Clinical Review

Is ETCO2 a predictor of PaCO2 in ventilated

neonates on the Neonatal Intensive Care Unit?
Suminthrra Naidu

Abstract
Background: A key focus to monitoring ventilator efficacy in critically ill neonates is to avoid the complications caused as a
result of hypocarbia and hypercarbia. The current gold standard for monitoring ventilator efficacy is by measuring Arterial
Blood Gases (ABGs). However more traditional methods of monitoring include Transcutaneous monitoring and Pulse oximetry
(SpO2). Several reports outline the limitations associated with the accuracy and applicability of these methods.
Aim: To explore the use and contribution of end-tidal carbon dioxide (ETCO2) monitoring in ventilated neonates on the
neonatal intensive care unit.
Search Methods: A comprehensive literature search was conducted by a selection of relevant search terms.
Selection criteria: The selection process was staged to ensure that credible robust evidence was selected against a set of pre-
defined inclusion and exclusion criteria. Following this stringent process seven papers were selected.
Main results: Through analysis of the data presented in the seven studies it can be deduced that extension of this method
of monitoring to the neonatal group is rendered more complicated by functional issues such as the relatively large dead
space, rapid breathing rates with low tidal volumes and ventilation-perfusion mismatch as a result of underlying lung disease.
However advances in capnometer technology have allowed for modifications to overcome such limitations. Rozycki et al
(1998), Wu et al. (2003), Bhat and Abhishek (2008) and Nangia et al (1997) found a good correlation existed between ETCO2
and PaCO2 in patients without underlying lung disease.
Conclusions: Whilst all the studies in this review showed that a correlation exists between ETCO2 and PaCO2, this correlation
was stronger in the groups with no underlying lung disease. In taking this forward it can be established that whilst ETCO2
cannot replace the gold standard, it can be used with caution as a valuable trending tool.

Key words: End-tidal carbon dioxide monitoring, capnography, ventilated neonate

Background perfusion injuries, air emboli and infec- While the more traditional approaches
tions. to monitoring ventilation are well estab-
Neonates ventilated on the neonatal in- Other modes of monitoring have evolved lished, they have their limitations, and
tensive care unit (NICU) are monitored in an attempt to provide continuous an awareness is required of the advance-
closely for efficacy of ventilation in an non-invasive measurements. These in- ments in technology to enable one to
attempt to reduce complications as- clude transcutaneous oxygen/carbon di- improve the way in which neonatal ven-
sociated with hypoxia, hypocarbia and oxide and SpO2 monitoring. Whilst these tilation is controlled. One such method
hypercarbia. This is achieved by a range methods of monitoring have become is End-tidal carbon dioxide monitoring
of monitoring methods such as arteri- standard practice in many neonatal units (ETCO2). This is a continuous non-inva-
al blood gases (ABGs), pulse oximetry they themselves are associated with lim- sive method of monitoring that is based
(SpO2) and transcutaneous (Tc) moni- itations such as those listed in the table on the principle that carbon dioxide (CO2)
toring. Whilst an ABG remains the gold below. will be detected during expiration from a
standard it only provides a snap shot
view of ventilator efficacy, and sampling
Table 1: Limitations of Transcutaneous and SpO2 Monitoring
is associated with complications such as

Transcutaneous Monitoring Pulse Oximetry

Correspondence author Thermal injury Information on Oxygenation only
Suminthrra Naidu; Advanced Long Stabilization time Signal quality
Neonatal Nurse Practitioner
Frequent calibrations required Pressure sores from probe
The Royal London Hospital e-mail
(n_simmi29@yahoo.com) Damage to skin with adhesive Motion artifacts
Slow response time Electromagnetic interferences

(Carter and Williams 2008)

Working Papers in Health Sciences 1:1 ISSN 2051-6266 / 20120005 1

 Clinical Review

correctly placed endotracheal tube (ETT).

