- Email: [email protected]
Transcutaneous oxygen measurement in paediatric respiratory disease
Br. J. Dis. Chest (1987) 81, 128
TRANSCUTANEOUS OXYGEN MEASUREMENT IN PAEDIATRIC RESPIRATORY DISEASE B. G. LOFI’US, W. BARRY AND .I. F. PRICE Department of Child Health and Chest Unit, King’s College Hospital, London, SE5 8RX
Summary
We have measured simultaneous arterial (Pao,) and transcutaneous oxygen tension (fcPoJ in 13 children, aged 8 months to 10 years, with respiratory disease. A Radiometer TCM2 transcutaneous oxygen monitor was used. At an electrode temperature of 44°C the regression equation was tcPo,=O.9 Pao,-10 and at 44.5”C tcPo,=O.9 Pao,-8. The 95% confidence limits were& 13 torr at both temperatures. We also studied the usefulness of the method in monitoring trends in oxygenation in individual patients. In six children paired observations were made on five or more occasions and in these &PO, followed changes in Pao, accurately. We conclude that this technique is not sufficiently reliable for predicting arterial values, but in individual patients it accurately reflects trends in oxygenation. INTRODUCTION There is a close relationship between transcutaneous and arterial oxygen values in the newborn (1). Few studies have assessed the technique in older children (2-5) and opinion varies as to the reliability of the method. We have studied the relationship between transcutaneous and arterial oxygen levels in a group of children with respiratory disease, assessed the influence of electrode temperature on that relationship, and evaluated the usefulness of the method in individual patients. Patients and Methods Thirteen patients aged 8 months to 10 years were studied. Three had bronchopulmonary dysplasia, two asthma, two bronchopneumonia, two bronchitis, two aspiration pneumonia, one pertussis and one upper airway obstruction. All had a clinical indication for the measurement of blood gases. No patient was shocked, hypotensive or hypothermic at the time of the study. Five had indwelling arterial cannulae. Nine children aged 8 months to 8 years (mean aged 2.4 years) were studied using an electrode temperature of 44°C and eight children aged l-10 years (mean age 4.3 years) with an electrode temperature of 44S”C. Four of the children were studied at both temperatures. Correspondence to: Dr B. G. Loftus, Department of Thoracic Medicine, King’s College Hospital, London, SE5 8RX.
LOFTUS,
BARRY,
PRICE:
TRANSCUTANEOUS
OXYGEN
MEASUREMENT
129
A Radiometer TCM2 transcutaneous oxygen monitor was used. The type of oxygen electrode used detects changes in arterial oxygen tension within 60 seconds (6). An automatic single-point in vitro calibration was performed with air as the calibration gas, the electrode was then placed on the upper chest, on an area of skin which had been cleaned with alcohol, and gently abraded with gauze. The electrode was allowed to stabilize for at least 20 minutes. Arterial blood was drawn from an indwelling cannula, or by direct radial arterial puncture, at a time when fluctuation of transcutaneous oxygen value had been less than 2 torr over 2 minutes. Inspired oxygen concentrations were constant during the procedure. Cannulae were flushed with heparinized saline. In order to remove flushing solution, 3-5 ml was first aspirated; then the sample was taken. Heparinized plastic syringes were used, 0.3-0.5 ml of blood being drawn up over 5-10 seconds. A drift check was performed on the electrode on completion of each study, and it was always within 5% of the original calibration value. Blood gaseswere immediately measured using a Radiometer ABL2 analyser. The analyser self calibrates hourly, and in addition was checked daily with reference solutionsQUALICHECK (Radiometer, Copenhagen). Paired values were not recorded if transcutaneous oxygen fluctuated by more than 5 torr as the sample was drawn, or in the minute afterwards. Pairs of values with Paoz greater than 110 mmHg were excluded from the group analyses to facilitate comparison of the two electrode temperatures over the most clinically relevant range, and because data above this level are not linearly related (4). Where two pairs of observations in an individual patient were similar or identical, one pair was excluded from group analysis to minimize the biasing effect of intrasubject correlation. No more than seven studies on any one individual were included in each analysis. Correlation coefficients, regression equations, and 95% confidence limits were calculated from the paired values of simultaneous tcPo, and Paoz using a linear regression model. Electrodes were not applied continuously for more than one hour, and the skin was inspected on completion of each study. Six patients had five or more sets of observations performed, over periods of days or months. For these we plotted the values obtained for each individual separatelywith a regression line. Numbers were too small for meaningful statistical analysis.
