INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE ZONAL MODEL describing regional pulmonary blood flow distribution states that in zone 1 regions, where alveolar pressure (PA) exceeds both the pulmonary arterial (Ppa) and pulmonary venous pressures (Ppv), alveolar capillaries are collapsed and there is no flow. Numerous investigators have, however, previously observed open vessels in alveolar corners under zone 1 conditions (3-6, 16, 20, 22, 23), and we found (13) that up to 15% of the normal resting cardiac output can go through lungs that are completely in zone 1. Using in vivo video microscopy through a pleural window, we directly observed that blood flowed through subpleural vessels under zone 1 conditions and that the open pathway used alveolar corner vessels (14). We subsequently demonstrated that the blood perfusing through the vessels that were open in zone 1 contributed to gas exchange (15).

Because others have suggested that there are anatomic differences between vessels located in subpleural lung regions and those located deeper in the interior (9, 18, 24, 25), we were concerned that our findings regarding zone 1 flow through subpleural vessels might not accurately reflect flow in deeper, interior vessels. Accordingly, we quantified regional pulmonary flow using radiolabeled microspheres and compared the relative subpleural and interior distributions, as well as the dependent-to-nondependent perfusion gradients, under both zone 1 and zone 2 conditions in excised lungs held upright or inverted. We also assessed the relative flow distributions and the dependent-to-nondependent gradient in upright and inverted lungs in vivo.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Eleven New Zealand White rabbits (2.6 ± 0.1 kg) were given pentobarbital sodium (30-60 mg/kg) and papaverine (3 mg/kg) intravenously, and afterward a tracheotomy was performed. The rabbits were ventilated with room air [tidal volume (VT) = 30 ml, frequency (f ) = 30 breaths/min].

In six of the rabbits, the jugular vein was cannulated. With the ventilator stopped at end exhalation, ~1,000,000 radiolabeled microspheres (15-µm diameter, labeled with either ¹³³Sn, ¹⁰³Ru, ⁹⁵Nb, ¹⁴¹Ce, or ⁴⁶Sc chosen randomly for the various conditions) were injected while three of the rabbits were held in a head-up position and three were held in an inverted, head-down position. The chest was then opened widely, and heparin (1,000 U) was administered by right ventricular puncture. Approximately 60 ml of blood were removed from the right atrium, and cannulas were inserted into the pulmonary artery and the left atrium via the right and left ventricles, respectively. The lungs were perfused with Tris-buffered Tyrode solution containing 6% dextran (Sigma Chemical, St. Louis, MO) and 15 mg/l papaverine until the venous outflow was clear (which required ~30-60 ml). The lung and heart were removed en bloc and positioned either upright (by suspending them by the trachea) or inverted (by hanging the lungs from their caudal surface using large, thick foam pads that were attached with tissue glue). Pulmonary vascular pressures were referenced to the most dependent part of the lungs.

Zone 2 conditions were set with PA = 25, Ppa = 40, and Ppv = 0 cmH₂O. Flow through the lungs was measured by timed collection from the venous outflow, after which the inflow circuit was switched to one containing Tyrode solution to which ~1,000,000 microspheres with a different radiolabel had been added. This second microsphere injection was accomplished while the same zone 2 conditions were maintained. We then returned to the original inflow circuit and remeasured flow from the venous outflow. Ppa was then reduced to 22.5 cmH₂O to produce zone 1 conditions throughout the entire lung (PA = 25, Ppa = 22.5, and Ppv = 0 cmH₂O). Flow was remeasured until outflow was constant. The inflow circuit was again switched to the one containing Tyrode solution plus ~1,000,000 microspheres with a third label. These were infused while the same zone 1 conditions were maintained, then the inflow circuit was returned to the original, and venous outflow was remeasured until it was constant.

In five rabbits, the lungs were prepared as described above and suspended either upright (n = 3) or inverted (n = 2) without a preceding in vivo microsphere injection. After establishment and measurement of flow under the same zone 1 conditions as described above (i.e., PA = 25, Ppa = 22.5, and Ppv = 0 cmH₂O), ~2,000,000 labeled microspheres were infused. Samples of perfusate obtained from left atrial effluentshowed that <0.1% of the microspheres passed through the lungs.

At the end of the experiments, Ppa was lowered to 0 cmH₂O, the lungs were held inflated at 25 cmH₂O for air drying for ~24 h, and then they were further dried in a vacuum oven at 60 °C for ~48 h. Lungs were then sliced into 1-cm transverse sections in isogravitational planes (Fig. 1) and subsequently divided into three approximately equal planes in the ventral-dorsal axis (Fig. 1, inset). This process produced ~36 sections from each pair of lungs. That portion of the lung tissue located <1 mm from the pleura was carefully dissected from the remaining tissue (Fig. 1). Tissue obtained from regions located directly under the foam pads used for the inverted lungs was discarded because the glue used to adhere the pads would preclude accurate weight corrections. Airways and vessels >1 mm in diameter were dissected away from the parenchyma and discarded. The average counts of all of the discarded tissue were 8.9 ± 0.6% of the total.

