British Journal of Anaesthesia 107 (4): 631–5 (2011)
Advance Access publication 23 June 2011 . doi:10.1093/bja/aer171
RESPIRATION AND THE AIRWAY
Dependence of the gradient between arterial and end-tidal
PCO2 on the fraction of inspired oxygen
H. Yamauchi 1,2*, S. Ito 2, H. Sasano 2, T. Azami 2, J. Fisher 3 and K. Sobue 2
1
TOYOTA-KAI Medical Corporation KARIYA TOYOTA General Hospital, Kariya, Aichi, Japan
Department of Anesthesiology and Medical Crisis Management, Nagoya City University, Graduate School of Medical Sciences, 1
Azakawasumi, Mizuho-cho, Mizuho-ku, Nagoya, Aichi 467-8601, Japan
3
Department of Anaesthesiology, University Health Network, 190 Elizabeth St, RFE IS-401, Toronto, Ontario, Canada
2
* Corresponding author. E-mail: [email protected]
Editor’s key points
† End-tidal CO2
concentrations usually
reflect arterial CO2
tensions.
† Absolute values depend
on the end-tidal to
arterial CO2 gradient and
alveolar dead space, but
were thought to be
independent of FIO2.
† Here, the arterial to
end-tidal CO2 gradient
increased significantly
with increased FIO2 in
anaesthetized ventilated
patients.
† This may reflect changes
in alveolar dead space,
pulmonary shunt, or
cardiac output.
′
Background. End-tidal PCO2 (PECO
) is routinely used in the clinical assessment of the
2
′
adequacy of ventilation because it provides an estimate of PaCO2. How well PECO
reflects
2
′
PaCO2 depends on the gradient between them, expressed as DPa2 ECO2. The major
′
determinant of DPa2 ECO
is alveolar dead space (VDalv). The fraction of inspired O2 (FIO2)
2
′
is not thought to substantially affect DPa2 ECO
in anaesthetized patients. We
2
′
hypothesized that a high FIO2 may indeed increase DPa2 ECO
by preferentially
2
vasodilating well-perfused alveoli, resulting in the redistribution of blood flow to these
alveoli from poorly perfused alveoli and an increase in VDalv. We therefore investigated
′
the effects of changes in FIO2 on DPa2 ECO
and VDalv.
2
Methods. With Institutional Review Board approval and informed consent, we studied 20 ASA
I –II supine patients undergoing elective lower abdominal surgery under combined general
and epidural anaesthesia. At constant levels of ventilation, FIO2 levels of 0.21, 0.33, 0.5,
′
0.75, and 0.97 were applied in a random order and DPa2 ECO
and VDalv were calculated.
2
′
values were, in order of ascending FIO2, {mean [standard error of the
Results. The DPa2 ECO
2
mean (SEM)]} 0.13 (0.04), 0.28 (0.08), 0.29 (0.09), 0.44 (0.11), and 0.53 (0.09) kPa. The
corresponding values of VDalv were 25.5, 33.8, 35.8, 48.9, and 47.4 ml. Each successive
′
hyperoxic level showed a significant increase in DPa2 ECO
except between the 0.33–0.5
2
and 0.75–0.97 FIO2 levels.
′
, in anaesthetized patients depends on FIO2.
Conclusions. These data demonstrate that DPa2 ECO
2
Keywords: pulmonary gas exchange; respiratory dead space; respiratory physiological
phenomena; ventilators, mechanical
Accepted for publication: 21 April 2011
Anaesthetists use end-tidal PCO2 (PE′CO2 ) values routinely to
monitor the adequacy of ventilation. The clinical utility of
PE′CO2 as a surrogate for arterial PCO2 (PaCO2 ) depends on the
gradient between them (DPa − E′CO2 ). This gradient depends
primarily on the degree of alveolar dead space (VDalv).
