INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The measurements of lung alveolar and dead space volumes using sinusoidally oscillating inert gases have been described in earlier studies (1). Recent studies have shown that inspired forcing oscillations of oxygen in the lung behave physically in a similar manner to a low-solubility inert gas like argon (5, 6). The use of oxygen as the test gas gives the technique a wider acceptability given that oxygen is always present and that it is easier to measure than argon.

In classic gas washout techniques, lung volume is derived from the exponential decay in gas concentration as an inert gas such as nitrogen changes from one steady state to another. The difference with the oscillation technique is that the inert gas concentration is forced to oscillate sinusoidally above and below a steady mean concentration. Measurements of lung volume can be made using this procedure because an inspired oscillation with a known amplitude when breathed into the lungs will be attenuated. The larger the lung volume, the greater is the attenuation of the expired oscillation amplitude. Other factors that attenuate the expired amplitude are the period of the oscillation and the rate of lung ventilation. If these two factors are known, then lung volume can be determined from the attenuation of the expired oscillation.

The current mathematical model used to derive airways dead space and end-expired alveolar volume from an inspired oxygen forcing oscillation assumes that alveolar ventilation is continuous and constant. Thus, the components of alveolar ventilation, tidal volume, airways dead space, and breathing rate are assumed to remain constant. When mechanical ventilation is used, this assumption could be considered valid. However, in spontaneously breathing subjects, both tidal volume and breathing rate vary from breath to breath, and it could be expected that these variations might exclude the use of the technique for measuring lung volumes.

The aim of the present study is to determine whether oscillations in the inspired oxygen concentration can be used to measure lung volumes, in human volunteers, under conditions where tidal volume is physiologically controlled by the freely breathing volunteer rather than a ventilator.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Instrumentation

Gas concentrations were measured using a calibrated mass spectrometer (VG Quadrupoles, Cheshire, UK), and gas flow was measured with a calibrated differential pressure flap-type pneumotachograph (Hamilton Medical Inc., Reno, NV) connected to a pressure transducer. The computer running the mass spectrometer also recorded the flow and gas concentration data at a sampling rate of ~ 40 ms. The sinusoidally varying oxygen concentration oscillation was generated by a microprocessor-controlled gas mixer (7). The gas mixer controls allowed the selection of the desired mean oxygen concentration, sinusoidal amplitude of the oscillation, length of the forcing period, and total gas flow.

Protocol

The gas mixer was set to deliver an FI_O2 of 0.26, with nitrogen used as the balance gas. The amplitude selected for the oscillation in FI_O2 was ± 0.04. For each of the nine adult volunteers (Table 1), the mixed gas was passed at 25 L/min along standard ventilator tubing to a breathing assembly that consisted of a T-piece with one arm vented to atmosphere and the other connected to the inlet of a one-way respiratory flap valve (8) (Figure 1). The volunteers were allowed to breath spontaneously and on inspiration drew breath through the valve into the nose via a nasal CPAP mask (Healthdyne Technologies, Marietta, GA). Between the flap valve and mask was inserted a flow transducer and a bacterial filter which also acted as a heat and moisture exchanger (Humid-vent; Gibeck Respiration AB, Upplands Väsby, Sweden). The mass spectrometer was used to sample gas continuously from a port built into the filter. The expired breath was passed back through the flap valve and exhausted via the outlet port through a length of ventilator tubing into a mixing box and then to the atmosphere. An illustration of the oxygen concentration oscillation under steady-state conditions and during the breathing of a forced inspiratory oxygen concentration sine wave are shown in Figure 2. During steady breathing of the oxygen oscillation, the mean inspired and the mean end- expired oxygen concentrations remain constant, but the breath-by-breath inspired and expired oxygen concentrations oscillate sinusoidally about their mean values with a small amplitude, . The tracing in Figure 2 shows the combined inspired and end-expired oscillations. The upper envelope shows the inspired forcing oscillation interrupted by the breathing pattern. The lower envelope is the resulting end-expired oscillation, which has a lower mean concentration due to oxygen consumption. It also has an attenuated amplitude, in comparison to the inspired oscillation, which is dependent upon the lung volume, alveolar ventilation, and forcing period.

View larger version (16K): [in this window] [in a new window]   Figure 1.   A simplified view of the apparatus used. The arrows indicate the direction of gas and data flow. The flap valves controlling flow are both shown in the open position.

