Measurement of the partial pressure of carbon dioxide in arterial blood (PaCO2) remains the criterion standard for evaluating the adequacy of ventilation. In certain populations, such as trauma patients with head injury, providing appropriate ventilation (avoiding either overventilation or underventilation) has been shown to save lives (1–2). Obtaining blood gas samples, however, is invasive and requires special equipment, which is often impractical in prehospital settings. Thus, a noninvasive means of estimating PaCO2 would be useful (3). Several such methods have been developed.
End-tidal carbon dioxide (EtCO2) monitoring is performed with increasing frequency in injured patients (4,5). However, this method may become inaccurate in unstable trauma patients (6). The PaCO2-EtCO2 gradient is related to the physiologic dead space as described by Enghoff modification of the Bohr equation (7). PaCO2-EtCO2 gradient increases with increasing physiologic dead space, due to lung injuries or decreased pulmonary perfusion.
Another noninvasive technique that can be used as a surrogate for PaCO2 is the measurement of the transcutaneous partial pressure of CO2 (tPCO2). This technique was pioneered by Severinghaus (8) in the 1960s using a temperature-stabilized heated electrode. Transcutaneous gas tension is a function of dermal capillary blood and in turn of arterial blood flow. Previous studies by Nishiyama et al. (9, 10) demonstrated good correlations between transcutaneous CO2 measurements and PaCO2 in adults undergoing general anesthesia. Several studies, performed in critically ill adults (11–14) as well as in adults undergoing noninvasive ventilation (15, 16), reported good correlation between tPCO2 measurements and PaCO2. Most notably, a study by Hinkelbein et al. (13) demonstrated the feasibility of tPCO2 monitoring in critically ill adults during interhospital transport.
Nevertheless, there is a paucity of studies comparing EtCO2 and tPCO2 monitoring in settings of evolving acute lung injury (ALI). Therefore, we carried out an animal study to evaluate the usability and accuracy of these methods in a clinically relevant porcine model of ALI secondary to smoke inhalation and burns.
MATERIALS AND METHODS
This study was approved by the US Army Institute of Surgical Research Animal Care and Use Committee. It was conducted in compliance with the Animal Welfare Act and the implementing Animal Welfare Regulations and in accordance with the principles of the Guide for the Care and Use of Laboratory Animals.
For this study, we selected a subgroup of eight female, nonpregnant, Yorkshire pigs from an ongoing study that investigates treatment of acute respiratory distress syndrome (ARDS) due to smoke inhalation and burns. Total intravenous anesthesia with ketamine HCL (20–30 mg/kg per hour), midazolam HCL (1.0–1.5 mg/kg per hour), and propofol (100 µg/kg per hour) was used during the experiment. All animals underwent tracheostomy and placement of central and arterial lines. After instrumentation, animals were allowed to recover for 2 to 3 h. After stabilization, they received inhalation injury with 22 breaths of pine-bark smoke at room temperature (Table 1). This model of smoke inhalation injury has been described previously (17). Smoke injury was followed immediately by a 40% total-body-surface-area, full-thickness flame burn. Upon completion of smoke inhalation and burn, animals were transferred to the animal intensive care unit where they were continuously monitored for the duration of the experiment. Total intravenous anesthesia levels were adjusted to effect as needed to attain no response to painful stimuli. In addition, buprenorphine HCl (0.1 mg/kg) was administered intramuscularly every 6 h for analgesia for the duration of the study. Resuscitation with lactated Ringer’s solution was performed by means of a computerized burn resuscitation decision support system for the first 24 h after burn. Subsequently, the fluid rate was adjusted manually to achieve a urinary output of at least 0.5 mL/kg per hour. Controlled mechanical ventilation (CMV) was initiated, with a tidal volume of 10 mL/kg and a respiratory rate adjusted to maintain PaCO2 between 35 and 45 mmHg. Before onset of ARDS, positive end-expiratory pressure was set to 5 cm H2O and remained constant. Fraction of inspired oxygen (FIO2) was titrated to achieve peripheral oxygen saturation (SpO2) at or above 90%. Fiberoptic bronchoscopy for pulmonary toilet was performed at 1, 2, 6, 12, and 24 h after injury, every 24 h thereafter, and anytime airway obstruction was clinically suspected. During suctioning and bronchoscopies, FIO2 was increased to 100%. After the onset of ARDS, animals were transitioned to low-tidal-volume CMV, and further ventilator changes, including adjustments of positive end-expiratory pressure and FIO2, were made in accordance with the ARDSnet protocol (18). The study was designed to continue for up to 7 days after injury. Experiments were terminated earlier if ARDS did not develop in 72 h after injury or if an animal reached terminal cardiopulmonary failure before 7 days. Animals were killed with intravenous injection of 20 mL Fatal Plus (Vortech Pharmaceuticals, Dearborn, Mich).
