INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MANY DISEASES AND ABNORMALITIES of the human lung begin with early permanent changes in the smallest air spaces (13). Numerous animal studies on acute effects of inflammatory substances, such as ozone, also corroborate the role of small airways in the early progression of lung disease (11). In addition, the growth of the airways relative to lung volume (VL) remains under investigation, with a few conflicting conclusions drawn from many studies (12). However, the in vivo determination of the caliber of the small airways in the human lung is extremely difficult. Aerosol-derived airway morphometry (ADAM) (2, 6) has been shown to be useful for providing a noninvasive means to accurately map the size of the lung air spaces.

Briefly, ADAM uses the gravitational settling of small, inhaled particles to infer the vertical distance [effective airway dimension (EAD)] that the particles must have settled to become lost to the airway wall. This method measures the dimensions of the air spaces as a function of the volumetric lung depth (VLD), i.e., the dimensions decrease as volume into the lung increases from the proximal to the distal airways. The location of the aerosol during the breath holds may not have been completely within a singular generation at a specific depth, but the aerosol is likely centered about that generation. Hence, the EADs are a volume average of the mean linear intercept of the air spaces at a depth populated mostly by air spaces of its respective generation. The technique has been limited, however, in its ability to associate air space dimensions with specific anatomic features.

In previous studies, we and others have shown that the dimension of the most distal airway generation, the alveolar region, can be reliably approximated by minimum EAD (EAD_min), by using ADAM under the appropriate experimental conditions, to detect even small changes between subjects due to age (3, 19). We hypothesized that, by using ADAM, we might also be able to identify in vivo the transitional bronchioles, a region that partitions the lung into two general anatomic areas marked by a difference in branching pattern.

One of the earliest lung morphometric studies that suggested the branching changes of the transitional bronchioles in the lung was the result of the construction of Weibel's lung "model A" that was based on data from lung cast measurements of the large airways and microscopic analysis of the parenchyma (18). The slope of airway size vs. generation was reported to be different between the proximal measurements, on the basis of dichotomous branching assumptions, and the distal acinar region, on the basis of microscopic linear cross-sectional observations of dissected lung tissue. Since then, others have refined the size and general location of this region in dissected lung tissue (e.g., Refs. 8, 15, 18). By analysis of the slope of EADs as a function of volumetric depth into the lung, we propose that the ADAM technique should be capable of detection of this transition. Furthermore, ADAM should detect a VLD associated with the transitional bronchioles that is similar to the anatomic dead space (ADS) as measured by the single-breath nitrogen washout procedure of Fowler (5). This procedure measures the volume of the conducting airways as the midpoint between purely conducting air space and alveolar gases.

In the present study, we used ADAM to characterize a structural region of the lung that is defined as the transition between two major areas of the lung with differing branching patterns. The in vivo information from ADAM gives the volume of the airways from the mouth to the transitional bronchioles (VLD_trans) and the diameter of the transitional airways (EAD_trans). To validate independently that VLD_trans is associated with the transitional bronchioles, we compared it with measures of ADS by single-breath nitrogen washout. These measurements were made in healthy nonsmoking adults at a VL of 70% of total lung capacity (TLC). We then investigated the influence of age, gender, and lung disease on VLD_trans and EAD_trans in subjects measured at 100% of TLC. By comparing the subjects recently studied at 70% of TLC with those studied at 100% of TLC, we were also able to estimate the effect of VL on VLD_trans and EAD_trans.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. The subjects were recruited and assigned to the five groups: 1) 20 normal adults (13 men/7 women; age 18-43 yr) for ADS and ADAM experiments performed at 70% of TLC; 2) 22 normal adults (8 men/14 women; age 19-40 yr) for ADAM experiments performed at 100% of TLC; 3) 26 normal children (10 boys/16 girls; age 8-18 yr) for ADAM experiments performed at 100% of TLC; 4) 13 patients (8 men/5 women) with moderately severe chronic obstructive pulmonary disease (COPD) for ADAM experiments performed at 100% of TLC; and 5) a group of 13 older normal adults (6 men/7 women) age-matched to the COPD patients for ADAM experiments performed at 100% of TLC. None of the subjects was a member of more than one group. Group 1 had their ADS and ADAM measurements made at an end-inspiratory VL of 70% of TLC. The other four groups were measured by ADAM at an end-inspiratory VL of 100% of TLC. Intergroup comparisons were made between two groups of normal adult subjects measured at either 70 or 100% of TLC; between normal children, adults, and older adults all performing at 100% of TLC; and between COPD patients and an age-matched group of the normal older adults, both performing at 100% of TLC. The normal subjects had no smoking history, no history of lung disease, and no recent history of acute respiratory infection or viral illness within the previous 4 wk. In each subject, we measured forced expiratory volumes (FEV)/flows and VL, inspiratory capacity (IC), and expiratory reserve volume by spirometry. Airway resistance (Raw), specific Raw (sRaw), and functional residual capacity (FRC) were measured by body plethysmography. TLC was defined as the sum of FRC and IC. Predicted FEV in 1 s (FEV₁) values were taken from Knudson et al. (10). Informed consent was obtained from each volunteer, and the study had the approval of the University of North Carolina Committee on the Protection of the Rights of Human Subjects.

