Elevations in End-Tidal CO₂ With CO₂ Use During Pediatric Endoscopy With Airway Protection: Is This Physiologically Significant?

Background:

Inflation of the gastrointestinal lumen is vital for proper visualization during endoscopy. Air, insufflated via the endoscope, is gradually being replaced with carbon dioxide (CO₂) in many centers, with the intention of minimizing post-procedural discomfort due to retained gas. Recent studies suggest that the use of CO₂ during pediatric esophagogastroduodenoscopy (EGD) with an unprotected airway is associated with transient elevations in exhaled CO₂ (end-tidal CO₂, EtCO₂), raising safety concerns. One possible explanation for these events is eructation of insufflation gas from the stomach.

Objectives:

To distinguish eructated versus absorbed CO₂ by sampling EtCO₂ from a protected airway with either laryngeal mask airway (LMA) or endotracheal tube (ETT), and to observe for changes in minute ventilation (MV) to exclude hypoventilation events.

Methods:

Double-blinded, randomized clinical trial of CO₂ versus air insufflation for EGD with airway protection by either LMA or ETT. Tidal volume, respiratory rate, MV, and EtCO₂ were automatically recorded every minute. Cohort demographics were described with descriptive characteristics. Variables including the percent of children with peak, transient EtCO₂ ≥ 60 mmHg were compared between groups.

Results:

One hundred ninety-five patients were enrolled for 200 procedures. Transient elevations in EtCO₂ of ≥60 mmHg were more common in the CO₂ group, compared to the air group (16% vs 5%, P = 0.02), but were mostly observed with LMA and less with ETT. Post-procedure pain was not different between groups, but flatulence was reported more with air insufflation (P = 0.004).

Conclusion:

Transient elevations in EtCO₂ occur more often with CO₂ than with air insufflation during pediatric EGD despite protecting the airway with an LMA or, to a lesser degree, with ETT. These elevations were not associated with changes in MV. Although no adverse clinical effects from CO₂ absorption were observed, these findings suggest that caution should be exercised when considering the use of CO₂ insufflation, especially since the observed benefits of using this gas were minimal.

What Is Known.

• Adult studies have demonstrated safety of carbon dioxide (CO₂) insufflation. Most have also shown fewer post-procedural gas-related symptoms when CO₂ gas is used for insufflation over air. These results have been reported with both esophagogastroduodenoscopy (EGD) and colonoscopy.

• Transient elevations of end-tidal CO₂ (EtCO₂) are observed in nonintubated children undergoing EGD without concurrent elevations in transcutaneous CO₂ levels, suggesting that eructated insufflation gas could be responsible.

What Is New

• Transient elevations in EtCO₂ persists in children undergoing EGD with CO₂ insufflation despite airway protection with laryngeal mask airway or endotracheal tube.

• CO₂ insufflation is not associated with changes in minute ventilation or other adverse physiologic or anesthesia events, suggesting that these elevations in EtCO₂ are not clinically significant.

Carbon dioxide (CO₂) is increasingly used over air for endoscopic insufflation in children, largely based on adult studies demonstrating its safety and, in most studies, decreased post-procedure gas-related discomfort (1–5). Pediatric studies remain few (6–11); some of these reported decreased post-procedural discomfort with CO₂ use, especially following colonoscopy (9–11). In one (6), pain reduction was only evident following adjusting for baseline abdominal pain, and no such benefit was observed in another (8). A recent meta-analysis concluded that there is limited evidence to recommend routine use of CO₂ in pediatric endoscopy (12), while another concluded that CO₂ gas is preferred for pediatric colonoscopy (13). Furthermore, peak end-tidal CO₂ (EtCO₂) measurement in 2 of the pediatric studies showed higher EtCO₂ in the CO₂ group, raising concerns about systemic absorption of CO₂ (6,8).