Ratio of the partial pressure of oxygen in arterial blood (PaO2) to the fraction of inspired oxygen (PFR) was measured at baseline, 2 h after injury, and every 6 h thereafter. A single set of baseline measurements was recorded after surgery but before the induction of smoke inhalation and burns. Baseline duration generally was 2 to 3 h and was adequate to provide stabilization time for all vital signs after surgical preparation. We defined ALI as a sustained PFR less than 300 mmHg, and ARDS as a sustained PFR less than 200 mmHg. As an index of pulmonary instability, we recorded the frequency of changes in the respiratory rate (as set on the ventilator) and in the FIO2. We defined hemodynamic instability as mean arterial pressure (MAP) of 60 mmHg or less (or a requirement for norepinephrine infusion to maintain MAP >60 mmHg).
A SenTec Digital Monitor and V-Sign digital sensor (SenTec Ag, Therwil, Switzerland), described previously (12, 15, 19), were used for tPCO2 monitoring. One-point tPCO2 sensor calibration was performed in vitro according to the manufacturer’s recommendations utilizing an 8% ± 0.05% CO2 standard gas before placement. Calibration was repeated every 6 to 8 h, as indicated by the device. The V-Sign sensor was attached to the base of the pig’s right auricle utilizing a single-use attachment ring. This area was chosen because the skin is not thick and because it is well perfused by a branch of external carotid artery. Before sensor placement, the attachment site was cleansed with isopropyl alcohol, and one drop of SenTec viscous contact gel was applied to the center of the sensor membrane. The V-Sign sensor was heated to 43.5°C. Use of a heating element within the electrode elevates local skin temperature and causes hyperemia. Application of heat to the skin surface, promotes “liquefaction” or changes in the matrix of the dermal lipid component in the stratum corneum layer, which increases gas diffusion through the skin (20, 21). It has been reported previously (22) that such changes begin to occur at 41°C and reach optimal levels for maximal gas diffusion at 44°C. Therefore, local hyperemia improves correlation between transcutaneous and arterial CO2 values. Previous publications (23, 24) indicated better PaCO2 correlation with increased electrode temperature; therefore, we used maximal temperature (43.5°C) allowed by the device. The SenTec monitor displays tPCO2 values in mmHg as measured at 37°C without applying a temperature correction.
For EtCO2 monitoring, a FilterLine H Set CO2 sidestream sampling line and airway adapter were used, along with a Capnostream 20 system (Oridion Medical, Jerusalem, Israel). The FilterLine was changed daily or sooner if it became contaminated with secretions.
An i-STAT blood gas analysis system (Abbot Point of Care, Princeton, NJ) and EG7+ cartridges were utilized for obtaining PaCO2 values. Paired measurements of EtCO2 and PaCO2, and of tPCO2 and PaCO2, were recorded every 6 h (as well as intermittently when clinically indicated) to assess the adequacy of ventilation. All measurements were done while animals were mechanically ventilated in the CMV mode.
SigmaPlot for Windows version 12.0 (Systat Software, San Jose, Calif) was used. Overlapping sets of data were analyzed based on (i) respiratory status (before vs. after the onset of ALI) and on (ii) hemodynamic status (stable vs. unstable). All results are expressed as mean ± SD. P < 0.05 was considered significant.
Linear regression was performed for pairs of PaCO2 and tPCO2 as well as for pairs of PaCO2 and EtCO2 data. Bias and 95% limits of agreement were calculated using Bland-Altman analysis (25). Student t tests and analyses of variance with post hoc Holm-Sidak multiple comparison tests were used.
A total of 454 h of animal intensive care unit care were provided during this study. We recorded 276 paired measurements of PaCO2 and EtCO2, and 250 paired measurements of PaCO2 and tPCO2, from eight animals. The average duration of monitoring was 56.8 ± 40.9 h. EtCO2 and tPCO2 measurements were made across a PaCO2 range of 25 to 85 mmHg. Animal data are presented in Table 1.
