Correlation of end-tidal carbon dioxide with transthoracic echocardiography derived cardiac output for the assessment of fluid responsiveness: A prospective observational study

Abstract

Background and Aims:

Judicious use of intravenous fluid therapy can be lifesaving in the intraoperative period. An assessment of fluid responsiveness is important for it. So, our study aimed to study the correlation between end-tidal carbon dioxide (EtCO₂) and transthoracic echocardiography-derived cardiac output for assessing fluid responsiveness in patients undergoing elective surgery under general anesthesia.

Material and Methods:

Patients who underwent elective lower abdominal or lower limb surgeries in a supine position under general anesthesia with positive pressure ventilation were included in this study. Cardiac output was calculated using transthoracic echo, and by measuring the diameter of the left ventricular outflow tract (LVOT), velocity time integral of LVOT (LVOT-VTI), and heart rate. Cardiac output (CO), EtCO₂, and hemodynamic and ventilatory parameters were analyzed by the operator before and 1 min after the infusion of 250 mL of normal saline.

Results:

EtCO₂ variations showed a weak correlation with the changes in CO induced by a fluid challenge (Spearman’s correlation r = 0.3, P = 0.005). When fluid responsiveness (FR) is defined as an increase in CO by > 15%, the AUROC of ∆ EtCO₂ was 0.638 (95% confidence interval [CI], 0.507–0.77). At a cut-off of ≥ 1 mmHg, it predicted FR status (responder vs. non-responder) with a sensitivity of 66% and a specificity of 64%. When percentage variation in EtCO₂ (percent ∆ EtCO₂) was considered, the AUROC was almost similar (0.618) (95% CI, 0.481–0.756), and it was not statistically significant (P = 0.093).

Conclusions:

Our study demonstrated a weak positive correlation between volume-induced changes in EtCO₂ and changes in CO in mechanically ventilated patients in the operating room. Variations in EtCO₂ can be used as an adjunct to guide hemodynamic optimization when no COcardiac output monitors are available.

Introduction

Intravenous (iv) fluid therapy aims to improve cardiac output (CO) and restore oxygen delivery to hypoperfused organs by fluid loading. Unfortunately, fluid is not always the solution to decreased perfusion, and a positive fluid balance is associated with increased lengths of hospital stays and mortality.[1] Furthermore, for physiological reasons, the hemodynamic response to a fluid challenge is not easily predictable because the expected improvement in CO is only observed in half the patients.[2] Hence, before administering a fluid bolus, it is crucial to determine whether volume loading will improve a patient’s perfusion or not. Targeting iv fluid therapy to fluid responsiveness (FR) and, therefore CO may optimize tissue-oxygenation and reduce the risk of tissue edema.

The application of invasive CO monitoring for guiding fluid resuscitation, previously with pulmonary artery catheterization and in recent times with less invasive devices, such as arterial pulse pressure analysis and esophageal Doppler, is limited and can be inconvenient for routine monitoring in the perioperative setting. Indeed, the use of CO measuring devices is also limited by their cost, an unfavorable risk–benefit balance (for indwelling devices), the lack of expertise of some users, and pathophysiological barriers.

The utility of non-invasive CO monitoring methods, which can rapidly identify FR and guide fluid therapy, has been studied in the past few years. In this context, volume-induced changes in end-tidal carbon dioxide (EtCO₂) are an attractive option to detect the changes in CO when no direct CO monitors are available. The amount of exhaled CO₂ depends on production by body tissues, pulmonary blood flow (i.e. CO), and alveolar ventilation. Hence, changes in EtCO₂ parallels changes in CO if alveolar ventilation is constant, as in patients with fully controlled mechanical ventilation, and if cell metabolism is stable.

The relationship between CO and EtCO₂ has been well defined in the scenario of cardiopulmonary resuscitation,[3] and the current resuscitation guidelines recommend the utilization of EtCO₂ in predicting a return of spontaneous circulation, monitoring the effectiveness of chest compressions and confirming the placement of the endotracheal tube.[4,5] Even though there are numerous published data focusing on the utility of EtCO₂ for the assessment of FR, most of them were performed in intensive care unit (ICU) settings[6,7,8,9,10] and in patients with spontaneous ventilation.[11,12] The ability of EtCO₂ variation after a fluid challenge to detect FR in the context of general anesthesia in the operating room has not been adequately investigated, and there are only a limited number of studies exploring the same.

So, our study aimed to study the correlation between EtCO₂ and transthoracic echocardiography-derived cardiac output for assessing fluid responsiveness in patients undergoing elective surgery under general anesthesia.