View larger version (39K): [in this window] [in a new window]   Fig. 1.   Rabbit lungs were sliced into 1-cm isogravitional planes after thorough drying. Subpleural tissue was separated from the deeper, interior tissue by microdissection. Inset: pattern of dorsal-ventral sectioning (see METHODS).

The radioactivity of each tissue piece was determined in a 3 × 3.25-in. sodium-well crystal gamma counter (Minaxi gamma counting system, model 5550, Packard, Downers Grove, IL). Each sample was corrected for decay time and spillover by the matrix inversion method (7). The highest count rates were an order of magnitude below the rate at which crystal "dead time" occurs. Relative flow to the subpleural vs. interior tissue was standardized by expressing regional flow as a fraction of total flow per gram of dried tissue.

Statistics. All data are presented as means ± SE. Student's paired t-test was used to compare the pleural vs. interior differences. ANOVA was used to compare flows in the two zonal conditions. Linear regressions were fit to the data to relate flow to lung height, with the most dependent and most nondependent slices omitted because these routinely had markedly reduced flows.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subpleural vs. interior flow distribution. Body position (upright or inverted) had no effect on the relative flow distribution between the subpleural and interior regions or on actual measured flow. Accordingly, the data from lungs held in both positions were grouped together and are presented in Table 1, which shows that the relative flow per gram to interior tissue was ~40% greater than that going to subpleural tissue, regardless of the zonal condition. In the excised lungs, measured flow under both zonal conditions was unaffected by the addition of microspheres. Flow in vivo was not measured, but we assume that each of the zonal conditions existed in some portion of the lung, presumably with the majority being in zones 2 and 3. 

Dependent to nondependent flow distribution. Mean weight-normalized relative flow to the subpleural and interior regions in each isogravitational slice are shown in two representative lungs in Fig. 2. Data from all 12 lungs are summarized in Table 2.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Representative mean data of weight-normalized relative flow per isogravitional slice for 1 upright (A) and 1 inverted (B) rabbit lung in all 3 conditions. , Interior tissue; , subpleural tissue. Lines represent linear regressions omitting data from the end slices (see RESULTS). cts, Counts.

Isogravitational perfusion heterogeneity. There was as much perfusion heterogeneity within most isogravitional slices as there was throughout the entire height of the lung (the individual data points making up the mean data for the representative lung shown in Fig. 2B are presented in Fig. 3). When the linear regressions were recalculated from the individual lung pieces rather than from the means of each slice, the correlations decreased considerably, and only 25-75% of the variation could be attributed to height (Table 2).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Representative data of weight-normalized relative flow per individual piece for each isogravitional slice for 1 inverted rabbit lung in all 3 conditions (same data averaged in Fig. 2B). Note the large heterogeneity of flows to individual pieces within an isogravitional slice.

Ventral-dorsal flow distribution. With the exception of one lung perfused under zone 1 conditions, flow was always greater in dorsal compared with ventral regions (Fig. 4). Flows for the subpleural and the interior tissues were combined because both showed the same trend. Ventral-dorsal sectioning was not done to an absolute thickness, and therefore a gradient of flow per centimeter could not be determined, but the increased flow in the dorsal regions was obvious.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Combined weight-normalized relative flow for pleural and interior tissue for each set of lungs divided into dorsal, middle, and ventral regions. Data for upright lungs (solid lines and symbols) and inverted lungs (dashed lines and open symbols) are shown.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The important findings of this study are that 1) zone 1 flow is not unique to the subpleural lung region, but it is considerably less than flow that occurs deeper in the lung interior; 2) the relative distribution of flow to subpleural vs. interior regions is remarkably similar regardless of whether the lung is in zone 1 or 2 conditions; 3) the previously demonstrated function of scale of measurement on the relationship between the dependent-to-nondependent flow gradient and lung height applies to flow going to both subpleural and parenchymal regions, under both zone 1 and 2 conditions; and 4) as has been shown in zones 2 and 3, zone 1 flow (to both subpleural and interior tissue) is also preferentially distributed to the dorsal lung regions of isogravitational planes.

Methods. Although the radiolabeled microspheres were well mixed in the perfusate just before infusion, their density of 1.4 g/cm³ makes them settle out of solution with time. To test that the settling of the beads at the slower flow rates encountered in zone 1 did not contribute to the flow distribution we observed, in one experiment we added blue-green fluorescent polystyrene beads (density = 1.055 g/cm³, 15-µm diameter) to the perfusate at the same time the radiolabeled microspheres were added, and we infused both under zone 1 conditions. Direct comparisons of the flow distribution observed showed little difference, suggesting that bead density did not adversely affect our results.

