- Email: [email protected]
Calibration of respiratory gas exchange measurements in inhalation toxicology studies
FUNDAMENTAL
AND
APPLIED
TOXICOLOGY
18,
144-148 ( 1992)
Calibration of Respiratory Gas Exchange Measurements in Inhalation Toxicology Studies WILLIAM J. MAUTZ Air Pollution Health Eflects Laboratory. Department of Community and Environmental Medicine, University of California, Irvine, California 92717
Received April 9, 1991;accepted July 18,1991
Calibration of Respiratory Gas Exchange Measurements in Inhalation Toxicology Studies. MAUTZ, W. J. (1992). Fundum. Appl. Toxicol. 18, 144-148.
The use of simplifying assumptions for determining respiratory gas exchange is associated with substantial errors in the estimates of oxygen consumption (vo’,,) and carbon dioxide production (vco2). Studies were done to estimate the magnitude of these errors under common exposure conditions, and a simple procedure that reveals these errors is described for calibrating an open flow respirometer. The errors associated with various simplifying assumptions ranged from -1 to -21% for voioz,0.1 to 15% for v,-oZ, and 4 to 45% for the respiratory exchange ratio (R). The calibration was performed with a standard calibration gas elevated in CO, and depressed in O2 relative to air and bled into the respirometer at a measured flow rate. Dilution of the gas into the respirometer airstream simulates the effect of respiratory gas exchange, and expected values of vob2,fco2, and R compared to measured values provide a check of accuracy of flow and gas fraction measurements, a test for system leaks, and a test for the effects of any simplifying assumptions on the calculation of respiratory gas exchange. o 1992 Society of Toxicology. Currently there is substantial interest in measuring metabolic gas exchange of mammals in toxicology studies. Such measurements are made in studies that include use of exercise during exposure (Mautz et al., 1985a, 1988; Malek and Alarie, 1989), measurements of deposition of inhaled compounds in the respiratory tract (Wiester et al., 1987) and investigations of ventilatory and metabolic rate responses to inhaled toxic compounds (Silver et al., 198 1; Leith, 1984; Mautz and Bufalino, 1989). Metabolic gas exchange is important as a measure of work load in exercise studies and as a potentially affected biological variable in response to toxicant exposures. Metabolic gas exchange is usually measured with an open flow respirometry system in which subjects are exposed in a chamber to a test compound in flowing air (Fig. 1). Flow rate through the system is measured either upstream or downstream of the subject. Volume fractional concentrations of metabolic gases, O2 and CO2 , are measured upstream and downstream and, with flow rate measurements, are used to 0272-0590/92 $3.00 Copyright All rights
0 1992 by the Society of Toxicology. of reproduction in any fom reserved.
calculate O2 uptake (vo2) and COZ production ( ~coZ). In some recent studies, the use of simplifying assumptions in equations for estimating metabolic gas exchange resulted in substantial systematic errors in the estimates. The purpose of this study was to briefly review the problem, make estimates of the magnitude of errors in such assumptions using gas exchange of a rat, and present a simple calibration procedure for eliminating these errors. An accurate determination of vo2 and pcoZ in an open flow respirometer requires that the effect of entry and exit of all gases be taken into account. Entering or exiting gases change both the fractional concentrations of metabolic gases and the airflows into and out of the animal chamber (Withers, 1977; Mautz et al., 1985b). For example, the production or consumption of each respiratory gas affects the downstream fractional content of the other respiratory gas. Thus, the downstream fractional O2 content is reduced both by O2 consumed by the subject and also by the dilution effect of CO2 production, while downstream fractional CO1 content is increased both by subject CO2 production and also by removal of OZ. Changes in water vapor due to evaporation from the subject will also alter the fractions of other gases and airflow. Flow rates and metabolic gas fractions can be measured from air scrubbed of water vapor or measurements can be numerically corrected for measured water vapor content. Using measurements of respirometer temperature and pressure, values of vo2 and I&oz are corrected to standard temperature (O’C), pressure (760 mm Hg), and dry (STPD) volume rates of change. Equations that correct measured O2 and CO2 fractions for the reciprocal effects of exchange of each gas on the downstream fraction of the other gas have been derived for a variety of specific gas analyzers and sampling arrangements (Depocas and Hart, 1957; Consolazio et al., 1963; Tucker, 1968; Hill, 1972; Lister et al., 1977; Withers, 1977; Fedak et al., 198 1; Mautz et al., 1985b). A general set of equations defining total flow rates in and out of a respirometer and component volume flow rates of respiratory gases was given by Mautz et al. (1985b). These equations can be solved simultaneously for specific sampling arrangements to yield expressions for V,, and pco2 in terms of a minimum number
144
RESPIRATORY
GAS EXCHANGE
UPSTREAM
of variables that account for the variety of effects on respiratory gas fractions. For a typical case in which flow downstream of the respirometer chamber is measured, water vapor is absorbed before respiratory gases are analyzed and CO2 is not absorbed; the relationships are 3
02
=
Fl,,(
v
1 -
FEco2)
ho2(1
-
=
vE
GAS (FIORD
TEMPERATURE
SAMPLES .
5c02
PRESSURE )
I
INHALATION
ANIMAL
EXPOSURE
EXPOSURE
MIXTURE
CHAMBER
FLOWMETER
A
&,,) (1)
E 1 -
vco2
-
145
MEASUREMENT
Fkoz(l
-
FIo2
F’oz) 1 -
-
%,z
&)*
%.I,( -
FI,,,
PUMP
1 -
FEoz) ’
t DOWNSTREAM
(2)
TEMPERATURE
GAS ‘FEO2’ WATER
CALIBRATION
where the symbols follow previously established conventions (Depocas and Hart, 1957; Hill, 1972; Withers, 1977): PO2 = pco2 = vE = FIo2 = FEoz= %02 = FECOZ = R=
rate of O2 consumption of the subject (STPD) rate of CO2 production of the subject (STPD) airflow out of the respirometer (STPD) dry volume fraction of O2 in inlet air dry volume fraction of O2 in outlet air dry volume fraction of CO2 in inlet air dry volume fraction of CO, in outlet air respiratory exchange ratio, ljco2/ po2.