Material and Methods

After obtaining approval from the institutional ethical committee (AIIMS/IEC/21/161 dated 09/04/2021) and registration with the Clinical Trial Registry of India (CTRI/2021/12/038508), this prospective observational study was conducted over 18 months from April 2021 to October 2022. Patients, both male and female, aged between 18 and 70 years, American Society of Anesthesiologists (ASA) class I and II, who underwent elective lower abdominal or lower limb surgeries in a supine position under general anesthesia with positive pressure ventilation, were included in the study. Patients who refused to consent or with signs/symptoms of fluid overload, cardiac failure, renal or hepatic failure, structural cardiac abnormalities, cardiac arrhythmias, pregnant patients, or those undergoing laparoscopic surgery or in thoracic surgeries etc., where the thorax is not accessible to anesthesiologists were excluded from the study.

All patients meeting the inclusion criteria were explained the procedure involved, and written consent was obtained. Upon arrival to the operating room, patients were positioned supine, and an advanced multiparameter monitor, including an electrocardiogram (ECG), non-invasive blood pressure (NIBP), and oxygen saturation (SpO₂) were attached, and baseline vital parameters were recorded. An appropriate-size intravenous catheter was inserted, preferably in an upper extremity vein, and intravenous fluid was started at the rate of 2 mL/kg/h.

The patients were pre-oxygenated with 100% O₂ using an appropriately sized face mask for 3 min. After pre-oxygenation, general anesthesia was induced with fentanyl 2 μg/kg, propofol 2 mg/kg, and vecuronium 0.1 mg/kg given intravenously. After induction, direct laryngoscopy was performed, and the trachea was intubated with an endotracheal tube (ET) of appropriate size. The tracheal position of the ET tube was confirmed by 5-point auscultation and EtCO₂ readings. Minute ventilation (MV) was adjusted to maintain EtCO₂ in the 35–45 mmHg range. Volume-controlled ventilation was set for a tidal volume (TV) of 6–8 mL/kg with positive end-expiratory pressure (PEEP) of 5 cm of H₂O. EtCO₂ was continuously measured at the tip of the ET tube using a side-stream infrared gas analyzer connected to the anesthesia machine. At baseline, heart rate (HR), systolic BP (SBP), diastolic BP (DBP), mean arterial pressure (MAP), EtCO₂, and CO were recorded.

All echocardiographic assessments were performed by an expert to ensure accuracy and uniformity. A portable ultrasound machine (SonoSite Edge II Ultrasound System, Fujifilm Inc.) with a phased array 1-5 MHz probe was used for the transthoracic echo. Cardiac output was calculated by measuring the diameter of the left ventricular outflow tract (LVOT), velocity time integral of LVOT (LVOT-VTI), and heart rate. The diameter of the LVOT was measured from the parasternal long-axis view during mid-diastole. LVOT-VTI was measured using a pulsed wave Doppler on the flow from the aortic valve to the apex in the apical five-chamber view. LVOT VTI was used to estimate stroke volume (SV) because it reflects the column of blood, which moves through the LV outflow tract during each systole, per the following equation:

Stroke volume = LVOT VTI x cross-sectional area of the LVOT.

Three measurements were taken for both LVOT VTI and the cross-sectional area of the LVOT, and an average was calculated. Assuming laminar flow through the LVOT, CO was calculated as:

CO = Stroke volume x heart rate.

In all patients, volume expansion was achieved by infusion of 250 mL of 0.9% normal saline over 10 min. During the 10 min of fluid administration, there was no modification of the ventilatory settings; also, no bolus or change in the rate of administration of vasoconstrictor and anesthetic agents was performed during this period.

CO, EtCO₂, and hemodynamic and ventilatory parameters were noted and measured by the operator 1 min after the infusion of 250 mL of normal saline. Fluid responsiveness (FR) was defined as an increase in CO > 15% from baseline after fluid bolus administration. Changes in EtCO₂ (∆EtCO₂) were noted in both fluid responders and non-responders and compared with the changes in CO (∆CO). We also calculated the number (percentage) of fluid responders and non-responders.

The fluid challenge could be repeated for the same patient with a fluid bolus of 250 mL of 0.9% normal saline in patients who required volume expansion intraoperatively at the discretion of the attending physician. The CO, EtCO₂, and hemodynamic and respiratory parameters were recorded before and after giving fluid bolus and were analyzed for correlation separately.

