The ventilatory responses of conscious dogs to isocapnic oxygen tests. A method of exploring the central component of respiratory drive and its dependence on O2 and CO2

Respiration Physiology (1980) 39, 183-197 © Elsevier/North-Holland Biomedical Press

THE

VENTILATORY

RESPONSES

TO ISOCAPNIC A METHOD RESPIRATORY

OF EXPLORING

OF CONSCIOUS

OXYGEN THE

TESTS.

CENTRAL

DRFVE AND ITS DEPENDENCE

A. UNGAW

DOGS

COMPONENT ON Oz AND

OF

COz

a n d P. B O U V E R O T

Laboratoire de Physiologie Respiratoire, Centre National de la Recherche Scientffi'que (Laboratoire assoei~ ~ l'Universit~ Louis Pasteur) 23 rue Becquerel, 67087 Strasbourg, France

Abstract. Conscious unrestrained dogs trained to breathe through a respiratory mask or, after chronic tracheostomy, through a cuffed endotracheal tube were studied in an altitude chamber operated in such a way that end-tidal Po2 was maintained at 100, 75 or 60 Torr. Each hypoxic experiment was completed within 1 h of the onset of hypoxia. At all levels of oxygenation, resting pulmonary ventilation (V), obtained from the tidal volume (VT) and ventilatory period (T), and alveolar gas tensions (PAo2, PAco~) were measured cycle-by-cycle before and during isocapnic O2-tests (IOT) at various steady levels of alveolar P¢o2 ranging from 30 to 48 Torr. For this, Pco~ in the inspired gas before and during lOT was adjusted so that Paco 2 remained unchanged in the course of the first few breaths which followed the switch to hyperoxia. In analysing the transient changes of V in the course of lOT, it was considered that an apnoea occurred'when there was no measurable deflection on the integrated pneumotachogram past a duration twice the control T from the beginning of the last recorded ventilatory cycle. (1) Control V vs. PAco~- relationships showed classic positive interaction between hypercapnia and hypoxia; (2) during lOT at PAo~ of 100, 75 or 60 Torr, an apnoea occurred, V invariably falling to zero, provided that PAco_, was below 38-35 Tort according to the level of oxygenation ; (3) above t.hat threshold PAco 2 value, the residual minimum ventilation (~Zres) observed during lOT was linearly related to PAco_, ; (4) ~'res vs. PAco_~relationships showed negative interaction between hypercapnia and hypoxia. It is concluded that (a) through isocapnic O2-tests , both the peripheral and central components of the ventilatory drive can be quantitatively estimated; (b) in conscious dogs, the pulmonary ventilation appears to be entirely driven by afferent activity from the arterial chemoreceptors, even in eueapnic normoxia ; (c) the lower minimum ventilation seen in the course of O2-tests from a hypoxic rather than a normoxic background is still observed at PAco 2 above normal, thus cannot be due only to hypocapnia related to preceding hypoxic hyperventilation but must be caused by a central respiratory inhibition directly or indirectly related to depressant effect of even moderate hypoxia. Arterial chemoreceptors Central hypoxic depression Control of breathing

Hypercapnia Hypoxia Oxygen breath tests

Accepted for publication 20 September 1979 I Present address: Department of Pharmacology, University of Edinburgh, 1 George Square, Edinburgh EH8 9JZ, United Kingdom. 183

184 "

A. UNGARAND P. BOUVEROT

A reflex ventilatory drive from the arterial chemoreceptors, if present in an animal or in man, can be demonstrated by studying breath-by-breath the transient changes of pulmonary ventilation (V) provoked by the sudden inhalation of pure oxygen (O2-test method; Dejours, 1957). A fall in ventilation, beginning a few seconds after the switch to 02, is interpreted as being due to the suppression of a chemoreceptor 02 drive (also called hypoxic drive) since (1) in anesthetized cats, a decrease of ~" starts about 3 sec after the onset of 02 breathing, together with a decrease in the discharge rate of sinus nerve chemoreceptor fibres (Leitner et al., 1965); (2) in the conscious dog, the transient fall in V seen under similar conditions no longer occurs after chronic chemodenervation (Bouverot et al., 1965). The fall in ventilation has been used to assess the magnitude of the 02 ventilatory drive acting reflexly through the arterial chemoreceptors (Dejours, 1962) in man (Dejours et al., 1958 ; Downes and Lambertsen, 1966; Lee and Bishop, 1974; Stockley, 1977), mammals (Bouverot et al., 1965; Favier and Lacaisse, 1978; Purves, 1966) and birds (Jones and Purves, 1970; BouVerot and S6bert, 1979). As an example, Bouverot et al. 0973) observed in conscious dogs that only two breaths of pure 02 caused "v"to fall abruptly within the first 20 sec following the switch to oxygen. V fell from 2.7 to 1.7 L BTPS. rain J in the normoxic animals (Pao,= 97 Torr), and from 3.5 to 1.3 L BTPS. rain ~ in acutely hypoxic animals (Pao, = 48 Tort). Here, the chemoreflex mechanism appeared to control 37~ and 63~o of the minute volume in normoxia and hypoxia, respectively. The residual minimum ventilation ('V'res) observed in the course of transient 02breathing can be interpreted as the result of various influences acting on the respiratory centres, influences which do not originate from arterial chemoreceptor stimulation. In the above-mentioned experiments of Bouverot et al. (1973), the slightly lower Vres from the hypoxic background (1.3 L BTPS. rain ~compared with 1.7 L BTPS.min ~ in normoxic animals) could plausibly be explained by the accompanying central hypocapnia since (1) Paco ~was 32 Torr in hypoxia and 37 Tort in normoxia, (2)the secondary increases of Pco, and H ~ concentration due to hypoventilation most probably did not influence the intracranial CO 2 receptors until several seconds later (Bouverot et al., 1965). On the other hand, from experiments in anesthetized animals, it is known that, after denervation of the arterial chemoreceptors, hypoxia depresses ventilation (Selladurai and Wright, 1932 ; Euler and Liljestrand, 1936; ~str6m, 1952" Cherniack et al.. 1970/71). In this study we have attempted to further distinguish between the central CO~ excitatory influence and acute hypoxia's depressant effect on ventilation. In conscious dogs studied in a hypobaric chamber, isoxic 9 vs. PAco~relationships were determined at three levels of oxygenation. By holding end-tidal Pco: constant in the course of transient O2-tests, '(/res vs. PAco~ response curves were also determined in the same experiments. The results indicate that the greater ventilatory depression seen in conventional O~-tests from a hypoxic rather than a normoxic background cannot be due to the hypocapnia related to hypoxic hyperventilation, but must be due to central ventilatory depression caused by even moderate hypoxia. ,