Initially adopted for anaesthetics, its use
to confirm airway patency and lung venti-
lation has expanded over the last decade
to include critical care, emergency med-
icine, field resuscitation and conscious
sedation settings (ASA 2005).
End-tidal CO2 Monitoring
Capnometers are available as either a
side-stream or mainstream sampling de-
vice. The use of capnography in neonatal
practice is less well documented because
of functional limitations including the ad-
ditional dead space, failure to reach an
expiratory plateau during rapid respira-
tory rates and the technical limitations of
ETCO2 devices to interpret CO2 in small
tidal volume states (Hammer 2006). Re-
cent advances in technology have ex-
plored ways in which to overcome such
limitations by modifying the capnometer Figure 1: The Time Capnogram – Represents the plot of the concentration of
to consider the dynamics of the neonatal carbon dioxide as measured in the airway sample against the time axis of the
lung model with resultant development phases of respiration
of a micro-stream device (Colman and
Krauss 1999). The clinical application of
the device can be interpreted in the cap-
nogram (Fig 1) which represents the con- A search strategy was undertaken cov-
centration of CO2 in the airway through ering the period May 2009 to February 10 studies were identified. To provide a
different phases of the respiratory cycle. 2010. The initial stages of this litera- robust review of evidence, studies that
ture search involved identifying primary mapped well on the hierarchy of evi-
A typical time capnogram can be consid- sources, grey literature and expert opin- dence were included in this review. Fol-
ered in two parts, i.e. an inspiratory and ions relating to the use of ETCO2. The lowing further evaluation, 3 papers were
an expiratory phase. The normal airway second stage of this search process was excluded as they focused around the clin-
CO2 values are 6-7%, which equates to formalised by entering search terms ical applications of capnography to the
45-55mmHg. The CO2 concentration namely: “Capnography or End-tidal CO2 neonatal speciality.
reflects cardiac output and pulmonary AND ventilated neonates”. The databases 7 studies were identified as eligible for
blood flow as the gas is transported by accessed included Medline, British Nurs- inclusion as they examined the correla-
the venous system to the right side of the ing Index, Cumulative Index of Nursing tion existing between ETCO2 and PaCO2.
heart and then pumped to the lungs by (CINAHL), Pubmed, Cochrane, Journals A key strength of these seven studies
the right ventricle (West 2008). When and Books at Ovid, Health Management where that they took place in tertiary
CO2 diffuses out of the lungs into the ex- Information Consortium (HMIC), The centres.
haled air, a device called the capnometer Excerpta Medical Database (EMBASE),
measures the partial pressure or maximal Biomed Central, Web of Knowledge,
concentration of CO2 at the end of exha- Turning Resource into Knowledge (TRIP),
lation. This represents the ETCO2 (Fig 1). Register for Clinical Trials, Health Infor-
mation Resources and the Electronic Ta-
Review Process ble of Contents (ZETOC).
In undertaking a literature search poten- Selection Criteria
tial confounding factors were considered.
Aim Studies were included in this review
Several key drivers have led to the devel- if they were primary research articles
opment of an evidence based healthcare, published after 1990 studying ventilated
requiring one to make use of best evi- neonates on a NICU and focused on the
dence to inform practice (Thomson and comparison of ETCO2 measurement with
Dowding 2002). PaCO2 values.
This review examines the applicability Studies were excluded if they included
and use of ETCO2 in ventilated newborn older infants, adults or animal studies.
infants on the NICU. To ensure that the search strategy incor-
porated recent advances in technology,
Search Strategy studies prior to 1990 were excluded.

Working Papers in Health Sciences 1:1 ISSN 2051-6266 / 20120005 2

 Author
Hierarchy of
Year Aim(s) of study Sample Methodology Results
Evidence
Country
27 infants with median birth DETCO2 had a good correlation with PaCO2
Kugelman, A. Prospective quantitative study
weight : 1835 grams (where n=222, r= 0.72, p<0.001).

