Eur J Appl Physiol (2001) 85: 272±279 DOI 10.1007/s004210100457

O R I GI N A L A R T IC L E

Susan A. Ward á Darryl Macias á Brian J. Whipp

Is breath-hold time an objective index of exertional dyspnoea in humans?

Accepted: 12 April 2001 / Published online: 7 July 2001 Ó Springer-Verlag 2001

Abstract Since dyspnoeic sensation (d) increases progressively with work rate (WR) and the duration of a volitional breath-hold (tBH) shortens, we wished to explore whether tBH might correlate suciently closely with d to provide a quantitative and descriptor-free index of respiratory sensation during dynamic exercise. Nine healthy males exercised on a cycle ergometer at a series of constant WRs, above and below the lactate threshold. Ventilatory and gas exchange variables were measured breath-by-breath. At each WR, breath-holds to the limit of tolerance were taken; d was recorded (visual-analog scale) immediately prior to and throughout each breath-hold. During breath-holds, d increased with time as a ``break-away'' monoexponential characteristic, reaching the maximum (100%) at the breakpoint. Despite end-tidal partial pressure of carbon dioxide at the break-point being higher and end-tidal partial pressure of oxygen being lower with increasing WR, the relationship between WR and tBH declined curvilinearly (i.e. with large falls in tBH occurring in the low WR range, but far smaller reductions at higher WRs). The tBH/minute ventilation relationship had a similar form. The relationship between pre-breath-hold d and tBH was also complex: the large reductions in tBH in the low WR range were associated with only modest increases in pre-BH d while, at higher WRs, the proS.A. Ward (&) Centre for Exercise Science and Medicine, Institute of Biomedical and Life Sciences, Glasgow University, Glasgow G12 8QQ, UK E-mail: [email protected] Tel.: +44-141-3306287 Fax: +44-141-3306345 S.A. Ward á D. Macias á B.J. Whipp Departments of Physiology and Anesthesiology, UCLA School of Medicine, Los Angeles, California 90024, USA B.J. Whipp Department of Physiology, St George's Hospital Medical School, London SW17 0RE, UK

gressively smaller decrements in tBH were associated with progressively larger increases in d. We therefore conclude that breath-hold duration is unlikely to provide a useful correlate of exertional dyspnoea during dynamic exercise. Furthermore, the relative prolongation of tBH at high WRs (accounting for the moreextreme levels of end-tidal gas tensions) may re¯ect the attention-diverting in¯uence of the exercise per se. Keywords Shortness-of-breath á Exertional dyspnoea á Ventilatory drive

Introduction The duration of a voluntary breath-hold in humans is normally determined by the intensity of the ``uncomfortable'' respiratory sensations reaching intolerable levels (Schneider 1930; Mithoefer 1965b; Lin 1982). It is of interest, therefore, that Nishino et al. (1996) have proposed recently that analysis of breath-hold manoeuvres might provide information regarding the genesis of dyspnoea. In addition to the stoicism of the subject, which is important in setting the upper limit for the stimuli that trigger the break from the breath-hold, the prevailing conditions immediately prior to the manoeuvre are also important. For example, when the initial condition of a stimulus such as hypoxia, hypercapnia or lung volume is far removed from the value that obtains at the limit of tolerance, then the time to reach the critical ``breaking'' level is prolonged. Consequently, breath-hold time (tBH) is longer with prior hyperoxia (Rodbard 1947; Klocke and Rahn 1959; AÊstrand 1960; Lin et al. 1974) and shorter with prior hypoxia (Hill and Flack 1910; Engel et al. 1946; Otis et al. 1948; AÊstrand 1960). Similarly, prior hyperventilation prolongs tBH, consequent to the initial hypocapnia (Hill and Flack 1910; Mithoefer 1965b). Respiratory mechanical in¯uences also in¯uence the breath-hold break-point: for example, tBH can be prolonged by rebreathing gas at the break-point that

273

does not alter the current levels of alveolar partial pressures of carbon dioxide and oxygen (PCO2 and PO2, respectively; Fowler 1954) or by starting the breath-hold from a high initial lung volume rather than a low one (Mithoefer 1965a). As exercise has been reported both to increase the dyspnoeic rating (d) in a broadly linear fashion with increasing work rate (WR; e.g. Stark 1988; Wilson and Jones 1991; Killian et al. 1992) and to shorten tBH (Rodbard 1947; AÊstrand 1960; Craig and Babcock 1962; Cummings 1962; Lin et al. 1974; Sterba and Lundgren 1985), we hypothesised that the shortening of tBH would re¯ect the degree of the dyspnoea induced by the exercise prior to the breath-hold. This depended upon the assumption that the shortening of tBH would re¯ect the in¯uence of the initial condition ± here the WR, in the ®rst instance ± rather than an alteration in the limiting value of the perceptual intensity at the break-point per se. We were interested, therefore, in determining the extent to which tBH during exercise might prove to be a close correlate and, hence, a potentially useful descriptor-free, quantitative index of the intensity of respiratory sensation prevailing prior to the breath-hold.

