Acute ventilatory responses to hypoxia during voluntary and

AIS 2089, pp. 161-168

Journal of Physiology (1994), 477.1

161

Acute ventilatory responses to hypoxia during voluntary and electrically induced leg exercise in man J. J. Pandit and P. A. Robbins University Laboratory of Physiology, Parks Road, Oxford OX1 3PT The acute ventilatory response to a brief period of hypoxia (AHVR) was measured in six subjects (a) at rest, (b) during electrically induced leg exercise (EEL), (c) during voluntary leg exercise at an external work rate matched to electrical exercise (EVI) and (d) during voluntary leg exercise at an internal work rate (i.e. metabolic rate) matched to electrical exercise (EV2). The end-tidal Po2 during hypoxia was 50 mmHg and the end-tidal Pco, was held constant at 1-2 mmHg above resting values throughout each of these four protocols. 2. EEL was produced by surface electrode stimulation of the quadriceps muscles so as to cause the legs to extend at the knee and lift a set of weights via a pulley system. During EV1, each subject lifted the same weight through the same height and at the same frequency as during his EEL protocol. During EV2, the weight, the height through which it was lifted and the frequency of voluntary contractions were altered to produce a similar 02 consumption and CO2 production as during EEL. 3. In each subject, end-tidal PCO2 values showed no change between the four protocols, and in three subjects in whom they were measured, arterial PCO2 values were also similar between the protocols. Venous lactate levels did not increase after EEL or EV2. 4. The AHVR during EEL (14-1 + 1-42 1 min'; mean + S.E.M) was significantly increased (Student's paired t test) compared with rest (7'55 + 1.10 1 min'; P < 0 003). The AHVR during EV2 was very similar to that during EEL (13-6 + 1P35 1 min1). The AHVR during EVI was significantly increased compared with rest (9-62 + 0-88 1 min'; P< 0 002) but significantly lower than during both EEL (P < 0 004) and EV2 (P < 0 008). 5. These results suggest that the increase in the acute ventilatory response to hypoxia which normally occurs during exercise in man can also occur in the absence of a drive to exercise from the cortex.

1.

During steady-state moderate exercise in man, the ventilation appears to be matched to the level of metabolic CO2 production such that alveolar or arterial PCO2 is kept approximately constant (Douglas, Haldane, Henderson & Schneider, 1913). It is generally accepted that a combination of humoral and neural factors are important in regulating this ventilatory response (Dejours, 1962; Asmussen, 1967; Cunningham, 1987). Humoral stimuli include factors such as blood gas oscillations (Yamamoto & Edwards, 1960); catecholamines (Cunningham, Hey, Patrick & Lloyd, 1963); arterial plasma potassium (Paterson, Robbins & Conway, 1989); and the CO2 flux itself (Yamamoto & Edwards, 1960; Phillipson, Duffin & Cooper, 1981), or unidentified factors related to it. The neural drive to breathe incorporates two distinct entities: first, a cortical 'irradiation' from the voluntary effort of exercise (Krogh & Lindhard, 1913), and second, peripheral feedback from the working muscles (McCloskey & Mitchell, 1972; Tibes, 1977; Bennett, 1984). Asmussen, Nielsen & Weith-Pedersen (1943), Adams, Garlick, Guz, Murphy & Semple (1984b) and Brice et al.

(1988 b) used electrical stimulation of the leg muscles in normal human subjects to eliminate the influence of cortical drive and found that the steady-state ventilation in euoxia was still matched to metabolic CO2 production. Furthermore, the last two groups of investigators repeated the experiments with paraplegic subjects, and concluded that reflex neural inputs from the electrically stimulated limbs were also unnecessary for this matching of ventilation to CO2 elimination which is so typical of the normal response to exercise (Adams, Frankel, Garlick, Guz, Murphy & Semple, 1984a; Brice et al. (1988a). It is well established that during moderate exercise, the normal respiratory response also involves an increase in the acute response to a brief period of hypoxia (Cunningham, Spurr & Lloyd, 1968; Weil, Byrne-Quinn, Sodal, Kline, McCullough & Filley, 1972; Masson & Lahiri, 1974). It has been suggested that this represents an augmented drive from the peripheral chemoreceptors during exercise, and this has also been examined by the use of short pulses of high inspiratory oxygen during exercise (Dejours, 1957). It