It is important to highlight that our injury model was fluid rather than static with rapidly declining respiratory status. The period between smoke/burn injury and onset of ALI was marked by pulmonary instability while animals progressed from healthy baseline conditions to ALI over 22.5 ± 10.6 h. The mean number of ventilator changes per hour was greater before onset of ALI (3.07 ± 1.53) than during ALI (1.84 ± 1.68, P = 0.02). Similarly, the average decrease in PFR in the first 24 h after injury was 199; in the second 24-h period, it was 100 (Fig. 1). There was a statistically significant decrease in PFR between baseline and 24 h, as well as between 12 and 24 h (P < 0.05), but not between 24 and 48 h (P > 0.05). During the study, two animals required norepinephrine infusion due to hypotension. During the study, we observed no thermal complications due to the placement of a heated tPCO2 sensor on the auricle.
Two partially overlapping sets of data were analyzed based on (i) pulmonary status (before vs. after the onset of ALI) and (ii) cardiovascular status (hemodynamically stable vs. unstable). The results of this analysis, describing the EtCO2-PaCO2 and tPCO2-PaCO2 relationships, are presented in Table 2 and are described in the following sections.
Before versus after the onset of ALI
Linear regression analysis of 105 EtCO2-PaCO2 pairs, recorded before the onset of ALI (Fig. 2A), demonstrated moderate correlation: PaCO2 = 13.45 + 0.68 * EtCO2 (R2 = 0.45, P < 0.0001). There was systemic underreading by capnography (P = 0.02). Bland-Altman analysis revealed a mean bias of 0.91 ± 3.77 mmHg (Fig. 2B).
Analysis of 171 EtCO2-PaCO2 pairs, recorded after ALI (Fig. 3A), demonstrated low correlation: PaCO2 = 41.51 + 0.34 * EtCO2 (R2 = 0.12, P < 0.0001). There was systemic underreading by capnography (P < 0.001). Mean bias increased to 14.84 ± 11.76 mmHg (Fig. 3B).
Analysis of 88 tPCO2-PaCO2 pairs, recorded before the onset of ALI (Fig. 2A), revealed moderate correlation: PaCO2 = 17.01 + 0.56 * tPCO2 (R2 = 0.51, P < 0.0001). No systemic difference was observed (P = 0.85). Mean bias was 0.09 ± 4.56 mmHg (Fig. 2C).
Analysis of 140 tPCO2-PaCO2 pairs, recorded after the onset of ALI (Fig. 3A), revealed strong correlation: PaCO2 = 7.74 + 0.83 * tPCO2 (R2 = 0.78, P < 0.0001). Slight systemic underreading was present (P = 0.026). Mean bias was 0.03 ± 5.44 mmHg (Fig. 3C).
Hemodynamically stable versus unstable
We repeated the analysis based on hemodynamic status. Linear regression of 233 EtCO2-PaCO2 pairs under hemodynamically stable conditions (Fig. 4A) demonstrated moderate correlation: PaCO2 = 5.32 + 1.03 * EtCO2 (R2 = 0.44, P < 0.0001). There was systemic underreading by capnography. The mean bias was 6.74 ± 8.26 mmHg (Fig. 4B).
Analysis of 41 EtCO2-PaCO2 pairs under hemodynamically unstable conditions (Fig. 5A) did not demonstrate a linear relationship: PaCO2 = 61.4 − 0.16 * EtCO2 (R2 = 0.03, P = 0.29). Systemic underreading was present (P < 0.001). Mean bias was 25.59 ± 15.30 mmHg (Fig. 5B).
Analysis of 212 tPCO2-PaCO2 pairs under hemodynamically stable conditions (Fig. 4A) demonstrated strong correlation: PaCO2 = 7.86 + 0.82 * tPCO2 (R2 = 0.80, P < 0.0001). There was no systemic difference between the two techniques (P = 0.29). Mean bias was −0.35 ± 4.75 mmHg (Fig. 4C).
Finally, analysis of 38 tPCO2-PaCO2 pairs under hemodynamically unstable conditions (Fig. 5A) demonstrated poor correlation: PaCO2 = 33.60 + 0.3 * tPCO2 (R2 = 0.39, P < 0.0001). There was systemic overreading (P < 0.001). Mean bias was −11.32 ± 14.87 mmHg (Fig. 5C).
