INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE OVERALL PULMONARY diffusing capacity (DL) is commonly determined by using carbon monoxide (CO) as an indicator. According to Roughton and Forster (19), the overall conductance of the alveolar-capillary CO transfer (DL_CO) can be partitioned

(1)

where Dm_CO represents the conductance of the alveolar-capillary membrane, _CO is the CO uptake conductance of red blood cells per unit blood volume, and Vc is the pulmonary capillary blood volume. Dm_CO can be determined by using the results of DL_CO measurements during hyperoxia and normoxia as well as _CO values derived from in vitro studies. The Dm_CO data as reported by Roughton and Forster are under debate at present. By using morphometric techniques, Crapo et al. (5) found 10-fold higher values. On the other hand, Hsia et al. (12) concluded from theoretical analyses that the Roughton-Forster model provided reasonable estimates of Dm_CO.

Apart from CO, two novel indicator gases have recently been introduced for estimating Dm: doubly labeled carbon dioxide (C¹⁸O₂) and nitric oxide (NO). Both gases are assumed to be quickly eliminated by chemical reactions within the red blood cells: C¹⁸O₂ because of isotopic exchange catalyzed by carbonic anhydrase (21) and NO by very rapid binding to hemoglobin, the reaction of which is 280 times faster than that of CO (4, 14). Therefore it is expected that the uptake of both gases is mainly limited by diffusion. Because the pathways of both gases include the alveolar-capillary membrane, the plasma layer, and the red blood cell membrane, as well as the mean diffusion distance within red blood cells to the point of reaction, overall DL of NO (DL_NO) and C¹⁸O₂ (DL_C¹⁸O2) are similar but not identical with Dm, even when assuming that reaction rates of both gases are infinitely high.

Because the DL_NO/DL_CO ratios in humans (8, 13) and in dogs (14) were higher than the respective ratio of Krogh's diffusion constants, it has been concluded that DL_NO indeed represents the membrane component Dm_NO. However, because the C¹⁸O₂ uptake may be influenced by isotopic exchange reactions (16, 21) and NO may react with tissues (14, 26), the interpretation of DL_C¹⁸O2 and DL_NO determinations has remained uncertain.

To assess which of the two indicator gases yields the closer approximation to Dm, we performed measurements of DL_C¹⁸O2 and DL_NO on rabbits to study the uptake kinetics of both gases as well as the influences exerted by chemical reactions.

	METHODS

Top Abstract Introduction Methods Results Discussion References

The experiments were carried out on six supine rabbits (mean body weight, 3.15 kg; range, 2.8-3.3 kg) that were anesthetized with pentobarbital sodium (19 mg · kg¹ · h¹ iv), paralyzed by alcuronium (0.09 mg · kg¹ · h¹ iv), endotracheally intubated (2.5-3 mm ID), and artificially ventilated by a computerized ventilatory servo system. The ventilator was designed to maintain steady mechanical ventilation and to enable single-breath maneuvers.

Indicator gas mixtures and protocol of experiments. To avoid the formation of NO₂ or its dimer N₂O₄ within the inspiratory gas and to facilitate the comparison of DL_C¹⁸O2 and DL_NO, two O₂-free gas mixtures (containing 0.9% C¹⁸O₂ in N₂ and 0.05% NO in N₂) were prepared and stored in gas-tight flexible aluminum bags. NO from the pure source was led through diluted KOH, subsequently collected with a KOH-containing syringe, and finally injected into the aluminum bag, which had been washed out repeatedly with N₂. To avoid reaction with water, pure C¹⁸O₂ was dried within a trap and led into a bag containing N₂.

