Airway lengthening after pneumonectomy (PNX) may increase diffusive resistance to gas mixing (1/DG); the effect is accentuated by increasing acinar gas density but is difficult to detect from lung CO-diffusing capacity (Dl CO). Because lung NO-diffusing capacity (Dl NO) is three- to fivefold that of Dl CO, whereas 1/DG for NO and CO are similar, we hypothesized that a density-dependent fractional reduction would be greater for Dl NO than for Dl CO. We measured Dl NOand Dl CO at two tidal volumes (Vt) and with three background gases [helium (He), nitrogen (N2), and sulfur hexafluoride (SF6)] in immature dogs 3 and 9 mo after right PNX (5 and 11 mo of age). At maturity (11 mo), background gas density had no effect on Dl NO, Dl CO, or Dl NO-to-Dl CO ratio in sham controls. In PNX animals, Dl NO declined 25–50% in SF6 relative to He and N2, and Dl NO/Dl CO declined ∼50% in SF6 relative to He at a Vt of 15 ml/kg, consistent with a significant 1/DG. At 5 mo of age, Dl NO/Dl CO declined 25–45% in SF6 relative to He and N2 in both groups, but Dl CO increased paradoxically in SF6 relative to N2 or He by 20–60%. Findings suggest that SF6, besides increasing 1/DG, may redistribute ventilation and/or enhance acinar penetration of the convective front.
- lung diffusing capacity
- carbon monoxide
- sulfur hexafluoride
in immature animals after pneumonectomy (PNX), the remaining lung rapidly expands into the empty thorax, followed by an accelerated increase in alveolar septal tissue volume (32), number of acinar airway segments and branch points (13), and the length and eventually diameter of conducting airways (4). Intra-acinar airways also lengthen. Because gas transport by diffusion becomes more important than by convection in distal acinar airways, post-PNX airway lengthening and reduction in total cross-sectional area may significantly increase the resistance to peripheral gas mixing by diffusion (1/DG) or reduce the gas phase conductance of an inspired bolus. Robertson et al. (25) showed that the fractional retention of an intravenously infused inert gas of high molecular weight is significantly greater than that of an inert gas of low molecular weight but similar partition coefficients, which suggests a diffusion-related mechanism of gas elimination. Using the multiple inert-gas elimination technique (MIGET), we found preferential alveolar retention of enflurane, the gas of the highest molecular weight used, in pneumonectomized adult dogs during heavy exercise, consistent with an increased 1/DG after PNX (12). However, the effect of 1/DG on lung diffusing capacity (Dl) has not been directly estimated.
Theoretically, 1/DG occurs in series with diffusive resistances of the tissue-plasma membrane barrier (1/DM) and the blood (1/Θ · Vc). Normally, 1/DG is considered negligible and has been omitted from the Roughton-Forster model of estimating Dl (27), i.e. Equation 1where Dm is membrane diffusing capacity, Θ is the rate of gas uptake by whole blood, and Vc is capillary blood volume. Increasing density of the background inert gas in distal airways will increase 1/DG and decrease Dl as long as distribution of ventilation or depth of acinar penetration of inspired CO or NO used to measure Dl remains unchanged by the change in gas density.
Because CO Dl (Dl CO) is equally sensitive to resistances offered by the alveolar tissue-plasma barrier and by erythrocytes, i.e., 1/Dm CO≈ 1/(ΘCO · Vc) (9,27), an increased 1/DG would cause a variable overestimation of Dm CO and Vc by the Roughton-Forster method that uses Dl CO measured at two levels of alveolar O2 tension. Recently, we were able to measure NO Dl (Dl NO) simultaneously with Dl CO by using a modified rebreathing technique (33). Because of the extremely rapid reaction rate of NO with hemoglobin, ΘNO becomes large enough that the term for erythrocyte resistance to NO uptake is extremely small, i.e., Dl NO predominantly reflects the resistance of the tissue-plasma barrier (8, 18,33) and any gas phase resistance Equation 2Dl NO is about four times that of Dl CO, but CO and NO have similar gas-phase diffusion coefficients, so that 1/DG for the two gases is approximately the same and the magnitude of change in 1/DGas gas density is altered is similar as well. Consequently, any change in 1/DG will cause about a fourfold greater fractional change in Dl NO than in Dl CO. Because of this amplified effect of gas density on Dl NO relative to Dl CO, it may be possible to detect changes in gas-phase resistance to diffusion for NO at rest that cannot be detected for CO. It was also anticipated that increased 1/DG would be easier to detect with a shallow than a deep inspired volume since the inspiratory convective front lies farther from the alveolar membrane with a shallow breath.