Our previously published randomized clinical trial at the University of Iowa, comparing CO₂ and air insufflation for esophagogastroduodenoscopy (EGD), revealed that the use of CO₂ in children resulted in an increased frequency of transient (usually ≤ 1 minute) elevations in EtCO₂ (≥ 60 mmHg) (8). These children did not have a protected airway, and we suspected systemic absorption of the insufflation gas. However, in a follow-up nested study, these peaks in EtCO₂ were not associated with elevations in transcutaneous CO₂ (TCO₂), providing reassurance that systemic absorption was not occurring (7). We concluded that the observed peaks in EtCO₂ during EGD were likely due to eructation of insufflated CO₂. Some evidence of absorption of CO₂ during colonoscopy has been reported by Dharmaraj et al (6), who noted EtCO₂ elevations during this procedure. Indeed, the reason for using CO₂ insufflation during colonoscopy is the expectation of gradual systemic absorption of the gas, reducing intestinal distention. However, use of CO₂ in upper endoscopy carries the potential additional hazard of inhalation of concentrated CO₂ escaping into the upper airway. Our prior observations of brief spikes of EtCO₂ > 60 mmHg without elevated systemic (transcutaneous) CO₂ suggests this to be a potential hazard.

To better examine the hypothesis that eructated CO₂ is responsible for transient elevations in EtCO₂ during EGD in nonintubated children, we designed this study to discover if frequent spikes in EtCO₂ would still be detected with a protected airway, sampling directly from an endotracheal tube (ETT) or laryngeal mask airway (LMA) to minimize contamination by eructated CO₂. We also sought to determine if changes in ventilation were associated with any observed elevations of EtCO₂ in these children, as decreased gas exchange could also cause increased systemic CO₂ and thus ETCO₂.

MATERIALS AND METHODS

Design

This is a double-blinded, prospective, randomized clinical trial that compared CO₂ versus air insufflation in children with a protected airway undergoing endoscopic procedures. The study was registered as a phase III clinical trial with Clinicaltrials.gov (NCT04541667) and approved by the Institutional Review Board (IRB) at University of Nebraska Medical Center.

Subject Enrollment

Children between 6 months to 19 years of age who were about to undergo an endoscopic procedure at the Children’s Hospital and Medical Center Omaha were recruited. In all cases, these were previously planned to be performed with a protected airway; either LMA or ETT. Patients were enrolled after obtaining age-appropriate consent, or consent with assent, as required by the IRB. Procedures included mostly diagnostic EGD-related procedures, including consecutive EGD and colonoscopy, EGD only, and enteroscopy. One patient who had colonoscopy only was enrolled but was not included in the outcome analysis. Patients with American Society of Anesthesiology (ASA) Physical Status Classification System of >2, children with chronic lung disease, and children who are wards of the state were excluded. Spanish-speaking families were enrolled with the help of interpreters. Families requiring the services of other language interpreters were excluded.

Randomization

Participants were randomized 1:1 for air versus CO₂ by the endoscopic nurse or surgical technologist using a sealed envelope system. The endoscopy gas delivery system was set up to deliver either air or CO₂ after randomization and was then covered to maintain blinding of the endoscopist. The anesthesia team, patients, and endoscopists were all blinded to gas being used. Additionally, the biostatistician was blinded to the randomization gas during analysis. The method of airway protection used (ETT or LMA) was left to the discretion of the anesthesiologist and was not randomized.

Data Collection

Data collected included patient demographics, patient’s medical problems or comorbidities, ASA status, indication for procedure, procedure duration, presence of a fellow trainee, year of training, and whether there were any procedure-related complications. Anesthesia-related characteristics were also recorded, including anesthesia duration, agents and doses used, and any anesthesia-related complications, such as hypotension, hypertension, or hypoxia. Anesthesia was provided by a pediatric anesthesiologist.

Ventilation was recorded continuously via direct connection to the LMA or ETT, including respiratory rate (RR) and tidal volume (TV). EtCO₂ levels were also continuously monitored. All these values were recorded every minute. Oxygenation was monitored continuously by pulse oximetry. Data was entered in a secure institutional database, REDCap.

