INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

AEROSOL BOLUS DISPERSION (ABD) has been proposed as a sensitive measure of small airway function (3, 5, 13, 15, 20). This technique consists of inserting a volumetrically small bolus of aerosol into a subject's tidal volume. As the aerosol travels through the airways of the lung, it becomes mixed with particle-free air, such that, on expiration, aerosol is dispersed over a larger volume of air than was the inhaled bolus. Relative to healthy nonsmokers, ABD has been found to be increased in asymptomatic smokers (3, 5, 15), in patients with cystic fibrosis (CF; Ref. 1), and in healthy young adults after acute ozone exposure (13). All studies have found ABD to increase with the volumetric penetration (V_p) of boluses, i.e., the volume of clean air inhaled following the center of a bolus (1, 3, 5, 7, 13, 15, 20). In the healthy lung, ABD is thought to be largely due to convective mixing and nonreversal of axial streaming (7, 20, 26). It has been speculated, however, that nonuniform distribution of ventilation due to regional differences in pulmonary resistance and compliance becomes an increasingly important factor affecting ABD in compromised lungs (1, 3, 4, 13, 19). Yet no studies have shown a direct relationship between dispersion and ventilation distribution indexes in normal subjects or diseased patients.

Regional time constants, regional lung volumes, and V_p may all play a role in determining the linear distance inhaled aerosol travels and the distribution of aerosol among airways at that distance. In a modeling study, Rosenthal (19) predicted that boluses penetrating a modest depth (V_p of 400 ml) into the lungs should be more sensitive to nonuniform ventilation between parallel compartments than should deeper boluses, which may preferentially go to better ventilated compartments, or shallow boluses not penetrating beyond the dead space. At such an intermediate depth, an inhaled bolus might divide between fast- and slow-ventilated regions so that the sequential emptying of regions may cause particles, adjacent on inspiration, to be separated by volume on expiration, thereby increasing ABD. For two compartments of equal volume in parallel, Rosenthal (19) predicted ABD to increase 1) with increasing disparity in regional time constants and 2) with the magnitude of time constants when the ratio of time constants is maintained.

Dynamic convective inhomogeneities also influence multiple-breath gas washouts. The washout of either a single compartment or parallel compartments with equal time constants should be a monoexponential function of time when constant flow rates are maintained (10, 18). On the other hand, disparities in time constants between parallel compartments may cause a multiexponential washout, the shape of which depends on the relative size of compartments, the concentration inhomogeneities between compartments before washout, and the magnitude of the difference in the time constants (18). Multiple-breath ¹³³Xe washout (MBW_Xe) has been utilized as a means of assessing regional ventilation distribution and whole lung washout (17, 21, 23-25). Patients with severe obstructive lung disease not only have slow total washouts but also have markedly nonuniform washout across the lungs (21).

We hypothesize that the increased ABD observed in CF patients, relative to normal subjects (1), is partially attributable to altered ventilation distribution in those patients. To investigate this hypothesis, we have compared aerosol bolus measurements with measures of MBW_Xe in healthy subjects and CF patients.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Subjects. Fourteen patients with CF (4 men, 10 women; 20-45 yr of age) and nine healthy nonsmoking male volunteers (18-40 yr of age) were recruited to participate in this study. To be recruited for the study, healthy subjects were required to have a ratio of forced expired volume in 1 s (FEV₁) to forced vital capacity (FVC) (FEV₁/FVC)  70% and FEV₁  90% of predicted, based on the equations of Knudson et al. (14). Healthy female subjects were not recruited to minimize the number of women of childbearing age exposed to radiation in our study. Patients with CF had mild-to-moderate airway disease with FEV₁ (%predicted) ranging from 49 to 77%. All subjects were free of acute respiratory infection for at least 2 wk before participation in the study. This study was approved by the Committee on the Protection of the Rights of Human Subjects, School of Medicine, University of North Carolina at Chapel Hill. Before participating in the study, all subjects were informed of study-associated risks and asked to sign a statement of informed consent.

