Abstract
BACKGROUND: The risk for severe hypoxemia during endotracheal intubation is a major concern in the ICU, but little attention has been paid to CO2 variability. The objective of this study was to assess transcutaneously measured partial pressure of CO2 (
) throughout intubation in subjects in the ICU who received standard oxygen therapy, high-flow nasal cannula oxygen therapy, or noninvasive ventilation for preoxygenation. We hypothesized that the 3 methods differ in terms of ventilation and CO2 removal.
METHODS: In this single-center, prospective, observational study, we recorded 
from preoxygenation to 3 h after the initiation of mechanical ventilation among subjects requiring endotracheal intubation. Subjects were sorted into 3 groups according to the preoxygenation method. We then assessed the link between 
variability and the development of postintubation hypotension.
RESULTS: A total of 202 subjects were included in the study. The 
values recorded at endotracheal intubation, at the initiation of mechanical ventilation, and after 30 min and 1 h of mechanical ventilation were significantly higher than those recorded during preoxygenation (P < .05). 
variability differed significantly according to the preoxygenation method (P < .001, linear mixed model). A decrease in 
by > 5 mm Hg within 30 min after the start of mechanical ventilation was independently associated with postintubation hypotension (odds ratio = 2.14 [95% CI 1.03–4.44], P = .039) after adjustments for age, Simplified Acute Physiology Score II, COPD, cardiac comorbidity, the use of propofol for anesthetic induction, and minute ventilation at the start of mechanical ventilation.
CONCLUSIONS: 
variability during intubation is significant and differs with the method of preoxygenation. A decrease in 
after the beginning of mechanical ventilation was associated with postintubation hypotension. (ClinicalTrials.gov registration NCT0388430.)
- intubation
- transcutaneous blood gas monitoring
- intensive care unit
- preoxygenation
- mechanical ventilation
- hypotension
Introduction
Endotracheal intubation is a frequently performed procedure in the ICU that has been shown to be associated with severe complications, including profound hypoxemia.1 Several studies on oxygenation parameters and potential strategies to prevent this life-threatening complication have been conducted.1-5 However, little data are available about the evolution of 
during endotracheal intubation.6-8 Nevertheless, it may be useful to estimate 
throughout intubation, as its variability may play a role in the development of postintubation hypotension.9
To date, the continuous monitoring of 
has been performed with end-tidal CO2 pressure (
) assessments. This measurement is the recommended method to confirm the appropriate placement of the endotracheal tube after intubation and to detect an esophageal intubation,10 but 
only provides data at the beginning of mechanical ventilation and not during preoxygenation and apneic oxygenation. Furthermore, it does not provide any reliable estimates of 
because factors such as spirometric changes, ventilation system leaks, respiratory dead space, and cardiac output can affect the 
gradient.11-13 The transcutaneously measured partial pressure of CO2 (
) is another reliable continuous method that is available to estimate 
in the ICU within a few minutes.13-16 Because 
estimation requires a transcutaneous sensor, this technique is not disrupted by spirometric changes and enables 
variability to be recorded during endotracheal intubation, including before the start of mechanical ventilation (ie, before, during, and after apnea).
Alveolar ventilation during this preoxygenation and its effect on 
may differ depending on whether subjects receive standard O2, high-flow nasal cannula oxygen therapy (HFNC), or noninvasive ventilation (NIV) during preoxygenation. NIV is associated with a high flow of oxygen and increases in the mean airway pressure and PEEP.17 HFNC is associated with a flow of oxygen exceeding the peak inspiratory flow and a positive pressure at approximately the 2–3 cm H2O level, and it may have an effect on CO2 removal from anatomic dead space. Standard O2 is associated with a lower flow than NIV and HFNC and does not create a positive pressure in the airways.1,4,18 
has significant effects on vascular tone. Rapid variability in 
throughout intubation may lead to an imbalance between the direct action of hypercapnic acidosis, which inhibits cardiac contractility and reduces vascular tone, and a counterbalanced response of the sympathoadrenal system, leading to an increase in cardiac output.19-22 We hypothesized that 
varies widely during endotracheal intubation in critically ill subjects and that the variability of 
differs by the preoxygenation method used. We therefore conducted a single-center, prospective, observational study to assess the variability in 
during endotracheal intubation in the ICU. The secondary aims were to assess the agreement of this measure with 
and to investigate the link between the variability in 
during intubation and the development of postintubation hypotension.
