INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

AFTER VOLUNTARY HYPERPNEA and resultant hyperventilation, healthy adult humans do not develop apnea, despite reduced arterial PCO₂ (12, 24). This phenomenon, referred to as respiratory afterdischarge or ventilatory short-term potentiation (VSTP), appears to be mediated by the activation of brain stem mechanism(s) (9, 17, 25). A major limitation to furthering our understanding of the importance of VSTP resides in the strict dependency on voluntary subject participation to elicit such responses. To overcome this limitation, Georgopoulos et al. (14) elegantly demonstrated the presence of VSTP in normal adult subjects by inducing short hypoxic exposures of 60- to 90-s duration followed by hyperoxia. Because PCO₂ is decreased during poikylocapnic hypoxia, a sudden switch to 100% O₂ should eliminate any previously operative hypoxic ventilatory drive and induce apnea. However, this is not the case, and a gradual reduction in ventilation occurs over the following 6-8 breaths, down to their baseline ventilatory level during previous room air breathing, but not below this level. The slow reduction in ventilation after the hyperoxic switch, instead of the theoretically expected complete cessation of breathing, was termed VSTP, in analogy to similar events that occur after voluntary hyperventilation (14). An additional strategy to assess short-term potentiation (STP) consists in the determination of the ventilatory decay characteristics after the transition from hypoxia to either normoxia or hyperoxia while isocapnic conditions are maintained (7). When this method is employed, the fast ventilatory decrease is considered to depict the loss of hypoxic drive at the peripheral chemoreceptors, while the slow decrease towards normoxic ventilation represents VSTP (23).

Original studies that described VSTP after a short hypoxic exposure did not address whether the actual degree of the hypoxic stimulus was a major determinant of the onset of VSTP and its temporal characteristics and magnitude. If, indeed, VSTP is stimulus-related, such information is crucial to delineation of appropriate physiological models of respiratory control, such that implementation of adequate gain characteristics will allow accurate prediction of respiratory system behavior (20).

In this study, we hypothesized that the degree of isocapnic hypoxia that precedes the sudden switch to isocapnic normoxia would modify VSTP characteristics.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Ten healthy adults (7 men; age 31.0 ± 3.7 yr) were studied after they signed an informed consent in regard to the experimental protocol, which had received approval from the Institutional Review Board.

Pulmonary function tests. To evaluate ventilatory responses adequately, one must initially ensure that no significant mechanical limitation to airflow is present. Therefore, pulmonary function studies were performed in the laboratory, which was located near sea level (mean atmospheric pressure, 761 mmHg). The best forced vital capacity (FVC), forced expiratory volume in 1 s, mean forced expiratory flow (FEF) during the middle half of FVC (FEF_25-75%), and maximal expiratory flow-volume curves were obtained from forced expiration into the wedge spirometer (DSIIa; Collins, Braintree, MA) and were corrected for BTPS. Individual test results were considered abnormal if they were more than ±2 SD from available reference values (6), and these results were thus excluded from the study. Mean FVC was 117 ± 8 (SE)% predicted value, forced expiratory volume in 1 s was 97 ± 3% predicted value, and FEF_25-75% was 88 ± 5% predicted value.

Ventilatory measurements. Subjects were studied while they were awake, sitting comfortably, wearing noseclips, and spontaneously breathing through a mouthpiece. All subjects were visually monitored to ensure that they did not fall asleep. O₂ saturation was continuously measured by pulse oximetry (Nellcor 3000, Hayward, CA). Subjects were connected via the mouthpiece to a Hans Rudolph pneumotachograph and to a two-way nonrebreather valve (Hans Rudolph, Kansas City, MO). PCO₂ was sampled continuously at the expiratory port of the two-way valve and was analyzed breath-by-breath by using a infrared microcapnometer (Columbus Instruments, Columbus, OH). The gas monitor was calibrated with gas mixtures of known CO₂ concentrations. The end-tidal expiratory CO₂ tension (PET_CO2) was held constant throughout the experiment at ~45 Torr. This was achieved by a custom-made microcomputer assembly (Lab-LC, National Instruments, Austin, TX), which provided control signals to the gas-flow controllers so that the CO₂ composition of the inspired-gas mixture could be adjusted to force PET_CO2 at the desired concentration. The dead space of this system was ~85 ml. During each test, expiratory flow was measured by using a heated pneumotach and a pressure transducer (Validyne, Northridge, CA). The signal was calibrated with a mechanically driven pump yielding 1,000-ml stroke volume at a frequency of 10 strokes/min. Corrections were made for changes in gas viscosity that were caused by the changes in O₂ concentration in the inhaled-gas mixture, which was warmed and humidified immediately before the inspiratory port of the two-way, nonrebreathing respiratory valve.