162

J. J. Pandit and P. A. Robbins

is not known which factors are involved in increasing the ventilatory sensitivity to hypoxia during exercise, but it is possible that some, or all, of the neural and humoral factors discussed above play a part. In particular, the role of voluntary effort or cortical irradiation in increasing the hypoxic chemoreflex sensitivity is unclear. The purpose of this study was to use the method of Adams et al. (1984b) and use electrical muscle stimulation in normal human subjects to investigate whether a cortical drive is necessary to augment the acute hypoxic ventilatory response (AHVR) during exercise. We compared the magnitude of the AHVR during electrical exercise not only with that in the resting state, but also with that during two voluntary exercise protocols: one matched with electrical exercise for external work rate, and one matched for internal work rate (i.e. metabolic rate).

METHODS Subjects Six normal males with a mean age of 23-2 years (range 21-29 years), height 1P82 m (range 1P78-1P91 m) and weight 72-2 kg (range 60-86 kg) were studied. They gave their informed consent to this study, and the study was approved by the Central Oxford Research Ethics Committee. Determination of resting end-tidal and blood gas tensions The end-tidal values for PCO2 (PET,C02) for each subject breathing air at rest and during electrical and voluntary exercise were obtained as the means over a 15 min period, after familiarization with the laboratory. End-tidal values for Pco, were also recorded in the few minutes prior to commencing each experimental period of each protocol. In three subjects (802, 905 and 911), three arterial blood samples from an indwelling radial catheter (20 G Angiocath, UT, USA) were drawn at 5 min intervals at rest, and during the electrical and both the voluntary exercise protocols to determine possible changes in arterial PCO2 (PaC002) with exercise (IL 1306 Blood Gas Analyser, Instrumentation Laboratories, Warrington, UK). In a fourth subject (835), it was possible to obtain only arterial blood samples at rest.

Main study Subjects were seated in a specially designed chair (Fig. 1). Their feet were strapped to a footplate which, when the leg was extended at the knee, lifted a set of weights (M) via a system of pulleys. The height (h) through which the weights were moved, and the frequency of muscular contraction (f) were

J. Physiol. 477.1

recorded during all three exercise protocols. The additional load (I) associated with the weight of the subject's leg and unloaded pulley system was estimated as the force registered on a spring balance pulling the footplate (with a subject in the chair and subject's feet attached) in the horizontal plane over a similar distance achieved during the exercise protocols. This load was used in the estimation of external work rate, as described below. The apparatus for electrical muscle stimulation was the same as that used by Adams et al. (1984b), but the method of stimulation was slightly different. Two pairs of carbonized rubber electrodes (Slendertone) were strapped to the anterior surfaces of each thigh. A layer of electrode paste was used to facilitate even stimulation of the quadriceps muscle group. The stimulus wave form consisted of a pulse width of 60 ,us with a repetition frequency of 40 Hz, modulated by a rectified sawtooth envelope (Adams et al. 1984b). The amplitude of the stimulus could be varied. Stimuli were delivered to each leg simultaneously at a rate of about forty per minute (range 38-50 min' to suit the comfort of individual subjects). The overall shape of each stimulus (i.e. the 'on' and 'off' ramp times) was also varied to suit individual subjects. After familiarization of the subjects to electrical stimulation, the output current was adjusted to provide the strongest contraction that could be achieved without any discomfort. Relaxation of the muscles was passive. The weight lifted during electrical exercise was adjusted to allow a reasonably free swing of the legs: it was found, for example, that if the weight was too heavy and the legs did not swing, electrical stimulation was painful. Subjects breathed through a mouthpiece with their noses occluded. Respiratory volumes were recorded by a turbine volume measuring device (Howson, Khamnei, O'Connor & Robbins, 1986), and flows were recorded by a pneumotachograph in series with it. Expired gas at the mouth was sampled continuously and analysed for Pco2 and Po2 by a mass spectrometer. The volumes, flows, Pco2 and Po2 at the mouth were recorded in real time with a 50 Hz sampling frequency using an IBM PC AT computer. This computer also executed a peak picking program to determine PET C0,' PETO2 and inspired and expired volumes and durations. The breath-by-breath end-tidal values were passed to a second IBM PC AT computer which compared these actual end-tidal values with the desired values and, by controlling a fast gas-mixing system, adjusted the composition of inspired gas to maintain the desired endtidal values independently of changes in the ventilation. Details of this dynamic end-tidal forcing system have been described in more detail elsewhere (Robbins, Swanson & Howson, 1982; Howson, Khamnei, McIntyre, O'Connor &

Robbins, 1987).

