INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE ACUTE VENTILATORY RESPONSE to euoxic hypercapnia has been shown to contain two components with differing dynamics (1, 3, 6, 13). The first component is rapid, with a time constant in the range of 8-26 s; it comprises 12-30% of the response and it is normally attributed to the effect of the hypercapnia on the carotid bodies. The second component is slower, with a time constant in the range of 65-180 s; it comprises 70-88% of the response and it is attributed to the effect of hypercapnia at the central chemoreceptors.

However, there is clear evidence that the above dynamics do not describe the whole of the response to hypercapnia and that a slower component exists. The existence of three components within the response had been suggested previously by Gelfand and Lambertsen (7). Khamnei and Robbins (10) showed that ventilation (E) rose throughout an entire 40-min period of constant elevated end-tidal PCO₂ (PET_CO2). No steady state was achieved under these conditions, contrary to predictions from the aforementioned time constants, suggesting that the accepted dynamics are not adequate to describe the whole response to CO₂.

The purpose of the present study was to attempt to determine how long E takes to reach steady state by studying an 8-h period of constant elevated PET_CO2 and to determine whether any slow components are identifiable at the relief of hypercapnia by studying the 8-h period immediately following the 8-h hypercapnia exposure.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Subjects. Ten healthy subjects (6 men, 4 women) aged between 18 and 27 yr volunteered to take part in the study. Their individual acute E-PET_CO2 sensitivities, determined after a 20-min exposure to hypercapnia by using our end-tidal forcing system (9, 11), are given in Table 1. The study requirements were fully explained in written and verbal forms to all participants in such a way that they were naive to the exact purpose of the experiment. Each subject gave informed consent before participation in the study. The research had approval from the Central Oxford Research Ethics Committee.

Experimental technique. The experiment was conducted with individual subjects inside a specially built chamber, which allowed them to be seated comfortably or move around if they wished. The design of this chamber avoided any need for a mouthpiece to control the end-tidal gases. The subjects wore a pulse oximeter attached to a finger to monitor arterial O₂ saturation. Respired gas was sampled (80 ml/min) via fine catheters held at the opening of each nostril by a nasal O₂-therapy mask. The samples were analyzed continuously for PO₂ and PCO₂ by mass spectrometry. The PO₂, PCO₂, and saturation data were sampled by computer at a rate of 50 Hz. Inspiratory and end-tidal values for PO₂ and PCO₂ were identified by computer and recorded for each breath. At the start of the experiment, the inspired-gas composition necessary to produce the desired end-tidal partial pressures was estimated and set manually before the subject entered the chamber. Once the subject had entered the chamber, the inspired composition was altered automatically every 5 min, or at manually overridden intervals, to minimize the error between the actual and desired end-tidal gases. This system has been described in greater detail elsewhere (8).

Measurements of E were made in the chamber by asking the subjects to breathe through a mouthpiece with their nose occluded. The mouthpiece was connected in series with a turbine volume-measuring device to measure respiratory volumes and to a pneumotachograph to record respiratory flows and timing information. Gas was sampled continuously from a port close to the mouth at a rate of 80 ml/min and analyzed by mass spectrometry for PO₂ and PCO₂. A computer was used to identify the ends of inspiratory and expiratory phases from the flow data and to record values for inspiratory and expiratory durations and volumes, for inspired PO₂ and PCO₂ (PI_CO2), and for end-tidal PO₂ (PET_O2) and PET_CO2.

Protocols. After one or two preliminary visits, during which accurate control measurements of PET_CO2 were made and subjects were familiarized with the apparatus, each participant visited the laboratory on two further occasions, each lasting 16 h. On each visit, one of two protocols was performed in a randomly determined order. In the hypercapnic protocol (protocol H), subjects were exposed to 8 h of hypercapnia, with PET_CO2 held at 6-7 Torr above the subject's control value (measured before the start of the experiment) and PET_O2 held at 100 Torr. This was followed by 8 h of poikilocapnic euoxia (PET_O2 = 100 Torr). In the control protocol (protocol C), subjects were exposed to 16 h of air as the inspired gas. E was measured over 5-min periods at the beginning of each protocol and then hourly throughout the exposure and recovery.

Statistical analysis. After the exclusion of the first minute of each ventilation record, the remaining 4 min of data were averaged to give a mean value for each hour of the experiment. The analysis was split into two parts: the first relating to the 8-h hypercapnic period and the second to the 8-h period following the relief of hypercapnia. To determine whether there was any significant change over time in response to hypercapnia, an analysis of variance (ANOVA) was undertaken by using the factors of subject and time. If there was a significant effect of time, then the ANOVA was repeated to determine the time by which a steady state had been reached and whether there were any further significant deviations from this steady state at later times. This was achieved by introducing factors for time sequentially from left to right, as indicated in the following linear model where V_ij is the dependent variable, µ is the mean, S indicates the contribution for a particular subject (index i), h₁-h₇ are the factors that indicate the contribution for hours 1-7 of the protocol (index j), and  is the residual error. Thus, for example, at the introduction of the term h₁, the null hypothesis is that the response does not differ across the 8-h period, whereas the alternate hypothesis is that the response at the first hour differs from that over the remaining hours.

View this table: [in this window] [in a new window]   Table 1.   E-PETCO2 sensitivities for individual subjects

All analyses were undertaken by using the SPSS software package.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Subjects. Of the 10 subjects studied, 8 provided data that were suitable for analysis. One female subject was rejected from the analysis process because of inadequate gas control brought about by technical difficulties, and one male subject hyperventilated unacceptably when asked to breathe through a mouthpiece. While in the chamber, subjects were generally comfortable, spending their time reading, watching television, or playing computer games, although some did report discomfort from the high levels of E during the last hours of hypercapnia.

