eNeonatal Review eNeonatal Review
January 2008: VOLUME 5, NUMBER 5

Editor's Note:

Over 800 Responses to our Survey

Thank you for answering the call…we received 818 responses! We appreciate your enthusiasm about eNeonatal Review and pledge to always maintain our high standards. We have chosen 3 remarkable comments and will announce the winners with next month’s issue.

The State of Non-Invasive CO2 Monitoring Techniques

In this Issue...

Blood gas status is used to determine the need for initiation, adjustment, and discontinuation of mechanical ventilatory support in the neonate. Carbon dioxide monitoring by transcutaneous (TcPCO2 ) and end tidal (PetCO2 ) measurements have been proposed as noninvasive surrogates to analysis of CO2 from arterial blood sampling. These monitoring techniques have not substituted for, but instead have been used as an adjunct to, standard monitoring of arterial (PaCO2 ) and capillary (PcapCO2 ) blood gases.

In this issue, we review recent reports of how these CO2 monitoring techniques have been used in preterm and term infants for a variety of conditions and applications.
COMMENTARY from our Guest Author
Course Directors

Edward E. Lawson, MD
Department of Pediatrics
Division of Neonatology
The Johns Hopkins University
School of Medicine

Christoph U. Lehmann, MD
Associate Professor
Department of Pediatrics
Division of Neonatology
The Johns Hopkins University
School of Medicine

Lawrence M. Nogee, MD
Associate Professor
Department of Pediatrics
Division of Neonatology
The Johns Hopkins University
School of Medicine

Mary Terhaar, DNSc, RN
Assistant Professor
Undergraduate Instruction
The Johns Hopkins University
School of Nursing

Robert J. Kopotic, MSN, RRT, FAARC
President of Kair Medical Innovations
San Diego, CA
Nelson Claure, MSc, PhD Commentary & Reviews:
Nelson Claure, MSc, PhD
Research Assistant
Professor of Pediatrics
Leonard M. Miller
School of Medicine at the
University of Miami
Director of the Neonatal Pulmonary Laboratory
Jackson Memorial
Medical Center
Miami, Florida
Guest Faculty Disclosure

Nelson Claure, MSc, PhD has disclosed no relevant financial relationships.

Unlabeled/Unapproved Uses

The author has indicated that there will be no reference to unlabeled/unapproved uses of drugs or products in the presentation.

Program Directors' Disclosures
At the conclusion of this activity, participants should be able to:

Discuss the current research assessing non-invasive CO2 monitoring techniques in preterm and term infants, under different conditions and for different purposes
Describe the reported advantages and limitations of non-invasive CO2 monitoring
Identify how the information presented can be used to optimize monitoring of preterm and term infants with respiratory failure
Program Information
CE Info
Credit Designations
Target Audience
Learning Objectives
Internet CME/CNE Policy
Faculty Disclosure
Disclaimer Statement

Length of Activity
1.0 hours Physicians
1 contact hour Nurses

Expiration Date
January 16, 2010

Next Issue
February 14, 2008

Step 1.
Click on the appropriate link below. This will take you to the post-test.

Step 2.
If you have participated in a Johns Hopkins on-line course, login. Otherwise, please register.

Step 3.
Complete the post-test and course evaluation.

Step 4.
Print out your certificate.

Physician Post-Test

Nurse Post-Test

Respiratory Therapists
Visit this page to confirm that your state will accept the CE Credits gained through this program or click on the link below to go directly to the post-test.

Respiratory Therapist Post-Test
eNeonatal Review Podcast eNeonatal Review is proud to continue our accredited
PODCASTS for 2008. Listen here.

In this audio interview, Nelson Claure, MSc, PhD, and Robert J. Kopotic, MSN, RRT, FAARC, discuss additional topics related to The State of Non-Invasive CO2 Monitoring Techniques.

Participants can now receive 0.5 credits per podcast after completing an online
. In addition to our monthly newsletters, there will be 6 podcasts throughout the year.