In this study, we evaluated the utility of two commonly used noninvasive carbon dioxide monitoring technologies in a severely injured, mechanically ventilated porcine model of ALI due to smoke inhalation and burns. Our principal findings were (i) the period of time between smoke/burn injury and the onset of ALI was characterized by pulmonary instability, manifested by a steady decrease in the PFR and by frequent ventilator changes and other interventions; (ii) both EtCO2 and tPCO2 were moderately correlated with PaCO2 during this period; (iii) after the onset of ALI, EtCO2 became relatively inaccurate, whereas tPCO2 did not; (iv) hemodynamic instability caused EtCO2 values to lose their linear relationship with PaCO2. Based on these observations, we conclude that both methods can be useful in the monitoring of patients with developing severe lung injury, but that caution should be used in the interpretation of results when patients are changing rapidly.
Good correlation between EtCO2 and PaCO2 across all ranges of dead space was reported by McSwain et al. (7). Our previous study (26) demonstrated close correlation between EtCO2 and PaCO2 in a porcine model of severe chest injury during periods of hemodynamic stability and during settings of hypoventilation and hyperventilation induced by varying tidal volumes and different respiratory rates in healthy swine. At the same time, results were not as promising when accuracy of capnography was evaluated in trauma patients (6). In addition, several studies compared accuracy of end-tidal and tPCO2 monitoring in human patients (13, 27, 28), favoring tPCO2. In our study, EtCO2 systemically underestimated PaCO2. Presence of systemic bias was confirmed by paired Student t test results (Table 2). This is to be expected, particularly in patients with developing ALI, in whom dead space progressively increases, thereby leading to an increase in PaCO2-EtCO2 gradient.
There was moderate correlation between tPCO2 and PaCO2 (R2 = 0.51) before ALI. This correlation became more linear (R2 = 0.78) after onset of ALI. We expected better correlation during pre-ALI stage, and such findings were surprising. We speculate that lower tPCO2-PaCO2 correlation before onset of ALI was possibly due to more frequent ventilator changes: approximately 3.07 ± 1.53 changes per hour in the initial phase of the experiment, compared with 1.84 ± 1.68 changes per hour after the onset of ALI. Also, more frequent suctioning was required in the first 24 h after injury due to copious secretions. Because it can cause alveolar collapse and loss of recruitment, repeated suctioning may have had a destabilizing effect on PaCO2 correlation. In addition, the tPCO2 sensor was more frequently disconnected and removed from the animals during the first 24 h of the experiment to accommodate bronchoscopies.
When the tPCO2-PaCO2 relationship was examined based on hemodynamic status, it was strongly linear (R2 = 0.80) with minimal bias (−0.35 ± 4.75 mmHg) as long as MAP remained greater than 60 mmHg. Our results confirm those reported previously in human studies (11, 13, 20, 27, 28) that demonstrated similar correlation. tPCO2 had a less linear correlation to PaCO2 during hemodynamic instability (R2 = 0.39). In addition, transcutaneous monitoring tended to overestimate PaCO2 under these conditions. Previous reporting (29, 30) indicated that vasoactive medications did not have a significant effect on tPCO2 monitoring accuracy. However, our results were different, possibly because of the sensor location on the right auricle. Although the animals were hypotensive, their ears were cold to touch, and there was a visible area of vascular congestion. Preferred locations for tPCO2 monitoring have been reported in humans (9, 10), but no such location has been defined in pigs.
Because we used a subset of animals for this work from a larger study, which pursued treatment of ARDS due to smoke inhalation and burns, the samples we obtained were convenience samples. This limited our ability to further investigate effects of suctioning, ventilator changes, and sensor disconnections on stabilizations/steady state times of the measured variables. Future studies designed specifically to address these limitations may be warranted.