STPD corrections are made for the measurement of flow rate out of the respirometer using measurements of temperature, water vapor content, and pressure,
where VEATP”= airflow out of the respirometer at ambient temperature, pressure, and humidity PB = barometric pressure of outflow airstream (mm Hg) PHzo= partial pressure of water vapor in outflow a&ream (mm Hg) TA = ambient temperature of outflow airstream (OK). Expressions for gas exchange that use simplifying assumptions about inputs and outputs of all gasesto the system and their effects on the fractional content of other gases will introduce errors that can be substantial. Wiester et al. (1987) estimated gas exchange of rats in flowing air by measuring upstream flow, estimating downstream flow by correcting for water vapor addition by the rat but not for the unequal exchange of O2 and CO*, and by assuming inlet air FIco2= 0.0. Malek and Alarie (1989) measured gas exchange of guinea pigs exercising in a chamber. The downstream airflow was split: one stream was chemically scrubbed of water vapor prior to determination of Ftio2, the other was scrubbed of water vapor and COZ prior to determination of FEo2, and the streams were rehydrated in wet test meters for flow measurements. The CO* content of upstream air was also assumed to be 0.0 in the calculation of vcoZ. fo2 was calculated
’
SAMPLES FECOp’ VAPOR)
FLOWMETER
GAS
FIG. 1. Open flow respirometer components with apparatus for adding calibration gas. Arrows show direction of a&low. Lines without amows show other sensor sample points. Respiratory gas fractions (Fo2 and Fcoz) are water vapor-free fractions.
as the product of C02-free downstream flow and the difference between upstream and downstream Fo2, also CO*-free. However, when CO2 is scrubbed before downstream flow rate and Fo2 measurements are made, correction for the difference between water vapor-free and COz-free airflow into the animal chamber and water vapor-free and COz-free airflow out of the animal chamber (the difference due to oxygen consumed) is required. This correction can be made with Eq. (4) as shown by Depocas and Hart (1957), Hill (1972), and Withers (1977), ., Vi,, - FL,) ‘02 = ‘E (1 - F;,,) ’ where, following the convention
(4)
of Withers (1977), p and
F’ are respectively CO*-free flow rate and C02-free gas frac-
tions. The absolute errors of using simplifying assumptions on estimates of vo2 and li co2 in the above studies cannot be estimated from the information available; however, such assumptions both overestimate vco2 and underestimate vo2. This results in large respiratory exchange ratios (R = vco2/ Vo2) ranging from 0.89 to 1.28 (Malek and Alarie, 1989) and 1.16 to 1.35 (Wiester et al., 1987). The R expected in steady-state conditions ranges from 0.7 to 1.O depending on the nutrient sources for metabolic energy production. While transient large increases or decreases in R can occur as a result of changes in the dissolved C02-bicarbonate pool in tissue fluids, or from metabolic conversions between fats and carbohydrates, laboratory measurements of R from rodents during rest or exercise fall within the expected 0.7- 1.O range (Mautz et al., 1985a,b; Mautz and Bufalino, 1989; Brooks and White, 1978). Respirometers are typically calibrated using standard calibration gases to set the baseline and span adjustments of
146
WILLIAM
J. MAUTZ
TABLE 1 Measured and Expected Gas Exchange Values and Error Difference Obtained during Calibration of an Open Flow Respirometer” Expected gas exchange Respirometer flow rate 3000
2000 1000
vo2
tic02
19.93 12.51 8.32 4.09 12.93 8.37 4.09 7.69 4.07
11.26 1.07 4.70 2.31 7.31 4.73 2.31 4.35 2.30
Measured gas exchange and percentage error R
0.565 0.565 0.565 0.565 0.565 0.565 0.565 0.566 0.565
vco2
vo2
19.89 12.45 8.35 4.03 12.92 8.15 4.05 7.82 3.98
-0.2 -0.5 +0.4 -1.5 -0.1 -2.6 -1.0 +1.7 -2.2
11.04 6.94 4.62 2.31 7.14 4.70 2.29 4.43 2.31
-2.0 -1.8 -1.7 0.0 -2.3 -0.6 -0.9 +I.8 +0.4
R
0.555 0.557 0.553 0.573 0.553 0.557 0.565 0.566 0.580
-1.8 -1.4 -2.0 fl.4 -2.2 -2.1 0.0 0.0 +2.1
u Data are means of three successivemeasurements. Flow rates and gas exchange rates are ml/min.
gas analyzers. With a simple additional calibration step described below, an open flow respirometry system can be checked for leaks, calibrated for the specific range of expected values of poZ and vco2, and evaluated for sensitivity to simplifying assumptions in applying equations for determination of metabolic rate. METHODS A typical open flow respirometry system (Fig. 1) was used in which room air or an experimental exposure mixture passed through an animal holding chamber and then through a rotameter flowmeter (Matheson, Horsham, PA) calibrated against a soap film flowmeter (Bubble Meter, SKC West, Fullerton, CA) for the measurement of total downstream flow rate. Temperature was measured at the flowmeter with a thermocouple probe (Sensortek, Clifton NJ) calibrated against a National Bureau ofStandards traceable mercury thermometer. Water vapor content of the downstream flow was measured with a dew point sensor (EG & G Model 9 11, Waltham MA) and used with a measurement of barometric pressure to correct respirometer flow to STPD conditions by Eq. (3) above. Gas samples were drawn upstream of the chamber for measurement of Fb2 and Fb, and downstream of the respirometer flowmeter for measurement of Fb2 and FEEm using a mass spectrometer (Perkin-Elmer, Model 1100, Pomona, CA) calibrated with gravimetric standard gases (Liquid Carbonics, Los Angeles, CA). The mass spectrometer did not sensewater vapor and thus measured dry gas fractions. Eqs. ( I) and (2) were then used to estimate PO2and vco2. The calibration test of the integrated performance of the respirometer was made by introducing a measured constant flow ( ri,,,) of a standard calibration gas into the animal chamber (Fig. 1). The calibration gas was, in comparison to air, elevated in COz and depressed in O2 content (4.97% C02, 13.01% 02, 0.95% Ar, and 81.07% Nz) and was metered into the respirometer through a soap film flowmeter. Because the flowmeter was lined with liquid film solution which adds water vapor to dry air, the calibration gas was first saturated with water vapor by bubbling through a series of two midget impingers (SKC West, Fullerton, CA) tilled with distilled water. Temperature was measured to correct measured flow rate to STPD. The bleeding of calibration gas into the empty animal chamber affected the downstream gas fractions in a fashion equivalent to the respiratory gas exchange of an animal, and this simulated respiratory gas exchange was determined only by gas fractions of the calibration gas, calibration gas input flow rate, and ambient air-gas fractions.