Quick Look
Current Knowledge
Endotracheal intubation is frequently performed in ICUs and is associated with complications such as hypotension. Little evidence is available about 
variability during this procedure. Physiological studies have shown a link between a 
decrease and the development of hypotension, but clinical studies have not confirmed these findings after intubation.
What This Paper Contributes to Our Knowledge
There is major variability in transcutaneously measured partial pressure of CO2 during endotracheal intubation in the ICU. A decrease in transcutaneously measured partial pressure of CO2 after the start of mechanical ventilation was associated with the development of hypotension postintubation, which is consistent with the results of physiological studies.
Methods
Trial Design
This prospective, observational study was conducted from May 2018 to June 2019 in a mixed 21-bed ICU in Rennes University Hospital, a teaching hospital in Rennes, France. This study was approved by the hospital’s ethics committee (18.32), and the database was declared to the Commission Nationale Informatique et Libertés. The study was registered with ClinicalTrials.gov (NCT0388430). Verbal and written information was given to the subjects, as required by the French law.
Subjects > 18 y old who were hospitalized in the ICU and undergoing endotracheal intubation were eligible for the study. Patients for whom tracheal intubation was needed immediately (ie, without enough time to set up the 
sensor) and patients who refused or whose relatives refused to participate in the study were not included. In addition, subjects could not be included twice in the study. The evaluation period spanned from the decision to intubate before preoxygenation to 3 h after the start of mechanical ventilation.
Measurement
A 
sensor (TCM5, Radiometer, Copenhagen, Denmark) was placed before preoxygenation on the subject’s chest, as recommended by the manufacturer. The transcutaneous measurement of 
is based on the phenomenon of CO2 gas diffusing through the skin. CO2 is detected by a sensor placed on the subject’s skin; the diffusion of CO2 is increased by heating the sensor between 42°C and 44°C. CO2 is finally measured electrochemically by determining the change in pH in an electrolyte solution separated from the skin by a highly permeable membrane. The change in pH is considered proportional to the logarithm of 
. A temperature correction is also performed to avoid errors in measurements related to CO2 produced by heating the skin.23-25 The sensor was placed at the upper part of the chest between the collarbone and the nipple. This sensor allows 
to be measured every second. As the sensor requires a few minutes for calibration, 
was considered stable when its value varied by < 1 mm Hg within 1 min. The sensor’s membrane was changed every 28 d, following the manufacturer’s recommendations, to obtain reliable measurements.
Procedures
The intubation and preoxygenation procedures were performed in the same way as in usual care because of the observational nature of our study. The type of preoxygenation device was therefore selected by the physician: standard O2 involved a nonrebreather mask or bag-valve-mask, while HFNC and NIV were the available alternatives. Preoxygenation occurs before the induction of anesthesia. In accordance with our usual care protocol, preoxygenation was performed in the semi-recumbent position at 30° for 3–5 min according to the guidelines previously reported for preoxygenation.26 Rapid-sequence induction using a hypnotic drug (ie, etomidate, propofol, pentothal, or ketamine) was performed, and either succinylcholine or rocuronium was administered before immediate intubation. Endotracheal intubation was performed by the resident or the physician in charge of the subject. The residents were considered junior operators, and the doctors of medicine were considered senior operators. During the apneic period before the first endotracheal intubation attempt, no bag-mask ventilation was provided. Its use was permitted after in case of emergency by the physician in charge. After endotracheal intubation, the subjects were ventilated using the volume control mode with a tidal volume of 6–8 mL/kg of predicted body weight. The breathing frequency, PEEP, and 
were left to the discretion of the physician. Sedation and analgesia, including either midazolam or propofol and morphine, were started immediately after intubation. A neuromuscular block with cisatracurium was introduced if it was considered necessary by the physician in charge of the subject. Throughout the procedure, fluid therapy was not mandatory but was provided when the physician requested it.
Definitions and Subgroup Analyses
Three groups were established according to the preoxygenation method chosen by the physician: standard O2, HFNC, and NIV. In the standard O2 group, 
was estimated as 0.21 + oxygen flow × 0.03.1-4
Postintubation hypotension was defined as the recorded systolic blood pressure being < 65 mm Hg or < 90 mm Hg despite 500–1,000 mL of fluid loading or the introduction or enhancement of vasoactive support.27,28 This measure was recorded within 3 h after endotracheal intubation.
