Role of Pv-aCO₂ gradient and Pv-aCO₂/Ca-vO₂ ratio during cardiac surgery: a retrospective observational study

Abstract

Introduction:

Arterial lactate, mixed venous O₂ saturation, venous minus arterial CO₂ partial pressure (P_v-aCO₂) and the ratio between this gradient and the arterial minus venous oxygen content (P_v-aCO₂/C_a-vO₂) were proposed as markers of tissue hypoperfusion and oxygenation. The main goals were to characterize the determinants of P_v-aCO₂ and P_v-aCO₂/C_a-vO₂, and the interchangeability of the variables calculated from mixed and central venous samples.

Methods:

35 cardiac surgery patients were included. Variables were measured or calculated: after anesthesia induction (T1), end of surgery (T2), and at 6–8 hours intervals after ICU admission (T3 and T4).

Results:

Macrohemodynamics was characterized by increased cardiac index and low systemic vascular resistances after surgery (p < 0.05). Hemoglobin, arterial-pH, lactate, and systemic O2 metabolism showed significant changes during the study (p < 0.05). P_v-aCO₂ remained high and without changes, P_v-aCO₂/C_a-vO₂ was also high and decreased at T4 (p < 0.05). A significant correlation was observed globally and at each time interval, between P_v-aCO₂ or P_v-aCO₂/C_a-vO₂ with factors that may affect the CO₂ hemoglobin dissociation. A multilevel linear regression model with P_v-aCO₂ and P_v-aCO₂/C_a-vO₂ as outcome variables showed a significant association for P_v-aCO₂ with S_vO₂, and BE (p < 0.05), while P_v-aCO₂/Ca-vO₂ was significantly associated with Hb, S_vO₂, and BE (p < 0.05) but not with cardiac output. Measurements and calculations from mixed and central venous blood were not interchangeable.

Conclusions:

P_v-aCO₂ and P_v-aCO₂/C_a-vO₂ could be influenced by different factors that affect the CO₂ dissociation curve, these variables should be considered with caution in cardiac surgery patients. Finally, central venous and mixed values were not interchangeable.

KEYWORDS
Anaerobiosis; Carbon dioxide; Cardiac surgical procedures; Perfusions

Introduction

Macrohemodynamic and metabolic variables drive cardiovascular support during cardiac surgery.¹, ² As described in other critically ill conditions, arterial lactate and mixed venous oxygen saturation (S_vO₂) have been used to evaluate tissue perfusion and oxygenation.³, ⁴

Parameters derived from carbon dioxide metabolisms (CO₂), such as the difference between venous and arterial CO₂ partial pressure (P_v-aCO₂) and the ratio between this gradient and the arterial minus venous oxygen content (P_v-aCO₂/C_a-vO₂), have been proposed as markers of hypoperfusion and anaerobic metabolism.⁵, ⁶, ⁷ However, the levels of these variables during cardiac surgery remain controversial.

In critically ill patients, reductions in both CO₂ production (VCO₂) and O₂ consumption (VO₂) are noted during oxygen supply dependence conditions. However, the decrease in VCO₂ was less pronounced than in VO₂ due to the increase in VCO₂ from anaerobic metabolism. Based on the modified Fick equation, VCO₂ is calculated as the product of cardiac output (CO) times venous minus arterial CO₂ content (VCO₂ = CO × C_v-aCO₂). Assuming a stable relationship between CO₂ partial pressure (PCO₂) and CO₂ content (CCO₂), the equation could be reformulated as VCO₂ = CO x P_v-aCO₂. Thus, P_v-aCO₂ increases in low cardiac output states and during microvascular dysfunction.⁸, ⁹

The P_v-aCO₂/C_a-vO₂ ratio has been proposed as a surrogate of the respiratory quotient (RQ) to identify the onset of anaerobic metabolism. Assuming stable conditions, CCO₂ could be replaced by PCO₂ in the formula. However, this ratio is influenced by other factors, such as changes in hemoglobin (Hb), lactate concentrations, base excess (BE), S_vO₂ (Haldane effect), and body temperature. Given that CO₂-derived parameters are measured under the influence of these factors, their physiologic meaning could be misleading. Despite these limitations, P_v-aCO₂ and P_v-aCO₂/C_a-vO₂ have been used in clinical practice to guide hemodynamic support and as outcome biomarkers.⁵, ⁸, ⁹ A better understanding of the limits and sources of errors of this practice should be considered in critically ill patients.¹⁰ Many investigations use central venous blood values (P_c-vCO₂) measured from the superior vena cava as equivalent to mixed venous blood measured from the pulmonary artery (P_vCO₂). The central venous-to-arterial CO₂ difference may represent one of the alternatives to the mixed-to-arterial CO₂ difference, but whether central venous blood values could subrogate mixed venous values is still debated.¹⁰, ¹¹, ¹²

We hypothesized that CO₂ calculated variables are dependent on physiologic alterations. Additionally, central and mixed venous blood parameters are not interchangeable.