The expected values of PO’,,and pco2 when adding the calibration gas were calculated using Eqs. 1 and 2 and substituting calibration gas flow, pc’,,,, (corrected to STPD) for I’r and calibration gas fractions FCAL02 and F cALcoIfor FEo2 and FEcoI, respectively. <,,, and F,, were the upstream air fractions. Expected values of PO2and Vco, thus obtained were compared to values derived from the downstream respirometer flowmeter and gas analysis aRer the calibration gas stream was diluted and mixed into the respirometer airstream at the empty animal chamber. The respirometer instruments and calculations were expected to yield respiratory gas exchange the same as the expected values derived from standardized gas fractions and calibration gas flow rate. Leaks in the respirometer system or errors in calibration of the respirometer flowmeter and in the STPD correction factor (Eq. 3) would introduce proportional errors in both lioz and tic,,. Errors in the calibration of the respiratory gas analyzer or use of simplifying assumptions in instrument configurations or respirometry calculations would introduce nonproportional errors (Withers, 1977) and differently affect vol and pco2 and thereby affect R. Calibrations were performed at several respirometer air flow rates and calibration gas flow rates. After calibration tests were performed, the gas exchange of a 260-g Sprague-Dawley rat was measured. These data were then used as an example to numerically examine the errors introduced by using simplifying assumptions in respiratory gas exchange calculations.
RESULTS The results of calibrating the respirometer are shown in Table 1 for three different respirometer flow rates and several calibration gas flow rates simulating gas exchange rates typical of small mammals. Percentage errors in measured po, and pcoZ compared to values expected from calibration gas flow depended on the accuracy of flow measurements and gas fraction measurements. Percentage errors in R, however, depended only on the accuracy of gas fraction measurements because flow parameters in the expression of R as the ratio of Eq. (2) to Eq. (1) cancel. Gas exchange of a 260-g Sprague-Dawley rat in 3000 ml/ min downstream flow at 22°C and 69% relative humidity was I?& = 7.82 ml/min, J&o2 = 6.41 ml/min, and R = 0.8 19. The effects of making certain simplifying assumptions on the calculation of this gas exchange are shown for a set
RESPIRATORY
GAS EXCHANGE
of sample parameters in Table 2. All of these assumptions tended to underestimate vo2 and overestimate vco2 resulting in large errors in R. The assumption that Fi,, = 0.0 produced an error in pco2 that was strongly dependent on the flow rate of air through the respirometer, and expressing to2 as the product of CC&free flow rate and COz-free oxygen fraction difference produced a large error in vo2. The error of neglecting unequal exchange of O2 and COZ in estimating downstream flow from measured upstream flow and water vapor addition was relatively small, and would approach 0 as R approached 1.O. DISCUSSION
147
MEASUREMENT
TABLE 2 Sensitivity of Respirometry Calculations to Simplifying Assumptions and Respirometer Flow Rate Percentage error Respirometer flow rate (ml min-‘)
vo2
PC’,,,
R
Case 1: Assume Fko2 = 0.0 1000
2000 3000 Case 2: Neglect the effect of unequal 0, and CO, exchange estimating r’, from i;. 1000 2000 3000 Case 3: Scrub CO2 prior to measurement of ilow and Fez. Assume Fb2 = 0.0. Use Vo2 = p! (F;, - F’&) instead of Eq. (4) 1000 2000 3000
-1.1 -2.2 -3.2
+9.9 t15
+6.0 +12 i-19
-3.6
+0.2
+3.9
+5.0
-3.7 +3.9 Determination of respiratory gas exchange in an open flow +0.09 -3.7 to.06 +4.0 respirometer as the product of air flow and upstream-downstream difference in fractional gas content requires, at minimum, a single flow measurement, either upstream or downstream of the respirometer chamber, and measurement of upstream and downstream gas fractions. The effect of each -21 +5.0 f33 -21 t10 +39 gas produced or consumed (0,) CO*, and H20) on the flow -21 t15 +45 rate and on the fractional content of the other gases needs to be accounted either by correcting flows and gas fractions for changes due to other gases (Eqs. 1, 2, and 3 for effects of C02, 02, and H20, respectively) or by chemical scrubbing tiplied by flow rate to yield estimates of voZ and vco2 (Eqs. of other gases from the airstream (Eq. 4 for the effects of 1 and 2), the resulting errors associated with these gas exCOz and see also Depocas and Hart, 1957; Hill, 1972; Withchange estimates are proportional to flow rate. In addition, ers, 1977; and Mautz et al., 1985b). it is more cumbersome to scrub CO* and H20 from the The calibration of an open flow respirometer with a stanairflow before measurement of flow and Fo2 if Eq. (4) is to dard gas mixture containing more COz and Iess O2 than air be applied to determine vol. COZ and Hz0 could instead is an effective method both for checking the combined acbe scrubbed only from the much smaller gas sample stream curacy of instruments for measuring gas exchange over the for O2 analysis rather than the entire respirometer airflow, range of expected values, testing for leaks in the system, and but then CO,-free and HzO-free respirometer flow rates must for testing the effects of any simplifying assumptions used be determined with separate measures of COZ and water vain the measurement. The technique is similar to that of Fedak por content. et al. (I 98 1) in which N2 was bled at a measured rate into a It is usually preferable in inhalation exposure studies to respirometer and its effect on FEozwas used to calibrate an measure flow downstream of the respirometer chamber and O2 analyzer for measurement of vo2. The advantage of the not impede the exposure mixture with an upstream flowpresent technique is that it includes simultaneous calibration meter, particularly if an aerosol exposure is involved. Reof the system for vcoZ and R as well as vo2. In addition, the spiratory gas exchange equations (Eq. 1 and 2) were thus sources of systematic errors in flow measurement vs gas fracpresented for a downstream flow measurement, but equation measurement can be qualitatively discriminated because tions for upstream flow measurement can be similarly dethe expected value of R is dependent on Fo2 and Fco2 but rived from the basic set of relations for volume flow rates of not on air flow through the respirometer, while voZ and pco, component respiratory gases (Mautz et al., 1985b) to yield will reflect errors in both flow and gas fraction measurements. Inhalation exposure studies present some special problems for incorporating gas exchange measurements. High flow rates may be required to deliver well-controlled concentrations of a toxicant to the breathing zone of a subject. High pco2 = PI %02(’ - FIOJ - Fko2(1 - F&x) (6) dilution air volumes make it more difficult to resolve changes ’ -F&3*-F&m in respiratory gas fractions due to metabolic gas exchange, and also increase the error associated with assuming E;,,, = where 1’1 = air-how into the respirometer (STPD). 0.0 (Table 2, Case 1). Because the values and associated errors If upstream airflow is measured, then all air passing of expressions containing respiratory gas fractions are multhrough the flowmeter must also pass through the animal