Hypoxic challenges. After an initial 2- to 4-min period of tidal breathing of an isocapnic normoxic gas mixture to establish a baseline, subjects were surreptitiously switched to the designated hypoxic gas mixture by a silent remote, pneumatic valve-switch mechanism (Hans Rudolph). The hypoxic gas mixture consisted of either 6% O₂ in N₂ or 10% O₂ in N₂, to which CO₂ was added by the computer controller to maintain PET_CO2 at the preset level. The hypoxic gas was administered for 60 s, after which time subjects were switched back to isocapnic normoxia. At least three cycles, as described above, were performed by each subject, and recovery periods of ~15 min were allowed between trials. Runs were discarded if PET_CO2 differed by >4 Torr from the individual resting PET_CO2 levels at any given time during the experiment. Hypoxic challenges with 6 and 10% O₂ were administered on different days and in random order.

Data analysis. Data are presented as means ± SD unless otherwise indicated. Respiratory analog data for each run were analyzed on a breath-by-breath basis by a custom-designed peak-trough detection procedure (Wavemetrics, Lake Oswego, OR). Ventilatory measures in isocapnic normoxia and hypoxia were compared by two-tailed paired Student t-tests. For determination of the time elapsed from the time of hypoxia cessation for each breath (t₀), the time at which individual breaths occurred was recorded for each run, and the average time elapsed was retained for each subject. For characterization of VSTP decay, individual E recovery rates while breathing normoxic gas were assessed by computer curve-fitting procedures, employing a two-exponential fit equation as previously described (21). Nonlinear techniques were used to calculate the parameters of the equation where y₀ is the individual baseline E value, A₁ and A₂ represent parameters, and k₁ and k₂ represent fast and slow rate constants, respectively. Improved goodness of the fit was found for a two-order exponential decay equation when individual E recovery is assessed. Thus the time constants (₁E = 1/k₁ and ₂E = 1/k₂) were used to compare E recovery patterns in 6 and 10% O₂. In those instances in which individual baseline E values were not attained, ₂E values were extrapolated on the basis of such asymptotic values.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All 10 subjects completed at least three runs with each hypoxic gas mixture. Mean respiratory measurements in isocapnic normoxia as well as peak ventilatory responses to 10 and 6% O₂ are shown in Table 1. Dose-dependent E increases occurred after the switch to hypoxic gases (P < 0.001, ANOVA) and resulted from similar contributions by VT and RR (Table 1). Mean PET_CO2 was 44.7 ± 1.7 and 45.1 ± 1.5 Torr at the end of the 10 and 6% O₂ hypoxic runs, respectively (Fig. 1; P = not significant).

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Breath-by-breath minute ventilation (E; top) and end-tidal CO2 pressure (PETCO2; bottom) measurements in 1 representative subject after transition to isocapnic normoxia (arrow) after 1-min exposure to isocapnic 10% O2 () and to 6% O2 (). Top: lines indicate 2-exponential best curve fit for each condition.

E after the switch to isocapnic normoxia was best accounted for by a two-order exponential decay equation (Fig. 2). The mean fast time constant (₁E) was 6.7 ± 1.3 s for the isocapnic normoxic recovery from 10% O₂ and 5.9 ± 1.4 s when 6% O₂ was administered (Table 2; P = not significant). In contrast, as shown in Table 2, the slow component of ventilatory decay or VSTP (₂E) was markedly prolonged in 6% compared with 10% O₂ (106.7 ± 11.3 vs. 38.2 ± 6.1 s, respectively; P < 0.0001).

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Ensemble averages of breath-by-breath E decay in 10 subjects after transition to isocapnic normoxia (arrow) after 1-min exposures to isocapnic 10% O2 () and 6% O2 (). Bars, SE.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, we demonstrate that the magnitude of the excitatory respiratory stimulus that precedes its cessation markedly modifies the slow time constant of ventilatory decay, i.e., VSTP, without affecting the fast time constant.

The exact mechanism of VSTP after hypoxic stimuli has yet to be elucidated, but it is probably mediated by activation of a pontomedullary region(s) that receives afferent information from peripheral chemoreceptors. Centrally received information is then processed and modified to sustain ventilation after gas switches, such that a smooth transition from one condition to the other will occur, and large respiratory oscillations, ranging from apnea to hyperventilation, will be prevented (9, 24).

In this study, we selected 1-min hypoxic runs to minimize the central inhibition of hypoxia. Indeed, Dahan and colleagues (7) have recently shown that if 3- to 5-min hypoxic exposures are allowed, VSTP is significantly attenuated, probably reflecting the inhibitory effect of hypoxia on VSTP. Thus hypoxic duration emerges as an important modifier of the ventilatory expression of STP. Our results with mild hypoxic runs of 1-min duration are in close agreement with those reported by Dahan et al., who found a ₁E of 4 s and a ₂E of 30 s. The present study shows that, in addition to the duration of the hypoxic period that precedes the normoxic switch, the magnitude of the hypoxic stimulus will markedly modify the VSTP characteristics.