Based on our data, transcutaneous capnometry is an acceptable trend-monitoring the tool in settings of lung injury in hemodynamically stable patients and may be useful in a prehospital environment. However, the current generation of tPCO2 sensors has some limitations. First, the tPCO2 sensor requires stabilization time after placement. Several articles (31, 32) recommended a 20-min stabilization time while using the SenTec Monitor. Second, to maintain accuracy, the sensor has to be regularly recalibrated in vitro. Recalibration frequency appears to depend on sensor temperature. At 43.5°C, recalibrations were required every 6 to 8 h. On average, recalibration can be completed in approximately 3 min. Afterward, the sensor is repositioned on a patient and again requires stabilization time. In our experience, about 25 to 30 min was spent on moving, cleaning, recalibrating, replacing, and waiting for the sensor to stabilize. In this study, because of previously described limitations, we could not establish the precise stabilization timing necessary for improved accuracy; further study should be considered to address this issue. Also, we should point out that a prolonged stabilization time may limit utilization of tPCO2 monitors in prehospital environments. Finally, the sensor membrane requires replacement every 42 days and between patients. Currently, several companies are developing next-generation solid-state tPCO2 sensors that may overcome these shortcomings. Given promising results obtained from the use of a transcutaneous CO2 sensor, these new developments will be a welcomed addition to critical care monitoring armamentarium as well as potentially opening new possibilities for servo controlling mechanical ventilators and extracorporeal life support devices.
In a porcine model of ALI due to smoke inhalation and burns, transcutaneous CO2 monitoring is an acceptable noninvasive surrogate for PaCO2 under hemodynamically stable conditions and can be useful as a trend monitoring tool. However, tPCO2 readings should be correlated with PaCO2 with increased frequency when a patient’s condition is dynamically changing, for example, during periods of hemodynamic instability. End-tidal CO2 monitoring offers an ease-of-use advantage over the current generation of transcutaneous CO2 monitors. However, EtCO2 readings should be correlated with PaCO2 with increased frequency during evolution of lung injury (changes in dead space).
The authors thank Oridion, Inc, for providing support for this study, under a Cooperative Research and Development Agreement with the US Army Institute of Surgical Research.
1. Davis DP, Idris AH, Sise MJ, Kennedy F, Eastman AB, Velky T, Vilke GM, Hoyt DB: Early ventilation and outcome in patients with traumatic brain injury. Crit. Care Med
34 (4): 1202–1208, 2006.
2. Neumar RW, Otto CW, Link MS, Kronick SL, Shuster M, Callaway CW, Kudenchuk PJ, Omato JP, McNally B, Silvers SM, et al.: Part 8: Adult Advanced Cardiovascular Life Support: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation
122 (18 suppl 3): S729–S767, 2010.
3. Cancio LC, Chung KK: The role of normoventilation in improving traumatic brain injury outcomes. US Army Med Dep J
Apr–Jun: 49–54, 2011.
4. Committee on Trauma, American College of Surgeons: Advanced Trauma Life Support for Doctors (Student Course Manual)
. 9th ed. Chicago, IL: American College of Surgeons, 2012.
5. Guidelines for the Field Management of Combat-Related Head Trauma.
New York, NY: The Brain Trauma Foundation; 2005. Available at: https://www.braintrauma.org/pdf/protected/btf_field_management_guidelines.pdf
. Accessed August 17, 2012.
6. Warner KJ, Cuschieri J, Garland B, Carlbom D, Baker David, Copass MK, Jurkovich GJ, Bulger EM: The utility of early end-tidal capnography in monitoring ventilation status after severe injury. J Trauma
66 (1): 26–31, 2009.
7. McSwain SD, Hamel DS, Smith PB, Gentile MA, Srinivasan S, Meliones JN, Cheifetz IM: End-tidal and arterial carbon dioxide measurements correlate across all levels of physiologic dead space. Respir Care
55 (3): 288–293, 2010.
8. Severinghaus JW. Methods of measurement of blood and gas carbon dioxide during anesthesia. Anesthesiology
21: 717–726, 1960.
9. Nishiyama T, Nakamura S, Yamashita K: Comparison of the transcutaneous oxygen and carbon dioxide tension in different electrode locations during general anesthesia. Eur J Anesthesiology
23: 1049–1055, 2006.
10. Nishiyama T, Kohono Y, Koishi K: Comparison of ear and chest probes in transcutaneous carbon dioxide pressure measurements during general anesthesia in adults. J Clin Monit Comput
25: 323–328, 2011.
11. Bendejelid K, Schutz N, Stotz M, Gerard I, Suter PM, Romand JA: Transcutaneous PCO2
monitoring in critically ill adults: clinical evaluation of a new sensor. Crit Care Med
33 (10): 2203–2206, 2005.
12. Rodriguez P, Lellouche F, Aboab J, Buisson CB, Brochard L: Transcutaneous arterial carbon dioxide pressure monitoring in critically ill adult patients. Intensive Care Med
134: 309–312, 2006.