Figure 1. Experimental apparatus Schematic diagram of the experimental set-up, showing the chair and pulley system which allow the external work rate during electrical and voluntary exercise to be estimated. P/ Active contraction Passive relaxation

J. P'hysiol. 477.1

Calculation of metabolic gas exchange

(VO2) and carbon dioxide production (Vco2) calculated (at standaid temperature and pressure, dry) oIn a breath-by-breath basis by first time-aligniing the r-ecords of the CO2 and 02 compositioIn at the mouth -with the record of respiratory flow as measured by the pneumotachograph, using the measured mass spectiometer delay. The as Xvet uncalibrated values of flowN, were then adjusted for changes in viscosity due to chaniges in gas composition. Next, for each half-breath, the respiratory flow from the pneumotachograph w-as calibrated usiing the volume measurement from the turbine device. rhen, for each gas, the difference between the amount bIreathed in and the amount breathed out -was calculated using a starting value foi the mean expired temperature, this temperatuie then being adjusted until it esulted in zero net nitr ogen exchange over the hole experimeintal period. Finally, the algorithm of Swanson (1980) w%as used to obtain an estimate of the breath-by-breath gas exchange at the pulmonary capillary. WAe have previously descriibed this method of calculating gas exchange in more detail (Pandit & Robbins, 1992).

Oxygen consumption were

Protocols After an initial 5 min to allowN a steady state to be reached, data collection was begun. Each experimental period was of 27 min duration. End-tidal Po2 was held at 100 mmHg for 5 min, followed by a total of three separate exposures to 3 mim of hypoxia (PET,o, 50 mmHg), separated by 5 min periods of euoxia (PET,o, 100 mmHg). After the last of the three hypoxic exposures, there was a final 3 min period of euoxia (Fig. 2). Throughout each protocol, PET,CO, was held 1-2 mmHg above resting values. Ther e were four protocols, and each of the six subjects unclertook each of these protocols six times, giving a total of 144 separate experimental periods, with a total of 432 separate hvpoxic exposures. Protocols A, B and C were undertaken in random order on different days, and protocol D was undertaken after the mean gas exchange data from protocol B had been calculated. In protocol A, the subject was at rest. Protocol B was undertaken during electrical muscle stimulation (EEL). For each experimental period of protocol C, the subject performed voluntary exercise, lifting the same weight through the same height and at the same frequency (using an electric metronome) as a matched period in protocol B: this protocol was therefore matched to EEI, for external work rate (EVI). In protocol 1), the subject performed voluntary exercise, but this time the weight, height lifted and frequency were adjusted to achieve similar Vco2 and VO2 as in protocol B: this protocol was therefore matched to EEL in terms of internal work rate (EV2). A blood sample for determination of venous lactate (YSI 23L Enzymatic Lactate Analyser, YSI Inc., Yellow Springs, OH, USA) was taken after each experimental period of electrical exercise (EEL) and its matching voluntary exercise period (EV2). Data analysis

VEX BET,CO2, PET,O2, VCO0 and 1702 for each experimental period were averaged over 1 min periods. The last minute of the initial 5 mnil of euoxia, before any hb poxia was administer ed, was taken to represent the 'baseline' VE. Thus, foI each subject, six such baseline

Data for

163

Voluntary control of AHVR during exercise

veentilations were obtained for each protocol, and these were combined to give the subject means for each protocol. The AHVR for each experimental period was calculated as the difference between the ventilation attained in the second minute of hypoxia during each hypoxic exposure, and the last miniute of euoxia, before the hypoxic exposure. Thus, for each subject, a total of eighteen separate values of AHVR were obtained for each of the four protocols. These values were then combined to give the mean AHVR for each subject for each protocol. AV'alues of V02 and Vco were averaged over each exper imental period, and then combined to give the relevant subject means for each protocol. External work rate for the protocols involving exercise was estimated by the equation,

External work = ghf(M+ I), where g is the acceleration due to gravity and M the weight lifted. The internal work rate during the exercise protocols was calculated using the measured VCo2 and VO to obtain the respiratory quotient, R, for each subject. The energy production per unit 02 consumptioin for this particular value of R was estimated from standard tables (Evans, 1941); and from the measured 1y02, the actual energy used was calculated. An estimate of efficiency was obtained from the ratio of the external work rate to the internal work rate for the protocol.

