Journal of Applied Physiology
Vol. 83, No. 2, pp. 543-549, August 1997
PULMONARY CIRCULATION AND LUNG FLUID BALANCE

Capillary recruitment and transit time in the rat lung

Robert G. Presson Jr., Thomas M. Todoran, Bracken J. De Witt, Ivan F. McMurtry, and Wiltz W. Wagner Jr.

Departments of Anesthesia, Physiology/Biophysics, and Pediatrics, Indiana University School of Medicine, Indianapolis, Indiana 46202; and Cardiovascular Pulmonary Laboratory, University of Colorado Health Sciences Center, Denver, Colorado 80262

ABSTRACT

INTRODUCTION

METHODS

RESULTS

DISCUSSION

ACKNOWLEDGEMENTS

FOOTNOTES

REFERENCES

ABSTRACT

Presson, Robert G., Jr., Thomas M. Todoran, Bracken J. De Witt, Ivan F. McMurtry, and Wiltz W. Wagner, Jr. Capillary recruitment and transit time in the rat lung. J. Appl. Physiol. 83(2): 543-549, 1997.Increasing pulmonary blood flow and the associated rise in capillary perfusion pressure cause capillary recruitment. The resulting increase in capillary volume limits the decrease in capillary transit time. We hypothesize that small species with relatively high resting metabolic rates are more likely to utilize a larger fraction of gas-exchange reserve at rest. Without reserve, we anticipate that capillary transit time will decrease rapidly as pulmonary blood flow rises. To test this hypothesis, we measured capillary recruitment and transit time in isolated rat lungs. As flow increased, transit time decreased, and capillaries were recruited. The decrease in transit time was limited by an increase in the homogeneity of the transit time distribution and an increased capillary volume due, in part, to recruitment. The recruitable capillaries, however, were nearly completely perfused at flow rates and pressures that were less than basal for the intact animal. This suggests that a limited reserve of recruitable capillaries in the lungs of species with high resting metabolic rates may contribute to their inability to raise O₂ consumption manyfold above basal values.

pulmonary microcirculation; indicator dilution; isolated rat lungs; video microscopy; digital image analysis

INTRODUCTION

DURING EXERCISE, an increase in pulmonary blood flow causes pulmonary capillary transit time to fall (4, 12, 16, 17, 26). If transit time becomes too short (<0.25 s), red blood cells leave the pulmonary capillaries incompletely saturated with O₂ (23). Increasing flow, however, elevates capillary transmural pressure, which recruits capillaries that were not perfused at rest and distends capillaries that were already perfused. The resulting increase in capillary blood volume has the important effect of reducing the rate of fall in capillary transit time (16, 17). Whereas a substantial recruitable capillary reserve has been demonstrated in larger species such as the dog (8, 16, 17, 25), it is uncertain whether a similar reserve exists in species such as the rat, which have a relatively higher resting metabolic rate and are thus more likely to utilize a larger fraction of gas-exchange reserve at rest. In the absence of a recruitable reserve, we hypothesize that capillary transit time will decrease rapidly as flow increases. To determine the extent of the recruitable reserve in the rat and to investigate how mobilization of this reserve affects the distribution of capillary transit times, we measured capillary recruitment and the capillary transit time distribution at baseline, during increased flow, and during increased venous pressure in isolated, pump-perfused rat lungs.

