nueva página del texto (beta)
Inglés (pdf)
Artículo en XML
Referencias del artículo
Traducción automática
Enviar artículo por email
Citado por SciELO 
Similares en
SciELO 
Permalink
Introduction
In recent decades, the survival of patients with congenital heart disease has improved significantly due to advances in perfusion techniques, anesthesia, cardiovascular surgery, and intensive care. However, the presence of low cardiac output syndrome (LCOS) remains one of the major determinants of post-operative morbidity1,2.
The decrease in cardiac output results in tissue hypoperfusion with increased tissue oxygen consumption. The main causes of decreased cardiac output after using a cardiopulmonary bypass are myocardial ischemia, ischemia-reperfusion injury, and the release of inflammatory mediators when blood contacts a foreign surface. Early detection of cardiac output and targeted therapies reduce the risk of tissue hypoxia and multiorgan dysfunction2.
Currently, the gold standard for measuring cardiac output are devices that use the thermodilution method, such as the Swan-Ganz catheter or the PICCO system. This technology also allows analysis of the contour of the invasive arterial pressure waveform. However, these devices are invasive and rarely available in intensive care units (ICUs) in low- and middle-income countries. For these reasons, several biomarkers have been used to assess tissue perfusion, such as central venous saturation, lactate, arteriovenous oxygen difference, and, more recently, arteriovenous CO2 difference1,3. In recent studies, the latter has been found to be associated with a higher risk of mortality and multiorgan dysfunction when it is > 6 mmHg in critically ill patients, either adults or children4,5.
Venous oxygen saturation (SvO2) reflects the balance between tissue oxygen delivery (DO2) and consumption (VO2). As oxygen availability decreases, cellular oxygen extraction increases to maintain energy production regardless of DO2. However, there is a "critical point" of oxygen delivery beyond which energy cannot be produced by cellular ATP formation, and VO2 becomes dependent on DO2, causing lactic acidosis5,6.
As oxygen extraction increases, venous saturation gradually decreases. Although the most common cause of decreased SvO2 is tissue hypoxia secondary to inadequate perfusion, other conditions, such as poor oxygenation or oxygen transport failure, can also affect SvO26-8. Other biomarkers, such as arterial lactate, are useful in the immediate post-operative period to assess the state of tissue perfusion. Arterial lactate is more useful when its behavior is analyzed over time. However, its disadvantage is that it can be elevated in other conditions, such as alkalosis and hyperglycemia, in addition to being a late marker of tissue hypoperfusion9.
Several studies have reported the utility of arteriovenous CO2 difference (ΔCO2) as a biomarker of tissue ischemia. Its main advantage is that it can be modified in the presence of tissue hypoperfusion, even when venous saturation and lactate remain normal6-8. Tissue CO2 elimination capacity is a function of cardiac output and tissue blood flow. According to the Fick equation, oxygen consumption (VO2) equals the product of cardiac output and arteriovenous O2 content difference (CaO2-CvO2). Similarly, CO2 consumption (VCO2) is the product of cardiac output and the difference in arteriovenous CO2 content10.
Decreased tissue blood flow (ischemic hypoxia) is associated with venous hypercapnia due to cellular anoxia resulting in increased CO2 production11. The increase in arterial CO2 depends on gas exchange and, because it is very efficient at the alveolar level, it is not significantly affected by decreased pulmonary flow. Hence, an increase in the CO2 difference is indicative of tissue hypoperfusion. Under physiological conditions, with adequate venous and systemic flow, the difference between arterial and venous CO2 should not exceed 6 mmHg12.
In adult patients, arteriovenous oxygen difference may be a good adjunct to hemodynamic assessment even when venous saturation is normal4,5.
This study aimed to analyze the relationship between ΔCO2 and post-operative evolution in pediatric patients undergoing surgery for congenital heart disease with an extracorporeal circulation pump and its correlation with DavO2, SvO2, and lactate.
Methods
Design
We conducted a longitudinal study in the cardiovascular intensive care unit (UCICV, for its Spanish acronym) of the INP. Patients from the neonatal stage to 18 years of age who underwent cardiovascular surgery for congenital malformations with an extracorporeal circulation pump between March 01, 2019, and February 28, 2021, were included in the study. Patients in whom sampling could not be completed were excluded from the study.
Data collection
Demographic variables included in the study were age, sex, weight, and cardiologic diagnosis. Surgical variables were surgery performed and surgical complexity according to RACH-1 and extended Aristotle scales13,14, extracorporeal circulation time, and aortic clamping time. Finally, the biochemical variables in the UCICV were arteriovenous oxygen difference (DavO2 = CaO2-CvO2)15, oxygen extraction index (IEO2 = DavO2/CaO2 × 100)15, ΔCO2 (venous CO2-arterial CO2), central venous saturation, and arterial lactate. For the calculation of arteriovenous difference and ΔCO2, arterial and venous blood gases were obtained from the central venous catheter. All measurements were collected on four moments: on admission to the UCICV, 6, 12, and 24 h later. The inotropic score was also calculated for each of these moments using the Wernosky formula16.