Statistical analysis The individual subject means for each of the variables were first subjected to an analysis of variance. If this indicated significant differences between the protocols, the differences between means of the variables from the group as a whole were assessed using a paired, two-tailed t test. Statistical significance for each variable was accepted at a value of P < 005 k-l (Bonferroni's correction) where k is the number of comparisons made for that variable (Krauth, 1988).

RESULTS Table 1 shows the mean values of end-tidal Pco2' as calculated from the values obtained in the air-breathing periods in each experiment before dynamic end-tidal forcing was started. Also shown are the arterial Pco, values from the four subjects in whom these were obtained. In each of the subjects, PET,co, did not change appreciably or consistently between the different protocols. Similarly, Pa,co2 did not change between protocols in the subjects in whom it was measured. Figure 2A shows the control of end-tidal gases in one representative subject (911) for the EEL protocol only. The gas-mixing system effected three rapid step reductions in while at the same time holding PET,CO, constant. PET,OP2 The ventilatory responses in the same subject are shown in Fig. 2B. The general form of the responses was similar in all subjects. The baseline ventilation was lowest in the protocol at rest, somewhat increased by EVI, and highest during EEL and EV2. Regarding the magnitudes of the acute hypoxic responses, the smallest occurred at rest. The AHVR appeared to be slightly increased by EVI. The largest AHVR clearly occurred during EEL and EV2, and

J. Physiol. 477.1


164

Table 1. Mean air-breathing values (mmHg ± S.E.M. for end-tidal PCO2 (PET,CO2' n = 6) and arterial Pco2 (Pa,co2, n = 3) for each subject in each of the four protocols, and the end-tidal value at which each subject was held by dynamic end-tidal forcing during the experiments (Clamp, PET,CO2) Rest

Subject 802 835 846 905 908 911

PET,CO2

Pa,C2

38-4 + 0-10 36-9 + 017 38-0 + 0-31 40'0 + 0-10 37-3 + 0 23 36-5 + 0-10

39-5 + 044 38-8 + 035

42-0 + 0-60

37-3 + 009

EVI

EEL

PET CO2 Pa 0C2 38-2 + 0-38 39-3 + 055 36-3 + 013 37-8 + 016 39-5 + 024 42-8 + 038 37-5 + 0 38 36-4 + 0-32 36-6 + 0-07

38-6 + 0O11 397 + 044 36-3 + 013 384 + 026 400 + 0-29 42-3 + 018 37-2 + 0 32 36-3 + 0'19 36-2 + 0'17

PET,CO2

Clamp

EV2 Pa 0C2

PET,CO2

Pa C2

37'0 + 0-31 36-6 + 0-31 36-1 +0O18

39 0 38-0

Table 3. Mean values of the acute hypoxic responses (1 min' ± S.E.M.) for each subject in each of the four protocols EEL EVI EV2 Rest Subject 6'62 + 0-49 9-21 + 0-51 8-84 + 0-34 802 4-82 + 0-42 13-6 + 0-58 14-2 + 0-64 8-32 + 0-66 835 5-21 + 0'43 10'7 + 0-40 11-0 + 0-61 9-36 + 0-35 846 7'26 + 0-45 17-2+ 1'06 9-46+0'48 7-14+044 905 17-2+0092 14-4 + 0-64 15'0 + 0-86 11-1 + 0-65 908 8'61 + 0-32 12-3 + 0-76

7.55 + 1l10

12-8 + 069 9 62 + 0 88

18-2 + 1-27 14-1 + 1-42

16-7 + 1.05 13-6 + 1-35

Table 4. Mean values (+ S.E.M.) for all six subjects combined for gas exchange (Vco2 and 1O ), external work rate, efficiency of work and change in plasma venous lactate before and after exercise (ALactate) for each of the four protocols EV2 EEL EVI Rest 245 + 13 298 + 13 388 + 14 * * t 401 + 15 Vc02 (ml min-) 306 + 17 377 + 21 * 476 + 21 * V2 (mlmin-') t 482 + 11 3.73 + 0-83 t 3.73 ± 0'83 * 9-06 + 1-15 External work rate (W) 53 + 0'60 * 14-8 + 1-33 13-9 + 1-78 * Efficiency (%) 0 07 + 0 04 t 0 03 + 0 05 ALactate (mmol I-') *