Geiger-Aginsky, D., Al infants intubated with a double The study further highlighted that DETCO2
Bader, D., Shoris, I., (490g-4790g) lumen ETT. Neonates had indwelling was a better predictor of PaCO2 in neonates
Riskin, A. To evaluate a nov- arterial lines. with underlying lung pathology.
el method of distal
end-tidal CO2 (DETCO2) Infants were connected simultane-
by comparison with ously to proximal ETCO2 and DETCO2
2008
Level 2b PaCO2 and with a more monitors and measurements were
standard method that compared to PaCO2.
measures mainstream
Israel Mean gestational age :
end-tidal CO2 in intubat-
ed infants. 32 weeks ( 24.8-40.8)

222 DETCO2- PaCO2 pairs

212 ETCO2-PaCO2 pairs
Good overall correlation and agreement
between ETCO2 and PaCO2 in surfactant
Singh, A. S. Retrospective chart review
treated, ventilated ELBW babies during the

Working Papers in Health Sciences 1:1 ISSN 2051-6266 / 20120005

first week of life.
The Pearsons Correlation Co-efficient was
Signal, N. 31 ELBW infants: < 1000g
0.71 with a 95% Confidence Interval.

To study the correlation The Bland Altman plot demonstrated a

good correlation and the study further
2005 and agreement between 21 male and 10 female All admissions to the NICU
ETCO2 and PaCO2 in ven- demonstrated an intraclass correlation
Level 2b tilated extremely low co-efficient of 0.81 (p<0.0001).
birth weight (ELBW)
Infants were mechanically ventilated
infants in the first week
Canada All ventilated and had indwelling arterial catheters
of life.
insitu.
All infants were connected to a main-
stream capnometer. 754 simultane-
ous ETCO2 and PaCO2 pairs collected
and analysed.
 Wu, Ch., Chou, H.C.,
This study did not show a difference in
Hseih, W.S., Chen, 61 Patients recruited. 20 Prospective non randomised single
measurements between preterm or term
W.K., Huang, P.Y. and term and 41 preterm infants. arm study.
neonates with underlying lung disease.
Tsao, P.N.
All patients were ventilated
130 ETCO2/PaCO2 pairs were ana- The researchers showed a positive correla-
2003 and had indwelling arterial
lysed from 61 patients. tion between ETCO2 and PaCO2.
catheters.

To estimate PaCO2 by Main reasons for intubation

Level 2b ETCO2 monitoring in the were either Respiratory An overall correlation showed ‘r’=0.83
Taiwan
NICU Distress Syndrome (RDS) or where p<0.001.
cardiac disease.

In the term group ‘r’ = 0.779, p< 0.001 and

in the preterm group ‘r’= 0.849, p< 0.001.

The overall ETCO2 bias (mean +/- SD) was

3.5 +/- 7.1mmHg and a 95% Confidence
Interval for the mean of 2.2-4.7.

Prospective non randomised single

Rozycki,H.J., 45 Newborn infants. Patients were stratified into two groups.
arm quantitative study.
All mechanically ventilated
411 ETCO2/PaCO2 pairs were ana-

Working Papers in Health Sciences 1:1 ISSN 2051-6266 / 20120005

Sysyn,G.D., and had indwelling arterial
lysed from 45 patients.
access.
However patients in the SUB groups and
A predefined subsample of The correlation co-efficient was 0.83, ALL groups regardless of gestational age
To determine the
Marshall,M.K., Malloy, infants with birthweight < with a 95% Confidence Interval and a or underlying lung disease demonstrated
accuracy and precision
R. and Wiswell, T.E. Level 2b 1000g, < 8 days old and who 63% accuracy in predicting hypocar- similiar correlation between ETCO2 and
of ETCO2 monitoring in
received surfactant therapy. bia and hypercarbia. PaCO2 (‘r’ = 0.833 and 0.821 respectively
NICU patients
with p<0.001).
Patients categorised into
SUB group and an ALL group The SUB group was incorporated into
1998
based on the abovemen- the ALL group.
tioned criteria.

