- Email: [email protected]
Influence of peripheral O2 tension on the ventilatory response to CO2 in cats
379
Rc,vfiration Phy,~iolo
I N F L U E N C E OF P E R I P H E R A L O. T E N S I O N ON THE V E N T l l A T O R Y R E S P O N S E TO CO~ IN CATS
J. H . G . M . V A N BEEK, A. B E R K E N B O S C H , J. DE G O E D E and C. N. OLIEVIER Leiden Universit.l M~'dica/ ('enlre, Dcparlmelll o! Phy~iolo.~,'.v am/P/{vsio/ov, ica/ Plo'vic.s. Leiden. T/w NeHwrlaml,~
Abstract. The effects of peripheral hypoxia on ventilation were investigated in 18 cats anaesthetized with chloralose urethane. The ponto-mcdullary region o f the brain was artificially perfused via a cannulated vertebral artery, using an extracorporeal circuit l"cd from a femoral artery. In this way the carbon dioxide tension (Pa)'ofl and the oxygen tension in /he blood supplying the brainstem could be imposed independently fiom the peripheral P(o: (Pa~'o:) and Po: (F'a/~:) in the systemic circulation, In all experiments the brainstem was kept hyperoxic. The steady-state ventilation ~/i could be described by Vl
=Sp.P~l~.o,+Sc-Pa(~o, - K
where Sp and Sc represent the peripheral and central sensitivity to carbon dioxide and K is a constanl. Sc and K were independent of the PAP,. In general, peripheral hypoxia increased and peripheral hyperoxia decreased Sp. compared to normoxia, h is concluded that: (I) there is no inleraction in the v e n t i l a t o r y r e s p o n s e b e t w e e n p e r i p h e r a l O 2 CO 2 a n d c e n t r a l C O 2 stimuli; and (2) a positivcinteraction in the ventilalory response between peripheral hypoxia and ( ' 0 2 originates from the arterial chemoreceptors. Central chemoreceptors Control of breathing Hypoxia
()~ and CO~ stimuhls interaction Peripheral chemoreceptors Ventilatory response to CO,
Hypoxia stimulates breathing in anaesthetized animals, but depresses ventilation after denervation of the arterial chemoreceptors (Lahiri, 1976). These findings suggest that there are effects of hypoxia at the level of the peripheral chemoreceptors and in the central nervous system. To understand the effects of hypoxia on ventilatory control it is necessary to know to what extent hypoxia influences central and peripheral CO_, responses. In afferent fibres from arterial chemoreceptors in cats an amplification of CO, ,4cc~7~led /br puhlicatiott 27 November 195'2 0034-5687/83/000-0000/$03.00 c~-~1983 Elsevier Biomedical Press
380
J.H. (;. M. VAN BEEK ~,t a/.
sensitivity by hypoxia has been l\mnd (Fitzgerald and Dehghani, 1982; Lahiri and DeLaney, 1975a), Giese e/ a/. (1978) did not find this amplification reflected in ventilatory responses of cats when stimulating the carotid bodies with low O, and high CO~. An interaction between stimuli acting oil central chemoreceptors and stimuli acting oil peripheral chemoreceptors could also play a role in determining the joint effect of O~ and CO~ on ventilation. A positive interaction between central and peripheral stimuli to breathing in anaesthetized cats has been reported by kahiri and DeLancy (1975b), but a negative interaction has been found by others (Gicse eta/., 1978; Berger eta/., 1978). An appropriate technique for investigating central and peripheral effects of O, and CO, oil ventilation is by artificial perfusion of the brainstem of the cat (Berkcnbosch e[ a/., 1979). The pcrfusion blood supplies a m o n g others the pons and medulla oblongata, where the central chemoreceptors and the integrating centres of the respiratory system are located. The O, and CO, tensions in this pcrfusion blood can be imposed independently from those in the systemic arterial blood, which is sensed by the peripheral arterial chemoreceptors. In this study we used tile artificial brainstem perfusion technique to investigate the effects of peripheral hypoxia. Direct hypoxic effects oil respiratory integrating centres and central chemosensitivc strnctures were excluded by keeping the pcrfusion blood hyperoxic. The principal aims were to investigate the dependence of peripheral and central CO, sensitivity oil peripheral hypoxia and the dependence of peripheral sensitivity on tile central level of stimulation with CO~. The effects of central hypoxia oll ventilation will be examined in subsequent studies. One of the main restllts of this study was that peripheral hypoxia enhanced the ventilatory sensitivity to peripheral CO~, but did not change the sensitivity to central CO~. Some o['these results have been presented in abstract form (van Beek et al.. 1981 ). Materials and methods
Experiments were performed oil 18 cats of both sexes weighing 2.7 to 4.3 kg.
SUR(ilCAI. PROCEDURES Anaesthesia was induced with 10 mg/kg ketamine intramuscularly, IBllowed by halothane or enflurane inhalation. After cannulation of the right femoral vein 20 mg/kg chloralose and 100 mg/kg urethane were slowly administered intravenously. Then the volatile anaesthetic was withdrawn. About tin hour later, infusion of a chloralosc urethane solution was started (1.5 i n g . k g ~-h ~ chloralose and 7.5 m g . k g t . h ~ urethane). A tracheal cannula was insertcd lbr the continuous measurement of ventilatory tlow and end-tidal P< <>.
P E R I P H E R A L HYPOXIA A N D V E N T I L A T I O N
381
The rectal temperature was measured with a thermistor and servo-controlled using a heating pad and an infrared lamp. The temperature range of all experiments was 36.8-38.4 °C. The right femoral artery was cannulated for monitoring blood pressure. The left femoral artery and vein were cannulated and an extracorporeal circuit (ECC), filled with Ringer's solution containing heparine, was fitted to provide a connection between the two cannulas. Blood from this femoral artery was pumped into the ECC and back to the femoral vein. After preparations in the neck region, a muscular branch of one vertebral artery was cannulated at the alar plate of the atlas. This vertebral artery was clamped on the proximal side of the cannulated branch and the other vertebral artery was also clamped. Blood from the femoral artery was then pumped via the ECC into the cannulated vertebral artery, The Pc <, and Po: of this blood (Pa{~,, and Pa)~:, also referred to as central CO, and central O, tension), could be imposed by means of an equilibrator, which was part of the ECC. These blood gas tensions were independent of those in the systemic circulation (Pa{io, and Pal;., also referred to as peripheral CO, and peripheral O, tension). At the usual infusion flow rate of 6 ml • min r, the pons, medulla oblongata and cerebellum are exclusively supplied with infused blood and there is an overflow into the circle of Willis (Berkenbosch el al., 1979). Details of the surgical procedures and the method of artificial brainstem perfusion have been described by Berkenbosch e t a / . (1979).