To learn more about podcasting and how to access this exciting new feature of eNeonatal Review, please visit this page.
Listen to our Podcast
Careful management is required to avoid the extremes of PaCO2 associated with mechanical ventilatory support. Noninvasive TcPCO2 and PetCO2 monitors have been proposed as means to reduce blood sampling while providing near continuous assessment. Currently, TcPCO2 is commonly used as an adjunct PaCO2 or PcapCO2, while PetCO2 monitoring is used less frequently.

Performing a TcPCO2 measurement consists of creating a skin-electrode unit of increased local perfusion by controlled hyperemia. The neonate’s thin epidermal layer has a small metabolic CO2 production, and diffusing CO2 molecules change the pH of an electrolyte solution. However, poor perfusion can result in overestimation due to reduced removal of the CO2 produced locally. Many TcPCO2 devices include a metabolism correction factor, and TcPCO2 is often post-calibrated to PaCO2 or PcapCO2. Further, in preterm infants, heating can produce injury, which limits application time at a single site.

Among the studies reviewed herein, Aliwalas et al reported acceptable TcPCO2 bias in infants of ≤ 28 weeks (w) of gestational age (GA), but warned of between-patient variability; Tingay et al reported a small overestimation bias in TcPCO2 that did not change at higher PaCO2 during transport; and Bernet-Buettiker et al reported that bias in estimation of PaCO2 with an ear TcPCO2 electrode was small but was accompanied by variability between patients, and that TcPCO2 was tightly related to PcapCO2. Although PcapCO2 may not always reflect PaCO2, this correlation is important because in preterm infants, invasive arterial lines are not commonly in place beyond the critical period of respiratory failure.

PetCO2 monitoring is performed by placing infrared sensors mainstream, or by sidestream gas sampling. Both the Aliwalas and Wu studies showed acceptable bias in PetCO2 with respect to PaCO2 in preterm infants after birth, and in older preterm and term infants using low-flow sidestream and mainstream measurements, respectively. Both studies showed important between-patient variability. However, they differed in their recommendations. Further, Tingay found PetCO2 underestimated PaCO2 during ground transport.

PetCO2 is more responsive to changes in PaCO2 than TcPCO2. However, preterm infants have a relatively large anatomical dead space and PetCO2 depends on tidal volume (VT) size.1 PetCO2 is also influenced by the arterial-alveolar gradient in infants with lung disease. Use of this gradient as a longitudinal correlate to advancing lung disease should be further examined.

Capnography may be informative in the preterm infant but can present some shortcomings. Claure et al described capnographic patterns associated with rebreathing, while others suggested patterns associated with lung disease.2 New mainstream PetCO2 sensors have small dead space volumes (<1 ml), and — although this has not been documented in small preterm infants — have been shown to induce rebreathing in preterm infants in a manner similar to flow sensors of comparable size. In larger preterm infants, a mainstream PetCO2 sensor of 2 ml dead space increased TcPCO2 due to rebreathing.3 High sidestream sampling flows can provide inaccurate data by diluting the end expiratory gas. Low-flow sidestream PetCO2 monitors have only recently become available.

In comparison to the older literature, these newer reports show improved bias with these monitoring techniques. However, there is poor precision, as indicated by wide between-patient variability. Under routine clinical conditions, the lack of certainty when a CO2 reading is outside the desirable range is an important limitation. In most cases, this finding requires repeat preparation and application procedures, and often requires additional blood sampling.

The availability of monitoring data over time is important, but the ability to trend such data has not been fully exploited. Preterm infants undergo different phases of respiratory disease, and an association has been found between extremes and fluctuations in PaCO2 with IVH.4 Detection of ensuing changes that lead to extreme hypo- or hypercarbia may be important triggers for intervention. In many centers, TcPCO2 monitoring is started ahead of a blood sample to assure stability, and monitoring often continues if the ventilatory support is changed. Although TcPCO2 or PetCO2 readings alone should not be used to adjust the support, they may give an early warning of impaired ventilation.

