Abstract
BACKGROUND: Capnometry detects hypoventilation earlier than pulse oximetry while supplemental oxygen is being administered. We compared the end-tidal CO2 (PETCO2) measured using a newly developed oxygen nasal cannula with a CO2-sampling port and the PaCO2 in extubated subjects after abdominal surgery. We also investigated whether the difference between PaCO2 and PETCO2 is affected by resting, by spontaneous breathing with the mouth consciously closed, and by deep breathing with the mouth closed.
METHODS: Adult post-abdominal surgery subjects admitted to the ICU were enrolled. After extubation, oxygen was supplied at 4 L/min using a capnometry-type oxygen cannula. The breathing frequency, PETCO2, and PaCO2 were measured after 30 min of oxygen supplementation. PETCO2 was continuously measured during rest, during breathing with the mouth consciously closed, and during deep breathing with the mouth closed. The difference between PETCO2 and PaCO2 during various breathing patterns was analyzed using the Bland-Altman method.
RESULTS: Twenty subjects were included. The bias ± SD (limits of agreement) for breathing frequency measured by capnometry compared with those obtained by direct measurement was 0.4 ± 3.6 (−6.7 to 7.4). In PETCO2 compared with PaCO2, the biases (limits of agreement) were 14.8 ± 8.2 (−1.3 to 30.9) at rest, 10.2 ± 6.4 (−2.3 to 22.7) with the mouth closed, and 7.7 ± 5.6 (−3.2 to 18.6) for deep breathing with the mouth closed. PETCO2 determined using the capnometry device yielded unreliable and widely ranging values under various breathing patterns. However, deep breathing with the mouth closed decreased the difference between PETCO2 and PaCO2, as compared with other breathing patterns.
CONCLUSIONS: PETCO2 measurements under deep breathing with mouth closed with a capnometry-type oxygen cannula improved the prediction of the absolute value of PaCO2 in extubated post-abdominal surgical subjects without respiratory dysfunction.
Introduction
Respiratory depression secondary to the residual effects of anesthetics, muscle relaxants, and opioids is a well-recognized hazard in the postoperative period. Therefore, the practice guidelines published by the American Society of Anesthesiologists in 2002 recommended monitoring of respiratory function during recovery for early detection of hypoxemia.1 Assessment of the airway patency, breathing frequency, and SpO2 during the postoperative recovery should be used. However, CO2 monitoring during spontaneous breathing was not clearly mentioned in the most recent guidelines updated in 2013.2
Hypoxia may be a late sign of hypoventilation while oxygen is being administered. Therefore, monitoring of the end-tidal carbon dioxide concentration (PETCO2) using capnometry may be more reliable than SpO2 in detecting hypoventilation.3 Capnography has been shown to be a more reliable predictor of hypoxemia than pulse oximetry. In sedated and postoperative patients receiving supplemental oxygen, capnography can detect hypoventilation faster and more reliably than pulse oximetry.4,5 Furthermore, Soto et al6 have reported that monitoring of nasal capnography is as reliable as plethysmography.
PETCO2 monitoring is achieved using 2 methods (ie, mainstream and sidestream monitoring). In mainstream monitoring, CO2 is detected directly during a ventilation circuit. On the other hand, sidestream approaches continuously draw CO2 from the patient's airway to the monitor. Therefore, CO2 detection by sidestream monitoring is sometimes unstable, due to occlusion of the sampling tube and unstable breathing in non-ventilated patients.7 Several studies have reported that mainstream monitoring was superior to sidestream monitoring for detection of PETCO2, in both ventilated and non-ventilated subjects.8–10 However, oral guidance could improve the performance of PETCO2 detection by using a nasal canulla for sidestream capnometry.10 Therefore, alteration of the breathing pattern from mouth breathing to nose breathing may improve the performance of PETCO2 detection and the difference between PETCO2 and PaCO2.