Adverse events during endotracheal intubation were classified as severe (eg, death due to any cause, cardiac arrest, 
< 80%, severe hypotension defined by a systolic blood pressure < 80 mm Hg) or moderate (eg, ventricular or supraventricular arrhythmia requiring intervention, esophageal or selective intubation, macroscopic aspiration, dental injury, difficult intubation defined by the need for > 3 laryngoscopies or the need to call another senior physician).1,27
ARDS was defined according to the Berlin definition.29 According to the Task Force criteria,30 the diagnosis of COPD was considered in subjects with symptoms of chronic cough and sputum production or dyspnea in addition to history of smoking. The diagnosis was retained when the postbronchodilator FEV1/FVC was ≤ 0.7 and not fully reversible on a previous spirometry. Subjects with no previous results for pulmonary function tests were screened for emphysema via chest radiograph and via scanner when available, and for intrinsic end-expiratory pressure during mechanical ventilation.
Obesity was defined as a body mass index of ≥ 30 kg/m2, and obesity-hypoventilation syndrome was defined as a combination of obesity, daytime hypercapnia, and sleep-disordered breathing.31 Obesity-hypoventilation syndrome was diagnosed before admission to the ICU. Restrictive disease was diagnosed in subjects with restrictive patterns on the pulmonary function test and a reduced vital capacity, residual volume, and total lung capacity.32 Subjects with bilateral bronchiectasis, asthma, and lung cancer were classified as “other” underlying respiratory disease. Chronic cor pulmonale was diagnosed when right-ventricular dilation and dysfunction associated with pulmonary hypertension were present.33 Right-ventricular dilatation was considered when the apical right ventricle at base was > 41 mm and right-ventricular dysfunction when tricuspid annular plane systolic excursion was < 17 mm or measurement of tissue Doppler of the free lateral wall < 0.095 ms−1.
Data Collection
was recorded at several time points during and after the intubation procedure: before preoxygenation, at the induction of anesthesia, at endotracheal intubation, at the initiation of mechanical ventilation, and at 30 min, 1 h, 2 h, and 3 h after the initiation of mechanical ventilation. The highest 
and the lowest 
observed during the endotracheal intubation procedure were recorded by a nurse or a physician not involved in the procedure. The time from anesthetic induction to intubation was also noted. The time required for the sensor calibration was retrieved from the sensor’s internal memory. All 
values recorded during the study were double checked against those recorded in the sensor’s internal memory. 
values reported were excluded if they varied by > 1 mm Hg within a second to avoid false measures related to noise in the signal. These values were considered incorrect. Technical issues leading to data collection failure, such as accidental or voluntary removal, were also recorded.
Hemodynamic and respiratory parameters were collected at the same time as were 
values. Demographic characteristics of the subjects, main reason for intubation, preoxygenation and intubation procedures, and complications were also recorded. In addition, we determined the severity of illness by using the Sequential Organ Failure Assessment (SOFA)34 and the Simplified Acute Physiology Score II (SAPS II) tools.35 When requested by the physician, the blood gas results obtained simultaneously with the 
measures were collected. The simultaneous 
value was noted. The 
values < 35 and > 45 mm Hg were reported, based on normal 
levels recorded in our laboratory.
End Points
The primary end points were the assessment of the variability of 
during endotracheal intubation and within the first 3 h of mechanical ventilation initiation and the comparison of the 
values performed according to the preoxygenation method used. The secondary end points were the reliability of 
to estimate 
and the relationship between the changes in 
and the onset of postintubation hypotension.
Statistical Analysis
Continuous variables are expressed as medians with interquartile ranges (IQR); comparisons between 2 groups were performed using the Mann-Whitney U test, and comparisons across 3 or more groups were performed with the Kruskal-Wallis test. The proportions were compared using the chi-square test or Fisher exact test, as required. Repeated measurements for 
that were recorded before, during, and after intubation were first compared according to whether the subjects underwent standard O2, HFNC, or NIV using one-way analysis of variance and the Bonferroni Dunn test for post hoc analysis. In addition, a linear effects mixed model was used to determine association of the preoxygenation method used with change of 
values during the study period. The continuous variables of age and body mass index were dichotomized for analysis of repeated measures. Subjects were differentiated by whether they were > 65 y old and whether they were obese. In addition, subjects were differentiated depending on whether minute ventilation increased by ≥ 100 mL/min from the value recorded at onset of mechanical ventilation and the value recorded after 30 min of mechanical ventilation; the threshold of 100 mL/min was arbitrarily chosen.