METHODS

Experimental preparation. Sprague-Dawley rats (n = 30) were anesthetized by intramuscular injection (0.06 ml/100 g body wt) of a mixture of ketamine (90 mg/ml) and xylazine (10 mg/ml), supplemented with additional doses (0.03 ml/100 g body wt) as needed to maintain surgical anesthesia. For each experiment, the perfusion circuit was primed with heparinized blood withdrawn from two donor animals by cardiac puncture. The trachea of a third animal (360-580 g) was cannulated with a 14-gauge needle stub via a tracheotomy, and the lungs were ventilated with a 6% CO₂-17% O₂-77% N₂ gas mixture and a tidal volume of 10 ml/kg at a rate of 60 breaths/min by using a Harvard rodent ventilator (model 683). A median sternotomy was performed, and the animal was given heparin (1,000 U/kg) by cardiac puncture. Shortly thereafter, the animal was exsanguinated by cardiac puncture, and the blood was added to the perfusion circuit. The main pulmonary artery was cannulated via the right ventricular outflow tract with a stainless steel cannula secured with a ligature around the aorta and main pulmonary artery. The left atrium was cannulated via the left ventricle with a plastic, multiorifice-tipped catheter, and perfusion was initiated at a flow rate of ~10 ml · kg¹ · min¹. The time from exsanguination to the initiation of perfusion was <5 min. Blood was pumped (Gilson Minipuls 3) into the pulmonary artery and drained passively from the left atrium into a water-jacketed reservoir (Fig. 1). The temperature of the reservoir was maintained at 37-38°C by a circulating water bath. The height of the reservoir could be altered to change venous pressure. A windkessel was located just distal to the pump to dampen vibrations and trap bubbles. Pulmonary arterial and venous pressures (Ppa and Ppv, respectively) were measured continuously with two transducers (Statham P23 XL) zeroed at the site of observation and connected with PE 200 tubing to the perfusion circuit ~4 cm from the ends of the arterial and venous cannulas. Airway pressure (Paw) was measured intermittently (Statham P23 XL transducer). The output from the pressure transducers was amplified (Gould model 13-G4615-52), processed by a 100-Hz low-pass analog filter, then sampled at 100 Hz by an analog-to-digital board (Computer Boards model C10-DA508-PGA) in a microcomputer (Dell 50-MHz 486). Blood gases were sampled from the pulmonary arterial line periodically and analyzed with an Instrumentation Laboratories model 1304. Sodium bicarbonate solution (1 meq/ml) was added to the venous reservoir as needed to neutralize metabolic acid.

Video microscopy. The animal was placed on a microscope stage in the right lateral decubitus position, and the left chest wall was excised to provide access to the left lung. End-expiratory pressure was maintained at 3 mmHg with a water overflow on the expiratory limb of the ventilator. The animal was raised until the pleural surface of the left lung came into contact with a transparent window. A 0.5 cm² area on the surface of the lung was observed through the window. The window pane was surrounded by a vacuum ring to prevent lateral movement of the observed area. The remainder of the pleural surface was covered with a thin sheet of plastic to prevent drying and to slow the transpleural diffusion of gases. The subpleural microcirculation under the window was observed with a modified Olympus BH2 reflectance microscope coupled to a Leitz Ultropak illuminator with a ×11 objective. Bright field illumination through the Ultropak illuminator was obtained with a 200-W mercury arc lamp mounted on an optical bench. This light source was heavily filtered to prevent tissue damage with a combination of dichroic infrared- reflecting filters, broad band-pass ultraviolet-absorbing filters, and a narrow band-pass interference filter to illuminate the field with the mercury green line (546 nm). Illumination for fluorescence microscopy was provided by a 100-W mercury arc mounted on the side arm of the BH2 microscope, which was also filtered by dichroic infrared-reflecting filters and ultraviolet-absorbing filters. Light from the 100-W arc passed through a blue band-pass exciter filter (410-480 nm) and a high-pass dichroic mirror (cutoff wavelength = 480 nm), which reflected the exciting light down through the objective onto the subpleural microcirculation beneath the window. Emitted light passed back through the objective, the dichroic mirror, and a yellow high-pass barrier filter (cutoff wavelength = 510 nm). Video recordings of the subpleural microcirculation were made with a Sony SVHS videorecorder (model SVO 5800) and a Cohu intensified charge-coupled device camera (model 5510), which was mounted on the microscope with a Nikon zoom CCTV adapter (model 79444).

Indicator dilution curves. In five animals, a microscopic field was selected that contained a single arteriole and a single venule of equal or smaller diameter. While observing this field with fluorescence microscopy, test injections of dye [0.1 ml of a solution of fluorescein isothiocyanate conjugated to dextran, mol mass 70 kDa (Sigma Chemical), 2 mg/ml 0.9% saline] were made with a pneumatically powered injector via a multiple-side hole cannula inserted into the perfusion circuit just proximal to the arterial cannula. The passage of dye through the subpleural microcirculation was videotaped, and elapsed time in milliseconds was recorded on the videotape by a Panasonic WJ-810 time-date generator that was triggered by the same switch that activated the injector.

Videotapes of these injections were replayed to observe the movement of dye from the arteriole across the capillaries into the venule. By requiring that 1) the dye proceed completely across the capillaries in the field before appearing in the venule and that 2) the dye disappear promptly from the venule once it had drained from the capillaries, we were reasonably sure that the selected arteriole and venule were effectively functioning as the sole inlet and outlet of the capillary bed being studied or, if other capillary beds contributed dye, that they were in phase with the observed capillaries (4, 16, 17). Deviation from this pattern caused us to seek another vessel pair. If no appropriate pair could be found, the animal was not used in the study.