The main goal was to assess the physiological determinants that affect P_v-aCO₂ and P_v-aCO₂/C_a-vO₂ during cardiac surgery. Second, we analyzed the agreement between mixed and central venous blood values to determine whether they are interchangeable.

Methods

The present study is a secondary retrospective analysis of a previous multicenter investigation,¹³ which was approved by the Institutional Bioethical Committee, and written consent was obtained from each patient. The objective in the original paper was to determine the clinical significance of the gradient SO₂ and lactate within the superior vena cava to the pulmonary artery in critical patients. Using this database, the present research focused on 35 cardiac surgeries performed in our institution.

We enrolled patients older than 18 years of age of either sex who required the insertion of a pulmonary artery catheter: LVEF <40%, combined surgery, pulmonary-hypertension and reintervention. Patients with known uncorrected valvular incompetence or intracardiac shunting were excluded. The pulmonary artery catheter monitoring (PAC-7.5 Edwards Life Sciences) was inserted in the right internal jugular vein, and a radial arterial line was placed according to standard clinical practice.

Anesthesia was induced with etomidate and fentanyl and maintained with isoflurane and fentanyl. Mechanical ventilation was established to maintain an arterial PCO₂ between 35 and 40 mmHg. CBP was achieved with a nonpulsatile flow using a membrane oxygenator at 32-33°C. The mean arterial pressure was maintained between 50 and 60 mmHg. Vasoactive and inotropic infusions were prescribed and titrated following standard practice in each patient. Before arterial and venous cannulation, 300mg.kg⁻¹ sodium heparin was administered intravenously (IV) to achieve an activated coagulation time greater than or equal to 480 seconds. Sodium heparin was later neutralized with protamine sulfate in a 1:1 proportion. All patients received 500 mg methylprednisolone during CPB. Body temperature was returned to 37°C before decannulation. Parenteral analgesia, sedation, and mechanical ventilation were sustained until the patients were hemodynamically stable, awake, and ready for weaning from mechanical ventilation.

Measurements and calculations were taken at T1, after anesthesia induction and before CPB; T2, after the end of surgery; T3 and T4, at 6- to 8-hour intervals after ICU admission.

Blood samples from the arterial line, internal jugular vein (proximal port), and pulmonary artery were drawn simultaneously and in duplicate. Hb concentration, blood gases, lactate, BE, and SO₂ were measured simultaneously in each vascular compartment (ABL-700, Radiometer, Denmark).

CO was measured in triplicate by thermodilution and was indexed by body weight. Systemic vascular resistance (SVR) was calculated. The physician in charge conducted hemodynamic and anesthetic treatment according to standard practice.

Mixed venous CO₂ content was calculated according to previously described formulas.¹⁴ The venous-arterial PCO₂ gradients were calculated from mixed and central venous blood and expressed as P_v-aCO₂, C_v-aCO₂, and P_cv-aCO2.

Arterial and venous oxygen contents were calculated using standard formulas. The venous-arterial CO₂ partial pressure/arterial-venous O₂ content ratio was calculated from central and pulmonary venous blood and expressed as P_v-aCO₂/C_a-vO₂ for mixed and P_cv-aCO₂/C_a-cvO₂ for central venous blood, respectively.

Statistical analysis

Data are presented as the mean and standard deviation (SD), standard error (SE) or absolute numbers and percentages (%). Variables over time were analyzed by repeated-measures ANOVA followed by the Bonferroni post hoc test. Linear correlations were studied between P_v-aCO₂ and P_v-aCO₂/C_a-vO₂ with hemodynamic and metabolic variables. Multilevel linear regression models with mixed effects for P_v-aCO₂ and P_v-aCO₂/C_a-vO₂ as outcome variables were performed, including CI, Hb, pH, BE, and S_vO₂ at different time intervals. Statistical significance was assumed when p < 0.05.

The interchangeability between central and pulmonary blood samples was studied using Bland-Altman analysis. The utility of this approach is limited when treatments or interventions affect the variables over time. Then, change or delta values were plotted using a Cartesian X-Y 4-quadrant plot, allowing the evaluation of the direction of the change or the concordance rate. The concordance rate was defined as the percentage of values included in the right superior and left inferior quadrants of the plots. The agreement between variables was considered weak when this percentage was less than 80%. The polar plot analysis allowed a more precise evaluation of the trend and magnitude of changes between the study (central venous blood) and reference variables (mixed venous blood). From the central point, changes in the calculated pairs of values are represented as vectors with defined angles and magnitudes. The mean angular bias (q) and the standard deviation represent all measured angles from the polar reference axis (0°). In addition, radial limits of agreement were estimated as the radial sector that contains 95% of the values (2SD). A mean angular value ±5° and a radial limit of agreement ±30° were the defined limits for the polar plot analysis.¹⁵

Results

Table 1 presents demographics characteristics. Norepinephrine was used in 2.7% of patients during T1, 50% in T2 and T3, and 36.1% in T4 (p < 0.05). Dobutamine or milrinone was used in 11.1% of patients during T1, 16.6% during T2 and T3, and 19.4% during T4 (ns). Drug infusion was titrated by the physician in charge. Three patients (8.6%) died.