ISOCAPNIC OXYGEN TESTS

185

Methods ANIMALS

Experiments were performed on 3 conscious adult mongrel dogs, two males (C and M; 19 and 17.5 kg respectively) and one female (F; 18.5 kg). The animals were trained to lie quietly, unrestrained, on a table in a hypobaric chamber, and to wear a mask (M and F) or a cuffed endotracheal cannula (C and F in which a chronic tracheostomy was performed about four months before the beginning of the experiments). The mask, a soft plastic cone, was placed on the dog's muzzle and sealed with a rubber strap behind the mouth; in dog F, the tracheal orifice was then occluded with an airtight bandage. The mask or tracheal tube was connected to low-resistance valves via a heated No. 2 Fleisch pneumotachograph. The total instrumental dead space was about 30 ml with the mask, and 35 ml with the cannula which had to be connected to the pneumotachograph via a soft rubber tube and a rigid conical junction. As the ventilatory responses of dog F with cannula or with mask were identical, the results were pooled.

MEASUREMENTS (1) Tidal volume (VT) and ventilatory period (T), from which were derived the respiratory.frequency (f) and minute volume of ventilation ((1") were measured cycle by cycle. The Fleisch pneumotachograph was connected to a differential inductance manometer (_+2 mbar). The output of the manometer fed simultaneously (1) an integration circuit and a photographic recorder, (2) a minicomputer-analog-to-digital converter. For each breath, the computing routine calculated VT (ml BTPS) from the output signal of the manometer and T (sec) by measuring the elapsed time between trigger pulses, f (min-~) determined as 60/T, and "q on a breath-by-breath basis by multiplying VT by 60/T. The ventilatory values were displayed on a video-terminal together with the time of occurrence of each expiration. Data were stored on magnetic tape for further statistical treatment. Before each experiment, the system was calibrated with a handdriven syringe which introduced and withdrew 10 fixed volumes of air at the usual frequency of the dog's respiration, and at the same mean gas temperature which was continuously read with a thermistor. We verified that ventilatory readings were overestimated by about 10% when oxygen, instead of air, was used in measurements (Grenvik et al., 1966). Since the error makes the measured residual minimum ventilation, "~res in the course of O2-tests, somewhat higher, it weights against the argument we will advance in the Discussion, thus strengthening the conclusion. (2) The 02 and CO 2 partial pressures in the ventilated gas were measured by using a fast polarographic O2-analyzer (OM-11 Beckman) and a fast infrared CO2-analyzer

186

A. UNGAR AND P. BOUVEROT

(LB-2 Beckman). Gas was continuously sampled, via a fine P.E. tube inserted between the respiratory valves and the pneumotachograph. The sampling gas flow was adjusted in such a way that linear outputs and good end-tidal plateaux for both CO2 and 02 were obtained over the range of ventilatory frequencies encountered in the experiments. The output signals of the analyzers were displayed simultaneously on the photographic recorder and on the X - Y calibrated coordinates of an oscilloscope. The latter was used to guide the manual addition of N 2 and CO, to the inspired gas mixture (see below).