40-0 38 0 40-0 42-0

Table 2. Mean values of the euxic baseline ventilations (1 min1 ± S.E.M.) for each subject in each of the four protocols EEL EVI Rest EV2 Subject 20 9 + 2 07 24-8 + 1-40 17-7 + 1-33 802 13-7 + 1'97 31-5 + 2-36 32-4 + 2-72 25-9 + 0 89 835 15-5 + 0 97 22-3 + 1-58 21-3 + 0 80 16-5 + 1-23 846 13-4 + 0 74 905 13-2 + 1-05 17-8 + 1-09 21'7 + 2'18 18-4 + 1-52 908 14-9 + 1-21 22-3 + 2'79 25-3 + 1-23 26-2 + 1-83 30 6 + 1-02 32-4 + 1-65 25 0 + 2-34 22-1 + 2-52 911 25-8 + 2-12 20'8 + 1-67 25'5 + 2'14 Mean 15-5 + 1-38

911 Mean

PET,CO2

38-9 + 015 40-1 + 038 36-8 + 0 26 37-7 + 026 40-4 + 0-27 42-6 + 0'28

Significant differences for comparisons between the columns on either side: t not significant; Student's t test with Bonferroni's corrections.

J. Physiol. 477.1

165

Voluntary control of AHVR during exercise B

A 120 r

50

50 1 100

F

40

40 cm

cm

I

I

E

E

E 30 _

80 F

*w

0

0~

20 30

F

60 F-

10 _

40 . -10

20 -5

5 10 Time (min)

0

15

20

25

-10

-5

0

5 10 Time (min)

15

Figure 2. Gas input and ventilatory responses in one subject A, end-tidal gas profiles for Po2 (V) and PCo, (0) for one subject (911). For clarity, only the mean profile from EEL is shown. The dynamic end-tidal forcing system effects three successive hypoxic steps of 3 min duration, each separated by 5 min of euoxia. Throughout the protocols, end-tidal Pco2 remains constant. B, mean ventilatory responses for all four protocols in the same subject (911). At rest (0); during EVI (V); during EEL (0); and during EV2 (v).

A

B l*l

.7-1

30 I

I

15

25

20 7C

X 15

10

E -

0 co cv

10

,< 5

m

0

Rest

EV1

EEL

EV2

Rest

EV1

EEL

EV2

Figure 3. Mean ventilatory responses and statistical comparisons A, mean values (± 2 S.E.M.) of euoxic baseline ventilation of all six subjects combined, for each protocol, and the comparisons between the protocols. n.s., no significant difference; *significant difference (Student's paired t test with Bonferroni's correction). Baseline ventilations in both EEL and EV2 are significantly different from ventilations in both rest and EVI. B, mean values ( + 2 S.E.M.) of the acute hypoxic ventilatory responses of all six subjects combined, for each protocol, and the comparisons between the protocols. Statistical comparisons are made as described above and in the text. Acute hypoxic ventilatory responses in both EEL and EV2 are significantly different from responses in both rest and EVI.

20

25

166


the magnitudes of these appeared to be similar. For all protocols the mean peak ventilation during hypoxia was reached in the second minute. The magnitudes of the three successive AHVRs within each protocol were similar. Table 2 shows the calculated values for euoxic baseline ventilation for each subject, and Fig. 3A shows the mean values for all six subjects combined, and the statistical comparisons between the protocols. Electrical exercise significantly increased the euoxic ventilation as compared with rest, and the baseline ventilation during EV2 was similar to that during EEL. Voluntary exercise matched to the external work rate of electrical stimulation (EVI) also slightly increased the baseline ventilation as compared with rest. The baselines during both EEL and EV2 were significantly higher than those during both rest and EVI. Table 3 shows the values of AHVR for each of the subjects, and Fig. 3B shows the mean values for all six subjects combined and the statistical comparisons between the protocols. Electrical exercise almost doubled the AHVR attained at rest. The AHVR during the matched voluntary protocol, EV2, was similar to that during EEL. Voluntary exercise matched to the external work rate of EEL did increase the AHVR modestly, as compared with rest (mean = 2-07 1 min'), and this value reached statistical significance. Both EEL and EV2 increased the AHVR significantly, as compared with both rest and EVI. Table 4 shows the gas exchange data, work rate and efficiency, and changes in lactate. During EVI, metabolic CO2 production and 02 consumption increased significantly by a small amount (53 ml min' and 71 ml min' respectively), even though external work rate during this protocol was small. Electrical exercise increased metabolism by a much greater amount, and while EV2 matched EEL adequately both in terms of Vco2 and V02, overall efficiency of work was much lower during electrical stimulation. There was no significant rise in lactate either after EEL or EV2.

