Utility of venous-to-arterial carbon dioxide changes to arteriovenous oxygen content ratios in the prognosis of severe sepsis and septic shock: A systematic review and meta-analysis

Abstract

Background:

Sepsis patients with insufficient tissue perfusion and hypoxia should be identified and resuscitated immediately. Recently, venous-to-arterial carbon dioxide pressure changes and the arteriovenous oxygen content difference ratio (Pcv-aCO2/Ca-vO2) as a predictor of tissue perfusion recovery and poor prognosis.

Objectives:

Pcv-aCO2/Ca-vO2 is a substitute for respiratory entropy, the elevation of which indicates a lack of tissue perfusion. Pcv-aCO2/Ca-vO2 can be used as an indicator to predict the prognosis of patients with sepsis or septic shock, but its prognostic value has not been fully evaluated. Here, we have performed a meta-analysis to assess its predictive value for mortality.

Methods:

Meta-analysis of Observational Studies in Epidemiology group guidelines were followed for this meta-analysis. We searched the comprehensive electronic databases of PubMed, EMBASE, Web of Science, and Cochrane libraries from inception to March 2019, using the terms including “venous-arterial,” “carbon dioxide,” “Shock, Septic,” and related keywords. The Newcastle-Ottawa scale was used for quality evaluation of the literature. A meta-analysis was performed using RevMan 5.3 and Stata 14.0 software to evaluate the effects of Pcv-aCO2/Ca-vO2 on short-term mortality, sequential organ failure assessment, and acute physiology and chronic health evaluation scores in patients with sepsis or septic shock.

Results:

The final analysis included 13 clinical studies involving a total of 940 subjects. The results of the meta-analysis showed that non-surviving patients had higher Pcv-aCO2/Ca-vO2 than survivors after fluid resuscitation (standardized mean difference = 0.68, 95% confidence interval = 0.24–1.12) and blood samples taken 6 h after resuscitation showed a greater risk of mortality (risk ratio = 1.89, 95% confidence interval = 1.48–2.41) and sequential organ failure assessment scores (mean difference = 1.58, 95% confidence interval = 0.88–2.28, P < 0.01) in patients with high Pcv-aCO2/Ca-vO2. These differences were statistically significant.

Conclusion:

This meta-analysis indicates that Pcv-aCO2/Ca-vO2 has predictive value for mortality in patients with sepsis or septic shock. Further studies are now required to determine the optimal threshold for predicting sepsis mortality.

Prospero Registration:

The protocol for this systematic review was registered on PROSPERO (CRD 42019128134).

Keywords

Pcv-aCO2/Ca-vO2 sepsis septic shock meta-analysis mortality

Introduction

Sepsis is defined as life-threatening organ dysfunction caused by dysregulation or infection.¹ Septic shock is a branch of sepsis with circulatory, cellular, and metabolic abnormalities that is associated with a higher risk of death than sepsis alone.^1,2 The primary management of septic shock patients is to correct hypoxemia and establish venous access for the early use of fluid replacement and antibiotic treatment.³ In recent years, sepsis associated mortality has decreased^4–6 but remains prevalent. In recent studies, sepsis-related mortality was reported as ⩾ 10%, and the mortality associated with septic shock was ⩾ 40%.¹ An array of indicators can predict the prognosis of these patients. In daily care, metabolic variables such as lactic acid and central venous oxygen saturation (ScvO2) are often used as substitutes for tissue hypoxia. Current resuscitation strategies rely on the rapid normalization of these parameters. Recently, venous-to-arterial carbon dioxide changes and the arteriovenous oxygen content difference ratio (Pcv-aCO2/Ca-vO2) is also an indicator.⁷ The relevant calculations of this value including:

\begin{array}{l} CaO 2 = 1.34 \times Hb \times SaO 2 + 0.003 \times PaO 2 \\ CvO 2 = 1.34 \times Hb \times ScvO 2 + 0.003 \times PvO 2 \end{array}

Pcv-aCO2/Ca-vO2 reflects the levels of respiratory entropy according to the Fick equation.⁸ The literature found that compared with lactic acid, Pcv-aCO2, and ScvO2, elevated Pcv-aCO2/Ca-vO2 in patients with septic shock is an accurate marker of global anaerobic metabolism, and high ratios are associated with poor prognosis.^9
–11 Changes in Pcv-aCO2/Ca-vO2 are more sensitive, so it becomes an attractive monitoring indicator. Gao et al.¹² believed that compared with the single application, the combined application of Pcv-aCO2/Ca-vO2 and lactic acid clearance can dynamically evaluate tissue perfusion to achieve precise treatment and improve patient prognosis. However, select studies have reported a lack of association of Pcv-aCO2/Ca-vO2 with patient mortality.^13,14 A randomized controlled trial found that Pcv-aCO2/Ca-vO2 guided resuscitation did not improve the prognosis of patients with sepsis or septic shock.¹⁵ Here, we performed a systematic review and meta-analysis to assess the prognostic value of Pcv-aCO2/Ca-vO2 in patients with sepsis or septic shock.