View larger version (48K): [in this window] [in a new window]   Figure 2.   An illustration of the changes in oxygen concentration observed during the breathing of oxygen forcing oscillations in a spontaneously breathing volunteer (Table 1, Subject 5). The left-hand panel shows the inspired and expired breath-by-breath change in oxygen concentration when breathing a gas mixture with a steady FIO2 of 0.27. The upper solid line shows the mean FIO2, and the lower solid line the mean FÉO2. The right-hand panel shows how the breathing pattern is induced by a sinusoidally oscillating FIO2. The FIO2 is forced to oscillate about its mean with an amplitude () of ± 4% over a 2-min period. This input oscillation is fed continuously to the flap valve. On inspiration, this gas is drawn into the lung for a few seconds before being expelled and the next inspiration begins. During the respiratory cycle, the gas oxygen concentration being supplied to the flap valve will have changed. So for each successive breath, the FIO2 will have either decreased or increased. This forms the upper envelope shown. If a continuous line is traced through all the FIO2 values, then the original input oscillation can be reconstructed. While the FIO2 is oscillating, the same amount of oxygen is being consumed as during steady-state oxygen breathing, and so the mean FÉO2 is the same in both cases. However, the amplitude of the resulting FÉO2 oscillation is altered by the airways dead space and alveolar volume. If an envelope is drawn through each individual breath-by-breath FÉO2 value, the amplitude of this oscillation can be seen (lower envelope).

Each volunteer was seated and allowed to breathe a gas mixture containing a steady oxygen concentration for 10 to 20 min. This allowed the resting volunteer to reach equilibrium and to establish a relaxed breathing rhythm through the tubing, valve, and mask. Once the end-expired oxygen concentration became steady (i.e., oxygen consumption was constant), the oxygen concentration and flow data covering 15 breaths were recorded by the computer. The steady-state mixed-expired oxygen concentration in the downstream mixing box was recorded separately by moving the mass spectrometer probe. The measured mean mixed-expired oxygen concentration was checked against the calculated mean mixed-expired oxygen concentration. The gas mixer was then set to deliver consecutively five different oxygen forcing sinusoids with periods of 1, 2, 3, 4, and 5 min. After each change in forcing period, ~ 2 to 5 min was allowed to establish a steady mean before recording at least two cycles worth of data for each forcing period. While the oxygen concentration was being collected, the tidal volume and breathing rate were recorded by the computer. Three repeats of a multiple-breath nitrogen washout procedure (3) were performed at the end of the study. Later, each seated volunteer had his or her thoracic gas volume (corrected to BTPS) measured by whole-body, volume-constant plethysmography (Masterscreen Body; Erich Jaeger GmbH & Co., Würzburg, Germany). This volume is equivalent to the FRC.

Calculation of Lung Volumes Using Oxygen Concentration Oscillations

From the gas data collected during the nitrogen washout procedure, airways dead space (VD) was calculated for each breath from the expired carbon dioxide concentration (9). The subject's FRC was estimated in three ways: by using oscillations of inspired oxygen; by calculating the total mass of nitrogen washed out from the lungs while breathing 100% oxygen, and by whole-body plethysmography. The components of FRC (VD and alveolar volume [VA]) were measured using oxygen oscillations. Before calculating the mixed-expired oxygen concentration, the raw flow and gas concentration data were reprocessed by the computer to realign automatically the flow signal with the delayed gas concentration signals. Typically it took 1.5 s for the gas to travel along the mass spectrometer sampling catheter. It was found empirically that a further fine adjustment of 120 ms was needed to provide the best data alignment. Then, for each breath, the inspired, mixed-expired, and end-expired oxygen concentrations were calculated. These variables, when plotted over time, produced sinusoidal oxygen concentration oscillations (Figure 2). The amplitudes and phase differences of these oscillations were used to calculate airways dead space (Equation 1).

(1)

where P, and PÉ, and PI are the amplitudes of the mixed-expired, end-expired, and inspired sine wave oscillation, respectively, and É and are the phase differences between the inspired forcing oscillation and the end-expired (taken as alveolar gas) and mixed-expired oscillations, respectively (4). In this instance, the mean tidal volume (VT) of all the breaths recorded for all forcing periods was used.

Alveolar volume was determined by first calculating, at any given forcing period, the amplitude ratio between the end-expired and inspired oscillations. This ratio is proportional to the alveolar volume, ventilation rate, and forcing period. When the ventilation rate and forcing period are known, the ratio is dependent upon alveolar volume. The smaller the ratio, the larger is the alveolar volume. This relationship is shown in Equation 2 (3).

(2)

where  = 2/T (T is the forcing period in minutes), _L is the lung ventilatory constant equal to VA/A, P is the pulmonary blood flow in L/min, and  is the gas solubility coefficient in blood. When mechanical ventilation is used the ventilation rate, A (A = [VT  VD] × breathing rate), can be calculated with some certainty, since the tidal volume and breathing rate are held steady by the ventilator. During spontaneous breathing, the tidal volume and breathing rate alter with each breath, and so mean values must be used.