Table 1
Main characteristics of cardiac surgery patients.

Age (years)	65 ± 10
BMI (kg.m−2)	28 ± 3
LVEF (%)	47 ± 13
CPB duration (min)	124 ± 32
Aortic clamp (min)	80 ± 29
Gender
Female	15/35
Male	20/35
Type of Surgery
Combined	10/35
CABG on-pump	12/35
CABG off-pump	6/35
Valvular replacement	6/35
Thoracic aortic surgery	1/35
Comorbidities
Arterial hypertension	17/35
Dyslipidemia	9/35
Smokers	11/35
Diabetes	7/35
Creatinine > 2 mg.dL−1	1/35

• Mean values ± SD, or fractions.
• BMI, body mass index; CABG, coronary artery bypass grafting; CPB, cardiopulomary bypass.

Hemodynamics and metabolic course

Systemic hemodynamics, O₂ metabolism and metabolic variables are summarized in Table 2. This patient exhibited a postoperative hemodynamic pattern characterized by high CI and low SVR (p < 0.05). The MAP was 75.1 ± 18.7 mm Hg and increased significantly to 86.7 ± 14.1 mm Hg (T4) compared to T2 (p < 0.05). The Hb concentration was significantly lower than baseline during T2 and T3 (p < 0.05). These changes were accompanied by a decrease in arterial oxygen content (C_aO₂) at the same time intervals (p < 0.05). Arterial pH decreased significantly during T2 and T3 (p < 0.05), whereas arterial lactate levels were higher than baseline throughout the study (p < 0.05). Although a transient trend toward lower BE values was noted in the postoperative period, these changes were not statistically significant compared to baseline. S_vO₂ and S_cvO₂ were significantly lower during T3 and T4 (p < 0.05). Oxygen delivery (DO₂) increased significantly during the postoperative period (T2 and T3) (p < 0.05), returning to basal values at T4. VO₂ increased and remained significantly higher than baseline levels after T2 (p < 0.05). The oxygen extraction ratio (EO₂) increased during T3 and T4 (p < 0.05).

Table 2
Hemodynamic, metabolic and CO₂derived variables at different time intervals.

	T1	T2	T3	T4
CI (L.min−1)	2.09 ± 0.62	2.71 ± 0.68a	2.86 ± 0.71a	2.81 ± 1.18a
SVRI (dyn.s/cm−5)	2670 ± 902	1872 ± 663a	2020 ± 648a	2224 ± 618b
MAP (mmHg)	75.1 ± 18.7	71.9 ± 13.5	79.4 ± 15.5	86.7 ± 14.1b
Hb (g.L−1)	11.8 ± 2.4	10.1 ± 1.8	9.9 ± 1.8a	10.7 ± 2.9
pH	7.40 ± 0.08	7.33 ± 0.09a	7.33 ± 0.08a	7.39 ± 0.1b
BE (mmol.L−1)	−1.3 ± 3.26	−4.3 ± 3.8	−4.8 ± 4.4	−2.4 ± 3.7
PaCO2 (mmHg)	36.9 ± 7.2	39.8 ± 9.8	37.9 ± 5.5	36.4 ± 5
Lactate (mmol.L−1)	1.27 ± 0.7	3.3 ± 1.4a	4.3 ± 3.5a	4.2 ± 4.9a
SvO2 (%)	77.0 ± 8.4	75.6 ± 6.5	64.6 ± 9.6a, b	62.1 ± 8.2a, b
ScvO2 (%)	77.3 ± 9.7	77.3 ± 7.3	67.6 ± 9.2a, b	65.2 ± 9.6a, b
CaO2 (mL.dL−1)	15.8 ± 3.3	13.6 ± 2.4	13.3 ± 2.4	14.1 ± 4.1
CvO2 (mL.dL−1)	12.3 ± 2.7	10.4 ± 2.3a	8.7 ± 2.3a	8.9 ± 3.4a
Ca-vO2 (mL.dL−1)	3.8 ± 1.5	3.2 ± 0.9	4.6 ± 1.1	5.0 ± 1.1a,b
CcvO2 (mL.dL−1)	12.4 ± 2.9	10.7 ± 2.3a	9.1 ± 2.3a	9.5 ± 3.5a
Ca-cvO2 (mL.dL−1)	3.6 ± 1.8	2.9 ± 1.1	4.1 ± 1.2	4.6 ± 1.5b
DO2 (mL.min−1 .m−2)	487 ± 197	663 ± 275a	590 ± 279a	513 ± 259
VO2 ( mL.min−1 .m−2)	108 ± 45	147 ± 51a	195±81a, b	179 ± 89a
EO2 (%)	23.3 ± 7.7	23.6 ± 6.3	34.5 ± 9.8a, b	36.5 ± 10a, b
Pv-aCO2 (mm Hg)	7.8 ± 2.5	7.5 ± 4.3	7.9 ± 2.7	7.2 ± 2.5
Cv-aCO2 (mL.dL−1)	5.7 ± 0.7	4.6 ± 1.0	4.7 ± 0.4	4.5 ± 0.5
Pv-aCO2/C(a-v)O2 (mm Hg.mL−1)	2.1 ± 0.6	2.5 ± 1.3	1.8 ± 0.7	1.4 ± 0.6a, b
Cv-aCO2/Ca-vO2 (mL.dL−1)	1.5 ± 0.1	1.5 ± 0.3	1.1 ± 0.1	0.9 ± 0.1a
Pcv-aCO2 (mmHg)	8.8 ± 3.0	8.0 ± 4.3	9.5 ± 3.3	7.5 ± 3.0
Pcv-aCO2/Ca-cvO2 (mmHg.mL−1)	2.8 ± 1.7	2.8 ± 1.3	2.2 ± 0.9	1.5 ± 0.5a, b