M E A S U R E M E N T OF V E N T I L A T O R Y RESPONSES
The tracheal gas flow was measured with a Fleisch 0 pneumotachograph, connected to the tracheal cammla. The flow signal was electronically integrated. CO, concentrations in the tracheal gas, arterial blood pressure, heart rate, perfusion pressure and rectal temperature were measured continuously. Blood gas tensions and pH of the systemic arterial blood and of the perfusion blood were also measured continuously with electrodes in the ECC. All signals were recorded on polygraphs and processed with a PDP 11/I0 minicomputer. For all signals averages over 20 breaths were calculated and stored. Details of the measurement procedures have been described by Berkenbosch el al. (1979). Ventilatory responses were recorded during artificial perfusion of tile pontomedullary region with a constant flow of hyperoxic blood from the equilibrator, which was supplied with a gas mixture of CO, and 601~,, O, in N, (CO, between 2 and 8?0). The cats inhaled a mixture of O, in N:, to which small flows of CO, and O, were added under manual control to maintain a desired arterial CO, level and O, level. For hyperoxic mixtures the O~ concentration was 60',,,, to attain hypoxic levels it was usually 10 to 12'~i,. Alter changing blood gas tensions, ventilation reached a near steady-state in 4 to 10 min for peripheral changes and in 10 to 20 min for central changes. To obtain
382
.I.H. (i. M. VAN BEEK cl
at.
some information on the magnitude of fluctuations around the steady-state level we determined two measuring points for each level, one near the maximum and the other near the minimum of ventilation.
ANAI,YSIS OF
I)AFA
Slopes and intercepts of response lilacs were estimated with linear regression analysis. Repeated measurements under different experimental conditions were compared with Student's paired t-test and the non-parametric Wilcoxon matched-pairs si-gnedranks test (Sokal and Rohlf, 1969). Since both tests yielded in essence the same results, only P values fi'om the paired t-test are given. Analysis of covariance was used to test for linear trends (Johnston, 1972).
pI~,OTO('OLS All experiments were perl\)rrned during central hyperoxia. Malay cats underwent more than one type of cxperimental protocol, of which there were 3: 1. The ventilatory response to changes in peripheral CO~ tension was determined at several levels o1" peripheral O, tension. This was done at one level o1" central CO~ tension in 4 cats. at 2 levels in 1 cat and at 3 levels in 3 cats. 11. The ventilatory response to changes in peripheral CO~ tension was determined at 2 to 6 central CO~ tensions in 11 cats. This was done during peripheral hyperoxia (Pal~ around 50 kPa) and hypoxia (Pa{'~, around 8 kPa). Ill. The ventilatory responsc to changes in central CO~ tension was determined during peripheral hyperoxia, hypoxia and occasionally during peripheral normoxia in 5 eats. The peripheral CO, tension was kept constant.
Results PROTOCOL 1. INFLUEN('EOF PERIPHERAt.()~ TENSION ON I'HE VENTILATORY RESPONSE TO PERIPHERAl, ('()~ In fig. 1 part of a recording of a peripheral COe response during peripheral hypoxia and hyperoxia is shown. At each oxygen level, ventilation increased with end-tidal CO,. The change from a hypoxic to a hyperoxic inspirate was followed by a large decrease in ventilation. The slopes of the CO~ response lines increased with a tall in peripheral O, tension as depicted in fig, 2. The ventilation X?E was described as a linear function of the peripheral CO, tension (Patio), vi:. X?~! = S p . Pa~'o, + K~
(1)
383
PERIPHERAL HYPOXIA AND VENTILATION
v, (~i}
t00
I
exp,
0.1
003
l FI Oz 0.6
.
FECOz[ 0 z,O
---m,, .--
arteriat
pressure {kPa) 250 hearf
rate (rain"1) ~0 perf
pressure IkPa)
0
Fig. 1. Ventilation at several levels of peripheral ('O~ tension as indicated by, k'l:((), during hypoxin (FI(), about 0.12: left of the arrow) and durii]g IDpcroxia Fio, a b o u t 0.6: right of the arrm~). At each oxygen level peripheral arterial oxygcn tension was kept constant by manipuhiting F i o . The central CO~ and O, tensions were 3. I and 50 kPa, respectivel.v.
~:xp 103 PJ,:, !kPit) -K/4"4/ " / "
6,1 o j / ~ 6"1
2 c-
0
0
5
10
pi,
~'co:, ( k P a )
Fig. 2. Vcntikttion ;is n function of P~ ~(), ~.tt several Pa~'): levels. Pa~(), ~as kept constant at 3.1) kPa. A I'irst series o1" measurements at Pa{'): = 6.1 kPa ( I ) ~ a s repeated at'let one hour (tEl). The measurel]]ClllS took two houl-s. Lilies dl-i.l~Al] ['lOll] linear regression LlllLI[ySiS.
although the responses during deep hypoxia were sometimes slightly convex (i.e., slope decreasing with increasing Patio,). In Eq. (1) Sp represents the peripheral CO, sensitivity and K, is a constant, depending on the central CO~ tension. The slope Sp and intercept K~ were estimated with linear regression analysis. There was no trend in K~ wilh the peripheral O, tension. Figure 3 shows Sp as a
384
,I. H. (i. M. VAN BEEK ~'z a/,
+
+
~+ +
E x
g 6
10
20
30
50
100
Pra~O~(kPa) Fig. 3. Linear log plot of normalized Sp (Sp divided b> mean Sp of Ihc cal at normoxia, i.e.. Pal' > 12 17 kPa) as a function of Pat'>. Each ,',ymbol represents a different cat. All results I'rom protocol l are plotled, lhc encircled Cll-C;i contain:, IO pohlts.
function o f Pal'~, for the 8 cats. The data for different central CO, levels were pooled, keeping in mind the finding that Sp does not depend on central CO, tension (see next section). For each cat Sp was normalized by dividing it by the mean value o f Sp of that cat at normoxia (Pa{~,, 12 17 kPa). On the averagc Sp increased when Pal',. decrcased. Whcn Pat',, was lowered frorn n o r m o x i a to hypoxia, ranging f l o r n 7 t o g k P a , the average increasc in Sp for t h e 8 cats w a s 0 . 1 2 L . m m ~-kPa (SE = 0.02, n = 10, P < 0.01): for hypoxia between 5 and 7 kPa the corresponding mean increase was 0.3 L - r a i n ' . k P a ~ ( S E = 0 . 1 , n = 8 , P < 0 . 0 5 ) . Only 2 of 33 Sp's were lower during peripheral hypoxia than during normoxia. The mean Sp decreased from 0.17 L - r a i n ' . k P a ~ at normoxia to 0.11 L . rain ~ - k P a ' ;,it hyperoxia (n = 12. P <0.01). Two very small negative vahies of Sp wcre measured tit hypcroxia.
I1. INFt.UENCt!OF I H E ( E N T R A L ('O~ RESPONSI! DURING ftYPEROXIA AND HYPOXIA
PROFO('OI.
T E N S I O N ON I H E P E R I P H E R A L ( O ,
A typical result from such an experiment is shown in fig. 4. The peripheral CO, responses in this figure were measured at peripheral O~ tensions maintained between 8.0 and 8.5 kPa. Each line was determined at a dilTerent central CO, level. The slopes o f the response lines were about the same, but the intercepts were strongly dependent on the central CO, tension. An analysis o f covariancc was perfornled to test Its1 linear trends in Sp with central CO~ tension, allowing each cat to have its own direction and magnitude o f trend (Johnston, 1972). Peripheral CO, sensitivity was not significantly inl]uenced by central CO, tension during peripheral hypoxia and hyperoxia (n = 11, P > 0.25). Furthermore, there was no systernalic change in K~ with a change in peripheral O~ tension from hyperoxia to Pal',, a r o u n d 8 kPa (n - 4 0 , P > 0.6).