The primary objective of our study was to correlate the changes in EtCO₂ with changes in CO (derived from left ventricular outflow tract velocity time integral) after volume expansion in patients undergoing lower abdominal or lower limb surgeries under general anesthesia with positive pressure ventilation. The secondary objectives were to study the efficacy of EtCO₂ as a marker of fluid responsiveness during perioperative volume expansion.

Sample size estimation

The sample size for estimating the correlation between ∆ EtCO₂ and ∆ CO was calculated according to the formula given by Hulley et al. (2013)[13]

Sample size N = [(Zα + Zβ)/C]² + 3,

where C = 0.5* ln[(1 + r)/(1-r)], r being the observed correlation in previous studies.

The sample size was based on a study by Güney Pınar et al. (2022),[11] who reported that the correlation between ∆ EtCO₂ and CO was 0.585. So, based on α error of 5% and a β error of 1% (Zα =1.96; Zβ =2.326)

[(1.96 + 2.326)/0.670]² + 3 = 43.92 ≈ 50

Thus, assuming 99% power and a 95% confidence interval (CI), the sample size for the study was estimated as 50.

Statistical analysis

Data were recorded in Microsoft Excel (2019) spreadsheet program. Statistical Package for Social Sciences (SPSS v23) (IBM Corp.) was used for data analysis. Descriptive statistics was elaborated in the form of means (standard deviations [SD]) and medians (interquartile ranges [IQR]) for continuous variables and frequencies/percentages for categorical variables.

Group comparisons for continuously distributed data were made using the independent sample “t”- test when comparing two groups, and one-way analysis of variance (ANOVA) when comparing more than two groups. Post-hoc pairwise analysis was performed using Tukey’s Honestly Significant Difference (HSD) Test in the case of one-way ANOVA. When the data were found to be non-normally distributed, appropriate non-parametric tests in the form of Wilcoxon test/Kruskal–Wallis test were used for these comparisons. Linear correlation between two continuous variables was explored using Pearson’s correlation (for normally distributed data) and Spearman’s correlation (for non-normally distributed data). Statistical significance was kept at P < 0.05.

Results

Fifty-eight patients who met the inclusion criteria were considered for this study. Out of these, 8 were excluded as no acoustic doppler window was available [Figure 1]. Rest 50 patients completed this study [Figure 1]. The demographic profile of patients is described in Table 1.

Figure 1.

Patient flow chart

Table 1.

Summary of patient characteristics

Basic details	Mean±SD* || Median (IQR†) || Min-Max‡|| Frequency (%)
Age (years)	38.40±13.75 || 37.50 (28.00-50.50) || 18.00-69.00
Age (years)
18-30	15 (30.0%)
31-40	14 (28.0%)
41-50	8 (16.0%)
51-60	11 (22.0%)
61-70	2 (4.0%)
Gender
Male	18 (36.0%)
Female	32 (64.0%)
Weight (Kg§)	63.02±7.61 || 60.00 (58.00-68.00) || 48.00-87.00
Height (cm||)	161.84±7.60 || 161.50 (156.25-167.50) || 145.00-178.00
BMI (Kg/m2)¶	24.09±2.70 || 23.61 (22.02-26.28) || 19.23-30.92
BMI
18.5-22.9	18 (36.0%)
23.0-24.9	16 (32.0%)
25.0-29.9	15 (30.0%)
30.0-34.9	1 (2.0%)
Co-morbidities
Nil	38 (76.0%)
HTN**	4 (8.0%)
DM††	3 (6.0%)
DM + HTN	2 (4.0%)
HTN + hypothyroidism	1 (2.0%)
Hypothyroidism	1 (2.0%)
Smoker	1 (2.0%)
ASA‡‡
I	38 (76.0%)
II	12 (24.0%)
Surgery
Lower limb	24 (48.0%)
Lower abdominal	26 (52.0%)

*Standard deviation, ^†Interquartile range, ^‡Minimum-Maximum, ^§Kilograms, ^||Centimeter, ^¶Body mass index, **Hypertension, ^††Diabetes, ^‡‡American society of anesthesiologists

We found that 49.2% of the total fluid challenges (n = 65) fulfilled the criteria of the responder status, and the rest 50.8% were non-responders. A summary of hemodynamic variables (HR and MAP), CO, and EtCO₂ before and after volume expansion in responders and non-responders are reported in Table 2. The absolute change in CO (∆CO) and percentage change in CO (percent-∆CO) in fluid responders and non-responders are described in Table 3. There was a significant difference between the two groups in terms of ∆CO (t = 10.319, P ≤ 0.001), with the mean ∆CO being highest in the responder group. The strength of association (point-biserial correlation) was calculated as 0.8 (large effect size). Similarly, there was a significant difference between the two groups in terms of percent-∆CO (t = 13.075, P ≤ 0.001), with the strength of association (point-biserial correlation) being 0.86 (large effect size).