On the basis of the above reasoning, we measured Dl NO and Dl COsimultaneously at two time points in immature dogs raised to maturity after right PNX using three background gases of different densities [helium (He), air (N2), and sulfur hexafluoride (SF6)] at two tidal volumes (15 and 45 ml/kg). We hypothesized that post-PNX elongation and dilatation of intra-acinar as well as conducting airways would increase 1/DG, which can be detected by a density-dependent reduction in Dl NO, whereas Dl COwould be relatively unaffected.
All procedures were approved by the Institutional Animal Care and Use Committee. Litter-matched male foxhounds underwent right PNX (n = 5) or right thoracotomy without lung resection (sham, n = 5) at 2 mo of age. At 3 (age 5 mo,n = 5 per group) and 9 mo (age 11 mo, n= 3 per group) after surgery, animals were anesthetized, intubated with a cuffed endotracheal tube, placed in the supine position, and mechanically ventilated sufficiently to eliminate spontaneous breathing. A latex balloon-tipped polyethylene catheter was inserted into the distal third of the esophagus, which was attached to a Validyne pressure transducer and Hewlett-Packard amplifier to measure changes in esophageal pressure. An identical setup was used to measure mouth pressure. An oximeter probe was placed on the tongue to monitor heart rate and O2 saturation.
The rebreathing test gas mixture, consisting of 0.3% CO, 0.3% methane, 0.8% acetylene, and 21% O2, in a balance of either He, N2, or SF6, or in a balance of O2, was humidified and drawn into a 6-liter Mylar reservoir bag. Just before each measurement, 40 parts/million of NO was added to the reservoir bag and thoroughly mixed via a mechanical pump. The desired volume of test gas mixture (15 or 45 ml/kg plus apparatus dead space) was drawn into a 3-liter calibrated syringe connected to the endotracheal tube via a manifold. The dog was first ventilated with the appropriate background gas (room air, or 21% O2 in balance of He, or 21% O2 in balance of SF6) for at least 3 min to ensure equilibration with resident alveolar gas. At a selected end expiration, the expiratory port was occluded. The animal was given three cumulative breaths of 15 ml/kg each (totaling 45 ml/kg) via the ventilator to expand the lungs and was then allowed to exhale passively to functional residual capacity. Next, the test mixture was delivered into the lung through the endotracheal tube, and the animal rebreathed in and out of the calibrated syringe for 16 s at a rate of 30 breaths/min. Gas concentrations at the mouth were continuously monitored by a chemiluminescence NO analyzer (NOA280, Sievers Instruments, Boulder, CO), infrared gas analyzers for CO, methane, and acetylene (Sensors, Saline, MI), and a Perkin Elmer 1100 mass spectrometer for O2, CO2, and N2. Analyzers were calibrated on the day of study according to manufacturer's specification. During rebreathing, all signals were digitized by computer.
All rebreathing measurements were done in duplicate. The order of the background gas (He, N2, or SF6) and delivered volume (15 or 45 ml/kg) was randomized. A baseline venous blood sample was drawn before beginning the experiment and at the end of measurements using each background gas; blood was analyzed for total hemoglobin, carboxyhemoglobin, and methemoglobin concentrations (OSM3, Radiometer, Copenhagen, Denmark). The instrument was calibrated for dog blood. Hematocrit was measured with a capillary tube centrifuge.