Aerosol bolus measurements. Detailed methodology for the bolus technique has been presented previously (7). Aerosol boluses were inserted into a 1.25-liter tidal volume, inhaled from functional residual capacity, at the V_p values of 250 ml (V_{p 250}) and 500 ml (V_{p 500}). V_p is the volume of air inhaled following the center, i.e., the volumetric mean, of a bolus. After a bolus inspiration, subjects exhaled to residual volume. Inspiratory and expiratory flow rates were kept constant at 0.4 l/s. Inhaled boluses were targeted to have a volumetric width of 150 ml. Experimentally, 95% of aerosol was confined in 180 ml of the inhaled breath. Boluses were composed of a 0.5-µm polydisperse (geometric SD = 1.5) aerosol of triphenyl phosphate (7, 13). All measurements were conducted with the subject in a seated position. Three to four maneuvers were conducted at each V_p. Flow and aerosol concentration from each maneuver were acquired at a 200-Hz sampling rate and were saved for subsequent data analysis.

¹³³Xe washout test. MBW_Xe was conducted with subjects seated in front of a gamma camera (Elscint Apex 415; large field of view). During the washout, subjects breathed a 1.2-liter tidal volume with tidal flows of 0.4 l/s to mimic the breathing pattern used for the bolus maneuver. Subjects rebreathed ¹³³Xe in air until equilibrium was reached, i.e., until a constant count rate was detected by the camera. At equilibrium, acquisition was initiated with a total of 32 images collected: 4 at equilibrium and 28 during a washout period where ¹³³Xe-free air was inhaled and exhaled ¹³³Xe was captured in a charcoal filter. Each image was acquired for 6 s, which was the subject's breathing period.

t

_ft

_st

1

1

Pulmonary function tests. Forced-expiratory maneuvers were conducted with the subject in a standing position. Three acceptable trials were obtained from each subject. The largest single FEV₁ and FVC values, regardless of trial, were used in the calculation of FEV₁/FVC and for percent predicted FEV₁. Percent predicted FEV₁ (14) was determined for comparisons with bolus data.

Data analysis (comparisons of tests). The primary focus of this study was to compare ABD with MBW_Xe indexes of ventilation distribution. Pearson correlation coefficients were utilized for comparisons between ABD for each V_p and A/B, SD_50%, , GoF, and F_s. As a secondary analysis, we also compared R and with these parameters. Because each bolus parameter (ABD, R, and, ) was measured at two depths into the lung, there were 10 comparisons between each bolus parameter and other indexes. A Bonferroni correction was applied, and a significance level of 0.005 was adopted. For comparison with the findings of Anderson et al. (1, 3), correlations were also determined between bolus parameters of R and ABD and both spirometry (%predicted FEV₁) and N₂. Statistical significance for within- and between-group differences were tested by dependent (paired) and independent two-tailed Student's t-test, respectively.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Table 1 provides a summary of mean data for the normal and CF groups. Bolus measurements and spirometry were available on all subjects. Nitrogen washout was not conducted in one CF patient because of instrument malfunction. The MBW_Xe data of one CF patient were inadvertently deleted during analysis so that indexes of A/B and SD_50% were not available. Typical aerosol boluses for a normal subject and CF patient are illustrated in Fig. 1. The MBW_Xe were slower and less monoexponential in patients than in normal subjects (Table 1).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Typical aerosol boluses for normal subject and cystic fibrosis (CF) patient with volumetric penetration of 500 ml (Vp 500). Aerosol bolus dispersion was increased and recovery was decreased in patient, relative to normal subject, at both volumetric penetration of 250 ml (Vp 250) and Vp 500. Patient's bolus may also be seen to rise in concentration earlier than that of normal subject. This shift toward the mouth may be due to early emptying of fast compartments.

ABD and R were significantly different between the CF and normal groups at both V_p levels (Table 1). Within groups, measurements made at V_{p 250} were significantly different from those at V_{p 500} for both ABD and R (P  0.02). As seen in our previous study (7), ABD and were correlated in both groups at V_{p 250} (P < 0.005) and V_{p 500} (P < 0.05). No associations were observed between ABD and R at either V_p or in either CF patients or normal subjects.