EXPERIMENTS

All animals fasted for 12 h before each experiment. The experiments were performed from January to April in the altitude chamber, in which temperature ranged from 17 to 21 °C. The inspiratory port of the valves was connected via a threeway tap to either of two wide T-tubes (inner diameter 2 cm) through which gases circulated in excess at constant identical flow rates (12 L . m i n ~). Gases were either ambient air (delivered by a compressor pump via a regularizing capacity) or compressed pure oxygen. CO2 and N 2 could be trickled into the air via a set of rotameters by manual adjustment of needle valves; similarly, CO 2 could be added to the oxygen. Humidifying gas-mixing chambers allowed gases to be practically saturated with water at room temperature before reaching the three-way tap. Via the tap, the dog could be abruptly switched from one T-tube to the other. Upstream from the inspiratory port, changes of pressure at the time of a switch were less than 1 mm H20; then, the gas flowing in excess through each T-tube escaped via its free arm. In preliminary experiments, the 02 T-line was replaced by a rubber bag in which 02 and CO 2 could be introduced at will. Three groups of experiments were performed at various levels of oxygenation in the altitude chamber, in normoxia (PB = 750 Torr) and in acute hypoxia (P~ -- 560 or 460 Torr approximately). The chamber was operated at low altitude or was depressurized within 3-5 min at a rate of 12-15 m . sec-~, and the pressure was adjusted in such a way that dogs breathing ambient air (FIo2 = 0.21) had steady end-tidal Po: values of 100, 75 or 60 Torr. At each level of oxygenation, experiments were carried out with CO 2 added to the inspired gas or not. When CO 2 was added in order to increase PAco2, enough N: was also added to hold PAo, constant if hyperventilation occurred. At steady values of PAo2 and PAco_, maintained for at least 7 rain, control ventilatory measurements were made and several O2-tests were performed. The dog was then switched for 5 breaths to the O2-1ine to which CO 2, in amounts determined by trial and error, was added when necessary, so that PAco2 did not change during the first few breaths of the test. If PAco_~changed, the test was rejected. Approximately 7 rain between successive O2-tests was necessary for PAo, to return to control value for at least 3 min. Each day's experiment on a given dog consisted of two successive studies at a

I S O C A P N I C O X Y G E N TESTS

187

single level of oxygenation and two different values of PAco 2. Most often, 5 02tests were performed in each daily experiment. Each animal was repeatedly studied during the experimental period. The three conscious resting dogs underwent a total of 342 O2-tests: 108 at 100 Torr PAo:, 84 at 75 Torr and 150 at 60 Torr. Every study under hypoxia was completed within 1 h.

A N A L Y S I S OF D A T A

In the analysis of each O2-test, the averaged values of VT, T, f and "~' during the 10 ventilatory cycles before the onset of 0 2 breathing were taken as the control values. In the standard O2-test (Dejours et al., 1958) the breath-by-breath changes in "¢ are calculated by dividing the tidal volume, VT, by the corresponding ventilatory period, T. In our initial experiments in hypoxia we observed that in many tests, after 2-3 O2-breaths, VT became unmeasurably small for a considerable time (see fig. 1A). We considered that the minimal minute ventilation, Vres, was effectively zero in these circumstances, and would be grossly overestimated by the standard analysis (compare the horizontal dashed line ad in fig. 1B to the solid line cd; see Discussion). In the present study, we therefore took ~'res to be zero whenever the following conditions were satisfied: (1) T was fairly constant before the test; (2) the period of the few ventilatory cycles immediately after the onset of O2-breathing hardly changed, compared with the mean control T value; (3) at the maximal ventilatory depression there was no measurable deflection on the integrated pneumotachogram for several seconds past a duration = 2 T from the beginning of the last recorded ventilatory cycle (ac in fig. 1). The results obtained with all the dogs at each level of oxygenation were pooled. For each group of experiments the linear regression of control "~/ on PAco 2 and its 95% confidence limits were computed. Since "¢res was zero for all tests where PAco: was below than 35-38 Tort, depending on the level of oxygenation, the regression of '¢res on PAco 2 was computed only for tests where PAco, was higher than 35 38 Torr.

Results I S O C A P N I C O X Y G E N TESTS

Figure 1A shows a typical O2-test in the conscious dog F previously made acutely hypoxic (PAo, = 60 Torr) and eucapnic (PAco 2 = 33 Torr) by breathing ambient air to which a small amount of CO 2 was added (Plco, = 5 Torr) at 4000 m in the altitude chamber (PB = 460 Torr). When the dog was given oxygen to breath (first arrow in fig. 1A), PAco 2 did not change because a similar amount of CO 2 was also added to the 02 gas line (see Methods). As the solid line in fig. 1B shows, the minute