The level of agreement between 
and 
was assessed using a Bland-Altman plot36 when the blood gas values were measured at the same time that the 
measurements were available. Receiver operating characteristic curves were constructed to determine the threshold values of 
associated with postintubation hypotension. 
during preoxygenation, at induction, at intubation, and at the initiation of mechanical ventilation and the difference in 
between the initiation of mechanical ventilation and 30 min after the start of mechanical ventilation (ie, δ
-MV30min) were used. The areas under the curve were compared based on their associated P values to determine which value had the best predictive ability for postintubation hypotension.
We conducted logistic regression analysis to determine whether this value was independently associated with postintubation hypotension. On the basis of previous studies,9,27,37-39 age, SAPS II score, COPD, cardiac comorbidity, the use of propofol for anesthetic induction, and minute ventilation at initiation of mechanical ventilation were selected as the variables to be used for adjustment. The results are expressed as odds ratios with 95% CIs. All probability values that are reported are two-sided. We considered P values < .05 to be statistically significant. The statistical analysis was performed using SPSS 25 (IBM, Armonk, New York) and MedCalc 18.6 (MedCalc BVBA, Ostend, Belgium).
Results
From February 2018 to March 2019, 303 patients were intubated in the ICU; 202 subjects were included in the study. The main reason for exclusion was that endotracheal intubation needed to be performed immediately in emergencies, without enough time to calibrate the 
sensor. Ninety-eight subjects (49%) received preoxygenation with standard O2, 43 (21%) received preoxygenation with HFNC, and 61 (30%) received preoxygenation with NIV (Fig. 1).
Flow chart. Measurements were considered unavailable to take if the sensor was already used for another intubation or if there was a technical issue (eg, calibration failure, medical team was too busy). HFNC = high-flow nasal cannula; NIV = noninvasive ventilation.
The baseline characteristics of the subjects were compared according to the preoxygenation method used (Table 1). In 5 subjects, COPD was diagnosed on the basis of the clinical and radiological criteria for suspected COPD only.30 For the other COPD subjects, the first pulmonary function test revealing obstructive syndrome were performed, depending on the subject, between 5 y and 6 months before admission to the ICU. Subjects with COPD or ARDS more frequently received preoxygenation with NIV and HFNC, respectively, than did other subjects. Age and body mass index differed significantly by preoxygenation method and were dichotomized as follows: individuals were grouped according to whether they were > 65 y old and whether they were obese. Subjects more frequently received preoxygenation with standard O2 when they were intubated for neurological reasons.
Baseline Subject Characteristics
Data regarding intubation conditions in the 3 groups are presented in Table 2. The number of laryngoscopy attempts was 1 (IQR, 1–2) in the 3 groups of subjects distinguished according to the preoxygenation method used. Ten junior operators failed intubation, which was then successfully performed by the senior operator in charge of the subject’s care. Severe complications occurred significantly more frequently in the HFNC group (P = .048) than in other 2 groups. The respiratory and hemodynamic parameters are shown in Table 3. Notably, the tidal volume per predicted body weight applied did not differ significantly across the 3 groups; in addition, the total PEEP was significantly higher in the NIV group, and the plateau pressure was significantly higher in the NIV and HFNC groups than in the standard O2 group within the 3 h after initiation of mechanical ventilation.
Intubation Parameters at Baseline
Respiratory and Hemodynamic Parameters
Calibration Process and Failure of the Sensor
Technical issues were observed for 13 of 202 subjects (6%) at some time points in the study period, despite appropriate initial calibration and availability of 
results at the time of preoxygenation and anesthetic induction. Consequently, we did not exclude these subjects from analysis. In 8 (4%) cases, the sensor was accidentally removed during care-related procedures. In 1 (0.5%) case, monitoring was interrupted due to electrocardiogram interference. In 2 (1%) cases, there were errors in measurement with major, rapid variability in 
without any sensor stabilization being obtained. In 2 (1%) cases, the technical problem was not specified by the physician. No physical complications related to the sensor, such as burns, were reported during the study. The median calibration time for the sensor was 6 min (interquartile range 5–8), with the values ranging from 1 min to 21 min.