148
WILLIAM
exposure chamber, and any samples of air for measurement of upstream gas fractions or water vapor must be drawn upstream of the flowmeter. The correction of flow rate to STPD conditions (Eq. 3) requires estimation of water vapor and temperature of air in the flowmeter. Thus, for any airflow measurement requiring correction to STPD conditions, the measurements of temperature and water vapor should be made close to the flowmeter. This will minimize errors due to temperature differences between the llowmeter and humidity-temperature sensors and to possible non-steady-state loss or gain of water vapor in the air flow path between these sensors. It is important to make measurements of upstream respiratory gas fractions rather than to assume their values. Upstream gas fractions may vary due to the effects of air purification of inhalation exposure systems and to experimental addition of toxicant compounds in carrier gases such as O2 or Nz . Introduction of systematic errors in respiratory gas exchange determinations due to such carrier gases is particularly important because the error will not occur in purified air controls and will appear as an apparent treatment effect of the toxicant compound. Sampling upstream and downstream gas fractions in succession for each calculation of respiratory gas exchange will also militate against errors introduced by gas analyzer baseline drift. Because gas analyzers are typically calibrated relative to upstream air, the gas fraction upstream-downstream difference is largely independent of the upstream gas fraction (Withers, 1977); however, corrections for the effect of changes in one gas on fractional content of another (Eq. 1, 2, 4, 5, and 6) do depend on upstream values. Thus, gas exchange measurements are not wholly independent of baseline drift, but the errors are small. Under the conditions of measurement of rat gas exchange in the present study, a baseline drift in Fo2 of 0.00 1 fractional unit would produce an error of only 0.02% in Vo2. Finally, this calibration technique, like any other, is subject to the errors in the measurements that compose it. The principal sources of error are in measurement of the calibration gas flow into the respirometer and measurement of upstream gas fractions. Flow measurement using a soap film flowmeter depends on stopwatch timing, measurement of room temperature and pressure for STPD correction, and ensuring the bubble meter and respirometer are at room pressure or that the respirometer system pressure is included in the STPD correction. Errors in measuring upstream gas fractions and the analytical accuracy of the standard calibration gas affect the calibration gas bleed determination of the expected values of gas exchange. However, if Fro, is measured to within f 0.0001 fractional unit and if calibration gas fractions are
J. MAUTZ
similarly accurate, the error uncertainty of PO2derived from the gravimetric standard gas mixture used in this study is confined to +-0.2%. ACKNOWLEDGMENTS This work was supported by the California Air Resources Board (AO129-33, A833-104) and the Electric Power Research Institute (RP 1962-1). I thank C. Bufalino and D. Daniels for technical assistanceand M. Kleinman and R. Phalen for reviewing the manuscript.
REFERENCES Brooks, G. A., and White, T. P. (1978). Determination of metabolic and heart rate responses of rats to treadmill exercise. J. Appl. Physiol. 45, 1009-1015. Consolazio, C. F., Johnson, R. E., and Pecora, L. J. (1963). Physiological Measurements ofMetabolic Functions in Man. McGraw-Hill. New York. Depocas, F., and Hart, J. S. (1957). Use of the Pauling oxygen analyzer for measurement of oxygen consumption of animals in open-circuit systems and in a short-lag, closed-circuit apparatus. J. Appl. Physiol. 10,388-392. Fedak, M. A., Rome, L., and Seeherman, H. (1981). One-step N,-dilution technique for calibrating open-circuit voz measuring systems. J. Appl. Physiol.
51, 772-776.
Hill, R. W. (1972). Determination of oxygen consumption by use of the paramagnetic oxygen analyzer. J. Appl. Physiol. 33,261-263. Leith, D. E. (1984). Mass transport in mammalian lungs: Comparative physiology. J. Toxicol. Environ. Health 13, 25 1-27 1. Lister, G., Jr., Hoffman, I. E., and Rudolph, A. M. (1977). Measurement of oxygen consumption: Assessing the accuracy of a method. J. Appl. Physiol. 43, 9 16-9 17. Malek, D. E., and Alarie, Y. (1989). Ergometer within a whole-body plethysmograph to evaluate performance of guinea pigs under toxic atmospheres. Toxicol. Appl. Pharmacol. 101, 340-355. Mautz, W. J., and Bufalino, C. (1989). Breathing pattern and metabolic rate responses of rats exposed to ozone. Resp. Physiol. 76, 69-78. Mautz, W. J., Kleinman, M. T., Phalen, R. F., and Cracker, T. T. (1988). Effects of exercise exposure on toxic interactions between inhaled oxidant and aldehyde air pollutants. J. Toxicol. Environ. Health 25, 165- 177. Mautz, W. J., McClure, T. R., Reischl, P., Phalen, R. F., and Cracker, T. T. (1985a). Enhancement of ozone-induced lung injury by exercise. J. Toxicol. Environ. Health 16, 84 l-854. Mautz, W. J., Phalen, R. F., McClure, T. R., and Bufalino, C. (1985b). A rodent treadmill for inhalation toxicological studies and respirometry. J. Appl. Physiol.
58,673-679.
Silver, E. H., Leith, D. E., and Murphy, S. D. (1981). Potentiation by triorthotolyl phosphate of acrylate ester-induced alterations in respiration. Toxicology 22, 193-203. Tucker, V. A. (1968). Respiratory exchange and evaporative water loss for the flying budgerigar. J. Exp. Biol. 48, 67-87. Wiester, M. J., Williams, T. B., Ring, M. E., Menache, M. G., and Miller, F. J. (1987). Ozone uptake in awake Sprague-Dawley rats. Toxicol. Appl. Pharmacol.
89,429-437.
Withers, P. C. (1977). Measurement of voZ, vco,, and evaporative water loss with a flow-through mask. J. Appl. Physiol. 42, 120-123.