The usefulness of CO2 monitoring may be enhanced when combined with ventilation data. As Claure et al reported (reviewed herein), in preterm infants, increases in PetCO2 and TcPCO2 correlated with rebreathing and occurred in parallel to a rise in total ventilation due to increased spontaneous breathing. In prior studies, during volume-targeted ventilation, a lower target VT led to an increase in TcPCO2 and spontaneous ventilation while total ventilation remained unchanged.5 Further, when preterm infants were switched from nasal CPAP to noninvasive pressure support, there was a greater increase in ventilation than among preterm infants that started with a higher baseline TcPCO2.6

Detection of exhaled CO2 has been used as an adjunct to standard methods to assess proper intubation. The report by Repetto et al suggests a means to more quickly detect improper intubation. However, PetCO2 alone may not be recommended, since conditions of reduced pulmonary circulation could lead to poor CO2 detection in spite of a correct intubation. A recent (2007) report recommends caution due to the risk of false-positive color change in colorimetric CO2 detectors in the presence of epinephrine, atropine, calfactant, and naloxone administered through the endotracheal tube.7

The recent peer-reviewed reports summarized herein, evaluating TcPCO2 and PetCO2 usefulness in estimating PaCO2, indicate that TcPCO2, in spite of its limitations, can be used as an acceptable estimate of CO2 measured with arterial and capillary blood gases. These newer data also suggest improved accuracy of PetCO2. However, most of these reports, similar to prior ones, do not justify substitution of PaCO2 analysis with these non-invasive methods. Nonetheless, the use of continuous CO2 monitoring as an adjunct to standard care has been shown beneficial in newborns in respiratory failure and, perhaps more particularly, in labile preterm infants by providing CO2 trending and variability data.


1. Greer KJ, Bowen WA, Krauss AN. End-tidal CO2 as a function of tidal volume in mechanically ventilated infants. Am J Perinatol. 2003;20(8):447-451.
2. Hagerty JJ, Kleinman ME, Zurakowski D, Lyons AC, Krauss B. Accuracy of a new low-flow sidestream capnography technology in newborns: a pilot study. J Perinatol. 2002;22(3):219-225.
3. McEvedy BA, McLeod ME, Kirpalani H, Volgyesi GA, Lerman J. End-tidal carbon dioxide measurements in critically ill neonates: a comparison of side-stream and mainstream capnometers. Can J Anesth. 1990(3);37:322-326.
4. Fabres J, Carlo WA, Phillips V, Howard G, Ambalavanan N. Both extremes of arterial carbon dioxide pressure and the magnitude of fluctuations in arterial carbon dioxide pressure are associated with severe intraventricular hemorrhage in preterm infants. Pediatrics. 2007; 119(2):299-305.
5. Herrera CM, Gerhardt T, Everett R, Musante G, Thomas C, Bancalari E. Effects of volume-guaranteed synchronized intermittent mandatory ventilation in preterm infants recovering from respiratory failure. Pediatrics. 2002:110(3):529-533.
6. Ali N, Claure N, Alegria X, D’Ugard C, Organero R, Bancalari E. Effects of non-invasive pressure support ventilation (NI-PSV) on ventilation and respiratory effort in very low birth weight infants. Pediatr Pulmonol. 2007; 42(8):704-710.
7. Hughes SM, Blake BL, Woods SL, Lehmann CU. False-positive results on colorimetric carbon dioxide analysis in neonatal resuscitation: potential for serious patient harm. J Perinatol. 2007;27(12):800-1.
Recommend to a Colleague

Newsletter Archive

Listen to our Podcast
Aliwalas LL, Noble L, Nesbitt K, Fallah S, Shah V, Shah PS. Agreement of carbon dioxide levels measured by arterial transcutaneous and end tidal methods in preterm infants ≤ 28 weeks gestation. J Perinatol. 2005; 25(1):26-29.

(For non-journal subscribers, an additional fee may apply for full text articles.)
View Journal Abstract View Full Article
Aliwalas et al sought to assess the agreement between measurements of TcPCO2 and sidestream PetCO2 with PaCO2 at 3 time points during the first 24 hours after birth (at 4, 12 and 24 h) in infants born at ≤ 28 w GA. Data were obtained from 27 ventilated infants with RDS who had a mean GA of 26.3 ± 1 w and a birth weight (BW) of 875 ± 14 grams (g). TcPCO2 was measured at 44°C and a metabolic correction factor of 5 mmHg was applied.

