- Email: [email protected]
A NEW METHOD FOR CALCULATION OF VENTILATORY DEADSPACE
CORRESPONDENCE
639
has a fresh gas inlet as shown by the right-hand arrow) and a fresh gas flow rate of 4-6 litre min"1, the ventilator fails to operate. Patient ventilation ceases and, after 20 s, the " Low Press" alarm is activated. In the conditions described it is not possible to restart the ventilator. The ventilator is designed to permit spontaneous ventilation and will not initiate a respiratory cycle if the respirometer unit detects "breathing". When the flow transducer is placed as shown with the Boyle Mk III absorber (and only the Mk III) and the flow rate described, the "end of breath detector" function "sees" the fresh gas flow as a continuous expiration and, consequently, disables the ventilator. This does not occur with flow rates of less than 4 litre min"1, or if the transducer is placed at the patient end of the circuit as the manufacturer recommends. However, it is our experience that it is not uncommon for anaesthetists to use flows in excess of 4 litre min"1 when initiating ventilation following induction of anaesthesia. The foregoing combination of circumstances initially proved elusive to detect, occasionally embarrassing, and is clearly not without risk. The manufacturer now acknowledges the problem, although, as they correctly point out, the Mk III absorber is old and soon will be withdrawn from service. Nevertheless, a substantial number are currently in service. Ohmeda have agreed to alter the software to eliminate the "end of breath detection" facility, thus overcoming the problem. Finally, it may be appropriate to remind users of the OAV 7710 that, if the gasflowis inadequate or if disconnected from the anaesthetic gas outflow, the ventilator will continue to function but will entrain room air without warning to the operator, with a consequent risk of patient awareness and reduced inspired oxygen concentration. J. THORBURN W. O. M. DAVIS
Glasgow A NEW METHOD FOR CALCULATION OF VENTILATORY DEADSPACE Sir,—Deadspace measurements are becoming common in research and clinical work. The single breath test for carbon dioxide (SBT-CO2) permits "on-line" measurement, and allows detailed analysis of derangements in gas exchange. Our original method for calculating deadspace from SBT-CO2 [1] produced deadspace values identical to those obtainable by collection of expired gas in a Douglas bag. However, because of rebreathing in the apparatus deadspace [2], both methods overestimate alveloar and physiological deadspace. We present some revised deadspace values, recalculated in order to give a better measure of alveolar gas exchange. Our original study [1], based on SBT-CO2, was performed with the aid of a computer. When combined with arterial blood sampling, SBT-CO2 allows calculation of the physiological deadspace fraction FD ph > s /FT: _ * "CO,
(1)
The alveolar deadspace fraction Fn alv /FT alv is obtained from: [/D phys -FD aw FT-FD3W
where FDa* is the airway deadspace.
(2)
When rebreathing occurs in the limbs of a Y-piece, part of the expired carbon dioxide registered by the "in-line " analyser returns to the patient in the next inspiration [2]. (This rebreathing can be reduced by using one-way valves, which are, however, unacceptable in clinical practice.) Thus the volume of expired carbon dioxide (FT,-,,,) registered per breath is greater than the volume that is eliminated; a greater volume than would be collected in a Douglas bag. The rebreathed volume of carbon dioxide (Fco2reb) is equivalent to approximately 24 ml of end-tidal gas [2]. In our original deadspace estimates [1], mixed expired Fco., (FECOJ was estimated from (KTQJ.— Fco,reb)/tidal volume
and was substituted in equation (1). This gave the same values as would have been obtained using a Douglas bag. As in a conventional Douglas bag experiment, FlCOt was regarded as zero. This method is theoretically incorrect. Rebreathing increases Fico, changing the denominator in equation (1). More importantly, the apparatus deadspace associated with rebreathing is attributed to the patient. Consequently, FDphys and FD"'V are overestimated. The revised method of calculation is intended to avoid this problem by measuring the patient's gas exchange rather than that of patient plus apparatus; the new deadspace values are thus smaller. In the new method, /•!„,, in equation (1) is obtained from "registered VTf.On divided by F T " , that is, no allowance is made for Fco 2 reb . Fico,; previously regarded as zero, is estimated from " Fco 2 reb divided by alveolar tidal volume (FT-FD1™)".