Outcomes

Primary

• Minute ventilation (MV) = TV × RR

• Hypercapnia; defined as EtCO₂ level ≥ 60 mmHg, at any point during the procedure.

Secondary

• Abdominal discomfort as assessed by the nurse or reported by the patient using FACES scale or Visual Analogue scale both pre- and post-procedure.

• Patient-reported pre- and post-procedure headache, nausea, emesis, abdominal bloating, flatulence, and belching.

• Procedure or anesthesia-related complications.

Post-procedure outcomes were reported or assessed in the recovery room when the participants awakened from anesthesia.

Statistical Analysis

Cohort characteristics were summarized using counts and percentages for categorical data, and medians and range for continuous data. The Mann-Whitney test was used to compare continuous variables between the groups, while Fisher exact test was used to compare categorical variables. The peak value of EtCO₂ over the duration of the procedure was determined for each patient. The percent of children with at least 1 elevation of EtCO₂ ≥ 60 mmHg was compared between groups. For MV, adjustment for the patient’s weight was done by dividing the patient’s mean MV by the patient’s weight and compared between groups. SAS software version 9.4 was used for data analysis (SAS Institute Inc., Cary, NC) and a P value <0.05 was considered statistically significant.

Power Calculation

When designing the study, we planned to enroll a sample size large enough to detect small differences in mean EtCO₂ between groups. A sample size of 75 participants per group was estimated to achieve 80% power to detect a small difference in mean EtCO₂ between groups assuming a significance level (alpha) of 0.05 using a 2-sided 2-sample t test. We aimed to enroll 100 patients in each group to accommodate for a 25% drop out rate.

RESULTS

A total of 195 patients were enrolled for 200 procedures from November 18, 2019 to July 6, 2021. One hundred two procedures were randomized to the CO₂ group, while 98 procedures were randomized to the air group. One patient had a colonoscopy-only procedure and was excluded from analysis. A procedure involving both an enteroscopy and a colonoscopy was excluded from the primary outcome analysis because the procedure duration was the longest in the cohort and did not match any other procedure in the other group. A procedure that did not have any recorded MV values was excluded from the MV analysis. Thirteen patients who were approached for the study declined participation and 1 family consented for the study but withdrew consent before study procedures were started. Therefore, a total of 197 procedures were included in all the analysis.

Characteristics of the Cohort (Table 1)

TABLE 1.

Characteristics of cohort

	CO2	Air	P value
Age, y, median (IQR)			0.92
Race/ethnicity (%)	12 (7)	12 (7)	0.83
Caucasian	82	86
African American/African	1	1
Hispanic	6	6
Other	1	1
Two or more races	10	6
Sex (%)			0.57
Male	45	41
Female	55	59
Weight, kg, median (IQR)	53.2 (40.8)	47 (31.7)	0.38
BMI/weight for length z score, median (IQR)	0.61 (2.08)	0.23 (2.05)	0.29
Duration of procedure, min, median (IQR)	32 (28)	31 (34)	0.77
EGD only + EGD portion of combined procedures	10 (7)	10 (7)	0.33
Colonoscopy	30 (16)	28 (16)	0.77
Anesthesia duration, min, median (IQR)	55.5 (30)	52.5 (36)	0.49
Primary GI symptom (%)			0.77
Abdominal pain	63	60	0.6
Diarrhea	20	23	0.64
Feeding or growth concerns	9	11	0.33
Dysphagia	7	11	0.38
Other	66	59
Procedure type (%)			0.97
EGD only	29	32
EGD/colonoscopy	67	68
Enteroscopy	1	0
Colonoscopy only	1	0
Personnel involved (%)			0.69
Faculty only	47.1	53.1
Faculty + 1st year fellow	17.7	16.3
Faculty + 2nd year fellow	18.6	19.4
Faculty + 3rd year fellow	16.7	11.2
ASA status (%)			0.32
ASA 1	21	28
ASA 2	79	72
Airway protection type (%)			0.07
LMA	71	83
ETT	29	17

Other symptoms include bloody stools, fatigue. BMI z score (children ≥ 2 years); weight for length z score (children < 2 years).