In a previous study, we found that an oxygen mask-based capnometry device was useful for measuring CO2 postoperatively.11 In this study, deep breathing reduced dissociation between PETCO2 and PaCO2. This newly developed capnometry-type nasal cannula (Japan Medical Next, Tokyo, Japan) has a unique sampling line inside the cannula for use in spontaneously breathing patients (Fig. 1). Each lumen of the cannula is divided into 2 parts, one part supplying oxygen and the other sampling air for the measurement of PETCO2. However, one potential drawback of this nasal cannula is that CO2 may not be detectable if the patients breathe through their mouths. Healthy adults normally breathe through the nose, but during the postoperative period, congestion of the nasal mucosa and the use of a nasal drainage tube may impede air flow through the nose as compared with the mouth. These alterations may prompt the patient to switch from nasal to mouth breathing, and both pediatric and adult patients commonly mouth-breathe at rest.12
Capnometry-type nasal cannula used in the study (A). The device has a unique sampling line inside the cannula for use in spontaneously breathing subjects. Each lumen is divided into 2 parts, one part supplying oxygen and the other measuring the end-tidal carbon dioxide concentration (B).
In the present study, we compared the breathing frequency measured directly with that measured by capnometry using a nasal cannula with a CO2 sampling port (capnometry-type nasal cannula) in subjects extubated after abdominal surgery. We also compared the PETCO2 and the pressure of carbon dioxide in the arterial blood (PaCO2). A previous study already reported that mainstream and sidestream measurements with an oral guide device could measure PETCO2 more accurately than a standard sidestream measurement.10 Therefore, in this study, we aimed not only to assess the accuracy of PETCO2 monitoring using this device, but also to determine whether the breathing pattern affects PETCO2 monitoring during post-anesthetic care. We investigated whether PaCO2 and PETCO2 differed significantly during resting, during spontaneous breathing with the mouth consciously closed, and during spontaneous breathing through the nose with the mouth closed.
QUICK LOOK
Current knowledge
Monitoring of SpO2 is common practice during postoperative recovery. In the clinical situation, capnometry is an alternative method to pulse oximetry for the early detection of hypoventilation. Although there is a standard respiratory assessment, the appropriate means of monitoring spontaneous breathing remain undetermined.
What this paper contributes to our knowledge
Various respiratory monitoring devices were assessed, and the present device presented results similar to those in a previous study. The PETCO2 values and breathing frequency were too unstable to allow for the prediction of the absolute values using this capnometry-type nasal cannula. Hence, although direct measurement is the accepted standard for respiratory assessment, this continuous respiratory monitoring device may support postoperative care. Moreover, deep breathing with the mouth closed improved the prediction of PETCO2 values using this capnometry-type nasal cannula.
Methods
Study Design
The ethics committee of our hospital approved the study design, and the protocol is registered at the University Hospital Medical Information Network Center (UMIN000014062). Informed consent was obtained from all enrolled subjects.
Study Setting and Population
The present study was conducted at a teaching hospital in Japan. Adult patients undergoing abdominal surgery and admitted to the ICU postoperatively were included. The nasogastric tube was removed before supplementation of oxygen by our oximetry device. Patients age <18 y, with an SpO2 of <95% during oxygen supplementation at 4 L/min delivered by a nasal cannula, who did not receive epidural anesthesia, who were obese (body mass index >30 kg/m2), and who were undergoing laparoscopy were excluded, because these conditions are likely to affect postoperative respiratory function, such as ventilation-perfusion mismatch.13,14 Patients who experienced apnea, dyspnea, or arterial desaturation (defined as SpO2 <95%) were also eliminated from the analysis. In the present study, we aimed to assess the accuracy of the measurements achieved by using the device. Therefore, we excluded patients who were considered at risk of postoperative respiratory dysfunction. Furthermore, the present study was implemented in the ICU; therefore, patients who were not admitted to the ICU were also excluded. The study period was 90 min for each subject.
Study Protocol
The capnometer was calibrated before oxygen administration. All subjects were extubated in the operating room, and oxygen was supplemented at 4 L/min during transfer to the ICU via a nasal cannula. The cannula was then replaced with a capnometry-type nasal cannula at the same oxygen flow. PETCO2 was measured with a sidestream with the mouth closed and a capnometer via spectrophotometry.