Methods Subjects Nine healthy male volunteers (Table 1) each provided signed informed consent, as approved by the institutional ethics committee, prior to participation in the investigation. None were experienced divers. Subjects ®rst underwent several familiarisation sessions in which all manoeuvres and procedures that were to be used in the study were practised; this was especially important for the breathhold manoeuvres. The subjects subsequently performed ramp-incremental and square-wave tests on an electromagnetically-braked cycle ergometer (Excalibur Sport, Lode, The Netherlands). Equipment Subjects breathed through a mouthpiece that was connected to a low-dead-space (90 ml), low-resistance (<1.5 cmH2O at 3 lás±1) turbine volume transducer (VMM, Alpha Technologies, Laguna Niguel, USA) for the measurement of inspiratory and expiratory volumes and ¯ows. A small aliquot of respired gas (1 mlás±1) was withdrawn continuously from the mouthpiece and analysed for oxygen, carbon dioxide and nitrogen concentrations by mass Table 1 Physical characteristics of the subjects. (V_ O2peak Peak oxygen uptake, Thla lactate threshold)

spectrometry (2000, Airspec, Biggin Hill, UK). The mass spectrometer was calibrated using precision-analysed mixtures. The time delay between the volume and gas concentration signals was measured by passing a bolus of gas through the system (Beaver et al. 1973). The electrical signals from these devices were sampled every 20 ms and processed on-line by a digital computer, after analog-todigital conversion, for the computation and display, breath-bybreath, of: minute ventilation (V_E , B.T.P.S.), oxygen uptake (V_ O2 S.T.P.D.), carbon dioxide output (V_ CO2 S.T.P.D.), the respiratory exchange ratio, ventilatory equivalents for carbon dioxide and oxygen (V_E =V_ CO2 and V_E =V_ O2 , respectively), and end-tidal PCO2 and PO2 (PETCO2 and PETO2, respectively; Beaver et al. 1981; Jenkins et al. 1989). The calibration and validation procedures have been described previously (Beaver et al. 1981). Heart rate was derived beat-to-beat, from the R±R interval of a standard six-lead electrocardiogram (5000, Quinton, Seattle, USA). Ratings of d were made at selected intervals in response to the standardised request (always given by the same experimenter) to ``Please rate the diculty of your breathing''. A visual-analog scale (Ward and Whipp 1989) was used as the rating device, with ``0'' representing ``nothing at all'', and ``100'' representing ``maximum''. This was achieved by the subjects rotating a cylindrical handgrip that was linked to a linear potentiometer, which caused a pointer to move along a horizontal scale that comprised a series of lightemitting diodes; this was positioned in front of the subject in a clearly visible position. The output of the potentiometer was sent to the computer and the device was reset to zero at the end of each rating. The data for each test were displayed on-line (Dynograph, Beckman, Anaheim, USA) and stored on digital tape for subsequent analysis. Protocols Each subject initially performed a 15-Wámin±1 ramp incremental exercise test to the limit of tolerance; subjects were constrained to maintain a pedalling frequency within the range of 60±80 revámin±1. This allowed the peak V_ O2 (V_ O2peak ) to be determined and the lactate threshold (Thla) to be estimated. V_ O2peak was taken as the mean V_ O2 over the ®nal 15 s of the ramp test. Thla was estimated non-invasively using a cluster of standard gas exchange indices (e.g. Whipp et al. 1986), as the V_ O2 at which: (a) the break-point in the V_ CO2 =V_ O2 relationship (the ``V-slope'' technique) occured and (b) when the V_E =V_ O2 and PETO2 started to increase systematically without an increase in V_E =V_ O2 and a fall in PETCO2. On subsequent occasions, subjects each completed steady-state tests at seven di€erent WRs (unloaded pedalling ± 0 W, 30 W, 60 W, 90 W, 120 W, 150 W, and 180 W), both below and just above Thla (typically, tests were duplicated). For the supra-Thla tests, WRs were constrained not to exceed 40% of ``delta'' (D, the di€erence between Thla and V_ O2peak ), thus ensuring that stable levels of V_E could be attained prior to breath-holding. Each test consisted of an initial 3-min period of unloaded cycling (0 W),