Material and methods

Literature search and study selection

This meta-analysis was conducted according to the Guidelines of the Meta-analysis of Observational Studies in Epidemiology group (MOOSE). Two researchers (WM and LTY) independently searched four databases including PubMed, EMBASE, the Web of Science and The Cochrane Library from inception to March 2019. The search terms included “venous-arterial,” “carbon dioxide,” “Shock, Septic” and related keywords. No restrictions on regions, publication types or languages were observed. According to the different databases, the terms were combined with keywords or free words for retrieval. References of all relevant articles were hand searched to uncover other potentially eligible studies.

Studies that met the following criteria were included: (1) observational studies (including prospective and retrospective studies); (2) sepsis or septic shock patients^1,16 (⩾ 18 years old); (3) the primary endpoint of the study: association of death with Pcv-aCO2/Ca-vO2 levels. Exclusion criteria included: letters, commentaries, case reports, conferences, abstracts, expert opinions, in vitro studies, animal experiments, and editorials. Studies in which the full text was unavailable or the original study data could not be converted were also excluded.

Data extraction and quality assessment

The two investigators (NZ and ZJZ) screened the studies according to title and abstract. For articles that met the inclusion criteria, the full texts were read and data were extracted according to previous data extraction terms. If disagreements occurred, differences were resolved by discussion, and the third reviewer (WM) judged the differences for which no solution was reached. When necessary, the original author of the article was contacted to obtain supplementary data. Continuous variables in the article, if not expressed as mean ± standard deviation, but expressed as the median and interquartile range, were converted by Luo et al.¹⁷ and Wan et al.’s¹⁸ methods. Extracted information included: first author name, year of publication, study design, patient’s number, study population characteristics, levels of Pcv-aCO2/Ca-vO2, reported outcomes, and information to assess the risk of bias.

Two authors (WM and NZ) used the Newcastle-Ottawa scale (NOS)¹⁹ to conduct risk bias assessments and quality assessment of the included articles without interference. Assessments consisted of three factors: representation of the study subjects, comparability between study groups, and the evaluation of outcome events. Risk scores for the scale ranged from 0–9. Studies with scores ⩾ 6 were considered low-risk. Differences were resolved through consultation with a third reviewer (LTY).

Statistical analysis

We performed quantitative analysis of the included studies when sufficient similarities were identified. Meta-analysis of the extracted data was analyzed using RevMan5.3 software provided by the Cochrane Collaboration and stata14.0. In other cases, descriptive analysis was used. For the categorical variables, the risk ratio (RR) was selected as the effect indicator, and the standardized mean differences (SMD) were used as the effect indicator of the continuous variable.²⁰ Prior to the meta-analysis, heterogeneity between the studies was assessed by the X² and I² statistic. If no heterogeneity was observed (P > 0.1, I² < 50%), a fixed-effect model for meta-analysis was performed. If heterogeneity was identified between studies (P < 0.1, I² ⩾ 50%), a random effects model was used for meta-analysis. Based on the characteristics of the included data, subgroup and sensitivity analysis were performed to eliminate heterogeneity and to calculate the combined effects. Funnel plots and the Egger et al.’s²¹ tests were used to detect publication bias.

Results

Study selection

A total of 334 articles were retrieved from the initial screening. The screening flow chart is shown in Figure 1. Duplicated literatures were removed. The abstracts and titles were screened according to inclusion and exclusion criteria, and 20 articles were obtained. Following the analysis of full texts, six studies were excluded as the outcome data failed to meet the inclusion requirements. A single study was excluded as it described before–after study data. Finally, the meta-analysis included 13 studies^{7,10,12
–14,22
–28} consisting of 940 participants.

Figure 1.

Flow chart of the literature selection process.

Quality assessment results

The NOS scores of the 13 studies are shown in Table 1. Scores ranged from 5 to 8 points. The overall quality of the studies was moderate, with two studies^13,28 having scores of 5, three studies^10,22,23 with scores of 6, five studies^{7,14,24
–26} with scores of 7, and three studies^12,27,29 with scores of 8. Since all subjects were diagnosed with sepsis or septic shock at the time of enrollment, no studies had scores of 4. No studies scored 9.

Table 1.

The NOS scores of the 13 studies.