We present here the new and old estimates for FDalv/FTolv and FDphys/FT (table I), and graphs (fig. 1) showing the new relationship between FDalv/FTalv and (Pa c0! -PE' C0a )/Pa CO2 . All the deadspace ratios are reduced, and the ratio (Par()i — PE'cnJ/Parlt: (where PEVO, is end-tidal Pco2) is now closer to FDalv/FTalv than previously. This relationship is of particular interest for the anaesthetist, as PaCOs — PE' C 0 2 may be regarded as a "poor man's" measure of alveolar deadspace. Which values should be referred to in future? When comparing our results with those of workers who have collected mixed expired gas, the old method of calculation may be preferred. However, the revised values for FDa'v/FTalv are more suitable as "reference values", representing a fairer measure of alveolar gas exchange under ideal conditions. The theoretical minimum value for Fn alv /FT alv in healthy patients TABLE I. Old and new deadspace ratios calculated from the 58 patients in the study of Fletcher and Jonson [1], The ventilator settings were: mean tidal volume 450 ml; frequency 17 b.p.m.; mean V r 743 ml at a frequency of 9 b.p.m. FD»»> S /FT
Small FT Median 5th percentile 95th percentile Large FT Median 5th percentile 95th percentile
F D O H 7V T
a
"
Old
New
Old
New
0.44 0.32 0.58
0.42 0.28 0.55
0.29 0.17 0.49
0.24 0.10 0.43
0.31 0.22 0.49
0.29 0.20 0.47
0.22 0.13 0.43
0.17 0.10 0.38
BRITISH JOURNAL OF ANAESTHESIA
640 0.6 "Small tidal volumes
/ /
/
m
ipLarge tidal volumes
/
/ 7 ''
0.5 0.4 "5 >
>** 0.3 0.2 0.1
'A / /
i
0.0
0.1 (Pa,'CO2
i
0.2
0.3
0.0
0.4
0.1
0.2
0.3
0.4
FIG. 1. Revised version of figure 8 in Fletcher and Jonson [1], showing the relationship between Fbalv/KTalv and (PaCOj — .PE'COj)/PaCOi at the two ventilator settings (see table I). The thick line represents the line of identity. Left: Small tidal volumes. Regression line and 95 % confidence intervals are shown. Zero PaCO2-PE'CO!, had it been achieved by any patient, is equivalent to a KDalv/KT"lv of about 0.06 (previously 0.13). The gradient of the regression line is 1.19 (previously 1.14);r = 0.94(0.95). Right: Large tidal volumes. There is a non-normal distribution. Several patients have small negative Paco> — PE' C 0 ! differences.
is approximately 0.03 [3]; children almost achieve this value [3, 4]. Use of single breath tests to analyse deadspace has clear advantages over the Douglas bag method. KDphys can be broken down into its component parts, and phase III slope, which gives information about the nature of pathological gas elimination, can be measured. As more workers switch to this methodology, it is important that we state clearly what we are measuring! R. FLETCHER B. JONSON
Lund
REFERENCES 1. Fletcher R, Jonson B. Deadspace and the single breath test for carbon dioxide during anaesthesia and controlled ventilation. British Journal of Anaesthesia 1984; 56: 109-119. 2. Fletcher R, Werner O, Nordstrom L, Jonson B. Sources of error and corrections in measurement of carbon dioxide elimination with the Siemens-Elema CO2 Analyser. British Journal of Anaesthesia 1983; 55: 177-185. 3. Fletcher R. Relationship between alveolar deadspace and arterial oxygenation in children with congenital cardiac disease. British Journal of Anaesthesia 1989;62: 168-176. 4. Fletcher R, Niklasson L, Drefeldt B. New and old methods for calculation of ventilatory deadspace. Anesthesia and Analgesia 1989; 68: 420-421.