In the CF patients, but not in the normal subjects, we found significant correlations between bolus parameters and ventilation indexes obtained by MBW_Xe (Table 2). The highly significant associations observed in the CF patients between ABD and , as well as F_s at V_{p 500}, are illustrated in Figs. 2 and 3, respectively. As shown in Table 2, weak associations (not significant after Bonferroni correction) were also observed between ventilation parameters ( and F_s) and both ABD at V_{p 250} and at V_{p 500}. The associations observed between R and GoF at V_{p 500} are illustrated in Fig. 4. Neither of the regional ventilation indexes, A/B and SD_50%, was correlated with bolus parameters. No significant correlations were observed between bolus parameters and MBW_Xe indexes in the normal subjects.

View larger version (8K): [in this window] [in a new window]   Fig. 2.   Total lung 133Xe washout rate () was correlated with aerosol bolus dispersion (ABD) Vp 500 (r = 0.76, P < 0.005) in CF patients.

View larger version (8K): [in this window] [in a new window]   Fig. 3.   ABD showed a strong correlation with fraction of 133Xe in a slow compartment (Fs) at Vp 500 (r = 0.89, P < 0.001) in CF patients. The strength of this relationship would be expected to diminish as difference in time constants between fast and slow compartments decreased.

View larger version (9K): [in this window] [in a new window]   Fig. 4.   Aerosol recovery and goodness of fit (GoF) were correlated at both Vp 250 (r = 0.73, P < 0.005) and Vp 500 (r = 0.72, P < 0.005). This relationship is shown at Vp 500 in CF patients.

Consistent with the results of Anderson et al. (1), significant correlations between spirometry and ABD were also observed in the CF patients: FEV₁ (%predicted) correlated with ABD at V_{p 500} (r = 0.63, P = 0.02) and to a lesser extent at V_{p 250} (r = 0.51, P = 0.06). As in smokers (3), there were no significant correlations between the single-breath nitrogen N₂ and bolus measurements observed in either normal subjects or CF patients.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

We investigated the influence of ventilation distribution on aerosol bolus measurements in health and disease. ABD, R, and were compared with MBW_Xe in normal subjects and CF patients. We made comparisons with five indexes of ventilation distribution from the MBW_Xe test: 1) , which is the total washout rate of the lungs and is decreased by poor ventilation of obstructed regions; 2) F_s, which is the fraction of the lung that was poorly ventilated; 3) GoF, which reflects the degree to which the lungs work as a single, presumably well-ventilated compartment and the uniformity of ventilation between compartments; 4) A/B, which indicates uniformity of ventilation between apical and basal regions of the lungs; and 5) SD_50%, which reflects the heterogeneity of washout rates across six gross regions of the lung. Comparisons between these indexes and ABD and R suggest that these bolus parameters are both affected by nonuniform ventilation distribution induced by airway obstruction.

Associations between bolus measurements and ventilation distribution were only observed in the CF patients. Spirometric data show that these patients had significant airway obstruction, relative to normal subjects, based on decreased FEV₁/FVC and the forced expiratory flow over the middle half of FVC (Table 1). Patients also exhibited marked nonuniform ventilation distribution, relative to normal subjects, as evidenced by slow washouts (increased ) and multiexponential ¹³³Xe washouts (decreased GoF and increased F_s). We have made no attempt to quantify the degree to which convection vs. diffusion-dependent inhomogeneities are affecting ventilation distribution. However, for particles in the size range of the experimental aerosol and larger, it is well accepted that aerosol mixing in the lung is predominantly convective (20, 26). Thus the associations between bolus parameters and the MBW_Xe indexes observed in the CF patients are likely attributable to the influence of convective inhomogeneities on these measurements.