PERIPHERAL HYPOXIA AND VENTILATION
385
e × p 003
c ( P~c% kPa) ~-2 20 .
c
1.0
L
3
_ _ L ~ _
L,
I
I
5
6
P~Pco: (kPa)
Fig. 4. Ventilation as a function of PaP.o, at three levels of Pa{o: Data for one cat rccordcd within two houl-s. Pa{~), \'¢:.ls8.t_ kPa. The points lor P'd(c. : c = 3.1 kPa arc from the slcad3,,-stalcs in fig. I.
PROTO(OL 111. INFLUENCE OF PERIPHERAL O, TENSION ON THE VENTII,ATORY P,ESPONSE TO CENTRAL CO~ Results of such an e x p e r i m e n t are shown in fig. 5. The slopes of these responses d u r i n g peripheral hypoxia a n d hyperoxia were a b o u t the same. In 2 of the 5 cats s o m e w h a t convex response lines were f o u n d (i.e., the lines become flatter as Pa{o: increases). However, in all 5 cats the response curve was linear to a good approxim a t i o n for the Pa[o ~ range from 2.5 t o 5 kPa. The linear regression estimates for the slope Sc of the response curves are given in table I. The hypothesis that this slope does n o t depend on peripheral s t i m u l a t i o n with hypoxia is considered in the next section, allowing for a d d i t i o n a l Sc pairs o b t a i n e d fl'om protocol 1 a n d 1I data. For this purpose we shall develop a simple m a t h e m a t i c a l model in the next section for describing the response to peripheral a n d central CO,. F r o m this we can estimate Sc when the central CO2 response was not directly measured.
MATHEMATICAL MODEL FOR VENTILATORY RESPONSE TO CO, AND DEPENDENCE OF ITS PARAMETERS ON PERIPHERAL O, IENSION Since the peripheral CO: sensitivity Sp is i n d e p e n d e n t of the central CO~ tension a n d the ventilation ~?E is to a first a p p r o x i m a t i o n a linear function of the central COt tension we can write for the intercept o f Eq. (1)
386
.I.H.(i.
M. \ " A N B E E K {,I a/.
exp 0 2 4 P~O:, ( k P a )
7.8
2.0 /
/
E 10
0
,
53/ /
4
2
6
P',I cO, ( k Pa) Fig. 5. Vcntilatioll
as tl ftLt~ction Of Pa~l > at t.'ot/stant f'ci{'o: and Pal+),. A t each l:'a~ ¢> tv, o levels Ior
peripheral ox_',,gcn wore, imposed.
Data l'or one cat recorded for one and ci quarter hours. Pa~> ',',as
49 k P a . • : p e r i p h e r a l h y p c r o x i a . P ~I'>= 53 k P a , l:'ai~<) = 7.5 k P a : • : p e r i p h e r a l h,q-~oxia, l'al'~ = 7.8 kF'a.
Pa['{i. - 7.0 kl)a.
1AI}IE
I
Central CO~ sensitivity. No. for ditTcrcni levels o f peripheral cu'Icl-ial oxygen Icnsion (l>a{'>) :.it conslaill peripheral arterial C O . lcnsion (Pal~o,). * indicates a convex i-cspollSO line in whicl~ case the range o f [:'ci{ o s c l e c i c d for l i n e a r r e g r e s s i o n w a s ttlkcll l'ronl 2.5 It) 5 k P a Expt. no.
O24
Pai ~~)
i k I'a I
Pal'> (kPa)
Sc (ml • rain
7.2
53
881
7.8 O42
7.3
28 4.S
862 41)4
*
171
*
O46
7.2
52
107
S. I
55
126
13.5
147
(~.~)
134
6.8
Ills
(~.S
668
52
936
13. I
<'45(~ ( ~0 *
7.9
7.7
588
4~,I
774
*
I 3.2
(~77
*
g.(I
3")9
*
i . kPa
II
PERIPHERAL HYPOXIA AND VENTILATION K 1 = Sc • PaSo. - K
387 (2)
where Sc is the central CO2 sensitivity and K is a constant. If we insert Eq. (2) into Eq. (1) we arrive at ~?E = Sp • PaP.o, + Sc • PaSo, - K
(3)
Eq. (3) was used by Heeringa et al. (1979), but only for overall hyperoxic conditions. From all protocol I and I1 experiments in which the peripheral CO_~ response was measured at more than one central CO, tension we could estimate the parameters Sc, Sp and K at each peripheral oxygen level with multiple regression analysis. The estimates are collected in Table 2. Sp decreases again with O, tension, but there is no systematic change in Sc and K. The difference in Sc between hyperoxia and hypoxia in tables 1 and 2 was not significant (n = 20, 0.20 < P < 0.40). The change in K between Pa{?,, around 50 kPa and around 8 kPa was not significant (see table 2, n = 11, P > 0.6), but the change in Sp was highly significant (n = 11, P < 0.01).
Discussion Our restllts could to a first approximation be described by Eq. (3). However, it is clear that the linearity of Eq. (3) breaks down if the ventilatory response to central and peripheral CO2 approaches the maximum breathing capacity of" an animal. Eq. (3) is therefore only valid for moderate stimuli. Under our experimental conditions with ventilations usually under 2.5 L • min ~ (about 0.8 L. min ~ - kg ~), deviations from linearity were generally small. We found that stimulation of the peripheral chemoreceptors with O, and CO, did not alter the central ventilatory sensitivity to CO,. Furthermore, we established that Sp, the peripheral CO~ sensitivity, did not change with the level of the central stimulation by CO,. Therefore, the responses to stimulation of breathing with central CO~ and with peripheral O, CO~ combinations are additive. This is reflected in Eq. (3) by the absence of the product term for central and peripheral CO, tensions and the finding that Sc is independent of the peripheral O, tension. Like the linearity of Eq. (3), additivity will break down if the combined responses to peripheral and central stimuli approach the maximum breathing capacity. Using the technique of artificial brainstem perfusion, Riedstra (1963) measured peripheral O_~ responses at several base excess levels and central CO, tensions. He concluded that the central CO~ influenced the critical value below which ventilation reacted to changes in peripheral O, tension. Lloyd (1966) has argued that Riedstra's data were consistent with addition of peripheral O, stimulation to central CO, stimulation, as we also found in this study, and that Riedstra's conclusion was due to his unconventional introduction of a critical Po:. We share Lloyd's view. Indeed, the ventilatory response to peripheral O~ does not manifestly show such a critical Po:. It should also be noticed that ventilatory responses are influenced by the base
388
J H. G. M. VAN
BEEK
el a/.
T A B L li "~
C c n l r a l CO, scnsilivity (So). p e r i p h e r a l ('O 2 sensitivity (Sp) :.llld K e s t i m a t e d by multiple regression allalysis Expi. no.
