Journal of Applied Physiology
Vol. 82, No. 2, pp. 632-643, February 1997
GAS EXCHANGE, MECHANICS, AND AIRWAYS

Mechanism of lobar alveolar pressure decline during forced deflation in canine regional emphysema

S. N. Mink

Sections of Respiratory Diseases and Critical Care Medicine, University of Manitoba, Winnipeg, Manitoba, Canada R3E-OZ3

ABSTRACT

INTRODUCTION

METHODS

RESULTS

DISCUSSION

APPENDIX

ACKNOWLEDGEMENTS

FOOTNOTES

REFERENCES

ABSTRACT

Mink, S. N. Mechanism of lobar alveolar pressure decline during forced deflation in canine regional emphysema. J. Appl. Physiol. 82(2): 632-643, 1997.A canine model of unilobar papain-induced emphysema was used to examine the extent to which differences in alveolar pressures (PA) would develop between an emphysematous right lower lobe (RLL) and normal left lower lobe (LLL) during forced vital capacity (FVC) deflation. RLL and LLL PA (PA_RLL and PA_LLL, respectively) were measured by the alveolar capsule technique. During forced deflation, PA and lobar flows were determined between 95 and 20% FVC. A choke point common to both lower lobes was observed at >40% FVC. The results showed that deflation compliance (C) was altered for the RLL such that <90% lobar vital capacity, C_RLL > C_LLL, whereas >90% lobar vital capacity, C_RLL < C_LLL. At 95 and 90% FVC, the initial RLL PA decline was greater than that for the LLL (P < 0.05). However, large differences in PA were prevented because of the effect of interdependence of regional expiratory flow (IREF). IREF caused a relative decrease in RLL flows and increase in LLL flows that limited PA differences. Between 80 and 50% FVC, as C_RLL became greater than C_LLL, and because of the initial effect of IREF, PA_RLL was ~PA_LLL. At 40% FVC, without IREF, lobar differences in PA widened. These findings indicate that IREF may affect the dynamics of flow limitation in regional lung disease.

maximum expiratory flow; nonuniform emptying; flow limitation

INTRODUCTION

IN PULMONARY EMPHYSEMA, it is often observed that alveolar destruction occurs nonhomogeneously throughout the lung and that some lung regions are severely diseased, whereas other regions are relatively spared (1, 18). This nonuniform process results in a lung in which mechanical properties are variable between regions. Lung mechanical properties, such as recoil pressure (Pel), frictional resistance, and bronchial pressure-airway behavior are the primary determinants of maximal expiratory flow (max) (8). Whereas alveolar destruction in emphysema characteristically results in a decrease in Pel and an increase in compliance over most of the vital capacity (VC) (1, 18), it was previously demonstrated in a papain model of emphysema that at very high lung volumes [VL; i.e., >90% total lung capacity (TLC)], compliance was actually lower than that found in healthy lungs (10, 11, 14). At these very high VL, it appears that nonelastic collagen is the predominant factor limiting further alveolar distension in this model, and therefore a relative lower compliance is found. Accordingly, the shape of the static pressure-volume curve is altered in papain-induced emphysema so that at very high VL compliance is lower, whereas over the remaining volumes, compliance is greater compared with values obtained from healthy lungs.

Such changes in compliance in regional emphysema would be expected to produce markedly nonuniform alveolar pressures (PA) during forced deflation. During the initiation of expiration, because compliance at very high VL is lower in emphysematous lung, the fall in PA (Pel + pleural pressure) of emphysematous regions may occur to a greater extent than that of normal lung units. On the other hand, with continuing deflation, compliance would relatively increase in emphysematous units so that, at low VL, PA would eventually rise to exceed that of normal lung units.

Whereas alterations in lung compliance in regional emphysema would provide the mechanism by which differences in PA would develop between regions, because frictional resistance does not appreciably increase in this model (10), it has been proposed that opposing this development of nonhomogeneous deflation would be the effect of regional interdependence of expiratory flow (IREF). Wilson et al. (20) proposed that if regions shared a common site of flow limitation (i.e., choke point), then there would be interdependence of flow between these regions. If differences in PA developed between regions, then the units with the higher PA would relatively increase their contribution to total flow to compensate for units with the lower recoil. Because flow would relatively increase from the region with the higher recoil, the fall in PA in this region would occur to a greater extent than without this compensation, and lung deflation would, to some degree, remain homogeneous. However, the extent to which IREF may be important in nonuniform lung disease to preserve homogeneity of PA has not been experimentally shown.

In the present study, lobar deflation was examined in a regional model of papain-induced emphysema (11). The alveolar capsule technique of Fredberg et al. (3) was used to monitor PA of the right lower lobe (PA_RLL) and left lower lobe (PA_LLL) during a forced VC maneuver (4, 5, 7, 13, 19). Emptying of the normal and emphysema lobes was compared at different fractions of the VC. The objectives were to determine how regional emphysema altered max and parameters of flow limitation and to assess the extent to which IREF contributed to these findings.

METHODS

3

RESULTS

A central choke point was found in 9 of the 12 emphysema experiments. In these nine experiments, a choke point was identified at the trachea at VL >40% whole lung VC and at lobar bronchi at the lower VL. These nine dogs were analyzed together and constituted the emphysema group. In the other three dogs, no central choke point was found, and these three dogs were analyzed as additional emphysema experiments. The six dogs in which normal saline was instilled constituted the control group. In the control group, choke points at the respective VL were identified at locations similar to those described in the emphysema group. Whole lung VC measured 2.2 ± 0.23 liters in the emphysema group vs. 2.1 ± 0.17 liters in the control group.