• CI, Cardiac Index; SVRI, Systemic Vascular Resistance Index; MAP, Mean Arterial Pressure; Hb, hemoglobin concentration; pH, arterial pH; BE, Base Excess; PaCO₂, arterial carbon dioxide partial pressure; Lactate, arterial lactate; S_vO₂, mixed venous hemoglobin saturation; S_cvO₂, central venous hemoglobin saturation; C_aO₂, arterial oxygen content; C_vO₂, mixed venous oxygen content; C_cvO₂, central venous oxygen content; DO₂, systemic oxygen delivery index; VO₂, systemic oxygen consumption index; EO₂, systemic oxygen extraction ratio. P_v-aCO₂, mixed venous minus arterial CO₂ partial pressure gradient; C_v-aCO₂, mixed venous minus arterial CO₂ Content; P_v-aCO₂/C_(a-v)O₂, mixed venous minus arterial PCO₂ /arterial minus mixed venous Oxygen content ratio; C_v-aCO₂/C_a-vO₂, mixed venous minus arterial CO₂ Content / arterial minus mixed venous Oxygen content ratio; P_cv-aCO₂, central venous minus arterial CO₂ partial pressure gradient; P_cv-aCO₂/C_a-cvO₂, central venous minus arterial PCO₂/arterial minus central venous oxygen content ratio. T1, initial evaluation; after general anesthesia and before surgery; T2, after surgery and extracorporeal circulation; T3 and T4, after ICU admission at 6-8 hours intervals. Mean values ± SD.
• ^a
  p < 0.05 vs T1.
• ^b
  p < 0.05 vs T2.

CO₂-derived measurements

CO₂-derived measurements are shown in Table 2. The P_v-aCO₂ and C_v-aCO₂ gradients did not exhibit significant changes compared to baseline. The P_v-aCO₂/C_a-vO₂ and C_v-aCO₂/C_a-vO₂ ratios did not change until T4, when the values were significantly lower than baseline (p < 0.05).

P_cv-aCO₂ and P_cv-aCO₂/C_a-cvO₂ showed higher absolute values but similar trends during the study.

To identify the overall influence of the different physiologic variables, we performed a global correlation for both variables. Figure 1 shows the scatter plot and correlation coefficients between P_vaCO₂ and the main hemodynamic and metabolic variables. A significant positive linear correlation was identified between P_vaCO₂ and arterial lactate and EO₂ (p < 0.05). A negative linear correlation was identified between the same variable and BE, S_vO₂, Hb, pH, and DO₂ (p < 0.05). However, r² values were always low, denoting a weak coefficient of determination between these variables. Figure 2 shows the global correlation between P_v-aCO₂ /C_v-aO₂ with the same variables. A significant positive linear correlation with CI and S_vO₂ (p < 0.05) was noted. On the other hand, a significant inverse linear correlation was found between P_v-aCO₂/C_v-aO₂ and Hb, arterial pH, BE, and EO₂ (p < 0.05), but all r² values were low.

Figure 1
Correlation between mixed venous minus arterial CO₂ partial pressure (P_v-aCO₂) with hemodynamic and metabolic parameters. CI, cardiac index; Hb, hemoglobin concentration; Lactate, arterial lactate; S_vO₂, mixed venous oxygen saturation; pH, arterial pH; BE, arterial base excess; EO₂, Systemic O₂ extraction ratio; DO₂, Systemic O₂ delivery.