End-tidal gases. Figure 1 shows the end-tidal gases recorded from each of the eight subjects while they were in the chamber, averaged every 5 min for each of the protocols. The mean values for PET_O2 and PET_CO2 (±SD) based on these 5-min averages are given in Table 2. Overall, good control of the end-tidal gases was achieved.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Control of end-tidal gases in chamber. End-tidal PO2 (PETO2; top), end-tidal PCO2 (PETCO2; bottom) averaged every 5 min from data collected breath by breath over 16 h for all 8 subjects during protocol H (hypercapnia; left) and protocol C (control; right).

Ventilation measurements. Figure 2 shows the mean ventilations measured hourly throughout each protocol and the PET_CO2 values associated with these measurements, both for the individual subjects and for the group means obtained by averaging across all eight subjects.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Hourly inspired PCO2, (PICO2) PETCO2, and ventilation (E) measurements. A: protocol H, results for individual subjects. B: protocol C, results for individual subjects. C: both protocols, data averaged over 8 subjects. , protocol H; , protocol C.

In protocol H, E was considerably elevated by hypercapnia. Much of this increase in E occurred during the first hour of hypercapnia, but there appeared to be a further increase in E during the second hour. ANOVA for the data from hours 1-8 demonstrated a significant effect of time (P < 0.005). After partitioning the variance according to the linear model described in METHODS, E was found to be significantly lower at the end of the first hour of hypercapnia compared with the remaining period (P < 0.001), but there were no further significant variations in E during the rest of the hypercapnic period. This was confirmed by undertaking ANOVA on the data for hours 2-8, for which there was no significant variation. Mean E-PET_CO2 sensitivities were calculated for each subject over the period t = 2-8 h and are shown in Table 1.

There was a slight fall in the PET_CO2 when compared with the target value during measurements made on the mouthpiece. This indicates that there was a degree of hyperventilation present associated with breathing through the mouthpiece. This, however, did not vary significantly either with time or between the protocols (ANOVA).

View larger version (9K): [in this window] [in a new window]   Fig. 3.   Fit of exponential to averaged ventilatory data. , Measured E; , calculated E; line, result of curve fit.

Changes in the chamber PI_CO2 made during these measurements provide further evidence to support the time course indicated by the E measurements. ANOVA indicated that PI_CO2 varied over the data from hours 1-8 (P < 0.005). After the variance was partitioned by using the linear model described in METHODS, PI_CO2 was found to be significantly lower at the end of the first hour compared with the remainder of the hypercapnic period (P < 0.001). There were no further differences during the remainder of the hypercapnic period. Again, this was confirmed by undertaking ANOVA on the data for hours 2-8.

After the relief of hypercapnia in protocol H, and after the large decrease in E between t = 8 h and t = 9 h, ANOVA on the data from hours 9-16 showed that there was no significant effect of time on E. This suggests that the recovery is complete within the first hour following the relief of hypercapnia.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Our results indicate the presence of a third, slow component of the ventilatory response to hypercapnia. However, no progressive changes in E were detected as occurring between t = 2 h and t = 8 h.

The mean ventilatory sensitivity to sustained hypercapnia in our experiments was 4.17 l · min¹ · Torr¹. This compares with values for acute sensitivities to hypercapnia of ~2.1 l · min¹ · Torr¹ from Bellville et al. (1) and of ~1.9 l · min¹ · Torr¹ from Dahan et al. (3). These differences might suggest that the magnitude of this slower component is ~50% of the whole response. However, we found that our particular group of subjects tended to have a higher acute (20 min) sensitivity to CO₂ of 3.05 l · min¹ · Torr¹ than reported in these other studies. Thus it is more likely that the magnitude of the slow component is ~25% of the total response.

Our data are insufficient to provide any very accurate time course for the slow component. However, an exponential curve can be fit to the mean E responses for protocol H, including a predicted chamber E at 20 min, calculated from the sensitivities measured at 20 min by using the end-tidal forcing system (Fig. 3). This yields a magnitude for the slow component as a function of the total response of ~34%, and a time constant of ~1 h. This third component is very different from that suggested by Gelfand and Lambertsen (7). The duration of their exposure to hypercapnia was only 8-10 min, and the time constant of their slowest component, which comprised 88% of the total response, was 89 s.

Our data lack a progressive rise in E over hours 2-8. This finding is consistent with that of Bisgard et al. (2) in the goat in which 4 h of hypercapnia were used, in this case localized to the carotid body.

We have very little idea of what particular mechanism might underlie the slow component of the ventilatory response to hypercapnia. Possibilities include a slow change in the stimulus at the central or peripheral chemoreceptors, either a slow resetting (i.e., left shift of stimulus response relation), or change in the sensitivity of the chemoreceptors, or a progressive neural adaptation in response to the elevated levels of ventilation. In the goat, no progressive change in carotid body discharge was detected with 4 h of hypercapnia localized to the carotid body (5), suggesting that the carotid body is not responsible for this rise in CO₂ sensitivity. Eger et al. (4) did investigate CO₂ responsiveness before and after 8 h of euoxic hypercapnia in humans. There was a suggestion that the slope of the E-PET_CO2 response curve had increased, but this did not reach significance. Smith et al. (12) reported an increase in the basal level of E after 5-10 h (typically 6 h) of forced hyperventilation, which persisted for up to 5 h afterward. Although this observation suggests that neural adaptation to hyperventilation could play some role, the persistence of the effect is not consistent with the more rapid return to baseline E that was observed at the relief of hypercapnia in our experiments.