ADS. Single-breath nitrogen washout was measured at 70% TLC, according to the technique of Fowler (5). Subjects inhaled a single breath of 100% O₂ (target flow rate of 800 ml/s) from residual volume (RV) to their predetermined end-inhalation volume and exhaled again (target flow rate of 400 ml/s) to RV without breath holds. A special mouthpiece, designed to reduce the volume of the oral cavity, was used. The mouthpiece was a metal tube, extending to the back of the mouth and surrounded by individually fitted silicone dental compound. This same mouthpiece was used for the EAD inhalations described below. ADS was determined as the midpoint of phase II of the nitrogen-washout exhalation curve by the equal areas method (7). The results of at least two measures were averaged for each subject.

EADs. The technique and theory used in these experiments is described below and in detail elsewhere (1, 14). In brief, the EADs were determined by analysis of exhaled aerosol recovery after inhalation of aerosol from FRC and breath holds for 0-10 s. If the lung is assumed to be composed of a system of randomly oriented tubes, the rate of decline (slope) of the particle recovery vs. breath-hold time is inversely proportional to the effective inner diameter of those tubes. A 1-µm monodisperse aerosol (mass median aerodynamic diameter, geometric SD = 1.1), composed of either diethylhexyl sebacate or carnauba wax, was condensed on salt nuclei generated by a condensation aerosol generator (MAGE) and delivered to a mouthpiece via a three-way valve and a check valve. The mouthpiece was fitted with instrumentation to record both flow, by using a pneumotachometer, and aerosol concentration, by using light scattering from a He/Ne laser to a photomultiplier tube. The subject was seated upright at the mouthpiece, and, during tidal breathing, the mouthpiece received filtered air. At the end inspiration of a suitable tidal breath, the three-way valve was switched to supply aerosol on demand, such that, on the next inhalation, the subject received the aerosol through the mouthpiece. The subject was instructed to inhale either to TLC or to a volume that equaled 70% of TLC. The rate of inhalation was controlled by the subject's following a visual signal proportional to a target flow rate of 1 l/s. At end inhalation of the aerosol, the subject was instructed to hold the breath for an interval of 0-10 s, followed by a controlled exhalation to RV. The maneuver was repeated 5-10 times for several different breath-hold times.