RESULTS Thirty-one observations in nine children at 44.O”C were analysed. The regression equation was tcPo,=O.9 Pao,-10 (r=0.92,95% confidence limits*13 torr). Twenty-eight observations in eight children at 44.5”C were analysed, and the regression equation was tcPo,=O.9 Pao,-8 (r=0.94, 95% confidence limits*13 torr; Fig. 1). The data points on individuals are shown in Fig. 2, with regression lines. The slope of this line varied between patients, but there was a close relationship at both electrode temperatures. No burns were seen where the electrodes were applied.
DISCUSSION In the children studied transcutaneous oxygen underestimated arterial oxygen values. This probably results from maturational changes in the skin with advancing postnatal age. The epidermal layer is thicker and the underlying capillaries less numerous (7). These changes may impair oxygen diffusion in response to local vasodilation, and may also increase oxygen consumption under the electrode-both leading to underestimation of the arterial value. A study of babies with bronchopulmonary dysplasia suggests that this discrepancy becomes apparent at 8-12 weeks postnatal age (8).
130
BR. J. DIS.
CHEST:
VOL.
81 NO.
loo-
2 /
n= 31
80-
/
60-
40/
2
k 1
20-
s
‘Z 5 5 0, 3 o 52 6 5 E
e b-
44°C
/
/ 0 loon=28
//
80-
60-
40-
/
/
20/
0
I 20
I 40 Arterial
44.5”C I 60 oxygen
I 80 tension
I 100
I 120
(torrl
Fig. 1. Relationship between tcPo, and Paoz in children with respiratory disease, using electrode temperatures of 44°C and 44S”C
At both electrode temperatures the 95% confidence limits were wide. This finding is apparent in previous studies though the confidence limits were not quoted, and inappropriate use of correlation coefficients gave rise to falsely optimistic claims regarding accuracy and reliability (2,4,5). This variability could be due to technical problems with the blood gas analyser or the transcutaneous monitor. We minimized errors associated with blood gas measurement by careful sampling, and by the use of a self-calibrating analyser which was checked daily with reference solutions, and is reproducible to within
LOFTUS,
BARRY,
12Or
80
j 6 .z a, 5 tT 2 2& 6 5 z e I-
PRICE:
TRANSCUTANEOUS
7 y.0 male Bronchopneumoma
I
120-
r
X’ X’
,*
/*
2 y.0. Asp
OXYGEN
MEASUREMENT
131
4 y.0 female Pertussis
, ./
female
1 y.0. female
F
pneumonia
B P.D.
80-
.’ /
.
40-
o-
’
120
’
’
I
I
44.5OC I
I
1 y.0. female
8 month
B.P D
Pneumonia
old
male
c
I 0
I 40
I
I 80 Arterial
I
44oc I 120 oxygen
J tenslon
(torr)
Fig. 2. Relationship between tcPo, and Pao, in six children with respiratory disease. Age, sex and diagnosis are indicated. For patient illustrated on the lower left values were recorded over a 4-month period-possibly accounting for the increased variation seen. Points marked X were excluded from group analysis 1.3 torr over the range studied. The transcutaneous monitor checks automatically for drift during the calibration procedure, ahd we confirmed that this drift was minimal. It is probable that the poor agreement between the methods reflects individual differences in skin characteristics. Factors which influence the relationship of &PO, to Pao,, such as the thickness of the epidermis, and the number of capillaries underlying the skin electrode, vary between patients. This variation could account for the wide scatter of data when groups of patients are analysed.
132
BR. J. DIS.
CHEST:
VOL.
81 NO.
2
Use of the higher electrode temperature increased the value of tcPo, for a given Pao, as would be expected (9). However, the accuracy of the method was not improved. We saw no burns after 1 hour, but one child who was continuously monitored for 4 hours at 44S”C, as part of clinical management developed a first degree burn. Because of the greater risk of burns at higher temperatures, we feel that 44°C is the optimal electrode temperature. Over short periods of time, in individual patients, the skin characteristics may exert a constant influence on the relationship between &PO, and Pao,. We found that trends in oxygenation were accurately reflected at both electrode temperatures, each patient’s line of regression presumably depending on individual skin characteristics. Transcutaneous oxygen monitoring may therefore be useful in the patient with rapidly fluctuating oxygen requirements, and in non-invasive assessment of oxygenation in response to treatment. Because of the limitations of the method and its poor performance in shocked patients (lo), it cannot replace the need for an indwelling arterial catheter in critically ill children. ACKNOWLEDGEMENTS The authors wish to thank Miss M. Rusbridge for preparing the illustrations, Morris for typing the manuscript.
and Miss A.
