Ventilation-Perfusion Inequality in Asymptomatic Asthma14 P. D. WAGNER, D. R. DANTZKER, V. E. IACOVONI, W. C. TOMLIN, and J. B. WEST
SUMMARY. Ventilation-perfusion ( V A / Q ) inequality was measured by a multiple inert gas elimination method in 6 asymptomatic patients with asthma and in a seventh patient during a severe asthmatic episode. Measurements were made before and at 5-min intervals after administration of aerosolized isoproterenol. All patients had some residual airway obstruction as measured during forced expirations. All except one patient had clearly bimodal distributions of V A / Q ratios during all phases of the study, as confirmed by an extensive exploration of distributions compatible with each set of inert gas data. One mode lay within the normal range of V A / Q , but the other, containing 19.8 per cent of the cardiac output on the average, was centered on a V A / Q ratio of only 0.07. There was essentially no shunt. Five min after the administration of isoproterenol, the blood flow to the low V A / Q mode approximately doubled, accounting for the observed decrease in arterial Po 2 . Breathing 100 per cent 0 2 had little effect on the distribution. The presence of a bimodal distribution of V A / Q ratios without shunt suggests that collateral ventilation may be an important mechanism determining the distribution of V A / Q ratios and preventing the development of shunts. This study also showed that in some asymptomatic asthmatic patients, as many as one half of the lung units may lie behind completely closed airways and have very low but finite V A / Q ratios as a result of collateral ventilation.
Introduction There has been considerable interest in recent years in gas exchange abnormalities that occur during episodes of asthma. In most studies, arterial hypoxemia has been found, and the mechanism is believed to be the existence of poorly (Received in original form April 21, 1978 and in revised form June 27,1978) 1 From the Department of Medicine, University of California at San Diego, La Jolla, Ca. 92093 and the Naval Regional Medical Center, San Diego, Ca. 92134. 2 Supported by Grants No. HL 17731, 07212, and from 00111 from the National Institutes of Health. 3 Presented in part at the Annual Meeting of the American Thoracic Society, Montreal, Canada, May 1975. 4 Requests for reprints should be addressed to Peter D. Wagner, M.D., Department of Medicine M-013A, School of Medicine, University of California at San Diego, La Jolla, Ca. 92093.
ventilated lung units of low ventilation-perfusion ( V A / Q ) ratio (1-3). Shunts as measured during breathing of 100 per cent 0 2 have generally been found to be small (2, 3). Although current evidence points to V A / Q inequality as the major cause of hypoxemia in asthma, little information exists concerning the characteristics of the distribution of V A / Q ratios. A second area of interest has been the response to inhaled aerosolized bronchodilators. Agents of this type have been commonly found to aggravate hypoxemia in asthmatic patients (4-7), and the mechanism is still debated. One hypothesis is that the j92-adrenergic effect of drugs such as isoproterenol causes preferential dilatation of blood vessels supplying the hypoxic regions of low V A / Q , thus augmenting the amount of V A / Q inequality. If this were true, one might expect 0 2 breathing to aggravate the degree of V A / Q inequality because release of hypoxic vasoconstriction in these units would also cause preferential dilatation of their blood ves-
AMERICAN REVIEW OF RESPIRATORY DISEASE, VOLUME 118, 1978
511
512
WAGNER, DANTZKER, IACOVONI, TOMLIN, AND WEST
sels and therefore an increase in the proportion of blood flow associated with small V A / Q units. This report examines the role of V A / Q inequality in asthma using measurements made with a multiple inert gas elimination technique (8-10) in 6 asymptomatic patients with longstanding asthma. A seventh patient with severe symptoms at the time of study was included for comparison. Specifically, we wished to determine: (1) the extent and pattern of maldistribution of ventilation and blood flow in asthmatic patients with some residual airway obstruction but without symptoms at the time of study, and (2) the response of the distribution of V A / Q ratios in these patients to the administration of aerosolized isoproterenol and to the breathing of 100 per cent 0 2 .
Materials and Methods Patients. Some general data on the 7 patients studied are shown in table 1. Patients 1 through 6 were asymptomatic at the time of study, with mild airway obstruction as measured during forced expirations, but Patient 7 was experiencing a severe asthmatic episode and was an inpatient at the time of the study. No patient had evidence of irreversible airway obstruction, bronchiectasis, or abnormal radiologic findings other than those of mild hyperinflation (except Patient 7, whose chest radiograph revealed somewhat increased basal lung markings). None had evidence of right heart failure or pulmonary hypertension, and in all except Patient 7 the baseline arterial Po 2 at the time of study was at least 70 mm Hg. Mean arterial Po 2 was 85 mm Hg. None of the asymptomatic patients had experienced an asthmatic episode or had required medical attention for at least 1 month before the study and none had had a recent upper respiratory infection. Preparation of patients. After informed consent had been obtained, 3 vascular lines were first placed. (1) A 7-gauge, Swanz-Ganz flow-directed catheter was inserted percutaneously under local anesthesia into an antecubital vein using an introducer (Desilets-Hoffman, USCI Billerica, Mass.). The introducer eliminated the need for a cutdown procedure for inserting the catheter. The catheter was advanced under continuous electrocardiographic and pressure monitoring first to the right ventricle and then into the pulmonary artery. Systolic, diastolic, and mean pressures were recorded in both locations. Final placement of the catheter in the pulmonary artery was such that the tip was neither too close to the pulmonic valve nor in the wedge position (known to cause sampling errors). (2) A 20-gauge Medicut cannula (Alloe, St. Louis, Mo.) was inserted under local anesthesia into the radial artery of the nondominant hand after ensuring adequate ulnar artery supply, (i) A 19-gauge butterfly needle was inserted
into a convenient peripheral arm vein. This line was placed on the arm opposite that used for the other lines. Inert gas method. The method for estimating V A / Q distributions was that described previously, based on the simultaneous elimination of 6 inert gases dissolved in saline, infused intravenously, and measured in arterial blood and expired gas (8-10). The details of the procedure were essentially unchanged from those reported previously in normal volunteers (9), except for direct sampling of mixed venous blood, and will be described here only in outline. A solution of the 6 gases (SF6, ethane, cyclopropane, halothane, ether, and acetone) was infused into the peripheral venous cannula continuously at the rate of 2.4 ml per min throughout the study (2 hours). No samples were collected during the first 30 min of infusion, the period required to ensure that steady-state concentrations of inert gas exchange had been reached (9-11). Thereafter, measurements were made both before and after the administration of aerosolized isoproterenol according to the protocol given subsequently. Inert gas concentrations in arterial and pulmonary arterial blood and in mixed expired gas were measured as described previously (10), and these values were entered into the computer program for estimation of representative V A / Q distributions (8). Distributions were generated using the enforced smoothing approach (8) (smoothing coefficient Z = 40). The data were subjected to an analysis (8) designed to investigate the degree of variation among distributions compatible with each set of data, and the representative distributions were interpreted accordingly. The results of this analysis are given in the Appendix. Sequence of measurements. After the 30-min waiting period for the infused gases to reach a steady state, duplicate baseline measurements were made. Isoproterenol was then administered in aerosolized form according to the standard regimen in operation at the Naval Regional Medical Center, San Diego (5 vital capacity inhalations, each 1 min apart, of a 1:200 solution of isoproterenol delivered by a "Maximist" nebulizer [De Vilbiss Co., Somerset, Pa.]). After the last inhalation of isoproterenol, inert gas measurements were made during a period of 20 min at 5-min intervals. An adequate steady state of pulmonary gas exchange after administration of isoproterenol was verified by continuous measurements of pulmonary artery pressure and heart rate, tidal volume, and respiratory frequency. These variables were constant to within at least ± 10 per cent (but usually to within ± 5 per cent) for 2 to 3 min before sampling at each point. The close agreement in time course between independently measured variables such as arterial blood gas pressures, cardiac output, and inert gas exchange (see Results) was supporting evidence for an adequate steady state.