In all groups, there was no difference in lobar pressure-volume behavior when curves were obtained statically and dynamically. The dynamic lobar pressure-volumes curves are shown in Fig. 1 during which all the other lobes were tied off. In the emphysema group (A), RLL volume was greater than LLL volume at all levels of PA. On the other hand, in the control group (B), pressure-volume curves were not different between the lower lobes. Although the dogs were randomized according to body weight, lobar volumes in the control group probably started out slightly larger than those in the emphysema group. This accounts for the slightly higher mean LLL volume in the control group, which was not significantly different between groups.

In the control and emphysema groups, lobar compliances were determined over multiple intervals of the pressure-volume curve. These results are shown in Table 1. In the emphysema group, for the emphysematous RLL, between 0 and 5 cmH₂O, lobar compliance was significantly higher than that found for the LLL, whereas it was significantly lower, between 10 and 20 cmH₂O and between 20 and 30 cmH₂O, respectively. In the control group, there were no differences in compliances between RLL and LLL in any intervals of pressure-volume curve.

Table 1. Lobar compliances in emphysema and control groups Transpulmonary Pressure Range 0-5 cmH2O 5-10 cmH2O 10-20 cmH2O 20-30 cmH2O Emphysema group (n = 9) RLL, l/cmH2O 0.120 ± 0.02* 0.026 ± 0.005  0.01 ± 0.002* 0.00034 ± 0.0002* LLL, l/cmH2O 0.069 ± 0.01  0.041 ± 0.008  0.022 ± 0.005  0.007 ± 0.002  Control group (n = 6) RLL, l/cmH2O 0.094 ± 0.009  0.047 ± 0.005  0.008 ± 0.001  0.002 ± 0.001  LLL, l/cmH2O 0.095 ± 0.008  0.035 ± 0.004  0.012 ± 0.005  0.002 ± 0.005 Values are means ± SE. RLL and LLL, right and left lower lobes, respectively. * Significantly different RLL vs. LLL within a group, P < 0.05 (paired t-test corrected for multiple comparisons). Significantly different between RLL vs. LLL groups, P < 0.05 (unpaired t-test corrected for multiple comparisons in which respective differences in RLL and LLL compliances were compared).

Figure 2 shows a dynamic PA vs. time curve obtained for a control (B; dog 4) and an emphysema dog (A; dog 6).

The mean values are shown in Fig. 3. At TLC, mean (±SE) PA_RLL and PA_LLL in the emphysema group averaged 30.4 ± 0.4 and 30 ± 1.2 cmH₂O, respectively, whereas corresponding values in the control group were 30.4 ± 1.4 vs. 30 ± 1.6 cmH₂O. In the emphysema group, over the course of deflation, there was significant interaction (P < 0.01) between PA and time. At 25 ms postdeflation, mean PA_RLL were lower than PA_LLL. Over the remaining course of the deflation, PA_RLL rose relative to PA_LLL. The LLL had emptied in all experiments by ~350 ms, whereas the RLL continued to empty in most experiments. In the control group, there was no interaction PA vs. time, and both lobes emptied uniformly over the remainder of the deflation.

In the emphysema group, the mean (±SE) time for RLL emptying (calculated from the initiation to the end of deflation) was greater than that found for the LLL [RLL: 450 ± 68 (SE) vs. LLL: 334 ± 10 ms, P < 0.05 between lobes; P < 0.05 between the two groups; note that one emphysema dog continued to empty until 800 ms; see legend for Fig. 3]. On the other hand, in the control group, the mean time for lobar emptying was similar for the two lobes (332 ± 37 ms for the RLL and 311 ± 50 ms for the LLL, resepectively).

In Fig. 4A, dynamic PA_RLL and PA_LLL measured in the emphysema group are plotted as a function of the whole lung VC. By two-way ANOVA, there was a significant interaction (P < 0.0001) between lobar PA (i.e., factor A) and VL (i.e., factor B). Interaction was examined to determine whether the relationship of PA to whole lung VC changed during the course of deflation. Interaction should be distinguished between the decreases in PA that occurred as VL decreased during deflation, which is given by factor B and which was statistically significant in both groups.

In the emphysema group (see Fig. 4A), at very high VL (i.e., 95 and 90% VC), PA_RLL were lower than PA_LLL; over the middle range of VL (i.e., 80 to 50% VC), PA_RLL were similar to PA_LLL; and at the lower VL, PA_RLL were higher than PA_LLL. In the control group (see Fig. 4B), there was no interaction between PA and VL, and PA_RLL were similar to PA_LLL at all VL examined.

In Fig. 5A, dPA/dt obtained in the emphysema group are plotted as a function of whole lung VC, and there was a significant interaction between dPA/dt and VL. In the emphysema group, at VL 40, whole lung VC and dPA/dt for RLL were significantly higher than dPA/dt for LLL. These findings were significantly different from those observed in the control group, in which there was no interaction between dPA/dt and VL.

In Fig. 6, whole lung maximum expiratory flow (max) are plotted for the emphysema and control groups over intervals of the VC. At VL >40%, whole lung VC and max were similar between groups, while at VL 40, whole lung VC and max obtained in the emphysema group were significantly lower than control group values.

Figure 7 shows the lobar flows calculated in the emphysema and control groups over the intervals of the whole lung VC. In the emphysema group, at 95 and 90% whole lung VC, RLL flows were less than LLL flows, whereas combined RLL and LLL flows were not different between groups. On the other hand, at VL 40% whole lung VC, RLL flows were greater that LLL values. In the control group, there were no differences in flows between lobes measured over the whole VC range.