AND
APPLIED
TOXICOLOGY
18,
144-148 ( 1992)
Calibration of Respiratory Gas Exchange Measurements in Inhalation Toxicology Studies WILLIAM J. MAUTZ Air Pollution Health Eflects Laboratory. Department of Community and Environmental Medicine, University of California, Irvine, California 92717
Received April 9, 1991;accepted July 18,1991
Calibration of Respiratory Gas Exchange Measurements in Inhalation Toxicology Studies. MAUTZ, W. J. (1992). Fundum. Appl. Toxicol. 18, 144-148.
The use of simplifying assumptions for determining respiratory gas exchange is associated with substantial errors in the estimates of oxygen consumption (vo’,,) and carbon dioxide production (vco2). Studies were done to estimate the magnitude of these errors under common exposure conditions, and a simple procedure that reveals these errors is described for calibrating an open flow respirometer. The errors associated with various simplifying assumptions ranged from -1 to -21% for voioz,0.1 to 15% for v,-oZ, and 4 to 45% for the respiratory exchange ratio (R). The calibration was performed with a standard calibration gas elevated in CO, and depressed in O2 relative to air and bled into the respirometer at a measured flow rate. Dilution of the gas into the respirometer airstream simulates the effect of respiratory gas exchange, and expected values of vob2,fco2, and R compared to measured values provide a check of accuracy of flow and gas fraction measurements, a test for system leaks, and a test for the effects of any simplifying assumptions on the calculation of respiratory gas exchange. o 1992 Society of Toxicology. Currently there is substantial interest in measuring metabolic gas exchange of mammals in toxicology studies. Such measurements are made in studies that include use of exercise during exposure (Mautz et al., 1985a, 1988; Malek and Alarie, 1989), measurements of deposition of inhaled compounds in the respiratory tract (Wiester et al., 1987) and investigations of ventilatory and metabolic rate responses to inhaled toxic compounds (Silver et al., 198 1; Leith, 1984; Mautz and Bufalino, 1989). Metabolic gas exchange is important as a measure of work load in exercise studies and as a potentially affected biological variable in response to toxicant exposures. Metabolic gas exchange is usually measured with an open flow respirometry system in which subjects are exposed in a chamber to a test compound in flowing air (Fig. 1). Flow rate through the system is measured either upstream or downstream of the subject. Volume fractional concentrations of metabolic gases, O2 and CO2 , are measured upstream and downstream and, with flow rate measurements, are used to 0272-0590/92 $3.00 Copyright All rights
0 1992 by the Society of Toxicology. of reproduction in any fom reserved.
calculate O2 uptake (vo2) and COZ production ( ~coZ). In some recent studies, the use of simplifying assumptions in equations for estimating metabolic gas exchange resulted in substantial systematic errors in the estimates. The purpose of this study was to briefly review the problem, make estimates of the magnitude of errors in such assumptions using gas exchange of a rat, and present a simple calibration procedure for eliminating these errors. An accurate determination of vo2 and pcoZ in an open flow respirometer requires that the effect of entry and exit of all gases be taken into account. Entering or exiting gases change both the fractional concentrations of metabolic gases and the airflows into and out of the animal chamber (Withers, 1977; Mautz et al., 1985b). For example, the production or consumption of each respiratory gas affects the downstream fractional content of the other respiratory gas. Thus, the downstream fractional O2 content is reduced both by O2 consumed by the subject and also by the dilution effect of CO2 production, while downstream fractional CO1 content is increased both by subject CO2 production and also by removal of OZ. Changes in water vapor due to evaporation from the subject will also alter the fractions of other gases and airflow. Flow rates and metabolic gas fractions can be measured from air scrubbed of water vapor or measurements can be numerically corrected for measured water vapor content. Using measurements of respirometer temperature and pressure, values of vo2 and I&oz are corrected to standard temperature (O’C), pressure (760 mm Hg), and dry (STPD) volume rates of change. Equations that correct measured O2 and CO2 fractions for the reciprocal effects of exchange of each gas on the downstream fraction of the other gas have been derived for a variety of specific gas analyzers and sampling arrangements (Depocas and Hart, 1957; Consolazio et al., 1963; Tucker, 1968; Hill, 1972; Lister et al., 1977; Withers, 1977; Fedak et al., 198 1; Mautz et al., 1985b). A general set of equations defining total flow rates in and out of a respirometer and component volume flow rates of respiratory gases was given by Mautz et al. (1985b). These equations can be solved simultaneously for specific sampling arrangements to yield expressions for V,, and pco2 in terms of a minimum number
144
RESPIRATORY
GAS EXCHANGE
UPSTREAM
of variables that account for the variety of effects on respiratory gas fractions. For a typical case in which flow downstream of the respirometer chamber is measured, water vapor is absorbed before respiratory gases are analyzed and CO2 is not absorbed; the relationships are 3
02
=
Fl,,(
v
1 -
FEco2)
ho2(1
-
=
vE
GAS (FIORD
TEMPERATURE
SAMPLES .
5c02
PRESSURE )
I
INHALATION
ANIMAL
EXPOSURE
EXPOSURE
MIXTURE
CHAMBER
FLOWMETER
A
&,,) (1)
E 1 -
vco2
-
145
MEASUREMENT
Fkoz(l
-
FIo2
F’oz) 1 -
-
%,z
&)*
%.I,( -
FI,,,
PUMP
1 -
FEoz) ’
t DOWNSTREAM
(2)
TEMPERATURE
GAS ‘FEO2’ WATER
CALIBRATION
where the symbols follow previously established conventions (Depocas and Hart, 1957; Hill, 1972; Withers, 1977): PO2 = pco2 = vE = FIo2 = FEoz= %02 = FECOZ = R=
rate of O2 consumption of the subject (STPD) rate of CO2 production of the subject (STPD) airflow out of the respirometer (STPD) dry volume fraction of O2 in inlet air dry volume fraction of O2 in outlet air dry volume fraction of CO2 in inlet air dry volume fraction of CO, in outlet air respiratory exchange ratio, ljco2/ po2.