Data analysis. The VLD of the branching transition (VLD_trans), given as a volume into the lung from the mouth, was determined from the ADAM data after a logarithmic transformation of the EADs and their respective VLDs. VLD_trans is defined as the point of discontinuity marked by an abrupt change in the slope of EAD vs. VLD (see Fig. 1). To determine the point of discontinuity, forward and backward point-by-point linear regressions were done on EAD vs. VLD data over the approximate range of VLD = 100-1,000 ml to obtain two series of linear correlation coefficients (r²). The two series were summed, and the point of the local maximum near the transition point, excluding the two end points, was denoted as VLD_trans. At less than VLD = ~100 ml, EAD measurements are quite variable, and the basic assumption of the ADAM mathematical model (that airways are randomly oriented) is less valid, while at volumes greater than ~1,000 ml, EADs are nearly constant with increasing lung depth and thus reflect constant alveolar dimensions. The ADAM measurement from one individual is given in Fig. 1 to illustrate the typical EADs at their respective lung depth. The lung depths on the horizontal axis are normalized to the end-inspiratory VL at the time of the experiment (VL), which, in this case, was 70% of TLC. For this subject, VLD_trans occurred at a lung depth of 450 ml. The associated EAD at this depth was 0.36 mm. At the maximum r² sum of 1.965, the forward regression (by sequentially including points beginning proximally at 140 ml) resulted in an r² = 0.983, and the backward regression (by sequentially including points beginning distally from 1,000 ml) resulted in an r² = 0.982. A similar routine was repeated for each subject. A relative parameter, rVLD_trans, was calculated from VLD_trans as a percentage of VL, i.e., normalized to either 70 or 100% of TLC. For the case illustrated in Fig. 1, the 450 ml of the VLD_trans correspond to 7.8% of an 8.2-liter TLC lung at a VL = 70% of TLC.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Effective airway dimensions (EADs) decrease linearly as a log function of depth into the lung relative to lung volume (VL) at time of breath holds for 2 regions of lung. Solid lines, least squares best fit through the 2 regions to illustrate the determination of volumetric lung depth (VLD) intersection (VLDtrans) and intersection of EADtrans. Data shown are from 1 subject. In this case, VL at time of measurement was 70% of total lung capacity (TLC).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Anthropometric data. The subjects' anthropometric data and the results of the pulmonary measurements for each group are summarized in Table 1. There were no significant differences with respect to age, FEV₁ predicted, sRaw, height, and weight between the normal adult subjects who performed at 70% of TLC and the normal adult subjects who performed at 100% of TLC. However, there was a slight difference in TLC: the TLC of the group at 70% of TLC was larger, on average (P = 0.02), than that of the group at 100% of TLC. This was attributable to differences in the ratio of men to women in each group (13 men/7 women and 8 men/14 women, respectively), because there was no difference in TLC between the men or women of the two groups. The COPD patients had a lower percent predicted FEV₁ and a higher sRaw compared with the age-matched normal subjects.

Average VLD_trans and EAD_trans. Table 2 lists the mean values for the five subject groupings for VLD_trans, rVLD_trans, EAD_trans, and the estimated airway generation at which the VLD_trans occurs. When VLD_trans is normalized to TLC, rVLD_trans was found to be remarkably constant for healthy normal subjects of differing age and VL, remaining at ~7.5% of an individual's TLC. No significant difference in rVLD_trans was found between normal subjects at lung inflations of 70 and 100% of TLC and between children and the two age groups of adults.

Developmental and age-related differences. There was a significant increase in VLD_trans between children and adults (from 306 to 392 ml; P < 0.001). It tended to decline again for the older adults, although not significantly, to 352 ml. rVLD_trans, however, did not significantly change with age from children through adults to older adults (7.3, 7.1, and 6.5%, respectively), although there was a tendency to decline with age. There was a significant correlation (P = 0.006) between VLD_trans and age within the group of children (Fig. 2), changing ~28 ml/yr from 8 to 18 yr, and thereafter remaining constant within the two groups of older adults. However, multiple regression analysis for VLD_trans as a function of TLC and age showed that only TLC (P < 0.001), and not age, is predictive of changes in VLD_trans. Indeed, for all three age groups taken together (children, adults, and older adults), VLD_trans increased with increasing TLC but not with age (Fig. 3). TLC increased only within the group of children at a rate of 320 ml/yr from 8 to 18 yr. The rate of increase of VLD_trans within the children followed that of TLC, i.e., 28/320 = 8.8%, which is only slightly greater than the value for rVLD_trans for children (7.3%). There was no correlation between rVLD_trans and TLC. There was a detectable gender difference within the group of children in the development of VLD_trans relative to TLC. The boys' rate of increase of VLD_trans was ~55 ml/yr (P = 0.02); this resulted in a 10% change in VLD_trans relative to a 532 ml/yr increase for TLC. The girls' increase in VLD_trans was only loosely correlated with age and increased by ~18 ml/yr (P = 0.087), while TLC increased by 236 ml/yr. This results in a change in VLD_trans of 7.6% relative to changes in TLC. Despite increasing VLD_trans with age, however, EAD_trans did not change with age for either the boys or the girls.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   VLDtrans increases with age for children; r = 0.53, n = 26.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   VLDtrans increases as a function of TLC for the 3 normal groups measured at 100% of TLC. For the regression of VLDtrans on TLC for the 3 normal groups: r = 0.72, n = 26 for children (); r = 0.72, n = 22 for adults (); and r = 0.88, n = 13 for older adults (). For all subjects combined, r = 0.72, n = 61.