Measurements
Arterial blood samples and PETCO2 measurements were obtained 3 times, once every 30 min. The breathing frequency was only measured with the mouth closed. PETCO2 was measured arbitrarily during rest (PETCO2 at rest), during breathing with the mouth consciously closed (PETCO2 with mouth closed), and during deep breathing with mouth closed (PETCO2 with deep breathing). PETCO2 was measured after a constant and normal-shaped capnography capnogram was confirmed at every 30 min of oxygen administration. PETCO2 was measured with a sidestream monitor (Microstream, AG-400R, Nihon-Kohden, Tokyo, Japan). If the PETCO2 was not detectable even with the nasal cannula appropriately fitted, the data point was recorded as such. All data measurements were collected by a critical care physician with training in respiratory monitoring and capnography.
Data Analysis
Quantitative variables are expressed as the median and interquartile range. The correlation between the breathing frequency by direct measurement and the breathing frequency measured with the capnometer while breathing with the mouth consciously closed was analyzed using the Pearson product-moment correlation coefficient for assessing validity and reliability.
Each variable was measured several times in each subject, and the mean bias, precision, and limits of agreement were estimated by the component of variance technique.15 The bias, precision, and limits of agreement between the PaCO2 and PETCO2 at rest, PETCO2 with mouth closed, and PETCO2 with deep breathing were calculated using the Bland-Altman method. The differences between PETCO2 and PaCO2 under various breathing conditions were analyzed with one-way analysis of variance. All tests were 2-tailed, and P < .05 was considered statistically significant. The statistical analyses were performed using Prism 6 for Mac OS X version 6.9b (GraphPad Software, San Diego, California).
The sample size was calculated based on a power analysis performed using G*power software. The G*power analysis indicated that 17 subjects were required for a 2-sided test with a significance of .05, power of 0.8, and estimated coefficient of determination (r2) of 0.36. Therefore, 23 subjects were enrolled in the present study.
Results
Twenty-three subjects were enrolled, of whom 3 subjects were excluded due to insufficient data collection. The remaining 20 subjects (13 men, 7 women), with a median age of 66 y (interquartile range, 61–72 y) and body mass index of 20–24 kg/m2, were included in the analysis (Table 1). All subjects tolerated the capnometry-type nasal cannula, and the SpO2 remained >95% during oxygen supplementation. The capnogram and PETCO2 at rest were not detectable in 3 subjects (3 of 20, 15%). A normal-shaped capnogram was displayed during closed mouth breathing in all subjects.
Clinical and Perioperative Subject Characteristics
The breathing frequencies during breathing with the mouth consciously closed measured by capnometry and directly measurement were significantly correlated (P < .001; r = 0.66 [0.48–0.78]; r2 = 0.43). The Bland-Altman analysis showed that the bias ± SD (limits of agreement) between the breathing frequency with the 2 different measurement approaches was 0.4 ± 3.6 (−6.7 to 7.4) (Fig. 2).
Bland-Altman plot of breathing frequency between direct and capnometer measurements using the nasal cannula. The bias ± SD (limits of agreement) between the breathing frequency with the 2 different measurement approaches was 0.4 ± 3.6 (−6.7 to 7.4).
The Bland-Altman analyses revealed that the biases ± SD (limits of agreement) between PaCO2 and either PETCO2 at rest, PETCO2 with mouth closed, and PETCO2 with deep breathing were 14.8 ± 8.2 (−1.3 to 30.9), 10.2 ± 6.4 (−2.3 to 22.7), and 7.7 ± 5.6 (−3.2 to 18.6), respectively (Fig. 3). The difference between PaCO2 and PETCO2 was significantly smaller during deep breathing with the mouth closed than at rest, with or without the mouth closed (P < .001, one-way analysis of variance).