188

A. U N G A R A N D P. B O U V E R O T

A

eco2, Torr 40

0 V T , LBTPS

a b c

d

0.4 t o I

I

I

I

I

I

I

B

~/, L BTPS.min -~

4

c

2

0

b

I

0

I

c

I

20

I

410

I

sec

Fig. I. Comparison of the breath-by-breath analyses of an "isocapnic O2-test" made with the standard method and the method used in this study. (A) F r o m above downwards: changes in CO 2 tensions in the ventilated gas, Pco~ ; integrated pneumotachogram as a function of time, VT; time = 10 sec; event marker which shows when and how long the three-way tap was turned on 02 (see Methods). Arrows indicate the first and last 02 breaths. Before and after the O2-test , the conscious dog breathed air at PB = 460 Torr; some CO 2 and N 2 was added to the air in order to maintain PAo: and PAco, constant at 60 and 33 Tort respectively. Some CO 2 was also added to oxygen so that PAco_, hardly changed in the course of the two first O2-breaths. The period of the 2nd breath consecutive to the start of 02 breathing would be a d according to the standard method of analysis. In this study, it was considered that the period for that 2rid breath, as indicated by ac, was twice a s g r e a t as the mean control T (close to ab) and that c d w a s an apnoeic time. (B) Transient changes in the minute volume, V, as a function of time in the course of the O,-test shown in (A). 9 was calculated at 60 VT/T (see Methods). Time zero indicates the start of the first inspiration under hyperoxia. Note that the standard method of analysis leads to earlier starting, overestimated minimal ventilation ('Vres, dashed line, ad) when compared with the value determined in this study (solid line, cd). Using ab as the ventilatory period of the 2nd O 2 breath would not change the residual minute volume, ~res, but would increase the measured ventilatory response latency (the dotted vertical line bb is located to the right of the solid vertical component of the aa line).

volume, "q, abruptly fell to zero after a short delay, and the dog did not breath for the 20 following seconds approximately (Vres = 0). Figure 2 shows the pooled ventilatory responses to 5 successive O2-tests performed on the same dog previously made slightly hypoxic (PAo, = 75 Torr) and hypercapnic

189

ISOCAPNIC OXYGEN TESTS

[';AiR CO~"I.

02 *CO~

V, L STPs.min-'

4 .~_ t._..!

_.[_~-4'

2

-'

I

0

0

10

I

"

I

I

20

30

sec

eAco2, Torr 40

~

eAo2, Torr

20O 100

0

_~ 0

I

I

I

10

20

30

sec

Fig. 2. Mean transient ventilatory effects of O -tests in a conscious dog breathing either through a mask or through an endotracheal tube and previously made slightly hypoxic. Ordinate: From the top down, breath-by-breath pulmonary ventilation (~'), partial pressure of C O in the end-tidal gas (PAco~), partial pressure of 02 in the end-tidal gas (PAo_3. Abscissa: time, starting from the onset of 0 2 breathing. Prior to the O2-tests, the animal breathed ambient air at 2 500 m (PB = 560 Torr) with some CO 2 added in order to keep PAco~ around 40 Torr and PAO, = 75 Torr. At time zero, the dog was switched to oxygen with some CO, added to prevent any significant change in PAco_~during the 2 or 3 first subsequent breaths. Squares and dashed line, values obtained in the dog breathing through a mask; circles and solid line, values obtained in the dog breathing through an endotracheal tube. Mean values _+SE (n = 5).

(PAco 2 ----40 T o r r a p p r o x i m a t e l y ) b y b r e a t h i n g a m b i e n t air w i t h s o m e C O 2 a d d e d at a b o u t 2500 m (PB = 560 T o r r ) . W h e t h e r t h e d o g b r e a t h e d t h r o u g h a m a s k ( s q u a r e a n d d a s h e d lines) o r t h r o u g h a n e n d o t r a c h e a l t u b e (circle a n d solid line) the results w e r e i d e n t i c a l . It c a n b e seen t h a t s w i t c h i n g the a n i m a l to o x y g e n w i t h s o m e C O 2 a d d e d did n o t s i g n i f i c a n t l y m o d i f y PAco_, v a l u e s for t h e 3 - 4 s u b s e q u e n t b r e a t h s . By c o n t r a s t , PAo2 i n c r e a s e d a b r u p t l y a n d w a s a b o v e 250 T o r r at the t i m e o f t h e o b s e r v e d maximal ventilatory depression.

190

A. U N G A R A N D P. B O U V E R O T

E F F E C T S O F V A R Y I N G PAco ~ ON C O N T R O L V A N D Vres AT T H R E E LEVELS OF OXYGENATION

The pooled results of experiments at PAo: of 100 Torr, 75 Torr and 60 Torr are presented in the three sections of fig. 3. The upper line of each section represents the steady-state regression of control ~' on PAco ~; the results are consistent with previous work showing the potentiation by hypoxia of the ventilatory response to CO 2 (Nielsen and Smith, 1952). The lower line of each section represents the regression of Vres on CO 2. The Vres vs. PAco: line for PAo, = 75 Torr (fig. 3B) is not significantly depressed from that for 100 Torr (fig. 3C), but at PAo, = 60 Torr (fig. 3A) the slope of the line is strongly depressed. At all levels of PAo~, Vres was zero at low Paco: values. The intercepts of the three Vres lines on the PAco, axis are not significant different from one another; the shallow slope of the line for PA<, = 60 Torr makes the estimate of that intercept imprecise. Discussion CRITIQUE OF METHODS

In the present study as in an experiment previously reported (Bouverot et al., 1973), conscious dogs were studied at identical altitudes of 4000 m (PB = 460 Torr). The transient ventilatory responses to O2-tests appear to be strikingly different in the two studies, the maximal ventilatory depression observed in our dogs being substantially greater than that reported by Bouverot et al. (1973, their fig. 1 and table 1). This difference however can easily be explained. First of all, the inhalation of 02 was restricted to two breaths in the earlier experiment, while it was continued for about five breaths in our dogs. Thus, it may be that discharge from the arterial chemoreceptors was not identically diminished in A

L B T PS'miFI-I 10

.: ..j " ...