Several studies have examined VSTP in nonisocapnic conditions. Gleeson and Sweer (18) reported that E decreases below baseline after 10% O₂ if hypoxia was terminated with 100% O₂. However, hypoventilation was not present when room air was employed instead of 100% O₂ to terminate the short hypoxic exposure. Discrepant findings were reported by Georgopolous and colleagues (14), who found no evidence of hypoventilation in nine subjects, when they used 8.5% O₂ followed by hyperoxia. One possible explanation for such an apparent discrepancy could reside in the different magnitude of PET_CO2 decrease in the various studies, whereby the larger the PET_CO2 reduction, the more likely is hypoventilation to occur in a hyperoxic background (14, 18). Thus the confounding contribution of hypocapnia will adversely affect any effort to obtain a true estimate of VSTP, such that tight control of an isocapnic state is necessary for accurate VSTP assessment (10, 11). In addition, it is difficult to compare the results derived from the present experiments that use a normoxic gas to terminate the hypoxic stimulus with those in which hyperoxia was employed (16), since hyperoxia has been shown to add an excitatory dimension to the ventilatory decay characteristics (7).

On the basis of most of the available animal and human experimental evidence on ventilatory STP, we propose that the central mechanism underlying VSTP is stimulus and state dependent and will also be modulated by the duration of the hypoxic exposure (4, 7, 9-11, 14, 18). For example, increased afferent peripheral chemoreceptor stimulation will induce more pronounced STP activation, which will be counterbalanced by the negative effects of more profound hypocapnia (if the PET_CO2 is not maintained constant) or by sleep (4, 7, 9-11, 14, 18).

Two additional considerations deserve comment. First, the actual PCO₂ in brain stem sites, during ventilatory transients such as those performed in our studies, may not be accurately represented by PET_CO2 measurements. If a linear correlation does not exist between PET_CO2 and brain stem PCO₂, the conclusions derived from the current study would be invalid or at least would be seriously challenged. However, we believe that the tight control of PET_CO2 throughout the experiments will prevent large fluctuations in brain stem PCO₂ levels. Nevertheless, despite such precautions, low arterial PO₂ will independently increase brain blood flow in a dose-dependent fashion and will modify tissue PCO₂ accordingly. Thus an attenuation, rather than an increase, in VSTP would have resulted from the latter consideration when 6% O₂ was employed. This obviously did not happen and suggests that, in our experimental setting, such concerns were of minor consequence, if any, to the magnitude of VSTP. Second, given a lung circulation time of 5-7 s (5, 22) and a respiratory frequency of 14-30 breaths/min, the hypoxic stimulus was effectively withdrawn after ~6-8 s of isocapnic normoxia. However, the real time required in each subject at each fraction of inspired O₂ for complete withdrawal of any residual hypoxic drive during the isocapnic normoxic transition was not measured, and this slightly biases the magnitude of ₁E, albeit without detracting from the validity of our comparisons between 6 and 10% O₂. Finally, short-term exposures to hypoxia (e.g., 10 min) have been shown by Gallman and Millhorn (13) to elicit in the cat a long-term facilitation of respiratory output that appears to be diencephalon dependent and will occur only when milder hypoxic stimuli are applied (arterial PO₂ >35 Torr). It is unclear whether such mechanisms that underlie long-term facilitation contributed to our findings.

The evolution of E decay paralleled that of VT, whereas RR seemed to follow a less predictable pattern. Similar findings have been previously described by several investigators in animals (9-11, 25) and in humans (1, 4, 14, 16, 18). In contrast, Fregosi (12) reported that STP was more dependent on RR in a submaximal, steady-state exercise background. Together, these results suggest that pontine locations underlying STP mechanisms may more heavily involve volume- rather than rhythm-related neurons.

Since the original description of VSTP after hyperpnea by Gesell and White (17), this important mechanism has been demonstrated after short-term hypoxia and has been found to be of similar magnitude in young and older humans (1, 14, 24). However, significant reductions in VSTP were reported in adult patients with the obstructive sleep apnea syndrome (15), and, more recently, in patients with congestive heart failure (2, 3). Stimulus dependency of VSTP mechanisms, as shown in the present study, would predict that ventilatory instabilities, such as periodic breathing, would be more likely to occur with more severe O₂ desaturation episodes (3, 4).

In summary, we have shown that ₂E as a measure of VSTP is stimulus dependent. The stimulus dependency of a stabilizing influence on respiratory output such as VSTP will further ensure swift transitions between moment-to-moment changes in excitatory and inhibitory inputs to respiratory drive.