At 95 and 90% whole lung VC, because in the emphysema group the respective PA_RLL were lower than PA_LLL, it was determined whether the reductions in RLL flows observed at these VL were appropriate for the corresponding lower PA. In this analysis, lobar flows calculated at 95 and 90% whole VC in the emphysema group were compared with respective flows in the control group, in which the PA of the emphysema group were slightly greater or equal to those in the control group. If, for comparable PA_RLL in the emphysema and control groups, RLL flows in the emphysema group were still lower than control group values, then other factors (such as interdependence of regional expiratory flow; see DISCUSSION) would need to be considered to explain the lower RLL flows found in the emphysema group. In the emphysema group, (see Fig. 4) the mean PA_RLL measured at 95% was ~13.5 cmH₂O, which was comparable to the PA of 13 cmH₂O found at 90% whole lung VC in the control group. At 90% whole lung VC, PA_RLL in the emphysema group was ~10 cmH₂O, which was slightly higher than the 8 cmH₂O PA_RLL found at 80% whole lung VC in the control group. The corresponding RLL flows measured at 95 and 90% whole lung VC in the emphysema group (see Fig. 7) were 2.3 ± 0.8 and 3.2 ± 0.8 l/s, respectively, whereas those in the control group at 90 and 80% whole lung VC were 5.5 ± 2.0 and 5.2 ± 0.6 l/s, respectively. When analyzed by two-way ANOVA, RLL flows measured in the emphysema group were still less than RLL flows in the control group (P < 0.01), even when PA_RLL in the emphysema group were slightly greater than those in the control group. Similarly, PA_LLL were approximately the same between groups at 95 and 90% whole lung VC, respectively (see Fig. 4). Yet, LLL flows measured in the emphysema group (see Fig. 7) were significantly greater than corresponding values in the control group (see DISCUSSION).

In both the control and emphysema groups, choke points were identified at the trachea for VL >40% whole lung VC, whereas for VL 40%, choke points were identified at approximately lobar bronchi. Table 2 shows the values of A^*, P^*, and Pend^* obtained at the multiple VL in the emphysema and control groups. There were no differences in these parameters between the two groups. In addition, RLL frictional resistances (see Table 2) were also not different in the two groups, although in the control group, frictional resistances measured at VL 40% whole lung VC tended to be greater than those in the emphysema group (see DISCUSSION).

Table 2. Choke-point variables in emphysema and control groups %Whole Lung Vital Capacity 95 90 80 70 60 50 40 30 20 Emphysema group P*, cmH2O 10.8 ± 4.8  6.4 ± 1.4   1.0 ± 1.4   3.5 ± 1.1   5.0 ± 1.1   7.7 ± 1.6   3.7 ± 1.4   3.0 ± 1.8   2.5 ± 1.1  Pend*, cmH2O 13.9 ± 2.4  9.7 ± 0.6  2.4 ± 1.1  0.7 ± 0.7  0.5 ± 0.9   0.9 ± 0.9  0.43 ± 1.0   0.9 ± 1.2   1.0 ± 1.0  A*, cm2 3.6 ± 1.4  4.1 ± 0.9  4.2 ± 1.1  3.3 ± 0.9  2.2 ± 0.45  1.7 ± 0.3  0.5 ± 0.25  0.30 ± 0.11  0.2 ± 0.11  Rfr, cmH2O · l1 · s 0.33 ± 0.22  0.525 ± 0.500  0.55 ± 0.49  0.659 ± 0.162  0.80 ± 0.400  0.74 ± 0.018  0.76 ± 0.126  0.32 ± 0.161  0.701 ± 0.360  Control group P*, cmH2O 12.0 ± 4.0  4.0 ± 1.4   2.3 ± 3.0   6.3 ± 2.9   7.6 ± 2.7   9.0 ± 0.04   6.6 ± 2.3   5.3 ± 0.3  0.33 ± 2  Pend*, cmH2O 15.8 ± 2.7  7.8 ± 1.3  0.8 ± 2.5   0.7 ± 1.9   2.1 ± 2.3   3.3 ± 1.7   3.2 ± 2.0   0.6 ± 1.1  2.0 ± 1.5  A*, cm2 3.9 ± 0.9  5.6 ± 1.4  3.0 ± 0.6  2.9 ± 0.6  2.7 ± 0.35  2.4 ± 0.35  0.4 ± 0.37  0.30 ± 0.5  0.5 ± 0.75  Rfr, cmH2O · l1 · s 0.35 ± 0.37  0.59 ± 0.392  0.55 ± 0.120  0.60 ± 0.119  0.70 ± 0.1  0.731 ± 0.86  1.76 ± 0.867  2.32 ± 0.85  1.78 ± 1.10 Values are means ± SE; n = 5, 3, and 3 dogs in emphysema group at 40, 30, and 20% whole lung vital capacity, respectively, and 5, 2, and 2 dogs in control group at 40, 30, and 20% whole lung vital capacity, respectively. P* and Pend*, lateral and end-on pressures at choke point, respectively; A*, cross section at choke point; Rfr, frictional resistance to right lobar bronchus.

In three dogs with unilobar emphysema, a tracheal choke point was not identified (additional emphysema experiments) so that there was no interdependence of flow between lobes. There was no apparent difference in the degree of emphysema produced in these dogs compared with that found in the emphysema group, as determined by RLL vs. LLL volume and compliance changes.