75

"Y''"Y

3~0

J

4LO

I

B s"

C

..

100

.,.~

315 415 end-tidal Pco2, Torr

3J5

I

i

45

Fig. 3. Control ~' and minimal ~" in the course of isocapnic O2-tests as a function of PAco z. Pooled data on three conscious dogs at three levels of oxygenation. Ordinate: minute volume, V. Abscissa: end-tidal or alveolar Pco 2 values. A, B a~d C represent experiments performed at PAo2 values of 60, 75 and 100 Torr, respectively, as indicated in each panel. In each section, the upper line represents the regression of steady control V on PAco 3 with 95','; confidence limits, and the lower line the regression of 'qres on PAco ~.

191

ISOCAPNIC OXYGEN TESTS

TABLE 1 Resting ventilatory values in a conscious dog (F) breathing either through a cuffed endotracheal tube or through a respiratory mask Tracheal tube PAo2 (Torr)

PAco: (Torr)

VT {ml BTPS)

100

40

224 8

0.4

(13) 75

T (sec)

V (ml BTPS/min)

PAco: (Torr)

VT (ml BTPS)

4.45

3 020

42

250

4.92

3 050

0.30

97

6

0.30

190

(13)

(13)

38.5

212

3.53

3 600

0.3

6

0.60

96

(8) 60

(13)

Mask

(8)

(8)

(8)

35.0

216

3.20

4050

0.3

9

0.60

70

(10)

(10)

(lO)

(10)

0.4

T (sec)

(17)

(17)

40

290

4.70

3 700

58

0.20

58

0.3

(8) 36 0.4

(8)

(8)

(17)

(ml BTPS/mIn)

(8)

(17)

(8)

263

3.67

4300

9

0.20

180

(8)

(8)

(8)

the two studies. Indeed in experiments preliminary to the present study, we found that dogs had to breathe at least three tidal volumes of oxygen for the ventilation to be maximally depressed from a normoxic background. To make sure this point was passed, our dogs breathed 02 for about 20-30 sec. Most often this time was long enough for five tidal volumes to be completed; however, when an apnoea occurred as shown in fig. 1A, 02 breathing was discontinued after 3 or 4 tidal volumes. A second point to be considered is that all the dogs studied in 1973 breathed through a respiratory mask while most of our present experiments were performed on dogs chronically tracheostomized and wearing a cuffed endotracheal tube. Although the possible influence of tracheostomy on alveolar ventilation and thermolysis was found to be slight (Flandrois et al., 1971), we checked whether ventilation was affected by the tracheal tube or not. As table 1 indicates and fig. 2 illustrates, neither the resting ventilatory values nor the transient ventilatory responses to O2-tests were significantly different in the dog breathing through a cuffed endotracheal tube and in the same animal wearing a respiratory mask, though PAco 2values tended to be slightly lower when the cannula was used. This latter observation may be related to the fact that tracheal cannulation curtailed the upper respiratory airways, thus reducing the anatomical dead space. This, however, was partly compensated by the larger instrumental dead space (see Methods). It may be concluded that, provided PAco: was the same in both situations, breathing through a tracheal tube or through a mask had no significant influence on our results. A final point relates to the method of analysing the transient ventilatory changes in the course of O,-test when an apnoea occurs. Apnoea simply means absence of breathing. As generally used, the term implies that the cessation of ventilatory movements is temporary. In the example shown in fig. 1, the second breath with follows

192

A. U N G A R A N D P. BOUVEROT

the switch to the hyperoxic gas mixture can be analysed in different ways, the related ventilatory period being taken as equal to ab, ac or ad. By our criteria (see Methods: Analysis of data), we consider that apnoea occurred between c and d, and thus that ~-res was zero (solid line cd in fig. 1B), because there was no measurable deflection on the integrated pneumotachogram for several seconds past a duration ac = 2 T from the beginning of that second breath, T being the control ventilatory period. By a less drastic criterion, we could say that apnoea started at b, considering that the actual ventilatory period of the second 02 breath was ab: this would not affect the value of Vres, but would slightly increase the latency of the ventilatory response (in fig. 1B the dotted vertical line bb is shifted to the right of the solid component of the vertical line aa) and prolong the apnoeic period (horizontal line bcd in fig. IB). If we now do what Bouverot et al. (1973) did, and consider that the ventilatory period of the second O: breath in fig. 1 was ad, then the latency for the maximal ventilatory response would be shorter but Vres would be overestimated, being greater than zero, and the apnoeic period obscured (vertical and horizontal dashed lines a a d i n fig. 1B). When reexamined on the basis of our present criteria, the data of Bouverot et al. (1973) indicate that their hypoxic animals most often developed an apnoea in the course of their double O2-tests; consequently, '¢res was necessarily zero. Then the apparent discrepancies between our results and those of these authors are reconciled; they are mainly due to different methods of analysis. We believe that the method used in the present study leads to results more consistent with the actual ventilatory changes during some O2-test (those which are accompanied by transient apnoea), and thus to a better understanding of the mechanisms controlling the ventilatory motor OUtl~Ut, now to be discussed.