The black level and gain of the camera and intensifier were adjusted according to these test injections to maximize the contrast between the baseline brightness of the microscopic field before the dye entered the circulation and the peak brightness during its passage through the microcirculation. The response of the video system was linear over the range of light intensities measured. Thus emitted light was proportional to the quantity of indicator injected.

After the test injections, a set of six injections of dye were made under each of the four conditions: 1) at a baseline flow rate (25 ± 4 ml · kg¹ · min¹) selected to produce a low-to-moderate flow rate by visual inspection through both the arteriole and venule, a rate similar to that previously reported for this preparation (13, 19); 2) at the maximum flow rate obtainable from our pump (69 ± 7 ml · kg¹ · min¹); 3) at baseline pump flow rate but with increased Ppv; and 4) again at baseline flow rate and baseline Ppv (same as condition 1). The order of maximum pump flow rate and increased venous pressure (conditions 2 and 3) was varied among the experiments. The two sets of baseline measurements (conditions 1 and 4) were averaged for subsequent comparison to high flow and high Ppv measurements. The average difference between the mean transit times of the two baseline sets of measurements (conditions 1 and 4) was 12 ± 2%. Ppa and Ppv values were recorded during each injection, and arterial blood gases were measured before and after each set of injections.

Indicator dilution curves were obtained by replaying the recordings of the injections and sampling image brightness at 30 Hz from rectangular areas over the arteriolar and venular lumens (vessel windows) and from areas over the adjacent alveoli (subtractor windows) by using a frame grabber board (Data Translation DT-2851) in a microcomputer (Dell 50 MHz 486). The sampling windows were movable and of adjustable size. By using methods described in detail elsewhere (17), the subtractor window signals were used to correct the vessel window signals for light emitted from the surrounding capillaries. An average arteriolar and venular curve was obtained for each set of experimental conditions by averaging curves from the set of six injections made under those conditions. The baseline segment of each of these average curves before indicator entered the vessel was set to zero, and the tail of each curve (~5% of the area under the curve) was extrapolated to baseline as a monoexponential function. Finally, the area under each curve was set to unity. These curves served as the input and output functions for the determination of the distribution of capillary transit times by deconvolution using damped least squares (3, 6). Mean capillary transit time (), was calculated from the transport functions (the area of which was also unity) as (16, 17)

(1)

where t₁ was the time when the intensity of the transport function curve (I) became greater than zero, and t_n was the time when the curve returned to baseline. The variance of the transit time (²), was calculated as (16, 17)

(2)

The relative dispersion (RD) was computed as /. The minimum transit time (t_min) for each transport function was calculated as the time when 1% of the indicator had crossed the capillary network, and the maximum transit time (t_max) was the time when 99% of the indicator had crossed.

Measurement of capillary recruitment. To determine how changes in capillary transit time were affected by capillary recruitment, we recorded the perfusion pattern of two to five alveoli for 1 min in the field where the transit time measurements were made (n = 5 animals) with the Ultropak illuminator under each set of experimental conditions. Recruitment measurements were made before the transit time measurements in some animals and after these measurements in others. To better define recruitment as a function of microvascular pressure in the rat, additional recruitment measurements were made over a range of perfusion pressures in five additional animals.

The videotapes from each 1-min recording were replayed, and the perfused capillary segments were traced onto separate sheets of plastic transparency film placed over the video monitor. A capillary segment was considered to be perfused if one or more red blood cells passed through the segment during the 1-min video recording. We measured the total length of the perfused capillaries from the tracings by using a digitizing pad (Houston Instruments Truegrid 1017), planimetry software (SigmaScan, Jandel Scientific), and a microcomputer (Gateway 66 MHz 486). The area of the observed alveolar walls was also measured by using the same system. The subpleural alveolar facets (n = 40) that we observed on the superior surface of the rat lung inflated to 3 mmHg had an average diameter of ~80 µm and an area of 5,000 µm². Therefore, we divided the wall area of each alveolus by 5,000 µm² to obtain the number of average-size alveolar walls in the observed alveolar facet. This normalization permitted us to compare results between individual alveoli and between animals. Division of the total length of perfused capillaries by the normalized alveolar area indicated how many times perfused capillaries crossed an average alveolar wall at its diameter. Defined mathematically, the capillary perfusion index (CPI) for the isolated rat lung is

(3)

The level of capillary recruitment can be readily estimated from the CPI. For example, a CPI of 80 µm can be visualized as a capillary path length that would cross the 80-µm diameter of an average alveolar facet once, whereas a CPI of 320 µm would mean that an average alveolar wall could be crossed four times at its diameter.

