INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

THE DESIGN PRESENTED IS A convenient way to keep end-tidal PCO₂ (PET_CO2) and PO₂ constant, even in the case of rapidly changing ventilation. Such a device is useful in several experimental circumstances, for instance when it is desired to have subjects voluntarily increase respiratory frequency (f) or tidal volume (VT) to examine perceptual consequences (e.g., Ref. 4) or to test respiratory muscle performance (e.g., Ref. 2). In other circumstances, breathing may change spontaneously (e.g., because of anxiety) and can cause undesirable secondary effects, which confound interpretation of the experiment. A prime example of this is found in modern techniques for functional brain imaging. Positron emission tomography and functional magnetic resonance imaging are techniques to measure local cerebral metabolism, thus inferring neural activity. Both techniques depend on changes in cerebral blood flow (1), which is strongly influenced by arterial PCO₂ (3). This contrivance also makes it convenient to measure steady-state ventilatory response to CO₂, as the desired CO₂ level is more easily maintained in the face of consequent changes in ventilation.

Sommer et al. (6) recently described a device that reduces changes in alveolar gases during hyperpnea. Our device, like theirs, limits the amount of fresh gas inspired and supplies additional gas containing higher PCO₂ and lower PO₂ for any additional ventilation. The Sommer et al. device, however, supplies additional gas from a tank containing an approximation of "average" alveolar gas composition, whereas our device supplies the actual alveolar gas expired on the previous breath, thus providing somewhat more precise regulation. Our device not only automatically adjusts for small (±10%) differences in resting PET_CO2 or end-tidal PO₂ among subjects but also allows us to make intentional large changes in PCO₂ (or PO₂) and hold each new level constant.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

Overview. The subject expires into a large tube that serves as an alveolar gas reservoir. Inspiratory gas initially comes from a bag reservoir supplied by a high-impedance constant-flow source. As the subject continues inspiration after the fresh gas from the bag has been exhausted, pressure drops in the system, and a spring-loaded valve at the distal end of the alveolar gas reservoir opens. The subject is thus able to breathe in more gas, but its composition is identical to alveolar gas. The minimum alveolar ventilation at which this device will hold constant PCO₂ is the alveolar ventilation, which establishes PCO₂ with no inspired CO₂.

Contrivance design. The contrivance, depicted in Fig. 1, comprises a manifold (T piece) connected to two tubes and a mouthpiece (or tightly sealed mask). One tube, the fresh gas inspiratory limb, provides gas from a 3-liter anesthetic bag via a one-way valve and a 1-m-long by 3.8-cm-ID tube. Fresh gas from a high-impedance flow (e.g., compressed gas tanks) is continuously fed to the anesthetic bag via a flowmeter; the flow through this meter determines the amount of inspired fresh gas or effective alveolar ventilation. The resistance of the line from bag to mouthpiece should be as low as possible. Depending on the requirements of the experiment, the fresh gas can be air, oxygen, or some other oxygen-containing blend to suit the requirements of the experiment (see Varying gas concentration below). In our laboratory, inspired fresh gas is heated and humidified (SCT 3000 Marquest Medical Products), and expiratory tubing is insulated to minimize cooling and condensation.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Schematic of contrivance. See text for detailed description. Pneumotachometers (P-Tach A and B) are not essential to the function of the system. Tube that forms the reservoir for expired gas is truncated. Solid arrowheads, steady flow; open arrowheads, cyclic flow; arrows with solid lines, inspiratory flow; arrows with dashed lines, expiration.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Pressure-flow characteristic of the rebreathing control valve. Opening pressure and resistance of valve is shown with 5 different spring preload settings. Flow was generated with a variable-controlled vacuum cleaner; each line shows result of slowly increasing power to the vacuum. Leftmost curve is at the usual setting for operation of the system, an opening pressure of 1 cmH2O. RB, flow permitted by the rebreathing control valve.

Operation of the system. Before the subject is connected to the mouthpiece, a normal resting PET_CO2 is measured with a fine nasal cannula while the subject sits quietly (5). It is our experience that subjects almost invariably increase ventilation behaviorally when breathing via a mouthpiece or mask. Thus to achieve normal resting PET_CO2 it is necessary to impose some rebreathing, even in the baseline state. After the subject is connected to the mouthpiece, the flow of fresh gas into the anesthetic bag is adjusted such that during baseline there is a small amount of alveolar gas reinspired from the expiratory limb. This should be confirmed both by visual verification that the bag fully collapses and by inspection of the tidal PCO₂ or PO₂ trace to confirm slight reinspiration of alveolar gas. Because baseline ventilation in the mouthpiece requires some reinspiration to achieve resting PET_CO2, the system can maintain constant alveolar concentrations in the face of small decreases in ventilation, as well as large increases in ventilation.

Varying gas concentration. For some experiments, it is desirable to vary alveolar gas concentration. This is easily accomplished by altering the content of the fresh gas supply. We find it convenient to use a gas blender to supply fresh gas (e.g., Siemens model 965). For instance, in an experiment requiring constant normoxia and variable PET_CO2, one could blend from a gas cylinder containing 21% O₂-balance N₂ and a tank containing 15% CO₂-21% O₂-balance N₂. The most rapid rise in PET_CO2 is accomplished by giving a few breaths of high inspired PCO₂ (PI_CO2) and then gradually reducing the inspired concentration to the required steady-state level. (In perceptual experiments, we generally avoid PI_CO2 greater than PET_CO2 to minimize possible taste cues.) The most rapid fall in PET_CO2 requires not only reducing PI_CO2 to zero but also temporarily increasing fresh gas flow. In experiments not requiring this speed and precision, inspired gas flow could simply be switched between two fixed concentrations.