Agreement Between
and
The Bland-Altman analysis of concordance between 
and 
is shown in Figure 2. The 
bias was 0.15 mm Hg, with the 95% limits of agreement ranging from 9.6 to –9.3 mm Hg. 
values > 100 mm Hg were noted in 5 subjects only and were distributed as follows: 102, 107, 116, 135, and 150 mm Hg. The values for 
recorded at the same time were 108, 101, 102, 133 and, 148, respectively.
Bland-Altman plot showing differences in arterial and transcutaneous measurements of partial pressure of CO2 (
) vs the mean of the 2 values among 131 critically ill subjects.
Variability During Endotracheal Intubation
The evolution of 
over time in the overall population is shown in Figure 3A. During intubation and the beginning of mechanical ventilation, we observed a significant increase in 
compared to that noted during the preoxygenation period. Then 
decreased but remained higher than that noted before preoxygenation. At initiation of mechanical ventilation, 151 subjects (75%) had 
values that were not between 35 and 45 mm Hg, including 111 (55%) subjects with 
values > 45 mm Hg.
Repeated measurement of 
before and after the initiation of mechanical ventilation. A: Among all the subjects in the study, and B: according to the preoxygenation method.*P < .05 vs preoxygenation (A) and vs preoxygenation in the standard O2 group (B). †P < .05 vs preoxygenation in the HFNC group (B). ‡P < .05 vs preoxygenation in the NIV group (B). §P < .05 across all 3 groups at all time points (B). 
= transcutaneous measurement of partial pressure of CO2; HFNC = high-flow nasal cannula oxygen therapy; NIV = noninvasive ventilation.
In 154 subjects (76%), median 
measured 1–3 h after the start of mechanical ventilation was significantly higher in subjects who received NIV (53 mm Hg [IQR 47–66]) than in those who received standard O2 (42 mm Hg [IQR 37–49], P < .001 after comparison), but 
did not significantly differ from the subjects with HFNC (44 mm Hg [IQR 39–55]). The variability in 
values according to the preoxygenation method is shown in Figure 3B. Results for univariate mixed model are shown in Table 4. Overall, the 
values differed significantly across the 3 preoxygenation groups and over the period studied (P < .001), and the values were affected by whether subjects had COPD (P < .001), obesity (P = .02), or a neurologic reason for intubation (P < .001), but not by whether they were > 65 y old, had ARDS, had underlying cardiac disease, had increased minute ventilation > 100 mL/min between initiation of and after 30 min of mechanical ventilation (P = .89), or were obese (P = .07). The maximum 
at endotracheal intubation was significantly higher in the NIV group than in the standard O2 group (P < .001). The extreme values recorded at this assessment time were 18–153 mm Hg in the standard O2 group, 22–120 mm Hg in the HFNC group, and 20–120 mm Hg in the NIV group. The preoxygenation method used remained independently associated with the variability of 
after COPD, obesity, and neurological reason for intubation were entered in the mixed model (F value 4.22, P = .01). The 
values decreased after the initiation of mechanical ventilation in the standard O2 and NIV groups and increased in the HFNC group.
Linear Mixed Model Effects for 
Variability
Variability and Development of Postintubation Hypotension
Postintubation hypotension occurred within 3 h of the beginning of mechanical ventilation in 37 subjects (38%) in the standard O2 group, 30 subjects (70%) in the HFNC group, and 40 (66%) subjects in the NIV group (P < .001). In the overall study population, 71 (35%) experienced postintubation hypotension within 1 h of intubation, and 107 subjects (53%) experienced it within 3 h.
The areas under the curve and the prognostic relevance of the 
values in predicting postintubation hypotension are listed in Table 5. Only the area under the curve for δ
-MV30min was found to be significant. The threshold value determined by the receiver operating characteristic curve was 5 mm Hg for δ
-MV30min. δ
-MV30min > 5 mm Hg was significantly associated with postintubation hypotension in both the unadjusted analysis (odds ratio 2.58 [95% CI 1.37–4.88], P = .003) and in the adjusted analysis (odds ratio 2.14 [95% CI 1.03–4.44], P = .039). Among the 71 subjects who experienced postintubation hypotension within 1 h of intubation, 44 (62%) with δ
-MV30min > 5 mm Hg developed hypotension after 30 min of mechanical ventilation. The SAPS II score was also associated with postintubation hypotension in the adjusted analysis (odds ratio 1.04 [95% CI 1.02–1.06], P < .001), as was the need for vasopressors before preoxygenation (odds ratio 10.76 [95% CI 1.32–87.96], P = .034).