REFERENCES 1. Huch R, Lubbers DW, Huch A. Reliability of transcutaneous monitoring of arterial Po2 in newborn infants. Arch Dis Child 1974;49:213-18. 2. Yahav J, Mindorff C, Levison H. The validity of the transcutaneous oxygen tension method in children with cardiorespiratory problems. Am Rev Resp Dis 1981;124:586-7. 3. Yip WC, Tay JS, Wong HB, Ho TF. Reliability of transcutaneous oxygen monitoring of critically ill children in a general paediatric unit. Clin Puediat 1983;22:431-5. 4. Monaco F, Nickerson BG, McQuitty JC. Continuous transcutaneous oxygen and carbon dioxide monitoring in the paediatric I.C.U. Crit Care Med 1982;10:765-6. 5. Bompard Y, Beaufils F, Azancot A, Asensi D. Continuous transcutaneous Po2 monitoring in vital distress of children. Birth Defects 1979;XV (4):383-6. 6. Schonfield T, Sargent CW, Bautista D. et al. Transcutaneous oxygen monitoring during exercise stress testing. Am Rev Resp Dis 1980;121:457-62. 7. Huch R, Huch A, Lubbers DW. Theoretical basisof transcutaneous oxygen pressure measurements. In: Huch R, Huch A, Lubbers DW, eds. Transcutaneous Po2. New York: ThiemeStratton, 1981:2-70. 8. Rome ES, Stork EK, Carlo WA, Martin RJ. Limitations of transcutaneous Po2 and Pco2 monitoring in infants with bronchopulmonary dysplasia. Pediatrics 1984;74:217-70. 9. Evans NTS, Naylor PRD. The systemic oxygen supply to the surface of human skin. Resp Physiol 1967;3:21. 10. Versmold HT, Lundarkamp M, Noltzman M, Strockhacken I, Riegel KP. Limitations of transcutaneous oxygen monitoring in sick neonates. Relation to blood pressure, blood volume, peripheral blood flow and acid-base status. Acta Anaesth Stand 1978;68 (Suppl):88-90.
Date accepted 20 June 1986
TRANSCUTANEOUS OXYGEN MEASUREMENT IN PAEDIATRIC RESPIRATORY DISEASE B. G. LOFI’US, W. BARRY AND .I. F. PRICE Department of Child Health and Chest Unit, King’s College Hospital, London, SE5 8RX
Summary
We have measured simultaneous arterial (Pao,) and transcutaneous oxygen tension (fcPoJ in 13 children, aged 8 months to 10 years, with respiratory disease. A Radiometer TCM2 transcutaneous oxygen monitor was used. At an electrode temperature of 44°C the regression equation was tcPo,=O.9 Pao,-10 and at 44.5”C tcPo,=O.9 Pao,-8. The 95% confidence limits were& 13 torr at both temperatures. We also studied the usefulness of the method in monitoring trends in oxygenation in individual patients. In six children paired observations were made on five or more occasions and in these &PO, followed changes in Pao, accurately. We conclude that this technique is not sufficiently reliable for predicting arterial values, but in individual patients it accurately reflects trends in oxygenation. INTRODUCTION There is a close relationship between transcutaneous and arterial oxygen values in the newborn (1). Few studies have assessed the technique in older children (2-5) and opinion varies as to the reliability of the method. We have studied the relationship between transcutaneous and arterial oxygen levels in a group of children with respiratory disease, assessed the influence of electrode temperature on that relationship, and evaluated the usefulness of the method in individual patients. Patients and Methods Thirteen patients aged 8 months to 10 years were studied. Three had bronchopulmonary dysplasia, two asthma, two bronchopneumonia, two bronchitis, two aspiration pneumonia, one pertussis and one upper airway obstruction. All had a clinical indication for the measurement of blood gases. No patient was shocked, hypotensive or hypothermic at the time of the study. Five had indwelling arterial cannulae. Nine children aged 8 months to 8 years (mean aged 2.4 years) were studied using an electrode temperature of 44°C and eight children aged l-10 years (mean age 4.3 years) with an electrode temperature of 44S”C. Four of the children were studied at both temperatures. Correspondence to: Dr B. G. Loftus, Department of Thoracic Medicine, King’s College Hospital, London, SE5 8RX.
