clinical patterns using combined O2 and CO2 variables in the pediatric post-CPB setting, offering a framework for future investigations into personalized hemodynamic management. Limitations This study has several limitations. Its single-center design and small sample size limit generalizability, and the absence of severe outcomes (e.g., mortality or emergent chest reopening) restricts the evaluation of CO2-derived variables in critical scenarios. Incomplete blood gas sampling due to catheter removal may have introduced bias, though the prospective design and systematic sampling mitigate this to some extent. The lack of pediatric-specific normative values for ΔCCO2 and ΔCCO2/AV-DO2 complicates interpretation, particularly in the post-CPB context. Moreover, the impact of a potentially residual shunt may need further studies, but it has not been suggested by other groups. Finally, the study did not establish specific cutoffs for CO2-derived variables, limiting immediate clinical applicability. Conclusion This study showed that ΔCCO2 holds potential as a predictor of prolonged mechanical ventilation in pediatric patients post-CPB, while ΔCCO2/AV-DO2 may enhance risk stratification when combined with lactate to define anaerobic metabolism’s. Those hypothesis-generating results underscore the need for further studies investigating multimodal hemodynamic assessment integrating CO2- and O2-derived variables to offer a promising approach to identifying at-risk patients and tailoring PICU management. Larger, multicenter studies are warranted to validate these findings, define appropriate CO2 thresholds, and explore the clinical impact of hemodynamic clinical patterns-guided interventions in pediatric cardiac surgery. Supplementary Information Additional file 1. Acknowledgements We thank Benoît Cousin for his help in the protocol design and review of the mathematical analysis regarding the CO2 variables. Abbreviations Arterio-venous oxygen difference Carbon dioxide Cardiopulmonary bypass Congenital heart disease Interquartile range Pediatric intensive care unit Oxygen Oxygen content in arterial blood Oxygen content in venous Risk adjustment for congenital heart surgery Ratio of ΔCCO2 with arterio-venous oxygen difference (AV-DO2) Veno-arterial CO2 content gradient Vaso-Inotropic Score Author contributions VLC, RJ and AP designed the study; VLC and AP wrote the protocol; VLC and RJ collected the data; CJ conducted the statistical analysis; VLC, RJ, CJ and AP interpreted the analysis; VLC and RJ wrote the original draft; VLC, RJ, TN, TS, MB, CJ, JW, AP contributed to writing, review, and editing the manuscript. All authors had full access to all data, read and approved the final manuscript.Graphical abstract was created in BioRender. Joye, R. (2025) https://BioRender.com/q7cdw8y Funding Open access funding provided by University of Geneva. None. Availability of data and materials The datasets analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The study was registered and approved by the local ethic committee (Commission Cantonale d’Ethique de la Recherche de l’Etat de Geneve), project number CCER 2022-00488. The study adheres to national regulations on medical research and complies with the Helsinki Declaration. According to Swiss Federal Law on research, consent was obtained for each patient from legal guardians. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Additional file 1. Data Availability Statement The datasets analyzed during the current study are available from the corresponding author on reasonable request.