VLD_trans and EAD_trans at 70 vs. 100% of TLC. Two groups of adults were compared when ADAM measurements were made at 70 and 100% of TLC (see Table 2). The average VLD_trans increased slightly but not significantly on increasing VL from 70 to 100% of TLC, and rVLD_trans was not significantly different for the two groups (7.9 ± 1.1 and 7.1 ± 1.8 % of TLC, respectively). EAD_trans, however, was smaller at 70% compared with 100% of TLC (0.418 ± 0.066 and 0.595 ± 0.175 mm, respectively), although the former group had a larger average TLC. When gender was taken into account (Table 3), only male VLD_trans increased significantly (from 385 ± 75 to 492 ± 116 ml; P = 0.017), while female VLD_trans only tended to increase (from 270 ± 22 to 334 ± 128 ml; P = 0.21). When the EADs at all depths of the lung (Fig. 4) are compared, the effect of lung inflation from 70 to 100% of TLC is to increase the caliber of the smaller airways. There appears to be little effect on the larger airways or the alveoli. Within these normal adult groups, EAD_trans did not correlate with TLC. At 70% of TLC, VLD_trans correlated significantly with TLC and height. ADS also correlated significantly with TLC and height. At 100% of TLC, VLD_trans correlated significantly only with TLC but not height.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   EADs were averaged for the 2 groups of normal adults who were measured at the 2 lung inflations: VL = 70 or 100% of TLC. Depth into the lung is relative to VL. Top line, upper 95% confidence interval for data at 100% of TLC; bottom line, lower 95% confidence interval for data at 70% of TLC.

COPD patients vs. age-matched normal subjects. rVLD_trans for the two groups were different: 6.5 ± 0.9% TLC for the normal subjects and 3.6 ± 1.1 %TLC for the patients, while absolute volumes differed also (352 ± 87 and 216 ± 64 ml, respectively). Both groups were measured at 100% of TLC. Within the COPD group, VLD_trans did not correlate with TLC, Raw, sRaw, or percent predicted FEV₁. VLD_trans, however, changed with rVLD_trans (r = 0.675, P = 0.011), unlike for the age-matched normal subjects, for which it did not change. EAD_trans was much larger in the COPD group (0.834 ± 0.398 mm) compared with the normal group (0.624 ± 0.232 mm).

VLD_trans is related to ADS and end of phase II. A significant correlation was observed between VLD_trans and ADS (P = 0.001) for the group of normal adults measured at a VL of 70% of TLC (Fig. 5). The slope of VLD_trans vs. ADS was linear, with the VLD_trans increase nearly twice that of ADS. The average VLD_trans was greater than ADS by 121 ml (345 ± 83 vs. 224 ± 34 ml). VLD_trans was closer in value to the end of phase II (360 ± 53 ml), where purely alveolated air begins in the nitrogen washout. The relationship of VLD_trans to end of phase II was linear and close to equality: VLD_trans = 0.93(end of phase II) + 9 ml (r = 0.599, P = 0.005). VLD_trans also correlated significantly with an individual's TLC (P < 0.001), as did ADS (P = 0.004) and end of phase II (P = 0.010).

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Relationship between VLDtrans and anatomic dead space for group of normal adults measured at 70% lung inflation. (VLDtrans = 1.71 ADS  39 ml; r = 0.70, P = 0.001, n = 20 subjects).