Subpleural vs. interior flow distribution. We found that flow to the lung tissue located <1 mm from the pleural surface was 65-71% of that in vessels located deeper in the interior (Table 1). This finding was observed in vivo and in zone 1 and 2 conditions in excised lungs, regardless of whether the lungs were positioned upright or inverted. We (7) and others (10) have previously shown that lung perfusion in vivo has a radial component with perfusion decreasing from more central to more peripheral regions. In 1937, Miller (18) described a capillary network on the surface of the pleura and around the bronchi that was one-fourth as dense as the network found in alveoli located deeper in the lung. Guntheroth and colleagues (9) confirmed this configuration using scanning-electron micrographs in the rat. Our findings are similar to those of Short and colleagues (24), who studied the effect of zonal conditions on vascular recruitment in subpleural vs. interior regions by counting the number of red blood cells per length of alveolar wall in rapidly frozen rat lungs after exposing them to alveolar pressures of 25, 12, or 4 cmH₂O (approximating zones 1, 2, and 3, respectively). They found that the number of red blood cells in the vessels located immediately beneath the pleural surface averaged ~55% of that seen in vessels located in a deeper microscopic plane in the same 15-µm-thick sections, regardless of zonal condition. The slightly greater fraction of subpleural flow that we observed can be explained by the facts that <1-mm-thick sections of "subpleural" tissue that we obtained by macrodissection included many vessels that would have been classified as interior vessels in the study by Short and colleagues and that these vessels receive a greater fraction of flow. Our results would overestimate the percentage of subpleural flow to the extent to which these interior vessels were included. Our results extend the observations of Short and colleagues by finding that the relative flow distributions measured in vitro (allowing a precise determination of and variation in the zonal conditions) are virtually identical with those occurring under the zonal conditions existing in vivo (Table 1). We were also able to demonstrate the effect of changing zonal conditions on the relative flow distribution within the same lung, and these observations confirm the comparisons made by Short and colleagues of the effect of different zonal conditions in individual lungs. Thus two independent methods confirm that flow through subpleural vessels is less than that occurring in vessels located deeper in the interior in all zonal conditions, and the effect of zonal conditions on flow through both subpleural and interior vessels is remarkably similar. Accordingly, the zone 1 flow that we previously observed using pleural windows in dogs (14) does not seem to be an artifact resulting from anomalous pleural vessels, but, rather, those observations underestimate the extent of zone 1 blood flow that occurs throughout the rest of the lung as a result of there being a less dense capillary network in this region. Models describing pulmonary blood flow in the lung must, therefore, account for the occurrence of flow through alveolar capillaries, despite the presence of zone 1 conditions.

Dependent-to-nondependent flow distribution. In excised lungs, pleural pressure is uniform and alveolar distension should, accordingly, be relatively uniform (with the exception of the effects of lung weight, which should increase the distension of nondependent regions for any given pleural pressure short of that producing maximum lung distension). Vascular tone was eliminated in our experiments by use of papaverine. Accordingly, regional perfusion differences should be determined by the intravascular hydrostatic pressure gradient and the vascular anatomy. We found a strong dependent-to-nondependent gradient for both subpleural and interior blood flow when the relationship was calculated using mean flows for isogravitional sections (Table 2, Fig. 2), and 80% or more of the flow variation (R²) seemed to be explained by the vertical gradient. The gradients were similar for both zonal conditions studied, regardless of whether the lungs were held upright or inverted. In the three lungs injected with microspheres in vivo, however, a dependent-to-nondependent gradient was observed when the lungs were positioned upright, but the slope was less and the vertical gradient only explained 65-75% of the variation (Table 2, Fig. 2). In contrast, the three lungs studied in the inverted position showed an antigravitational distribution of perfusion in both the subpleural and interior regions (Table 2, Fig. 2). We presume this difference resulted from compression of cephalad lung regions when rabbits are held in the inverted position.

Stop-flow pressures. We estimated that flow in both the subpleural and interior regions would occur 7-8 cm above the most dependent region when the lung was perfused under zone 1 conditions (Table 2, x-intercept). Because the arterial pressure at the bottom of these lungs was 22.5 cmH₂O, the arterial pressure at the point of vessel closure would be ~15 cmH₂O, or 10 cmH₂O below the alveolar pressure of 25 cmH₂O. This zone 1 stop-flow pressure is very similar to what we previously observed when examining alveolar corner vessels with the pleural window (14) and in other studies of rabbit lungs done in situ (12, 13). Recent morphometric studies of capillaries in the three zonal conditions found that alveolar septal vessels were closed in zone 1, whereas alveolar corner were still dilated (4). Because alveolar corner vessels are still patent at the stop-flow pressure, we assume that closure is occurring further downstream, presumably in small pulmonary venules.