Assessment of the relationship between cerebral and splanchnic oxygen saturations measured by near‐infrared spectroscopy and direct measurements of systemic haemodynamic variables and oxygen transport after the Norwood procedure

Abstract

Objectives

To evaluate the clinical utility of near‐infrared spectroscopic (NIRS) monitoring of cerebral (Sco₂) and splanchnic (Sso₂) oxygen saturations for estimation of systemic oxygen transport after the Norwood procedure.

Methods

Sco₂ and Sso₂ were measured with NIRS cerebral and thoracolumbar probes (in humans). Respiratory mass spectrometry was used to measure systemic oxygen consumption (V̇o₂). Arterial (Sao₂), superior vena caval (Svo2) and pulmonary venous oxygen saturations were measured at 2 to 4 h intervals to derive pulmonary (Qp) and systemic blood flow (Qs), systemic oxygen delivery (Do₂) and oxygen extraction ratio (ERo₂). Mixed linear regression was used to test correlations. A study of 7 pigs after cardiopulmonary bypass (study 1) was followed by a study of 11 children after the Norwood procedure (study 2).

Results

Study 1. Sco₂ moderately correlated with Svo₂, mean arterial pressure, Qs, Do₂ and ERo₂ (slope 0.30, 0.64. 2.30, 0.017 and −32.5, p < 0.0001) but not with Sao₂, arterial oxygen pressure (Pao₂), haemoglobin and V̇o₂. Study 2. Sco₂ correlated well with Svo₂, Sao₂, Pao₂ and mean arterial pressure (slope 0.43, 0.61, 0.99 and 0.52, p < 0.0001) but not with haemoglobin (slope 0.24, p > 0.05). Sco₂ correlated weakly with V̇o₂ (slope −0.07, p  =  0.05) and moderately with Qs, Do₂ and ERo₂ (slope 3.2, 0.03, −33.2, p < 0.0001). Sso₂ showed similar but weaker correlations.

Conclusions

Sco₂ and Sso₂ may reflect the influence of haemodynamic variables and oxygen transport after the Norwood procedure. However, the interpretation of NIRS data, in terms of both absolute values and trends, is difficult to rely on clinically.

Adequate systemic oxygen delivery (Do₂) balanced with systemic oxygen consumption (V̇o₂) is crucial to the care of any child requiring intensive care but is particularly difficult to assess after the Norwood procedure.¹ Direct measurement of Do₂ and V̇o₂ is most desirable but is rarely performed outside of investigational protocols, and methods often use inappropriate surrogates. Consequently, indirect markers of oxygen balance, such as superior vena caval oxygen saturation (Svo₂), arterial (Sao₂) and venous oxygen saturation difference and blood lactate, are commonly used to estimate the adequacy of Do₂ in these patients.²,³,⁴ The disadvantages of these techniques include the need for repeated blood sampling and the necessarily intermittent nature of the data accrual.

Near‐infrared spectroscopy (NIRS) provides a non‐invasive, continuous method to monitor regional tissue oxygenation.⁵,⁶ This technique depends on the transparency of biological tissue to light in the infrared region of the spectrum of tissue chromophores, such as haemoglobin and cytochrome aa3. Changes in absorption at several wavelengths can be converted into signals of oxyhaemoglobin, deoxyhaemoglobin and oxidised cytochrome aa3. Furthermore, newer generations of NIRS devices permit quantitative measurement of the ratio of oxyhaemoglobin to total haemoglobin, representing the tissue oxygenation index as an absolute term, independent of a tissue path length factor.⁷,⁸