ASA = American Society of Anesthesiology; BMI = body mass index; EGD = esophagogastroduodenoscopy; ETT = endotracheal tube; GI = gastrointestinal; IQR = interquartile range; LMA = laryngeal mask airway.

Patient characteristics, including age, body mass index z score, weight (in kilograms), gender, race/ethnicity, ASA status, procedure, and anesthesia duration were similar in both groups. About 57% of the cohort were females and 84% Caucasians. Seventy-six percent of the cohort had an ASA status of 2. Around half of the procedures in each group involved a trainee while half was done by faculty alone. LMA was used to protect the airway in 77% of the procedures and ETT in the remainder. The median age was 12 years in both groups and about 9% of the cohort was below 4 years.

Outcomes

Primary

The percentage of participants who experienced any episode of hypercapnia (EtCO₂ ≥ 60 mmHg) was higher in the CO₂ group when compared to the air group (21% vs 9%, P= 0.02; see Fig. 1). Out of the 136 participants who had consecutive EGD and colonoscopy, 29% experienced hypercapnia at some point during their procedure, 23% in the CO₂ group, versus 8% in the air group (P = 0.02). Hypercapnia events were more common in the upper endoscopy portion of the procedure in the CO₂ group (21%) when compared to the air group (5%) (P = 0.02) but were not significantly different during the colonoscopy portion of procedures (10% vs 5%, P= 0.33). Episodes of hypercapnia in the CO₂ group were significantly more common than air with the use of LMA (22% vs 10 %, P = 0.045). This was less apparent in the ETT group (17% in the CO₂ group vs 0% in the air group, P = 0.14). The median value of MV was similar in both groups; 75.9 mL/kg/min in the CO₂ group and 74.3 mL/kg/min in the air group (P = 0.45). Median MV measured over 5 minutes before peak EtCO₂ was 53.28 mL/kg/min in the CO₂ group versus 40.23 mL/kg/min in the air group (P = 0.20) and was 68.5 mL/kg/min in the CO₂ group during the 5 minutes after episodes of hypercapnia when compared to 49.5 mL/kg/min in the air group (P = 0.067). Multiple spikes of EtCO₂ ≥ 60 mmHg were observed more often in the CO₂ group (P = 0.03) and, although many of the EtCO₂ spikes lasted no more than 1 minute, 2 patients in the CO₂ group had prolonged episodes of elevated EtCO₂, lasting up to 10 minutes.

FIGURE 1.

Percent of procedures with transient spike in EtCO₂ level ≥ 60 mmHg. *: P value <0.05; ns: P value is not significant. EtCO₂ = end-tidal carbon dioxide.

Secondary

Patient-Reported or Nursing-Assessed Outcomes

Table 2 lists the percentages of pre- and post-procedure patient-reported or nursing-assessed outcomes in both groups. Although most of the pre- and post-procedure patient-reported symptoms were similar between groups, patient reported pre-procedure bloating was higher in the air group when compared to the CO₂ group (9% vs 2%, P = 0.02). Additionally, more patients in the air group reported flatulence post-procedure (42% vs 20%, P = 0.004). Furthermore, the percent change in flatulence from pre- to post-procedure remained significantly higher in the air group (33.8%) when compared to the CO₂ group (17.5%) with P= 0.034 for all procedures and 42.6% (air) versus 19.4% (CO₂) for EGD/colonoscopy procedures with P = 0.013. There was no significant change in flatulence between groups from pre- to post-procedure in EGD-only procedures. Patient reported pre-procedure abdominal pain was associated with post-procedure abdominal pain but was not different between groups.

TABLE 2.