Area Under the Curve for 
Values and Occurrence of Postintubation Hypotension
Discussion
This study is a proof-of-concept study, and further studies are required. The 
values recorded during endotracheal intubation within 3 h after initiation of mechanical ventilation in the ICU varied significantly. The first period of variability was observed at the time of apneic oxygenation, as previously reported during general anesthesia.40 Some subjects experienced a quick increase in the 
level, and some subjects exhibited extreme values of hypercapnia during endotracheal intubation. 
variability differed significantly according to whether subjects received standard O2, NIV, or HFNC for preoxygenation, and we noted an association between a decrease in 
within the first 30 min of mechanical ventilation and postintubation hypotension.
Thus, 
appears to be a suitable substitute for 
and can safely be measured during endotracheal intubation, although it cannot provide the exact 
value.13,15,16,41,42 The sensor provided several advantages over those used for 
and 
measurements. First, the sensor quickly provided 
values using a noninvasive method and recorded CO2 values continuously, which cannot be done with blood gas analysis. Furthermore, unlike capnometry, use of the sensor enables the estimation of 
during the apneic period after anesthetic induction and appears to be more accurate for 
estimation than 
in many situations encountered in the ICU, including shock and acute respiratory failure with hypercapnia.11,12,15
The issue that arises is the clinical impact of changes in 
during endotracheal intubation, which can affect the outcome for some subjects. We observed a significant increase in 
values during intubation, and most of the subjects had abnormal values during the study period. Some subjects experienced extreme 
values during intubation. This finding may be explained by apneic oxygenation, which is known to increase 
and its association with susceptibility to hypercapnia in subjects with respiratory diseases. 
should be monitored among patients in whom deleterious effects of 
variability can occur, such as those with brain injury43 or individuals placed on extracorporeal membrane oxygenation.44,45
A clinical implication may be the relationship between CO2 variability and the occurrence of postintubation hypotension. Rapid hypercapnia correction is commonly considered to lead to this complication.46 However, there is little evidence available to support this assumption. Whereas cardiac output and sympathetic tone increase with an increase in CO2,19-21,47 no human studies have demonstrated the link between the correction of CO2 during intubation and the occurrence of postintubation hypotension. Franklin et al9 reported an association between hypercapnia before intubation and the occurrence of hypotension, suggesting a link between CO2 variability and postintubation hypotension. In our study, we observed a relationship between a decrease in 
values within half an hour after the start of mechanical ventilation and the occurrence of postintubation hypotension, independent of the effects of other factors such as a history of COPD or minute ventilation at the initiation of mechanical ventilation. Of note, the recorded 
value was not associated with hypotension at any time during the intubation period. We believe that our results suggest that the risk of hypotension is related less strongly to high values of 
(and consequently to a high level of 
) than to a sharp decrease in 
values and 
induced by the correction of hypoventilation following the initiation of mechanical ventilation. Our results are therefore consistent with studies that have reported that a high level of CO2 or respiratory acidosis is associated with higher cardiac output.19-21
We can hypothesize that a decrease in 
or the correction of acidosis leads to a drop in cardiac output and may induce hypotension. The definition of postintubation hypotension used in this work can be criticized. There are several definitions available28 in the literature, and an evaluation after 1 h is often preferred.27,48 Using this definition, Jaber et al27 reported a rate of cardiovascular collapse of approximately 25% within the first hour after endotracheal intubation. Because we recorded postintubation hypotension within the first 3 h after endotracheal intubation and initiation of mechanical ventilation, the postintubation hypotension incidence was higher than that reported in other studies. Indeed, we found that 53% of the subjects exhibited postintubation hypotension in this time period, with 36 subjects (18%) developing postintubation hypotension after 1 h of mechanical ventilation.