STPD corrections are made for the measurement of flow rate out of the respirometer using measurements of temperature, water vapor content, and pressure,
where VEATP”= airflow out of the respirometer at ambient temperature, pressure, and humidity PB = barometric pressure of outflow airstream (mm Hg) PHzo= partial pressure of water vapor in outflow a&ream (mm Hg) TA = ambient temperature of outflow airstream (OK). Expressions for gas exchange that use simplifying assumptions about inputs and outputs of all gasesto the system and their effects on the fractional content of other gases will introduce errors that can be substantial. Wiester et al. (1987) estimated gas exchange of rats in flowing air by measuring upstream flow, estimating downstream flow by correcting for water vapor addition by the rat but not for the unequal exchange of O2 and CO*, and by assuming inlet air FIco2= 0.0. Malek and Alarie (1989) measured gas exchange of guinea pigs exercising in a chamber. The downstream airflow was split: one stream was chemically scrubbed of water vapor prior to determination of Ftio2, the other was scrubbed of water vapor and COZ prior to determination of FEo2, and the streams were rehydrated in wet test meters for flow measurements. The CO* content of upstream air was also assumed to be 0.0 in the calculation of vcoZ. fo2 was calculated
’
SAMPLES FECOp’ VAPOR)
FLOWMETER
GAS
FIG. 1. Open flow respirometer components with apparatus for adding calibration gas. Arrows show direction of a&low. Lines without amows show other sensor sample points. Respiratory gas fractions (Fo2 and Fcoz) are water vapor-free fractions.
as the product of C02-free downstream flow and the difference between upstream and downstream Fo2, also CO*-free. However, when CO2 is scrubbed before downstream flow rate and Fo2 measurements are made, correction for the difference between water vapor-free and COz-free airflow into the animal chamber and water vapor-free and COz-free airflow out of the animal chamber (the difference due to oxygen consumed) is required. This correction can be made with Eq. (4) as shown by Depocas and Hart (1957), Hill (1972), and Withers (1977), ., Vi,, - FL,) ‘02 = ‘E (1 - F;,,) ’ where, following the convention
(4)
of Withers (1977), p and
F’ are respectively CO*-free flow rate and C02-free gas frac-
tions. The absolute errors of using simplifying assumptions on estimates of vo2 and li co2 in the above studies cannot be estimated from the information available; however, such assumptions both overestimate vco2 and underestimate vo2. This results in large respiratory exchange ratios (R = vco2/ Vo2) ranging from 0.89 to 1.28 (Malek and Alarie, 1989) and 1.16 to 1.35 (Wiester et al., 1987). The R expected in steady-state conditions ranges from 0.7 to 1.O depending on the nutrient sources for metabolic energy production. While transient large increases or decreases in R can occur as a result of changes in the dissolved C02-bicarbonate pool in tissue fluids, or from metabolic conversions between fats and carbohydrates, laboratory measurements of R from rodents during rest or exercise fall within the expected 0.7- 1.O range (Mautz et al., 1985a,b; Mautz and Bufalino, 1989; Brooks and White, 1978). Respirometers are typically calibrated using standard calibration gases to set the baseline and span adjustments of
146
WILLIAM
J. MAUTZ
TABLE 1 Measured and Expected Gas Exchange Values and Error Difference Obtained during Calibration of an Open Flow Respirometer” Expected gas exchange Respirometer flow rate 3000
2000 1000
vo2
tic02
19.93 12.51 8.32 4.09 12.93 8.37 4.09 7.69 4.07
11.26 1.07 4.70 2.31 7.31 4.73 2.31 4.35 2.30
Measured gas exchange and percentage error R
0.565 0.565 0.565 0.565 0.565 0.565 0.565 0.566 0.565
vco2
vo2
19.89 12.45 8.35 4.03 12.92 8.15 4.05 7.82 3.98
-0.2 -0.5 +0.4 -1.5 -0.1 -2.6 -1.0 +1.7 -2.2
11.04 6.94 4.62 2.31 7.14 4.70 2.29 4.43 2.31
-2.0 -1.8 -1.7 0.0 -2.3 -0.6 -0.9 +I.8 +0.4
R
0.555 0.557 0.553 0.573 0.553 0.557 0.565 0.566 0.580
-1.8 -1.4 -2.0 fl.4 -2.2 -2.1 0.0 0.0 +2.1
u Data are means of three successivemeasurements. Flow rates and gas exchange rates are ml/min.
gas analyzers. With a simple additional calibration step described below, an open flow respirometry system can be checked for leaks, calibrated for the specific range of expected values of poZ and vco2, and evaluated for sensitivity to simplifying assumptions in applying equations for determination of metabolic rate. METHODS A typical open flow respirometry system (Fig. 1) was used in which room air or an experimental exposure mixture passed through an animal holding chamber and then through a rotameter flowmeter (Matheson, Horsham, PA) calibrated against a soap film flowmeter (Bubble Meter, SKC West, Fullerton, CA) for the measurement of total downstream flow rate. Temperature was measured at the flowmeter with a thermocouple probe (Sensortek, Clifton NJ) calibrated against a National Bureau ofStandards traceable mercury thermometer. Water vapor content of the downstream flow was measured with a dew point sensor (EG & G Model 9 11, Waltham MA) and used with a measurement of barometric pressure to correct respirometer flow to STPD conditions by Eq. (3) above. Gas samples were drawn upstream of the chamber for measurement of Fb2 and Fb, and downstream of the respirometer flowmeter for measurement of Fb2 and FEEm using a mass spectrometer (Perkin-Elmer, Model 1100, Pomona, CA) calibrated with gravimetric standard gases (Liquid Carbonics, Los Angeles, CA). The mass spectrometer did not sensewater vapor and thus measured dry gas fractions. Eqs. ( I) and (2) were then used to estimate PO2and vco2. The calibration test of the integrated performance of the respirometer was made by introducing a measured constant flow ( ri,,,) of a standard calibration gas into the animal chamber (Fig. 1). The calibration gas was, in comparison to air, elevated in COz and depressed in O2 content (4.97% C02, 13.01% 02, 0.95% Ar, and 81.07% Nz) and was metered into the respirometer through a soap film flowmeter. Because the flowmeter was lined with liquid film solution which adds water vapor to dry air, the calibration gas was first saturated with water vapor by bubbling through a series of two midget impingers (SKC West, Fullerton, CA) tilled with distilled water. Temperature was measured to correct measured flow rate to STPD. The bleeding of calibration gas into the empty animal chamber affected the downstream gas fractions in a fashion equivalent to the respiratory gas exchange of an animal, and this simulated respiratory gas exchange was determined only by gas fractions of the calibration gas, calibration gas input flow rate, and ambient air-gas fractions.