United States
 A total of 152 samples from
arterial catheters were ana-
Nangia, S. lysed from babies with birth This study showed that ETCO2 correlates
weights from 900g to 3400g, closely with PaCO2 in most clinical situa-
To determine the corre- with gestational ages from tions in neonates. Patients with underlying
Prospective non randomised quan-
lation between ETCO2 28 weeks to 42, ventilated lung disease had the lowest correlation
Saili, A. and Level 2b titative study. A total of 152 ETCO2/
and PaCO2 in various for various indications such coefficient (‘r’ =0.55), whilst patients
Dutta, A.K. PaCO2 pairs analysed.
clinical situations. as Severe Birth Asphyxia who were ventilated for non pulmonary
1997 (SBA), Meconium Aspira- reasons demonstrated a higher correlation
tion Syndrome (MAS), and coefficient (‘r’=0.96).
New Dehli,
Hyaline Membrane Disease
India (HMD).
The overall correlation coefficient was
A total of 133 ETCO2/PaCO2
Prospective nonrandomised quantita- 0.73, with a p < 0.001. The ETCO2 value
Bhat, Y.R. and pairs were analysed from 32
tive study. was lower than the corresponding PaCO2
ventilated newborns.
value in 86.5% pairs.
‘r’ more than or equal to 0.92 in neonates
Mean gestational age was Total of 133 ETCO2/PaCO2 pairs from ventilated for sepsis, asphyxia and apnoea
Abhishek, N. To determine the 34.6 +/- 3.8 weeks. 32 patients. of prematurity. ‘r’ =0.67 in the HMD group
correlation and agree- and ‘r’ = 0.69 in MAS group.
ment between ETCO2
Level 2b Neonates who received surfactant had a
and PaCO2 in newborns Birthweight was 2200 grams
2008 ventilated for various +/- 780 grams. better ‘r’ value than those than did not
(0.76 vs 0.60).

Working Papers in Health Sciences 1:1 ISSN 2051-6266 / 20120005

clinical conditions.
The researchers stratified the
patients according to reasons for
ventilation. Patients in the pulmonary
India
group were further stratified into
those who received surfactant and
those who did not.
 20 patients (13 in the pulmo- Low flow capnography accurately meas-
Prospective nonrandomised pilot
Haggerty, J.J., nary group and 7 in the non ured alveolar CO2 in newborns without
study.
pulmonary group). pulmonary disease.

Eligible infants were me-

chanically ventilated and had Capnography was performed through
Kleinman, M.E.,
indwelling arterial catheters a side port in the proximal ETT.
insitu.

The measured PeTCO2 -PaCO2 gradients

PeTCO2 was measured for one minute
Zurakowski, D., were much higher in newborns with un-
pre and post ABG sampling.
To evaluate the accuracy derlying pulmonary disease.
of a new low flow side
stream capnography 2 groups were identified:
and analyse the compo- newborns who were receiv-
Level 2b
nents of the capnogram ing ventilation for underly-
Lyons,A.C. and in ventilated newborns ing pulmonary disease and
with and without pul- newborns receiving post
monary disease. operative mechanical ven-
tilation with no underlying
pulmonary disease.
Newborns in the pulmonary group
(n=13) and newborns in the control 4 Waveforms were quantified which may
Krauss, B group (n=7)were matched for birth- be useful differentiating patients with pul-

Working Papers in Health Sciences 1:1 ISSN 2051-6266 / 20120005

weight, gestational age and postnatal monary disease from those without.
age.
2002
Mean PeTCO2 - PaCO2 gradients were
United States
different between the two groups.