Study	Selection (maximum****)				Comparability (maximum**)	Outcome (maximum***)			Total score (maximum 9)
	Representativeness of the exposed cohort	Selection of the non-exposed cohort	Ascertainment of exposure	Demonstration that outcome of interest was not present at start of study^a	Comparability of cohorts on the basis of the design or analysis	Assessment of outcome	Was follow-up long enough for outcomes to occur^b	Adequacy of follow-up of cohorts^c	Total score (maximum 9)
Dubin et al.^{13 ¤}	*	*			*	*		*	5
Gao et al.¹²	*	*	*		**	*	*	*	8
Gong et al.²⁹	*	*	*		**	*	*	*	8
He et al.^{22 ¤}	*	*	*		*	*		*	6
Mekontso-Dessap et al.¹⁰	*	*	*		*	*	*		6
Mesquida et al.²³	*	*	*		*	*		*	6
Mesquida et al.⁷	*	*	*		**	*		*	7
Ospina-Tascón et al.¹⁴	*	*	*		*	*	*	*	7
Rivera Solís et al.²⁷	*	*	*		**	*	*	*	8
Sánchez Nava et al.^{28 ¤}		*			**	*		*	5
Shaban et al.²⁴	*	*	*		*	*	*	*	7
Trejo Arteaga et al.26	*	*	*		*	*	*	*	7
Zhou et al.²⁵	*	*	*		**	*	*		7

All subjects enrolled in the study had sepsis or septic shock, so all studies did not have a star.

The follow-up time was set to 28 days, and the study was less than 28 days without a star.

The study of loss of follow-up rate >20% has no star.

The study used the quality assessment criteria for case–control studies in NOS, as detailed in Appendix 1.

Publication bias was explored using the Egger et al.’s²¹ test and funnel plots (P < 0.1 indicated significant bias; Figure 2). The included studies respectively contained Pcv-aCO2/Ca-vO2 levels for survivors and non-survivors (Figure 2(a)) and 28-day mortality in patients with high Pcv-aCO2/Ca-vO2 and low Pcv-aCO2/Ca-vO2 values (Figure 2(b)). The funnel plots and Egger’s test revealed no significant publication bias (P = 0.191), (P = 0.05).

Figure 2.

(a) Included studies containing Pcv-aCO2/Ca-vO2 levels for survivors and non-survivors and (b) included articles containing 28-day mortality in patients with high Pcv-aCO2/Ca-vO2 and low Pcv-aCO2/Ca-vO2.

Study characteristics

The characteristics of the 13 included studies are shown in Table 2. Of the 940 patients, 442 were males, 274 were females, with gender not reported for 224 subjects.^10,14 A single study²⁶ was obtained by searching the relevant literature. Other studies^{7,10,12
–14,22
–25,27
–29} were sourced from databases. All studies had English abstracts, 12 of which were published from 2015 to 2018.^{7,12
–14,22
–29} Three studies were published in Spanish,^26
–28 two were published in Chinese,^12,29 and eight were published in English.^{7,10,13,14,22
–25} The included studies were all observational, of which six were retrospective^{10,12,23,25,28,29} and seven were prospective.^{7,13,14,22,24,26,27} Excluding the study by Dubin and Mekontso-Dessap, the subjects were from intensive care units (ICU). Of the 13 standards-compliant studies, six used the 2013 guidelines to diagnose sepsis and septic shock,^{12,14,22,23,25,29} and one used the 2016 version⁷ which was unavailable in other studies. All subjects in these 10 studies were patients with sepsis or septic shock, of which 45.3% were pulmonary infections, with abdominal infections accounting for 23.0%. The source of urinary tract, blood flow, and soft tissue infections were 11.4%, 7.4%, and 4.8%, respectively.^{7,12
–14,22,23,25,27
–29} The other studies included other shock types including hemorrhagic and cardiogenic shock (Table 2), septic patients accounted for 63.6% of these study cohort.^10,24,26 The follow-up time differed between studies. Two studies were followed up for 30 days,^10,27 six studies were followed for 28 days,^{12,14,24
–26,29} four studies followed the ICU stay,^7,22,23,28 and one was followed up in hospital.¹³ Of the 13 studies, nine extracted the mean ± standard deviation of Pcv-aCO2/Ca-vO2 of surviving and non-surviving patients (Figure 3). A total of five studies extracted patient mortality data based on the cut-off value (Figure 4). We analyzed these two data sets independently.

Table 2.

Characteristics of the included literature.