Statistical analysis. We used analysis of variance for randomized blocks to test , ², RD, t_min, t_max, CPI, blood gases, and hemodynamic variables for differences among experimental groups, followed by Tukey's honestly significant difference test for pairwise comparisons. When indicated by Bartlett's test, log transforms were performed prior to analysis of variance to obtain equal variances among the groups (22). The level of significance was P < 0.05 for all statistical tests. All observations are reported as means ± SE.

RESULTS

When flow was increased, and t_max decreased significantly, but t_min did not change (Fig. 2, Table 1). Conversely, when venous pressure was increased at baseline flow, and t_min increased significantly, but t_max did not change (Fig. 2, Table 1). In both cases, the distribution of transit times became more homogeneous (Figs. 3 and 4), as shown by a decrease in the RD (Table 1).

Table  1.   Transit time variables Variable Control Increased Ppv Increased Mean transit time, s 3.9 ± 0.2  6.7 ± 1.0* 2.4 ± 0.1* Minimum transit time, s 1.3 ± 0.1  2.7 ± 0.4* 1.1 ± 0.1  Maximum transit time, s 12.3 ± 0.7  18.6 ± 2.8  6.5 ± 0.3* Variance, s2  5.5 ± 0.7  12.1 ± 3.7  1.3 ± 0.1* Relative dispersion 0.57 ± 0.02  0.48 ± 0.01* 0.46 ± 0.02* CPI, µm 265 ± 35  372 ± 48* 391 ± 39* Values are means ± SE (n = 5 animals). Ppv, venous pressure; , flow; CPI, capillary perfusion index. * Significantly different from control (P < 0.05 from analysis of variance for randomized blocks, followed by Tukey's honestly significant difference test for pairwise comparisons).

Capillaries were recruited both when flow was increased and when Ppv was increased at baseline flow (Table 1). Capillary recruitment increased from low levels, when the estimated microvascular pressure was close to Paw, to an apparent plateau at ~9 mmHg (Fig. 5). We estimated microvascular pressure as the average of Ppa and Ppv when Ppv was greater than Paw. However, when Ppv was less than Paw, we used the average of Ppa and Paw.

Ppa increased significantly when flow was increased, but not when Ppv was increased without a change in flow rate (Table 2). Ppv was significantly higher than baseline when the venous reservoir was raised without changing the pump flow rate and also when flow was increased as a result of resistance in the venous cannula (Table 2). In both cases (high flow or high Ppv with baseline flow), Ppv exceeded Paw moving the lungs from zone 2 (Ppa > Paw > Ppv) into zone 3 (Ppa > Ppv > Paw). There were no significant differences in blood gases among the experimental groups (Table 2).

Table  2.   Cardiorespiratory variables Variable Control Increased Ppv Increased Pump flow rate, ml · min1 · kg1 25 ± 4  25 ± 4  69 ± 7  Ppa, mmHg 10.4 ± 0.8  13.1 ± 0.7  16.0 ± 1.5* Ppv, mmHg 0.7 ± 0.2  6.1 ± 0.4* 5.4 ± 1.4* PaO2, Torr 126 ± 8  129 ± 8  126 ± 7  PaCO2, Torr 36 ± 1  34 ± 1  35 ± 1  pH 7.40 ± 0.01  7.42 ± 0.01  7.41 ± 0.01 Values are means ± SE (n = 5). Ppa, pulmonary arterial pressure; PaO2 and PaCO2, arterial PO2 and PCO2, respectively. * Significantly different from control (P < 0.05 from analysis of variance for randomized blocks, followed by Tukey honestly significant difference test for pairwise comparisons).