Definition 1-sec forced PO2; PaC02 Numbers
13
42
24
26
7.51 0.72
7.45 0.81 13
7.54 0.93
109
70
(122) (116) (119) (124) (120) (105) (96) (92)
20
7.42 0.80
38
81
5.5
4.1
93 34
9.0 6.4
16.2
10.0
6.3
8.5 2.5 4.3 5.7 5.5 3.9 71 3.4
8.2
(132) (173) (144) (111) (111) (79) (74) (37) 0.8
8.8 3.3 4.9 5.2 5.2 3.0 58 1.6
27
7.39 0.73 41
27
39
8.0 52 40
14.0
-
1.4(27) 1.3 (25) 0.5(11) 38 (45) 0.3 (6)
-
7.42 0.91
35
7.47 0.83
78
5.9
79
7.8
2.9
7.3 (105) 2.5(140) 4.5 (130) 5.3 (104) 4.7 (92) 2.6 (60) 55 (66) 1..4 (28)
5.8
(92) (67) (120) (110) (110) (100) (87) (62) 9.4
4.4
5.6 1.3 3.6 4.3 4.3 3.2 74 2.8
M M
F
M
1.9
4.4 (86) 1.5(96) 2.7(113) 2.9 (88) 2.8 (85) 1.9 (68) 68 (81) 1.2 (28)
M
21
178.0 79.5
25
7
178.0 72.3
26
6
175.0 62.7
47
5
178.0 80.5
4
Patients
-
(115)* (215) (189) (128) (116) (75) (66) (47)
F
37
175.0 77.3
3
5.6
5.6 2.9 3.6 4.1 3.7 2.1 57 2.2
F
30
162.5 54.5
23
157.5 70.5
2
of abbreviations: T L C = total lung capacity; RV = residual volume; FRC = functional residual capacity; VC = vital capacity; FVC = forced vital capacity; F E V j = expiratory volume; MMEF = maximal mid-expiratory flow rate; Raw = airway resistance; V E = minute ventilation; QT = total cardiac output; Pa02 = arterial = arterial PCO2; (A-a) DO2 = alveolar-arterial difference in diffusing capacity for 0 2 . in parentheses indicate per cent of predicted values.
(A-a) D 0 2 , mm Hg
R
Pa02» m m Hg PaC02' m m H 9 pH, units
Current pulmonary function T L C , liters R V , liters FRC, liters VC, liters FVC, liters F E V x , liters FEVlf%FVC M M E F , liter/sec Raw, cm H2O per liter per sec V E , liter/min CYr, liter/min
Sex
Age, years Height, cm Weight, kg
1
ANTHROPOMETRIC, HISTORICAL, A N D PULMONARY FUNCTION DATA
TABLE 1
5 ^
H
> £
2 0
£' >
Ox
>
LSTH
514
WAGNER, DANTZKER, IACOVONI, TOMLIN, AND WEST
Patients 3, 4, 5, and 7 were then given 100 per cent 0 2 to breathe, and a final set of measurements was obtained after their breathing 0 2 for 30 min. Each set of measurements was collected with the patient in a supine position, and consisted of the following steps in sequence: (1) collection of heparinized, 10- to 15-ml arterial and pulmonary arterial blood samples, (2) collection in duplicate of 20-ml expired gas samples, (3) collection of heparinized, 3-ml arterial and pulmonary arterial blood samples and a separate sample of mixed expired gas for measurement of Po 2 , Pco 2 , and (in blood) pH (radiometer electrodes were used), (4) a single measurement of cardiac output by indicator dilution using a Gilford densitometer, (5) recording of minute ventilation and respiratory frequency minute by minute using a calibrated Wright respirometer, (6) recording of systolic, diastolic, and mean pulmonary artery pressures, and (7) performance of both a relaxed and a forced expiratory maneuver recording expired volume as a function of time.
Results Baseline Conditions Distribution of VA/Q ratios. The inert gas partition coefficients, retention ratios and excretion ratios for each of the 6 inert gases in each of the 7 patients in their baseline states are given in table 2. Retention is the ratio of mixed arterial to mixed venous partial pressure, and excretion is the ratio of mixed expired to mixed venous partial pressures. For details of table 2, see Appendix. T h e corresponding representative V A / Q distributions recovered in all patients are given in figure 1. In Patient 1, the distribution was narrow and similar to those seen in normal volunteers (9), except for the presence of a small (4.0 per cent) shunt. In the remaining patients, the representative distributions were bimodal.