Lung volumes were calculated from methane dilution and expressed at btps conditions. Cardiac output was calculated from the exponential disappearance of acetylene with respect to methane, corrected for the intercept of CO disappearance (29). Dl CO and Dl NO were calculated from the exponential disappearance of CO and NO, respectively, with respect to methane (28, 33). Dl CO was corrected to a standard hemoglobin concentration of 14.6 g/dl by the method of Dinakara et al. (6). Results were normalized by body weight, expressed as means ± SE, and compared between groups with respect to background gases by repeated-measures ANOVA. Differences were considered significant if P < 0.05.
Table 1 shows body weight, hemoglobin concentration, and lung volumes measured by the rebreathing technique. Average body weight was higher in PNX animals 9 mo after surgery than in corresponding sham animals. However, the average weight gain between 4 and 10 mo after surgery among the three animals that were studied at both time points was similar (64 ± 11% for PNX group and 42 ± 13% for sham group; P = 0.25 by repeated-measures ANOVA). Hemoglobin concentrations were similar between groups and ages. Mean alveolar Po 2results during rebreathing, end-expiratory, and end-inspiratory lung volume per kilogram of body weight were similar between groups at both delivered volumes and both ages.
Table 2 shows the physical characteristics of the inspired background gas mixtures used to prebreathe the animals before measurement of Dl CO and Dl NO by the rebreathing technique.
Table 3 shows Dl NO measured in different background gases and at different intervals after surgery. In sham animals, Dl NO increased with increasing tidal volume. At each tidal volume, Dl NO was not significantly affected by changing background gas density. In PNX animals, the absolute magnitude of Dl NO was similar to that in sham animals at each tidal volume; however, there was a statistically significant density-dependent decline in Dl NO from a He background gas mixture to an SF6 background at a tidal volume of 45 ml/kg 3 mo after PNX and at a tidal volume of 15 ml/kg 9 mo after PNX. For the most part, Dl NO measured in a N2 carrier was intermediate between that in a carrier of SF6 and He commensurate with the expected continuity based on the diffusion coefficients for NO in the different carriers.
Table 4 shows Dl CO measured in different background gases and at different times after surgery. Increasing tidal volume caused a small increase in Dl CO in both sham and PNX dogs. There were no significant differences in Dl CO between measurements made in a background of He and those in a background of N2. However, there was an unexpected density-dependent increase in Dl CO in both PNX and sham animals between measurements made in a background of SF6 and that in a background of He or N2 at 3 mo after surgery (5 mo of age).
The ratio of Dl NO to Dl CO measured in a background of N2at a tidal volume of 45 ml/kg averaged 4.06 ± 0.31 for sham animals and 4.31 ± 0.70 for PNX animals 3 mo after surgery (P > 0.05 between groups); the corresponding ratios 9 mo after surgery were 2.84 ± 0.46 for sham and 3.38 ± 0.23 for PNX animals (P > 0.05 between groups; Table5). In both sham and PNX groups, the Dl NO-to-Dl CO ratio was significantly lower in the high-density background of SF6at both tidal volumes 3 mo after surgery compared with a background of N2 or He. At 9 mo after sham surgery, the Dl NO-to-Dl CO ratio was no longer affected by background gas density; however, at 9 mo after PNX, there was still a marked density-dependent decrease in Dl NO-to-Dl CO ratio at a tidal volume of 15 ml/kg in an SF6 background compared with the same tidal volume in a background of N2 or He (P < 0.05).
It is unlikely that airway NO exchange significantly influenced the results in Tables 3 and 5. Diffusing capacity and production of NO from airways (23, 36) are ∼100 times and 2–4 times smaller than Dl NO and alveolar NO production, respectively. Virtually all resistance to NO uptake resides in the alveolar membrane, whereas at rest half of the resistance to CO uptake is in the membrane and half is in capillary blood (9, 27). Membrane diffusing capacity increases with lung volume, but Vc does not (19); hence, Dl NO increased much more with lung volume than Dl CO.