Study	Study design	Source of infection (n)	Cutoff	Sample size (n)	Male (%)	Diagnostic Criteria	SOFA Day1	APACHE II	Sampling time	Follow-up (days)	Non-survivors (n,%)	RR	NOS score
Dubin et al.¹³	Prospective	Pneumonia 7; Urinary 4; Blood stream infection 2; Abdominal 9; Bone 1	NR	23	57	NR	10 ± 3	21 ± 6	MAP ⩾ 65 mmHg*	Hospital mortality	13 (57)	NR	5
Gao et al.¹²	Retrospective	Pneumonia 82; urinary 8; blood stream infection 11; soft tissues 5; abdominal 27; other 12	1.8	145	67	2001 SCCM/ESICM/ACCP/ATS/SIS	9.0 ± 3.7	18.3 ± 4.3	T6	28-day mortality	33 (22.8)	3.888 (2.443,6.189)	8
Gong et al.²⁹	Retrospective	Pneumonia 72; urinary 5; blood stream infection 16; abdominal 4; other 7	1.0	104	65	2001 SCCM/ESICM/ACCP/ATS/SIS	11.00 ± 3.67	20.39 ± 4.54	T6	28-day mortality	61 (58.7)	2.888 (1.676,4.976)	8
He et al.²²	Prospective	Pneumonia 19; soft tissues 3; abdominal 30; blood stream infection 4; brain 1; other 4	1.6	61	66	2001 SCCM/ESICM/ACCP/ATS/SIS	11 ± 4	21.98 ± 9.34	T24	ICU mortality	12 (20)	5.812 (1.17, 28.872)	6
Mekontso-Dessap et al.¹⁰	Retrospective	Septic shock 47; lung injury 30; cardiogenic shock 12	1.4	89	NR	NR	NR	NR	T0	1-month mortality	52 (58)	NR	6
Mesquida et al.²³	Retrospective	Pneumonia 8; urinary 4; soft tissues 5; abdominal 14; Other 4	1.4	35	63	2001 SCCM/ESICM/ACCP/ATS/SIS	9 ± 3	NR	MAP ⩾ 65 mmHg and ScvO2 ⩾ 70%^*	ICU mortality	10 (29)	NR	6
Mesquida et al.⁷	Prospective	Pneumonia 12; urinary 4; soft tissues 6; abdominal 18; other 12	NR	52	60	2016 Sepsis-3	11 ± 3	22 ± 9	MAP ⩾ 65 mmHg^*	ICU mortality	21 (40)	NR	7
Ospina-Tascón et al.¹⁴	Prospective	Pneumonia 41; urinary 25; soft tissues 15; abdominal 37; no specific parts 14; other 3	NR	135	NR	2001 SCCM/ESICM/ACCP/ATS/SIS	11.25 ± 4.67	23.47 ± 7.65	T6	28-day mortality	58 (43)	1.43 (0.86,2.30)	7
Rivera Solís et al.²⁷	Prospective	Pneumonia 10; urinary 9; soft tissues 1; abdominal 14; other 3	1.4	37	54	NR	NR	NR	T24	30 day mortality	14 (38)	5.49 (1.07,28.09)	8
Sánchez Nava et al.²⁸	Retrospective	Pneumonia 7; urinary 20; soft tissues 2; abdominal 10	1.4	39	56	NR	10.08 ± 3.30	NR	T24	ICU mortality	14 (36)	NR	5
Shaban et al.²⁴	Prospective	Septic shock 32; hemorrhagic shock 3; cardiogenic shock 2; septic shock with cardiogenic shock 13	0.25	50	54	NR	NR	26.9 ± 7.3	T6	28-day mortality	31 (62)	NR	7
Trejo Arteaga et al.²⁶	Prospective	Septic shock 13; hemorrhagic shock 9; cardiogenic shock 3; mixed shock 1	1.4	26	50	NR	9.48 ± 4.84	NR	T6	28-day mortality	17 (65.4)	NR	7
Zhou et al.²⁵	Retrospective	Pneumonia 93; urinary 9; blood stream infection 24; abdominal 15; other 3	1.4	144	62	2001 SCCM/ESICM/ACCP/ATS/SIS	11.27 ± 2.76	22.34 ± 3.93	T6	28-day mortality	74 (51.4)	2.177 (1.749,2.711)	7

SOFA: sequential organ failure assessment; APACHE: acute physiology and chronic health evaluation; RR: risk ratio; NOS: Newcastle-Ottawa scale; NR: not reported; Day1: day 1; T0: time to place the central venous catheter; T6: the sixth hour after fluid resuscitation; T24: the 24th hour after fluid resuscitation; ICU: intensive care units.

Data presented as mean ± SD.

To achieve this standard after fluid resuscitation, then blood sample collection.

Figure 3.

Pcv-aCO2/Ca-vO2 levels in survivors and non-survivors (a: subjects were all sepsis or septic shock patients; b: subjects included other types such as cardiogenic shock).

Figure 4.

Mortality in patients with high Pcv-aCO2/Ca-vO2 (high ratio) and low Pcv-aCO2/Ca-vO2 (low ratio) (a: subjects were all sepsis or septic shock patients; b: subjects included other types such as cardiogenic shock).