Rearranging Equation 2 gives

(3)

where Z represents the (1 + P/A) term in Equation 2. With a low solubility gas like oxygen, its low value of  (0.02) tends to make Z² close to 1 (6).

This estimate of VA (from Equation 3), derived from a continuous-ventilation mathematical model, is subject to error when the data are obtained from tidally breathing subjects (10). Because of this error, VA is overestimated by the continuous-ventilation model but can be corrected to the "true" alveolar volume, VA', according to Equation 4 (10).

(4)

The corrected, or "true" alveolar volume, VA', is then combined with VD calculated in Equation 1 to derive the FRC. The volumes calculated for the nitrogen washout and oxygen oscillation techniques are at ambient pressure and temperature.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung Ventilation

In the nine volunteers, the overall mean VT was 0.71 L with a range from 0.58 to 0.92 L and the overall mean breathing rate was 12 with a range from 9 to 16 breaths/min (Table 1). Within each data collection period, VT varied little and the mean SD was 0.04 L (Table 1). The mean difference between the FI_O2, 0.259 ± 0.089 (range: 0.245 to 0.269), and the mean FÉ_O2, 0.215 ± 0.017 (range: 0.183 to 0.238), was 0.044 ± 0.016.

Airways Dead Space

A comparison between the airways dead space measured by the single-breath carbon dioxide technique, VD_SB, and by the oxygen oscillation technique, VD_OX, at different forcing periods, shows the optimal forcing period is 3 min, as this provides the closest agreement between the two measurements (Figure 3). At this period, and after subtracting the nasal mask volume, the means for VD_SB and VD_OX were 0.19 ± 0.03 and 0.17 ± 0.04 L, respectively. These values were not significantly different (p < 0.05) (Figure 4). VD_OX at forcing periods longer than 3 min underestimated VD_SB by 18%, while at the shortest forcing period of 1 min (not shown on Figure 3) the underestimation increased to ~ 45%.

View larger version (11K): [in this window] [in a new window]   Figure 3.   A comparison between airways dead space measured by the single-breath carbon dioxide technique and the oxygen oscillation technique at inspired forcing periods of 2, 3, 4, and 5 min. The ordinate shows the percentage difference, [(VDOX  VDSB)/ VDSB] × 100, between the two techniques. The mean ± SD is shown for nine volunteers. Data for the 1-min forcing period are not shown. See text for details.

View larger version (10K): [in this window] [in a new window]   Figure 4.   A comparison between airways dead space measured by the single-breath carbon dioxide technique and the oxygen oscillation technique at the optimal forcing period of 3 min. The solid line shows the mean difference between the two techniques. The abscissa shows the mean VD, (VDSB + VDOX)  2, of the two techniques.

Alveolar Volume

By subtraction of VD_SB from the FRC measured by the multiple-breath nitrogen washout technique (MBNW), the alveolar volume. VA_MBNW, was calculated and then compared with the end-expired alveolar volume, VA_OX, measured using inspired oxygen oscillations (Figure 5). At all forcing periods, VA_OX was within 85% of VA_MBNW and, with a forcing period of 3 min, the means were not different (p > 0.05) at 3.11 ± 0.7 and 2.87 ± 0.82 L, respectively (Figure 5).

View larger version (13K): [in this window] [in a new window]   Figure 5.   A comparison between the end-expired alveolar volume measured by inspired oxygen oscillations with different forcing periods and the calculated alveolar volume. Alveolar volume was calculated by subtracting VDOX from the VAMBNW. The ordinate shows the percentage difference between the two techniques, [(VAOX  VAMBNW)/VAMBNW] × 100. See text for abbreviations and Figure 3 for plot details.

Functional Residual Capacity

The FRC measured by plethysmography (FRC_PLETH) was compared with the FRC measured by the MBNW technique (FRC_MBNW), and this is shown in the upper panel of Figure 6. The mean FRC_MBNW was significantly (p < 0.05) smaller than the FRC_PLETH, with the mean FRC_PLETH being 3.34 ± 0.84 L (range: 2.15 to 4.40 L) compared with the FRC_MBNW of 3.00 ± 0.85 L (range: 1.86 to 4.23 L).