Larson and Severinghaus1 reported that the administration
of O2 to awake sitting subjects led to an increase in
DPa − E′CO2 . They attributed this finding to the vasodilating
effect of O2 on blood vessels at the base of the lungs,
which results in the expansion of the apical VDalv. However,
Whitesell and colleagues2 failed to demonstrate such a
relationship in supine subjects under the influence of
general anaesthesia. In that study, the fraction of inspired
O2 (F IO2 ) was not well controlled and the range examined
was narrow. Therefore, in this study, we used a prolonged
steady state over a wide range of F IO2 levels to determine
whether hyperoxia increases DPa − E′CO2 and VDalv in supine
subjects during general anaesthesia.
Methods
This study was approved by the ethics committee of Nagoya
City University Graduate School of Medical Science, and
informed written consent was obtained from each participant. The subjects were ASA I/II grade patients undergoing
elective lower abdominal surgery; no patient received premedication. An epidural catheter was placed into the lower thoracic epidural space before induction of general anaesthesia.
After preoxygenation of the lungs with 100% O2, anaesthesia
was induced with propofol (1.5–2.0 mg kg21), fentanyl (1.5 –
2.0 mg kg21), and rocuronium (0.6 –1.0 mg kg21). Anaesthesia was maintained with air-and/or-oxygen–sevoflurane
& The Author [2011]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved.
For Permissions, please email: [email protected]
BJA
Statistical analysis
Statistical analysis was performed using SigmaStat (HULINKS,
Tokyo, Japan). The values of DPa − E′CO2 , VDalv, PaCO2 , and
PE′CO2 were analysed at different time intervals by one-way
analysis of variance for repeated measures followed by
Fisher’s least significant difference test. The values are
expressed as mean [standard error of the mean (SEM)]. A
value of P,0.05 was considered to be statistically significant.
Results
Study characteristics
Twenty subjects (15 women) were studied. Their characteristics were as follows: [mean (range)]: age, 60.5 (36– 78) yr;
height, 159.4 (138 –180) cm; and weight, 53.5 (35– 79) kg.
The settings of the artificial ventilators were as follows:
[mean (range)]: VT, 453.8 (350 –550) ml; f, 10.2 (8–12)
min21. The types of surgeries performed are listed in
Table 1. Three subjects had mild obstructive lung impairment
(Table 1).
632
Table 1 Patient characteristics (n¼20) expressed as mean
(range), mean (SD) or number. RFT, respiratory function test
Value
Age (yr)
60.5 (36 –78)
Sex (male:female)
05:15
Height (cm)
159.4 (11.0)
Weight (kg)
53.5 (11.5)
Tidal volume (ml)
453.8 (66.5)
Respiratory frequency (min21)
10.2 (1.1)
Types of surgery (cases)
Uterus
10
Colon
6
Rectum
3
Bladder
1
Abnormal RFT (cases)
3 (all FEV1s of 60 – 70%)
0.7
†#‡
†#‡
0.6
P a–E¢CO2 (kPa)
(1.5 –2.5%) combined with a continuous epidural infusion of
ropivacaine (0.375%) at a rate of 4–6 ml h21. Sevoflurane
concentration was kept constant and complete muscle relaxation was maintained during the measurement phase of our
study. Standard anaesthesia monitors were used. End-tidal
gas was sampled, and PCO2 , PO2, and sevoflurane concentrations were determined (GF-100; Nihon Kohden, Tokyo,
Japan). The end-tidal values and F IO2 were calculated by
the standard anaesthetic monitors. A respiratory monitor
(NICOTM; Novametrix, CT, USA) which calculated VDalv values
by single-breath CO2 analysis3 was installed in the respiratory
circuit. An indwelling radial arterial catheter allowed for continuous arterial pressure monitoring and intermittent arterial
blood sampling. All the monitors were calibrated before use.