Results of the included studies

Pcv-aCO2/Ca-vO2 levels in survivors and non-survivors

A total of 510 patients from nine studies were analyzed,^{7,10,13,14,22
–24,26,28} including 282 survivors and 228 non-survivors (Figure 3). Heterogeneity between the results were 79%, and the random effects model was used for analysis. Meta-analysis showed higher Pcv-aCO2/Ca-vO2 in non-surviving patients with sepsis or septic shock (SMD = 0.68, 95% confidence interval (CI) = 0.24–1.12, Z = 3.04, P = 0.002; Figure 3). Subgroup analysis was performed according to patient type. All patients in group A had sepsis or septic shock,^{7,13,14,22,23,28} and group B included other types including cardiogenic shock.^10,24,26 The pooled SMD of the random effects model (I² = 84%, P < 0.001) in group A was 0.78 (0.15–1.41; Figure 3A) and 0.68 in group B (0.24–1.12, I² = 0, P = 0.84, Figure 3B). Upon exclusion of the Ospina-Tascon study, the heterogeneity became 49%. As the source of heterogeneity was variable, representative methods provided by the article differed from other studies, with median and interquartile spacing.

Mortality in patients with high Pcv-aCO2/Ca-vO2 (high ratio) and low Pcv-aCO2/Ca-vO2 (low ratio)

Five studies compared 28-day mortalities in patients with different Pcv-aCO2/Ca-vO2 ratios^{12,25
–27,29} consisting of a total of 456 patients (261 in the high-ratio group and 195 in the low-ratio group; Figure 4). No heterogeneity was observed between the studies (I² = 0, P = 0.64). The two groups were combined using a fixed-effects model. The results showed that the 28-day mortality rates in the high-ratio group were higher, and the differences were statistically significant (RR = 1.89, 95% CI = 1.48–2.41, P < 0.01; Figure 4).

SOFA and APACHE II scores in patients with high vs low Pcv-aCO2/Ca-vO2

Three studies reported the effects of Pcv-aCO2/Ca-vO2 levels on the first day SOFA and APACHE scores^12,25,29 that included 393 patients (Figure 5). The meta-analysis showed that the differences between APACHE II in the two groups were not statistically significant (MD = 0.50, 95% CI = −4.42 to 5.42, P = 0.84), and the heterogeneity between the studies was large (I² = 97%, P < 0.01; Figure 5(b)). The sources of heterogeneity were subjective according to scoring processes or time. The SOFA scores on the first day showed statistically significant differences between the groups (MD = 1.58, 95% CI = 0.88–2.28, P < 0.01) and the heterogeneity was mild (I² = 21%, P = 0.28; Figure 5(a)).

Figure 5.

(a) SOFA scores in patients with high vs low Pcv-aCO2/Ca-vO2 and (b) APACHE II scores in patients with high vs low Pcv-aCO2/Ca-vO2.

Discussion

This was the first meta-analysis to demonstrate the value of Pcv-aCO2/Ca-vO2 in predicting mortality in sepsis patients in the ICU. Non-surviving patients had higher Pcv-aCO2/Ca-vO2 values following measures to ensure fluid resuscitation and the maintenance of a stable circulatory system at the time of admission (SMD = 0.68, 95% CI = 0.24–1.12). Blood samples from 6 h after resuscitation showed a higher risk of death (RR = 1.89, 95% CI = 1.48–2.41) and higher SOFA scores (MD = 1.58, 95% CI = 0.88–2.28, P < 0.01) in patients with high Pcv-aCO2/Ca-vO2 levels.

Insufficient tissue perfusion and hypoxia at the cellular level in patients with sepsis are important indicators of mortality. Patients with these characteristics should be immediately identified and resuscitated. Previous studies are based on central venous oxygen saturation (ScvO2), blood lactate levels and central venous-arterial carbon dioxide partial pressure differences (Pcv-aCO2) as a predictor of tissue perfusion recovery and poor prognosis.^30,31 Using ScvO2 to predict patient prognosis is less accurate, as when ScvO2 reaches normal levels, patients maintain insufficient tissue perfusion and poor prognosis. Hyperlactemia suggests that the patients have insufficient tissue perfusion or hypoxia at the cellular level, but under the influence of liver dysfunction or drugs, the body is not hypoxic and hyperlactemia will occur. A lag in the increase of lactic acid is also observed.^32,33 Central venous and arterial carbon dioxide partial pressure differences (Pcv-aCO2) are widely used because as they are simple and easy to measure.^34,35 However, although tissue perfusion is sufficient, Pcv-aCO2 will continue to increase due to the Haldane effect.³⁶ It is possible that Pcv-aCO2 is normal in patients with high cardiac output and significant hypoperfusion, such as sepsis, as sufficient blood flow exists to remove CO₂ in the tissue. The value of the Pcv-aCO2/Ca-vO2 index in the assessment of tissue hypoperfusion and poor prognosis in patients with sepsis is currently under investigation.