Table 2.

Summary of hemodynamic variables, cardiac output, and EtCO₂ before and after volume expansion in responders and non-responders (n=65) (values are mean (SD)

	Before volume expansion	After volume expansion	P
Heart rate (bpm*)
Responder	82.69 (10.28)	78.38 (8.66)	<0.001
Non-responder	76.52 (10.62)	74.39 (10.46)	<0.001
MAP (mmHg†)
Responder	71.25 (7.72)	76.84 (7.15)	<0.001
Non-responder	75.18 (9.02)	78.82 (9.45)	<0.001
Cardiac output (mL/min‡)
Responder	4101.86 (774.70)	5187.10 (1038.27)	<0.001
Non-responder	4053.89 (955.88)	4413.33 (1058.67)	<0.001
EtCO2 (mmHg†)
Responder	32.91 (1.38)	33.69 (1.00)	<0.001
Non-responder	33.39 (1.69)	33.73 (1.04)	0.139

*Beats per minute, ^†millimeter of mercury, ^‡milliliter per minute

Table 3.

Comparison of two subgroups (responders vs. non-responders) of the variable fluid responsive status in terms of change in cardiac output (∆CO) and change in EtCO₂ (∆EtCO₂) ((n=65)

∆CO§	Fluid responsive status		t-test or Wilcoxon–Mann–Whitney U test
	Responder	Non-responder	t	P
Mean (SD*)	1085.24 (354.58)	359.45 (183.32)	10.319	<0.001
Median (IQR†)	982.96 (858.84-1371.42)	355.71 (240.94-523.76)
Min–max‡	438.77-2135.44	4.33-711.42
Percent-∆CO§			t	P
Mean (SD*)	26.37 (6.45)	8.84 (4.05)	13.075	<0.001
Median (IQR†)	24.85 (21.88-32.09)	9.91 (5.07-12.43)
Min–max‡	16.79-39.04	0.14-14.18
∆EtCO2§			W	P
Mean (SD*)	0.78 (0.83)	0.33 (1.19)	674.000	0.045
Median (IQR†)	1 (0-1)	0 (0-1)
Min–max‡	-1-2	-2-3
Percent-∆EtCO2§			W	P
Mean (SD*)	2.45 (2.59)	1.14 (3.63)	653.000	0.093
Median (IQR†)	3.03 (0-3.12)	0 (0-3.12)
Min–max‡	-2.86-6.9	-5.71-9.38

*Standard deviation, ^†Interquartile range, ^‡Minimum-Maximum, ^§change in cardiac output

The absolute change in EtCO₂ (∆EtCO₂) and percentage change in EtCO₂ (percent-∆EtCO₂) in fluid responders and non-responders are described in Table 3. There was a significant difference between the two groups in terms of ∆EtCO₂ (W = 674.000, P = 0.045), with the median ∆EtCO₂ being the highest in the responder group. The strength of association (point-biserial correlation) was 0.22 (small effect size). However, there was no significant difference between the groups in terms of percent-∆EtCO₂ (W = 653.000, P = 0.093), and the strength of association (point-biserial correlation) was 0.21 (small effect size)

We observed a weak positive correlation of 0.3 between ∆EtCO₂ and ∆CO, which was statistically significant (rho = 0.34, P = 0.005) [Supplementary Figure 1 ^{(1.6MB, tif)} ]. For every 1 unit increase in ∆EtCO₂, the ∆CO increases by 128.76 units. So, as per the results of our study, the increase in ∆EtCO₂ with a 1 unit increase in CO would be negligible. There was a weak positive correlation between percent-∆CO and percent-∆EtCO₂, which was statistically significant (rho = 0.31, P = 0.011) [Supplementary Figure 2 ^{(396KB, tif)} ]. For every 1 unit increase in percent-∆CO, the percent-∆EtCO₂ increases by 0.08 units. Conversely, for every 1 unit increase in percent-∆EtCO₂, the percent-∆CO increases by 0.87 units.