Figure 2
Correlation between mixed venous minus arterial CO₂ partial pressure/arterial minus mixed venous oxygen content ratio (P_v-aCO₂/C_a-vO₂). CI, cardiac index; Hb, hemoglobin concentration; Lactate, arterial lactate; S_vO₂, mixed venous oxygen saturation; pH, arterial pH; BE, arterial base excess; EO₂, Systemic O₂ extraction ratio; DO₂, Systemic O₂ delivery.

Although the coefficients of determination remained weak, the statistical significance of the correlations varied over time. P_v-aCO₂ did not correlate with any of these variables at baseline, but it was significantly correlated with lactate (T2) (p < 0.05) or with Hb, lactate, S_vO₂, BE and EO₂ at the T3 (p < 0.05) and T4 time points (p < 0.05). In addition, the P_v-aCO₂ /C_a-vO₂ ratio was significantly correlated with Hb, S_vO₂, BE and EO₂ at T1 (p < 0.05); S_vO₂ and EO₂ at T2 (p < 0.05); Hb, lactate and BE at T3 (p < 0.05); and CI, Hb and BE at the end of the study (T4) (p < 0.05).

Table 3 summarizes the multilevel linear regression model with mixed effects for the P_v-aCO₂ difference and P_v-aCO₂/C_a-vO₂ ratio as outcome parameters, including CI, Hb, arterial pH, BE, and S_vO₂ for each analysis. S_vO₂ and BE were the main significant determinants for P_v-aCO₂ (r² = 0.25), whereas Hb, S_vO₂, and BE were the statistical determinants for the P_v-aCO₂/C_a-vO2 ratio (r² = 0.33) (p < 0.05).

Table 3
Multilevel linear regression model with mixed effects for P_v-aCO₂ difference and P_v-aCO₂/C_a-vO₂ ratio.

	Coefficient	SE	p-value	95% CI
Pv-aCO2
Constant	11.4	2.00	< 0.001	5.70 to 7.19
SvO2 (%)	−0.082	0.027	0.002	−0.135 to −0.029
BE (mmol.L−1)	−0.238	0.072	0.001	−0.381 to −0.096
r2 = 0.25; p < 0.001
Pv-aCO2/Ca-vO2
Constant	0.98	0.624	0.118	−2.55 to 2.24
SvO2 (%)	0.046	0.007	< 0.001	0.030 to 0.061
Hb (g.L−1)	−0.222	0.035	< 0.001	−0.292 to −0.152
BE (mmol.L−1)	−0.084	0.022	< 0.001	−0.128 to 0.041
r2 = 0.33; p < 0.001

• S_vO₂, mixed venous oxygen saturation; Hb, hemoglobin concentration; BE, base excess; SE, standard error.

Bland-Altman, Quadrant plot and Polar plot analysis

The bias of the difference in the Bland-Altman analysis for P_v-aCO₂ and P_cv-aCO₂ was 0.59, whereas the limits of agreement were −3.7 to 4.9 mm Hg (2SD). P_v-aCO₂/C_a-vO₂ and P_cv-aCO₂/C_a-cvO₂ were also analyzed. The mean difference value was 0.38, whereas the limits of agreement ranged from −1.13 to 1.89 (2SD).

In Fig. 3, the upper panel shows the 4-quadrant and polar plot analysis that describes the trend and magnitude of changes between AP_v-aCO₂ and AP_cv-aCO₂ gradients. The concordance rate was 73 %, and the mean polar angle was −1.5° ± 38°. The radial limit of agreement was 76°. The AP_v-aCO₂/C_a-vO₂ and AP_cv-aCO₂/C_a-cvO₂ ratios are shown in the lower panel. The concordance rate calculated from the 4-quadrant plot was 81%. The mean polar angle was 3.3° ± 34°, and the radial limit of agreement was 68°.

Figure 3
Upper panel: 4-quadrant plot and polar plot analysis showing concordance rate, mean polar angles, and radial limits of agreement between delta venous minus arterial CO₂ partial pressure gradients measured from mixed venous and central venous blood samples (AP_v-aCO₂, AP_cv-aCO₂). Lower panel: 4-quadrant plot and polar plot analysis showing concordance rate, mean polar angles, and radial limits of agreement between delta venous minus arterial CO₂ partial pressure/arterial minus venous oxygen content ratios measured from mixed venous and central venous blood samples. (AP_v-aCO₂/C_a-vO₂, AP_cv-aCO₂/C_a-cvO₂).

Discussion

First, cardiovascular and metabolic alterations at different time intervals during cardiac surgery were not followed by significant changes in the P_v-aCO₂ gradient, and the only significant difference was a decrease in the P_v-aCO₂/C_v-aO₂ ratio at the end of the study.