Table 2: Detail of selected studies

 Clinical Review

Methodology server bias through use of the same res- that capnometers were calibrated as
piratory therapist to collect all data. per manufacturer’s recommendations.
Study Design Kugelman et al (2008) and Haggerty et al
Data Analysis (2002) did not report any control for cali-
This review focuses on analysing the re- bration of these monitors. All studies in
lationship between ETCO2 and PaCO2 Ethics this review used the ‘gold standard’ (ABG)
which requires potential confounding as a reference interval. The researchers
factors to be tightly controlled. All stud- Rozycki et al (1998), Wu et al (2003 ) and further validated the studies by allowing
ies in this review were therefore correla- Haggerty et al (2002) explicitly demon- for upper and lower thresholds for CO2
tional and non randomised. strated in their studies the application concentrations to be determined, provid-
To ensure the validity of results and of the ethical principles of research (ICN ing reference ranges for hypocarbia and
minimise subjective bias, all patients 2003). hypercarbia.
were evaluated using both the reference
standard (ABG) and the test of interest Type of capnometer used Patient population characteristics
(ETCO2). Although this method of allo-
cation can potentially lead to systematic The capnometers used in the clinical set- Singh and Singhal (2005), Kugelman et al
bias, the researchers stratified patients ting utilise either side-stream or main- (2008), Bhat and Abhishek (2008), Hag-
into cohorts based on patient character- stream sampling. Kirpalani et al (1991) gerty et al (2002) and Rozycki et al (1998)
istics. showed that side-stream sampling un- demonstrated a good degree of homoge-
derestimates PaCO2 due to the relative- neity about gestational age, postconcep-
Data Collection ly low tidal volumes and rapid breathing tual age, birth weight, gender, diagnosis
rates in neonates resulting in falsely low and management strategies. In the study
The manner in which data is obtained can ETCO2 readings. Another disadvantage is conducted by Wu et al (2003), it is estab-
be susceptible to influences that can alter the delay between sampling and meas- lished that the researchers did not pro-
the results (Polit and Beck 2010). In ad- urement of ETCO2 (Pascucci et al 1989). vide information about baseline charac-
dition, clearly defined standard protocols However McEvedy et al (1990) showed teristics of the patients recruited into the
for data collection are important in avoid- that both mainstream and sidestream study. Nangia et al (1997) demonstrated
ing verification and workup bias (Dawes technology produced similiar results. a good degree of generalisability by divid-
et al 2005). In all the studies, patients For the studies in this review, Rozycki ing the group into 3 cohorts.
received both ABG and ETCO2 monitor- et al (1998), Wu et al (2003), Bhat and
ing, thereby ensuring workup bias was Abhishek ( 2008) and Singh and Singhal Confounding Factors
avoided. However, none of the studies (2005) used mainstream sampling, whilst
discussed a standard protocol of data Nangia et al (1997) used side-stream The researchers had to demonstrate tight
gathering or analysis. For interventional sampling and Haggerty et al (2002) used control of confounding factors for both
trials, the accepted analysis is by inten- micro-stream sampling. However Kugel- ETCO2 and PaCO2 measurements so that
tion to treat. The equivalent in this study man et al (2008) aimed to demonstrate the relationship between these could be
would equate to ensuring that all record- that distal ETCO2 (DETCO2) was a better understood, thereby eliminating any fac-
ed samples were assessed and that each predictor of PaCO2 than mainstream tors that may obscure the relationship
paired ETCO2 and PaCO2 was assessed. ETCO2. The use of different methods of and contribute to type 1 and type 2 er-
Wu et al (2003) refers to measurements ETCO2 monitoring used across the vari- rors (Polit and Beck 2010). In order to as-
being scattered throughout the course of ous studies allowed for a comparison to sess for confounding factors, Rozycki et al
the disease but no further information is be made between the different types of (1998) established 2 groups based on the
provided to allow for reproducibility out- capnography . difference between PaCO2 and ETCO2
side this study. Bhat and Abhishek (2008) where the difference of </- 5mmHg was
and Rozycki et al (1998) stated that ABGs Reliability and accuracy of the defined as a low bias group and a differ-
were performed at the discretion of the capnometer ence of >/- 5mmHg was defined as a high
clinician, whilst Singh and Singhal (2005) bias group. Similiarly Singh and Singhal
have not provided any information on Reliabiity refers to the accuracy and (2005), Kugelman et al (2008), Wu et al
how data was collected. This limitation consistency of the information obtained (2003) and Bhat and Abhishek (2008)
lends itself to various degrees of errors (Polit and Beck 2010). One method of demonstrated an assessment of con-
in measurement and biases. Failure to quantifying this would be through prop- founding factors by assessing for high and
establish uniformity in data gathering in- er calibration and application of the low bias groups.
fluences the reliability of results. In the capnometer to the breathing circuit. In
study performed by Nangia et al (1997), addition, assessment of sensitivity and Results
it is unclear whether the tests were per- specificity will strengthen the reliability
formed independently of each other of the capnometer allowing the end user An important feature of clinical trials is
thereby possibly lending to reviewer bias. to interpret these findings with a degree that they should be comparative. All stud-
However Haggerty et al (2002) provided a of confidence. ies in this review used a control group of
basic guide for data collection which cor- Rozycki et al (1998), Wu et al (2003), Bhat patients without co-existing lung disease,
related with waveform analysis. Wu et al and Abhishek (2008), Nangia et al (1997) i.e. non pulmonary group. The aim of this
(2002), attempted to minimise inter-ob- and Singh and Singhal (2005) reported review was to determine whether ETCO2