NIRS has been extensively evaluated in the cerebral⁶,⁹,¹⁰,¹¹,¹²,¹³ and splanchnic circulations of newborn infants.¹⁴,¹⁵,¹⁶ Because of its relative ease of use, it is also increasingly used in intensive care units as surrogates of Svo₂ and systemic oxygenation.¹⁷,¹⁸,¹⁹ Good correlations have been generally reported between splanchnic (Sso₂) or cerebral oxygen saturation (Sco₂) and Svo₂ in various patient groups.¹⁷,¹⁸,¹⁹ But there are few data in children with congenital heart disease, either preoperatively or postoperatively. Importantly, all the previous studies have used interindividual single‐point assessments in a relatively large population, and NIRS has not been validated against directly measured systemic haemodynamic variables and oxygen transport. Thus, we performed two studies. Study 1 examined pigs after cardiopulmonary bypass (CPB) with normal circulation. Study 2 was a clinical study of neonates during the 72 h after the Norwood procedure. The Norwood group was chosen because it is a particularly challenging subset of patients in which adequate Do₂ is difficult to assess and would be highly advantageous to assess non‐invasively. Our original hypothesis was that NIRS would accurately reflect systemic oxygen transport when compared with direct measurements. We therefore obtained continuous NIRS measurements of Sco₂ and Sso₂, and continuous measurement of V̇o₂ and, in combination with blood gases, derived repeated and quantitative measurements of pulmonary (Qp) and systemic blood flows (Qs), Do₂ and oxygen extraction ratio (ERo₂). We examined the correlation of Sco₂ and Sso₂ with each of the elements, as well as interindividual and intraindividual variability, to determine the clinical usefulness of NIRS for monitoring systemic haemodynamic function and oxygen transport in patients after the Norwood procedure.

MATERIALS AND METHODS

Study 1

After review and approval by the Institutional Animal Care and Use Committee of the Research Institute in The Hospital for Sick Children, Toronto, Canada, seven Yorkshire pigs weighing 18.5 (1.6) kg were studied. Techniques for anaesthesia, CPB and postoperative management were as described elsewhere.²⁰ Briefly, the animals were studied during general anaesthesia, with inhaled isoflurene (2%) and intravenous infusion of pancuronium (0.8 μg/kg/min). After median sternotomy the pigs underwent a total of 3 h of CPB at 32°C. After rewarming the animal were weaned from CPB, with continuous infusion of dopamine 5–10 μg/kg/min when necessary.

Study 2

Patients

This study was approved by the institutional Research Ethics Board. Written informed consent was obtained from the parents of 11 children (10 boys, aged from 4 to 92 days, median 7 days) undergoing the Norwood procedure between April and October 2004. Table 1 shows the patients' demographics.

Table 1 Clinical data for the 11 patients.

Patient	Age (days)	Weight (kg)	BSA (m2)	CPB (min)	ACC (min)	Circulatory arrest (min)	Cerebral perfusion (min)	Diagnosis
1	7	3.5	0.23	108	47	12	35	HLHS, AS, MS
2	4	3.7	0.25	151	100	35	53	HLHS, AS, MS
3	7	4	0.26	105	47	3	44	HLHS, AS, MS,
4	16	3.5	0.24	133	39	34	0	HLHS, endocardial fibroelastosis of LV, AS, MS
5	7	4.2	0.27	122	62	3	60	HLHS, AS, MS
6	12	3.5	0.23	165	75	13	59	DILV, TGA
7	6	3.5	0.23	172	82	9	70	HLHS, AA, MA
8	92	4.1	0.26	105	58	30	17	DILV, TGA,
9	7	4	0.25	167	64	17	44	HLHS, AS, MS
10	6	2.9	0.2	142	62	1	62	HLHS, AS, MS
11	9	3.6	0.24	109	50	4	44	HLHS, AS, MS

AA, aortic atresia; AS, aortic stenosis; ACC, aortic cross clamp time; BSA, body surface area; CPB, cardiopulmonary bypass; DILV, double inlet left ventricle; HLHS, hypoplastic left heart syndrome; LV, left ventricle; MA, mitral atresia; MS, mitral stenosis; TGA, transposition of great arteries.

Intraoperative procedures

All patients were intubated with cuffed endotracheal tubes (microcuff Heidelberg paediatric; Microcuff GmbH, Weinheim, Germany). General anaesthesia was maintained with inhaled isoflurane, intravenous fentanyl and pancuronium bromide. Low‐flow CPB and selective cerebral perfusion was used in 10 of 11 patients. A standard Norwood procedure with 3.5 mm right modified Blalock–Taussig shunt was used.²¹ Phenoxybenzamine 0.25 mg/kg was given at initiation of CPB. Milrinone (100 μg/kg) was given before termination of CPB. Dopamine (5 μg/kg/min) was initiated for the immediate time around cessation of CPB and was subsequently discontinued if the haemodynamic and ventricular functions were good. A pulmonary venous line was inserted into the orifice of the right upper pulmonary vein.