Second, the P_v-aCO₂ gradient and P_v-aCO₂/C_v-aO₂ ratio were weak but significantly associated with factors that affect the CO₂ hemoglobin dissociation curve. Furthermore, the dependence of these factors varied over time. Finally, we evidenced poor agreement between central and mixed venous calculations.

As frequently observed during cardiac surgery, patients developed a hyperdynamic pattern.¹⁶ In this context, lactic acidosis was not related to a low cardiac output state but was more likely to be an imbalance between oxygen delivery and consumption. Persistent hyperlactatemia could also result from accelerated aerobic glycolysis under the effects of endogenous or exogenous catecholamines.¹⁷, ¹⁸, ¹⁹

P_v-aCO₂ and P_v-aCO₂/C_a-vO₂ as hemodynamic and metabolic markers

Hemodynamic and metabolic derangements were not accompanied by significant P_v-aCO₂ gradient alterations. Our findings agree with other studies that showed a weak correlation between the P_v-aCO₂ difference and some physiologic variables (CI, pH, Hb, BE, and S_vO₂).²⁰, ²¹ It has been shown that this variable was not able to detect significant changes in systemic blood flow and global O₂-derived metabolism in cardiac surgery patients.¹, ²² Several confounding factors may influence the CO₂ hemoglobin dissociation curve and the relationship between PCO₂ and CCO₂ in venous blood in unstable critically ill patients.²³, ²⁴ Accordingly, the coefficients of determination between P_v-aCO₂ and Hb, pH, BE, lactate, and S_vO₂ were significant but weak, either when calculated globally or separately, at each time interval. Furthermore, the multilevel linear regression model identified BE and S_vO₂ as weak but significant determinants of the P_v-aCO₂ gradient.

The P_v-aCO₂/C_a-vO₂ ratio also remained relatively stable compared to baseline until the end of the study when there was a significant reduction. Until this time, ratios were greater than the threshold described (> 1.4), indicating the onset of anaerobic metabolism. However, this issue remains controversial, and some studies failed to identify such an association after blood transfusion in hemorrhagic shock. VO₂ and the RQ were measured by analysis of normalized expired gases, but P_v-aCO₂/C_a-vO₂ remained high perhaps due to persistent hyperlactatemia.²⁵ The P_v-aCO₂/C_a-vO₂ ratio is a composite calculation affected by many pathophysiological changes. The same considerations mentioned for the P_v-aCO₂ gradient could have affected the capacity of the ratio to identify tissue hypoxia. In this case, a weak but significant correlation was also found with S_vO₂, CI, Hb, arterial pH, and BE. The multilevel linear regression model revealed that Hb, S_vO₂, and BE were the main determinants affecting the ratio.²⁶ Transient changes in Hb concentration affected P_v-aCO₂/C_a-vO₂ independent of the occurrence of anaerobic metabolism. Hemodilution may affect the P_v-aCO₂/C_v-aO₂ gradient through changes in CO₂ dissociation from Hb. Anemic hypoxia increases oxygen extraction and may result in reductions in C_v-aO₂. Thus, one mechanism can explain the effect of Hb changes on P_v-aCO₂, whereas two mechanisms can affect the P_v-aCO₂/C_v-aO₂ ratio.

S_vO₂ and BE also affect the balance between dissolved and combined CO₂. Furthermore, lactate was a significant determinant of the P_v-aCO₂ gradient and P_v-aCO₂/C_v-aO₂ ratio during T2 and T3 when lactic acidosis was present. The changing behavior of the correlations at different time points further reinforces the concept that both are more dependent on the variables that modify the dissociation of CO₂ from Hb compared with CI or DO₂.

During CPB, the CO₂ hemoglobin dissociation curve shifts downward in both arterial and venous blood, affecting CO₂ transport even after CPB ends.²⁷ A primary determinant of this change is hemodilution.¹⁰ Restoration of blood CO₂ transport capacity does not occur immediately after hemoglobin correction. Among other factors, metabolic acidosis, changes in body temperature, and the Haldane effect could shift the CO₂ hemoglobin dissociation curve and the relationship between PCO₂ and CCO₂.²⁷

During CPB, the temperature decreased to 32–33°C followed by a rewarming phase up to 37°C at the end of the procedure. Thus, by the time these patients were evaluated, body temperature was within the normal range. Hypothermia increases CO₂ solubility, and rewarming might cause the release of dissolved CO₂ from the tissues, also affecting the P_v-aCO₂ gradient.²⁸ Sudden changes in body temperature affect VO₂, CO₂ production, and transport, and these alterations could remain several hours after surgery.

The hyperdynamic state and vasoactive/inotropic drug infusion could be associated with the maldistribution of blood flow, and the heterogeneous circulation could slow or impair CO₂ removal from peripheral tissues.²⁹

By the end of the study, sedoanalgesia and mechanical ventilation were gradually diminished, and spontaneous breathing recovered. The higher VO₂ with increased EO₂ from T3 may represent increased O₂ demands and trend with decreased S_vO₂ and increased C_v-aO₂. By this time, arterial pH and BE were within normal ranges, suggesting preserved systemic aerobic metabolism.³⁰

Lack of agreement between central and mixed venous blood CO₂-derived parameters

An additional source of error should be considered when the analysis of SO₂ and the CO₂-derived variables are made from central venous blood samples.