Subject

Age (years)

Body mass (kg)

Height (cm)

Thla (lámin±1)

V_ O2peak (lámin±1)

1 2 3 4 5 6 7 8 9 Mean SE

23 24 49 23 24 33 22 28 30 28.4 2.8

72 77 86 73 70 77 75 78 70 75.3 1.6

175 188 183 176 165 176 182 187 175 178.6 2.4

2.9 3.2 2.2 1.7 2.3 2.0 2.8 2.3 1.5 2.32 0.19

4.2 4.9 3.2 2.8 4.3 2.9 4.5 3.8 2.7 3.70 0.27

274 followed by a period of cycling at a given WR, typically for 10 min (i.e. sucient to allow ventilation to stabilise). The order of the tests was assigned randomly among subjects, with each subject completing no more than one supra-Thla test on a particular day. At selected intervals during these tests (typically two to three times during the steady state period of each constant-load test), subjects performed a breath-hold to the limit of tolerance. No encouragement or other feedback was provided by the experimenters during the manoeuvre. Speci®cally, subjects were instructed to: exhale to residual volume, and then inhale rapidly to total lung capacity; and then to hold their breath for as long as possible. tBH was taken as the time between the onset of the maximal inspiration and the break-point of the breath-hold. Prior to being instructed to undertake the breath-hold manoeuvre, subjects were asked to provide a rating of d. About 20 s was allowed for the subject to register the level of the rating, after which the potentiometer was reset to zero by one of the experimenters. As the instruction to start the breath-hold was given, subjects were again asked to rate the diculty of their breathing, this time continuously for the entire duration of the breath-hold without resetting the potentiometer. Statistics Mean response values are expressed ‹1 standard error (SE) and are presented as mean (SE). Di€erences between individual responses (e.g. for tBH at a given WR) were tested for statistical signi®cance (P<0.05) by a standard two-way analysis of variance that included a paired-comparisons analysis for selected pairs of variables.

Results Subjects were able to perform the breath-hold manoeuvres reliably. For example, over a WR range extending from 0 W to 210 W, subjects demonstrated average di€erences of only 1.51 (0.8) s between initial and repeat breath-holds performed on di€erent occasions; this di€erence was not signi®cant. A typical online recording of an individual breath-hold manoeuvre is presented in Fig. 1, for each of three WRs. The termination of the breath-hold is evident in the restoration of rhythmic respiratory volume and gas tension excursions. Note that: (a) tBH was shorter at higher WRs, declining in this example from 43 s at 30 W to 16 s at 150 W; and (b) for the ®rst post-breath-hold expiration, PETCO2 was higher and PETO2 lower than the values obtained immediately prior to the breath-hold, consistent with retention of metabolically-produced carbon dioxide and continual extraction of oxygen from the trapped lung gas during the breath-hold (e.g. Otis et al. 1948; Lanphier and Rahn 1963; Mithoefer 1965b; Ferrigno and Lundgren 1999). V_E increased with WR, the slope of the relationship becoming steeper at higher WRs (e.g. Fig. 2A, C, ®lled symbols), re¯ecting an augmented response above Thla to e€ect respiratory compensation for the metabolic acidosis (e.g. Wasserman et al. 1967; Sutton and Jones 1979). The decline in tBH with increasing WR (Table 2) was consistently curvilinear with upward concavity, as shown by the representative display (Fig. 2A: open symbols) and the group-mean display (Fig. 2C: open symbols). That is, for a given WR increment, the reduction in tBH

Fig. 1 Representative (subject 3) on-line recording of inspired volume (VT,I), respired partial pressure of carbon dioxide (PCO2) and respired partial pressure of oxygen (PO2) for a breath-hold manoeuvre performed to the limit of tolerance at work rates of 30 W (C), 90 W (B) and 150 W (A). Following a forced expiration to residual volume, each breath-hold was initiated by a maximal rapid inspiration to total lung capacity; this volume was held for the duration of the breath-hold (as signi®ed by an absence of respiratory rhythmicity and depicted by the horizontal arrow). At the breath-hold break-point, the subject exhaled rapidly back to residual volume before resuming regular breathing

was less striking in the higher WR range, despite the proportionally greater operating levels of V_E (Fig. 2A, C). This curvilinear characteristic was preserved when tBH was plotted as a function of V_E , both in individual response pro®les and the group mean response (Fig. 3). As WR increased, the values of PETCO2 on the ®rst post-breath-hold expiration rose (Table 2; Fig. 2B, D: open symbols), and those of PETO2 fell (Table 2; Fig. 2B, D: ®lled symbols). Again, these changes were more