721
Body 'acight (kg)
Pal'), (kPa)
3.7
39
Sc (rill.rain
i .kPa
I)
Sp (ml.miu
i .kpa
i)
K (ml.min
643
.~_ "~ 8
.:;_"%1_~
7.6
( ~__
~01
2446
938
3.9
55 14.6 9.7
1023 1184 1080
172 277 483
4840 5581 52115
OOl
3. I
51) 8.2
I 185 791
102 155
5002 3245
1103
2.9
50 8.2
920 867
171 264
3568 2703
004
4.3
53 8. l
803 978
101 379
28211 3637
1114
3.7
53 8.4
452 468
49 395
15711 2503
1118
3.7
61 7.6
253 167
7{) 325
966 723
(119
3.9
57 7.7
419 249
50 I 12
1154 289
11211
3.5
57 7.7
631 855
I I0 358
3149 4622
022
3.9
55 7.5
457 4~
79 164
1710 1271
1124 *
3.4
57 7.6
1528 1664
l 12 68il
6494 8157
1/13
2.7
50 13.4 8.1i 6. I
459 483 111()1 383
39 75 253 342
I lfl I 876 222O 377
105
3.7
54 13.3 6.0
851 686 I 1)
141 169 682
3869 2839 2681
* The I-csull of cxpi. 024 given in l'ablc 1 was from an earlier slate in this cXpOl'ilricnl and ~,as not included in the Illultipic regression amll',.sis.
excess which was around 0 mmol/L in Riedstra's experiments and - 7 lnmol/L ( S D = 2 mmol/L, range - l l . 0 t o - 1 . 7 mmol/L) in ours. Lahiri and DeLaney (1975b) concluded from the restllts of their experiments oil cats that there is a positive interaction between the peripheral and central drive to ventilation. In order to demonstrate such an interaction they had to apply "overall"
PERIPHERAL HYPOXIA AND VENTILATION
389
hypoxia, while in our experiments the brainstem was always supplied with hyperoxic blood. It may well be that an interaction between central and peripheral stimuli to breathing is brought about by hypoxia in the brainstem. In addition, there may be an interaction between central hypoxia and central hypercapnia. Berger el ell. (1978) and Giese e l al. (1978) concluded that central and peripheral stimuli showed negative interaction (hypoadditivity). It should be noted that ventilations often approached or exceeded 0.8 L . rain ' • kg ~ m their experiments. We limited ourselves to ventilations below this value and our conclusions only pertain to these lower levels of ventilation. Therefore, their results and ours are not necessarily in contradiction. Fitzgerald and Dehghani (1982) and Lahiri and DeLaney (1975a) measured neural signals in the sinus nerve while applying different CO, and O, tensions in the arterial blood. There was sensitivity to CO, at hyperoxia and this sensitivity increased when the arterial O, tension was lowered to normoxic and hypoxic values. Aortic bodies also responded to increases of COn and their responses showed positive interaction between O, and CO~ (Fitzgerald and Dehghani, 1982: Lahiri el al., 1979). We measured a similar pattern in the ventilation when applying combined O, CO, stimuli peripherally, keeping the central CO, constant (see figs. 2 and 3). As shown by Fitzgerald and Dehghani (1982) the peripheral chemoreceptor responses to CO, become convex at deeper levels of hypoxia, but they still respond to lowering of the Po~. Our ventilatory peripheral CO_, responses also tended to be convex during severe hypoxia (see fig. 2). This saturation of the response to CO, of the peripheral chemoreceptors could explain why we found lower Sp values during severe hypoxia than during milder hypoxia in some cases (see fig. 3). It may also be the reason why Giese er al. (1978) did not find a positive interaction in ventilation when they stimulated the carotid body, since their levels of hypoxia were in general deeper than the ones we used. However, such phenomena could also reflect saturation in other parts of the peripheral chemoreflex arc. The available experimental results suggest strongly that the positive interaction between the peripheral O, and CQ, stimuli, as reflected in the ventilatory response, takes place at the arterial chemoreceptors and can already be measured in the neural signals carried in their afferents. These signals constitute a peripheral drive, which is combined with the central CO, drive in the integrating centres of the respiratory control system. The neural output of these centres is translated into ventilation. At the level of ventilatory output, the peripheral and central drives are combined additively.
Acknowledgements This research was supported by the Foundation for Medical Research F U N G O , which is subsidized by the Netherlands Organization for the Advancement of Pure Research (Z.W.O.). We are indebted to L. Philips for his skilful work in the surgical
390
J . H . G . M . VAN BEEK
el
a/.
preparation of the animals, Dr. M . E . Wise for advice on statistics and the text and J. H. van der Perk and M. M. Rancuret for technical assistance.
References Beck, J. H. G. M. van, A. Berkenbosch, J. de Goede and C. N. Olievier ( 1981 ). Peripheral hypoxaemia and ventilatory response to C()~ in cats. Era..I. R~,slm.. Di.~. 62 (Suppl.): 146 147. Berger, W., K. Bcrger, J. Berndt and K. Giese (1978h Interaction of peripheral and central respiratory drives in eats. 1. Effects of sodium cyanide as a peripheral chemoreceptor stimulus at difl'erent levels ot'CSF pH. pllii~,,er's Arch. 374:205 210. Berkenbosch. A., J. Heeringa. C N. Olicvier and E.W. Kruyl (1979). Artificial perfusion of lhe ponto-medullary region of cats. A method for separation of central and peripheral effects of chemical slimnlation of ventilation. Respir. Ptn'.~iol. 37:347 364. Fitzgerald, R.S. and G . A . Dehghani (1982). Neural responses of the cat carotid and aortic bodies to hypercaprda and hypoxia..I. AppI. t'ln'siol. 52:596 ~01. (;icse, K.. J. Berndl and W. Bcrgcr (1978). lnlcraction of central and peripheral drives in cats. 11. Peripheral and central interaction of hypoxia and hypercapnia, p/lii,k, er's Arch. 374:21 I 217. Heeringa. ,l., A. Berkenbosch, J. de (}ocde and C. N. Olievier (1979). Relative contributions of central and peripheral chemoreceplors to Ihe ventilatory response to ('O 2 during hyperoxia. Re.~;:~ir. Phlsiol. 37:365 379. Johnston, J. (1972). Econometric Methods. Tokyo, McGrax~-Hill Kogakusha, pp. 192 207. Lahiri, S. and R.G. DeLancy (1975a). Stinaulus inleraction in the responses of carotid body chemoreceptor single afferent fibres. Re.~pir. Physiol. 24:249 266. Lahiri. S. and R . G . DeLancy (1975b). Relationship between carolid chemoreceptor activity and ventilation in tile cat. Re.spir. P/nsiol. 24:267 286. Lahiri, S. (1t)76). Depressant effect of acute and chronic hypoxia in ventilation. In: Morphology and Mechanisms of Chemoreceplors, cditcd by' A. S. Paintal. Delhi. Vallabhbhai Patel ('hest Instinlle, pp. 138 146. Lahiri. S., E. Mnlligan, r. Nishino and A. Mokashi (197t)). Aortic body chcnlorcccptor responses to changes in P ~ o : a n d P o : i n l h e c a l . J. Apl~l. Phl,~'iol. 47:S58 g66. Lloyd. B.B. (It)66). The interactions between hypoxia and other ventilator> stimuli. In: Proc. Int. Syrup. Cardiowlsc. Rcspir. Effects of Hypoxia, edited by J. D. Itatcher and D. B. Jennings. Basle, Karger, pp. 146 165. Riedstra. J.W. (1963). Influence ol'ccnlral and peripheral P{o: (pH) on tile vcntilatory response to hypoxic chemoreceptor stimulalion, t o : , Pk.rsiol. Pharma('ol. Need. 12:407 452. Sokal. R. R. ~lnd F.J. Rohlf (1%9). Biometry. San Francisco, Frccman, pp. 3~)t) 401.