Working Papers in Health Sciences 1:1 ISSN 2051-6266 / 20120005 7

 Clinical Review

is a good predictor of PaCO2. One such Kugelman et al (2008) found that DETCO2 In analysing the data presented, it has
method of quantifying this relationship was a reliable predictor of PaCO2 where become clear that whilst ETCO2 moni-
would be to examine the relationship ‘r’ = 0.72 with a p value <0.001, and toring cannot replace the gold standard
between these two variables. All the re- mainstream ETCO2 was a poor predictor of measuring PaCO2 it can provide use-
searchers, with the exception of Haggerty of PaCO2 with ‘r’= 0.21 and a p value of ful information if used as a trending tool.
et al (2002) demonstrated this relation- <0.005. This is further supported by Rozycki et
ship using correlation coefficients. Hag- All the reviewed studies strengthened the al (1998), Wu et al (2003), Bhat and Ab-
gerty et al used the Pearson’s Correlation. outputs of their studies by using a control hishek (2008)and Nangia et al (1997) who
In a study conducted by Wu et al (2003) non pulmonary group. In all these stud- found a good correlation exists between
of 60 neonates, the researchers observed ies, patients in the non pulmonary group ETCO2 and PaCO2 in patients without
a correlation in both term infants (44 showed a better correlation between underlying lung disease. Whilst the re-
samples, r = 0.78, p<0.001) and preterm ETCO2 and PaCO2. searchers were able to demonstrate a
infants (86 samples, r = 0.85, p<0.001). good degree of transferability and gener-
Similarly, Rozycki et al (1998) in a study Discussion alisability through appropriate methods
of 45 newborns demonstrated a good of sampling and methodology, the data
correlation between ETCO2 and PaCO2 From conducting this review it can be was further validated by statistical analy-
in ventilated neonates including preterm deduced that limitations exist when ap- sis relevant to this type of study.
infants, showing correlation coefficients plying ETCO2 monitoring to ventilated In the research presented by Haggerty
of ‘r’ = 0.833 and 0.821 respectively with neonates. Advances in technology have et al (2002), Singh and Singhal (2005)
a 95% CI for the biases and p< 0.001. In allowed modification of capnometers to and Kugelman et al (2008), although the
the study conducted by Bhat and Ab- address the challenges posed by the ne- researchers were able to demonstrate
hishek (2008), the researchers were able onatal physiology. This is supported by a good correlation between ETCO2 and
to show a correlation and agreement a study done by Hopper et al in a rab- PaCO2, inconsistencies lie in the way in
between ETCO2 and PaCO2 by demon- bit model. Here researchers were able which data was collected and communi-
strating a correlation coefficient of ‘r’ = to show that inducing lung injury with cated. Singh and Singhal (2005) limited
0.73, with a CI of 95%, where p< 0.001. meconium instillation reduced the cor- the study to extremely low birth weight
Although a good correlation was demon- relation between ETCO2 and PaCO2 from infants making the findings less transfer-
strated, the correlation between ETCO2 0.94 to 0.80 and raised the bias. able, thus making it difficult to extend
and PaCO2 was greater in the group who comparison to infants outside this cohort.
were ventilated for sepsis, asphyxia, and A common finding in the studies in this Whilst Haggerty et al (2002) demonstrat-