275

Fig. 2 Representative (subject 3; A, B) and group mean responses (C, D) of breath-hold time (tBH; m) and minute ventilation (V_E ) immediately prior to the breath-hold (d), versus work rate (A and C), and end-tidal partial pressure of carbon dioxide (m) and endtidal partial pressure of oxygen (d) for the ®rst exhalation following the breath-hold (End-BH PETCO2 and End-BH PETO2, respectively), versus work rate (B and D). Curves of best-®t are superimposed. Group mean responses are shown ‹1 SE

evident in the lower WR range. This observation, consistent across all subjects, suggests that more ``extreme'' conditions of humoral drive could be tolerated at higher WRs. Subjects were also able to provide d pro®les reliably during the breath-hold manoeuvres. After a short delay, d started to increase from the baseline of 0% in a ``break-away'' fashion. In all cases, subjects chose to rate to the maximum value of 100% at the breath-hold break-point. With increasing WR, d attained the maximum rating of 100% more rapidly, as shown in Fig. 4A for a representative subject (subject 3); for ease of comparison, the individual pro®les of d versus time at each of the three WRs presented (30 W, 90 W, 150 W) have been superimposed. The d-time pro®les were each well ®t to a standard semi-logarithmic (exponential) function (Fig. 4B), yielding time constants (s) of 12 s, 7 s and 4 s at WRs of 30 W, 90 W and 150 W, respectively, in the representative subject shown here. For the group as a whole, s averaged 14.6 (0.9) s at 30 W, 8.7 (1.0) s at 90 W, and 4.5 (0.5) s at 150 W; these values were all signi®cantly di€erent from each other. In all instances, d responded in a qualitatively similar fashion to that presented in Fig. 4A.

The pro®le of pre-breath-hold (or control) d values expressed relative to tBH was also essentially consistent among subjects (Fig. 5). That is, small increases in prebreath-hold d at low WRs were associated with large reductions in tBH (e.g. in the representative example shown here, Fig. 5A, as d increased from 3% to 5%, tBH decreased from 85 s to 25 s). However, as WR rose, progressively larger increases in d were associated with progressively smaller decrements in tBH. We hypothesised that were the factors (e.g. respiratory-mechanical; humoral) that might be assumed to dictate the dyspnoeic intensity prevailing at any particular WR (i.e. prior to a breath-hold) also to establish a mean rate of stimulus (S) development during the breath-hold (i.e. dS/dt), then the relationship between d and tBH during exercise should be hyperbolic: dS=dt  tBH ˆ K. We found the curvilinear relationship between pre-breath-hold d and tBH to be described adequately by such a function (Fig. 5), but not by a simple monoexponential function.

Discussion An essential requirement of this investigation is that subjects were able to reliably (a) rate the intensity of d over a range of steady-state WRs, and (b) perform a breath-hold from total lung capacity to the limit of tolerance. We and others (e.g. Stark et al. 1981; Adams et al. 1985; Ward and Whipp 1989; Wilson and Jones 1991) have indicated previously that the reproducibility of respiratory sensation ratings during exercise is good, even over periods as long as several weeks. We were also

276 Table 2 Group mean responses during exercise. (tBH breath-hold time, PETO2 end-tidal partial pressure of oxygen, PETCO2 end-tidal partial pressure of carbon dioxide, End-BH PETO2 PETO2 at the end of breath-hold, End-BH PETCO2 PETCO2 at the end of breath-hold, d dyspnoeic sensation) Work rate (W)

tBH (s)

10 30 60 90 120 150 180

59.1 32.5 30.2 23.9 17.1 12.6 11.1

d (%) (6.0) (3.3) (2.8) (1.7) (1.8) (1.8) (1.3)

3.8 4.4 7.0 14.3 17.9 31.5 37.2

(0.7) (0.8) (1.4) (4.0) (2.4) (2.2) (2.9)

V_E lámin±1

PETO2 (mmHg)

PETCO2 (mmHg)

End-BH PETO2 (mmHg)

End-BH PETCO2 (mmHg)

18.1 19.9 26.3 33.0 41.2 50.6 61.0

109.4 105.0 104.8 104.0 103.7 104.0 105.5

39.2 40.9 41.5 43.4 45.3 44.0 43.2

84.0 82.5 77.0 77.5 71.2 72.0 71.9

53.1 52.7 55.8 60.1 63.0 63.9 65.2

(1.3) (1.5) (1.5) (3.1) (3.0) (5.0) (5.8)