TABLE 2 INERT GAS D A T A FOR A L L PATIENTS G as Patient Variable
\
SF6
Ethane
Cyclopropane
Halothane
Ether
Acetone
0.00495 0.04767 0.00362 247.33
0.07778 0.14603 0.05094 186.30
0.50197 0.44871 0.21226 80.65
2.18463 0.77094 0.38384 103.21
10.81710 0.94381 0.46623 360.46
223.41200 0.99717 0.48455 6939.39
0.00467 0.00687 0.00189 394.08
0.07788 0.10553 0.02840 252.51
0.44604 0.34004 0.12003 90.75
2.14524 0.65092 0.30535 77.59
11.17224 0.89991 0.45597 208.46
209.56800 0.99417 0.49777 3368.23
R E WT
0.00848 0.00688 0.00285 394.08
0.08838 0.08027 0.02745 305.49
0.55146 0.22845 0.14371 122.80
2.67246 0.53543 0.41935 77.44
1 2.49269 0.83988 0.67565 140.88
337.94600 0.99342 0.75086 3982.79
4
X R E WT
0.00501 0.01057 0.00352 388.02
0.09460 0.19457 0.05407 124.97
0.54362 0.46787 0.20526 77.05
2.57690 0.78844 0.38684 109.84
13.55520 0.94759 0.50408 390.17
278.87000 0.99734 0.52656 7373.61
5
\ R E WT
0.00353 0.01525 0.00214 385.36
0.08331 0.21786 0.04003 116.67
0.50126 0.47833 0.16065 77.57
2.21684 0.75202 0.33774 99.03
11.40896 0.93944 0.42445 343.92
345.94000 0.99793 0.44076 9461.44
6
X R E WT
0.00467 0.01454 0.00352 390.34
0.08484 0.17875 0.05328 151.44
0.49490 0.44755 0.20907 77.83
2.38966 0.75347 0.45047 92.34
13.92790 0.94429 0.59327 366.63
364.10500 0.99769 0.64258 8501.53
7
\
0.00635 0.07254 0.00337 194.45
0.09870 0.37919 0.03506 75.29
0.53580 0.54625 0.13909 74.90
2.3970 0.77070 3.1444 96.10
13.24695 0.93297 0.50799 312.34
282.70500 0.99655 0.55831 5684.00
1
R E WT 2
\ R E WT
3
\
R E WT
Definition of abbreviations'- X. = partition coefficient; R = retention ratio (mixed arterial to mixed venous partial pressure); E = excretion ratio (mixed expired to mixed venous partial pressure); WT = inverse of the square root of minimal variance of retention estimated from the combined retention and excretion data.
515
V A / Q INEQUALITY IN ASTHMA
A total of 9 measurements was made in Patient 1, and in every case the result was a narrow unimodal distribution. A total of 48 measurements was made in the remaining 6 patients, and in every case the result was a bimodal pattern. T h e reproducibility of the fundamental pattern and the results of analysis to determine variability among compatible solutions (see Appendix) left little doubt as to the fundamental pattern of inequality observed in these patients, namely, that the distribution was bimodal. Only 1 patient (No. 1) had any shunt ( V A / Q of 0), and even in this case the magnitude was very small. On the other hand, in the 6 patients with bimodal distributions the findings of no shunt, but of considerable amounts of perfusion in regions of very low V A / Q , were striking. In Patients 2 through 7, the fraction of cardiac output associated with the mode of low V A / Q was 9.7, 7.5, 21.4, 24.8, 19.5 and 44.8 per cent, respectively. V A / Q ratios at the peak of these modes were 0.11, 0.03, 0.10, 0.05, 0.07, and 0.03, respectively. Patients 2 and 3 (table 1) were hyperventilating at the time, as is clear from the relationships between Pco 2 and pH. From the pH values of table 1, it would appear that minute ventilation
O
0
001
01
0
0.01
0.1
10
1.0
100 1000
0
001
01
10.0 1000
0
0.01
0.1
10
1.0
increased by 50 to 70 per cent. Because these patients continued to hyperventilate for the duration of the study, it is not known whether the distributions were altered as a result. There is some evidence that increases in ventilation of this magnitude do not alter distributions much in disease. For example, in patients with chronic obstructive pulmonary disease with similar areas of low V A / Q ratios (12), no consistent change in pattern was observed during exercise in which ventilation increased by at least a factor of 2 from the resting values. This finding supports the suggestion that the modest degree of hyperventilation observed did not greatly influence the pattern of V A / Q inequality. It should also be pointed out that the bimodal patterns observed occurred in all patients (except No. 1) and thus are not related to the degree of ventilation. Areas of high V A / Q ratios were not observed in any patient. This was consistent with the pa hologic finding in asthma that the primary changes are in the airways. This would be expected to cause inequality of ventilation and chiefly areas of low V A / Q ratios. Presumably, any inequality of blood flow is through active mechanisms such as hypoxic vasoconstriction that
100 1000
0
001
01
10.0 1000
0
0.01
0.1
10
10
100 100.0
0
0.01
0.1
1.0
10.0 100.0
10.0 100.0
VENTILATI0N-PERFUSI0N RATIO
Fig. 1. Representative distributions obtained under baseline conditions in each of the 7 patients. These results suggest that in all except Patient 1, 2 distinct populations of gas exchange units were present, namely, those with ventilationperfusion (VA/Q) ratios in the normal range and those with VA/Q,ratios between 0.01 and 0.1. The latter population contains approximately 20 per cent of the cardiac output. Except in Patient 1, there was no shunt. Areas of high V A / Q ratio did not exist.
516
WAGNER, DANTZKER, IACOVONI, TOMLIN, AND WEST
would serve to decrease the degree of V A / Q inequality rather than cause high V A / Q units. Arterial Po2. Baseline arterial Po 2 values (table 1) were surprisingly high for the amount of V A / Q inequality measured. Thus, in all but one of the asymptomatic patients (No. 2) Po 2 was 78 mm Hg or greater, whereas the inert gas data suggested the presence of large amounts of V A / Q inequality. It is possible to determine how much hypoxemia a given distribution will produce if the mixed venous and inspired Po 2 and Pco 2 are known (12, 13). T h e result of this calculation with a comparison of the measured arterial Po 2 in each case is shown in figure 2. (This figure also contains similar calculations made under conditions following bronchodilator therapy, which will be discussed subsequently). In figure 2, the solid squares from which each arrow originates represent prebronchodilator conditions. There is generally good agreement between the measured arterial Po 2 (abscissa) and that calculated from the inert gas data (ordinate). Thus, if mixed venous and inspired Po 2 and Pco 2 are taken into account,
50
60
70
80
90
100
110
MEASURED ARTERIAL P0 2 , mmHg Fig. 2. Relationship between measured arterial Pox and arterial Po* calculated from the recovered distribution of ventilation-perfusion ( V A / Q ) ratios taking into account measured values for mixed venous and inspired Pd* and Pco*. Closed squares from which arrows originate are baseline values before administration of bronchodilator. Open circles are values obtained 5 min after inhalation* of isoproterenol. Closed circles are values obtained at 10, 15, and 20 min after bronchodilator. There was generally good agreement between measured and calculated values for Po2, which suggests that the observed degree of V A / Q inequality completely explained the hypoxemia in these patients.