CO₂ is a byproduct of oxygen metabolism. In aerobic conditions, the levels of CO₂ produced in the body should not exceed the amounts consumed of O₂. In patients with sepsis, the tissues were hypoperfused and the cells are hypoxic. During this time, the global oxygen supply (VO₂) is reduced, the VCO₂ due to aerobic metabolism is reduced, and the VCO₂ due to anaerobic metabolism is increased. The magnitude of the decrease in VCO₂ is smaller than the decrease in oxygen consumption, thus the VCO₂/VO₂ > 1. According to the Fick equation, the ratio of the differences in CO₂ content between mixed venous and arteries to the differences in arteriovenous oxygen content (Cv-aCO2/Ca-vO2) reflect the body’s respiratory entropy (VCO₂/VO₂).⁸ However, the complexity of the formula for the calculation of CO₂ limits its clinical application. Clinically, central venous blood samples were used instead of mixed venous blood samples, and the Pcv-aCO2/Ca-vO2 is substituted for Cv-aCO2/Ca-vO2. In studies from Mekontso-Dessap et al.¹⁰ in 2002, the Pcv-aCO2/Ca-vO2 ratio was a more reliable global anaerobic metabolism marker than other O2- and CO2-derived parameters. The change in Pcv-aCO2/Ca-vO2 in patients with septic shock is more sensitive than lactate, and the sensitivity and specificity for identifying hypoxia are higher. According to the VCO2/VO2, Gong et al.²⁹ believed that the Pcv-aCO2/Ca-vO2 ratio > 1 is perceived as abnormal. The study from Monnet et al.³⁷ showed that a Pcv-aCO2/Ca-vO2 ⩾ 1.8 can predict increases in oxygen consumption in fluid responders, the critical values in the studies by Gao et al.¹² was dependent on this. Mesquida et al.²³ and Mekontso-Dessap et al.¹⁰ obtained cut-off values based on the receiver operating characteristic (ROC) curve, when Pcv-aCO2/Ca-vO2⩾1.4, the patient’s lactate clearance rates decreased and the survival times were shorter. Four studies later conducted based on the cutoff of 1.4 reported that the mortality rates of patients increased when this ratio increased.^25
–28 In the subgroup analysis of mortality, SOFA, and APACHE II, we divided patients into high and low Pcv-aCO2/Ca-vO2, but did not specify specific cut-off values. Although the best clinical observations and critical values are uncertain currently, the most frequently used and most valuable Pcv-aCO2/Ca-vO2 threshold in the included studies was 1.4. The most common sampling time in literatures was 6 h after fluid resuscitation and blood pressure maintenance in patients with septic shock (Table 2).^{12,14,24
–26,29} The authors did not specifically explain the reasons for the time points selected. It is likely that according to resuscitation guidelines, the initial recovery time for patients with septic shock is 6 h, and so most authors choose this time point. A single study concluded that for patients undergoing cardiac surgery, monitoring this indicator 2 h after surgery is more meaningful. Whether this time point is important for septic shock patients remains undefined.³⁸

Meanwhile, we found that Pcv-aCO2/Ca-vO2 was also higher in non-survivors compared to survivors, not only during septic shock, but in patients with other types of shock including cardiogenic shock. However, the lack of studies and small sample size question these conclusions^10,24,26 (Figure 3). The study by Du et al. suggests that Pcv-aCO2/Ca-vO2 represents postoperative global anaerobic metabolism and predictive responses to oxygen delivery (DO₂) challenge in patients undergoing cardiac surgery, while Abou-Arab et al. obtained the opposite conclusion through experimentation.^38–40 Recent studies published in 2019 proposed that Pcv-aCO2/Ca-vO2 is an important predictor of postoperative major organ morbidity and mortality in patients undergoing cardiac surgery, and that the ratio may represent a useful resuscitation target.⁴¹ The role of Pcv-aCO2/Ca-vO2 in other critically ill patients requires further validation studies. Three studies,12,25,29 including Zhou and colleagues used lactic acid, Pcv-aCO2/Ca-vO2, and lactic acid in combination with Pcv-aCO2/Ca-vO2 to construct ROC curves that were used to predict 28-day mortality in patients with septic shock. The area under the curve (AUC) of the combined application of Pcv-aCO2/Ca-vO2 and lactic acid was more accurate than lactic acid or Pcv-aCO2/Ca-vO2 alone. Rivera Solís et al.²⁷ reported that after resuscitation, patients with ScvO2 < 70% and a ratio > 1.4 have a higher mortality rate. To-date, as accurate indicators of tissue perfusion and hypoxia are lacking, in critically ill patients, clinicians should pay attention to this indicator combined with lactic acid clearance and central venous blood oxygen saturation. This can minimize any influencing factors to achieve early resuscitation, and to judge the effects of early resuscitation treatment.

This meta-analysis has inherent limitations. The included studies were sparse and the majority of articles had small sample sizes, questioning the credibility of the conclusions. In addition, the rates of false positive results were higher, leading to study bias. The research environment of the studies included in this article was based on those of the ICU. For patients with sepsis in other environments, the conclusions require confirmation. Based on our current data, we only provided commonly used cut-off values for this indicator at 1.4, with the usual sampling time point being the 6th hour after admission for treatment. It is therefore difficult to propose the optimal cut-off values of Pcv-aCO2/Ca-vO2 that can predict poor patient prognosis.