Rc,vfiration Phy,~iolo
I N F L U E N C E OF P E R I P H E R A L O. T E N S I O N ON THE V E N T l l A T O R Y R E S P O N S E TO CO~ IN CATS
J. H . G . M . V A N BEEK, A. B E R K E N B O S C H , J. DE G O E D E and C. N. OLIEVIER Leiden Universit.l M~'dica/ ('enlre, Dcparlmelll o! Phy~iolo.~,'.v am/P/{vsio/ov, ica/ Plo'vic.s. Leiden. T/w NeHwrlaml,~
Abstract. The effects of peripheral hypoxia on ventilation were investigated in 18 cats anaesthetized with chloralose urethane. The ponto-mcdullary region o f the brain was artificially perfused via a cannulated vertebral artery, using an extracorporeal circuit l"cd from a femoral artery. In this way the carbon dioxide tension (Pa)'ofl and the oxygen tension in /he blood supplying the brainstem could be imposed independently fiom the peripheral P(o: (Pa~'o:) and Po: (F'a/~:) in the systemic circulation, In all experiments the brainstem was kept hyperoxic. The steady-state ventilation ~/i could be described by Vl
=Sp.P~l~.o,+Sc-Pa(~o, - K
where Sp and Sc represent the peripheral and central sensitivity to carbon dioxide and K is a constanl. Sc and K were independent of the PAP,. In general, peripheral hypoxia increased and peripheral hyperoxia decreased Sp. compared to normoxia, h is concluded that: (I) there is no inleraction in the v e n t i l a t o r y r e s p o n s e b e t w e e n p e r i p h e r a l O 2 CO 2 a n d c e n t r a l C O 2 stimuli; and (2) a positivcinteraction in the ventilalory response between peripheral hypoxia and ( ' 0 2 originates from the arterial chemoreceptors. Central chemoreceptors Control of breathing Hypoxia
()~ and CO~ stimuhls interaction Peripheral chemoreceptors Ventilatory response to CO,
Hypoxia stimulates breathing in anaesthetized animals, but depresses ventilation after denervation of the arterial chemoreceptors (Lahiri, 1976). These findings suggest that there are effects of hypoxia at the level of the peripheral chemoreceptors and in the central nervous system. To understand the effects of hypoxia on ventilatory control it is necessary to know to what extent hypoxia influences central and peripheral CO_, responses. In afferent fibres from arterial chemoreceptors in cats an amplification of CO, ,4cc~7~led /br puhlicatiott 27 November 195'2 0034-5687/83/000-0000/$03.00 c~-~1983 Elsevier Biomedical Press
380
J.H. (;. M. VAN BEEK ~,t a/.
sensitivity by hypoxia has been l\mnd (Fitzgerald and Dehghani, 1982; Lahiri and DeLaney, 1975a), Giese e/ a/. (1978) did not find this amplification reflected in ventilatory responses of cats when stimulating the carotid bodies with low O, and high CO~. An interaction between stimuli acting oil central chemoreceptors and stimuli acting oil peripheral chemoreceptors could also play a role in determining the joint effect of O~ and CO~ on ventilation. A positive interaction between central and peripheral stimuli to breathing in anaesthetized cats has been reported by kahiri and DeLancy (1975b), but a negative interaction has been found by others (Gicse eta/., 1978; Berger eta/., 1978). An appropriate technique for investigating central and peripheral effects of O, and CO, oil ventilation is by artificial perfusion of the brainstem of the cat (Berkcnbosch e[ a/., 1979). The pcrfusion blood supplies a m o n g others the pons and medulla oblongata, where the central chemoreceptors and the integrating centres of the respiratory system are located. The O, and CO, tensions in this pcrfusion blood can be imposed independently from those in the systemic arterial blood, which is sensed by the peripheral arterial chemoreceptors. In this study we used tile artificial brainstem perfusion technique to investigate the effects of peripheral hypoxia. Direct hypoxic effects oil respiratory integrating centres and central chemosensitivc strnctures were excluded by keeping the pcrfusion blood hyperoxic. The principal aims were to investigate the dependence of peripheral and central CO, sensitivity oil peripheral hypoxia and the dependence of peripheral sensitivity on tile central level of stimulation with CO~. The effects of central hypoxia oll ventilation will be examined in subsequent studies. One of the main restllts of this study was that peripheral hypoxia enhanced the ventilatory sensitivity to peripheral CO~, but did not change the sensitivity to central CO~. Some o['these results have been presented in abstract form (van Beek et al.. 1981 ). Materials and methods
Experiments were performed oil 18 cats of both sexes weighing 2.7 to 4.3 kg.
SUR(ilCAI. PROCEDURES Anaesthesia was induced with 10 mg/kg ketamine intramuscularly, IBllowed by halothane or enflurane inhalation. After cannulation of the right femoral vein 20 mg/kg chloralose and 100 mg/kg urethane were slowly administered intravenously. Then the volatile anaesthetic was withdrawn. About tin hour later, infusion of a chloralosc urethane solution was started (1.5 i n g . k g ~-h ~ chloralose and 7.5 m g . k g t . h ~ urethane). A tracheal cannula was insertcd lbr the continuous measurement of ventilatory tlow and end-tidal P< <>.
P E R I P H E R A L HYPOXIA A N D V E N T I L A T I O N
381
The rectal temperature was measured with a thermistor and servo-controlled using a heating pad and an infrared lamp. The temperature range of all experiments was 36.8-38.4 °C. The right femoral artery was cannulated for monitoring blood pressure. The left femoral artery and vein were cannulated and an extracorporeal circuit (ECC), filled with Ringer's solution containing heparine, was fitted to provide a connection between the two cannulas. Blood from this femoral artery was pumped into the ECC and back to the femoral vein. After preparations in the neck region, a muscular branch of one vertebral artery was cannulated at the alar plate of the atlas. This vertebral artery was clamped on the proximal side of the cannulated branch and the other vertebral artery was also clamped. Blood from the femoral artery was then pumped via the ECC into the cannulated vertebral artery, The Pc <, and Po: of this blood (Pa{~,, and Pa)~:, also referred to as central CO, and central O, tension), could be imposed by means of an equilibrator, which was part of the ECC. These blood gas tensions were independent of those in the systemic circulation (Pa{io, and Pal;., also referred to as peripheral CO, and peripheral O, tension). At the usual infusion flow rate of 6 ml • min r, the pons, medulla oblongata and cerebellum are exclusively supplied with infused blood and there is an overflow into the circle of Willis (Berkenbosch el al., 1979). Details of the surgical procedures and the method of artificial brainstem perfusion have been described by Berkenbosch e t a / . (1979).