apnoea versus those ventilated for Hya- review was that in neonates with un- ed a good degree of agreement by eval-
line Membrane Disease or Meconium As- derlying lung disease ETCO2 poorly cor- uating the use of the capnometer across
piration Syndrome. A similiar finding was related with PaCO2. In looking at this two groups, the researchers did not pro-
demonstrated in all studies. This leads to group further it was concluded that the vide any confidence limits or intervals
the conclusion that conditions with pa- patients were predominantly preterm and a correlational relationship between
renchymal lung pathology produce a low- with surfactant deficient lung disease. ETCO2 and PaCO2 was not established.
er correlation. Surfactant is necessary for lung alveoli
to overcome surface tension and remain Summary
Singh and Singhal (2005) conducted a ret- open. Without adequate surfactant, the
rospective chart review of extremely low shearing forces exerted trying to open Pulse oximetry and transcutaneous mon-
birth weight infants (ELBW) of 754 paired alveoli by either the infant’s own inspir- itoring have been accepted methods of
samples using simple linear regression. atory effort or by a mechanical ventilator monitoring in the NICU for decades. In-
The researchers reported an intraclass can induce lung damage. The deficiency terpretation of SpO2 is a representation
correlation of 0.81. in surfactant decreases lung compliance of oxygenation only and can be flawed by
Bhat and Abhishek (2008) and Haggerty and functional residual capacity result- limitations such as signal to noise ratio in-
et al (2002), demonstrated a higher gra- ing in an increased physiological dead duced as a result of poor perfusion and
dient between ETCO2 and PaCO2 when space leading to ventilation/perfusion motion artefact. Whilst technologic and
comparing infants ventilated for pulmo- mismatch and poor tidal volumes. This engineering advances have addressed
nary disease to those ventilated for non pathological picture may explain why these known shortcomings, no definitive
pulmonary conditions. Similarly Nangia ETCO2 poorly correlates with PaCO2 in solution has been derived to date to en-
et al (1997) studied 152 samples in pre- neonates with underlying lung disease. sure an accurate SpO2 reading for any
term infants <32 weeks gestational age The studies under review showed that patient under any circumstances. Whilst
and reported a correlation of 0.55 in the preterm infants who received surfactant SpO2 monitoring provides valuable data
HMD group. This finding conforms to the therapy showed an improved correlation regarding oxygenation its contribution to
theory that ETCO2 poorly predicts PaCO2 between ETCO2 and PaCO2. monitoring ventilation efficacy is limited
in ventilated neonates with lung This fits with the theory that following as it provides no data about CO2 or met-
disease (Watkins and Weindling 1987) surfactant administration, surface ten- abolic balance that is otherwise obtained
and would be consistent with the disease sion is reduced, more alveoli are recruit- from arterial blood gas measurement.
pattern of these babies i.e. atelectasis ed,and tidal volumes improve thereby
with consequent ventilation-perfusion reducing ventilation-perfusion mismatch. Transcutaneous monitoring represents a
mismatch. simple non invasive technique for contin-