Bland-Altman limits of agreement were large enough to make them unacceptable for clinical decisions. A Fourquadrant plot demonstrated a weak concordance rate (73%) for the central and mixed venous delta PCO₂ differences. The concordance rate for the delta PCO₂/CO₂ ratio was 81%, which is in the limit of concordance acceptance. When completing the study with the polar plot method, the radial limits of agreement were extremely high for both variables. These findings along with the initial Bland-Altman approach confirm that trends between these variables were not interchangeable. Cardiac surgery courses with sudden and significant metabolic and hemodynamic changes that may differently affect the upper part of the body, including the central nervous system, compared to the infradiaphragmatic region, mainly the splanchnic area.¹¹

Limitations of the study. The retrospective characteristics of the analysis may represent a limitation. We recognized that RQ is the gold standard used to identify the onset of anaerobic metabolism, and P_v-aCO₂/C_v-aO₂ is a proper surrogate, not P_cv-aCO₂/C_a-cvO₂.

Nevertheless, P_cv-aCO₂/C_a-cvO₂ is typically used for this purpose. Thus, the goal of this study was to show the poor agreement between the ratio calculated from either mixed venous or central venous samples. We agree that a proper analysis should consider the CO₂ contents instead of pressures. Although these calculations were performed, it should be emphasized that any algorithm for CO₂ content calculations has severe limitations. Douglas et al.¹⁴ reported an excellent correlation between measured and calculated CO₂ contents. Despite this, the corresponding bias and 95% limits of agreement between the measured and calculated CO₂ contents were 0.02 and 4.66 mL/100 mL, respectively. Consequently, the measured and calculated CO₂ contents are not interchangeable. This explains for the frequent negative values of calculated C_v-aCO₂ found with the Douglas formula.

Information about body temperature was not available and precluded a more precise analysis of arterial and venous CO₂ content. Finally, the number of cases was relatively low.

Conclusions

In this population, the P_v-aCO₂ gradient and P_v-aCO₂/C_a-vO₂ ratio did not change significantly throughout the study and were dependent on the effects of changing physiological conditions. Many pathophysiological changes could affect the relationship between PCO₂ and CCO₂, making these measurements less sensitive to changes in systemic blood flow. Additionally, simultaneous measurements made from central and mixed venous blood showed poor agreement. Therefore, CO₂-derived variables should be cautiously used to guide hemodynamic support and to monitor tissue oxygenation during cardiac surgery.