Fig. 3 Representative (subject 3; A) and group mean responses (B) of tBH, versus V_E immediately prior to the breath-hold. Curves of best-®t are superimposed. Group mean responses are shown ‹1 SE

con®dent that our subjects performed the breath-hold manoeuvre appropriately. This required familiarisation on several occasions prior to the start of the investigation, after which our subjects were able to perform to a reproducible end-point. In no instance was there any evidence of small respiratory movements (evidenced as air¯ow or lung volume ¯uctuations) towards the end of the manoeuvre, a strategy commonly employed by breath-hold divers as a means of extending dive duration (e.g. Ferrigno and Lundgren 1999). The familiarisation period also minimised any in¯uence of short-term ``apnoeic'' training e€ects (i.e. sequences of breath-holds undertaken with relatively truncated recovery periods, for example 10 min or less) that have been shown to

(1.6) (1.6) (1.3) (1.9) (1.5) (2.4) (2.0)

(1.0) (0.9) (0.8) (1.2) (0.9) (1.6) (1.8)

(1.3) (0.9) (1.3) (1.5) (1.9) (2.2) (1.5)

(1.0) (0.8) (0.9) (1.9) (1.3) (1.5) (1.9)

improve breath-holding ability (Heath and Irvine 1968; Nishino et al. 1996). The characteristics of the ventilatory and tBH responses to exercise do not, of themselves, appear to be surprising; V_E progressively increasing and tBH progressively decreasing as a function of WR (Fig. 2A, C). What we found most surprising in our results, however, was the relationship between these variables. That is, the slope of incremental change of V_E was not associated with a progressively greater decrement of tBH. Rather, the concaveupwards incremental response of V_E was associated with a decremental tBH pro®le that was also consistently concave-upwards (Fig. 2A, C) i.e. the proportionally greater increments of respiratory drive at high WRs were associated with a proportionally reduced shortening of tBH. Consequently, the relationship between tBH and V_E was also concave upwards (Fig. 3). This pattern is consistent with a ``desensitisation'' to the factors contributing to respiratory drive under these conditions. This view is supported by the fact that the subjects were able to breath-hold to progressively higher PETCO2 values and lower PETO2 values with increasing WR (Fig. 2B, D), which is consistent with earlier observations (e.g. AÊstrand 1960; Craig and Babcock 1962; Cummings 1962). For example, in the representative subject shown here, the break-point PETCO2 was 55 mmHg at 30 W, but was some 66 mmHg at 180 W; the equivalent values for PETO2 were 83 and 72 mmHg, respectively. Furthermore, at the higher supra-Thla WRs (e.g. 150 W or more, typically), this was accomplished from pre-breathhold PETCO2 and arterial PCO2 levels that were slightly lower because of respiratory compensation (of the order of 2±4 mmHg at 180 W). It is clear from our results, therefore, that tBH during exercise provides no simple index of either the stimuli that determine a perceived (at least, as reported) level of dyspnoea, or of that level itself. That is, the curvilinear characteristics of the d/tBH relationship (Fig. 5) were such that small increases in pre-breath-hold d values over the lower WR range were associated with large decrements of tBH, while subsequently large increases in d resulted in only small further reductions in tBH. The relatively long tBH observed at high WRs (and, therefore, relative to the prevailing pre-breath-hold V_E and d) resulted in greater magnitudes of the ostensible humoral

277

Fig. 5 Representative (subject 3; A) and group mean responses (B) of dyspnoeic sensation immediately prior to the breath-hold, versus tBH. Curves of best-®t are superimposed. Group mean responses are shown ‹1 SE

Fig. 4 A Representative (subject 3) continuous on-line recordings of dyspnoeic sensation during a breath-hold manoeuvre performed to the limit of tolerance at work rates of 30 W (d), 90 W (m) and 150 W (j), and (B) with best-®t semi-logarithmic (exponential) ®ts. The pro®les have been superimposed to illustrate the shortening and steepening of the dyspnoea pro®le with increasing work rate

stimuli (i.e. end-breath-hold PETCO2 was higher, and PETO2 was lower). One reason to account for the fact that tBH is not a simple proportional function of the prevailing intensity of the pre-breath-hold d may be that the tBH re¯ects not only this ``initial'' condition and the magnitude of the limiting perception (i.e. assumed to be 100%, with the ``100'' being unchanged at the various end-points), but also the rate of change of the stimuli (including both humoral and respiratory-mechanical) during the breathhold (e.g. AÊstrand 1960; Lanphier and Rahn 1963; Mithoefer 1965b). It is this latter component that is likely to be variable at di€erent WRs, and highly variable among subjects at any particular WR. Furthermore, our simple hyperbolic model described earlier for the d/tBH relationship during exercise (i.e. with the