639
has a fresh gas inlet as shown by the right-hand arrow) and a fresh gas flow rate of 4-6 litre min"1, the ventilator fails to operate. Patient ventilation ceases and, after 20 s, the " Low Press" alarm is activated. In the conditions described it is not possible to restart the ventilator. The ventilator is designed to permit spontaneous ventilation and will not initiate a respiratory cycle if the respirometer unit detects "breathing". When the flow transducer is placed as shown with the Boyle Mk III absorber (and only the Mk III) and the flow rate described, the "end of breath detector" function "sees" the fresh gas flow as a continuous expiration and, consequently, disables the ventilator. This does not occur with flow rates of less than 4 litre min"1, or if the transducer is placed at the patient end of the circuit as the manufacturer recommends. However, it is our experience that it is not uncommon for anaesthetists to use flows in excess of 4 litre min"1 when initiating ventilation following induction of anaesthesia. The foregoing combination of circumstances initially proved elusive to detect, occasionally embarrassing, and is clearly not without risk. The manufacturer now acknowledges the problem, although, as they correctly point out, the Mk III absorber is old and soon will be withdrawn from service. Nevertheless, a substantial number are currently in service. Ohmeda have agreed to alter the software to eliminate the "end of breath detection" facility, thus overcoming the problem. Finally, it may be appropriate to remind users of the OAV 7710 that, if the gasflowis inadequate or if disconnected from the anaesthetic gas outflow, the ventilator will continue to function but will entrain room air without warning to the operator, with a consequent risk of patient awareness and reduced inspired oxygen concentration. J. THORBURN W. O. M. DAVIS
Glasgow A NEW METHOD FOR CALCULATION OF VENTILATORY DEADSPACE Sir,—Deadspace measurements are becoming common in research and clinical work. The single breath test for carbon dioxide (SBT-CO2) permits "on-line" measurement, and allows detailed analysis of derangements in gas exchange. Our original method for calculating deadspace from SBT-CO2 [1] produced deadspace values identical to those obtainable by collection of expired gas in a Douglas bag. However, because of rebreathing in the apparatus deadspace [2], both methods overestimate alveloar and physiological deadspace. We present some revised deadspace values, recalculated in order to give a better measure of alveolar gas exchange. Our original study [1], based on SBT-CO2, was performed with the aid of a computer. When combined with arterial blood sampling, SBT-CO2 allows calculation of the physiological deadspace fraction FD ph > s /FT: _ * "CO,
(1)
The alveolar deadspace fraction Fn alv /FT alv is obtained from: [/D phys -FD aw FT-FD3W
where FDa* is the airway deadspace.
(2)
When rebreathing occurs in the limbs of a Y-piece, part of the expired carbon dioxide registered by the "in-line " analyser returns to the patient in the next inspiration [2]. (This rebreathing can be reduced by using one-way valves, which are, however, unacceptable in clinical practice.) Thus the volume of expired carbon dioxide (FT,-,,,) registered per breath is greater than the volume that is eliminated; a greater volume than would be collected in a Douglas bag. The rebreathed volume of carbon dioxide (Fco2reb) is equivalent to approximately 24 ml of end-tidal gas [2]. In our original deadspace estimates [1], mixed expired Fco., (FECOJ was estimated from (KTQJ.— Fco,reb)/tidal volume
and was substituted in equation (1). This gave the same values as would have been obtained using a Douglas bag. As in a conventional Douglas bag experiment, FlCOt was regarded as zero. This method is theoretically incorrect. Rebreathing increases Fico, changing the denominator in equation (1). More importantly, the apparatus deadspace associated with rebreathing is attributed to the patient. Consequently, FDphys and FD"'V are overestimated. The revised method of calculation is intended to avoid this problem by measuring the patient's gas exchange rather than that of patient plus apparatus; the new deadspace values are thus smaller. In the new method, /•!„,, in equation (1) is obtained from "registered VTf.On divided by F T " , that is, no allowance is made for Fco 2 reb . Fico,; previously regarded as zero, is estimated from " Fco 2 reb divided by alveolar tidal volume (FT-FD1™)".