M E A S U R E M E N T OF V E N T I L A T O R Y RESPONSES
The tracheal gas flow was measured with a Fleisch 0 pneumotachograph, connected to the tracheal cammla. The flow signal was electronically integrated. CO, concentrations in the tracheal gas, arterial blood pressure, heart rate, perfusion pressure and rectal temperature were measured continuously. Blood gas tensions and pH of the systemic arterial blood and of the perfusion blood were also measured continuously with electrodes in the ECC. All signals were recorded on polygraphs and processed with a PDP 11/I0 minicomputer. For all signals averages over 20 breaths were calculated and stored. Details of the measurement procedures have been described by Berkenbosch el al. (1979). Ventilatory responses were recorded during artificial perfusion of tile pontomedullary region with a constant flow of hyperoxic blood from the equilibrator, which was supplied with a gas mixture of CO, and 601~,, O, in N, (CO, between 2 and 8?0). The cats inhaled a mixture of O, in N:, to which small flows of CO, and O, were added under manual control to maintain a desired arterial CO, level and O, level. For hyperoxic mixtures the O~ concentration was 60',,,, to attain hypoxic levels it was usually 10 to 12'~i,. Alter changing blood gas tensions, ventilation reached a near steady-state in 4 to 10 min for peripheral changes and in 10 to 20 min for central changes. To obtain
382
.I.H. (i. M. VAN BEEK cl
at.
some information on the magnitude of fluctuations around the steady-state level we determined two measuring points for each level, one near the maximum and the other near the minimum of ventilation.
ANAI,YSIS OF
I)AFA
Slopes and intercepts of response lilacs were estimated with linear regression analysis. Repeated measurements under different experimental conditions were compared with Student's paired t-test and the non-parametric Wilcoxon matched-pairs si-gnedranks test (Sokal and Rohlf, 1969). Since both tests yielded in essence the same results, only P values fi'om the paired t-test are given. Analysis of covariance was used to test for linear trends (Johnston, 1972).
pI~,OTO('OLS All experiments were perl\)rrned during central hyperoxia. Malay cats underwent more than one type of cxperimental protocol, of which there were 3: 1. The ventilatory response to changes in peripheral CO~ tension was determined at several levels o1" peripheral O, tension. This was done at one level o1" central CO~ tension in 4 cats. at 2 levels in 1 cat and at 3 levels in 3 cats. 11. The ventilatory response to changes in peripheral CO~ tension was determined at 2 to 6 central CO~ tensions in 11 cats. This was done during peripheral hyperoxia (Pal~ around 50 kPa) and hypoxia (Pa{'~, around 8 kPa). Ill. The ventilatory responsc to changes in central CO~ tension was determined during peripheral hyperoxia, hypoxia and occasionally during peripheral normoxia in 5 eats. The peripheral CO, tension was kept constant.
Results PROTOCOL 1. INFLUEN('EOF PERIPHERAt.()~ TENSION ON I'HE VENTILATORY RESPONSE TO PERIPHERAl, ('()~ In fig. 1 part of a recording of a peripheral COe response during peripheral hypoxia and hyperoxia is shown. At each oxygen level, ventilation increased with end-tidal CO,. The change from a hypoxic to a hyperoxic inspirate was followed by a large decrease in ventilation. The slopes of the CO~ response lines increased with a tall in peripheral O, tension as depicted in fig, 2. The ventilation X?E was described as a linear function of the peripheral CO, tension (Patio), vi:. X?~! = S p . Pa~'o, + K~
(1)
383
PERIPHERAL HYPOXIA AND VENTILATION
v, (~i}
t00
I
exp,
0.1
003
l FI Oz 0.6
.
FECOz[ 0 z,O
---m,, .--
arteriat
pressure {kPa) 250 hearf
rate (rain"1) ~0 perf
pressure IkPa)
0
Fig. 1. Ventilation at several levels of peripheral ('O~ tension as indicated by, k'l:((), during hypoxin (FI(), about 0.12: left of the arrow) and durii]g IDpcroxia Fio, a b o u t 0.6: right of the arrm~). At each oxygen level peripheral arterial oxygcn tension was kept constant by manipuhiting F i o . The central CO~ and O, tensions were 3. I and 50 kPa, respectivel.v.
~:xp 103 PJ,:, !kPit) -K/4"4/ " / "
6,1 o j / ~ 6"1
2 c-
0
0
5
10
pi,
~'co:, ( k P a )
Fig. 2. Vcntikttion ;is n function of P~ ~(), ~.tt several Pa~'): levels. Pa~(), ~as kept constant at 3.1) kPa. A I'irst series o1" measurements at Pa{'): = 6.1 kPa ( I ) ~ a s repeated at'let one hour (tEl). The measurel]]ClllS took two houl-s. Lilies dl-i.l~Al] ['lOll] linear regression LlllLI[ySiS.
although the responses during deep hypoxia were sometimes slightly convex (i.e., slope decreasing with increasing Patio,). In Eq. (1) Sp represents the peripheral CO, sensitivity and K, is a constant, depending on the central CO~ tension. The slope Sp and intercept K~ were estimated with linear regression analysis. There was no trend in K~ wilh the peripheral O, tension. Figure 3 shows Sp as a
384
,I. H. (i. M. VAN BEEK ~'z a/,
+
+
~+ +
E x
g 6
10
20
30
50
100
Pra~O~(kPa) Fig. 3. Linear log plot of normalized Sp (Sp divided b> mean Sp of Ihc cal at normoxia, i.e.. Pal' > 12 17 kPa) as a function of Pat'>. Each ,',ymbol represents a different cat. All results I'rom protocol l are plotled, lhc encircled Cll-C;i contain:, IO pohlts.
function o f Pal'~, for the 8 cats. The data for different central CO, levels were pooled, keeping in mind the finding that Sp does not depend on central CO, tension (see next section). For each cat Sp was normalized by dividing it by the mean value o f Sp of that cat at normoxia (Pa{~,, 12 17 kPa). On the averagc Sp increased when Pal',. decrcased. Whcn Pat',, was lowered frorn n o r m o x i a to hypoxia, ranging f l o r n 7 t o g k P a , the average increasc in Sp for t h e 8 cats w a s 0 . 1 2 L . m m ~-kPa (SE = 0.02, n = 10, P < 0.01): for hypoxia between 5 and 7 kPa the corresponding mean increase was 0.3 L - r a i n ' . k P a ~ ( S E = 0 . 1 , n = 8 , P < 0 . 0 5 ) . Only 2 of 33 Sp's were lower during peripheral hypoxia than during normoxia. The mean Sp decreased from 0.17 L - r a i n ' . k P a ~ at normoxia to 0.11 L . rain ~ - k P a ' ;,it hyperoxia (n = 12. P <0.01). Two very small negative vahies of Sp wcre measured tit hypcroxia.
I1. INFt.UENCt!OF I H E ( E N T R A L ('O~ RESPONSI! DURING ftYPEROXIA AND HYPOXIA
PROFO('OI.
T E N S I O N ON I H E P E R I P H E R A L ( O ,
A typical result from such an experiment is shown in fig. 4. The peripheral CO, responses in this figure were measured at peripheral O~ tensions maintained between 8.0 and 8.5 kPa. Each line was determined at a dilTerent central CO, level. The slopes o f the response lines were about the same, but the intercepts were strongly dependent on the central CO, tension. An analysis o f covariancc was perfornled to test Its1 linear trends in Sp with central CO~ tension, allowing each cat to have its own direction and magnitude o f trend (Johnston, 1972). Peripheral CO, sensitivity was not significantly inl]uenced by central CO, tension during peripheral hypoxia and hyperoxia (n = 11, P > 0.25). Furthermore, there was no systernalic change in K~ with a change in peripheral O~ tension from hyperoxia to Pal',, a r o u n d 8 kPa (n - 4 0 , P > 0.6).