The outcome variables recorded were the presence of LCOS, multiple organ failure (dysfunction of two or more organs)17, time on mechanical ventilation, days of hospitalization, and mortality in the UCICV. LCOS was diagnosed based on clinical assessment of the following variables: tachycardia, capillary refill time > 2 s, weak pulses, marmoreal skin, urine output < 1 mL/kg/h, venous saturation < 60%, lactate > 3 mmol/L, and need for inotropic support.
Analysis
Data were collected in Excel format (Microsoft Office 2022) and then exported to STATA 17.0 (StataCorp) for statistical analysis. Qualitative variables were presented as frequencies and proportions, while quantitative variables were presented as medians and interquartile ranges. The statistical tests used to identify associations were χ2 and Wilcoxon for qualitative and quantitative variables, respectively. In addition, Pearson correlation was used for continuous variables, specifically the relationship between ΔCO2 and the other tissue hypoperfusion variables. Finally, logistic regression was used to estimate the odds ratio (OR) of the combined hypoperfusion variables. In all cases, p < 0.05 was considered statistically significant.
Results
Eighty-two patients were included in the study: 61% were male. The median age was 17 months. Fifty-nine percent of patients had ΔCO2 ≥ 6 mmHg. Patients with risk adjustment in congenital heart surgery system surgical complexity ≥ 3 had ΔCO2 ≥ 6 mmHg. The median cardiopulmonary bypass (CPB) time was 120 min, with no significant difference between patients with ΔCO2 > 6 mmHg or patients with ΔCO2 < 6 mmHg (Table 1).
Table 1 Demographic characteristics of patients who underwent cardiovascular surgery for congenital malformations
| Variables | Total | ΔCO2 < 6 mmHg | ΔCO2 ≥ 6 mmHg | p-value |
|---|---|---|---|---|
| Patients | 100% (82) | 41% (34) | 59% (48) | - |
| Male | 61% (50) | 59% (20) | 63% (30) | 0.737 |
| Age (months) | 17 (3-51) | 12 (3-49) | 19 (6-60) | 0.496 |
| Age < 1 year | 41% (34) | 47% (16) | 38% (18) | 0.387 |
| Univentricular physiology | 23% (19) | 24% (8) | 23% (11) | 0.948 |
| RACHS ≥ 3 | 57% (47) | 50% (17) | 63% (30) | 0.260 |
| Aristotle ≥ 3 | 46% (38) | 41% (14) | 50% (24) | 0.430 |
| ECT (minutes) | 120 (79-171) | 115 (79-176) | 121 (81-171) | 0.865 |
| Aortic clamping (minutes) | 82 (37-111) | 82 (33-124) | 83 (38-110) | 0.984 |
ECT: extracorporeal circulation time; RACHS: risk adjustment in congenital heart surgery system.
As shown in table 2, inotropic score ≥ 5 (p < 0.001), lactate ≥ 2 mmol/L (p = 0.027), and DavO2 > 5 (p = 0.048) were significantly associated with ΔCO2 ≥ 6 mmHg. Patients with a higher ΔCO2 showed a longer UCICV stay (p = 0.043) and a higher frequency of multiorgan dysfunction (p = 0.073) but not a longer mechanical ventilation time (p = 0.627) (Table 2). Patients in whom ΔCO2 persisted above 6 mmHg at 12 h after admission to the UCICV also had higher IEO (p = 0.057) and persisted with higher inotropic values (p = 0.085).