DISCUSSION The main conclusion from this study is that the increase in the acute ventilatory response to hypoxia which normally occurs during exercise in man can also occur in the absence of a voluntary drive to exercise from the cortex.

Is electrical exercise involuntary? A number of methods were used to try to establish whether the subjects voluntarily contracted their leg muscles during electrical stimulation. First, all the subjects were comfortable during electrical stimulation and denied that they had co-operated voluntarily with the electrical stimulus. We stopped the experiment if any discomfort whatsoever was reported. Second, we used the method suggested by Adams et at. (1984 b) to assess voluntary involvement by occasionally switching off the current discretely during EEL: a voluntary contraction might

J. Physiol. 477.1

have been expected if the subject had been driving his muscles in tune with the stimulator; however, this did not occur. Occasionally, we also told the subject that the experiment and data collection had finished, but continued the muscle stimulation: if any part of the leg movement had been voluntary, we might have expected a change in the height through which the weight was lifted (Fig. 1). Again, this did not occur. Indeed, careful observation of the height lifted and the frequency of contractions revealed a remarkable constancy during all experimental periods of electrical stimulation. In addition, we kept careful observation of the legs themselves. Electrical stimulation produces a characteristic generalized contraction of the quadriceps, very different from that produced during voluntary contraction. Relaxation during electrical stimulation was passive: we might have expected to see an active contraction of the hamstrings had the subject been making voluntary efforts. It was also noticed that on some occasions, the strength of contraction of the right and left quadriceps was slightly unequal during electrical stimulation: such inequality would be unexpected during normal voluntary activity, and was not seen in the same subjects during EVI and EV2. Observation of the subjects' feet showed that with electrical stimulation, the muscles around the ankle joint were relaxed and flaccid, thus keeping the foot in a natural partial plantar flexion during leg extension; during the voluntary protocols (EVI and EV2), the foot was, in contrast, kept partially dorsiflexed as the leg was raised. Any dorsiflexion of the foot during electrical quadriceps activation, therefore, might have suggested an active component to contraction. Finally, the overall efficiency of work during EEL was much lower than during EVI or EV2 (Table 4): a significant voluntary contribution during EEL would have been unlikely to yield such large differences in efficiency. We are therefore satisfied that exercise during EEL was involuntary. While we cannot entirely exclude all possibility that some volitional reinforcement might have been involved, we feel that, if it occurred, its contribution to the overall results was extremely small.

Control of Pco0 and lactic acid stimuli When comparing the magnitudes of the acute ventilatory hypoxic responses between different protocols, it is important to ensure that the Pco, and lactic acid stimuli between the protocols are similar. In the three exercise protocols in this study, the overall work load was relatively low, and the end-tidal, air-breathing PCO2 (and the arterial PCo2 in the three subjects in whom it was measured) was not appreciably changed by the level of exercise, as compared with rest (Table 1). Since the dynamic end-tidal forcing system held PET,co, at the same level in all the protocols, the level of arterial PCO2 should also have been similar between protocols.