PERIPHERAL HYPOXIA AND VENTILATION
385
e × p 003
c ( P~c% kPa) ~-2 20 .
c
1.0
L
3
_ _ L ~ _
L,
I
I
5
6
P~Pco: (kPa)
Fig. 4. Ventilation as a function of PaP.o, at three levels of Pa{o: Data for one cat rccordcd within two houl-s. Pa{~), \'¢:.ls8.t_ kPa. The points lor P'd(c. : c = 3.1 kPa arc from the slcad3,,-stalcs in fig. I.
PROTO(OL 111. INFLUENCE OF PERIPHERAL O, TENSION ON THE VENTII,ATORY P,ESPONSE TO CENTRAL CO~ Results of such an e x p e r i m e n t are shown in fig. 5. The slopes of these responses d u r i n g peripheral hypoxia a n d hyperoxia were a b o u t the same. In 2 of the 5 cats s o m e w h a t convex response lines were f o u n d (i.e., the lines become flatter as Pa{o: increases). However, in all 5 cats the response curve was linear to a good approxim a t i o n for the Pa[o ~ range from 2.5 t o 5 kPa. The linear regression estimates for the slope Sc of the response curves are given in table I. The hypothesis that this slope does n o t depend on peripheral s t i m u l a t i o n with hypoxia is considered in the next section, allowing for a d d i t i o n a l Sc pairs o b t a i n e d fl'om protocol 1 a n d 1I data. For this purpose we shall develop a simple m a t h e m a t i c a l model in the next section for describing the response to peripheral a n d central CO,. F r o m this we can estimate Sc when the central CO2 response was not directly measured.
MATHEMATICAL MODEL FOR VENTILATORY RESPONSE TO CO, AND DEPENDENCE OF ITS PARAMETERS ON PERIPHERAL O, IENSION Since the peripheral CO: sensitivity Sp is i n d e p e n d e n t of the central CO~ tension a n d the ventilation ~?E is to a first a p p r o x i m a t i o n a linear function of the central COt tension we can write for the intercept o f Eq. (1)
386
.I.H.(i.
M. \ " A N B E E K {,I a/.
exp 0 2 4 P~O:, ( k P a )
7.8
2.0 /
/
E 10
0
,
53/ /
4
2
6
P',I cO, ( k Pa) Fig. 5. Vcntilatioll
as tl ftLt~ction Of Pa~l > at t.'ot/stant f'ci{'o: and Pal+),. A t each l:'a~ ¢> tv, o levels Ior
peripheral ox_',,gcn wore, imposed.
Data l'or one cat recorded for one and ci quarter hours. Pa~> ',',as
49 k P a . • : p e r i p h e r a l h y p c r o x i a . P ~I'>= 53 k P a , l:'ai~<) = 7.5 k P a : • : p e r i p h e r a l h,q-~oxia, l'al'~ = 7.8 kF'a.
Pa['{i. - 7.0 kl)a.
1AI}IE
I
Central CO~ sensitivity. No. for ditTcrcni levels o f peripheral cu'Icl-ial oxygen Icnsion (l>a{'>) :.it conslaill peripheral arterial C O . lcnsion (Pal~o,). * indicates a convex i-cspollSO line in whicl~ case the range o f [:'ci{ o s c l e c i c d for l i n e a r r e g r e s s i o n w a s ttlkcll l'ronl 2.5 It) 5 k P a Expt. no.
O24
Pai ~~)
i k I'a I
Pal'> (kPa)
Sc (ml • rain
7.2
53
881
7.8 O42
7.3
28 4.S
862 41)4
*
171
*
O46
7.2
52
107
S. I
55
126
13.5
147
(~.~)
134
6.8
Ills
(~.S
668
52
936
13. I
<'45(~ ( ~0 *
7.9
7.7
588
4~,I
774
*
I 3.2
(~77
*
g.(I
3")9
*
i . kPa
II
PERIPHERAL HYPOXIA AND VENTILATION K 1 = Sc • PaSo. - K
387 (2)
where Sc is the central CO2 sensitivity and K is a constant. If we insert Eq. (2) into Eq. (1) we arrive at ~?E = Sp • PaP.o, + Sc • PaSo, - K
(3)
Eq. (3) was used by Heeringa et al. (1979), but only for overall hyperoxic conditions. From all protocol I and I1 experiments in which the peripheral CO_~ response was measured at more than one central CO, tension we could estimate the parameters Sc, Sp and K at each peripheral oxygen level with multiple regression analysis. The estimates are collected in Table 2. Sp decreases again with O, tension, but there is no systematic change in Sc and K. The difference in Sc between hyperoxia and hypoxia in tables 1 and 2 was not significant (n = 20, 0.20 < P < 0.40). The change in K between Pa{?,, around 50 kPa and around 8 kPa was not significant (see table 2, n = 11, P > 0.6), but the change in Sp was highly significant (n = 11, P < 0.01).
Discussion Our restllts could to a first approximation be described by Eq. (3). However, it is clear that the linearity of Eq. (3) breaks down if the ventilatory response to central and peripheral CO2 approaches the maximum breathing capacity of" an animal. Eq. (3) is therefore only valid for moderate stimuli. Under our experimental conditions with ventilations usually under 2.5 L • min ~ (about 0.8 L. min ~ - kg ~), deviations from linearity were generally small. We found that stimulation of the peripheral chemoreceptors with O, and CO, did not alter the central ventilatory sensitivity to CO,. Furthermore, we established that Sp, the peripheral CO~ sensitivity, did not change with the level of the central stimulation by CO,. Therefore, the responses to stimulation of breathing with central CO~ and with peripheral O, CO~ combinations are additive. This is reflected in Eq. (3) by the absence of the product term for central and peripheral CO, tensions and the finding that Sc is independent of the peripheral O, tension. Like the linearity of Eq. (3), additivity will break down if the combined responses to peripheral and central stimuli approach the maximum breathing capacity. Using the technique of artificial brainstem perfusion, Riedstra (1963) measured peripheral O_~ responses at several base excess levels and central CO, tensions. He concluded that the central CO~ influenced the critical value below which ventilation reacted to changes in peripheral O, tension. Lloyd (1966) has argued that Riedstra's data were consistent with addition of peripheral O, stimulation to central CO, stimulation, as we also found in this study, and that Riedstra's conclusion was due to his unconventional introduction of a critical Po:. We share Lloyd's view. Indeed, the ventilatory response to peripheral O~ does not manifestly show such a critical Po:. It should also be noticed that ventilatory responses are influenced by the base
388
J H. G. M. VAN
BEEK
el a/.
T A B L li "~
C c n l r a l CO, scnsilivity (So). p e r i p h e r a l ('O 2 sensitivity (Sp) :.llld K e s t i m a t e d by multiple regression allalysis Expi. no.