In conclusion, Pcv-aCO2/Ca-vO2 is a predictor of poor prognosis in patients with sepsis or septic shock. For patients with severe sepsis or septic shock, active and effective fluid resuscitation should be performed as soon as possible. On the basis of the MAP, ScvO2 and lactate clearance rates reaching baseline standards, further Pcv-aCO2/Ca-vO2 assessments are required to guide fluid resuscitation and improve the patient’s prognosis.

Footnotes

Appendix 1 Acknowledgements

The authors gratefully acknowledge the help provided by Dr Dongyue Wang of the image processing. Thanks, are also due to Dr. Chao Xu, who gave them much valuable advice in the early stages of this work.

Author contributions

M.W. helped design the study, perform the data analysis, resolve discussion and write the manuscript. T.L. helped design the study, perform the data analysis and interpret the data.

Z. N. helped resolve discussion and modify the grammar and content of the manuscript. J.Z. helped resolve discussion and modify the grammar and content of the manuscript. D.Q. helped design the study and write the manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Min Wang

References

Singer

Deutschman

Seymour

, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016; 315(8): 801–810.

Shankar-Hari

Phillips

Levy

, et al. Developing a new definition and assessing new clinical criteria for septic shock: for the third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016; 315(8): 775–787.

Rhodes

Evans

Alhazzani

, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock: 2016. Intensive Care Med 2017; 43(3): 304–377.

Meyer

Harhay

Small

, et al. Temporal trends in incidence, sepsis-related mortality, and hospital-based acute care after sepsis. Crit Care Med 2018; 46(3): 354–360.

Stevenson

Rubenstein

Radin

, et al. Two decades of mortality trends among patients with severe sepsis: a comparative meta-analysis*. Crit Care Med 2014; 42(3): 625–631.

Kaukonen

Bailey

Suzuki

, et al. Mortality related to severe sepsis and septic shock among critically ill patients in Australia and New Zealand, 2000-2012. JAMA 2014; 311(13): 1308–1316.

Mesquida

Saludes

Pérez-Madrigal

, et al. Respiratory quotient estimations as additional prognostic tools in early septic shock. J Clin Monit Comput 2018; 32(6): 1065–1072.

Ospina-Tascon

Hernandez

Cecconi

Understanding the venous-arterial CO₂ to arterial-venous O₂ content difference ratio. Intensive Care Med 2016; 42(11): 1801–1804.

Mallat

Lemyze

Meddour

, et al. Ratios of central venous-to-arterial carbon dioxide content or tension to arteriovenous oxygen content are better markers of global anaerobic metabolism than lactate in septic shock patients. Ann Intensive Care 2016; 6(1): 10.

10.

Mekontso-Dessap

Castelain

Anguel

, et al. Combination of venoarterial PCO₂ difference with arteriovenous O₂ content difference to detect anaerobic metabolism in patients. Intensive Care Med 2002; 28(3): 272–277.

11.

Saludes

Proença

Gruartmoner

, et al. Central venous-to-arterial carbon dioxide difference and the effect of venous hyperoxia: a limiting factor, or an additional marker of severity in shock? J Clin Monit Comput 2017; 31(6): 1203–1211.

12.

Gao

Cao

Central venous-arterial carbon dioxide tension to arterial-central venous oxygen content ratio combined with lactate clearance rate as early resuscitation goals of septic shock. Zhonghua Yi Xue Za Zhi 2018; 98(7): 508–513.

13.

Dubin

Pozo

Kanoore Edul

, 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: 445–450.

14.

Ospina-Tascón

Umaña

Bermúdez

, et al. Combination of arterial lactate levels and venous-arterial CO₂ to arterial-venous O₂ content difference ratio as markers of resuscitation in patients with septic shock. Intensive Care Med 2015; 41(5): 796–805.

15.

Tang

Liu

, et al. P(v-a)CO2/C(a-v)O2-directed resuscitation does not improve prognosis compared with SvO2 in severe sepsis and septic shock: a prospective multicenter randomized controlled clinical study. J Crit Care 2018; 48: 314–320.

16.

Dellinger

Levy

Rhodes

, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med 2013; 39(2): 165–228.

17.

Luo

Wan

Liu

, et al. Optimally estimating the sample mean from the sample size, median, mid-range, and/or mid-quartile range. Stat Methods Med Res 2018; 27(6): 1785–1805.

18.

Wan

Wang

Liu

, et al. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol 2014; 14: 135.

19.

Wells

Shea

O’Connell

, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses, http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp

20.

Deeks

JJ.

Issues in the selection of a summary statistic for meta-analysis of clinical trials with binary outcomes. Stat Med 2002; 21(11): 1575–1600.

21.

Egger

Davey Smith

Schneider

, et al. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315(7109): 629–634.

22.