The patients’ lungs were ventilated using a volumecontrolled mode (Aestiva/5; Datex-Ohmeda, Buckinghamshire, UK), I:E (inspiratory-to-expiratory) ratio, 1:2 without
PEEP. F IO2 levels of 0.21, 0.33, 0.5, 0.75, and 0.97 were
applied by altering the ratio of air to O2 (total inspiratory
flow, 6 litre min21) in a random order. The order in each
case was decided by drawing a card in sequence blindly
from five cards on which F IO2 levels were written. Tidal
volume (VT) and respiratory frequency ( f ) were both adjusted
to maintain PE′CO2 at 4.00 –4.65 kPa in the first F IO2 level of
each case. They were unchanged after the adjustment
during the study. After establishing a steady haemodynamic
and ventilatory state, only FIO2 levels were changed whereas
keeping both VT and f constant. When the levels were +1%
the target F IO2 level, each F IO2 level was maintained for 15
min or more. At the end of each F IO2 period, arterial blood
was withdrawn for over 1 min and blood gas analysis was
carried out immediately (ABL 725; Radiometer, Tokyo,
Japan). DPa − E′CO2 was obtained from the temperaturecorrected PaCO2 and PE′CO2 values and VDalv was calculated
by the respiratory monitor.
Yamauchi et al.
0.5
†
0.4
†
0.3
0.2
0.1
0
0.21
0.33
0.5
F IO2
0.75
0.97
Fig 1 The relationship between FIO2 and DPa − E′CO2 . DPa − E′CO2
significantly increased between all FIO2 levels, except between
the 0.33 –0.5 and 0.75 –0.97 levels. †P,0.05 vs 0.21. #P,0.05 vs
0.33. ‡P,0.05 vs 0.5.
The mean (SEM) values of DPa − E′CO2 at 0.21, 0.33, 0.5, 0.75,
and 0.97 levels of F IO2 were 0.13 (0.04), 0.28 (0.08), 0.29
(0.09), 0.44 (0.11), and 0.53 (0.09) kPa (n¼20), respectively.
DPa − E′CO2 values differed significantly at all FIO2 levels,
except between the 0.33–0.5 and 0.75– 0.97 levels.
DPa − E′CO2 increased consistently with an increase in F IO2
(Fig. 1). The corresponding values of VDalv were 25.5, 33.8,
35.8, 48.9, and 47.4 ml (mean), VDalv increased by 86% at
F IO2 = 0.97 compared with the value at F IO2 = 0.21 (Fig. 2).
The mean PaCO2 values at F IO2 levels of 0.21, 0.33, 0.5,
0.75, and 0.97 were 4.55, 4.52, 4.56, 4.54, and 4.76 kPa,
respectively. The corresponding mean values of PE′CO2 were
4.42, 4.24, 4.27, 4.10, and 4.23 kPa, respectively (Fig. 3).
The value of PaCO2 when the F IO2 was 0.97 was statistically
higher than the values at other F IO2 levels (Fig. 3).
Discussion
The major finding of this study is the dependence of
DPa − E′CO2 on F IO2 , which is associated with VDalv. We found
BJA
The dependence of DPa2 E′CO2 on FIO2
60
VDalv (ml)
50
†
†
0.33
0.5
F IO2
†#‡
†#‡
0.75
0.97
40
30
20
10
0
0.21
P a CO2 and P E¢ CO2 (kPa)
Fig 2 The relationship between FIO2 and VDalv. The VDalv values
differed significantly between all FIO2 levels, except between
the 0.33 – 0.5 and 0.75 – 0.97 levels. When FIO2 was 0.97, VDalv
increased by 86% when compared with the value when FIO2
was 0.21. †P,0.05 vs 0.21. #P,0.05 vs 0.33. ‡P,0.05 vs 0.5.
P a CO
2
5
P E¢ CO2
†
4.8
4.6
4.4
4.2
4
#
‡
0.21
0.33
0.5
F IO2
0.75
0.97
Fig 3 The relationships between FIO2 and PaCO2 , or PE′CO2 . The
value of PaCO2 when FIO2 was 0.97 was statistically higher than
the values at other FIO2 levels. The value of PE′CO2 at the FIO2
level of 0.21 was statistically higher than the values at other
FIO2 levels, and the value at the FIO2 of 0.75 was statistically
lower than the values at other FIO2 levels. #P,0.05 vs 0.33, 0.5,
0.75, and 0.97. ‡P,0.05 vs 0.21, 0.33, 0.5, and 0.97. †P,0.05 vs
0.21, 0.33, 0.5, and 0.75.
that VDalv increased to a small but measurable degree with
increasing FIO2 , resulting in an increase in DPa − E′CO2 , which
is assumed to be constant when PE′CO2 is used to assess the
adequacy of ventilation. Few studies have reported on the
effect of the phase change in F IO2 on DPa − E′CO2 and VDalv.