the hypoxemia can be explained by the amount of V A / Q inequality. The relatively well-preserved arterial Po 2 in the face of moderately severe V A / Q inequality in these examples can be attributed to generally large amounts of cardiac output, ventilation, or both (contributing differently in different subjects; table 1). Patient 7 was a good example of the influence of factors other than the amount of V A / Q inequality in arterial Po 2 . Almost 50 per cent of the pulmonary blood flow was associated with areas of very low V A / Q , and the arterial Po 2 was 52 mm Hg. This situation can be contrasted with measurements made in patients with heart failure after myocardial infarction (14) in which similar degrees of hypoxemia were observed, but the amount of perfusion associated with areas of low (or zero) V A / Q ratios was only 10 to 15 per cent of the cardiac output. The difference was the high cardiac output in Patient 7 as compared to the low values seen in heart failure. Changes after Inhalation of Aerosolized Isoproterenol Large transient changes were observed in the degree of airway obstruction, in inert gas concentrations, and in both hemodynamic and blood gas data after inhalation of isoproterenol in all but the severely affected patient (No. 7). With the exception of the measurement of airway obstruction, all of these changes had returned to baseline values by 10 min after inhalation of isoproterenol, and remained at or near baseline values throughout the remainder of the study. Shown in figure 3 is the complete sequence of representative distributions obtained in 1 typical patient (No. 4) together with the measured values for arterial Po 2 . Similar changes were seen in Patients 2, 3, 5, and 6, who also had bimodal V A / Q distributions before bronchodilator therapy. By contrast, Patient 1, with a narrow distribution before bronchodilator therapy, showed little response to isoproterenol, (an increase in shunt from 4.0 to 7.4 per cent) and the severely affected patient (No. 7) showed essentially no response in any of the variables measured. T h e individual changes in perfusion of the low V A / Q mode and in arterial Po 2 are shown in figure 4. The most striking finding in the V A / Q distribution after inhalation of isoproterenol was that the perfusion of the poorly ventilated low V A / Q mode transiently increased by a factor of approximately 2. Thus, although airway obstruc-
517
VA/Q INEQUALITY IN ASTHMA
1
Ventilation
DM. ASTHMA 5 MIN. POST
1.2 -
IBRONCHODILATOR
10 -
PQ0 2 = 7 0
08 -
or
BLOOD FLOW ,
L/m
c
06 -
o
f
0.4 -
Blood Flow \
0.2 -
> 0
0.01
0.1
1.0
10.0
VENTILATION-PERFUSION 1.4 i 1.2
,No Shunt 0 -
100.0
£„ 0 0.01
RATIO
1.2 Ventilation
1.0
\
1.0 •
0.8
0.8
0.6
0.6
0.4
0.4 I
K01
> 0.1
LO
10.0
VENTILATION-PERFUSION RATIO
y ,,
1.0
100
100.0
RATIO
D.M ASTHMA 20 MIN. POST BRONCHODILATOR Poo2=84
11
Ventilation
Blood FlowT\
ft
\
0.2
5
Ventilation
V\
i ~o\
VENTILATION-PERFUSION 1.4 •
D.M. ASTHMA 10 MIN. POST BRONCHODILATOR
J \
um.mr'. IAJ , 0.01
0.1
LO
1
VENTILATION-PERFUSION
100
100.0
RATIO
Fig. 3. Sequence of representative distributions obtained in Patient 4, showing the baseline result before administration of bronchodilator and those at varying time intervals after the administration of isoproterenol. At 5 min after the administration of the drug, although the bimodality was still evident, the mode of lower ventilation-perfusion ( V A / Q ) ratio contained approximately twice as much perfusion as under baseline conditions. By 10 min (and thereafter), the distribution was no different from the baseline configuration. Similar results were obtained in Patients 2,3, 5, and 6, but no change was observed in the symptomatic Patient 7.
tion was decreased by administration of isoproterenol, as judged by maximal mid-expiratory flow (MMEF) (figure 5) and 1-sec forced expiratory volume (FEVj), the amount of V A / Q inequality was transiently increased by the drug. At first sight, a 2-fold increase in the blood flow of units constituting the low V A / Q mode should decrease their V A / Q ratio by a factor of approximately 2. However, from figure 5 it can be seen that the mean V A / Q ratio of lung units in that mode increased at the same time as the perfusion increased. T h e reason for the increase in mean V A / Q is that ventilation of these units increased relatively more than did blood flow. Increasing the perfusion of the low V A / Q units worsens the over-all V A / Q relationships, but increasing their ventilation improves them. T h e net result, however, is a worsening of gas exchange because the amount by which V A / Q ratios were increased was insignificant in terms of alveolar Po 2 . Thus, alveolar Po 2 in the low V A / Q units remained close to mixed venous
values, whereas the blood flow-weighted contribution of these units to the value for arterial Po 2 approximately doubled. After the transient effects had subsided (by 10 min) the amount of inequality was indistinguishable from that present before administration of isoproterenol (figure 3). T h e mean results in Patients 2, 3, 4, 5, and 6 are summarized in figure 5. Here, hemodynamic, gas exchange, and flow rate measurements are shown as changes from baseline (that is, values before administration of isoproterenol). It can be seen that the time courses of each variable except MMEF coincide and indicate maximal changes 5 min after bronchodilator therapy. Effects of Breathing 100 Per Cent 02: Patients 3,4,6, and 7 No systematic changes in pattern or extent of V A / Q inequality were observed in the 4 patients who were given 100 per cent 0 2 to breathe for 30 min. This finding is of particular interest
518
WAGNER, DANTZKER, IACOVONI, TOMLIN, AND WEST
O h- < > CD 100
tios became unventilated after an equivalent period of breathing 100 per cent 0 2 . The absence of shunt is supported by direct measurements of arterial Po 2 , which was no less than 550 mm Hg and usually in the range of 600 mm Hg. In our hands, even young, normal subjects were not seen to have an arterial Po 2 greater than this during breathing of 100 per cent 0 2 , and the probable reasons have been discussed previously (9).