Table 2 Tissue perfusion variables on admission to the UCICV and outcome variables
| Variables | Total (n = 82) | ΔCO2 < 6 mmHg on admission (n = 34) | ΔCO2 ≥ 6 mmHg on admission (n = 48) | p-value |
|---|---|---|---|---|
| Lactate ≥ 2 (mmol/L) | 74% (61) | 62% (21) | 83% (40) | 0.027 |
| Venous saturation ≤ 60 | 39% (32) | 35% (12) | 42% (20) | 0.560 |
| DavO2 > 5 | 51% (42) | 38% (13) | 60% (29) | 0.048 |
| IEO2 > 25 | 79% (65) | 71% (24) | 85% (41) | 0.166 |
| Inotropic score ≥ 5 | 55% (45) | 29% (10) | 73% (35) | < 0.001 |
| UCICV stay | 10 (6-21) | 7.5 (4-19) | 13 (7-27.5) | 0.043 |
| Mechanical ventilation time | 144 (82.5-504) | 120 (72-480) | 144 (96-528) | 0.627 |
| Extubation on the ward | 37% (30) | 44% (15) | 31% (15) | 0.233 |
| Acute kidney injury | 28% (23) | 26% (9) | 29% (14) | 0.789 |
| Multiple organ failure | 11% (9) | 3% (1) | 17% (8) | 0.073 |
| Low cardiac output syndrome | 51% (62) | 56% (19) | 67% (32) | 0.321 |
| Ischemia | 5% (4) | 3% (1) | 6% (3) | 0.638 |
| Arrhythmias | 29% (24) | 24% (8) | 33% (16) | 0.336 |
| Bleeding | 13% (11) | 9% (3) | 17% (8) | 0.348 |
| Infection | 40% (33) | 29% (10) | 48% (23) | 0.092 |
| Arrest | 6% (5) | 3% (1) | 8% (4) | 0.397 |
| Death in UCICV | 4.8% (4) | 0% (0) | 8.3% (4) | 0.138 |
| Hospital death | 6.1% (5) | 3% (1) | 8.3% (4) | 0.397 |
DavO2: arteriovenous oxygen difference; IEO2: oxygen extraction index; UCICV: cardiovascular intensive care unit.
A stronger correlation was found between ΔCO2 and tissue hypoperfusion variables such as lactate (r = 0.59, p < 0.001) and DavO2 (r = 0.28, p = 0.009) at 12 h after admission (Table 3). Similarly, when analyzing the total number of samples collected (n = 326), a statistically significant direct correlation was observed between ΔCO2 and arterial lactate (p < 0.001), DavO2 (p < 0.009), and inotropic score (p < 0.001). In contrast, an inverse correlation, also statistically significant, was found between ΔCO2 and venous oxygen saturation (r = 0.32, p < 0.001) (Table 4).
Table 3 Correlation 12 h after admission between lactate, DavO2, inotropic score, and venous saturation with ΔCO2
| At 12 h (n = 81) | ΔCO2 (r) | p-value |
|---|---|---|
| Lactate | 0.59 | < 0.001 |
| DavO2 | 0.28 | 0.009 |
| Inotropic score | 0.52 | < 0.001 |
| Venous saturation | −0.28 | 0.009 |
DavO2: arteriovenous oxygen difference.
Table 4 Correlation among, lactate, DavO2, inotropic score, and venous oxygen saturation of collected samples
| Total (n = 326) | ΔCO2 (r) | p-value |
|---|---|---|
| Lactate | 0.25 | < 0.001 |
| DavO2 | 0.22 | < 0.001 |
| Inotropic score | 0.36 | < 0.001 |
| Venous saturation | -0.32 | < 0.001 |
DavO2: arteriovenous oxygen difference.
When the combined analysis between tissue hypoperfusion variables was performed, patients with ΔCO2 > 6 mmHg and lactate > 2 mmol/L at the time of ICU admission were 9.7 times more likely to develop multiorgan dysfunction. Similarly, patients with ΔCO2 > 6 mmHg and an inotropic score ≥ 10 were 12.6 times more likely to die (Table 5).
Table 5 Combined analysis between tissue hypoperfusion variables, complications, and outcomes
| Combined variables | Risk factor | Proportion | p-value | OR | 95% CI |
|---|---|---|---|---|---|
| ΔCO2 ≥ 6 + Lactate ≥ 2 | Infection | 54% versus 27% | 0.013 | 3.2 | 1.2-7.9 |
| Multiple organ failure | 20% versus 2% | 0.029 | 9.7 | 1.15-81 | |
| ΔCO2 ≥ 6 + score ≥ 10 | Death in ICU | 17% versus 2% | 0.033 | 12.6 | 1.2-129 |
| ΔCO2 ≥ 6 + DavO2 ≥ 5 | Infection | 60% versus 29% | 0.01 | 3.7 | 1.4-9.5 |
CI: confidence interval; DavO2: arteriovenous oxygen difference; ICU: intensive care unit; OR: odds ratio.
There were four deaths, all of them in the group of patients with ΔCO2 > 6 mmHg at UCICV admission (Table 2 and Figure 1).
Discussion
Although the determination of ΔCO2 is relatively easy to perform in the ICU, few studies have analyzed its relevance as a biomarker of tissue hypoperfusion in post-operative pediatric patient with congenital heart disease. Most studies have been performed in patients with sepsis. Ospina-Tascón et al.18 found that patients with septic shock who continued to have ΔCO2 > 6 mmHg at 6 h after initiating fluid and vasoactive support had higher SOFA severity scale scores and higher mortality (p > 0.0001). Lactate clearance was slower in this group of patients (p < 0.001). However, in this study, the correlation between cardiac index measured by thermodilution and ΔCO2 was weak (r = 0.25, p < 0.01). In pediatric patients with severe sepsis, Fernández-Sarmiento et al. also found no correlation between ejection fraction measured by transthoracic echocardiography and ΔCO2 (r = 0.13)11. In contrast, in a group of adult cardiovascular surgery patients, Takami and Masumoto found a moderate correlation between cardiac index and ΔCO2 (r = 0.35, p < 0.001) and between venous saturation and ΔCO2 (r = 0.35, p < 0.001)10.