Mass spectrometry. We used a respiratory magnetic sector mass spectrometer (M3; Varian MAT, Bremen, Germany) modified to measure isotopic ratios also (23). The relevant gases (NO, O₂, CO₂, Ar, and C¹⁸O₂) were recorded, setting three ion collectors at the following mass-to-charge ratios (m/e): 30 for NO, 32 for O₂, and 44 for CO₂. We determined CO₂ at the first plate of a double-ion collector set at m/e = 44. By repeatedly changing the accelerating voltage (peak jump), we detected C¹⁸O₂ at the second plate of this double collector (m/e = 48). In the same way, Ar (m/e = 40) was measured at the CO₂-44 ion collector. The signal-to-noise ratios for C¹⁸O₂ and NO were 1,800:1 at 8,700 parts/million C¹⁸O₂ and 325:1 at 500 parts/million NO. A particular feature of the analyzing procedure is that dry sample gas was repeatedly compared with a reference gas that differed only in its C¹⁸O₂ or NO content. Drift errors and cross-talk effects were thereby avoided (7, 14), background of the mass peaks was subtracted, and C¹⁸O₂ concentration was obtained in terms of the difference compared with natural abundance.

Calculations for DL. We used the alveolar partial pressures (PA) for C¹⁸O₂ and NO (PA_C¹⁸O2 and PA_NO, respectively) obtained from the end-tidal portion of the gas sample. These values were processed by performing the DL_C¹⁸O2 and DL_NO calculations on the basis of three equations defining gas transfer during inspiration, breath holding, and expiration, as previously described (22). DL was calculated by a trial-and-error approximation method. A detailed description of the underlying model and derivation of the equations is given elsewhere (22).

	RESULTS

Top Abstract Introduction Methods Results Discussion References

A total of 63 single-breath experiments for C¹⁸O₂ were carried out in six rabbits. The lung volume before inflation (RV) averaged 14.1 ± 1.8 (SD) ml, the lung volume after inflation was 64.1 ± 1.8 ml, dead space averaged 6.93 ± 0.9 ml, and static compliance was 2.23 ± 0.2 ml/cmH₂O. The breath-holding periods ranged from 0 to 17 s. In Fig. 1, the ratio of the PA_C¹⁸O2 at overall time t and time 0 (PA_C¹⁸O2/PA_0,C¹⁸O2) is related to the overall time (including inflation, breath holding, and deflation) for the C¹⁸O₂ disappearance from alveolar gas. PA_0,C¹⁸O2 (ranging between 6 and 7 Torr) was derived by calculating the dilution of inspired C¹⁸O₂ within the alveolar volume. The ¹⁸O label was removed according to the relationship

(2)

where t is time in seconds.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Ratio of alveolar partial pressures (PA) of C18O2 at times t and time 0 (PA0,C18O2) as related to overall time (t) available for disappearence of C18O2 from alveolar gas. Data are from 63 single-breath experiments performed on 6 rabbits.

The coefficient 0.9965 was calculated as 1  0.0035, because a reasonable extrapolation to time 0 from the measurements (Fig. 1) was not feasible. During the initial phase (t <3 s), the ratio PA_C¹⁸O2/PA_0,C¹⁸O2 was reduced to <0.01. The DL_C¹⁸O2 was calculated from the fast component, as explained previously (21, 22). The smallest PA_C¹⁸O2 values, measured during the slow phase, came close to 30 times the level of natural abundance of ¹⁸O.

The NO results from 24 measurements are similarly illustrated as a plot of PA_NO/PA_0,NO vs. t in Fig. 2. In each animal, the overall time available for NO removal was set at 4, 6, 8, and 10 s. During a time span of 10 s, >99% of the inhaled NO disappeared from the alveolar gas. Nonlinear regression analysis provided the following PA_NO/PA_0,NO-to-time relationship

(3)

which is plotted as a curve.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Ratio of PA of NO (PANO) at time t and time 0 (PA0,NO) as related to overall time (t) available for disappearence of NO from alveolar gas. Data are from 24 single-breath experiments performed on 6 rabbits.