Q
Discussion
LU Q
O 2
200
SUMMARY. Ventilation-perfusion ( V A / Q ) inequality was measured by a multiple inert gas elimination method in 6 asymptomatic patients with asthma and in a seventh patient during a severe asthmatic episode. Measurements were made before and at 5-min intervals after administration of aerosolized isoproterenol. All patients had some residual airway obstruction as measured during forced expirations. All except one patient had clearly bimodal distributions of V A / Q ratios during all phases of the study, as confirmed by an extensive exploration of distributions compatible with each set of inert gas data. One mode lay within the normal range of V A / Q , but the other, containing 19.8 per cent of the cardiac output on the average, was centered on a V A / Q ratio of only 0.07. There was essentially no shunt. Five min after the administration of isoproterenol, the blood flow to the low V A / Q mode approximately doubled, accounting for the observed decrease in arterial Po 2 . Breathing 100 per cent 0 2 had little effect on the distribution. The presence of a bimodal distribution of V A / Q ratios without shunt suggests that collateral ventilation may be an important mechanism determining the distribution of V A / Q ratios and preventing the development of shunts. This study also showed that in some asymptomatic asthmatic patients, as many as one half of the lung units may lie behind completely closed airways and have very low but finite V A / Q ratios as a result of collateral ventilation.
Introduction There has been considerable interest in recent years in gas exchange abnormalities that occur during episodes of asthma. In most studies, arterial hypoxemia has been found, and the mechanism is believed to be the existence of poorly (Received in original form April 21, 1978 and in revised form June 27,1978) 1 From the Department of Medicine, University of California at San Diego, La Jolla, Ca. 92093 and the Naval Regional Medical Center, San Diego, Ca. 92134. 2 Supported by Grants No. HL 17731, 07212, and from 00111 from the National Institutes of Health. 3 Presented in part at the Annual Meeting of the American Thoracic Society, Montreal, Canada, May 1975. 4 Requests for reprints should be addressed to Peter D. Wagner, M.D., Department of Medicine M-013A, School of Medicine, University of California at San Diego, La Jolla, Ca. 92093.
ventilated lung units of low ventilation-perfusion ( V A / Q ) ratio (1-3). Shunts as measured during breathing of 100 per cent 0 2 have generally been found to be small (2, 3). Although current evidence points to V A / Q inequality as the major cause of hypoxemia in asthma, little information exists concerning the characteristics of the distribution of V A / Q ratios. A second area of interest has been the response to inhaled aerosolized bronchodilators. Agents of this type have been commonly found to aggravate hypoxemia in asthmatic patients (4-7), and the mechanism is still debated. One hypothesis is that the j92-adrenergic effect of drugs such as isoproterenol causes preferential dilatation of blood vessels supplying the hypoxic regions of low V A / Q , thus augmenting the amount of V A / Q inequality. If this were true, one might expect 0 2 breathing to aggravate the degree of V A / Q inequality because release of hypoxic vasoconstriction in these units would also cause preferential dilatation of their blood ves-
AMERICAN REVIEW OF RESPIRATORY DISEASE, VOLUME 118, 1978
511
512
WAGNER, DANTZKER, IACOVONI, TOMLIN, AND WEST
sels and therefore an increase in the proportion of blood flow associated with small V A / Q units. This report examines the role of V A / Q inequality in asthma using measurements made with a multiple inert gas elimination technique (8-10) in 6 asymptomatic patients with longstanding asthma. A seventh patient with severe symptoms at the time of study was included for comparison. Specifically, we wished to determine: (1) the extent and pattern of maldistribution of ventilation and blood flow in asthmatic patients with some residual airway obstruction but without symptoms at the time of study, and (2) the response of the distribution of V A / Q ratios in these patients to the administration of aerosolized isoproterenol and to the breathing of 100 per cent 0 2 .
Materials and Methods Patients. Some general data on the 7 patients studied are shown in table 1. Patients 1 through 6 were asymptomatic at the time of study, with mild airway obstruction as measured during forced expirations, but Patient 7 was experiencing a severe asthmatic episode and was an inpatient at the time of the study. No patient had evidence of irreversible airway obstruction, bronchiectasis, or abnormal radiologic findings other than those of mild hyperinflation (except Patient 7, whose chest radiograph revealed somewhat increased basal lung markings). None had evidence of right heart failure or pulmonary hypertension, and in all except Patient 7 the baseline arterial Po 2 at the time of study was at least 70 mm Hg. Mean arterial Po 2 was 85 mm Hg. None of the asymptomatic patients had experienced an asthmatic episode or had required medical attention for at least 1 month before the study and none had had a recent upper respiratory infection. Preparation of patients. After informed consent had been obtained, 3 vascular lines were first placed. (1) A 7-gauge, Swanz-Ganz flow-directed catheter was inserted percutaneously under local anesthesia into an antecubital vein using an introducer (Desilets-Hoffman, USCI Billerica, Mass.). The introducer eliminated the need for a cutdown procedure for inserting the catheter. The catheter was advanced under continuous electrocardiographic and pressure monitoring first to the right ventricle and then into the pulmonary artery. Systolic, diastolic, and mean pressures were recorded in both locations. Final placement of the catheter in the pulmonary artery was such that the tip was neither too close to the pulmonic valve nor in the wedge position (known to cause sampling errors). (2) A 20-gauge Medicut cannula (Alloe, St. Louis, Mo.) was inserted under local anesthesia into the radial artery of the nondominant hand after ensuring adequate ulnar artery supply, (i) A 19-gauge butterfly needle was inserted
into a convenient peripheral arm vein. This line was placed on the arm opposite that used for the other lines. Inert gas method. The method for estimating V A / Q distributions was that described previously, based on the simultaneous elimination of 6 inert gases dissolved in saline, infused intravenously, and measured in arterial blood and expired gas (8-10). The details of the procedure were essentially unchanged from those reported previously in normal volunteers (9), except for direct sampling of mixed venous blood, and will be described here only in outline. A solution of the 6 gases (SF6, ethane, cyclopropane, halothane, ether, and acetone) was infused into the peripheral venous cannula continuously at the rate of 2.4 ml per min throughout the study (2 hours). No samples were collected during the first 30 min of infusion, the period required to ensure that steady-state concentrations of inert gas exchange had been reached (9-11). Thereafter, measurements were made both before and after the administration of aerosolized isoproterenol according to the protocol given subsequently. Inert gas concentrations in arterial and pulmonary arterial blood and in mixed expired gas were measured as described previously (10), and these values were entered into the computer program for estimation of representative V A / Q distributions (8). Distributions were generated using the enforced smoothing approach (8) (smoothing coefficient Z = 40). The data were subjected to an analysis (8) designed to investigate the degree of variation among distributions compatible with each set of data, and the representative distributions were interpreted accordingly. The results of this analysis are given in the Appendix. Sequence of measurements. After the 30-min waiting period for the infused gases to reach a steady state, duplicate baseline measurements were made. Isoproterenol was then administered in aerosolized form according to the standard regimen in operation at the Naval Regional Medical Center, San Diego (5 vital capacity inhalations, each 1 min apart, of a 1:200 solution of isoproterenol delivered by a "Maximist" nebulizer [De Vilbiss Co., Somerset, Pa.]). After the last inhalation of isoproterenol, inert gas measurements were made during a period of 20 min at 5-min intervals. An adequate steady state of pulmonary gas exchange after administration of isoproterenol was verified by continuous measurements of pulmonary artery pressure and heart rate, tidal volume, and respiratory frequency. These variables were constant to within at least ± 10 per cent (but usually to within ± 5 per cent) for 2 to 3 min before sampling at each point. The close agreement in time course between independently measured variables such as arterial blood gas pressures, cardiac output, and inert gas exchange (see Results) was supporting evidence for an adequate steady state.