Furqan et al. found that ΔCO2 > 6 mmHg was also associated with low venous saturation in pediatric patients undergoing surgery for congenital heart disease. Both studies concluded that ΔCO2 could be used as a good predictor of cardiac index in patients undergoing CPB surgery (r = 0.47, p = 0.0011)19, observations that are consistent with our findings showing that 67% of patients with ΔCO2 > 6 mmHg had LCOS.
Venous saturation is a surrogate for cardiac output and can be used as a biomarker of tissue hypoperfusion to guide treatment goals in the septic shock patient. However, an adequate correlation between ΔCO2 and SvO2 has not been demonstrated in these patients. In a study conducted by Mallat et al. in adults with septic shock, patients in whom ΔCO2 persisted > 6 mmHg, even when venous saturation was optimized to above 70%, continued to have lactate > 2 mmol/L, allowing identification of those patients in whom perfusion was still suboptimal6. These findings could be explained by the fact that ΔCO2 is a biomarker of ischemic hypoxia and is not affected by changes in oxygen availability or blood hemoglobin levels, as is the case with venous saturation. Interestingly, in our group of patients, the correlation found between ΔCO2 and venous saturation was adequate (r = −0.32, p < 0.001), similar to that described by Furqan et al. (OR = 0.340)19. In this group of patients, ΔCO2 could be used with venous saturation to assess systemic perfusion.
Among other markers of tissue hypoperfusion, our study found that patients with ΔCO2 > 6 mmHg had lactate > 2 mmol/L and arteriovenous oxygen differences > 5 on admission (p = 0.027, p = 0.048), findings similar to those described by Rhodes et al. in pediatric patients undergoing cardiovascular surgery. These authors found a good correlation between ΔCO2 and DavO2 (r = 0.55, p < 0.01)1, suggesting that ΔCO2 can be used in conjunction with DavO2 as a biomarker of tissue ischemia during the hemodynamic assessment of patients with suspected low cardiac output. However, these authors found no correlation between ΔCO2 and lactate (r = 0.02, p = 0.85), in contrast to our study, where patients with ΔCO2 > 6 mmHg had higher lactate (p = 0.027). With these findings, we suggest that ΔCO2 can be used as a tool for early detection of tissue hypoperfusion.
Regarding ventilatory support, we found that patients with ΔCO2 > 6 mmHg did not have longer mechanical ventilation time (p = 0.627) or longer hospital stay (p = 0.49), results similar to a study by Akamasu et al. in 114 pediatric patients who underwent extracorporeal circulation pump surgery for congenital heart disease. There were no differences between patients with ΔCO2 under or over 6 mmHg on admission (p = 0.80)20. In contrast, we found an association between ΔCO2 > 6 mmHg and a longer stay in the cardiovascular ICU (p = 0.043).
Several studies have analyzed whether there is an association between ΔCO2 > 6 mmHg and patient outcome. In the study by Rhodes et al.1, patients who had an inotropic score > 15 on admission, at least one cardiorespiratory arrest event, an unplanned surgical reintervention, or the need for extracorporeal membrane support had ΔCO2 > 6 mmHg (mean 8.3 mmHg). In the same study, those who did not have any of these events had a mean ΔCO2 of 5.6 mmHg. Insom et al. also found that ΔCO2 > 6 mmHg was associated with a higher inotropic score (r = 0.21, p = 0.02)21.
Logistic regression analysis determined that ΔCO2 > 6 mmHg at admission was independently associated with higher mortality OR 1.3 (95% CI [confidence interval], 1.07-1.31), similar to Mukai et al. in patients undergoing on-pump cardiac surgery (area under the curve 0.804, 95% CI, 0.688-0.921)22. In our study, patients with admission ΔCO2 > 6 mmHg had higher inotropic scores (p < 0.001) and a higher incidence of multiorgan failure: eight patients from the ΔCO2 > 6 mmHg group compared to only one patient from the ΔCO2 < 6 mmHg group.
Patients with ΔCO2 > 6 mmHg and inotropic score ≥ 10 were 12.6 times more likely to die. Unfortunately, four patients died, all from the ΔCO2 > 6 mmHg group.
ΔCO2 is a useful marker of tissue hypoperfusion. Patients with ΔCO2 > 6 mmHg on admission have higher lactate, DavO2, and inotropic values. Early monitoring of ΔCO2 is a useful biomarker to identify patients at increased risk of post-operative morbidity and mortality, especially when accompanied by a high inotropic score.