The DL_C¹⁸O2 and DL_NO data from each animal are shown in Table 1. The overall values (means ± SD) are 14.0 ± 1.1 ml · mmHg¹ · min¹ (for DL_C¹⁸O2; n = 34) and 2.2 ± 0.1 ml · mmHg¹ · min¹ (for DL_NO; n = 24). The reproducibility, determined as the coefficient of variation (SD/mean value) in each animal, was 7.5% for DL_C¹⁸O2 and 4.4% for DL_NO. Because of the ratio of RV-to-inspiratory volume (14/50 ml) in the animals, the end-tidal PO₂ values (as determined within the end-tidal portion of the alveolar gas sample) averaged 21 ± 3 Torr. The far right column of Table 1 contains the DL_C¹⁸O2/DL_NO ratios. The DL_C¹⁸O2/DL_NO ratio averaged 6.3 ± 0.5 (±SD). There was no significant correlation between the individual DL_NO or DL_C¹⁸O2 values and the DL_C¹⁸O2/DL_NO ratios.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Critique of methods. The technique of applying ¹⁸O-labeled CO₂ and NO for determinations of diffusing capacity has been analyzed previously (14, 21, 22). A serious problem was the high removal rate of both gases, in particular of C¹⁸O₂. Therefore, we performed very short single-breath maneuvers and, instead of only considering breath holding (as usually applied for determining DL_CO), the phases of inspiration and expiration were also taken into account. However, there was no significant difference between DL values obtained at different end-expiratory partial pressure levels and with different breath-holding times. This result indicates that the problem of calculating DL from very short single-breath experiments was reasonably solved by applying the three-equation solution (22). More than 99% of inspired C¹⁸O₂ disappears from alveolar space within 3 s (Fig. 1). Thereafter, the small remaining residue is removed very slowly during a second phase. The first phase of label decay is assumed to be primarily caused by alveolar-capillary transfer of C¹⁸O₂; however, the second phase is as yet poorly understood. A similar characteristic has been previously found in humans (22). In this paper, the phenomenon of slow label removal was assumed to be caused by admixtures regained from dead spaces or from the bicarbonate and water pools of the tissues lining the airways and alveoli. Nioka et al. (16) did not find this slow phase of label decay in rebreathing studies of C¹⁸O₂ uptake performed on isolated buffer-perfused guinea pig lungs. This discrepancy could be caused by the fact that the sources of C¹⁸O₂, which account for the effects in single-breath experiments, are unloaded during rebreathing. We determined the DL_C¹⁸O2 from the fast phase of C¹⁸O₂ removal, correcting the respective PA_C¹⁸O2 values by subtracting the partial pressure of the remaining residues. The influence of pulmonary tissues was neglected. This may lead to errors in DL_C¹⁸O2 determinations (21, 22) and is therefore briefly discussed below.

Comparison of DL_C¹⁸O2 and DL_NO determinations. C¹⁸O₂ as well as NO have been applied for assessing Dm, the membrane component of DL (2, 8, 14, 21). The question arises which of the two indicator gases yields the closer approximation to Dm. This question cannot be answered by simply comparing the DL values.

Factors affecting the DL_C¹⁸O2/DL_NO ratio. The almost complete disappearance of C¹⁸O₂ is caused by isotopic exchange reactions that have been described in detail in previous papers (16, 22). In brief, C¹⁸O₂ reacts with water, H₂¹⁶O, forming bicarbonate, HC¹⁸O¹⁶O¹⁸O. During the back reaction, there is a one-third probability of regaining C¹⁸O₂ but a two-thirds probability for producing C¹⁶O¹⁸O and ¹⁸O-labeled water (H₂¹⁸O). This isotopic exchange occurs very quickly within the red blood cells (because of high catalytic activity of carbonic anhydrase) and, to a smaller extent, within pulmonary tissues. This means that the ¹⁸O label and thus, in a broader sense, also C¹⁸O₂, disappears into the large O₂ pool of body water, thereby reducing the back pressure of C¹⁸O₂ to very low values. Because DL_C¹⁸O2 is calculated by assuming the back pressure to be zero, it represents a lower limit of Dm_CO2

(4)