Working Papers in Health Sciences 1:1 ISSN 2051-6266 / 20120005 8

 Clinical Review

uous monitoring of O2 and CO2. Accura-

cy of this method of monitoring is limit- References Kugelman, A., Zeiger-Aginsky,D,
ed by factors such as thermal injury, skin D., Bader, D.,Shoris, I. and Riskin,
maturity, long stabilisation times and the American Academy of Respiratory A.(2008) ‘A novel method of distal
need for frequent calibrations and site Care (2004) ‘Capnography/ end-tidal CO2 capnography in
rotation. Capnometry During Mechanical intubated infants: a comparison
Ventilation: Revision and Update’, with arterial CO2 and with
Whilst it can be appreciated that advanc- Respiratory Care, 49, pp 534-539. proximal mainstream end-tidal
es in technology have allowed the cap- CO2’,Paediatrics, 122, pp 1219-
nometer to be more applicable to the ne- Bhat, Y.R. and Abhishek, N. (2008) 1224.
onatal population, it cannot replace the ‘Mainstream end-tidal carbon
gold standard of ABG monitoring. ABGs dioxide monitoring in ventilated McEvedy, B.A., McLeod, M.E.,
not only provide valuable data measur- neonates’, Singapore Medical Kirpalani, H., Volgyesi, G.H. and
ing efficacy of ventilation but can also Journal, 49 (3), pp 199-203. Lerman, J. (1990) ‘End-tidal
provide a range of valuable data such as carbon dioxide measurements in
electrolytes, blood glucose, lactate etc. Carter, R. and Williams, J.S. (2008) critically ill neonates: a comparison
‘Oximetry Update: Applications, of side-stream and mainstream
All the studies in this review have ex- Limitations, and New Technologies’, capnometers’, Canadian Journal of
plored the correlational relationship be- Respiratory Therapy for decision Anaesthesia; 37, pp 322-326.
tween ETCO2 and PaCO2. It has become makers in respiratory care, Available
clear that the accuracy of ETCO2 in pre- at: http://www.rtmagazine.com/ National Audit Office (2007)
dicting PaCO2 in ventilated newborn in- issues.articles/ Improving Quality and Safety.
fants is limited by various components Available at: http://www.nao.org.
specific to neonatal pulmonary physiolo- Colman, Y. and Krauss, B. (1999) uk/publications/0607/primary_
gy and the presence or absence of co-ex- ‘Microstream capnography care_governance.aspx (Accessed in
isting parenchymal lung disease. How- technology: a new approach to an September 2009).
ever, by accepting and understanding its old problem’, Journal of Clinical
limitations, capnometry is easily applied Monitoring and Computing, 15(6), Nangia, S., Saili, A. and Dutta,
and does provide continuous trending pp 403-409. A.K. (1997) ‘End tidal carbon
data that can rapidly alert clinicians to dioxide monitoring-Its reliability
significant changes in PaCO2 that may Dawes, M., Philip, D., Alastair, G., in neonates’, Indian Journal of
require prompt respiratory assessment. Jonathan, M., Kate, S. and Robin, S. Paediatrics, 64, pp 389-394.
Further advances in capnometer design (2005) Evidence- Based Practice: A
e.g. micro-stream sampling, will necessi- Primer for Health Care Professionals. Pascucci, R.C., Schena, J.A. and
tate further evaluation of this modality in London: Elsevier Limited. Thompson, J.E. (1989) ‘Comparison
clinical practice anticipating that the cur- of a sidestream and mainstream
rent performance will be improved upon. Hagerty, J.J., Kleinman, M.E., capnometer in infants’, Critical Care
The ability to continuously display ETCO2 Zurakowski, D., Lyons, A.C. and Medicine,17, pp 560-562.
can therefore potentially provide a valua- Krauss, B. (2002) ‘Accuracy of a new
ble adjunct to our monitoring and assist low-flow side stream capnography Polit, D.F. and Beck, C.T. (2006)
in optimising the management of the technology in newborns’, Journal of Essentials of Nursing Research:
ventilated newborn infant on the neona- Perinatology, 22 (3), pp 219-225. Methods, Appraisal, and Utilization.
tal unit. Crawfordsville:R. R. Donnelley.
Hammer, J. (2006).’Ventilator
Strategies-What monitoring is Rozycki, H.J., Sysyn, G.D., Marshal,
helpful?’, Paediatric Respiratory M.K., Malloy, R. and Wiswell, T.E.
Reviews, 7, pp 183-185. (1998) ‘Mainstream end-tidal
8. International Council of carbon dioxide monitoring in
Nurses (2003) ICN. Available at: the neonatal intensive care unit’,
http://www.icn.ch/ (Accessed in Paediatrics, 101, pp 648-53.
January 2010).
Singh, S.A. and Singhal, N (2006)
Kirpalani, H., Kechagrias, S. and ‘Does End -tidal carbon dioxide
Lerman, J. (1991) ‘Technical and measurement correlate with
clinical aspects of capnography arterial carbon dioxide in extremely
in neonates’, Journal of Medical low birth weight infants in the first
Engineering and Technology, week of life?’, Indian Pediatrics,43
15(4/5), pp 154-161. ,pp 20-25.

Working Papers in Health Sciences 1:1 ISSN 2051-6266 / 20120005 9

 Clinical Review

Thompson, C. and Dowding, D.

(2002) Clinical Decision Making
and Judgement in Nursing. London:
Elsevier Limited.

Wu, C.H., Chou, H.C., Hseih, W.S.,

Chen, W.K., Huang, P.Y. and Tsao, P.N.
(2003)’Good estimation of arterial
carbon dioxide by end-tidal carbon
dioxide monitoring in the neonatal
intensive care unit’, Pediatric
Pulmonology, 35 (4), pp 292-5.

Watkins, A.M. and Weindling, A.M.

(1987) ‘Monitoring of end tidal CO2
in the neonatal intensive care unit’,
Archives Dis Child, 62 (8), pp 837-
839.

Working Papers in Health Sciences 1:1 ISSN 2051-6266 / 20120005 10