Safety precautions. It is standard practice in our laboratory never to connect any gas supply to the breathing apparatus that does not contain adequate oxygen to support life. It is also routine to monitor arterial saturation with a pulse oximeter whenever a subject is connected to any breathing circuit.

Test of efficacy. We tested the system's ability to maintain constant PET_CO2 by instructing subjects to target breathing to a metronome at various frequencies while they were being coached to breathe at particular VT values. This provided a range of ventilations from 5 to 60 l/min achieved with eight different combinations of VT and f by using VT values of 0.5-3 liters and f values of 10-40 breaths/min. Transitions between patterns of ventilation were made within a few breaths, and each new pattern was held for 3 min during which PET_CO2 was monitored for deviations from the set level.

Measurements and recordings. Subjects breathed via the mouthpiece and wore a nose clip. Tidal PCO₂ and PO₂ were monitored continuously (Datex-Engstrom Cardio Cap II). Flow was sampled at the mouth and at the rebreathing control valve (Fleisch Pneumotachograph no. 2. and Validyne MP45 with 2-cmH₂O diaphragm). Airway opening pressure was measured at the mouthpiece (Validyne MP45 with 56-cmH₂O diaphragm). Analog signals were digitized at a sample rate of 20 samples · s¹ · channel¹ and recorded by using data-acquisition software (Windaq/200, DATAQ Instruments, Akron, OH).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

Efficacy. Fig. 3 shows pertinent pressures, flows, and tidal gases during a range of conditions of VT and f voluntarily achieved by one subject. In steady-state conditions, the system was capable of maintaining PET_CO2 within ±2 Torr of baseline in the worst case of five subjects tested, and ±1 Torr was the typical performance. We also examined performance when the system was challenged by two- to fivefold rapid transitions between levels of ventilation. We used as a measure of performance the variability of PET_CO2 during the 20 breaths surrounding transitions. The worst case showed a standard deviation of 1 Torr, and the best, 0.2 Torr. Median standard deviation of 25 transitions was 0.5 Torr. Typical transitions are shown in Fig. 4. Our measurements show improved precision of PET_CO2 regulation over those of Sommer et al. (6), but we do not know whether operation of their system could be further improved.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   System variables during 5 breathing patterns generated voluntarily by the subject. TOT, flow at the mouth; RB, reinspired alveolar gas flow from reservoir (i.e., flow permitted by rebreathing control valve, which was set at 1-cmH2O opening pressure); PAW, airway opening pressure measured at the mouth; PCO2, concentration of CO2 measured at the mouthpiece. Tracing shows steady end-tidal level (defined by dotted line) and shows evidence of reinspired CO2 during inspiration. Panels a, b, and c: tidal volume was constant at 1 liter and frequency increased from 10 to 15 to 20 breaths/min, respectively. Panels c and d: frequency was constant at 20 breaths/min and tidal volume changed from 1 to 2 liters. Panels d and e: tidal volume was constant at 2 liters and frequency decreased from 20 to 10 breaths/min. Panels c and e: same ventilation (20 l/min) achieved with very different tidal volume-frequency combinations. Data span the range from 10 to 40 l/min. Inspiratory pressure peaks range from 2 to 4 cmH2O. Note that RB increases with increasing ventilation, whether due to frequency or tidal volume change.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Typical examples of rapid changes () in ventilation. Each breath is represented by 1 data point. Top: change from 12 to 70 l/min, accomplished by increasing tidal volume (VT) from 0.5 to 3 liters at constant respiratory frequency (f). Bottom: change from 26 to 16 l/min accomplished by decreasing f from 30 to 19 breaths/min at constant VT. Notice that end-tidal PCO2 is essentially unchanged throughout these transitions.

Limitations. This device is relatively insensitive to changes in VT and f, because most of the fresh gas is inspired first; however, neither the present device nor that of Sommer et al. (6) can prevent a rise in PET_CO2 if the subject reduces real alveolar ventilation below minimum alveolar ventilation. This may occur if the subject reduces minute ventilation, or if the subject adopts a breathing pattern of very small VT with high f, so that much of the fresh gas penetrates no further than the anatomic dead space. Whereas this pattern of breathing is unphysiological in humans, it implies that the device will not work in animals that pant (although the problem should be ameliorated by the effects of high-f mixing). Furthermore, these devices do not automatically adapt to changes in metabolic CO₂ production.

Troubleshooting common problems. This system has been in use in our laboratory for 2 yr. We encountered several problems while setting up and learning to use this device. Most problems result in a fall in PET_CO2 at high ventilation. Possible causes are as follows. 1) VT in large subjects may exceed the volume of the expired gas reservoir; this is easily remedied by extending the length of the reservoir tube. 2) Inward leaks (especially when a face mask is used) compromise operation by allowing added fresh gas into the system. 3) The threshold valve may open prematurely because of a) valve malfunction or b) excessive resistance in the fresh gas line, which causes the pressure drop to exceed threshold before the bag collapses. This malfunction can be detected by setting fresh air flow just above minute ventilation (so the bag does not fully collapse) and observing whether PCO₂ remains zero throughout inspiration. 4) There is excessive fresh gas flow at baseline.

Conclusion. The contrivance presented here is simple to implement, provides somewhat more precise regulation of end-tidal gases than previous designs, and conveniently allows one end-tidal gas to be independently varied.