PART OF CONTROL x/DRIVEN BY THE CHEMOREFLEX MECHANISM

In conscious dogs, the Pao~ threshold for stimulation of the arterial chemoreceptors has been estimated to be about 200 Torr (Bouverot et al., 1965). Taking into consideration the alveolar-arterial oxygen differences at various levels of oxygenation, which remain to be determined in the conscious dog, it may be safely assumed that the Po2 of the arterial blood rose above this threshold value in all the O2-tests performed in the present study. Indeed, fig. 2 shows that PAo~ abruptly increased to 250-300 Torr within the first 15 sec following the switch to 02 in the dogs breathing air at 2500 m altitude. Similar observations were made at 140 m or 4000 m. Moreover, in the latter situation, PAo~ rose to 300 Torr within an even shorter time (mean value 12 + 2 sec) because of the previous hypoxic hyperventilation and, consequently, of a faster 02 wash-in during transient oxygen breathing. It may, therefore be considered that the discharge from the arterial chemoreceptors was transiently suppressed in the course of all the O2-tests of the present study. The observed fall in ventilation would then indicate how much of the control ventilation had been driven by this afferent activity.

ISOCAPNIC OXYGEN TESTS

193

However, this would lead to underestimation of the chemoreflex drive of ventilation for two reasons. First, as discussed by Dejours (1962), a rise in H ÷ concentration due to an increase of HbO 2 cannot be avoided in the course of O2-tests. Also unavoidable, though delayed, is the CO2-H + stimulus which progressively builds up due to the hypoventilation during 02 breathing. Secondly, as said in Methods, our measurements were overestimated by about 10~o when oxygen circulated through the pneumotachograph calibrated with air; as a consequence, the minimal ventilation during O2-tests was also overestimated, and thus the maximal fall of ventilation was underestimated. With these limitations in mind, such an underestimate of the reflex ventilatory drive originating in the arterial chemoreceptors can be visualized in each panel of fig. 3 by the vertical distance between the upper and lower solid lines. This vertical distance represents the maximal change in "¢ observed, at constant PAco:, in the course of our isocapnic 3- to 5-breath O2-tests. From fig. 3 it is obvious that (1) at any PAco: value, the lower the level of oxygenation, the greater is the chemoreflex drive, a well-known phenomenon, (2) when PAco2 decreased below 35-38 Torr depending the level of oxygenation, the pulmonary ventilation appears to be entirely driven by afferent activity from the arterial chemoreceptors, since the minimal ventilation was zero when this activity disappeared in the course of transient O2-breathing (see lower solid line in each panel of fig. 3). MINIMAL 'V D U R I N G ISOCAPNIC O2-TESTS

As discussed above, there is a strong probability that Pao~ rose above the threshold for arterial chemoreceptors during our O2-tests. Thus, the minimal or residual ventilation (Vres) then observed must be the net result of two opposite excitatory and inhibitory effects: (1) other excitatory influences which do not originate from arterial chemoreceptor stimulation, (2) a direct depression of the brain stem respiratory complex by hypoxia (St. John and Wang, 1977), and possibly also an indirect cortical inhibitory influence (Tenney et al., 1971). Focusing attention on the lower solid line in the three panels of fig. 3, we can see that : (1) below a PAco ~,value around 35-38 Torr according to the level of oxygenation, "¢res was zero as already mentioned; (2) above this PAco ~ value, X)res increased linearly with PAco_~but, when the level of oxygenation then decreased from 100 to 60 Torr PAo:, the slope of the ~rres vs. PAco_~relationship decreased. The specific mechanisms underlying these changes are not clear yet. However, several points call for discussion. The first observation, that ~)res was zero below a threshold PAco_, value, suggests that central inhibitory influences then overcame other excitatory influences which did not originate from arterial chemoreceptor stimulation, in particular the central excitatory action of the CO~-H + stimulus. It is worth noting that there was no measurable Vres in our normoxic dogs (PAo, = 100 Torr; fig. 3C) when PAco_, was below 38 Torr. Such a high Po_~threshold for central depression at quite normal Pco~ is rather surprising. However, in surgically chemodenervated cats in which