Figure 8 shows the PA_RLL and PA_LLL vs. whole lung VC curves obtained in these three experiments. Unlike in the emphysema group, these relationships were inconsistent between experiments. In dogs 2 and 3, choke points were identified at lobar bronchi between 95 and 20% whole lung VC. In contrast to the emphysema group in which, during the initiation of deflation, PA_RLL fell to a greater extent than PA_LLL, in dogs 2 and 3 PA_RLL were greater than PA_LLL by ~2 cmH₂O during the entire deflation period (see DISCUSSION). In dog 1, choke points were identified in mainstem bronchi at VL >40% whole lung VC. In a manner similar to that found in the emphysema group, in dog 1 during initial deflation PA_RLL fell more rapidly than PA_LLL. However, unlike in the emphysema group, PA_RLL remained much lower than PA_LLL by ~2-3 cmH₂O until very late in deflation (30% whole lung VC), after which PA_RLL became >PA_LLL.

The VL at which PA_RLL became equal to PA_LLL was compared in the emphysema group and additional emphysema experiments. In the emphysema group, this VL averaged 69 ± 4% whole lung VC, whereas in the additional emphysema experiments it occurred at the end of the VC maneuver and averaged 10 ± 10% whole lung VC (P < 0.001 between groups). Moreover, in the additional emphysema experiments, max at all VL appeared lower than those measured in the control group and were significantly reduced at 95 and 90% whole lung VC (see Table 3).

Table 3. Whole lung maximum expiratory flow in additional emphysema experiments %Whole Lung Vital Capacity 95 90 80 70 60 50 40 30 20 7.2 ± 1.0* 8.1 ± 0.7* 7.5 ± 1.4  5.4 ± 1.2  4.6 ± 1.2  3.5 ± 1.2  2.9 ± 0.9  2.5 ± 1.0  1.5 ± 0.6 Values are means ±SE in l/s; n = 3 dogs. * Significantly different vs. control group, P < 0.05 (unpaired t-test corrected for multiple comparisons).

DISCUSSION

In the emphysema group, the results showed that only at the extremes of the whole lung VC maneuver were PA different between the emphysematous and normal lower lobes. During early expiration, because compliance of the emphysematous lobe was lower at high Ptp, a gradient in PA was quickly established between the lower lobes, and, at 95 and 90% whole lung VC, PA_RLL were lower than PA_LLL (see Fig. 4). As deflation continued, PA_RLL rose relative to PA_LLL, and between 80 and 50% whole lung VC, PA_RLL approximated PA_LLL. Eventually, however, PA_RLL became higher than PA_LLL so that, at 40% whole lung VC, the expirate came predominantly from the emphysematous RLL for the remainder of the deflation. Moreover, in the emphysema group, despite the changes in lobar PA that occurred between lobes, max_tot were not different between the emphysema and control groups until the lower VL were reached. The results obtained in the emphysema group are examined below in terms of possible mechanisms that would regulate emptying of the normal and emphysematous regions over the course of the VC maneuver in this model.

It could be argued that in the emphysema group, the changes in PA and flows that were observed between lobes during deflation reflected only the effect of the RLL compliance changes and that there was no evidence that flow interaction between lobes had contributed to these findings. In that case, compared with the LLL, because the compliance of the emphysematous RLL lobe was smaller at high Ptp, PA of the emphysematous lobe decreased faster during early deflation. However, because the compliance of the emphysematous lobe was relatively greater at low Ptp, RLL emptying continued whereas that of the LLL had already ceased. There would be no need to invoke other mechanisms such as flow interdependence between regions to explain any of the findings observed in the emphysema group.

However, the results shown in Fig. 7, in which lobar flows are plotted as a function of %whole lung VC, contradict this view. In the emphysema group, compared with the values obtained in control group, there were marked changes in RLL and LLL flows at the high lung regions that would support a role for a flow interdependent mechanism between regions. The rationale for this conclusion is as follows.

As previously discussed, PA_RLL were less than PA_LLL during early expiration because of the lower compliance of the emphysematous lobe (see Fig. 4). Wilson et al. (20) indicated that when regions share a common downstream airway, flow from each region depends on the driving pressure for other regions, and flow from the region with the higher driving pressure is favored. For a mean PA (averaged among the different regions), flows from the region with the higher PA would increase, whereas flow from the other would decrease compared with results obtained when PA were similar between regions. Solway et al. (17) showed similar results in a transistor model of nonhomogeneous airflow obstruction. When regions shared a common choke point, if flows from one region were reduced by an obstruction, then the other region would increase its flow rate. In my laboratory, a similar conclusion was reached in a canine model of nonhomogeneous airflow obstruction (12).

In the emphysema group, at 95 and 90% whole lung VC, because a choke point was common to both lower lobes and because the downstream pressure (i.e., P^*) was the same for both lobes, in terms of the results of Wilson et al. (20) the LLL would be favored for flow because PA_LLL > PA_RLL. Therefore, in the emphysema group, LLL flows were significantly greater than RLL flows at 95 and 90% whole lung VC (see Fig. 7).

Yet, in the emphysema group, how does one know whether this reduction in RLL flows would have occurred without invoking a mechanism such as IREF? That is, if flow from both lobes were not interdependent, and if PA_RLL < PA_LLL because of a lower RLL compliance during early deflation, then what would have been the resulting RLL and LLL flows in the emphysema group?