Gender relationships of VLD_trans, rVLD_trans, EAD_trans, and generation. Table 3 lists the values for VLD, rVLD_trans, EAD_trans, and estimated airway generation, as analyzed by gender. For the normal adult groups, of the four parameters, only VLD_trans is significantly different between men and women. Men's VLD_trans was significantly greater than that of the women, consistent with a larger TLC in men, and remained at ~7.3% of TLC. For adults, both men and women, VLD_trans increased similarly (by 128 and 124%, respectively), while their VL increased by 143% (from 70 to 100% of TLC). With respect to rVLD_trans, there was no difference between males and females for any of the groups when VL increased from 70 to 100% of TLC, with the exception of rVLD_trans for adult women, which decreased from 8.0 to 6.8% of TLC (P = 0.030). EAD_trans values for men and women were not different within any of the study groups. For both men and women, EAD_trans increased significantly (from 0.413 to 0.526 and from 0.426 to 0.634 mm, respectively), while increasing VL from 70 to 100% of TLC. Boys and girls were not analyzed, because they were not matched for age and developmental stage. There was no difference in any of the parameters between men and women who were COPD patients.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have analyzed lung morphometry by ADAM in various subject types for detection of a transition in airway branching patterns associated with the transitional bronchioles. This transition was originally reported by Weibel in the model A (18) on the basis of lung casts and microscopic dissection. The in vivo EAD measurements by ADAM provide EAD as a function of VLD. When these values were plotted on logarithmic scales, two distinct regions of the lung were evident according to differing slopes of EAD vs. VLD. The point at which the proximal slope differs from the distal slope, we designated as VLD_trans, with its associated airway size EAD_trans. We made comparative analyses of both VLD_trans and EAD_trans as a function of age, gender, and lung disease. These values were also compared with single-breath nitrogen washout measurements, another physiological index of transition from conducting airways to alveolated air spaces.

The different slopes of the EAD vs. VLD relationship (Fig. 1) were surmised to be due to a transition in the pattern of branching that occurs at the transition of bronchioles to ducts where predominantly dichotomous branching of the airways transitions into a budding of the alveolar ducts. Because ADS by the Fowler technique (5) is defined as the midpoint of the transition from airway dead space gas and purely alveolar gas, VLD_trans should be similar to, but slightly deeper into the lung, than that detected by ADS as discussed below. At this region of the bronchial tree, two things happen simultaneously with respect to gas transport and aerosol deposition. First, for gas transport, there is a rapid increase in diffusion of the gas into the alveolar volume, which is complete at the end of phase II. Second, at this point, there is a dramatic change in the type of branching from the nearly regular dichotomous type of the airways to the irregular budding type of the transitional bronchioles; this produces a change in the rate of decrease of airway caliber. We found that there was a good correlation of VLD_trans with ADS (P = 0.001) and end of phase II (P = 0.005), and, at a VL that is 70% of TLC, average VLD_trans was 345 ml, whereas the measured dead space for our subjects was ADS = 224 ml and 360 ml at the end of phase II. The variability of VLD_trans for a given ADS is consistent with variations calculated from our random errors in volume measurement. The volume of VLD_trans greater than ADS is consistent with the volume contained within the airways from generation 17 through the transitional bronchioles' generation 19 of Weibel's model. On the basis of this model, our VLD_trans occurs at generation 19. VLD_trans, therefore, should be equivalent to the volume measured at the end of phase II of the nitrogen washout. Both VLD_trans and ADS correlated with an individual's TLC (P < 0.001 and 0.004, respectively), but multiple regression with the use of both parameters indicated that only TLC is important in predicting VLD_trans. EAD_trans did not change with TLC, nor did EAD_min [a parameter of alveolar air space dimension (19)] within any of the normal groups; this indicates that an individual's lung capacity is determined primarily by the number of bronchiole-alveolar ducts rather than the size of the branches and alveoli.