The respiratory monitor used in our study utilizes the
single breath test for CO2 (SBT-CO2) analysis to calculate
VDalv. This method was validated in animal experiments4
and the values thus calculated did not differ from those calculated using the Bohr–Enghoff equation in ventilated
newborns.5
DPa − E′CO2 depends on factors that increase the fluctuation of alveolar PCO2 (PACO2 ). PACO2 usually fluctuates over
the ventilatory cycle, whereas PaCO2 reflects the much
dampened time-averaged PACO2 . At end-exhalation, PACO2 is
maximal and approaches the PCO2 of mixed venous blood.6
Thus, increases in the fluctuation of PACO2 , caused by
changes in VT7,8, f 8,9, and CO2 production (VCO2),7,9 increase
DPa − E′CO2 . In this study, VT and f were kept constant. VCO2
also likely remained constant, because the subjects were
anaesthetized and maintained under normothermic conditions. The changes in DPa − E′CO2 were due to the changes
in PE′CO2 , because PaCO2 remained constant, except when
F IO2 was 0.97. Hardman and Aitkenhead10 indicated a high
correlation between DPa − E′CO2 /PaCO2 and VDalv/alveolar
tidal volume in the presence of intrapulmonary shunt or
VCO2 in their model analysis. Thus, the changes in
DPa − E′CO2 were most likely due to the dilution effect of
PACO2 with gas from VDalv. We presume that the mechanism
of the increase in VDalv with the increase in FIO2 is that high
F IO2 reduces vascular resistance in high perfusion alveolar
areas, resulting in the redistribution of blood flow away
from low perfusion alveolar areas as proposed by Larson
and Severinghaus.1
Intrapulmonary shunt can also influence DPa − E′CO2 . It
leads to the increase in DPa − E′CO2 as a consequence of
the increase in PaCO2 with venous admixture. Although this
gap with the shunt has nothing to do with real dead
space, it is erroneously interpreted as VDalv, which has
been referred to as ‘shunt dead space’. Thus, the calculated
VDalv includes the ‘shunt dead space’ and is overestimated.11
However, the influence of the shunt on the calculated VDalv
is fairly small in low shunt values. According to Mecikalski
and colleagues,12 even if the shunt rate is 20%, ‘shunt
dead space’ is only a small percentage of the VT. Similarly,
Suter and colleagues13 reported that ‘shunt dead space’
amounted to only 1–2% of the VT. A major cause for intrapulmonary shunt during general anaesthesia is the development of atelectasis.14 Breathing pure O2 also causes
absorption atelectasis. Dantzker and colleagues15 reported
that breathing pure O2 induced absorption atelectasis
within 9 min and that pulmonary shunt increased abruptly
when FIO2 increased above 0.8. Marntell and colleagues16
reported that intrapulmonary shunt created by F IO2 above
0.95 remained larger when F IO2 was returned to 0.21, indicating that absorption atelectasis produced during O2-rich
breathing persisted throughout anaesthesia in horses.
Brismar and colleagues17 found that in chest computed
tomography study, atelectasis rapidly developed in the
dependent lung regions during the induction of general
anaesthesia, probably with pure O2 preoxygenation, but it
did not progress with time or increased F IO2 . General anaesthesia was induced after preoxygenation using 100% O2,
which is a standard clinical practice. Some absorption
atelectasis must have been produced before the beginning
of the study. Although an inflation of the lungs to airway
pressure of 3.92 kPa (40 cm H2O)18 or an application of
PEEP of 0.98 kPa (10 cmH2O)19 can eliminate the atelectasis,
we did not perform them. The atelectasis produced on the
induction of anaesthesia must have persisted, even if F IO2
was changed. Thus, the atelectasis would have little
impact on the phase increases in DPa − E′CO2 . However, it
633
BJA
remains possible that an increment of the shunt effect due
to the redistribution of pulmonary flow with the changes in
F IO2 was associated with the increase in DPa − E′CO2 . The
increment of the shunt effect should cause the increase in
DPa − E′CO2 with the increase in PaCO2 . In this study, the
increase in DPa − E′CO2 with the increase in PaCO2 seen just
at F IO2 of 0.97 would reflect the reversible addition of the
shunt effect.