Postoperative management

The central temperature (oesophageal) was maintained at 36–37°C. Postoperative monitoring included arterial, superior vena caval and pulmonary venous pressures and heart rate. Sedation was maintained by a continuous intravenous infusion of morphine and intermittent injections of a muscle relaxant (pancuronium) and lorazepam. Infants were ventilated with volume control and pressure support. Ventilation volume and rate were adjusted to maintain Paco₂ between 40–50 mm Hg. Inotropic agents, vasoactive drugs (milrinone, dopamine, phenoxybenzamine and vasopressin) and volume infusions (5% albumin or blood) were given according to our standard protocol.²²

Methods of measurement

Sco₂ and Sso₂

NIRS probes consist of a near‐infrared light emitter optode and a receiver optode with a distance of 5 cm. In pigs, they were placed on the right and left sides of the forehead. In patients, we chose to replicate previously published techniques of probe placement in this group of patients.¹³ Briefly, the probes were placed on the patient's forehead in the midline (Sco₂) and slightly to the right of the midline on the thoracic–lumbar flank (Sso₂). The probes were monitored by a dual‐detector device (INVOS 5100A; Somanetics, Troy, Michigan, USA) and recordings were made at 1 min intervals.

V̇o₂

V̇o₂ was measured continuously with an AMIS2000 mass spectrometer (Innovision A/S, Odense, Denmark). This is a sensitive and accurate method that permits simultaneous measurements of multiple gas fractions. We have described the details elsewhere.²³

Calculations of systemic haemodynamic variables and oxygen transport

Blood samples were taken from the arterial, superior vena cava and pulmonary vein lines for the measurements of blood gases. Qp and Qs were then calculated by the direct Fick method: Qp  =  V̇o₂/(Cpvo₂ − Cao₂) and Qs  =  V̇o₂/(Cao₂ −Cvo₂), where Cao₂, Cpvo₂ and Cvo₂ indicate systemic arterial (equal to pulmonary arterial), pulmonary venous and superior vena caval oxygen contents, respectively. Do₂ and ERo₂ were calculated by standard equations: Do₂  =  Qs × Cao₂ and ERo₂  =  V̇o₂/Do₂. All values in patients were indexed to body surface area and in pigs, to body weight.

Study protocols

The animal study (study 1) was performed during the first 6 h after CPB. Six sets of measurements were obtained at 1, 3 and 6 h, with a 10 min interval between each set of measurements. The clinical study (study 2) was performed during the first 72 h after the patient's arrival in the cardiac intensive care unit. Values of haemodynamic function, oxygen transport and central body temperature were collected at 2 h intervals during the first 24 h and at 4 h intervals during hours 25 through 72. Sampling was avoided if sedation, paralysis and ventilatory or haemodynamic treatment were changed within the prior 15 min.

Data analysis

Data are expressed as mean (SD). Interrelationships between the measures were sought by using mixed linear regression analysis for repeated measures without regard to time. When a significant correlation was found (p < 0.05), interindividual differences were further analysed. The extent of the correlation was indicated by the intercept and slope values. All data were analysed with SAS statistical software V.8 (SAS Institute, Inc, Cary, North Carolina, USA).

RESULTS

Pigs

Two pigs died before the end of the 6 h study period. As the NIRS measures of Sco₂ on the two sides were similar, the mean values were used for analysis.

Patients

Sco₂ was obtained in all the patients and Sso₂ in five patients, in three of whom measurements were stopped at 28 and 48 h, respectively, due to technical issues in two patients (patients 1 and 3) and extubation in the other (patient 11). There was no incidence of circulatory collapse or death during the study period. All patients survived to hospital discharge. Extubation was done between 2–16 days (median seven days) after the procedure except in one child who had vocal cord complications. Extubation for that infant was done at 90 days, after a bidirectional cavopulmonary anastomosis (table 1).

Correlations of Sco₂ and Sso₂ with systemic haemodynamic variables and oxygen transport

Table 2 and figs 1–3 detail the results of the correlations of Sco₂ and Sso₂ with haemodynamic variables and oxygen transport in the animal and clinical studies.

Table 2 Correlations (mixed linear regression) of Sco₂, Sso₂ and Svo₂ with haemodynamic and oxygen transport variables for the entire group and individually in 7 pigs and 11 patients.