LOFTUS,
BARRY,
PRICE:
TRANSCUTANEOUS
OXYGEN
MEASUREMENT
129
A Radiometer TCM2 transcutaneous oxygen monitor was used. The type of oxygen electrode used detects changes in arterial oxygen tension within 60 seconds (6). An automatic single-point in vitro calibration was performed with air as the calibration gas, the electrode was then placed on the upper chest, on an area of skin which had been cleaned with alcohol, and gently abraded with gauze. The electrode was allowed to stabilize for at least 20 minutes. Arterial blood was drawn from an indwelling cannula, or by direct radial arterial puncture, at a time when fluctuation of transcutaneous oxygen value had been less than 2 torr over 2 minutes. Inspired oxygen concentrations were constant during the procedure. Cannulae were flushed with heparinized saline. In order to remove flushing solution, 3-5 ml was first aspirated; then the sample was taken. Heparinized plastic syringes were used, 0.3-0.5 ml of blood being drawn up over 5-10 seconds. A drift check was performed on the electrode on completion of each study, and it was always within 5% of the original calibration value. Blood gaseswere immediately measured using a Radiometer ABL2 analyser. The analyser self calibrates hourly, and in addition was checked daily with reference solutionsQUALICHECK (Radiometer, Copenhagen). Paired values were not recorded if transcutaneous oxygen fluctuated by more than 5 torr as the sample was drawn, or in the minute afterwards. Pairs of values with Paoz greater than 110 mmHg were excluded from the group analyses to facilitate comparison of the two electrode temperatures over the most clinically relevant range, and because data above this level are not linearly related (4). Where two pairs of observations in an individual patient were similar or identical, one pair was excluded from group analysis to minimize the biasing effect of intrasubject correlation. No more than seven studies on any one individual were included in each analysis. Correlation coefficients, regression equations, and 95% confidence limits were calculated from the paired values of simultaneous tcPo, and Paoz using a linear regression model. Electrodes were not applied continuously for more than one hour, and the skin was inspected on completion of each study. Six patients had five or more sets of observations performed, over periods of days or months. For these we plotted the values obtained for each individual separatelywith a regression line. Numbers were too small for meaningful statistical analysis.
RESULTS Thirty-one observations in nine children at 44.O”C were analysed. The regression equation was tcPo,=O.9 Pao,-10 (r=0.92,95% confidence limits*13 torr). Twenty-eight observations in eight children at 44.5”C were analysed, and the regression equation was tcPo,=O.9 Pao,-8 (r=0.94, 95% confidence limits*13 torr; Fig. 1). The data points on individuals are shown in Fig. 2, with regression lines. The slope of this line varied between patients, but there was a close relationship at both electrode temperatures. No burns were seen where the electrodes were applied.
DISCUSSION In the children studied transcutaneous oxygen underestimated arterial oxygen values. This probably results from maturational changes in the skin with advancing postnatal age. The epidermal layer is thicker and the underlying capillaries less numerous (7). These changes may impair oxygen diffusion in response to local vasodilation, and may also increase oxygen consumption under the electrode-both leading to underestimation of the arterial value. A study of babies with bronchopulmonary dysplasia suggests that this discrepancy becomes apparent at 8-12 weeks postnatal age (8).
130
BR. J. DIS.
CHEST:
VOL.
81 NO.
loo-
2 /
n= 31
80-
/
60-
40/
2
k 1
20-
s
‘Z 5 5 0, 3 o 52 6 5 E
e b-
44°C
/
/ 0 loon=28
//
80-
60-
40-
/
/
20/
0
I 20
I 40 Arterial
44.5”C I 60 oxygen
I 80 tension
I 100
I 120
(torrl
Fig. 1. Relationship between tcPo, and Paoz in children with respiratory disease, using electrode temperatures of 44°C and 44S”C
At both electrode temperatures the 95% confidence limits were wide. This finding is apparent in previous studies though the confidence limits were not quoted, and inappropriate use of correlation coefficients gave rise to falsely optimistic claims regarding accuracy and reliability (2,4,5). This variability could be due to technical problems with the blood gas analyser or the transcutaneous monitor. We minimized errors associated with blood gas measurement by careful sampling, and by the use of a self-calibrating analyser which was checked daily with reference solutions, and is reproducible to within
LOFTUS,
BARRY,
12Or
80
j 6 .z a, 5 tT 2 2& 6 5 z e I-
PRICE:
TRANSCUTANEOUS
7 y.0 male Bronchopneumoma
I
120-
r
X’ X’
,*
/*
2 y.0. Asp
OXYGEN
MEASUREMENT
131
4 y.0 female Pertussis
, ./
female
1 y.0. female
F
pneumonia
B P.D.