Age-related changes in VLD_trans were investigated for children (from 8 yr old) to older adults (to 80 yr old). VLD_trans increased in volume for the children (from 8 to ~18 yr) and then remained constant during the adult ages. Multiple regression analysis indicated that the developmental increase in the children's VLD_trans was concomitant primarily with an increase in lung size with age, a result similar to that reported by Hart et al. (9) for ADS vs. FRC. In fact, VLD_trans correlated best with TLC for all of the age groups of subjects, although there were no significant age-related changes in TLC for adults. Also, there was no change with age in VLD_trans when it was normalized to an individual's TLC; VLD_trans remained at ~7.2% of TLC throughout the ages investigated. This indicates that it remains a nearly constant proportion of VL during development after 8 yr of age and with large differences in TLC (2.4-8.0 liters). There was a small drop, however, in the normalized VLD_trans for the older adults (to 6.5% of TLC). This may reflect the beginnings of airway disease with advanced age, as indicated by a slightly reduced FEV₁ (P = 0.07 compared with young adults). In regard to the size of the airways at this transition, we found little or no significant age-related change in EAD_trans, both within the age groups and for the subjects taken as a whole, although there was a trend for EAD_trans to increase with age (from 0.530 mm in children, through 0.595 mm in adults, to 0.624 mm in the older adults). It is possible that individual variability in EAD_trans (and measurement uncertainty) masks the small dimensional changes in these airways during development. Also, there may be some significant changes in airway branching during development in children who are younger than 8 yr of age. It may be of some practical use to perform longitudinal studies of EAD_trans similar to the studies of VL growth curves.

On the other hand, small age-related developmental differences in VLD_trans by gender were detected for the children. The change in VLD_trans with age relative to the change in TLC was higher in boys than in girls (10 vs. 7.6%, respectively). This indicates that the growth of the boys' airways relative to TLC is greater than that of the girls in this age range. A similar conclusion was drawn by Merkus et al. (12) from their longitudinal study of flow-volume data.

EAD_trans increased from 0.413 to 0.595 mm when increasing VL from 70 to 100% of TLC, unlike the nearly constant value of EADs measured at ADS in beagle dogs over a wider range of lung inflations (16). The authors of that study concluded that the gaseous diffusion front of the ADS measurement was shifted slightly peripherally at higher lung inflations. The peripheral shift of ADS thus produced EAD measurements made at progressively smaller airways, apparently just enough to produce constant EAD values, although the airways were distended after higher inflations. Our EAD_trans value, on the other hand, is based on an anatomic structure defined by the branching pattern of the airways and, as such, constitutes a fixed generation that does not shift position relative to the EADs.

Although VLD_trans did not significantly change with lung inflation from 70 to 100% of TLC, it tended to increase slightly in volume (from 345 to 392 ml), which is consistent with an increase of 55 ml observed in our measurements of ADS on increasing VL in the same individuals from 70 to 100% of TLC and is comparable with changes in ADS found by others (17). This change in VLD_trans represented a slight decrease in the relative proportion of VL (from 7.9 to 7.1% of TLC). This would indicate that the dimensions of the conducting airways change without increasing the airway volume significantly. This apparent contradiction can be explained by a closer inspection of Fig. 4, which compares the EADs for the entire lung averaged for the two VL groups. The largest airways do not significantly increase in size, whereas the largest relative change occurs near the transitional bronchioles and at the air spaces slightly deeper into the lung, where, by shear number, relative volume change is very large. A similar finding was concluded by Shepard et al. (17) by using CO₂ exhalations for determining dead space volumes at varying VL values. This would explain why VLD_trans does not increase very much with increasing VL, because little of the VL change occurs in the larger airways. Greater VL change occurs in the smaller airways that are slightly deeper into the lung than at VLD_trans, where the distensibility of the airways is nearly equal to that of the alveoli.

Interestingly, the alveolar dimensions at 25% lung depth (EAD_min) did not change between 70 and 100% inflation. This would be consistent with alveolar recruitment from newly ventilated acini that were not fully ventilated at the lower VL, rather than enlargement of already ventilated regions, and with traction on the transitional bronchioles by the recruited alveoli that results in an increase in duct caliber but little increase in average alveolar dimension. This phenomenon requires more investigation, with repeated data in one group of subjects, each studied at several VL values from RV to TLC.