Dependent variable	Independent variable	Intercept	Group slope	p Value	Range of individual slopes	p Value for interindividual slope difference
Pig data
Sco2	Svo2	28.1	0.30	<0.0001	0.14–0.67	0.0005
	Sao2	57.6	−0.11	0.82
	Pao2	48.0	−0.01	0.60
	Haemoglobin	52.4	−0.36	0.28
	MAP	18.6	0.64	<0.0001	−0.71–1.16	<0.0001
	CO	39.7	2.30	<0.0001	1.56–6.42	0.0006
	Do2	39.0	0.017	<0.0001	0.02–0.07	0.006
	V̇o2	39.9	0.01	0.48
	ERo2	59.7	−32.5	<0.0001	−65.4–16.4	0.002
Patient data
Sco2	Svo2	28.5	0.43	<0.0001	0.17–0.97	<0.0001
	Sao2	4.6	0.61	<0.0001	0.14–2.13	0.0001
	Pao2	12.1	0.99	<0.0001	0.36–1.93	<0.0001
	Haemoglobin	47.5	0.24	0.54
	MAP	26.0	0.52	<0.0001	0.14–1.11	0.21
	Qp	43.6	3.72	0.001	0.86–6.72	0.03
	Qs	42.1	3.22	<0.0001	0.27–13.64	0.002
	Do2	42.4	0.03	<0.0001	0.01–0.10	0.001
	V̇o2	56.8	−0.07	0.046	−0.33–0.39	<0.0001
	ERo2	61.1	−33.2	<0.0001	−91.7–−5.3	<0.0001
Sso2	Svo2	49.2	0.26	<0.0001	0.11–0.54	0.32
	Sao2	33.9	0.39	0.0001	Infinite likelihood
	Pao2	46.9	0.41	0.002	0.19–0.83	0.22
	Haemoglobin	58.2	0.30	0.48
	MAP	51.7	0.22	0.03	Infinite likelihood
	Qp	60.0	1.48	0.23
	Qs	59.3	1.73	0.03	0.46–10.3	0.28
	Do2	58.2	0.015	0.004	0.006–0.066	0.11
	V̇o2	63.3	−0.007	0.84
	ERo2	68.9	−20.3	0.0002	Infinite likelihood
Svo2	Qp	47.1	2.49	0.02	−9.3–10.8	0.04
	Qs	40.5	5.82	<0.0001	3.1–11.3	0.02
	Do2	36.4	0.05	<0.0001	0.032–0.097	0.001
	V̇o2	69.7	−0.20	<0.0001	−0.45–0.028	0.008
	ERo2	77.9	−83.2	<0.0001	−91.9–−69.4	0.18

Do₂, systemic oxygen delivery; ERo₂, oxygen extraction ratio; MAP, mean arterial pressure; Pao₂, arterial oxygen pressure; Qp, pulmonary blood flow; Qs, systemic blood flow; Sao₂, arterial oxygen saturation; Sco₂, cerebral oxygen saturation; Sso₂, splanchnic oxygen saturation; Svo₂, superior vena caval oxygen saturation; V̇o₂, systemic oxygen consumption.

Figure 1 Correlations between cerebral oxygen saturation (Sco₂) and systemic haemodynamic and oxygen transport variables of arterial oxygen saturation (Sao₂), superior vena caval oxygen saturation (Svo₂), arterial partial oxygen pressure (Pao₂), mean arterial pressure (MAP), pulmonary (Qp) and systemic blood flows (Qs), systemic oxygen consumption (V̇o₂), oxygen delivery (Do₂) and oxygen extraction ratio (ERo₂) in patients during the first 72 h after arrival in the cardiac intensive care unit.

Figure 2 Correlations between splanchnic oxygen saturation (Sso₂) and systemic haemodynamic and oxygen transport variables of superior vena caval oxygen saturation (Svo₂), arterial oxygen saturation (Sao₂), arterial partial oxygen pressure (Pao₂), mean arterial pressure (MAP), pulmonary (Qp) and systemic blood flows (Qs), oxygen delivery (Do₂), systemic oxygen consumption (V̇o₂) and oxygen extraction ratio (ERo₂) in patients during the first 72 h after arrival in the cardiac intensive care unit.

Figure 3 Trends for cerebral oxygen saturation (Sco₂, solid line) and superior vena caval oxygen saturation (Svo₂, dotted line) for each of the 11 patients during the first 72 h after arrival in the cardiac intensive care unit (ICU).

Study 1

We obtained 106 sets of measurements in the seven pigs. Sco₂ ranged from 30–59%, Svo₂ from 46.6–86.1%, Sao₂ from 91.1–100% and arterial oxygen pressure (Pao₂) from 60–237 mm Hg.