First of all, it can be observed in Fig. 7 that LLL flows obtained in the emphysema group at 95 and 90% whole lung VC were greater than corresponding LLL flows in the control group. In the canine lung, flows are fairly constant from TLC to ~50% VC (see Fig. 6) so that there would be few changes in LLL flows expected in the emphysema group for the small differences in PA_LLL found between groups at these VL. Accordingly, why should LLL flows increase in the emphysema group if not due to a mechanism such as IREF?

Second, at 95% whole lung VC, PA_RLL in the emphysema group averaged ~13 cmH₂O and RLL flows averaged ~2.5 l/s. On the other hand, in the control group, at 90% VC, PA_RLL measured ~11 cmH₂O, and RLL flows measured ~5 l/s. Thus, for similar PA_RLL measured in the two groups, RLL flows were much lower that those found in the control group. It appears that in the emphysema group, LLL flows increased by ~2 l/s (~30%), whereas RLL flows decreased by ~2 l/s, and these reciprocal changes are exactly what was predicted by Wilson et al. (20).

Moreover, Wilson et al. (20) showed that when resistance was unchanged between regions (see Table 2 at high VL), the relative flows between these regions (_RLL  LLL)/_avg) would be determined by their corresponding differences in PA [(PA_RLL  PA_LLL)/(PA_avg  P^*)]. In the emphysema group, PA between lobes at 95% whole lung VC was nearly equal to 3 cmH₂O (see Fig. 4), whereas PA_avg  P^* was 15  10.8 = 4.2 cmH₂O (see Table 2). This would give a pressure ratio of 0.71, which was slightly less than the measured flow ratio of 0.88 calculated from the lobar flows in Fig. 7 [i.e., (2.5  6.5)/4.5]. Accordingly, the reciprocal changes in LLL and RLL flows observed in the emphysema group are in agreement with the predictions of Wilson et al. (20) and support the conclusion that there was interaction of expiratory flows between lobes that preserved uniform PA decline in the emphysema group.

Furthermore, the results are also in agreement with those of Topulos et al. (19), who showed that PA differences between lobes develop very early during the course of deflation. In the condition in which a common choke point is present, near-maximal interlobar pressure differences coincided with peak flow. Thus the maximum effect of IREF was observed at high VL in the emphysema group.

Accordingly, without a flow interdependent mechanism, RLL flow measured at 95 and 90% VC in the emphysema group would have nearly doubled or doubled, and LLL would have decreased by ~30% compared with the results observed in Fig. 7. In turn, there would have been a marked widening of the PA difference between lobes at 95 and 90% whole lung VC. For instance, at 90% whole lung VC, ~0.200 liter had expired in the emphysema group [i.e., 10% expired × whole lung VC (~2 liters) = 0.2 liter], and from Fig. 4, PA_RLL was 9 cmH₂O and PA_LLL was 11 cmH₂O. On the basis of the respective RLL and LLL pressure-volume curves shown in Fig. 1, of the 0.200 liter expired, 40 ml came from the emphysematous lobe, 100 ml came from the LLL, and 60 ml came from the other normal lobes. In the condition under which RLL flow doubled, and where LLL flow decreased by 30%, then the amount expired by the emphysematous lobe would have increased to ~80 ml, and the amount expired by the LLL would have decreased to ~70 ml. From Fig. 1, the resulting RLL and PA_LLL would have been 7.5 and 17.5 cmH₂O, respectively. Thus, in the case where regional interdependence of flow did not occur, the PA difference at 90% whole lung VC would have widened from 2 to 10 cmH₂O.

If, at 90% whole lung VC, PA_RLL and PA_LLL were 7.5 and 17.5 cmH₂O, respectively, and if this analysis were continued from 90 to 80% whole lung VC, then another 0.200 liter would be expired. If there were no IREF, then on the basis of the assumptions and estimations described in the Appendix, at 80% whole lung VC, PA_RLL would be ~5 cmH₂O and PA_LLL would be ~8 cmH₂O. Similarly, it is also estimated in the Appendix that at 70% whole VC, PA_RLL and PA_LLL would be 4 and 6 cmH₂O, respectively; at 60% whole lung VC, PA_RLL and PA_LLL would be 2.5 and 4 cmH₂O, respectively; and at 50% whole lung VC, PA_RLL and PA_LLL would be 2 and 3 cmH₂O, respectively (see Appendix).

Thus, in the condition in which IREF was absent (IREF-absent condition), over the mid-VC range, the PA difference between the emphysematous and normal lobes would have widened compared with what was observed in the emphysema group. It is noteworthy that in this emphysema model, the changes in RLL compliances produced would dictate that the initial fall in PA_RLL would be greater than that found for the LLL, after which PA_RLL would rise. However, without IREF, the PA difference predicted to occur between 60 and 80% whole lung VC would be ~1.5-4 cmH₂O between lobes, rather than the 0 difference shown in Fig. 4. Without IREF, uniform PA among lobes would not occur until <50% whole lung VC. Accordingly, this study points out that IREF played an important role in the maintenance of the relatively uniform PA decline over the high mid-VC range (i.e., 80 to 50% whole lung VC) in the emphysema group.

In the emphysema group, Fig. 6 shows that at >40% whole lung VC, the respective max_tot values were not different from those found in the control group. Between 80 and 50% whole lung VC, because PA were not different between the emphysema and control groups and because RLL resistances were also not different between groups, the pressure-head measured at the choke point (Pend^*) was unchanged in the emphysema group (see Table 2). Because wave speed was reached at the same airway site with a similar Pend^* in the two groups, wave-speed variables and hence max were unchanged at these VL in the emphysema group (2).