References

• ¹
  Pölönen P, Ruokonen E, Hippeläinen M, et al. A prospective, randomized study of goal-oriented hemodynamic therapy in cardiac surgical patients. Anesth Analg. 2000;90:1052–9.
• ²
  Tripodaki ES, Tasoulis A, Koliopoulou A, et al. Microcirculation and macrocirculation in cardiac surgical patients. Crit Care Res Pract. 2012;2012:654381.
• ³
  Schumacker PT, Cain SM. The concept of a critical oxygen delivery. Intensive Care Med. 1987;13:223–9.
• ⁴
  Zhang Z, Xu X. Lactate clearance is a useful biomarker for the prediction of all-cause mortality in critically ill patients: a systematic review and meta-analysis. Crit Care Med. 2014;42:2118–25.
• ⁵
  Vallet B, Teboul JL, Cain S, et al. Venoarterial CO2 difference during regional ischemic or hypoxic hypoxia. J Appl Physiol. 2000;89:1317–21.
• ⁶
  Mallat J. Use of venous-to-arterial carbon dioxide tension difference to guide resuscitation therapy in septic shock. World J Crit Care Med. 2016;5:47.
• ⁷
  Lamsfus-Prieto J, de Castro-Fernández R, Hernández-García AM, et al. Valor pronóstico de los parámetros gasométricos del dióxido de carbono en pacientes con sepsis. Una revisión bibliográfica. Rev Esp Anestesiol Reanim. 2016;63:220–30.
• ⁸
  Mesquida J, Saludes P, Gruartmoner G, et al. Central venous-to-arterial carbon dioxide difference combined with arterial-to-venous oxygen content difference is associated with lactate evolution in the hemodynamic resuscitation process in early septic shock. Crit Care. 2015;19:126.
• ⁹
  Ducey JP, Lamiell JM, Gueller GE. Arterial-venous carbon dioxide tension difference during severe hemorrhage and resuscitation. Crit Care Med. 1992;20:518–22.
• ¹⁰
  Dubin A, Ferrara G, Kanoore Edul VS, et al. Venoarterial PCO2-to-arteriovenous oxygen content difference ratio is a poor surrogate for anaerobic metabolism in hemodilution: an experimental study. Ann Intensive Care. 2017;7:65.
• ¹¹
  Riva JA, Bouchacourt JP, Kohn WE, et al. Las tendencias en el tiempo de las saturaciones de oxígeno en la vena cava superior y la arteria pulmonar no son equivalentes en cirugía cardiaca. Rev Esp Anestesiol Reanim. 2015;62:140–4.
• ¹²
  Chawla LS, Zia H, Gutierrez G, et al. Lack of equivalence between central and mixed venous oxygen saturation. Chest. 2004;126:1891–6.
• ¹³
  Gutierrez G, Comignani P, Huespe L, et al. Central venous to mixed venous blood oxygen and lactate gradients are associated with outcome in critically ill patients. Intensive Care Med. 2008;34:1662–8.
• ¹⁴
  Douglas AR, Jones NL, Reed JW. Calculation of whole blood CO₂ content. J Appl Physiol. 1988;65:473–7.
• ¹⁵
  Critchley LA, Yang XX, Lee A. Assessment of trending ability of cardiac output monitors by polar plot methodology. J Cardiothorac Vasc Anesth. 2011;25:536–46.
• ¹⁶
  Fischer GW, Levin MA. Vasoplegia during cardiac surgery: current concepts and management. Semin Thorac Cardiovasc Surg. 2010;22:140–4.
• ¹⁷
  Minton J, Sidebotham DA. Hyperlactatemia and cardiac surgery. J Extra Corpor Technol. 2017;49:7–15.
• ¹⁸
  Naik R, George G, Karuppiah S, et al. Hyperlactatemia in patients undergoing adult cardiac surgery under cardiopulmonary bypass: causative factors and its effect on surgical outcome. Ann Card Anaesth. 2016;19:668–75.
• ¹⁹
  Haanschoten MC, Kreeftenberg HG, Arthur Bouwman R, et al. Use of postoperative peak arterial lactate level to predict outcome after cardiac surgery. J Cardiothorac Vasc Anesth. 2017;31:45–53.
• ²⁰
  Heinze H, Paarmann H, Heringlake M, et al. Measurement of central and mixed venous-to-arterial carbon dioxide differences in cardiac surgery patients. Appl Cardiopulm Pathophysiol. 2011;15:29–37.
• ²¹
  Morel J, Grand N, Axiotis G, et al. High veno-arterial carbon dioxide gradient is not predictive of worst outcome after an elective cardiac surgery: a retrospective cohort study. J Clin Monit Comput. 2016;30:783–9.
• ²²
  Abou-Arab O, Braik R, Huette P, et al. The ratios of central venous to arterial carbon dioxide content and tension to arteriovenous oxygen content are not associated with overall anaerobic metabolism in postoperative cardiac surgery patients. PLoS One. 2018;13:1–11.
• ²³
  Dubin A, Estenssoro E, Murias G, et al. Intramucosal-arterial PCO2 gradient does not reflect intestinal dysoxia in anemic hypoxia. J Trauma. 2004;57:1211–7.
• ²⁴
  Jakob SM, Kosonen P, Ruokonen E, et al. The haldane effect – an alternative explanation for increasing gastric mucosal PCO2 gradients? Br J Anaesth. 1999;83:740–6.
• ²⁵
  Ferrara G, Edul VSK, Canales HS, et al. Systemic and microcirculatory effects of blood transfusion in experimental hemorrhagic shock. Intensive Care Med Exp. 2017;5:24.
• ²⁶
  Dubin A, Pozo MO, Kanoore Edul VS, et al. Poor agreement in the calculation of venoarterial PCO2 to arteriovenous O2 content difference ratio using central and mixed venous blood samples in septic patients. J Crit Care. 2018;48:445–50.
• ²⁷
  Cavaliere F Impaired carbon dioxide transport during and after cardiopulmonary bypass. Perfusion. 2000;15:433–9.
• ²⁸
  Cavaliere F, Martinelli L, Guarneri S, et al. Arterial-venous PCO2 gradient in early postoperative hours following myocardial revascularization. J Cardiovasc Surg (Torino). 1996;37:499–503.
• ²⁹
  Carrel T Englberger L, Mohacsi P, et al. Low systemic vascular resistance after cardiopulmonary bypass: incidence, etiology, and clinical importance. J Card Surg. 2000;15:347–53.
• ³⁰
  Williams J, McLean A, Ahari J, et al. Decreases in mixed venous blood O2 saturation in cardiac surgery patients following extubation. J Intensive Care Med. 2020;35:264–9.

Publication Dates

• Publication in this collection
  23 Oct 2023
• Date of issue
  2023

History

• Received
  26 Oct 2020
• Accepted
  29 July 2021
• Published
  15 Aug 2021