J. Physiol. 477.1

Voluntary control of AHVR during exercise

The constancy of PET,CO2 and Paco2 between rest and electrically induced exercise is consistent with the finding of Adams et al. (1984b), who measured end-tidal Pco, in all subjects and Pa,Co, in two of their subjects. Brice et al. (1988b) measured the temporal patterns of Paco2 during electrical leg exercise, and found a slight tendency to hypocapnia in the first 30 s of electrical exercise, but found that eucapnia was achieved in the steady state. We found no significant rise in lactate levels during EEL or EV2 (Table 4). This result differs somewhat from that of Adams et al. (1984b), who obtained a mean lactate level of 2 17 mmol I' after EEL, and with that of Brice et al. (1988b) who reported a small but significant rise of about 0-5 mmol 1'. These differences may be related to the intensity of electrical stimulation: in both these previous studies, the increase in JIco2 was of the order of 300 ml min'. Our stimulation achieved a more modest increase in Ic02 of about 150 ml min', and so avoided the problem of any lactic acidosis.

Physiological significance of the results The purpose of this study was to assess whether voluntary control is necessary for the increase in ventilatory response to hypoxia which occurs during exercise. Our conclusion is *that it is not. Voluntary exercise at the lower work rate (EV1) increased the ventilatory response to hypoxia by a small amount (about 21 min'). However, the AHVR was increased much more by the electrical exercise protocol EEL, although the external work rate was the same as during EVI (Fig. 3). Furthermore, it is interesting that the AHVR obtained during EV2 was so similar to that during EEL. Although the metabolic rates in these two protocols were matched, there was in addition voluntary effort during EV2 which might be expected to act independently to increase the magnitude of the AHVR, as compared with EEL. It appears, therefore, that internal work rate is the important variable, rather than external work rate or how the work is achieved. This result suggests that Vco2 or factors related to it are important in determining the ventilatory response to hypoxia during exercise in man. This study does not indicate which precise factors may be involved, since the increased Vco2 may increase the stimulus from blood gas oscillations; or the sensation of 'electricity' in normal subjects could evoke a sympathetic response and catecholamine release; or electrical stimuli to

the muscle might release potassium and increase the levels in arterial blood. Any, or all, of these humoral stimuli could be proportional to the work rate and may increase the chemoreflex sensitivity to hypoxia by a direct action on the carotid bodies. Finally, this study does not exclude the possible involvement of nervous afferents from the muscles: these would also carry information about the amount of internal work being done (whether voluntary or involuntary) and could mediate an increased sensitivity to hypoxia more centrally in the chemoreflex pathway.

167

REFERENCES ADAMS, L., FRANKEL, H., GARLICK, J., Guz, A., MURPHY, K. & SEMPLE, S. J. G. (1984a). The role of spinal cord transmission in the ventilatory response to exercise in man. Journal of Physiology 355, 85-97. ADAMS, L., GARLICK, J., GUZ, A., MURPHY, K. & SEMPLE, S. J. G. (1984 b). Is the voluntary control of exercise in man necessary for the ventilatory response? Journal of Physiology 355, 71-83. ASMUSSEN, E. (1967). Exercise and the regulation of ventilation. Circulation Research 20 and 21, suppl. 1, 132-145. ASMUSSEN, E., NIELSEN, M. & WEITH-PEDERSEN, G. (1943). Cortical or reflex control of respiration during muscular work? Acta Physiologica Scandinavica 6,168-175. BENNETT, F. M. (1984). A role for neural pathways in exercise hyperpnea. Journal of Applied Physiology 56,1559-1564. BRICE, A. G., FORSTER, H. V., PAN, L. G., FUNAHASHI, A., HOFFMAN, M. D., MURPHY, C. L. & LOWRY, T. F. (1988a). Is the hyperpnea of muscular contractions critically dependent on spinal afferents? Journal of Applied Physiology 64, 226-233. BRICE, A. G., FORSTER, H. V., PAN, L. G., FUNAHASHI, A., LOWRY, T. F., MURPHY, C. L. & HOFFMAN, M. D. (1988 b). Ventilatory and Pa,co2 responses to voluntary and electrically induced leg exercise. Journal of Applied Physiology 64, 218-225. CUNNINGHAM, D. J. C. (1987). Studies on arterial chemoreceptors in man. Journal of Physiology 384,1-26. CUNNINGHAM, D. J. C., HEY, E. N., PATRICK, J. M. & LLOYD, B. B. (1963). The effect of noradrenaline infusion on the relation between pulmonary ventilation and the alveolar Po, and PCO2 in man. Annals of the New York Academy of Sciences 109, 756-770. CUNNINGHAM, D. J. C., SPURR, D. & LLOYD, B. B. (1968). The drive to ventilation from arterial chemoreceptors in hypoxic exercise. In Arterial Chemoreceptors, ed. TORRANCE, R. W., pp. 301-324.