721
Body 'acight (kg)
Pal'), (kPa)
3.7
39
Sc (rill.rain
i .kPa
I)
Sp (ml.miu
i .kpa
i)
K (ml.min
643
.~_ "~ 8
.:;_"%1_~
7.6
( ~__
~01
2446
938
3.9
55 14.6 9.7
1023 1184 1080
172 277 483
4840 5581 52115
OOl
3. I
51) 8.2
I 185 791
102 155
5002 3245
1103
2.9
50 8.2
920 867
171 264
3568 2703
004
4.3
53 8. l
803 978
101 379
28211 3637
1114
3.7
53 8.4
452 468
49 395
15711 2503
1118
3.7
61 7.6
253 167
7{) 325
966 723
(119
3.9
57 7.7
419 249
50 I 12
1154 289
11211
3.5
57 7.7
631 855
I I0 358
3149 4622
022
3.9
55 7.5
457 4~
79 164
1710 1271
1124 *
3.4
57 7.6
1528 1664
l 12 68il
6494 8157
1/13
2.7
50 13.4 8.1i 6. I
459 483 111()1 383
39 75 253 342
I lfl I 876 222O 377
105
3.7
54 13.3 6.0
851 686 I 1)
141 169 682
3869 2839 2681
* The I-csull of cxpi. 024 given in l'ablc 1 was from an earlier slate in this cXpOl'ilricnl and ~,as not included in the Illultipic regression amll',.sis.
excess which was around 0 mmol/L in Riedstra's experiments and - 7 lnmol/L ( S D = 2 mmol/L, range - l l . 0 t o - 1 . 7 mmol/L) in ours. Lahiri and DeLaney (1975b) concluded from the restllts of their experiments oil cats that there is a positive interaction between the peripheral and central drive to ventilation. In order to demonstrate such an interaction they had to apply "overall"
PERIPHERAL HYPOXIA AND VENTILATION
389
hypoxia, while in our experiments the brainstem was always supplied with hyperoxic blood. It may well be that an interaction between central and peripheral stimuli to breathing is brought about by hypoxia in the brainstem. In addition, there may be an interaction between central hypoxia and central hypercapnia. Berger el ell. (1978) and Giese e l al. (1978) concluded that central and peripheral stimuli showed negative interaction (hypoadditivity). It should be noted that ventilations often approached or exceeded 0.8 L . rain ' • kg ~ m their experiments. We limited ourselves to ventilations below this value and our conclusions only pertain to these lower levels of ventilation. Therefore, their results and ours are not necessarily in contradiction. Fitzgerald and Dehghani (1982) and Lahiri and DeLaney (1975a) measured neural signals in the sinus nerve while applying different CO, and O, tensions in the arterial blood. There was sensitivity to CO, at hyperoxia and this sensitivity increased when the arterial O, tension was lowered to normoxic and hypoxic values. Aortic bodies also responded to increases of COn and their responses showed positive interaction between O, and CO~ (Fitzgerald and Dehghani, 1982: Lahiri el al., 1979). We measured a similar pattern in the ventilation when applying combined O, CO, stimuli peripherally, keeping the central CO, constant (see figs. 2 and 3). As shown by Fitzgerald and Dehghani (1982) the peripheral chemoreceptor responses to CO, become convex at deeper levels of hypoxia, but they still respond to lowering of the Po~. Our ventilatory peripheral CO_, responses also tended to be convex during severe hypoxia (see fig. 2). This saturation of the response to CO, of the peripheral chemoreceptors could explain why we found lower Sp values during severe hypoxia than during milder hypoxia in some cases (see fig. 3). It may also be the reason why Giese er al. (1978) did not find a positive interaction in ventilation when they stimulated the carotid body, since their levels of hypoxia were in general deeper than the ones we used. However, such phenomena could also reflect saturation in other parts of the peripheral chemoreflex arc. The available experimental results suggest strongly that the positive interaction between the peripheral O, and CQ, stimuli, as reflected in the ventilatory response, takes place at the arterial chemoreceptors and can already be measured in the neural signals carried in their afferents. These signals constitute a peripheral drive, which is combined with the central CO, drive in the integrating centres of the respiratory control system. The neural output of these centres is translated into ventilation. At the level of ventilatory output, the peripheral and central drives are combined additively.
Acknowledgements This research was supported by the Foundation for Medical Research F U N G O , which is subsidized by the Netherlands Organization for the Advancement of Pure Research (Z.W.O.). We are indebted to L. Philips for his skilful work in the surgical
390
J . H . G . M . VAN BEEK
el
a/.
preparation of the animals, Dr. M . E . Wise for advice on statistics and the text and J. H. van der Perk and M. M. Rancuret for technical assistance.
References Beck, J. H. G. M. van, A. Berkenbosch, J. de Goede and C. N. Olievier ( 1981 ). Peripheral hypoxaemia and ventilatory response to C()~ in cats. Era..I. R~,slm.. Di.~. 62 (Suppl.): 146 147. Berger, W., K. Bcrger, J. Berndt and K. Giese (1978h Interaction of peripheral and central respiratory drives in eats. 1. Effects of sodium cyanide as a peripheral chemoreceptor stimulus at difl'erent levels ot'CSF pH. pllii~,,er's Arch. 374:205 210. Berkenbosch. A., J. Heeringa. C N. Olicvier and E.W. Kruyl (1979). Artificial perfusion of lhe ponto-medullary region of cats. A method for separation of central and peripheral effects of chemical slimnlation of ventilation. Respir. Ptn'.~iol. 37:347 364. Fitzgerald, R.S. and G . A . Dehghani (1982). Neural responses of the cat carotid and aortic bodies to hypercaprda and hypoxia..I. AppI. t'ln'siol. 52:596 ~01. (;icse, K.. J. Berndl and W. Bcrgcr (1978). lnlcraction of central and peripheral drives in cats. 11. Peripheral and central interaction of hypoxia and hypercapnia, p/lii,k, er's Arch. 374:21 I 217. Heeringa. ,l., A. Berkenbosch, J. de (}ocde and C. N. Olievier (1979). Relative contributions of central and peripheral chemoreceplors to Ihe ventilatory response to ('O 2 during hyperoxia. Re.~;:~ir. Phlsiol. 37:365 379. Johnston, J. (1972). Econometric Methods. Tokyo, McGrax~-Hill Kogakusha, pp. 192 207. Lahiri, S. and R.G. DeLancy (1975a). Stinaulus inleraction in the responses of carotid body chemoreceptor single afferent fibres. Re.~pir. Physiol. 24:249 266. Lahiri. S. and R . G . DeLancy (1975b). Relationship between carolid chemoreceptor activity and ventilation in tile cat. Re.spir. P/nsiol. 24:267 286. Lahiri, S. (1t)76). Depressant effect of acute and chronic hypoxia in ventilation. In: Morphology and Mechanisms of Chemoreceplors, cditcd by' A. S. Paintal. Delhi. Vallabhbhai Patel ('hest Instinlle, pp. 138 146. Lahiri. S., E. Mnlligan, r. Nishino and A. Mokashi (197t)). Aortic body chcnlorcccptor responses to changes in P ~ o : a n d P o : i n l h e c a l . J. Apl~l. Phl,~'iol. 47:S58 g66. Lloyd. B.B. (It)66). The interactions between hypoxia and other ventilator> stimuli. In: Proc. Int. Syrup. Cardiowlsc. Rcspir. Effects of Hypoxia, edited by J. D. Itatcher and D. B. Jennings. Basle, Karger, pp. 146 165. Riedstra. J.W. (1963). Influence ol'ccnlral and peripheral P{o: (pH) on tile vcntilatory response to hypoxic chemoreceptor stimulalion, t o : , Pk.rsiol. Pharma('ol. Need. 12:407 452. Sokal. R. R. ~lnd F.J. Rohlf (1%9). Biometry. San Francisco, Frccman, pp. 3~)t) 401.