Definition 1-sec forced PO2; PaC02 Numbers
13
42
24
26
7.51 0.72
7.45 0.81 13
7.54 0.93
109
70
(122) (116) (119) (124) (120) (105) (96) (92)
20
7.42 0.80
38
81
5.5
4.1
93 34
9.0 6.4
16.2
10.0
6.3
8.5 2.5 4.3 5.7 5.5 3.9 71 3.4
8.2
(132) (173) (144) (111) (111) (79) (74) (37) 0.8
8.8 3.3 4.9 5.2 5.2 3.0 58 1.6
27
7.39 0.73 41
27
39
8.0 52 40
14.0
-
1.4(27) 1.3 (25) 0.5(11) 38 (45) 0.3 (6)
-
7.42 0.91
35
7.47 0.83
78
5.9
79
7.8
2.9
7.3 (105) 2.5(140) 4.5 (130) 5.3 (104) 4.7 (92) 2.6 (60) 55 (66) 1..4 (28)
5.8
(92) (67) (120) (110) (110) (100) (87) (62) 9.4
4.4
5.6 1.3 3.6 4.3 4.3 3.2 74 2.8
M M
F
M
1.9
4.4 (86) 1.5(96) 2.7(113) 2.9 (88) 2.8 (85) 1.9 (68) 68 (81) 1.2 (28)
M
21
178.0 79.5
25
7
178.0 72.3
26
6
175.0 62.7
47
5
178.0 80.5
4
Patients
-
(115)* (215) (189) (128) (116) (75) (66) (47)
F
37
175.0 77.3
3
5.6
5.6 2.9 3.6 4.1 3.7 2.1 57 2.2
F
30
162.5 54.5
23
157.5 70.5
2
of abbreviations: T L C = total lung capacity; RV = residual volume; FRC = functional residual capacity; VC = vital capacity; FVC = forced vital capacity; F E V j = expiratory volume; MMEF = maximal mid-expiratory flow rate; Raw = airway resistance; V E = minute ventilation; QT = total cardiac output; Pa02 = arterial = arterial PCO2; (A-a) DO2 = alveolar-arterial difference in diffusing capacity for 0 2 . in parentheses indicate per cent of predicted values.
(A-a) D 0 2 , mm Hg
R
Pa02» m m Hg PaC02' m m H 9 pH, units
Current pulmonary function T L C , liters R V , liters FRC, liters VC, liters FVC, liters F E V x , liters FEVlf%FVC M M E F , liter/sec Raw, cm H2O per liter per sec V E , liter/min CYr, liter/min
Sex
Age, years Height, cm Weight, kg
1
ANTHROPOMETRIC, HISTORICAL, A N D PULMONARY FUNCTION DATA
TABLE 1
5 ^
H
> £
2 0
£' >
Ox
>
LSTH
514
WAGNER, DANTZKER, IACOVONI, TOMLIN, AND WEST
Patients 3, 4, 5, and 7 were then given 100 per cent 0 2 to breathe, and a final set of measurements was obtained after their breathing 0 2 for 30 min. Each set of measurements was collected with the patient in a supine position, and consisted of the following steps in sequence: (1) collection of heparinized, 10- to 15-ml arterial and pulmonary arterial blood samples, (2) collection in duplicate of 20-ml expired gas samples, (3) collection of heparinized, 3-ml arterial and pulmonary arterial blood samples and a separate sample of mixed expired gas for measurement of Po 2 , Pco 2 , and (in blood) pH (radiometer electrodes were used), (4) a single measurement of cardiac output by indicator dilution using a Gilford densitometer, (5) recording of minute ventilation and respiratory frequency minute by minute using a calibrated Wright respirometer, (6) recording of systolic, diastolic, and mean pulmonary artery pressures, and (7) performance of both a relaxed and a forced expiratory maneuver recording expired volume as a function of time.
Results Baseline Conditions Distribution of VA/Q ratios. The inert gas partition coefficients, retention ratios and excretion ratios for each of the 6 inert gases in each of the 7 patients in their baseline states are given in table 2. Retention is the ratio of mixed arterial to mixed venous partial pressure, and excretion is the ratio of mixed expired to mixed venous partial pressures. For details of table 2, see Appendix. T h e corresponding representative V A / Q distributions recovered in all patients are given in figure 1. In Patient 1, the distribution was narrow and similar to those seen in normal volunteers (9), except for the presence of a small (4.0 per cent) shunt. In the remaining patients, the representative distributions were bimodal.