The intraclass correlation coefficient indicated moderate agreement between TcPCO2 and PaCO2 and between PetCO2 and PaCO2 at the 3 time points during the first 24 hours. The bias observed in these two methods was relatively small (2.2, 4.4, and 2.6 mmHg for TcPCO2, and -0.3, 2.4, and 1.9 mmHg for PetCO2 at 4, 12 and 24h, respectively). However, the precision reflected a wide between-patient variation. Interestingly, both bias and precision did not change with time. The comparisons (TcPCO2 – PaCO2 and PetCO2 – PaCO2) were not influenced by BW, mean airway pressure, mean blood pressure, or application site of the transcutaneous probe. The investigators acknowledge that the lack of observed effect of these factors may be related to the sample size.

This well-conducted study presents valuable data obtained during the initial phase of respiratory failure. The investigators cautiously conclude that agreement is moderate but not sufficient to recommend substitution of the standard practice of PaCO2 sampling. Nonetheless, these results show smaller measurement bias compared to older reports. Further, although these data included 3 time points during the first 24 hours, within-patient trending data were not given. Such information could provide additional insights on the ability of these monitoring methods to follow trends. Moreover, longitudinal data beyond day 1 would have provided additional valuable information.
Recommend to a Colleague

Newsletter Archive

Listen to our Podcast
Tingay DG, Stewart MJ, Morley CJ. Monitoring of end tidal carbon dioxide and transcutaneous carbon dioxide during neonatal transport. Arch Dis Child Fetal Neonatal Ed. 2005; 90(6):F523-526.

(For non-journal subscribers, an additional fee may apply for full text articles.)
View Journal Abstract View Full Article
The aim of this study was to assess the accuracy and reliability of end tidal CO2 during neonatal transport. Sidestream PetCO2 measurements were compared to arterial and transcutaneous measurements in mechanically ventilated infants during neonatal transport. PetCO2, TcPCO2, and PaCO2 data were obtained from 21 infants with GA between 26 to 40 w, age at study 1.8 to 61.2 hours, who underwent ground transport. Recordings of TcPCO2 and PetCO2 were calibrated to simultaneous PaCO2 values.

Although PetCO2, TcPCO2 and PaCO2 were linearly related, the investigators found that PetCO2 underestimated PaCO2 by a consistent bias of 7.8 mmHg, while TcPCO2 overestimated PaCO2 by 0.97 mmHg. This finding was more evident among infants with severe lung disease. Further, the bias observed with TcPCO2 and PetCO2 did not change with PaCO2. Two-thirds of the TcPCO2 values were within 5.3 mmHg of the paired PaCO2 value, while 81% of TcPCO2 values were within 7.5 mmHg of PaCO2. In contrast, only 48% of PetCO2 values were within 7.5 mmHg of PaCO2.

This study indicates that TcPCO2 is sufficiently accurate for PaCO2 estimation during neonatal transport, while sidestream PetCO2 readings should be used with caution as they (on average) underestimated PaCO2. The investigators suggest TcPCO2 as the preferred method for CO2 monitoring. These data could be extrapolated to the NICU, where conditions may facilitate the use of these devices in situations comparable to neonatal transport. Finally, although not provided, additional information may have been gained by stratified analysis, since the study included such a wide range of GA.
Recommend to a Colleague

Newsletter Archive

Listen to our Podcast
Bernet-Buettiker V, Ugarte MJ, Frey B, Hug MI, Baenziger O, Weiss M. Evaluation of a new combined transcutaneous measurement of PCO2/pulse oximetry oxygen saturation ear sensor in newborn patients. Pediatrics. 2005; 115(1):e64-68.

(For non-journal subscribers, an additional fee may apply for full text articles.)
View Journal Abstract View Full Article
The aim of this study was to evaluate a newly developed combination TcPCO2/pulse oximetry (SpO2) sensor that obtains measurements from the neonate’s ear. This sensor works at a temperature of 42° C at the point of contact to the skin. The investigators compared TcPCO2 data to PaCO2 readings from 30 infants of a median age of 3.5 (1-28) d and at 38.3 (28.7-40.7) w post-conception.