The expected values of PO’,,and pco2 when adding the calibration gas were calculated using Eqs. 1 and 2 and substituting calibration gas flow, pc’,,,, (corrected to STPD) for I’r and calibration gas fractions FCAL02 and F cALcoIfor FEo2 and FEcoI, respectively. <,,, and F,, were the upstream air fractions. Expected values of PO2and Vco, thus obtained were compared to values derived from the downstream respirometer flowmeter and gas analysis aRer the calibration gas stream was diluted and mixed into the respirometer airstream at the empty animal chamber. The respirometer instruments and calculations were expected to yield respiratory gas exchange the same as the expected values derived from standardized gas fractions and calibration gas flow rate. Leaks in the respirometer system or errors in calibration of the respirometer flowmeter and in the STPD correction factor (Eq. 3) would introduce proportional errors in both lioz and tic,,. Errors in the calibration of the respiratory gas analyzer or use of simplifying assumptions in instrument configurations or respirometry calculations would introduce nonproportional errors (Withers, 1977) and differently affect vol and pco2 and thereby affect R. Calibrations were performed at several respirometer air flow rates and calibration gas flow rates. After calibration tests were performed, the gas exchange of a 260-g Sprague-Dawley rat was measured. These data were then used as an example to numerically examine the errors introduced by using simplifying assumptions in respiratory gas exchange calculations.
RESULTS The results of calibrating the respirometer are shown in Table 1 for three different respirometer flow rates and several calibration gas flow rates simulating gas exchange rates typical of small mammals. Percentage errors in measured po, and pcoZ compared to values expected from calibration gas flow depended on the accuracy of flow measurements and gas fraction measurements. Percentage errors in R, however, depended only on the accuracy of gas fraction measurements because flow parameters in the expression of R as the ratio of Eq. (2) to Eq. (1) cancel. Gas exchange of a 260-g Sprague-Dawley rat in 3000 ml/ min downstream flow at 22°C and 69% relative humidity was I?& = 7.82 ml/min, J&o2 = 6.41 ml/min, and R = 0.8 19. The effects of making certain simplifying assumptions on the calculation of this gas exchange are shown for a set
RESPIRATORY
GAS EXCHANGE
of sample parameters in Table 2. All of these assumptions tended to underestimate vo2 and overestimate vco2 resulting in large errors in R. The assumption that Fi,, = 0.0 produced an error in pco2 that was strongly dependent on the flow rate of air through the respirometer, and expressing to2 as the product of CC&free flow rate and COz-free oxygen fraction difference produced a large error in vo2. The error of neglecting unequal exchange of O2 and COZ in estimating downstream flow from measured upstream flow and water vapor addition was relatively small, and would approach 0 as R approached 1.O. DISCUSSION
147
MEASUREMENT
TABLE 2 Sensitivity of Respirometry Calculations to Simplifying Assumptions and Respirometer Flow Rate Percentage error Respirometer flow rate (ml min-‘)
vo2
PC’,,,
R
Case 1: Assume Fko2 = 0.0 1000
2000 3000 Case 2: Neglect the effect of unequal 0, and CO, exchange estimating r’, from i;. 1000 2000 3000 Case 3: Scrub CO2 prior to measurement of ilow and Fez. Assume Fb2 = 0.0. Use Vo2 = p! (F;, - F’&) instead of Eq. (4) 1000 2000 3000
-1.1 -2.2 -3.2
+9.9 t15
+6.0 +12 i-19
-3.6
+0.2
+3.9
+5.0
-3.7 +3.9 Determination of respiratory gas exchange in an open flow +0.09 -3.7 to.06 +4.0 respirometer as the product of air flow and upstream-downstream difference in fractional gas content requires, at minimum, a single flow measurement, either upstream or downstream of the respirometer chamber, and measurement of upstream and downstream gas fractions. The effect of each -21 +5.0 f33 -21 t10 +39 gas produced or consumed (0,) CO*, and H20) on the flow -21 t15 +45 rate and on the fractional content of the other gases needs to be accounted either by correcting flows and gas fractions for changes due to other gases (Eqs. 1, 2, and 3 for effects of C02, 02, and H20, respectively) or by chemical scrubbing tiplied by flow rate to yield estimates of voZ and vco2 (Eqs. of other gases from the airstream (Eq. 4 for the effects of 1 and 2), the resulting errors associated with these gas exCOz and see also Depocas and Hart, 1957; Hill, 1972; Withchange estimates are proportional to flow rate. In addition, ers, 1977; and Mautz et al., 1985b). it is more cumbersome to scrub CO* and H20 from the The calibration of an open flow respirometer with a stanairflow before measurement of flow and Fo2 if Eq. (4) is to dard gas mixture containing more COz and Iess O2 than air be applied to determine vol. COZ and Hz0 could instead is an effective method both for checking the combined acbe scrubbed only from the much smaller gas sample stream curacy of instruments for measuring gas exchange over the for O2 analysis rather than the entire respirometer airflow, range of expected values, testing for leaks in the system, and but then CO,-free and HzO-free respirometer flow rates must for testing the effects of any simplifying assumptions used be determined with separate measures of COZ and water vain the measurement. The technique is similar to that of Fedak por content. et al. (I 98 1) in which N2 was bled at a measured rate into a It is usually preferable in inhalation exposure studies to respirometer and its effect on FEozwas used to calibrate an measure flow downstream of the respirometer chamber and O2 analyzer for measurement of vo2. The advantage of the not impede the exposure mixture with an upstream flowpresent technique is that it includes simultaneous calibration meter, particularly if an aerosol exposure is involved. Reof the system for vcoZ and R as well as vo2. In addition, the spiratory gas exchange equations (Eq. 1 and 2) were thus sources of systematic errors in flow measurement vs gas fracpresented for a downstream flow measurement, but equation measurement can be qualitatively discriminated because tions for upstream flow measurement can be similarly dethe expected value of R is dependent on Fo2 and Fco2 but rived from the basic set of relations for volume flow rates of not on air flow through the respirometer, while voZ and pco, component respiratory gases (Mautz et al., 1985b) to yield will reflect errors in both flow and gas fraction measurements. Inhalation exposure studies present some special problems for incorporating gas exchange measurements. High flow rates may be required to deliver well-controlled concentrations of a toxicant to the breathing zone of a subject. High pco2 = PI %02(’ - FIOJ - Fko2(1 - F&x) (6) dilution air volumes make it more difficult to resolve changes ’ -F&3*-F&m in respiratory gas fractions due to metabolic gas exchange, and also increase the error associated with assuming E;,,, = where 1’1 = air-how into the respirometer (STPD). 0.0 (Table 2, Case 1). Because the values and associated errors If upstream airflow is measured, then all air passing of expressions containing respiratory gas fractions are multhrough the flowmeter must also pass through the animal