TABLE 2 INERT GAS D A T A FOR A L L PATIENTS G as Patient Variable
\
SF6
Ethane
Cyclopropane
Halothane
Ether
Acetone
0.00495 0.04767 0.00362 247.33
0.07778 0.14603 0.05094 186.30
0.50197 0.44871 0.21226 80.65
2.18463 0.77094 0.38384 103.21
10.81710 0.94381 0.46623 360.46
223.41200 0.99717 0.48455 6939.39
0.00467 0.00687 0.00189 394.08
0.07788 0.10553 0.02840 252.51
0.44604 0.34004 0.12003 90.75
2.14524 0.65092 0.30535 77.59
11.17224 0.89991 0.45597 208.46
209.56800 0.99417 0.49777 3368.23
R E WT
0.00848 0.00688 0.00285 394.08
0.08838 0.08027 0.02745 305.49
0.55146 0.22845 0.14371 122.80
2.67246 0.53543 0.41935 77.44
1 2.49269 0.83988 0.67565 140.88
337.94600 0.99342 0.75086 3982.79
4
X R E WT
0.00501 0.01057 0.00352 388.02
0.09460 0.19457 0.05407 124.97
0.54362 0.46787 0.20526 77.05
2.57690 0.78844 0.38684 109.84
13.55520 0.94759 0.50408 390.17
278.87000 0.99734 0.52656 7373.61
5
\ R E WT
0.00353 0.01525 0.00214 385.36
0.08331 0.21786 0.04003 116.67
0.50126 0.47833 0.16065 77.57
2.21684 0.75202 0.33774 99.03
11.40896 0.93944 0.42445 343.92
345.94000 0.99793 0.44076 9461.44
6
X R E WT
0.00467 0.01454 0.00352 390.34
0.08484 0.17875 0.05328 151.44
0.49490 0.44755 0.20907 77.83
2.38966 0.75347 0.45047 92.34
13.92790 0.94429 0.59327 366.63
364.10500 0.99769 0.64258 8501.53
7
\
0.00635 0.07254 0.00337 194.45
0.09870 0.37919 0.03506 75.29
0.53580 0.54625 0.13909 74.90
2.3970 0.77070 3.1444 96.10
13.24695 0.93297 0.50799 312.34
282.70500 0.99655 0.55831 5684.00
1
R E WT 2
\ R E WT
3
\
R E WT
Definition of abbreviations'- X. = partition coefficient; R = retention ratio (mixed arterial to mixed venous partial pressure); E = excretion ratio (mixed expired to mixed venous partial pressure); WT = inverse of the square root of minimal variance of retention estimated from the combined retention and excretion data.
515
V A / Q INEQUALITY IN ASTHMA
A total of 9 measurements was made in Patient 1, and in every case the result was a narrow unimodal distribution. A total of 48 measurements was made in the remaining 6 patients, and in every case the result was a bimodal pattern. T h e reproducibility of the fundamental pattern and the results of analysis to determine variability among compatible solutions (see Appendix) left little doubt as to the fundamental pattern of inequality observed in these patients, namely, that the distribution was bimodal. Only 1 patient (No. 1) had any shunt ( V A / Q of 0), and even in this case the magnitude was very small. On the other hand, in the 6 patients with bimodal distributions the findings of no shunt, but of considerable amounts of perfusion in regions of very low V A / Q , were striking. In Patients 2 through 7, the fraction of cardiac output associated with the mode of low V A / Q was 9.7, 7.5, 21.4, 24.8, 19.5 and 44.8 per cent, respectively. V A / Q ratios at the peak of these modes were 0.11, 0.03, 0.10, 0.05, 0.07, and 0.03, respectively. Patients 2 and 3 (table 1) were hyperventilating at the time, as is clear from the relationships between Pco 2 and pH. From the pH values of table 1, it would appear that minute ventilation
O
0
001
01
0
0.01
0.1
10
1.0
100 1000
0
001
01
10.0 1000
0
0.01
0.1
10
1.0
increased by 50 to 70 per cent. Because these patients continued to hyperventilate for the duration of the study, it is not known whether the distributions were altered as a result. There is some evidence that increases in ventilation of this magnitude do not alter distributions much in disease. For example, in patients with chronic obstructive pulmonary disease with similar areas of low V A / Q ratios (12), no consistent change in pattern was observed during exercise in which ventilation increased by at least a factor of 2 from the resting values. This finding supports the suggestion that the modest degree of hyperventilation observed did not greatly influence the pattern of V A / Q inequality. It should also be pointed out that the bimodal patterns observed occurred in all patients (except No. 1) and thus are not related to the degree of ventilation. Areas of high V A / Q ratios were not observed in any patient. This was consistent with the pa hologic finding in asthma that the primary changes are in the airways. This would be expected to cause inequality of ventilation and chiefly areas of low V A / Q ratios. Presumably, any inequality of blood flow is through active mechanisms such as hypoxic vasoconstriction that
100 1000
0
001
01
10.0 1000
0
0.01
0.1
10
10
100 100.0
0
0.01
0.1
1.0
10.0 100.0
10.0 100.0
VENTILATI0N-PERFUSI0N RATIO
Fig. 1. Representative distributions obtained under baseline conditions in each of the 7 patients. These results suggest that in all except Patient 1, 2 distinct populations of gas exchange units were present, namely, those with ventilationperfusion (VA/Q) ratios in the normal range and those with VA/Q,ratios between 0.01 and 0.1. The latter population contains approximately 20 per cent of the cardiac output. Except in Patient 1, there was no shunt. Areas of high V A / Q ratio did not exist.
516
WAGNER, DANTZKER, IACOVONI, TOMLIN, AND WEST
would serve to decrease the degree of V A / Q inequality rather than cause high V A / Q units. Arterial Po2. Baseline arterial Po 2 values (table 1) were surprisingly high for the amount of V A / Q inequality measured. Thus, in all but one of the asymptomatic patients (No. 2) Po 2 was 78 mm Hg or greater, whereas the inert gas data suggested the presence of large amounts of V A / Q inequality. It is possible to determine how much hypoxemia a given distribution will produce if the mixed venous and inspired Po 2 and Pco 2 are known (12, 13). T h e result of this calculation with a comparison of the measured arterial Po 2 in each case is shown in figure 2. (This figure also contains similar calculations made under conditions following bronchodilator therapy, which will be discussed subsequently). In figure 2, the solid squares from which each arrow originates represent prebronchodilator conditions. There is generally good agreement between the measured arterial Po 2 (abscissa) and that calculated from the inert gas data (ordinate). Thus, if mixed venous and inspired Po 2 and Pco 2 are taken into account,
50
60
70
80
90
100
110
MEASURED ARTERIAL P0 2 , mmHg Fig. 2. Relationship between measured arterial Pox and arterial Po* calculated from the recovered distribution of ventilation-perfusion ( V A / Q ) ratios taking into account measured values for mixed venous and inspired Pd* and Pco*. Closed squares from which arrows originate are baseline values before administration of bronchodilator. Open circles are values obtained 5 min after inhalation* of isoproterenol. Closed circles are values obtained at 10, 15, and 20 min after bronchodilator. There was generally good agreement between measured and calculated values for Po2, which suggests that the observed degree of V A / Q inequality completely explained the hypoxemia in these patients.