Patient reported clinical outcomes (pre- and post-procedure)

Post-procedure patient reported outcomes and hypercapnia % procedures	CO2	Air	P value
Belching pre-procedure (%)
All procedures	5	1	0.38
EGD only	0	0
EGD + colonoscopy	6	2	0.37
Belching post-procedure (%)
All procedures	14	12	0.81
EGD only	0	0
EGD + colonoscopy	18	15	0.80
Bloating pre-procedure (%)
All procedures	2	13	0.02
EGD only	0	0
EGD + colonoscopy	3	16	0.02
Bloating post-procedure (%)
All procedures	11	16	0.47
EGD only	6	0	1
EGD + colonoscopy	13	20	0.32
Flatulence pre-procedure (%)
All procedures	4	9	0.3
EGD only	6	0	1
EGD+ colonoscopy	3	11	0.14
Flatulence post-procedure (%)
All procedures	20	42	0.004
EGD only	19	0	0.23
EGD+ colonoscopy	21	53	0.0005
Nausea pre-procedure (%)
All procedures	13	6	0.15
EGD only	7	10	1
EGD + colonoscopy	16	5	0.05
Nausea post-procedure (%)
All procedures	7	8	1
EGD only	4	3	1
EGD + colonoscopy	7	9	0.76
Emesis pre-procedure (%)
All procedures	4	2	0.68
EGD only	3	3	1
EGD + colonoscopy	3	2	1
Emesis post-procedure (%)
All procedures	1	1	1
EGD only	3	0	0.48
EGD + colonoscopy	0	2	0.48
Headache pre-procedure (%)
All procedures	7	5	0.77
EGD only	7	3	0.61
EGD + colonoscopy	7	6	1
Headache post-procedure (%)
All procedures	13	13	1
EGD only	7	17	0.42
EGD + colonoscopy	16	11	0.46
Abdominal pain pre-procedure (%)
All procedures	24	23	1
EGD only	13	21	0.65
EGD + colonoscopy	27	24	0.83
Abdominal pain post-procedure (%)
All procedures	19	21	0.84
EGD only	13	0	0.48
EGD + colonoscopy	19	26	0.5
Hypercapnia (EtCO2) ≥ 60 mmHg %
All procedures	21	8	0.016
EGD only	16	5	0.02
EGD + colonoscopy	9	5	0.33

EGD = esophagogastroduodenoscopy; EtCO₂ = end-tidal carbon dioxide.

Procedure and Anesthesia-Related Outcomes

There were no procedural complications during this study. Anesthesia-related complications were minimal and transient and were similar in both groups. A few patients (3 in the CO₂ group and 2 in the air group) experienced hypotension, 2 patients in both groups had hypertension, and 1 patient in the air group developed tachypnea. Procedure duration was similar in both groups.

DISCUSSION

The findings from this clinical trial add to existing knowledge and fill a gap created by earlier pediatric studies of endoscopic insufflation using CO₂ (6–11). Our previous study (8) revealed that transient EtCO₂ spikes (≥60 mmHg) are seen when CO₂ is used for insufflation during pediatric upper endoscopy. We now show that these also occur with the use of a protected airway, although more common with LMA than ETT. Observed peaks in EtCO₂ were not associated with changes in MV or adverse events. The only observed benefit of CO₂ use was decreased post-procedure flatulence.

Although EtCO₂ has been shown to be unreliable in the detection of changes in ventilation in nonintubated patients (14–16), it provides an accurate estimate of ventilation in intubated patients (17) during routine anesthesia. Nevertheless, the use of CO₂ as an insufflation gas during endoscopy is an additional variable that must be considered during EtCO₂ monitoring. The LMA has been shown to be as effective for ventilation and gas exchange as the ETT for healthy children undergoing EGD (18). Therefore, we theorized that protecting the airway with either an LMA or ETT should minimize the possibility of contamination of sampled end-tidal gas by eructated CO₂ as we had proposed to be the cause of elevated EtCO₂ with CO₂ endoscopic insufflation (7,8).