In the emphysema group, although IREF maintained a relatively uniform PA decline between lobes over the mid-whole lung VC range, what is the evidence that IREF contributed to the unchanged max_tot? In the IREF-absent condition, it was previously shown (see above) that at VL >40% whole lung VC, differences in PA would further widen compared with values found in the emphysema group. Then, if IREF were not present, what would be the effect of this widening on max_tot and lobar flows compared with values obtained during homogeneous deflation? At 90, 80, and 70% whole lung VC, it seems unlikely that there would be much of a difference. Because flows for both lobes are relatively constant (i.e., 5 l/s) between PA of 3 and 20 cmH₂O (see Fig. 7B), substitution of the PA calculated to occur in IREF-absent condition for those measured during homogenous deflation would not alter the sum of lobar max very much. In either case, the sum of RLL and LLL flows would average ~10 l/s.

However, at 60 and 50% whole lung VC, there is evidence that IREF could play a role in the maintenance of max_tot. During the IREF-absent condition (see Appendix), it was predicted that at 60% whole lung VC, PA_RLL would be 2.5 cmH₂O and PA_LLL would be 4 cmH₂O, respectively. At a PA of 2.5 cmH₂O (i.e., which corresponds to a VL between 40 and 30% whole lung VC), RLL flows would be ~2.0 l/s (see data in Fig. 7 at 30% whole lung VC). At a PA_LLL of 4 cmH₂O (i.e., which corresponds to 60% VC in Fig. 4), LLL flows would be ~3.0 l/s (see data at 40% VC in Fig. 7). One can observe that the combined lobar flows of 5.0 l/s would be <8.5 l/s measured at 60% whole lung VC during homogeneous deflation, in which RLL and LLL flows were 5 and 3.5 l/s, respectively (see Fig. 7B).

At 50% whole lung VC, in the IREF-absent condition (see Appendix), PA_RLL was predicted to be 2 cmH₂O (which corresponds to 30% VC in Fig. 4), and PA_LLL was predicted to be 3.0 cmH₂O (which corresponds to between 50 and 40% VC in Fig. 4). The respective flows obtained from Fig. 7B would be 1.5 and 3 l/s. The combined predicted lobar flows of 4.5 l/s are <8 l/s that were measured at 50% whole lung VC when deflation was homogeneous in Fig. 7.

Thus this analysis shows that there would be little change in max_tot at high VL (i.e., >60% whole lung VC), if IREF were not present. Because the canine flow-volume curve is relatively constant over a large range of PA, widening of the PA difference that would occur in the IREF-absent condition would not substantially change the sum of lobar flows compared with those found during homogeneous deflation. However, as the lung deflates, choke points start to jump into lobar bronchi, and then small differences in PA may cause large differences in lobar flows. In Fig. 7B, particularly for the RLL, flows decreased from 5 to 2 l/s when PA reached a critical value of <3 cmH₂O. Then, IREF would prevent PA_RLL from reaching this critical value at too high a VL, and in turn, this effect would preserve RLL flows at 60 and 50% whole lung VC.

In the emphysema group, the sum of RLL and LLL flows measured at 60% whole lung VC in Fig. 7 were slightly greater than what would be predicted during the IREF-absent condition (6 vs. 5 l/s). At 50% whole lung VC, the sum of lobar flows in the emphysema group (4.5 l/s) were similar to those calculated for the IREF-absent condition. In terms of this analysis, preservation of max_tot was observed mainly at 60% whole lung VC. Thus the effect of IREF on max in this emphysema model was relatively modest.

The canine lung might be relatively insensitive to IREF because over the upper one-half of the VC maneuver, flows are relatively constant, despite large decreases in lung recoil. The extent to which IREF will preserve regional flows depends on the precise nature of the pressure-flow relationships that are found in the different regions, which will, in turn, determine the reciprocal changes in regional flow that will occur. For instance, in the human lung, flow is more sensitive to changes in VL. In that case, without IREF, an increase in PA in the normal lobe might not offset the decrease in PA in the emphysema lobe so that a decrease in max_tot would be observed over a greater proportion of the VC maneuver in nonhomogeneous disease. Moreover, in some experiments in the present study (see additional emphysema experiments), a central choke point did not occur despite the fact that a lesion similar to one found in the emphysema group was produced. Thus, under different conditions, the effect of IREF on max_tot may vary in importance. It is difficult to make general inferences about the effect of IREF on max_tot in regional lung disease without making the measurements.

At VL 40% whole lung VC, a central choke point was not found in the emphysema group, and max_tot measured were less than control group values (see Fig. 6). At these VL, choke points were identified at lobar bronchi and there was no interdependence of flow between regions. In the emphysema group, because at the high VL most of the early deflation had come from the LLL and because RLL compliance was increased relative to LLL compliance, PA_LLL < PA_RLL at the low VL (see Fig. 4). Because frictional resistance is not increased in this model, the pressure head at the RLL choke point [(PA  frictional pressure (Pfr)] would be greater than that for the LLL. Hence, RLL flow was less than LLL flow. In the emphysema group, combined flows from the RLL and LLL were less than those found in the control group so that max_tot were smaller at the low VL in the emphysema group.

Furthermore, at VL 40% whole lung VC, frictional resistances measured in the control group appeared to be slightly greater than those in the emphysema group, although the numbers were small and there were wide SE in the two groups. At these VL, because PA_RLL values obtained in the emphysema group were slightly higher than corresponding values found in the control group, it is likely that airway diameter was larger and therefore frictional resistances were lower in the emphysema group. Finally, it is important to note that when choke points move into lobar bronchi, flow from the subtended region would be governed by local choke point pressure-area behavior. This behavior is difficult to predict and may vary between regions. This behavior may be abnormal in emphysema because of destruction of tissue attachments to bronchi (10) (see below).