Gas Transport in the Airways
Down to the 8th generation of conducting airways, gas transport is almost purely convective (1, 34, 38). During laminar flow, a cone-shaped concentration front carries inspired gases deeper into more distal conducting airways, where mixing occurs by a combination of axial convection and radial diffusion (Taylor's dispersion) down to about the 12th generation (1, 38). Axial diffusion becomes increasingly important at higher generations. Interactive diffusion and convection near terminal bronchioles also facilitate gas mixing by making mean gas tensions more homogeneous among acini. Within small acinar airways distal to terminal bronchioles, the convective concentration front becomes so slow as to be almost stationary; beyond this quasistationary front, axial diffusion dominates intra-acinar gas transport and mixing (20,38). Intra-acinar diffusion facilitates a more homogeneous distribution of alveolar gas tensions than predicted by bulk flow alone (20). Model simulations by Paiva and Engel (21,22) suggest that inhomogeneity of gas concentrations within a single acinus, although present at rest normally, is functionally small and contributes little to the alveolar-arterial O2 tension difference (A-aDo 2). The normal acinus functions as a single gas exchange unit, consistent with reports by Heller and Schuster (10) in rabbits, by Worth et al. (41) in adult dogs, and in our present results.
Gas Phase Resistance to Diffusion
Theoretically, diffusive transport within the acinus could be impaired 1) when bulk penetration of inspired gas is inadequate so that the convective front lies a greater distance from the alveolar membrane or 2) when acinar airways are elongated, as by PNX, leading to a longer diffusion path to the alveolar membrane. In both instances, diffusing capacity would be reduced. Cotton and colleagues (3, 7) varied the penetration of an inspired bolus by changing the breath-hold time and preinspiratory lung volume and found little overall effect on the three-equation single-breath Dl CO(Dl -3EQ). However, when the expirate was divided into timed samples, increasing breath-hold times increased Dl CO estimated from the early alveolar samples but reduced Dl CO estimated from the later alveolar samples (30). In patients with asthma and emphysema, Dl -3EQ increased with increasing breath-hold time and correlated with the phase II slope of the single-breath N2 washout curve (7). These findings were interpreted as indicating impaired acinar gas transfer. Thus peripheral inhomogeneity in gas mixing exerts a small but detectable effect on Dl CO even in normal humans, and the effect is exaggerated in lung disease. Theoretically, the effect of inhomogeneity on Dl CO should be minimized by the rebreathing technique where better gas mixing is achieved during measurement.
Expected Effects of Changing Gas Phase Density
We expected that raising acinar gas density should increase 1/Dg for both NO and CO similarly. However, because Dl NO normally is about fourfold greater than Dl CO, the fractional decline in Dl NO from a given change in gas density should be about fourfold greater than that in Dl CO. The ratio of Dl NO to Dl CO should correspondingly fall as acinar gas density increases, as shown in Table 5. On the other hand, increasing acinar gas density with SF6 has been repeatedly shown to enhance resting O2 transport, as evidenced by a progressive reduction of A-aDo 2, as gas density increases (2, 14, 17, 39, 41). Increasing gas density also improves resting steady-state CO uptake in the lung (5, 16), although single-breath Dl CO in adult dogs is not affected (41). That raising acinar 1/Dgshould improve efficiency of O2 and CO exchange seems counterintuitive. Two explanations have been suggested to explain this:1) deeper convective penetration into the acinus and2) improved gas mixing among acini.
Deeper convective penetration into the acinus.
Increasing gas density raises resistance to radial diffusion (Taylor's dispersion) (38) and delays obliteration of the cone-shaped concentration front within conducting airways, leading to a more distal convective penetration of the concentration front, i.e., closer to the alveolar membrane, thereby reducing the path length for gas phase diffusion and lowering 1/DG. This mechanism opposes the density-dependent increase in 1/DG caused by a lower diffusion coefficient for NO and CO. The net effect depends on which mechanism dominates. Altered Taylor's dispersion has been invoked to explain the density-dependent increase of steady-state CO uptake observed in human subjects (5, 16). However, model simulations and experimental data (10, 20, 37) suggest that normal resting A-aDo 2 is predominantly caused by ventilation-perfusion heterogeneity among acini rather than by diffusion limitation within an acinus. Hence, this explanation does not fully account for the reported effect of SF6 on resting A-aDo 2.