Sco₂ was not correlated with Sao₂ (slope −0.11, p  =  0.82), Pao₂ (slope −0.01, p  =  0.60) or haemoglobin (slope −0.36, p  =  0.28). Sco₂ was correlated with Svo₂ (slope 0.30, p < 0.0001) but with large interindividual variations (slope 0.14 to 0.67, p  =  0.0005). Sco₂ was also correlated with mean arterial pressure (slope 0.64, p < 0.0001) with large interindividual variations (slope −0.71 to 1.16, p < 0.0001) and moderately correlated with Qs, Do₂ and ERo₂ (slope 2.3, 0.017, −32.5, respectively, p < 0.0001 for all) with significant interindividual variations (table 2). Sco₂ was not correlated with V̇o₂ (slope 0.01, p  =  0.48) (table 2).

Study 2

We obtained 247 sets of measurements for Sco₂ and 103 sets for Sso₂. For the entire group, Sco₂ ranged from 25–79% and Sso₂ from 40–83%; Svo₂ ranged from 24–74%, Sao₂ from 55–86% and Pao₂ from 25.0–50.6 mm Hg.

Unlike in pigs, in patients Sco₂ was correlated with both Sao₂ and Pao₂ (slope 0.61 and 0.99, respectively, p < 0.0001 for both) but with large interindividual variations (slope 0.17–0.97 for Sao₂; 0.36–1.93 for Pao₂; p < 0.0001 for both). Sco₂ was not correlated with haemoglobin (slope 0.24, p  =  0.54). Similarly to the data in pigs, Sco₂ was correlated with Svo₂ (slope 0.43, p < 0.0001) but with large interindividual variations (slope 0.17–0.97 for Svo₂; 0.14–2.13 for Sao₂; p < 0.0001 for both). It was also correlated with mean arterial pressure (slope 0.52, p < 0.0001) but, interestingly, interindividual variations were insignificant (slope 0.14–1.11, p  =  0.21). Sco₂ was weakly correlated with V̇o₂ (slope −0.07, p  =  0.05) and moderately correlated with Qp, Qs, Do₂ and ERo₂ (slope 3.1, 3.2, 0.03, −33.2, respectively, p < 0.0001 for all except for Qp, p  =  0.001) with significant interindividual variations (table 2, fig 1).

Sso₂ was similarly, although more weakly, correlated with haemodynamic and oxygen transport variables, with the coefficients being about half those with Sco₂. Interindividual variations were also large, although they did not achieve significance, probably due to the smaller sample size. Sso₂ was not correlated with V̇o₂ (slope −0.007, p  =  0.84) (table 2, fig 2).

The correlations of Sco₂ and Sso₂ with Qs, Do₂, V̇o₂ and ERo₂ were much weaker than those of Svo₂ with these variables (table 2).

Lastly, examples of individual patient plots (fig 3) of the NIRS and the directly derived data show notably variable intraindividual relationships that changed with time and at individual time points.

DISCUSSION

This is the first comprehensive investigation of the relationship between NIRS measurements of Sco₂ and Sso₂, as a non‐invasive clinical haemodynamic monitor of systemic oxygenation, and the directly measured systemic haemodynamic and oxygen transport variables. We assessed the utility of NIRS after CPB in two different circulations—namely, the normal biventricular circulation in an animal model and the more complex parallel circulation, with residual arterial desaturation, in children after the Norwood procedure. Our data showed that, in patients with varied and relatively low values of Pao₂ and Sao₂, Sco₂ was closely correlated with both of these values, but not in pigs with fully saturated arterial oxygenation. Most important, Sco₂ was, similarly in both groups, closely correlated with mean arterial pressure, loosely correlated with Svo₂, Sao₂, Qp, Qs, Do₂ and ERo₂, poorly correlated with V̇o₂ and not significantly correlated with either haemoglobin or Qp. Whereas the correlation between NIRS‐derived oxygen saturations and many of the directly measured indices were highly significant, the relatively loose correlations at an absolute level, combined with wide interindividual variability, cast doubt on the potential clinical utility of such measurements.