In the additional emphysema experiments, regional interdependence of flow was not a factor in the regulation of RLL and LLL flow. In dog 1, choke points were identified at mainstem bronchi. In this dog, as in the emphysema group, PA_RLL fell faster than PA_LLL at the beginning of expiration. The difference between PA_RLL and PA_LLL averaged 2-3 cmH₂O over most of the VC. This difference in PA appeared bigger than values found in the emphysema group, and moreover, PA_RLL did not equal PA_LLL until 30% whole lung VC was reached. This greater nonuniformity of deflation is consistent with what was predicted in the previous analysis when regional interdependence of flow was not a factor in controlling lobar emptying in the emphysema group (see above).

In dogs 2 and 3, lobar choke points were observed throughout the whole VC range. In these dogs, in contrast to what was observed in the emphysema group, PA_RLL did not decrease faster than PA_LLL at the beginning of expiration. In dogs 2 and 3, the changes in RLL mechanical properties were similar to those produced in the emphysema group. The reason for the slower PA_RLL decline in dogs 2 and 3 is that, when lobar choke points are found in emphysematous dogs, choke point pressure-area behavior may be altered. For a given pressure head, choke point area and hence max are less than what would be predicted from a normal airway. In a previous study (13), this finding was ascribed to loss of bronchial tissue attachments due to destruction in this papain model. Thus in dogs 2 and 3 during early deflation, PA_RLL decline was less than that found for the LLL, despite reduced compliance at high Ptp in the emphysema lobe.

The present study shows that despite the changes in compliance produced in the emphysematous lobe, PA decline over the high mid-vital range was relatively uniform in the emphysema group. This occurred because of a flow interdependence mechanism that helped to maintain PA pressures between regions quite homogeneous. In the emphysema lobe, at high VL, compliance was lower than normal values, whereas at low VL compliance was higher than normal values. On the other hand, there were no changes in frictional resistances between lobes. The papain model resembles panacinar emphysema rather than centrilobular emphysema (10, 18). Centrilobular emphysema is associated with cigarette smoking and often is accompanied by an increase in frictional resistance (10, 18). Because frictional resistance may cause upstream movement of choke points (9, 16, 17), a common choke point among lobes may not occur in human disease, and the present results must be applied cautiously to humans.

The present study shows that when a common choke point was present, a flow interdependence mechanism limited the extent to which differences in PA developed between regions in a canine model of regional emphysema. IREF appeared to maintain choke points centrally over a greater proportion of the whole lung VC maneuver compared with what would occur without IREF. This preserved max_tot at VL at which, without IREF, choke points would move from central locations into lobar airways. These results show that IREF may affect the dynamics of flow limitation in regional lung disease.

ACKNOWLEDGEMENTS

This work was supported by the Medical Research Council of Canada.

FOOTNOTES

Address for reprint requests: S. N. Mink, GF-221, Health Science Center, 700 William Ave., Winnipeg, Manitoba, Canada R3E-OZ3.

Received 28 September 1995; accepted in final form 30 September 1996.

APPENDIX

Estimation of PA_RLL and PA_LLL in IREF-Absent Condition

If, at 90% whole lung VC, the predicted PA_RLL and PA_LLL in the emphysema group were 7.5 and 17.5 cmH₂O, respectively, then flow from the RLL at a PA of 7.5 cmH₂O would be 5 l/s, and flow from the LLL at a PA of 17.5 cmH₂O would be also be 5 l/s. This is because lobar flows predicted to occur without IREF would be those found in the control group. Then, at a PA_RLL of 7.5 cmH₂O, whole lung VC and hence lobar VC taken from Fig. 4B would be 70% because deflation is homogeneous in the control group. In turn, lobar flow measured at 70% whole lung VC from Fig. 7B would be ~5 l/s. Similarly, at a PA_LLL of 17.5 cmH₂O (which corresponds to a whole lung and lobar VC >95%; see Fig. 4), LLL flow would also be ~5 l/s (from Fig. 7B).

Furthermore, max_tot is the sum of RLL flow, LLL flow, and the combined flows from the remaining lobes. The combined flows from the remaining lobes between 90 and 80% whole lung VC would be [60% × LLL flow (5 l/s)] = 3 l/s. The sum of RLL flow, LLL flow, and flow from the remaining lobes would yield a predicted max_tot of ~13 l/s. In that case, the time for deflation between 90 and 80% whole lung VC would be 0.015 s (i.e., 0.2 liter  13 l/s). Of the 0.2 liter expired between 90 and 80% whole lung VC, 75 ml (i.e., 5 l/s × 0.015 s) would come from the RLL, and 75 ml (i.e., 5 l/s × 0.15 s) would come from the LLL, whereas the remainder would occur from the other lung lobes. From the RLL and LLL pressure volumes curves obtained in the emphysema group (see Fig. 1A), expiration of an additional 75 ml from the respective lobes would yield PA_RLL of ~5 cmH₂O and PA_LLL of ~8 cmH₂O.

Similarly, between 80 and 70% whole lung VC, the RLL at a PA of 5 cmH₂O would be deflating at ~5 l/s (Figs. 4B, 7B), whereas LLL flow at a PA of 8 cmH₂O would also be deflating at ~5 l/s. Then, predicted max_tot would again be 13 l/s. Of the 0.2 liter expired, 75 ml would come from the RLL (5 l/s × 0.015 s) and 75 ml (5 l/s × 0.015 s) would come from the LLL. The resulting PA_RLL would be ~4.0 cmH₂O, and the PA_LLL would be ~6.0 cmH₂O.