product of dS/dt and tBH being constant, regardless of the value of tBH) would suggest that the value of the curvature constant K is notionally equivalent to the maximum tolerable breath-hold perception. A further complexity in trying to determine the levels of humoral stimuli at the breath-hold limit is that the actual humoral stimuli at their sites of detection (i.e. presumably the conventional sites of central and peripheral chemoreception) would re¯ect the in¯uence of the their transit from the lungs, and would therefore be those obtaining one transit time earlier, at the current levels of vascular ¯ow (e.g. Craig and Babcock 1962; Mithoefer 1965b; Lin et al. 1974). It has been demonstrated that the transit time from the lungs to the carotid bodies is shorter during exercise (Dejours et al. 1958; Lin et al. 1974; Petersen et al. 1978; Ward and Bellville 1983); for example, decreasing from 7±9 s at rest to 4 s at a WR of 125 W (Petersen et al. 1978). However, similar information is not available with respect to the central sites of chemosensitivity, although no di€erence in the V_E response latency to square-wave hyperoxic PETCO2 forcings (i.e. with peripheral chemosensitivity suppressed) between rest and 75 W was discernible in the study of Ward and Bellville (1983). There are also problems concerning the stimulus levels themselves. It is widely recognised that end-tidal gas tensions will be markedly di€erent from the corresponding arterial values during exercise (e.g. Jones et al. 1979; Whipp et al. 1990). As a result of a breath-hold,

278

however, these di€erences are likely to be markedly reduced, if not abolished, on the ®rst (and rapid) postbreath-hold exhalation. This is a because (a) intra-pulmonary di€usional gas exchange during the breath-hold should obviate the e€ects of regional variation of alveolar ventilation-to-perfusion (e.g. Dempsey et al. 1984; Hammond et al. 1986), and (b) the contribution from continuing gas exchange across the alveolar-capillary membrane will be reduced as the alveolar gas tensions approach or, for PCO2, even exceed, mixed venous levels (e.g. Lanphier and Rahn 1963; Mithoefer 1965b). While our interpretations are bound by the ontological constraint of assuming that the perceptions reported as dyspnoea during the exercise are closely related to those that develop during the breath-holding, the assumption that the maximum value reported for the break-point perception (i.e. ``100'' with our visual-analog scale strategy) represents the same integral magnitude of the stimulus intensity associated with the various physiological sources of a€erent drive may well not be the case. That is, the attention to the motor demands of the activity diverting attention from breathing itself might well serve to o€set a component of the acuity of the sensation reported as dyspnoea (Douglas and Haldane 1909; Hill and Flack 1910). It has been reported, for example, that tBH can be prolonged in resting subjects if they are doing mental arithmetic (Alpher et al. 1986) or squeezing a rubber ball (Bartlett 1977; Alpher et al. 1986); interestingly, however, this was reported not to be the case during passive exercise (Craig and Babcock 1962). We believe that this attention-diverting effect of the dynamic exercise may subserve a similar role. A further in¯uence on tBH has been reported to be the involuntary inspiratory muscle contractions that can develop during breath-holding when breathing is avoided by closure of the glottis (e.g. Agostoni 1963; Lin et al. 1974; Whitelaw et al. 1981). Several investigators have argued that such contractions exacerbate respiratory perceptions during breath-holding (e.g. Campbell and Guz 1981; Whitelaw et al. 1981). Whether the in¯uence of such contractions becomes progressively more marked with increasing WR, perhaps re¯ecting the more pronounced baseline respiratory drive, cannot be resolved in the present investigation, however. We found that the kinetics of the perceptual response during the breath-hold to be interesting. The relatively slow development of d early in the breath-hold is consistent with the resting observations of Campbell and Guz (1981) and Nishino et al. (1996). Unlike Flume et al. (1994), however, we found no evidence in our subjects of the variability of the intra-breath-hold pro®le that these investigators referred to as a ``signature''. As expected, the intra-breath-hold d kinetics were consistently faster as the WR increased. However, apart from their consistency, what was surprising about these pro®les was that d rose at all during the breath-hold. That is, subjects were asked to provide ratings that were presumed to re¯ect the ``diculty of their breathing''; however, the subjects provided these responses at a time when no overt