194

A. U N G A R A N D P. B O U V E R O T

Paco~ was maintained at 32 Torr, Lahiri (1976) found a threshold Pao~ for ventilatory depression around 130 Torr. On the other hand, Koepchen et al. (1976) as well as St. John and Wang (1977) reported a decrease in the activity of inspiratory neurons as Pao: was decreased to about 90 Torr. It would not be unreasonable to admit that hypoxic central depression can actually affect ventilation in so-called normoxic conditions (Miller and Tenney, 1975). Although PAco: did remain constant until the ventilatory response developed during our isocapnic O~-tests, this does not necessarily mean that Pco~ in the brain stem remained constant. In particular, the changes in cerebral blood flow which have been described during hypocapnic hypoxia (Severinghaus et al., 1966) and hypercapnic hypoxia (Cohen et al., 1967) may well have been responsible for a decreased central Pco~. However we believe that this effect was small. Indeed, Bouverot and Bureau (1975) found that, in chemodenervated conscious dogs which remained hypercapnic when made hypoxic, a shift of the cerebrospinal fluid towards alkalosis, explicable by concomitant increase in the cerebral blood flow, was not significant before 2 h hypoxic exposure. In the present study, all measurements were completed within 1 h after starting hypoxia. The second observation that the slope of the Vres vs. PAco~ relationship decreased with PAo~ indicates a negative interaction between hypercapnia and hypoxia when the chemoreflex mechanism is suppressed. When '¢res is plotted as a function of PAo~ (fig. 4), it may be seen that the ventilation which proceeded from nonchemoreflex mechanisms (1) was zero at PAco~= 35 Torr and remained small at P A c o = 40 Torr; (2)was quite important at a maintained PAco: of 45 Torr in normoxia, but decreased considerably as PAo: was decreased. This latter observation can be explained by a central hypoxic inhibition that affects the central sensitivity to CO,. A similar phenomenon has been previously described by Cherniack et al. (1970/71, 1976) and Lahiri (1976) who concluded that hypercapnia and hypoxia act synergistically in producing ventilatory depression, The mechanism for this remains to be elucidated.

Vres

L BTPs'min -~

PAco 2 ,Torr

4

~E]

[]

~

o

6•0

o

J

45

A

40 o

8L0

i

35

L 100 PAo2, Torr

Fig. 4. Residual minute ventilation, ~'res, in the course of isocapnic O2-tests as a function of PAo:. Pooled data on three conscious dogs at three levels of alveolar P('o:. Data redrawn from fig. 3.

ISOCAPN1C OXYGEN TESTS

195

A last point deserves to be examined, at least as a working hypothesis. Suprapontine reticular influences on respiratory control are well recognized (Cohen and Hugelin, 1965), but their role in the hypoxic ventilatory response has not received much attention. Discharging arterial chemoreceptors afferents are known to excite respiratory neurons via specific pathways. On the other hand, the same afferents are known to excite the reticular activating system which, in turn, via ascending pathways. causes cortical arousal. The cortex, in turn, as emphasized by Tenney et al. (1971), is known to exert descending inhibitory influences on the medullary respiratory centres. Thus, it would not be unreasonable to consider that the arterial chemoreceptors exert two opposite effects on the respiratory neuronal pool: (1) a direct excitatory effect, acting after a short delay via specific paucisynaptic pathways, and (2) an indirect inhibitory effect acting after a longer delay via non-specific polysynaptic pathways, Then it might be hypothe6z~d that only the first excitatory effect would disappear in the course of O2-test, while the previously initiated but delayed inhibitory effect could continue to act. Then, the lower the level of oxygenation before the O2-test, the stronger might be the inhibitory influence from the cortex at the time the transient O2-breathing suppresses the direct stimulating action from the arterial chemoreceptors. Whatever the mechanisms involved, the results of the present study indicate that isocapnic O2-tests prolonged for several breaths can be used to assess both peripheral and central components of the ventilatory drive in resting conscious subjects. They also indicate, in the conscious dog, that depressant ventilatory effects can exist at modest levels of hypoxia and hypercapnia.

Acknowledgements The authors thank Dr. R. G o o d e for his participation in some experiments. They also gratefully acknowledge the technical assistance of Miss Lacaisse, and the help of J.P. Gendner in the computer operations.

References A,str6m, A. (1952). On the action of combined carbon dioxide excess and oxygen deficiency in the regulation of breathing. Acta PhysioL Stand. 27: Suppl. 98. Bouverot, P., R. Flandrois, R. Puccinelli and P. Dejours (1965). Etude du r61e des ch6mor6cepteurs art6riels dans la r6gulation de la respiration pulmonaire chez le Chien 6veill6. Arch. Int. Pharmacodyn. 157:253 271. Bouverot, P., V. Candas and J.P. Libert (1973). Role of the arterial chemoreceptors in ventilatory adaptation to hypoxia of awake dogs and rabbits. Respir. Phy,~'iol. 17:209 219. Bouverot, P. and M. Bureau (1975). Ventilatory acclimatization and csf acid-base balance in carotid chemodenervated dogs at 3550 m. Pfliigers Arch. 361 : 1%23. Bouverot, P. and P. S6bert (1979). O2-chemoreflexdrive of ventilation in awake birds at rest. Respir. Physiol. 37: 201-218.