Between 70 and 6O% whole lung VC, PA_RLL of 4 cmH₂O would again be deflating at ~5 l/s and PA_LLL of 6 cmH₂O would also be deflating at ~5 l/s. The predicted max_tot would again be ~13 l/s. Seventy-five milliliters would be expired from each lobe, and PA_RLL and PA_LLL would fall to 2.5 and 4 cmH₂O, respectively. Finally, between 60 and 50% whole lung VC, RLL flow measured at PA of 2.5 cmH₂O would fall to 1.5 l/s, whereas LLL flow at PA of 4 cmH₂O would be 3.0 l/s. The predicted max_tot (6.8 l/s) would be the sum of the RLL flow (2.0 l/s), LLL flow (3.0 l/s) and flow from the remaining lobes (0.60 × 3.0 l/s). The time to deflate the next 0.2 liter would be 0.025 s. The RLL would expire 63 ml and the LLL 88 ml. PA_RLL would fall to 2 cmH₂O, and PA_LLL would fall to 3 cmH₂O.

REFERENCES

1.	Christie, R. V. The elastic properties of the emphysematous lung and their clinical significance. J. Clin. Invest. 13: 295-321, 1934.
2.	Dawson, S. V., and E. A. Elliott. Wave-speed limitation on expiratory flowa unifying concept. J. Appl. Physiol. 43: 498-515, 1977.
3.	Fredberg, J., D. Keefe, G. Glass, R. Castile, and I. Frantz. Alveolar pressure non-homogeneity during small amplitude-high frequency oscillations. J. Appl. Physiol. 57: 788-800, 1984.
4.	Georgopoulos, D., A. Gomez, and S. Mink. Factors determining lobar emptying during maximal and partial forced deflations in nonhomogeneous airway obstruction in dogs. Am. J. Respir. Crit. Care Med. 149: 1241-1247, 1994.
5.	Georgopoulos, D., S. N. Mink, L. Oppenheimer, and N. R. Anthonisen. Heterogeneity of maximal lobar emptying rates in dogs with compensatory lung growth. J. Appl. Physiol. 67: 1164-70, 1989.
6.	Jadue, C, H. Greville, J. Coalson, and S. Mink. Forced expiration and HeO2 response in canine peripheral airway obstruction. J. Appl. Physiol. 58: 1788-1801, 1984.
7.	McNamara, J. R., G. Castile, G. Glass, and J. Fredberg. Heterogeneous lung emptying during forced expiration. J. Appl. Physiol. 63: 1648-1657, 1987.
8.	Mead, J., J. M. Turner, P. T. Macklem, and J. B. Little. The significance of the relationship between lung recoil and maximum expiratory flow. J. Appl. Physiol. 22: 95-108, 1967.
9.	Mink, S. Mechanism of reduced maximum expiratory flow in methacholine-induced bronchospasm in dogs. J. Appl. Physiol. 55: 897-912, 1983.
10.	Mink, S. Expiratory flow limitation and the response to breathing a helium-oxygen gas mixture in a canine model of pulmonary emphysema. J. Clin. Invest. 73: 1321-1334, 1984.
11.	Mink, S. N., A. Gomez, L. Whitley, and J. J. Coalson. Hemodynamics in dogs with pulmonary hypertension due to emphysema. Lung 164: 41-54, 1986.
12.	Mink, S. N., H. Greville, A. Gomez, and J. Eng. Expiratory flow limitation in dogs with regional changes in lung mechanical properties. J. Appl. Physiol. 64: 162-173, 1988.
13.	Mink, S. N., S. G. Holtby, D. J. Berezanski, L. Oppenheimer, and N. R. Anthonisen. Heterogeneity of maximal lobar empty ing rates in dogs with compensatory lung growth. J. Appl. Physiol. 67: 1164-1170, 1989.
14.	Mink, S. N., H. W. Unruh, and L. Oppenheimer. Vascular and interstitial mechanics in canine pulmonary emphysema. J. Appl. Physiol. 59: 1704-1715, 1985.
15.	Mink, S., M. Ziesmann, and L. D. H. Wood. Mechanisms of increased maximum expiratory flow during HeO2 breathing in dogs. J. Appl. Physiol. 47: 490-502, 1979.
16.	Pedersen, O. F., R. G. Castile, J. Drazen, and R. Ingram, Jr. Density dependence of maximum expiratory flow in the dog. J. Appl. Physiol. 53: 397-404, 1982.
17.	Solway, J., J. J. Fredberg, R. Ingram, O. Pedersen, and J. Drazen. Interdependence of regional lung emptying during forced expiration: a transistor model. J. Appl. Physiol. 62: 2013-2025, 1987.
18.	Thurlbeck, W. M., J. A. M Henderson, R. G. Fraser, and D. V. Bates. Chronic obstructive lung disease: a comparison between clinical, roentgenologic, function, and morphologic criteria in chronic bronchitis, emphysema, asthma, and bronchiectasis. Medicine 49: 81-145, 1970.
19.	Topulos, G. P., G. Nielan, G. Glass, and J. J. Fredberg. Interdependence of regional expiratory flows limits alveolar pressure differences. J. Appl. Physiol. 69: 1413-1418, 1990.
20.	Wilson, T. A., J. Fredberg, J. Rodarte, and R. Hyatt. Interdependence of regional expiratory flow. J. Appl. Physiol. 59: 1924-1928, 1985.