Long

Liu

, et al. The prognostic value of central venous-to-arterial CO₂ difference/arterial-central venous O₂ difference ratio in septic shock patients with central venous O₂ saturation ⩾80. Shock 2017; 48(5): 551–557.

23.

Mesquida

Saludes

Gruartmoner

, 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.

24.

Shaban

Salahuddin

Kolko

, et al. The predictive ability of PV-ACO2 gap and PV-ACO2/CA-VO2 ratio in shock: a prospective, cohort study. Shock 2017; 47(4): 395–401.

25.

Zhou

Song

Gong

, et al. Persistent hyperlactatemia-high central venous-arterial carbon dioxide to arterial-venous oxygen content ratio is associated with poor outcomes in early resuscitation of septic shock. Am J Emerg Med 2017; 35(8): 1136–1141.

26.

Trejo Arteaga

Martínez Zubieta

Cerón Díaz

, et al. Relación de la diferencia de presión venoarterial de dióxido de carbono sobre contenido arteriovenoso de oxígeno (ΔPCO₂/Ca-vO₂) versus lactato como marcadores pronóstico en pacientes en estado de choque. Rev Asoc Mex Med Crit Ter Int 2016; 30(2): 119–126.

27.

Rivera Solís

Sánchez Díaz

Martínez Rodríguez

, et al. Clasificación clínica de la perfusión tisular en pacientes con choque séptico basada en la saturación venosa central de oxígeno (SvcO2) y la diferencia venoarterial de dióxido de carbono entre el contenido arteriovenoso de oxígeno (ΔP(v-a)CO2/C(a-v)O2). Med Crit 2016; 30(5): 283–289.

28.

Sánchez Nava

Muñoz Ramírez

MDR

Chávez Pérez

CIGI.

Utilidad de los parámetros gasométricos como predictores de mortalidad en pacientes con choque séptico. Med Crit 2017; 31(3): 128–135.

29.

Gong

Song

Zhou

, et al. Venous-to-arterial carbon dioxide difference to arteriovenous oxygen content difference ratio combined with lactate to predict prognosis of patients with septic shock. Zhonghua Nei Ke Za Zhi 2016; 55(9): 673–678.

30.

Nguyen

Rivers

Knoblich

, et al. Early lactate clearance is associated with improved outcome in severe sepsis and septic shock. Crit Care Med 2004; 32(8): 1637–1642.

31.

Wittayachamnankul

Chentanakij

Sruamsiri

, et al. The role of central venous oxygen saturation, blood lactate, and central venous-to-arterial carbon dioxide partial pressure difference as a goal and prognosis of sepsis treatment. J Crit Care 2016; 36: 223–229.

32.

Hernandez

Bruhn

Castro

, et al. Persistent sepsis-induced hypotension without hyperlactatemia: a distinct clinical and physiological profile within the spectrum of septic shock. Crit Care Res Pract 2012; 2012: 536852.

33.

Rimachi

Bruzzi

Carvahlo

Orellano-Jimenez

, et al. Lactate/pyruvate ratio as a marker of tissue hypoxia in circulatory and septic shock. Anaesth Intensive Care 2012; 40(3): 427–432.

34.

Dres

Monnet

Teboul

JL.

Hemodynamic management of cardiovascular failure by using PCO(2) venous-arterial difference. J Clin Monit Comput 2012; 26(5): 367–374.

35.

Van Beest

Lont

Holman

, et al. Central venous-arterial pCO(2) difference as a tool in resuscitation of septic patients. Intensive Care Med 2013; 39(6): 1034–1039.

36.

Jakob

Kosonen

Ruokonen

, et al. The Haldane effect—an alternative explanation for increasing gastric mucosal PCO2 gradients? Br J Anaesth 1999; 83(5): 740–746.

37.

Monnet

Julien

Ait-Hamou

, et al. Lactate and venoarterial carbon dioxide difference/arterial-venous oxygen difference ratio, but not central venous oxygen saturation, predict increase in oxygen consumption in fluid responders. Crit Care Med 2013; 41(6): 1412–1420.

38.

Moussa

Durand

Leroy

, et al. Central venous-to-arterial PCO2 difference, arteriovenous oxygen content and outcome after adult cardiac surgery with cardiopulmonary bypass: a prospective observational study. Eur J Anaesthesiol 2019; 36(4): 279–289.

39.

Abou-Arab

Braik

Huette

, 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(10): e0205950.

40.

Long

Wang

, et al. The use of the ratio between the veno-arterial carbon dioxide difference and the arterial-venous oxygen difference to guide resuscitation in cardiac surgery patients with hyperlactatemia and normal central venous oxygen saturation. Chin Med J 2015; 128(10): 1306–1313.

41.

Mukai

Suehiro

Kimura

, et al. Comparison of the venous-arterial CO₂ to arterial-venous O₂ content difference ratio with the venous-arterial CO₂ gradient for the predictability of adverse outcomes after cardiac surgery. J Clin Monit Comput 2020; 34: 41–53.