Previous studies have evaluated single‐point comparisons between Sco₂ or Sso₂ and Svo₂ in individual children in a larger population.¹⁷,¹⁸,¹⁹ Good correlations have generally been reported, but large interindividual differences are apparent, even in these studies. Our study was the first to evaluate NIRS in the setting of postoperative complex congenital heart disease, comparing the non‐invasive data with directly measured indices of oxygen transport during the first three days after the Norwood procedure, and confirmed the previous findings. Our data showed that an increase in Sco₂ or Sso₂ of 1% explained 0.43% or 0.26%, respectively, of the increase in Svo₂. However, interindividual variations were large, ranging from 0.17–0.97% for Sco₂ and from 0.11–0.54% for Sso₂. Furthermore, the individual trends shown in fig 3 show an inconsistent relationship between Sco₂ and Svo₂ throughout the measurement period. Although differences in absolute values and even wide interindividual variability may be tolerated if NIRS is used for trend analysis, the second issue of an unreliable intraindividual relationship between NIRS and Svo₂, for example, casts doubt on its potential utility as a precise tool for monitoring haemodynamic trends.

This is, however, not surprising. NIRS measures the equilibrium of oxyhaemoglobin and deoxyhaemoglobin in a mixture of veins, arteries and capillaries in the underlying tissue and reflects a regional state of oxygenation. Although NIRS has been extensively used to monitor cerebral and splanchnic oxygenation in various clinical situations including during CPB, deep hypothermic circulatory arrest¹³,²⁴ and in other high risk newborns,⁶,¹⁵,¹⁹,²⁵ and has been found to be helpful in predicting cerebrovascular dysfunction²⁴,²⁵ and splanchnic ischaemia,¹⁵ it has rarely been rigorously examined for its validity, particularly in the setting of complex parallel circulations such as those that exist after the Norwood procedure. However, whereas the venous portion predominantly determines NIRS measurement of the underlying tissue oxygenation, direct measurements, such as jugular bulb oxygen saturation and hepatic venous oxygen saturation, are well known to correlate variably with NIRS measurement of that organ.¹¹,¹⁴ This may be a particular problem in children with complex congenital heart disease, where the contribution of the venous portion has been shown to range from 60–100% of Sco₂ with varied systemic oxygen saturation as seen in our patients after the Norwood procedure.¹² It must also be remembered that NIRS measures oxygenation in a small part of the target organ, and the regional venous oxygen saturation reflects the balance of oxygen delivery and consumption of the whole organ. Svo₂, conversely, reflects the balance of systemic oxygen transport, presumably explaining the discrepancy that Yeh T Jr et al²⁶ observed between jugular bulb oxygen saturation and Svo₂ in children undergoing CPB. Extracerebral tissue factors, such as ischaemia,²⁷ oedema and the location of the probe,²⁸ may also affect the robustness of NIRS signals.

Limitations

The superior vena cava was used to measure systemic venous saturation for the calculations of Qs and Do₂. This measure does not account for potential differences in inferior vena cava saturation²⁹,³⁰ and may at least partly account for the poorer correlations with Sso₂. Conversely, it could be argued that, by sampling upper body venous saturation, we have a more representative measurement, incorporating cerebral blood flow. Nonetheless, jugular bulb saturation would have been better in this regard but was not appropriate in this clinical protocol.³¹

Furthermore, we chose the usual position (posterior flank) to assess Sso₂ by NIRS. Recently, Fortune et al¹⁵ used the site of just below the umbilicus to monitor splanchnic oxygenation and found a sensitivity of 90% to detect splanchnic ischaemia in neonates during apnoeic episodes. Sso₂ measured in such a way may reflect more Do₂ and V̇o₂, and thus warrants further investigation in children with heart disease.

Conclusions

NIRS measurement of Sco₂ and Sso₂ reflects the changes in haemodynamic variables and oxygen transport during the early postoperative period after the Norwood procedure. However, large interindividual differences and intraindividual temporal variability make interpretation difficult and may limit the utility of NIRS as a continuous monitor of systemic haemodynamic function and oxygen transport in critically ill patients.

Abbreviations

Cao₂ - systemic arterial oxygen contents

CPB - cardiopulmonary bypass

Cpvo₂ - systemic pulmonary venous oxygen contents, Cvo₂, systemic superior vena caval oxygen contents

Do₂ - systemic oxygen delivery

ERo₂ - oxygen extraction ratio

NIRS - near‐infrared spectroscopy

Pao₂ - arterial oxygen pressure

Qp - pulmonary blood flow

Qs - systemic blood flow

Sao₂ - arterial oxygen saturation

Sco₂ - cerebral oxygen saturation

Sso₂ - splanchnic oxygen saturation

Svo₂ - superior vena caval oxygen saturation

V̇o₂ - systemic oxygen consumption