The area under the receiver operating curve (ROC) curve (AUROC) for ∆EtCO₂ predicting fluid responsive status: responder vs. non-responder was 0.638 (95% CI: 0.507–0.77), thus demonstrating poor diagnostic performance. It was statistically significant (P = 0.045). At a cut-off of ∆ EtCO₂ ≥ 1, it predicted fluid responsive status: responder with a sensitivity of 66% and a specificity of 64% [Figure 2]. The AUROC for percent-∆EtCO₂ predicting fluid responsive status: responder vs. non-responder was 0.618 (95% CI: 0.481–0.756), thus demonstrating poor diagnostic performance. It was not statistically significant (P = 0.093). At a cut-off of percent-∆EtCO₂ ≥ 2.941, it predicts fluid responsive status: responder with a sensitivity of 66%, and a specificity of 64% [Figure 3]. However, the cut-off and the diagnostic parameters were unreliable as the test was not statistically significant.

Figure 2.

ROC curve analysis showing diagnostic performance of ∆ EtCO₂ in predicting fluid responsive status: Responder vs. non-responder (n = 65)

Figure 3.

ROC curve analysis showing diagnostic performance of percent-∆EtCO₂ in predicting fluid responsive status: responder vs. non-responder (n = 65)

Discussion

The present study was undertaken in 50 patients admitted to a tertiary care hospital, posted electively for lower limb or lower abdominal surgeries under GA. A total of 65 volume expansions were performed in 50 patients, and the correlation of changes in CO measured by TTE (LVOT VTI derived CO) with changes in EtCO₂ was studied. Our study demonstrated that EtCO₂ variations showed a weak (statistically significant) correlation with the changes in CO induced by a fluid challenge (Spearman’s correlation r = 0.3, P = 0.005). We also studied the efficacy of EtCO₂ as a marker of FR after VE by ROC analysis. When FR is defined as an increase in CO by > 15%, the AUROC of ∆ EtCO₂ was 0.638 (95% CI, 0.507–0.77). At a cut-off of ≥ 1 mmHg, it predicted FR status (responder vs. non-responder) with a sensitivity of 66% and a specificity of 64%. When percentage variation in EtCO₂ (percent ∆ EtCO₂) was considered, the AUROC was almost similar (0.618) (95% CI; 0.481–0.756), and it was not statistically significant (P = 0.093).

Previous studies performed in the ICU setting and operating room have evaluated variations in EtCO₂ as a surrogate for changes in CO during volume expansion, Passive Leg Raising (PLR), or increase in Positive End-Expiratory Pressure (PEEP) level.[7,8,9,10] Toupin et al.[14] in 2016 showed that in patients undergoing cardiac surgery, a positive response to volume expansion was associated with a rise of at least 2 mm Hg in EtCO₂ following PLR. They observed a low positive predictive value (PPV) (54%) and a high negative predictive value (NPV) (86%). In their study, the calculated correlation (Pearson’s correlation coefficient) between EtCO₂ variation and the cardiac index (CI) variation after the PLR maneuver was 0.47 (95% CI, 0.29–0.62; P ≤ 0.01). Our work differs from this study in many ways. The study population (cardiac surgical patients with valvular heart disease vs. ASA I, II patients undergoing lower limb/lower abdominal surgeries), the method used for VE (PLR maneuver vs. 250 mL crystalloid infusion), and the means of CO measurement (thermodilution method vs. LVOT VTI-derived CO by TTE) were different, which could have affected the results.

Another study by Jacquet-Lagrèze et al.[15] in 40 patients who underwent major non-cardiac surgery under GA found that an increase of more than 2 mmHg of EtCO₂ (increase > 5.8%) accurately predicted a positive response to a 500 mL colloid VE (specificity = 96%, sensitivity = 60%; AUROC = 0.80; 95% CI, 0.65–0.96). The correlation between CI variation and EtCO₂ variation they observed was r = 0.566 (P < 0.001). However, a variation in EtCO₂ < 5.8% proved ineffective in distinguishing responders and non-responders. Unlike ours, their study group included high-risk patients (with heart failure or significant arteriopathy) and emergency high-risk surgeries (peritonitis, hip fracture etc.). That was of importance because the fluid bolus administered in a patient with acute circulatory failure tries to restore a compromised hemodynamic system. In contrast, in a patient without hemodynamic failure, volume expansion (VE) aims to maximize SV and CO. Furthermore, the fluid used for VE was 500 mL colloid, and the CO measurements were performed with an esophageal doppler.