We present here the new and old estimates for FDalv/FTolv and FDphys/FT (table I), and graphs (fig. 1) showing the new relationship between FDalv/FTalv and (Pa c0! -PE' C0a )/Pa CO2 . All the deadspace ratios are reduced, and the ratio (Par()i — PE'cnJ/Parlt: (where PEVO, is end-tidal Pco2) is now closer to FDalv/FTalv than previously. This relationship is of particular interest for the anaesthetist, as PaCOs — PE' C 0 2 may be regarded as a "poor man's" measure of alveolar deadspace. Which values should be referred to in future? When comparing our results with those of workers who have collected mixed expired gas, the old method of calculation may be preferred. However, the revised values for FDa'v/FTalv are more suitable as "reference values", representing a fairer measure of alveolar gas exchange under ideal conditions. The theoretical minimum value for Fn alv /FT alv in healthy patients TABLE I. Old and new deadspace ratios calculated from the 58 patients in the study of Fletcher and Jonson [1], The ventilator settings were: mean tidal volume 450 ml; frequency 17 b.p.m.; mean V r 743 ml at a frequency of 9 b.p.m. FD»»> S /FT
Small FT Median 5th percentile 95th percentile Large FT Median 5th percentile 95th percentile
F D O H 7V T
a
"
Old
New
Old
New
0.44 0.32 0.58
0.42 0.28 0.55
0.29 0.17 0.49
0.24 0.10 0.43
0.31 0.22 0.49
0.29 0.20 0.47
0.22 0.13 0.43
0.17 0.10 0.38
BRITISH JOURNAL OF ANAESTHESIA
640 0.6 "Small tidal volumes
/ /
/
m
ipLarge tidal volumes
/
/ 7 ''
0.5 0.4 "5 >
>** 0.3 0.2 0.1
'A / /
i
0.0
0.1 (Pa,'CO2
i
0.2
0.3
0.0
0.4
0.1
0.2
0.3
0.4
FIG. 1. Revised version of figure 8 in Fletcher and Jonson [1], showing the relationship between Fbalv/KTalv and (PaCOj — .PE'COj)/PaCOi at the two ventilator settings (see table I). The thick line represents the line of identity. Left: Small tidal volumes. Regression line and 95 % confidence intervals are shown. Zero PaCO2-PE'CO!, had it been achieved by any patient, is equivalent to a KDalv/KT"lv of about 0.06 (previously 0.13). The gradient of the regression line is 1.19 (previously 1.14);r = 0.94(0.95). Right: Large tidal volumes. There is a non-normal distribution. Several patients have small negative Paco> — PE' C 0 ! differences.
is approximately 0.03 [3]; children almost achieve this value [3, 4]. Use of single breath tests to analyse deadspace has clear advantages over the Douglas bag method. KDphys can be broken down into its component parts, and phase III slope, which gives information about the nature of pathological gas elimination, can be measured. As more workers switch to this methodology, it is important that we state clearly what we are measuring! R. FLETCHER B. JONSON
Lund
REFERENCES 1. Fletcher R, Jonson B. Deadspace and the single breath test for carbon dioxide during anaesthesia and controlled ventilation. British Journal of Anaesthesia 1984; 56: 109-119. 2. Fletcher R, Werner O, Nordstrom L, Jonson B. Sources of error and corrections in measurement of carbon dioxide elimination with the Siemens-Elema CO2 Analyser. British Journal of Anaesthesia 1983; 55: 177-185. 3. Fletcher R. Relationship between alveolar deadspace and arterial oxygenation in children with congenital cardiac disease. British Journal of Anaesthesia 1989;62: 168-176. 4. Fletcher R, Niklasson L, Drefeldt B. New and old methods for calculation of ventilatory deadspace. Anesthesia and Analgesia 1989; 68: 420-421.