The persistence of brief episodes of elevated EtCO₂ despite use of airway protection in this study needs explanation. Reduced ventilation of the patients was likely not responsible, as MV was not diminished prior to events. Although we did not monitor systemic absorption in this study, our prior published work (7) showed no systemic hypercarbia, measured by TCO₂ monitoring, in association with increased EtCO₂ events. Thus, it is possible that there is no significant absorption of CO₂ from the upper gut, or that it is so gradual as to not be detectible by that technique. Another possible explanation is that some of the insufflation gas may leak into the trachea despite the presence of an ETT or LMA. Since fraction of CO₂ gas in the exhaled breath is very small, any small trace contamination from exogenous source (leakage of insufflated CO₂) causes significant elevation in ETCO₂. The brevity (<1 minute) of these events, and the apparent higher incidence of EtCO₂ elevations in the LMA group, compared to the ETT group, suggests this hypothesis. It is certainly possible that this device is more likely than the ETT to allow eructated CO₂ to enter the lower airway. However, the selection of ETT versus LMA was not randomized, and further studies comparing ETT versus LMA, would be required to confirm or refute this hypothesis. Regardless of the actual mechanism, it is clear that transient elevations of EtCO₂ occur during CO₂ insufflation. Fortunately, no patient in either group experienced any adverse events associated with the use of CO₂.

One can question that, even in the absence of adverse effects, are the benefits of CO₂ insufflation worth any associated respiratory risk? Adult studies report improved patient comfort following CO₂ insufflation for a variety of procedures (1,19–21), including upper endoscopy (21). Many of these studies found only subjectively reported benefits of CO₂, such as decreased patient-reported bloating or discomfort, without objective findings by observers (22). Findings from pediatric studies are not consistent. Some reported decreased patient complaints of discomfort with CO₂ insufflation for colonoscopy (6,9,11). Kresz et al (10) found a higher need for additional narcotics in the air group during the procedure, a highly unusual occurrence in our practice, regardless of the insufflation gas used. Dharmaraj et al (6) only found a decreased post-procedural abdominal discomfort in the CO₂ group on multivariate analysis after adjusting for a pre-operative diagnosis of abdominal pain. In this study, we did not find a difference in patient-reported or nursing-assessed abdominal pain, even in patients who underwent consecutive EGD/colonoscopy, and after adjusting for a pre-operative diagnosis of abdominal pain. Although patients in the air group reported more flatulence after the procedure than the CO₂ group, they also reported more bloating pre-procedure.

CONCLUSIONS

In conclusion, we observed a significantly higher number of brief EtCO₂ spikes in children receiving CO₂ instead of air for endoscopic insufflation, a difference that occurs even with airway protection, and more commonly with LMAs than with ETTs (8). These EtCO₂ spikes were not caused by reduced MV and were not physiologically significant, given the lack of any associated adverse events. In addition, our previous study (7) showed that these transient spikes of EtCO₂ were not associated with any detectable elevation of transcutaneous (systemic) CO₂ levels. Nevertheless, given the lack of substantial observed benefits of CO₂ insufflation for upper endoscopy in children, the increased costs of its use (23), and the possible safety concerns suggested by transient spikes in EtCO₂ levels of ≥60 mmHg, the routine use of CO₂ for all pediatric upper endoscopy procedures must be questioned. For more complex or prolonged procedures, such as gastrostomy placement, enteroscopy, or use of electrocautery, its potential benefits may be more substantial. In these cases, minimizing the amount of peritoneal gas accumulation or the risk spark-induced combustion may be significant.

Acknowledgments

We would like to thank the pediatric gastroenterology faculty and fellows, pediatric anesthesiology faculty and fellows, pediatric endoscopy suite staff, and recovery nurses at Children’s Hospital and Medical Center particularly Nancy Nielsen RN who led the study randomization, the study research coordinators: Evan Roberts MBA, Kym Abraham RN, Denise Hoover BS (Pediatric Research Office), data management team: Alicia Bremer MSN APRN and Cynthia Hug BS (Children’s Hospital and Medical Center Omaha), and Elizabeth Patterson GI administrative assistant for all their help with this study.