Neglect of removal of C¹⁸O₂ by pulmonary tissues could even lead to an overestimation of DL_C¹⁸O2. The reaction rate of pulmonary tissues can roughly be estimated by comparing the product volume × catalytic factor with that of red blood cells within pulmonary capillaries. Nioka et al. (16) determined the CO₂ reaction rate in guinea pig lung tissue to be 9.2/s, which is 51 times the uncatalyzed reaction rate at 37°C (6). Applying this value to our results, along with reasonable data for tissue volume [13.6 ml (18)], volume of pulmonary capillary blood [3 ml (27)], hematocrit (0.4), and catalytic factor of red blood cells [15,000 (human data); Ref. 6], it can be calculated that the reaction rate of pulmonary tissues amounts to only 4% of the corresponding value in red blood cells. Apart from this estimation, pulmonary tissues also exhibit a capacity, and neglecting it would cause an underestimation of DL_C¹⁸O2. By neglecting reactions as well as capacity, both effects may partially compensate each other. If both effects were significant, DL_C¹⁸O2 determinations should depend on breath-hold duration; this was not the case in the present study.

Influence of inhomogeneities. Determinations of DL are based on the assumption of functionally homogeneous lungs. By simulating rebreathing experiments in studies of the influences of various inhomogeneities, Meyer et al. (14) have shown that sufficiently large inhomogeneities can significantly affect the DL_NO/DL_CO ratio. Because single-breath experiments are supposed to be more strongly influenced by inhomogeneities, and the DL_C¹⁸O2/DL_NO ratio exceeds DL_NO/DL_CO by a factor of two, even larger effects could have affected our experiments. Therefore, we modeled single-breath maneuvers in a manner fairly similar to that described by Meyer et al., investigating unequal distributions of tidal volume, RV, DL, and combinations of these parameters for two parallel alveolar compartments. We obtained decreases in the effective DL_C¹⁸O2/DL_NO below the DL_C¹⁸O2/DL_NO values of a homogeneous lung for various types of inhomogeneities; these were most pronounced when DL values were distributed inhomogeneously. Distributing 80% of DL to the first compartment and 20% to the second compartment produced reductions of DL_C¹⁸O2 by 56%, DL_NO by 36%, and DL_C¹⁸O2/DL_NO by 31% compared with homogeneous conditions. Marked inhomogeneities thus have to be assumed to explain a significant deviation between DL_C¹⁸O2/DL_NO and Krogh's diffusion-constant ratio. Moreover, the inhomogeneities must have been of similar extent in all six rabbits studied; otherwise, we should have found a correlation between the interindividual DL values and the DL ratios. Therefore, we conclude that inhomogeneities of a parallel type could have exerted only a minor influence on the DL_C¹⁸O2/DL_NO ratio.

(5)

1

1

1

1

1

1

Influence of pulmonary tissues on DL_NO determinations. Calculations of diffusing capacity depend on the alveolar distribution volume (VA), which was determined on the basis of the Ar dilution. Previous determinations by Meyer et al. (14), in rebreathing experiments on dogs, have shown that VA for NO exceeded VA for the poorly soluble gas He by up to 36%. They considered three major reasons as possibly accounting for the excess of the VA of NO over the VA of He.

1

1

4

Comparison with data in other studies. On the basis of Eq. 1, Dm_NO can be estimated by using the DL_NO value of the present study, _NO = 4 ml · ml¹ · mmHg¹ · min¹, as calculated from Carlsen and Comroe (3) and pulmonary capillary blood volume, 3 ml, yielding Dm_NO  2.7 ml · mmHg¹ · min¹. However, measurements of red blood cell reaction kinetics by rapid-mixing technique, as previously performed (3), may have been seriously biased by diffusion limitation from unstirred layers (11, 14). Therefore, true _NO may have been considerably higher (8) and Dm_NO would represent an upper limit rather than the true value. This is confirmed when using the data of _O2, as has been measured previously (10). Because both solubility and reaction rate with hemoglobin are greater for NO than for O₂, _O2 is expected to represent an underestimate of _NO, providing _NO >14 ml · ml¹ · mmHg¹ · min¹. By using DL_NO = 2.2 ml · mmHg¹ · min¹ and pulmonary capillary blood volume = 3 ml again, such _NO estimate reveals Dm_NO  2.3 ml · mmHg¹ · min¹, hence representing a value that is <5% higher than the overall DL_NO. Altogether, DL_NO should provide the closest underestimate of Dm that can be obtained from measurements in vivo.