ABD in the CF patients was found to be associated with measures of ventilation distribution in a depth-dependent fashion, i.e., ventilation distribution had a greater influence on ABD at the deeper V_p (V_{p 500}) into the lung. In the CF patients at V_{p 500}, ABD was correlated with F_s (r = 0.89, P < 0.001) and  (r = 0.76, P < 0.005). The  is decreased by the poor ventilation of obstructed regions in the CF patients, and the fraction of the lung that is poorly ventilated is reflected by F_s. The strong correlations between ABD and both  and F_s suggest that, with increasing severity of obstruction, there is an increasing probability of an inspired bolus being divided between fast and slow (i.e., obstructed) regions and that differing time constants between these regions will cause sequential emptying, thereby enhancing ABD. The late expiration of aerosol from the slow regions may also exhibit itself as a long tail on the exhaled bolus and account for the positive association observed between F_s and (r = 0.56, P < 0.05) in the patients at V_{p 500}. At V_{p 500}, on the basis of the correlations with F_s and , ABD is clearly influenced by inhomogeneities in the distribution of ventilation between compartments as well as the pattern or sequence of lung filling and emptying.

At the shallower V_p, ABD did not show the strong associations with ventilation distribution apparent at V_{p 500}. At V_{p 250}, ABD was only weakly associated with F_s (r = 0.60, P < 0.05) and  (r = 0.58, P < 0.05). These weaker associations at V_{p 250} suggest that deeper aerosol bolus inhalations may be important to detect nonuniform ventilation distribution brought about by airway obstruction. It may be that V_{p 250} was close enough to the end of inspiration so as to have the inhaled aerosol preferentially enter slow filling regions. Aerosol inhaled into slow regions may be exhaled out later than expected because convective inhomogeneities may alter the sequence of lung emptying from that of "last in, first out" to "last in, last out" (10, 16). The CF patients showed peripheral (50 ± 30 ml) that was significantly greater than the (22 ± 25 ml) observed in normal subjects at V_{p 250} (P < 0.05). Despite weak associations with our measures of ventilation distribution, ABD was still significantly greater in patients than in normal subjects at V_{p 250} (P < 0.05).

The R correlated well with GoF at both V_{p 250} (r = 0.73, P = 0.003) and V_{p 500} (r = 0.72, P = 0.003). As the lungs shift from a single, presumably well-ventilated system toward a multicompartment system with healthy and obstructed lung regions, R decreases (Fig. 4). Particles that entered poorly ventilated regions will be exhaled slower and later than particles that entered healthy regions, even though the two sets of particles were similarly localized within the inspired bolus. Increased residence time in slow regions and sites of airway obstruction, which may have induced the ventilatory inhomogeneities, could enhance deposition (12). Indeed, an increased regional deposition fraction for poorly ventilated areas has been reported for aerosols in the 0.5- to 1.0-µm size range and has been attributed to increased residence time (23, 25). It has been demonstrated in healthy children with total lung capacity (TLC) ranging from 2.1 to 5.3 liters that, for a given V_p into the lung, deposition of inhaled particles increases with decreasing TLC, whereas ABD is unaffected (22). The enhanced deposition by sedimentation and diffusion is attributed to smaller dimensions associated with smaller TLC. Similar to the effect of TLC on deposition, the preferential ventilation suspected at V_{p 250} may produce an increase in deposition by taking aerosol to smaller airway dimensions than would occur with uniform ventilation patterns.

There were no correlations found between ABD or R and regional measures of ventilation distribution, SD_50% and A/B, nor were these latter parameters correlated with other indexes of ventilation distribution. Both SD_50% and A/B were significantly increased in patients compared with normal subjects, indicating larger interregional disparities in ventilation within the patients. Specifically, A/B and SD_50% both increase with decrements in ventilation to the apical regions of the lungs relative to the basal regions. Gross interregional differences in ventilation, i.e., partial obstruction of an entire lung or multiple lobes, should, theoretically, increase ABD, and they have been shown clinically in one patient to do so (4, 19). For CF patients in our study, however, the damage to the lungs was not so distinct between lung lobes [e.g., a unilateral bronchial stenosis (4)] so as to lead to correlations between ABD and the indexes SD_50% and A/B.

In conclusion, the results of this study show that aerosol bolus measurements of ABD and R are related to convective inhomogeneity of ventilation in patients with CF. The influence of uneven ventilation on ABD in these patients was greater when boluses were delivered deeper (500 vs. 250 ml) into the lungs. On the other hand, given the lack of correlations between indexes of ventilation distribution and ABD in normal subjects, it appears that convective inhomogeneity of ventilation is not a major mechanism of dispersion in the normal lung.