rhythmic respiratory movements were taking place. They therefore consistently appeared to rate the sensation regardless of the question that had been posed. A further complication is the observation that, regardless of WR, subjects did not immediately ``rate'' d at the start of the breath-hold to levels comparable with those obtained at that WR when a breath-hold was not being performed. That is, d increased essentially monotonically from the start-point at the ``0'' base-line through to the ``100'' maximum (Fig. 5), although the primary exercise-related d readings were clearly above ``0''. That is, the e€ect of the deep breath rapidly reduced the prevailing level of d to zero: this was the case at all WRs, although for an interval that was an inverse function of the WR. This suggests that events related to respiratory perception in the context of the breath-hold were, in fact, distinct to some degree from those related to the exercise per se. We were, however, unable to extend our analysis into considerations of the inter-relationships between the pro®le of developing dyspnoeic sensation and the subdivisions of the breath-hold. Two phases of the breathhold have been identi®ed (e.g. Agostoni 1963; Lin et al. 1974; Ferrigno and Lundgren 1999): (a) an initial ``easygoing'' phase in which the glottis is closed and intrathoracic pressure remains stable, the termination of which ± the ``physiological breaking point'' ± is argued to be determined by nonsubjective factors such as arterial PCO2 and lung volume, and (b) a subsequent ``struggle'' phase ± highly in¯uenced by psychological factors ± during which involuntary rhythmic contractions of the inspiratory muscles develop, causing the airway to open and the breath-hold to terminate (the ``conventional breaking point''). We conclude, therefore, that there is no suciently useful relationship between the intensity of the prevailing degree of dyspnoea (as rated on a visual-analog scale) during muscular exercise and the tolerable duration of a volitional breath-hold that would allow the latter to serve as a meaningful descriptor-free index of the exercise-related perception. Interestingly, the relationship between tBH and WR (taken here as a proxy variable for the associated respiratory drive) was, however, consistently such that the proportionally greater rate of ventilatory increase at high WRs was associated with a relative prolongation of the breath-hold. We believe that the attention-diverting in¯uence of the exercise per se is a more likely explanation for the relative prolongation of the breath-hold at high WRs (accounting for the more-extreme levels of alveolar and presumably arterial gas tensions), than the a€erent respiratoryrelated drive from the limbs and/or central command of the motor act serving as inhibitory signals themselves.

References Adams L, Chronos N, Lane R, Guz A (1985) The measurement of breathlessness induced in normal subjects: validity of two scaling techniques. Clin Sci 69:7±16

279 Agostoni E (1963) Diaphragm activity during breath-holding: factors related to its onset. J Appl Physiol 18:30±36 Alpher VS, Nelson RB III, Blanton RL (1986) E€ects of cognitive and psychomotor tasks on breath-holding span. J Appl Physiol 61:1149±1152 AÊstrand PO (1960) Breath holding during and after muscular exercise. J Appl Physiol 15:220±222 Bartlett D Jr (1977) E€ects of Valsalva and Mueller maneuvers on breath-holding time. J Appl Physiol 42:717±721 Beaver WL, Wasserman K, Whipp BJ (1973) On-line computer analysis and breath-by-breath graphical display of exercise function tests. J Appl Physiol 34:128±132 Beaver WL, Lamarra N, Wasserman K (1981) Breath-by-breath measurement of true alveolar gas exchange. J Appl Physiol 51:1662±1675 Campbell EJM, Guz A (1981) Breathlessness. In: Hornbein TF (ed) Regulation of human respiration. Dekker, New York, pp 1181± 1195 Craig AB Jr, Babcock SA (1962) Alveolar CO2 during breathholding and exercise. J Appl Physiol 17:874±876 Cummings, EG (1962) Breath holding at beginning of exercise. J Appl Physiol 17:221±224 Dejours P, Labrousse Y, Raynaud J, Girard F, Teillac A (1958) Stimulus oxygeÁne de la ventilation au repos et au cours de l'exercice musculaire, a basse altitude (50 m), chez l'homme. Rev Fr Etude Clin Biol III:105±123 Dempsey JA, Hanson P, Henderson K (1984) Exercise-induced alveolar hypoxemia in healthy human subjects at sea level. J Physiol (Lond) 355:161±175 Douglas CG, Haldane JB (1909) The regulation of normal breathing. J Physiol (Lond) 38:420±440 Engel GL, Ferris EB, Webb JP, Stevens CD (1946) Voluntary breathholding. II. The relation of the maximum time of breathholding to the oxygen tension of the inspired air. J Clin Invest 25:729±733 Ferrigno M, Lundgren CEG (1999) Human breath-hold diving. In: Lundgren CEG, Miller JN (eds) The lung at depth. Dekker, New York, pp 529±585 Flume PA, Eldridge FE, Edwards LJ, Houser LM (1994) The Fowler breathholding study revisited: continuous rating of respiratory sensation. Respir Physiol 95:53±66 Fowler WS (1954) Breaking point of breath-holding. J Appl Physiol 6:539±545 Hammond MD, Gale GE, Kapitan KS, Ries A, Wagner PD (1986) Pulmonary gas exchange in humans during exercise at sea level. J Appl Physiol 60:1590±1598 Heath JR, Irvine CJ (1968) An increase in breath-hold time appearing after breath-holding. Respir Physiol 4:73±77 Hill L, Flack M (1910) The in¯uence of oxygen inhalations on muscular work. J Physiol (Lond) 40:347±372 Jenkins JS, Valcke CP, Ward DS (1989) A programmable system for acquisition and reduction of respiratory physiological data. Ann Biomed Eng 17:93±108 Jones NL, Robertson, DG, Kane JW (1979) Di€erence between endtidal and arterial PCO2 in exercise. J Appl Physiol 47:954±960 Killian KJ, Summers E, Jones NL, Campbell EJM (1992) Dyspnea and leg e€ort during incremental cycle ergometry. Am Rev Respir Dis 145:1339±1345 Klocke FJ, Rahn H (1959) Breath holding after breathing of oxygen. J Appl Physiol 14:689±693