Improved gas mixing among acini.
Simulation by Paiva and Engel (20, 21) suggests that at branch points of distal conducting airways near terminal bronchioles, a combination of convection and diffusion enhances gas mixing among acini that share a common dead space. Because SF6 has a lower kinematic viscosity than either N2 or He and in addition to its higher density, SF6 will generate greater turbulence at branch points despite a low Reynold's number. These properties of SF6 can enhance interacinar gas mixing and reduce A-aDo 2 by improving uniformity of respired gas concentrations among acini without impairing intra-acinar diffusive O2 uptake. A similar mechanism could explain the density-dependent enhancement of steady-state CO uptake at rest. Because steady-state CO uptake is highly sensitive to nonuniformity of alveolar ventilation with respect to Dl CO(15), improving interacinar gas mixing could enhance steady-state CO uptake and reduce A-aDo 2.
Gas Phase Resistance During Maturation and After PNX
In sham animals at full maturity (11 mo of age), altering gas density had no effect on Dl NO or Dl NO-to-Dl CO ratio, i.e., negligible 1/DG at either tidal volume. In PNX animals at full maturity, Dl NO and Dl NO-to-Dl CO ratio at a 15 ml/kg tidal volume progressively fell (P < 0.05), whereas Dl CO remained unchanged as gas density increased. These response patterns are consistent with our hypothesis of an increased 1/DG after PNX, which is only detected at a low tidal volume but not at a higher tidal volume (45 ml/kg) presumably because bulk acinar penetration of the inspirate is faster and deeper at a high tidal volume. The patterns of density dependence of Dl NO in immature (5 mo old) sham and PNX animals are similar to those observed in their respective adult groups. One difference is that in 5-mo-old PNX animals, an increased 1/DG is evident only at a high tidal volume (45 ml/kg) but not at a low tidal volume (15 ml/kg), although the Dl NO-to-Dl CO ratio declined significantly with increasing gas density at both tidal volumes.
In contrast to Dl NO, the response of Dl CO to changing gas density in immature dogs differed from that in adult dogs. In both sham and PNX animals at 5 mo of age, Dl CO increased by an average of 50% as gas density increased, whereas Dl NO either remained unchanged or decreased. Dl NO-to-Dl CO ratio declined by an average of 35% with increasing gas density. The magnitude of density-dependent increase in Dl COwas similar regardless of tidal volume or whether PNX had been performed; this effect totally disappeared by 11 mo of age. Worth et al. (41) also found no effect of SF6 on Dl CO in adult dogs. We propose two possible explanations for the paradoxical increase in Dl CO. 1) As already discussed, a high gas density may enhance acinar gas penetration by impeding Taylor's dispersion, thereby reducing 1/DG and increasing Dl CO. The effect would be more pronounced in the smaller lungs of immature animals, particularly if conducting and acinar airways grow at different rates and reach maximal length and cross-sectional area at different times during maturation.2) Increasing gas density can alter the distribution of ventilation by increasing the turbulent resistance due to convective acceleration between airways of different diameters (26) and by increasing the resistance imposed by changing directions of flow. If ventilation becomes preferentially redistributed to regions of higher Dl, i.e., from upper to lower lobes, Dl CO and Dl NO would both increase. This effect might disappear in adult animals if the distribution of Dl became more uniform with maturation. Owing to the exaggerated 1/DG imposed by SF6, both of these mechanisms would increase Dl COmore than Dl NO so that the ratio of Dl NO to Dl CO would fall as was observed (Table 5). These explanations remain speculative but raise testable hypotheses that can be further investigated.