A hyperoxia-induced reduction in cardiac output may contribute to an increase in DPa − E′CO2 and VDalv. Anderson and
colleagues20 reported that cardiac output decreased by
10.3% when F IO2 was changed from 0.21 to 1.0. Johann
and colleagues21 reported that cardiac output decreased by
10.6% after a change in F IO2 from ≤0.6 to 1.0 in patients
after coronary artery bypass surgery. The decrease in
cardiac output increases DPa − E′CO2 and VDalv by causing
the decrease in CO2 elimination from alveoli and the redistribution of pulmonary blood flow. Isserles and Breen22
reported that during constant minute ventilation and tissue
VCO2, there were reductions in PE′CO2 of 7.4% and PaCO2 of
4.7% which resulted in the increase in DPa − E′CO2 when the
reduction in cardiac output was 10%. If both PE′CO2 and
PaCO2 are 5.33 kPa, DPa − E′CO2 should increase by 0.13
kPa. Although we did not detect any changes in heart rate
and arterial pressure during this study or find any changes
in cardiac output in another experiment changing F IO2 in
the same manner as this study, it is likely that a small
decrease in cardiac output contributed to the increases in
DPa − E′CO2 and VDalv partially in high F IO2 . However, in this
study, we did not identify a decrease in PaCO2 with the
decrease in cardiac output in high F IO2 . We believe that
was because the effect on PaCO2 of the increase in the
shunt exceeded that of the reduction in cardiac output at
F IO2 of 0.97.
Isoflurane increases VDalv more than propofol under
positive pressure ventilation.23 Although sevoflurane we
used may have the same mode of action as isoflurane, its
concentration was kept constant throughout this study. Ventilation– perfusion relationships during epidural analgesia
show no significant changes when compared with the
control state.24 Therefore, neither sevoflurane nor epidural
analgesia would have influenced our results. Although
there is a gradual increase in VDalv with an increased duration
of general anaesthesia,25 this effect is eliminated by varying
the F IO2 levels in a random order.
In summary, we demonstrated that DPa − E′CO2 depends
on F IO2 in anaesthetized ventilated patients. Not only the
true VDalv but also intrapulmonary shunt and cardiac output
may be associated with the magnitude of DPa − E′CO2 in the
presence of variable FIO2 . Although the effect of F IO2 on
DPa − E′CO2 is small, it should be taken into account when
using PE′CO2 as a surrogate for PaCO2 .
Conflict of interest
None declared.
634
Yamauchi et al.
References
1 Larson CP, Severinghaus JW. Postural variations in dead space
and CO2 gradients breathing air and O2. J Appl Physiol 1962; 17:
417– 20
2 Whitesell R, Asiddao C, Gollman D, Jablonski J. Relationship
between arterial and peak expired carbon dioxide pressure
during anesthesia and factors influencing the difference. Anesth
Analg 1981; 60: 508– 12
3 Fletcher R, Jonson B, Cumming G, Brew J. The concept of dead
space with special reference to the single breath test for carbon
dioxide. Br J Anaesth 1981; 53: 77– 88
4 Arnold JH, Thompson JE, Arnold LW. Single breath CO2 analysis:
description and validation of a method. Crit Care Med 1996; 24:
96– 102
5 Wenzel U, Wauer RR, Schmalisch G. Comparison of different
method for dead space measurements in ventilated newborns
using CO2-volume plot. Intensive Care Med 1999; 25: 705– 13
6 Dubois AB, Britt AG, Fenn WO. Alveolar CO2 during the respiratory
cycle. J Appl Physiol 1952; 4: 535– 48
7 Jones NL, Robertson DG, Kane JW. Difference between end-tidal
and arterial PCO2 in exercise. J Appl Physiol 1979; 47: 954–60
8 Ito S, Mardimae A, Han J, et al. Non-invasive prospective targeting
of arterial PCO2 in subjects at rest. J Physiol 2008; 586: 3675– 82
9 Lamb TW, Anthonisen NR, Tenney SM. Controlled frequency
breathing during muscular exercise. J Appl Physiol 1965; 20:
244– 8
10 Hardman JG, Aitkenhead AR. Estimating alveolar dead space
from the arterial to end-tidal CO(2) gradient: a modeling analysis.