13. Hinkelbein J, Floss F, Denz C, Krieter H: Accuracy and precision of three different methods to determine PCO2
) during interhospital ground transport of critically ill and ventilated adults. J Trauma
65 (1): 10–18, 2008.
14. Nicolini A, Ferrari M: Evaluation of transcutaneous carbon dioxide monitor in patients with acute respiratory failure. Ann Thorac Med
6 (4): 217–220, 2011.
15. Storre J, Steurer B, Hans-Joachim K, Dreher M, Windisch W: Transcutaneous PCO2
monitoring during initiation of noninvasive ventilation. Chest
132 (6): 1810–1816, 2007.
16. Senn O, Clarenbach F, Kaplan V, Maggiorini M, Bloch KE: Monitoring carbon dioxide tension and arterial oxygen saturation by a single earlobe sensor in patients with critical illness or sleep apnea. Chest
128 (3): 1291–1296, 2005.
17. Batchinsky A, Burkett S, Zanders T, Chung KK, Regn DD, Jordan BS, Necsoiu C,, Nguyen R, Hanson MA, Morris MJ,, et al.: Comparison of airway pressure release ventilation to conventional mechanical ventilation in the early management of smoke inhalation injury
in swine. Crit Care Med
39 (10): 2314–2321, 2011.
18. The Acute Respiratory Distress Syndrome Network: Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med
342: 1301–1308, 2000.
19. Kocher S, Rohlin R, Tschupp A: Performance of a digital PCO2
ear sensor. J Clin Monit Comput
18 (2): 75–79, 2004.
20. Eberhard P: The design, use, and results of transcutaneous carbon dioxide analysis: current and future directions. Anesth Analg
105 (6): s48–s52, 2007.
21. Franklin ML: Transcutaneous measurement of partial pressure of oxygen and carbon dioxide. Respir Care Clin N Am
1 (1): 119–131, 1995.
22. Lubbers DW: Theoretical basis of the transcutaneous blood gas measurements. Crit Care Med
9 (10): 721–733, 1981.
23. Nishiyama T, Nakamura S, Yamashita K: Effects of electrode temperature of the new monitor, TCM4, on the measurement of transcutaneous oxygen and carbon dioxide tension. J Anesth
20 (4): 331–334, 2006.
24. Sorensen L, Brage-Andersen L, Greisen G: Effects of the transcutaneous electrode temperature on the accuracy of transcutaneous carbon dioxide tension. Scand J Clin Lab Invest
71 (7): 548–552, 2011.
25. Bland JM, Altman DJ: Statistical methods for assessing agreements between two methods of clinical measurements. Lancet
1 (8476): 307–310, 1986.
26. Isbell CL, Batchinsky AI, Hetz KM, Baker WL, Cancio LC: Correlation between capnography and arterial carbon dioxide before, during and after severe chest injury in swine. Shock
37 (1): 103–109, 2012.
27. Hirabayashi M, Fujiwara C, Ohtani N, Kagawa S, Kamide M: Transcutaneous PCO2
monitors are more accurate than end-tidal PCO2
monitors. J Anesth
23 (2): 198–202, 2009.
28. Phan CQ, Tremper KK, Lee SE, Barker SJ: Noninvasive monitoring of carbon dioxide: a comparison of the partial pressure of transcutaneous and end-tidal carbon dioxide with the partial pressure of arterial carbon dioxide. J Clin Monit
3 (3): 149–154, 1987.
29. Palmisano B, Severinghaus J: Transcutaneous PCO2
: a multicenter study of accuracy. J Clin Monit
6 (3): 189–195, 1990.
30. Janssens JP, Howarth-Frey C, Chevrolet JC: Transcutaneous PCO2
to monitor noninvasive mechanical ventilation in adults: assessment of a new transcutaneous PCO2
. 113 (3): 768–773, 1998.
31. Domingo C, Canturri E, Lujan M, Moreno O, Espuelas H, Marin A: Transcutaneous measurement of partial pressure of carbon dioxide and oxygen saturation: validation of the SenTec Monitor. Arch Bronconeumol
42 (5): 246–251, 2006.
32. Janssens JP, Perrin E, Bennani I, DeMuralt B, Titelion V, Picaud C: Is continuous transcutaneous monitoring of PCO2
) over 8 h reliable in adults? Respir Med
95 (5): 331–335, 2001.