View larger version (7K): [in this window] [in a new window]   View larger version (7K): [in this window] [in a new window]   Figure 6.   Upper panel: A comparison between FRC measured by whole-body plethysmography and by the multiple-breath nitrogen washout technique (closed squares). The abscissa is the mean FRC measured by the two techniques. The dotted line shows the zero percent difference line, and the solid line shows the mean difference between FRCPLETH and FRCMBNW. Lower panel: A comparison between FRC measured by whole-body plethysmography and by the oxygen oscillation technique, at a forcing period of 3 min (open circles). The abscissa is the mean FRC measured by the two techniques. The dotted line shows the zero percent difference line, and the solid line shows the mean difference between FRCPLETH and FRCOX.

The FRC_PLETH was also compared with the FRC measured using the inspiratory oxygen oscillations (FRC_OX), with the latter FRC calculated by adding together the optimal VD_OX and VA_OX, as shown in the lower panel of Figure 6. The mean FRC_OX of 3.28 ± 0.72 L (range: 2.03 to 4.22 L) was not significantly different (p < 0.05) from the mean FRC_PLETH (Figure 6, lower panel). The FRCs measured by the MBNW and inspiratory oxygen oscillations were not significantly different (p < 0.05) from each other.

During these studies, it was noted that the spontaneous breathing rate was inversely related to the FRC (FRC_OX r² = 0.62, FRC_PLETH r² = 0.76). No significant relationships were found between the FRC and VT or A or the other variables shown in Table 1.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The sitting volunteers found that breathing through the nasal mask and tubing was harder than normal breathing, but once the breathing rhythms steadied after ~ 15 min they breathed comfortably, with the tidal volume varying on average by 100 ml and the breathing rate by 2 breaths/min during the data collection period. Another indication of stability was that the mean end-expired oxygen concentration remained steady throughout the study (see Figure 2). If the present breathing system was used on volunteers who required a higher ventilation rate, then the resistance to breathing might become a problem. However, this could be overcome by using wider-bore tubing.

The period dependency of the VD measurements, derived by the oxygen oscillation technique, is the same as that observed in studies of mechanically ventilated dogs, with the optimal forcing periods being between 2 and 3 min in both instances (3, 6). Although dogs have smaller lungs than humans, they also have lower lung ventilation rates, and so the ratio between ventilation and volume is similar to that seen in humans. The underestimation of VD when using a 1-min oxygen oscillation is also similar to earlier findings (3, 4). The error at short forcing periods possibly results from the problems of accurately measuring the amplitudes and phase angles of the end-expired and mixed-expired oscillations, which at shorter periods become progressively smaller. At shorter forcing periods, the expired variables in Equation 1 progressively diminish. As a result, small measurement errors in the expired gas concentration and flow can lead to large variations in the estimation of VD. The measurement of VA by FI_O2 oscillations was possible at all the forcing periods used, although with the 1-min period the measurement error was also at its greatest (Figure 5).

A comparison of the three measuring techniques show that plethysmography and oxygen oscillations provide the same measurements of FRC (Figure 6). Other studies have shown that the nitrogen washout technique measures a smaller FRC than does plethysmography (11, 12). The mean difference of 0.4 L between FRC_PLETH and FRC_MBNW in this study could be due to FRC_PLETH being overestimated because of trapped gas either in the gastrointestinal tract or in the measuring box, but we have no way of verifying this hypothesis. The mean FRC of 3.3 L obtained by the oxygen oscillation technique in these studies compares well with the FRC of 2.9 L measured in 13 sitting healthy adults of a similar age, height, and weight (13). Earlier estimates of FRC were reported to lie between 2.5 and 2.75 L but these measurements seem to have been made while the subjects were supine. The FRC has been shown to decrease under these conditions (13).

The exact relationship between breathing rate and FRC is a complex one and is related to the inspiratory and expiratory time and the tidal volume (14). This relationship implies that subjects with the largest FRC have the slowest breathing rates. A simple explanation for this observation is that those with the greatest FRC also have the greatest volume available for gas exchange, and thus critical carbon dioxide concentrations that stimulate ventilation would take longer to be reached.

In healthy adults, our studies have shown that the oxygen oscillation technique can be used to measure simultaneously the component parts of FRC (airways dead space and alveolar volume). The advantage of this technique over other techniques for measuring lung volume is that it has the potential to be used on a continuous basis without interruption of the subject's ventilation pattern. Providing that oxygen consumption remains constant, a measure of the components of FRC can be updated at the optimal forcing period of 3 min. Although it is possible to make multiple measurements with the two comparator techniques used in this study, the MBNW technique requires several minutes between tests to reset the nitrogen concentration and plethysmography requires the subject to make conscious breathing maneuvers for the measurements to be made.

In conclusion, the studies show that it is possible to extend the use of inspiratory forcing oscillations of oxygen to measure accurately lung volumes in spontaneously breathing adults.