Anesth Analg 2003; 97: 1846–51.
11 Tang Y, Turner MJ, Baker AB. Effects of alveolar dead-space, shunt
and V/Q distribution on respiratory dead-space measurements. Br
J Anaesth 2005; 95: 538–48.
12 Mecikalski MB, Cutillo AG, Renzetti AD Jr. Effect of right-to-left
shunting on alveolar dead space. Bull Eur Physiopathol Respir
1984; 20: 513– 9
13 Suter PM, Fairley B, Isenberg MD. Optimum end-expiratory airway
pressure in patients with acute pulmonary failure. N Engl J Med
1975; 292: 284– 9
14 Bendixen HH, Hedley-Whyte J, Laver MB. Impaired oxygenation in
surgical patients during general anesthesia with controlled ventilation: a concept of atelectasis. N Engl J Med 1963; 269: 991– 6
15 Dantzker DR, Wagner PD, West JB. Instability of lung units with
low VA/Q rations during O2 breathing. J Appl Physiol 1975; 38:
886– 95
16 Marntell S, Nyman G, Hedenstierna G. High inspired oxygen concentrations increase intrapulmonary shunt in anaesthetized
horses. Vet Anaesth Analg 2005; 32: 338– 47
17 Brismar B, Hedenstierna G, Lundquist H, Strandberg A,
Svensson L, Tokics L. Pulmonary densities during anesthesia
with muscular relaxation—a proposal of atelectasis. Anesthesiology. 1985; 62: 422– 8
18 Hedenstierna G, Rothen HU. Atelectasis formation during
anesthesia: causes and measures to prevent it. J Clin Monit
Comput 2000; 16: 329– 35
19 Tokics L, Hedenstierna G, Stranberg A, Brismar B, Lundquist H.
Lung collapse and gas exchange during general anesthesia:
effect of spontaneous breathing, muscle paralysis, and positive
end-expiratory pressure. Anesthesiology 1987; 66: 157–67
20 Anderson KJ, Harten JM, Booth MG, Kinsella J. The cardiovascular
effects of inspired oxygen fraction in anaesthetized patients. Eur
J Anaesthesiol 2005; 22: 420– 5
The dependence of DPa2 E′CO2 on FIO2
21 Johann M, Keith J, John K. Normobaric hyperoxia reduces cardiac
index in patients after coronary artery bypass surgery.
J Cardiothorac Vasc Anesth 2005; 19: 173–5
22 Isserles SA, Breen PH. Can changes in end-tidal PCO2
measure changes in cardiac output? Anesth Analg 1991; 73:
808– 14
23 Praetel C, Banner MJ, Monk T, Gabrielli A. Isoflurane inhalation
enhances increased physiologic deadspace volume associated
BJA
with positive pressure ventilation and compromises arterial oxygenation. Anesth Analg 2004; 99: 1107–13
24 Lundh R, Hendenstierna G, Johansson H. Ventilation-perfusion
relationships during epidural analgesia. Acta Anaesthesiol Scand
1983; 27: 410–41
25 Askrog VF, Pender JW, Smith TC, Eckenhoff JE. Changes respiratory dead space during halothane, cyclopropane, and nitrous
oxide anesthesia. Anesthesiology 1964; 25: 342–52
635