TcPCO2 data were also compared to PcapCO2 readings in a second group of 30 infants of a median age of 9.1 (1-28) d and at 37.9 (29.9-41.0) w post conception. In the TcPCO2 - PaCO2 comparison group, the ear TcPCO2 sensor overestimated PaCO2 with a bias of 3.21 mmHg and a precision of 6.02 mmHg (2 SD of the mean difference). In this group 23/30 infants were receiving conventional or high frequency ventilation and only 3 infants had a BW <1500g. In the second group, the TcPCO2 - PcapCO2 comparisons revealed that the ear TcPCO2 sensor overestimated PcapCO2 with a bias of only 0.67 mmHg and 8.07 mmHg precision. In this group, only 2/30 infants were receiving conventional ventilation.

These results indicate that this new TcPCO2 device can be used to detect PaCO2 with an acceptable bias, an improvement in the measurement as compared to older data. However, precision data indicating limits of agreement suggest caution in using TcPCO2 alone or as a primary factor to guide an intervention. Interestingly, TcPCO2 measurements were close to those obtained by PcapCO2. This is likely due to the fact that both estimate capillary bed CO2 tension. Although capillary blood sampling may not always reflect arterial blood gases, it is the only method available for infants without invasive arterial catheters. The sensor temperature of 42° C is sufficient for TcPCO2, and is more appealing in small infants with immature skin than the 44° C, typically needed for monitoring TcPO2. However, this study did not include infants under 28 w GA; therefore, these results cannot be extrapolated to this population. Further, the application of this sensor in small preterm infants remains to be evaluated.
Recommend to a Colleague

Newsletter Archive

Listen to our Podcast
Wu CH, Chou HC, Hsieh WS, Chen WK, Huang PY, Tsao PN. Good estimation of arterial carbon dioxide by end-tidal carbon dioxide monitoring in the neonatal intensive care unit. Pediatr Pulmonol. 2003; 35(4):292-295.

(For non-journal subscribers, an additional fee may apply for full text articles.)
View Journal Abstract View Full Article
The objective of this investigation was to evaluate mainstream PetCO2 measurements for estimation of PaCO2 in term and preterm neonates. The study population included 20 term and 41 preterm ventilated infants of a median GA of 31.4 (22.8 - 42.2) w, who were studied at a median age of 13 days. Paired values of PetCO2 and PaCO2 were compared in both groups of infants.

In the term infants, PetCO2 underestimated PaCO2 with a bias of 3.5 ± 9.0 mmHg. Measurements in the preterm infant group also showed that PetCO2 underestimated PaCO2 with a bias of 3.4 ± 6.0 mmHg. Precision (2 SD of the mean difference) was 18 mmHg and 12 mmHg for term and preterm infants, respectively. The assessment of correlation between PetCO2 and PaCO2 showed an r-value of 0.78 and 0.85 for term and preterm infants respectively.

These data indicate good correlation and an acceptable bias in the estimation of PaCO2 with mainstream PetCO2 monitoring in term and preterm ventilated infants. This report describes one of the better correlations reported in the literature, and based on the findings, the investigators recommend use of PetCO2 instead of PaCO2 to reduce blood loss. The reported precision, however, indicates significant between-patient variability — which was surprisingly better in the preterm than in the term group of infants.
Recommend to a Colleague

Newsletter Archive

Listen to our Podcast
Claure N, D’Ugard C, Bancalari E. Elimination of ventilator dead space during synchronized ventilation in premature infants. J Pediatr. 2003; 143(3):315-320.

(For non-journal subscribers, an additional fee may apply for full text articles.)
View Journal Abstract View Full Article
In this study, the investigators sought to document the effects of increased instrumental dead space caused by mainstream flow sensors, and to develop a method to reduce these effects during synchronized ventilation in preterm infants. The effects were assessed by measurements of ventilation, TcPCO2, and measurements of end-inspiratory and end tidal CO2 from sidestream capnography. Ten ventilated preterm infants with a