Bland-Altman analysis of the end-tidal carbon dioxide concentration (PETCO2) during resting (A), with mouth closed (B), and during deep breathing (C) while administering oxygen supplementation at 4 L/min. The biases ± SD (limits of agreement) were 14.8 ± 8.2 (−1.3 to 30.9) for PETCO2 during resting, 10.2 ± 6.4 (−2.3 to 22.7) for PETCO2 with mouth closed, and 7.7 ± 5.6 (−3.2 to 18.6) for PETCO2 during deep breathing.
Discussion
In the present study, we evaluated an oxygen nasal cannula-based capnometry device in terms of PETCO2 and breathing frequency monitoring in extubated subjects after abdominal surgery. Moreover, PETCO2 was measured under various breathing patterns. The capnography capnogram was not detectable in 3 of 20 subjects when they were breathing in the resting condition, whereas we could observe the capnogram in all subjects when we asked them to breathe with their mouths closed.
In the present study subjects, appropriate pain control could support deep breathing with a consciously closed mouth. As expected, rapid shallow breathing due to pain and respiratory dysfunction was not detected in any of the subjects. This suggests that some postoperative subjects are breathing mainly through the mouth and have respiratory difficulties due to insufficient pain control which poses an obvious and significant problem when we use this device clinically.
This is also in contrast to the capnometry-type face mask that we investigated previously,11 because the PETCO2 could be monitored in all subjects, regardless of whether the mouth was open or closed. Moreover, we also found that the breathing frequencies measured directly and by the nasal cannula correlated well (r = 0.66) when the subjects were breathing with their mouths closed. Moreover, the Bland-Altman method showed that the difference (ie, bias) ± SD (limits of agreement) of the breathing frequency between 2 different measurement was 0.4 ± 3.6 (−6.7 to 7.4). The reason for the discrepancy in breathing frequency as measured by capnometry and measured directly may be that expiratory flow was intermittently discontinued, resulting in a double count upon capnography. Furthermore, the breathing frequency measured by capnometry is not counted for 1 min, but rather calculated by the length of expiration, breath by breath. This method of calculation may hence be affected by differences in the breathing frequency between capnography and direct measurement. In the present study, the breathing frequency was measured once for each measurement at arbitrary time points. Usage of the mean breathing frequency by repeated measurements for 1 min might improve the correlation of breathing frequency between the 2 methods. Although the value of breathing frequency with the capnogram is unreliable, the oxygen nasal cannula-based capnometry device may help detect hypoventilation under continuous measurement.
In addition, we also compared the PETCO2 and PaCO2 under various breathing conditions and found that the differences (biases) were 14.8, 10.2, and 7.7 mm Hg during resting breathing, mouth closed breathing, and deep breathing, respectively. These values appear somewhat greater than the reported differences between PETCO2 and PaCO2 in intubated subjects, which average 5 mm Hg but can be as high as 15 mm Hg.16,17 However, they compare favorably with those measured with the previously investigated oxygen mask-based capnometry,11 for which we reported that the differences were 12.6 and 9.1 mm Hg during resting breathing (regardless of the mouth condition) and deep breathing, respectively.
Interestingly, the difference between the PaCO2 and PETCO2 gradually decreased from resting breathing to mouth closed breathing to deep breathing. The difference depends mainly on 3 factors: the dead space (defined as the tidal volume component that does not participate in gas exchange), ventilation/perfusion matching, and the amount of exhaled breath available for sampling. The reason for the greater difference during resting breathing than during the mouth closed condition may be explained by the amount of exhaled breath available for sampling. It is possible that variable fractions of the tidal volume were exhaled through the mouth during the resting condition, leaving an insufficient amount for sampling through the nasal cannula CO2-monitoring port. In this case, the sampled gas may be contaminated by the supplied oxygen and ambient air, and the PETCO2 would be lower. The reasons for the smaller difference between the PETCO2 and PaCO2 during deep breathing when compared with mouth closed breathing may include all 3 of these factors. Even when the mouth is closed and the entire tidal volume is exhaled through the nose, some breaths might be too small to prevent contamination by the ambient air during resting breathing. However, this was less likely during deep breathing, because deep breathing increase the inspiratory and expiratory flow rate, which decrease contamination by the ambient air. More importantly, the difference between PETCO2 and PaCO2 is heavily dependent on the ratio of the dead space to the tidal volume. Obviously, the tidal volume during deep breathing is greater than that during resting breathing, whereas the dead space remains relatively unchanged. Therefore, it is not surprising that the difference between the PETCO2 and PaCO2 was the smallest during deep breathing.