80-
.’ /
.
40-
o-
’
120
’
’
I
I
44.5OC I
I
1 y.0. female
8 month
B.P D
Pneumonia
old
male
c
I 0
I 40
I
I 80 Arterial
I
44oc I 120 oxygen
J tenslon
(torr)
Fig. 2. Relationship between tcPo, and Pao, in six children with respiratory disease. Age, sex and diagnosis are indicated. For patient illustrated on the lower left values were recorded over a 4-month period-possibly accounting for the increased variation seen. Points marked X were excluded from group analysis 1.3 torr over the range studied. The transcutaneous monitor checks automatically for drift during the calibration procedure, ahd we confirmed that this drift was minimal. It is probable that the poor agreement between the methods reflects individual differences in skin characteristics. Factors which influence the relationship of &PO, to Pao,, such as the thickness of the epidermis, and the number of capillaries underlying the skin electrode, vary between patients. This variation could account for the wide scatter of data when groups of patients are analysed.
132
BR. J. DIS.
CHEST:
VOL.
81 NO.
2
Use of the higher electrode temperature increased the value of tcPo, for a given Pao, as would be expected (9). However, the accuracy of the method was not improved. We saw no burns after 1 hour, but one child who was continuously monitored for 4 hours at 44S”C, as part of clinical management developed a first degree burn. Because of the greater risk of burns at higher temperatures, we feel that 44°C is the optimal electrode temperature. Over short periods of time, in individual patients, the skin characteristics may exert a constant influence on the relationship between &PO, and Pao,. We found that trends in oxygenation were accurately reflected at both electrode temperatures, each patient’s line of regression presumably depending on individual skin characteristics. Transcutaneous oxygen monitoring may therefore be useful in the patient with rapidly fluctuating oxygen requirements, and in non-invasive assessment of oxygenation in response to treatment. Because of the limitations of the method and its poor performance in shocked patients (lo), it cannot replace the need for an indwelling arterial catheter in critically ill children. ACKNOWLEDGEMENTS The authors wish to thank Miss M. Rusbridge for preparing the illustrations, Morris for typing the manuscript.
and Miss A.
REFERENCES 1. Huch R, Lubbers DW, Huch A. Reliability of transcutaneous monitoring of arterial Po2 in newborn infants. Arch Dis Child 1974;49:213-18. 2. Yahav J, Mindorff C, Levison H. The validity of the transcutaneous oxygen tension method in children with cardiorespiratory problems. Am Rev Resp Dis 1981;124:586-7. 3. Yip WC, Tay JS, Wong HB, Ho TF. Reliability of transcutaneous oxygen monitoring of critically ill children in a general paediatric unit. Clin Puediat 1983;22:431-5. 4. Monaco F, Nickerson BG, McQuitty JC. Continuous transcutaneous oxygen and carbon dioxide monitoring in the paediatric I.C.U. Crit Care Med 1982;10:765-6. 5. Bompard Y, Beaufils F, Azancot A, Asensi D. Continuous transcutaneous Po2 monitoring in vital distress of children. Birth Defects 1979;XV (4):383-6. 6. Schonfield T, Sargent CW, Bautista D. et al. Transcutaneous oxygen monitoring during exercise stress testing. Am Rev Resp Dis 1980;121:457-62. 7. Huch R, Huch A, Lubbers DW. Theoretical basisof transcutaneous oxygen pressure measurements. In: Huch R, Huch A, Lubbers DW, eds. Transcutaneous Po2. New York: ThiemeStratton, 1981:2-70. 8. Rome ES, Stork EK, Carlo WA, Martin RJ. Limitations of transcutaneous Po2 and Pco2 monitoring in infants with bronchopulmonary dysplasia. Pediatrics 1984;74:217-70. 9. Evans NTS, Naylor PRD. The systemic oxygen supply to the surface of human skin. Resp Physiol 1967;3:21. 10. Versmold HT, Lundarkamp M, Noltzman M, Strockhacken I, Riegel KP. Limitations of transcutaneous oxygen monitoring in sick neonates. Relation to blood pressure, blood volume, peripheral blood flow and acid-base status. Acta Anaesth Stand 1978;68 (Suppl):88-90.
Date accepted 20 June 1986