Lanphier EH, Rahn H (1963) Alveolar gas exchange during breath holding with air. J Appl Physiol 18:478±482 Lin YC (1982) Breath-hold diving in terrestrial mammals. Exerc Sport Sci Rev 10:270±307 Lin YC, Lally DA, Moore TO, Hong SK (1974) Physiological and conventional breath-hold breaking points. J Appl Physiol 37:291±296 Mithoefer JC (1965a) The breaking point of breath holding. In: Rahn H, Yokoyama T (eds) Physiology of breath-hold diving and the Ama of Japan, National Academy of Sciences National Research Council Publication 1341, National Academy of Science, Washington DC, pp 195±205 Mithoefer JC (1965b) Breath-holding. In: Fenn WO, Rahn H (eds) Handbook of physiology. Respiration. Vol II. American Physiological Society, Washington DC, pp 1011±1025 Nishino T, Sugimori K, Ishikawa T (1996) Changes in the period of no respiratory sensation and total breath-holding time in successive breath-holding trials. Clin Sci 91:755±761 Otis AB, Rahn H, Fenn WO (1948) Alveolar gas changes during breath holding. Am J Physiol 152:674±686 Petersen ES, Whipp BJ, Drysdale DB, Cunningham DJC (1978) Testing a model. In: Fitzgerald R, Lahiri S, Gautier H (eds) Control of respiration during sleep and anesthesia. Plenum, New York, pp 335±342 Rodbard S (1947) The e€ect of oxygen, altitude and exercise on breath-holding time. Am J Physiol 150:142±148 Schneider EC (1930) Observation on holding the breath Am J Physiol 94:464±470 Stark RD (1988) Dyspnoea: assessment and pharmacological manipulation. Eur Respir J 1:280±287 Stark RD, Gambles SA, Lewis JA (1981) Methods to assess breathlessness in healthy subjects. Clin Sci 6l:429±439 Sterba JA, Lundgren CEG (1985) Diving bradycardia and breathholding time in man. Undersea Biomed Res 12:139±150 Sutton JR, Jones NL (1979) Control of pulmonary ventilation during exercise and mediators in the blood: CO2 and hydrogen ion. Med Sci Sports 11:198±203 Ward SA, Bellville JW (1983) Peripheral chemore¯ex suppression by hyperoxia during moderate exercise in man. In: Whipp BJ, Wiberg DM (eds) Modelling and control of breathing. Elsevier, New York, pp 54±61 Ward SA, Whipp BJ (1989) E€ects of peripheral and central chemore¯ex activation on the isopnoeic rating of breathing in exercising humans. J Physiol (Lond) 411:27±43 Wasserman K, Van Kessel AL, Burton GG (1967) Interaction of physiological mechanisms during exercise. J Appl Physiol 22:71±85 Whipp BJ, Ward SA, Wasserman K (1986) Respiratory markers of the anaerobic threshold. Adv Cardiol 35:47±64 Whipp BJ, Lamarra N, Ward SA, Davis JA, Wasserman K (1990) Estimating arterial PO2 from ¯ow-weighted and time-averaged alveolar PCO2 during exercise. In: Swanson GD, Grodins FS (eds) Respiratory control: a modelling perspective. Plenum, New York, pp 91±99 Wilson RC, Jones PA (1991) Long-term reproducibility of Borg scale estimates of breathlessness during exercise. Clin Sci 80:309±312 Whitelaw WA, Mcbride B, Amar J, Corbet K (1981) Respiratory neuromuscular output during breath-holding. J Appl Physiol 50:435±443