Our results suggest that it may be difficult to predict the PaCO2 from the PETCO2 due to the wide limits of agreement between the PaCO2 and PETCO2, irrespective of the breathing pattern (32 mm Hg [resting] vs 23 mm Hg [with mouth closed] vs 22 mm Hg [deep breathing]). This is despite the fact that we studied a relatively homogeneous population of subjects (ie, open abdominal surgery, presence of epidural anesthesia, exclusion of obesity). The lung mechanics change following abdominal and thoracic surgery.18 After upper abdominal surgery, the vital capacity is typically reduced to ∼40% of its baseline value and remains depressed for at least 10–14 d. The functional residual capacity is also decreased to ∼70% for several days, before gradually improving to a normal level by 1 week postoperatively.18 These changes in respiratory function affect the dead space and are the likely cause of the wide range of differences between PETCO2 and PaCO2. Another potential reason for the wide range of differences between PETCO2 and PaCO2 is the effect of contamination of the oxygen flow and dilution of the CO2 value. Paul and colleagues et al19 reported a significant inverse linear relationship between PETCO2 and the oxygen flow. Even at a low oxygen flow, supplied at 4 L/min, contamination between the expired respiratory gases and oxygen flow might affect the PETCO2 value. Further research will be required in other critical care populations, with various pain, sedative, and supplied oxygen levels, to compare our results with those obtained using other capnometry devices.
Finally, in the present study, it should be noted that we mainly discussed the accuracy of the PETCO2 value, because the use of absolute values of breathing frequency and PETCO2 may be difficult for detecting hypoventilation in the post-anesthesia care unit because of poor accuracy. Moreover, these evaluations of breathing frequency and PETCO2 with various breathing patterns require bedside staff for measurements. Therefore, this device may not be useful for routine respiratory monitoring in the postoperative care unit. However, breathing patterns, such as deep breathing with the mouth closed, improved detection of the absolute value of PETCO2.
This study has several limitations. First, the sample size was small, with only 20 subjects, and none of the enrolled subjects had a respiratory comorbidity or anthropometric condition, such as obesity. Second, the subject population was limited to those undergoing abdominal surgery, and the monitoring device was a nasal cannula with sidestream; therefore, the results may not be applicable to other surgical populations and other monitoring devices, such as the nasal/oral sampling technique. Third, the data collection was not blinded. Potentially, the physician in charge may have predicted that the PETCO2 would increase following deep breathing, which may have biased the results.
Conclusions
Closing the mouth and deep breathing may help decrease the difference between PETCO2 and PaCO2, thereby enabling more accurate prediction of the absolute PaCO2 value. Although the capnometry-type oxygen cannula device may not be useful for continuous respiratory monitoring due to unstable measurement, PETCO2 measurements under deep breathing with the mouth closed with a capnometry-type oxygen cannula improve the prediction of the absolute value of PaCO2 in extubated post-abdominal surgical subjects without respiratory dysfunction.
Acknowledgments
We thank Editage for English language editing.
Footnotes
- Correspondence: Shunsuke Takaki MD PhD, 3-9 Fukuura Kanazawaku Yokohama 236-0004, Kanagawa, Japan. E-mail: shunty5323{at}gmail.com.
Drs Takaki and Fukuchi presented a version of this paper at the annual congress of Euroanaesthesia 2015, held May 30-June 2, 2015, in Berlin, Germany.
The authors have disclosed no conflicts of interest.
- Copyright © 2017 by Daedalus Enterprises