Relating Gas Phase Resistance to Airway Structure
In normal resting dogs, there is no significant acinar 1/DG at 5 or 11 mo of age. However, there is a significant increase in acinar 1/DG in dogs 3 and 9 mo after PNX as puppies, consistent with post-PNX changes in airway structure we previously assessed by in vivo computer tomomography imaging and by postmortem morphometry (4, 13). High-resolution computer tomomography scan performed on the present animals showed elongated conducting airways 4 mo after PNX, whereas delayed airway dilatation occurred between 4 and 10 mo after PNX (4). These dimensional changes after PNX can reduce acinar penetration of an inspired breath, particularly when tidal volume is small, thereby increasing the diffusion path length from the convective front to the alveolar membrane independently of acinar geometry. When it is assumed that intra-acinar pathways also lengthened by 3 mo post-PNX, total peripheral airway cross-sectional area would be small relative to length and intra-acinar 1/DG distal to the convective front would be high as observed. By 9 mo post-PNX, compensatory growth of the remaining alveolar septal tissue (31) as well as the number of respiratory bronchioles (13) increase intra-acinar airway cross-sectional area and partially counterbalance the effect of airway lengthening. Such a compensatory mechanism may explain why the increased 1/DG is only evident at small tidal volumes in the mature animals after PNX.
Other Evidence of Gas Phase Resistance to Diffusion After PNX
Another method of detecting 1/DG is from the relative excretion rates of infused inert gases of differing molecular weights by MIGET. Specifically, a systematically higher retention of the inert gas with the highest molecular weight (enflurane) than expected from the enforced curve-fitting algorithm is indication of diffusion-limited alveolar gas mixing (37). In normal human subjects exercising at sea level and simulated high altitude, Torre-Bueno et al. (35) found no significant 1/DG by MIGET at rest or during exercise. However, in exercising horses, Hopkins et al. (11) found that infused enflurane was preferentially retained, which suggests a significant 1/DG. A similar finding was reported in mechanically ventilated American alligators (24).
We previously demonstrated a preferential retention of enflurane in adult dogs during exercise 1 yr after right PNX but not in control dogs (12). Combined with a density-dependent reduction of Dl NO, these findings provide strong evidence of increased 1/DG during post-PNX adaptation. The increased 1/DG, although detectable from NO uptake, was not of a sufficient magnitude to impair CO uptake. Similarly, the preferential enflurane retention in adult dogs after PNX did not contribute significantly to measured A-aDo 2 during exercise (12). Because post-PNX compensatory septal tissue growth is less vigorous in dogs pneumonectomized as adults than as puppies, the effect of acinar expansion and airway elongation on 1/DG should be greater in pneumonectomized adults. Thus a greater 1/DG after PNX, although detectable at rest and during exercise, does not cause a significant impairment in gas exchange.
In conclusion, our composite data demonstrate a negligible 1/DG in normal dog lungs regardless of maturity. After PNX, 1/DG is abnormally increased as evidenced by a density-dependent reduction of Dl NO, consistent with previously reported structural adaptation in the conducting as well as intra-acinar airways. The unexpected finding of a density-dependent increase of Dl CO in immature normal and PNX dogs cannot be fully explained at present but may be due to opposing effects of gas density on the convective penetration of an inspirate and on the diffusion coefficient of CO, or due to a density-dependent redistribution of ventilation preferentially to regions of high Dl CO during rebreathing. The density dependence of Dl CO disappears as the animals reach full maturity, suggesting that uniformity of Dl CO improves with postnatal lung growth and development.
The authors thank Heather Stanley and Richard T. Hogg for technical assistance and expertise in animal care.
This study was supported by National Heart Lung and Blood Institute Grants R01 HL-40070, HL-54060, HL-45716, and HL-62873.
Address for reprint requests and other correspondence: C. C. W. Hsia, Dept. of Internal Medicine, Univ. of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, TX 75390-9034.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published January 31, 2003;10.1152/japplphysiol.00525.2002
- Copyright © 2003 the American Physiological Society