the hypoxemia can be explained by the amount of V A / Q inequality. The relatively well-preserved arterial Po 2 in the face of moderately severe V A / Q inequality in these examples can be attributed to generally large amounts of cardiac output, ventilation, or both (contributing differently in different subjects; table 1). Patient 7 was a good example of the influence of factors other than the amount of V A / Q inequality in arterial Po 2 . Almost 50 per cent of the pulmonary blood flow was associated with areas of very low V A / Q , and the arterial Po 2 was 52 mm Hg. This situation can be contrasted with measurements made in patients with heart failure after myocardial infarction (14) in which similar degrees of hypoxemia were observed, but the amount of perfusion associated with areas of low (or zero) V A / Q ratios was only 10 to 15 per cent of the cardiac output. The difference was the high cardiac output in Patient 7 as compared to the low values seen in heart failure. Changes after Inhalation of Aerosolized Isoproterenol Large transient changes were observed in the degree of airway obstruction, in inert gas concentrations, and in both hemodynamic and blood gas data after inhalation of isoproterenol in all but the severely affected patient (No. 7). With the exception of the measurement of airway obstruction, all of these changes had returned to baseline values by 10 min after inhalation of isoproterenol, and remained at or near baseline values throughout the remainder of the study. Shown in figure 3 is the complete sequence of representative distributions obtained in 1 typical patient (No. 4) together with the measured values for arterial Po 2 . Similar changes were seen in Patients 2, 3, 5, and 6, who also had bimodal V A / Q distributions before bronchodilator therapy. By contrast, Patient 1, with a narrow distribution before bronchodilator therapy, showed little response to isoproterenol, (an increase in shunt from 4.0 to 7.4 per cent) and the severely affected patient (No. 7) showed essentially no response in any of the variables measured. T h e individual changes in perfusion of the low V A / Q mode and in arterial Po 2 are shown in figure 4. The most striking finding in the V A / Q distribution after inhalation of isoproterenol was that the perfusion of the poorly ventilated low V A / Q mode transiently increased by a factor of approximately 2. Thus, although airway obstruc-
517
VA/Q INEQUALITY IN ASTHMA
1
Ventilation
DM. ASTHMA 5 MIN. POST
1.2 -
IBRONCHODILATOR
10 -
PQ0 2 = 7 0
08 -
or
BLOOD FLOW ,
L/m
c
06 -
o
f
0.4 -
Blood Flow \
0.2 -
> 0
0.01
0.1
1.0
10.0
VENTILATION-PERFUSION 1.4 i 1.2
,No Shunt 0 -
100.0
£„ 0 0.01
RATIO
1.2 Ventilation
1.0
\
1.0 •
0.8
0.8
0.6
0.6
0.4
0.4 I
K01
> 0.1
LO
10.0
VENTILATION-PERFUSION RATIO
y ,,
1.0
100
100.0
RATIO
D.M ASTHMA 20 MIN. POST BRONCHODILATOR Poo2=84
11
Ventilation
Blood FlowT\
ft
\
0.2
5
Ventilation
V\
i ~o\
VENTILATION-PERFUSION 1.4 •
D.M. ASTHMA 10 MIN. POST BRONCHODILATOR
J \
um.mr'. IAJ , 0.01
0.1
LO
1
VENTILATION-PERFUSION
100
100.0
RATIO
Fig. 3. Sequence of representative distributions obtained in Patient 4, showing the baseline result before administration of bronchodilator and those at varying time intervals after the administration of isoproterenol. At 5 min after the administration of the drug, although the bimodality was still evident, the mode of lower ventilation-perfusion ( V A / Q ) ratio contained approximately twice as much perfusion as under baseline conditions. By 10 min (and thereafter), the distribution was no different from the baseline configuration. Similar results were obtained in Patients 2,3, 5, and 6, but no change was observed in the symptomatic Patient 7.
tion was decreased by administration of isoproterenol, as judged by maximal mid-expiratory flow (MMEF) (figure 5) and 1-sec forced expiratory volume (FEVj), the amount of V A / Q inequality was transiently increased by the drug. At first sight, a 2-fold increase in the blood flow of units constituting the low V A / Q mode should decrease their V A / Q ratio by a factor of approximately 2. However, from figure 5 it can be seen that the mean V A / Q ratio of lung units in that mode increased at the same time as the perfusion increased. T h e reason for the increase in mean V A / Q is that ventilation of these units increased relatively more than did blood flow. Increasing the perfusion of the low V A / Q units worsens the over-all V A / Q relationships, but increasing their ventilation improves them. T h e net result, however, is a worsening of gas exchange because the amount by which V A / Q ratios were increased was insignificant in terms of alveolar Po 2 . Thus, alveolar Po 2 in the low V A / Q units remained close to mixed venous
values, whereas the blood flow-weighted contribution of these units to the value for arterial Po 2 approximately doubled. After the transient effects had subsided (by 10 min) the amount of inequality was indistinguishable from that present before administration of isoproterenol (figure 3). T h e mean results in Patients 2, 3, 4, 5, and 6 are summarized in figure 5. Here, hemodynamic, gas exchange, and flow rate measurements are shown as changes from baseline (that is, values before administration of isoproterenol). It can be seen that the time courses of each variable except MMEF coincide and indicate maximal changes 5 min after bronchodilator therapy. Effects of Breathing 100 Per Cent 02: Patients 3,4,6, and 7 No systematic changes in pattern or extent of V A / Q inequality were observed in the 4 patients who were given 100 per cent 0 2 to breathe for 30 min. This finding is of particular interest
518
WAGNER, DANTZKER, IACOVONI, TOMLIN, AND WEST
O h- < > CD 100
tios became unventilated after an equivalent period of breathing 100 per cent 0 2 . The absence of shunt is supported by direct measurements of arterial Po 2 , which was no less than 550 mm Hg and usually in the range of 600 mm Hg. In our hands, even young, normal subjects were not seen to have an arterial Po 2 greater than this during breathing of 100 per cent 0 2 , and the probable reasons have been discussed previously (9).
Q
Discussion
LU Q
O 2
200