Blackwell, Oxford. DEJOURS, P. (1957). Interet methodologique de l'etude d'un organisme vivant a la phase initiale de rupture d'un equilibre physiologique. Comptes rendus des Seances de l'Academie des Sciences Paris 245,1946-1948. DEJOURS, P. (1962). Chemoreflexes in breathing. Physiological Reviews 42, 335-358. DOUGLAS, C. G., HALDANE, J. S., HENDERSON, Y. & SCHNEIDER, E. C. (1913). Physiological observations made on Pike's Peak, Colorado, with special reference to adaptation to low barometric pressures. Philosophical Transactions of the Royal Society B 203,185-318. EVANS, C. L. (1941). Starling's Principles of Human Physiology, 8th edn, pp. 816-854. J. & A. Churchill, London. HoWSON, M. G., KHAMNEI, S., MCINTYRE, M. E., O'CONNOR, D. F. & ROBBINS, P. A. (1987). A rapid computer-controlled binary gas-mixing system for studies in respiratory control. Journal of Physiology 394, 7P. HOWSON, M. G., KHAMNEI, S., O'CONNOR, D. F. & ROBBINS, P. A. (1986). The properties of a turbine device for measuring respiratory volumes in man. Journal of Physiology 382, 12P. KRAUTH, J. (1988). Distribution-free statistics: An applicationoriented approach. In Techniques in the Behavioral and Neural Sciences, vol. 2, ed. HUSTON, J. P., pp. 36-37. Elsevier, Amsterdam. KROGH, A. & LINDHARD, J. (1913). The regulation of respiration and circulation during the initial stages of muscular work. Journal of Physiology 47,112-136. MCCLOSKEY, D. I. & MITCHELL, J. H. (1972). Reflex cardiovascular and respiratory responses originating in exercising muscle. Journal of Physiology 224,173-186. MASSON, R. G. & LAHIRI, S. (1974). Chemical control of ventilation during hypoxic exercise. Respiration Physiology 22, 241-262. PANDIT, J. J. & ROBBINS, P. A. (1992). Ventilation and gas exchange during sustained exercise at normal and raised CO2 in man. Respiration Physiology 88,101-112.

168


PATERSON, D. J., ROBBINS, P. A. & CONWAY, J. (1989). Changes in arterial plasma potassium and ventilation during exercise in man. Respiration Physiology 78, 323-330. PHILLIPSON, E. A., DUFFIN, J. & COOPER, J. D. (1981). Critical dependency of respiratory rhythmicity on metabolic C02 load. Journal of Applied Physiology 50, 45-54. ROBBINS, P. A., SWANSON, G. D. & HOWSON, M. G. (1982). A prediction-correction scheme for forcing alveolar gases along certain time-courses. Journal of Applied Physiology 52, 1353-1357. SWANSON, G. D. (1980). Breath-to-breath considerations for gas exchange kinetics. In Exercise Bioenergetics and Gas Exchange, ed. CERRETELLI, P. & WHIPP, B. J., pp. 211-222. Elsevier, Amsterdam. TIBES, U. (1977). Reflex inputs to the cardiovascular and respiratory centers from dynamically working muscle. Circulation Research 41, 332-341. WEIL, J. V., BYRNE-QUINN, E., SODAL, I. E., KLINE, J. S., MCCULLOUGH, R. E. & FILLEY, G. F. (1972). Augmentation of chemosensitivity during mild exercise in normal man. Journal of Applied Physiology 33, 813-819. YAMAMOTO, W. S. & EDWARDS, McI. W. (1960). Homeostasis of carbon dioxide during intravenous infusion of carbon dioxide. Journal of Applied Physiology 15, 807-818.

Acknowledgements We are grateful to Professor A. Guz and Dr K. Murphy of the Department of Medicine, Charing Cross Hospital for the loan and instruction in the use of the electrical muscle stimulator. We also thank Mr B. Howse and Mr D. O'Connor for their skilled technical assistance. This study was supported by the Wellcome Trust. Jaideep J. Pandit is a Wellcome Trust Medical Graduate Training Fellow. Received 2 February 1993; accepted 7 October 1993.

J. Physiol. 477.1