We also compared patients with COPD with an age-matched group of normal subjects. The disease produced a marked decrease in VLD_trans, from 352 to 216 ml, consistent with volume reduction due to airway narrowing. As a proportion of total lung size, rVLD_trans also decreased with the disease (from 6.5 to 3.6% of TLC), while the size of the transition, EAD_trans, increased from 0.624 to 0.834 mm. These changes were similar to but greater than those found with aging in adults, as discussed above. The decrease of rVLD_trans that we observed in our patients was not simply caused by an increase in TLC typical of COPD. TLC increased 7.5% in men and 9% in women with COPD compared with age-matched normal subjects. rVLD_trans, however, changed much more, decreasing by 49% in men and 40% in women with COPD. In addition, there was no correlation of rVLD_trans with TLC within any of the groups. We found no correlation of these changes with two typical parameters of severity of the disease, sRaw and FEV₁, although, as a group, these parameters were significantly different from normal values (Table 1). We also found a simultaneous decrease in VLD_trans and an increase in transitional bronchiole caliber (EAD_trans) observed in these individuals. This may be consistent with loss of airway tissue and disruption of the branching pattern at and just proximal to the transitional bronchioles, such that the site of transition from regular branching to budding is shifted mouthward. It is also possible that severe ventilation inhomogeneity on exhalation of the aerosol produces an overlapping morphometry of relatively disease-free sections deeper in the lung, with the more proximal airways of sections containing significant damage. For this case, these large changes seen in VLD_trans and EAD_trans, along with EAD_min, may be used to quantify the aggregate volume of lung tissue damage summed from the various discrete locations in the lung. Also, air spaces in damaged acinar regions may not fill with aerosol sufficiently to obtain reliable EADs from those distal regions. More studies need to be undertaken to quantify the relationship of VLD_trans and EAD_trans with the type and severity of lung diseases.

We found no unexpected differences between men and women with respect to VLD_trans and EAD_trans. The only difference was found in VLD_trans, which was greater in men than women, consistent with a larger lung capacity in men. As a proportion of TLC, the difference in VLD_trans was insignificant. An increase in VLD_trans caused by an increase in VL from 70 to 100% of TLC was similar in both men and women, although the VLD_trans in women decreased somewhat as a proportion of VL.

VLD_trans correlates closely with both ADS, as measured by nitrogen washout, and with lung size for normal lungs. In fact, VLD_trans correlated to a better degree with TLC than did our measurements of dead space. This indicates that VLD_trans may be a more precise measurement of the airway volume than is ADS. Also, it is not surprising that VLD_trans is greater than ADS (345 ± 83 vs. 224 ± 34 ml, respectively), because the two parameters are measured by different methods. ADS measurements rely on the transport of a gas into a large volume by convection and diffusion, whereas ADAM is a function of convective transport of particles to a location in the lung. The diffusion of gases at the airway-alveolar interface during transport is likely to diminish the size of the ADS measurement. Factors that increase the time for gaseous diffusion to occur, such as reduced flow rate and breath holds, have been shown to decrease ADS measurements (4, 17). Particle position, on the other hand, is relatively unaffected by diffusion. The two volumes, ADS and VLD_trans, are correlated, because the morphology producing the two measurements (increased alveolarization and branching changes) are nearly anatomically coincident. We found that VLD_trans was not significantly different in value compared with the end of phase II of the nitrogen washout (345 ± 83 vs. 360 ± 53 ml), which is the point on the washout curve that begins to be purely alveolated gas on exhalation. In conjunction with EAD_min, as a measure of alveolar dimensions (19), a determination of VLD_trans and EAD_trans permits ADAM to be more useful for identification of changes in lung airway volume and diameter during development and disease.

We have shown that in vivo ADAM measurements identify and characterize a distinct site in the lung, described by VLD_trans and EAD_trans, that is likely associated with the transitional bronchioles of the human lung. The ability to characterize this region in vivo is important, because it is believed to be the initial site of lung damage that is associated with smoking products and air pollutant exposure.