We compared 3 groups that were established according to the preoxygenation technique chosen by the physician. As in previous studies,48 standard O2 was used more frequently than the other techniques. It is noteworthy that, unlike Bailly et al,48 who studied the impact of preoxygenation on pulse oximetry, we included a higher proportion of subjects who underwent HFNC in our cohort. This difference is probably due to this technique being used increasingly more often in the ICU.1,3-5 In our study, 
varied differently over time in the 3 groups, and standard O2 and NIV had the same profile trend, while HFNC had a different profile. These differences probably occurred because there was considerable cardio-respiratory disease heterogeneity across the 3 groups. On the one hand, the use of NIV was linked to a higher rate of COPD with hypercapnia. On the other hand, there were more cases of ARDS in the HFNC group. Impaired respiratory system compliance among these subjects could have led to the occurrence of hypercapnia after the start of mechanical ventilation, explaining the increasing pattern in 
over the study period.49 In more than half of the subjects who experienced postintubation hypotension within 1 h after the start of mechanical ventilation, a δ
-MV30min > 5 mm Hg was noted before hypotension occurred. Thus, we believe that the decrease in 
after intubation is a predictor of hypotension after the beginning of mechanical ventilation and not a consequence of hypotension. However, we also noted that the preoxygenation method has an impact on 
variability independent of the underlying diseases of the patient.
The main strength of our study lies in the exclusion of patients who had various diagnoses and underwent different preoxygenation techniques. In addition, their care in terms of the intubation processes was not modified. We can therefore presume that the variability in CO2 recorded reflects the variability that occurs daily in a medical ICU. Furthermore, the use of this 
sensor during endotracheal intubation in the ICU is an innovative monitoring method that can lead to new clinical research perspectives.
Our study has several limitations to note. It is a single-center study, and it may not be possible to generalize our results to ICUs caring for subjects with particular pathologies or with specific preoxygenation practices. Lung function measurements were not available in all subjects for the diagnosis of the underlying respiratory disease in all subjects, which may have resulted in misclassification of some subjects. The evaluation of the agreement between 
and 
was not possible for all subjects due to the observational nature of the study. Moreover, it would have been more accurate to assess the correlation between 
and 
at all of the time points in the study. As the calibration of the sensor required a few minutes, we were not able to include extreme emergency intubations; consequently, the use of 
monitoring is excluded in emergency circumstances. It would have been interesting to assess the CO2 variability among these subjects because they usually have higher severity scores and are therefore at a higher risk for postintubation hypotension. In addition, given that not all subjects had an arterial catheter, blood pressure measurements were recorded intermittently, so some cases of postintubation hypotension may have been missed. Fluid management was not homogenized at the time of intubation, which could have caused confusion in the analysis of the relationship between CO2 variability and hemodynamic alterations, as it could impact postintubation hypotension incidence,27 although a recent study failed to show an effect of fluid bolus on the occurrence of cardiovascular collapse.50
Additional investigations should focus on controlling 
variability, decreasing the incidence of postintubation hypotension and improving the outcomes among subjects in whom the variability of 
has a potential deleterious effect, such patients with brain injury43 and those receiving with extracorporeal membrane oxygenation.45
Conclusions
Our study describes the evolution of transcutaneous CO2 throughout intubation among critically ill subjects. We observed major changes in 
during and after endotracheal intubation. Among the 3 groups established according to the preoxygenation technique used, the variability in 
differed. We found that 
showed the same trend when standard O2 or NIV was used for preoxygenation but was higher when NIV was used. In contrast, 
had a different trend when HFNC was used for preoxygenation. We also highlighted an association between a decrease in 
after the start of mechanical ventilation and the development of postintubation hypotension. Because we performed a proof of concept study, additional studies need to be conducted to verify our results and specifically, to determine whether postintubation hypotension and the deleterious variability in CO2 can be prevented by the continuous monitoring of 
during endotracheal intubation.
Footnotes
- Correspondence: Aurélien Frérou MD, Service des Maladies Infectieuses et Réanimation Médicale, Hôpital Pontchaillou, 2 rue Henri Le Guilloux, 35033 Rennes cedex 9, France. E-mail: aurelien.frerou{at}gmail.com
A version of this paper was presented by Dr Frérou at the 46th Annual Congress of La Société de Réanimation de Langue Française, held January 23–25, 2019, in Paris.
The authors have disclosed no conflicts of interest.
- Copyright © 2021 by Daedalus Enterprises
References