196

A. UNGAR AND P. BOUVEROT

Cherniack, N.S., N.H. Edelman and S. Lahiri (1970/71). Hypoxia and hypercapnia as respiratory stimulants and depressants. Respir. Physiol. I 1 : 113 126. Cherniack, N. S., S.G. Kelsen and S. Lahiri (1976). The effects of hypoxia and hypercapnia on central nervous system output. In: Respiratory Adaptations, Capillary Exchange and Reflex Mechanisms, edited by A. S. Paintal and P. Gill-Kumar. Delhi, Vallabhbhai Patel Chest Institute, pp. 312-325. Cohen, M. I. and A. Hugelin 0965). Suprapontine reticular control of intrinsic respiratory mechanisms. Arch. Ital. Biol. 103: 317-334. Cohen, P. J., S.C. Alexander, T. C. Smith, M. Reivich and H. Wollman (1967). Effects of hypoxia and normocarbia on cerebral blood flow and metabolism in conscious man. J. Appl. Physiol. 23: 183-189. Dejours, P. (1957). Int6rEt m6thodologique de l'6tude d'un organisme vivant ~ la phase initiale de rupture d'un ~quilibre physiologique. C.R. Acad. Sci. (Paris) 245:1946 1948. Dejours, P., Y. Labrousse, J. Raynaud, F. Girard and A. Teillac (1958). Stimulus oxyg~ne de la ventilation au repos et au cours de l'exercice musculaire fi basse altitude (50 m) chez l'HorrOne. Rev. Fr. Etudes Clin. Biol. 3: 105-123. Dejours, P. (1962). Chemoreflexes in breathing. Physiol. Rev. 42: 335-358. Downes, J. J. and C.J. Lambertsen (1966). Dynamic characteristics of ventilatory depression in man on abrupt administration of 02. J. Appl. Physiol. 21 : 447-453. Euler, U.S. yon and G. Liljestrand (1936). Chemical stimulation of carotid sinus and the regulation of respiration. Skand. Arch. Physiol. 74: 101-128. Favier, R. and A. Lacaisse (1978). Stimulus oxygene de la ventilation chez le rat 6veill6. J. Physiol. (Paris) 74:411~17. Flandrois, R., J. R. Lacour and H. Osman (1971). Control of breathing in the exercising dog. Respir. Physiol. 13: 361-371. Grenvik, A., U. Hedstrand and H. Sj6gren (1966). Problems in pneumotachography. Acta Anaesth. Stand. 10:147 155. Jones, D. R. and M. J. Purves (1970). The effect of carotid body denervation upon the respiratory responses to hypoxia and hypercapnia in the duck. J. Physiol. (London) 211 : 295 309. Koepchen, H. P., D. Klussendorf, J. Borchert, D. W. Lessmann, A. Dinter, C. Frank and D. Sommer (1976). Types of respiratory neuronal activity pattern in reflex control of ventilation. In: Respiratory Adaptations, Capillary Exchange and Reflex Mechanism, edited by A. S. Paintal and P. Gill-Kumar. Delhi, Vallabhbhai Patel Chest Institute, pp. 291 311. Lahiri, S. (1976). Depressant effect of acute and chronic hypoxia on ventilation. In: Morphology and Mechanisms of Chemoreceptors, edited by A. S. Paintal. Delhi, Vallabhbhai Patel Chest Institute, pp. 138 145. Lee, K.D. and J.M. Bishop (1974). The reflex hypoxic respiratory drive in patients with chronic bronchitis. Clin. Sci. Mol. Med. 46: 347-356. Leitner, L. M., 8. Pag6s, R. Puccinelli and P. Dejours (1965). Etude simultan6e de la ventilation et des d6charges des ch~mor6cepteurs du glomus carotidien chez le chat. I. Au cours d'inhalations br~ves d'oxyg6ne pur. Arch, Int. Pharmacoc(vn. 154:421 426. Miller, M.J. and S. M. Tenney (1975). Hyperoxic hyperventilation in carotid-deafferented cats. Respi~: Physiol. 23:23 30. Nielsen, M. and H. Smith (1952). Studies on the regulation of respiration in acute hypoxia. Acta Physiol. Scand. 24: 293-313. Purves, M.J. (1966). The effect of a single breath of oxygen on respiration in the newborn lamb. Respir. Physiol. 1 : 297-307. Selladurai, S. and S. Wright (1932). Mode of action of respiratory stimulants. I. Mode of action of oxygen lack. Q. J. Exp. Physiol. 22: 233-248. Severinghaus, J.W., H. Chiodi, E.I. Eger, B. Brandstater and T.F. Hornbein (1966). Cerebral blood flow in man at high altitude. Circ. Res. 19: 274-282. St. John, W. M. and S. C. Wang (1977). Response of medullary respiratory neurons to hypercapnia and isocapnic hypoxia. J. Appl. Physiol. 43: 812-821.

ISOCAPN1C OXYGEN TESTS

197

Stockley, R. A. (1977). The estimation of the resting reflex hypoxic drive to respiration in normal man. Respir. Physiol. 31: 217-230. Tenney, S. M., P. Scotto, L.C. Ou, D. Bartlett and J. E. Remmers (1971). Suprapontine influences on hypoxic ventilatory control. In: High Altitude Physiology: Cardiac and Respiratory Aspects, edited by R. Porter and J. Knight. London, Churchill, pp. 89-97.