148
WILLIAM
exposure chamber, and any samples of air for measurement of upstream gas fractions or water vapor must be drawn upstream of the flowmeter. The correction of flow rate to STPD conditions (Eq. 3) requires estimation of water vapor and temperature of air in the flowmeter. Thus, for any airflow measurement requiring correction to STPD conditions, the measurements of temperature and water vapor should be made close to the flowmeter. This will minimize errors due to temperature differences between the llowmeter and humidity-temperature sensors and to possible non-steady-state loss or gain of water vapor in the air flow path between these sensors. It is important to make measurements of upstream respiratory gas fractions rather than to assume their values. Upstream gas fractions may vary due to the effects of air purification of inhalation exposure systems and to experimental addition of toxicant compounds in carrier gases such as O2 or Nz . Introduction of systematic errors in respiratory gas exchange determinations due to such carrier gases is particularly important because the error will not occur in purified air controls and will appear as an apparent treatment effect of the toxicant compound. Sampling upstream and downstream gas fractions in succession for each calculation of respiratory gas exchange will also militate against errors introduced by gas analyzer baseline drift. Because gas analyzers are typically calibrated relative to upstream air, the gas fraction upstream-downstream difference is largely independent of the upstream gas fraction (Withers, 1977); however, corrections for the effect of changes in one gas on fractional content of another (Eq. 1, 2, 4, 5, and 6) do depend on upstream values. Thus, gas exchange measurements are not wholly independent of baseline drift, but the errors are small. Under the conditions of measurement of rat gas exchange in the present study, a baseline drift in Fo2 of 0.00 1 fractional unit would produce an error of only 0.02% in Vo2. Finally, this calibration technique, like any other, is subject to the errors in the measurements that compose it. The principal sources of error are in measurement of the calibration gas flow into the respirometer and measurement of upstream gas fractions. Flow measurement using a soap film flowmeter depends on stopwatch timing, measurement of room temperature and pressure for STPD correction, and ensuring the bubble meter and respirometer are at room pressure or that the respirometer system pressure is included in the STPD correction. Errors in measuring upstream gas fractions and the analytical accuracy of the standard calibration gas affect the calibration gas bleed determination of the expected values of gas exchange. However, if Fro, is measured to within f 0.0001 fractional unit and if calibration gas fractions are
J. MAUTZ
similarly accurate, the error uncertainty of PO2derived from the gravimetric standard gas mixture used in this study is confined to +-0.2%. ACKNOWLEDGMENTS This work was supported by the California Air Resources Board (AO129-33, A833-104) and the Electric Power Research Institute (RP 1962-1). I thank C. Bufalino and D. Daniels for technical assistanceand M. Kleinman and R. Phalen for reviewing the manuscript.
REFERENCES Brooks, G. A., and White, T. P. (1978). Determination of metabolic and heart rate responses of rats to treadmill exercise. J. Appl. Physiol. 45, 1009-1015. Consolazio, C. F., Johnson, R. E., and Pecora, L. J. (1963). Physiological Measurements ofMetabolic Functions in Man. McGraw-Hill. New York. Depocas, F., and Hart, J. S. (1957). Use of the Pauling oxygen analyzer for measurement of oxygen consumption of animals in open-circuit systems and in a short-lag, closed-circuit apparatus. J. Appl. Physiol. 10,388-392. Fedak, M. A., Rome, L., and Seeherman, H. (1981). One-step N,-dilution technique for calibrating open-circuit voz measuring systems. J. Appl. Physiol.
51, 772-776.
Hill, R. W. (1972). Determination of oxygen consumption by use of the paramagnetic oxygen analyzer. J. Appl. Physiol. 33,261-263. Leith, D. E. (1984). Mass transport in mammalian lungs: Comparative physiology. J. Toxicol. Environ. Health 13, 25 1-27 1. Lister, G., Jr., Hoffman, I. E., and Rudolph, A. M. (1977). Measurement of oxygen consumption: Assessing the accuracy of a method. J. Appl. Physiol. 43, 9 16-9 17. Malek, D. E., and Alarie, Y. (1989). Ergometer within a whole-body plethysmograph to evaluate performance of guinea pigs under toxic atmospheres. Toxicol. Appl. Pharmacol. 101, 340-355. Mautz, W. J., and Bufalino, C. (1989). Breathing pattern and metabolic rate responses of rats exposed to ozone. Resp. Physiol. 76, 69-78. Mautz, W. J., Kleinman, M. T., Phalen, R. F., and Cracker, T. T. (1988). Effects of exercise exposure on toxic interactions between inhaled oxidant and aldehyde air pollutants. J. Toxicol. Environ. Health 25, 165- 177. Mautz, W. J., McClure, T. R., Reischl, P., Phalen, R. F., and Cracker, T. T. (1985a). Enhancement of ozone-induced lung injury by exercise. J. Toxicol. Environ. Health 16, 84 l-854. Mautz, W. J., Phalen, R. F., McClure, T. R., and Bufalino, C. (1985b). A rodent treadmill for inhalation toxicological studies and respirometry. J. Appl. Physiol.
58,673-679.
Silver, E. H., Leith, D. E., and Murphy, S. D. (1981). Potentiation by triorthotolyl phosphate of acrylate ester-induced alterations in respiration. Toxicology 22, 193-203. Tucker, V. A. (1968). Respiratory exchange and evaporative water loss for the flying budgerigar. J. Exp. Biol. 48, 67-87. Wiester, M. J., Williams, T. B., Ring, M. E., Menache, M. G., and Miller, F. J. (1987). Ozone uptake in awake Sprague-Dawley rats. Toxicol. Appl. Pharmacol.
89,429-437.
Withers, P. C. (1977). Measurement of voZ, vco,, and evaporative water loss with a flow-through mask. J. Appl. Physiol. 42, 120-123.