It is noteworthy that the findings of our study corroborated with the results of a recent study by George s et al.,[16] who investigated the utility of EtCO₂ as a marker of FR after VE in neurosurgery patients in the operating room. They found out that changes in EtCO₂ > 1.1% induced by infusion of 250 mL crystalloid could identify FR with a sensitivity of 62.9% (95% CI, 62.5–63.3%) and a specificity of 77.8% (95% CI, 77.6–78.1%). The AUROC for ∆EtCO₂ after VE was 0.683 (95% CI, 0.680–0.686). The correlation they observed between variations in EtCO₂ and CI was weak (r = 0.454, P < 0.001). Our study differed from this study in that we used LVOT VTI-derived CO by TTE instead of the continuous CO monitor used by them to measure SVI and CI.

In intensive care, a significant correlation has been shown between variations in CO and variations in EtCO₂ following volume expansion. Monnet et al.,[8] in a study conducted on 40 mechanically ventilated ICU patients, observed that an increase in EtCO₂ ≥ 5% predicted an increase in CI ≥ 15% (measured by a PiCCO device) by VE with a sensitivity of 71% (95% CI, 48–89%) and specificity of 100% (95% CI, 82–100%). In a similar study, Young et al.[9] compared EtCO₂ and volumetric EtCO₂ (VCO₂) to CO (NICOM bioreactance CO monitor) in 34 mechanically ventilated ICU patients after a PLR or volume loading maneuver. The authors reported that EtCO₂ increased by 5.9 ± 7.6% in volume responders compared with 1.4 ± 4.4% in non-responders (P = 0.02) and an AUROC of 0.67 (95% CI, 0.63–0.69). In a study conducted in mechanically ventilated ICU patients with circulatory shock, Lakhal et al.[10] observed that changes in EtCO₂ outperformed other surrogates (PP, SBP, MAP, HR, and femoral flow) for changes in CO during the fluid challenge. ∆EtCO₂ had an AUROC of 0.82 (IQR: 0.73–0.90), which was significantly higher than the AUROC for other parameters. They found out that a value of ∆ EtCO₂ > 1 mmHg had good positive (5.0 [2.6–9.8]) and fair negative (0.29 [0.2–0.5]) likelihood ratios.

In many ways, our research differs from these earlier investigations. Most of them took place in ICUs and/or included patients who were receiving vasopressors, had acute circulatory failure, or both. That could be the reason why the observed correlation between changes in EtCO₂ and CO was stronger. It was previously noticed that the link between EtCO₂ and CO is predicted to be weaker when there is no circulatory insufficiency.[17] It is also worth mentioning that, in previous studies, PLR-induced changes in EtCO₂ and CO demonstrated a better correlation than during fluid challenge-induced VE. This can be explained by the fact that the preload increase induced by a mini fluid challenge is inferior compared with a PLR maneuver.

There were certain limitations in our study. Firstly, this was a single-center study conducted on the Indian sub-group of the population only, and variation may be observed with studies conducted on other population groups and demographic characteristics. Secondly, the study involved adult ASA I and II patients undergoing elective lower limb/lower abdominal surgery in a supine position under GA with controlled ventilation only. Thirdly, we used TTE for the measurement of CO instead of continuous CO monitors due to the lack of availability of the equipment. The use of continuous CO monitoring would have given a beat-by-beat assessment of cardiac output. That could have avoided the time lag from the beginning of the CO assessment and the calculation of the result. Finally, the cut-off of 15% change in CO to discriminate responders from non-responders could be too high in some of the fluid challenges; also, the amount of fluid we used to induce VE was only 250 mL compared to a 500 mL bolus or PLR maneuver, both of which could have affected our results. So, further studies with diverse patient profiles are required to evaluate the diagnostic performance of change in EtCO₂ as amarker of fluid responsiveness and also to compare the efficacy of mainstream versus side stream CO₂ analyzer.

Conclusions

Our study demonstrated a weak positive correlation between volume-induced changes in EtCO₂ and changes in CO in mechanically ventilated patients in the operating room. Variations in EtCO₂ can be used as an adjunct to guide hemodynamic optimization when no CO monitors are available. In the future, with the development of more accurate ETCO₂ monitors, the sensitivity of this non-invasive parameter for assessing fluid responsiveness can be improved, making it more clinically dependable. However, as of now, the diagnostic performance of change in EtCO₂, as a stand-alone parameter, when used in elective surgical patients is poor.

Conflicts of interest

There are no conflicts of interest.

Supplementary Figure 1

Correlation between ∆ CO and ∆ EtCO₂ (n = 65)

Supplementary Figure 2

Correlation between percent-∆CO and percent-∆EtCO₂ (n = 65)

Funding Statement

Nil.