DISCUSSION

This study demonstrates two forms of gas-exchange reserve in the rat lung, reserve capillary length and unperfused capillaries. Reserve capillary length results from the fact that red blood cells become saturated with O₂ in ~0.25 s (23). Under resting conditions, however, red blood cells have traversed only a fraction of the capillary network in 0.25 s, leaving the distal portion of the capillaries unused for gas exchange. In the present study, mean transit time fell from 3.9 to 2.4 s as flow increased. Thus on average, red blood cells traveled farther across the capillaries before saturation was complete, thereby utilizing reserve capillary length. The second type of reserve took the form of capillaries, which, unperfused at rest, became perfused with increasing pressure and flow. These newly recruited capillaries added directly to the gas-exchange surface area. We observed a 50% rise in the level of capillary recruitment, as shown by the CPI, over the range of pressures and flows studied.

Recruitment of capillaries not only increased gas-exchange surface area but also increased the volume of blood contained in the capillaries. The increased capillary blood volume in turn prolonged capillary transit time thus helping prevent transit times from becoming too short for complete saturation of red blood cells with O₂. Capillary volume can be calculated from the relationship

(4)

From this relationship, the relative change in capillary volume can be calculated

(5)

In the present study, we measured transit time with a plasma marker. When flow was increased from baseline to maximum, capillary plasma volume increased by 70% of control

(6)

We also measured a 50% increase in the CPI with this flow change. We assume that the 50% increase in capillaries newly perfused by red blood cells as measured by the CPI was associated with a similar increase in capillary blood volume. Because capillary blood volume is a subset of capillary plasma volume, we conclude that, over the range of pressure and flow studied, recruitment accounted for a large fraction of the 70% increase in capillary volume as measured by a plasma marker. The remainder of the increase (the difference between 70 and 50%) could have resulted from distension of already perfused capillaries or increased perfusion of capillaries with plasma alone.

Another mechanism limiting the decrease in capillary transit time that occurred with increasing flow was a more homogeneous distribution of transit times (RD decreased from 0.57 to 0.46; Table 1). Unlike the increase in capillary volume, which limited the reduction of all transit times in the distribution equally, the narrowing of the transit time distribution preferentially spared the fastest transit times, i.e., the mean and maximum transit times both decreased, but the minimum transit time did not decrease (Table 1). Thus the fastest transit times, the ones most likely to become too short, were spared when the transit time distribution narrowed.

The transit time distribution also became more homogeneous when Ppv alone was increased (RD fell from 0.57 to 0.48). Because the transit time distribution became more homogeneous whether flow or Ppv was raised, we think that the responsible mechanism was an increased microvascular pressure that distended the capillaries more uniformly. This concept is consistent with observations of capillary flow in the microscopic field. At baseline, flow was intermittent through some capillary segments, whereas flow in others was steady. We considered the intermittently perfused segments to be partially closed and thus have higher resistance and slower transit times than the continuously perfused segments. As microvascular pressure increased, whether by increasing the flow rate or by raising the venous reservoir, flow through intermittently perfused segments became continuous, indicating they had opened more completely. This would cause the resistance and transit times of these segments to become more like those of the continuously perfused segments. We also observed that newly recruited segments tended to be perfused continuously. The alteration toward homogeneous perfusion with increasing microvascular pressure should decrease the RD of transit times, a result that occurred. From these observations we conclude that the increased homogeneity of the transit time distribution resulted from increased microvascular pressure and not increased flow rate per se.

Whereas these results demonstrate a recruitable reserve in the rat, our maximum flow rate, 69 ± 7 ml · kg¹ · min¹, was a fraction of the baseline cardiac output reported for the intact rat, ~250 ml · kg¹ · min¹ (20). Similarly, the average Ppa during increased flow was 16 mmHg and during increased Ppv was 13 mmHg, both below the resting average of ~20 mmHg for the intact rat (20). Although the highest pressure and flow utilized were below the normal baseline for the intact rat, we found a plateau in capillary recruitment at a microvascular pressure of ~9 mmHg (Fig. 5), suggesting recruitment was approaching completion. Thus reserve in the form of recruitable capillaries was utilized at relatively low flows and pressures. This is consistent with the idea that smaller species, such as the rat, with relatively high resting metabolic rates utilize a larger fraction of their total gas-exchange surface area reserve at baseline than do larger species. For example, rats increase their O₂ consumption 3-fold from rest to maximal exercise (2, 9), whereas dogs increase their O₂ consumption 25-fold (11, 15, 24). Although the maximum O₂ consumption of the rat (70 ml · kg¹ · min¹) is comparable to that of a well-trained human athlete (9), the rat is closer to maximum O₂ consumption at rest than other species and thus has less reserve.

The transit times we reported here are plasma transit times, which, because of the Fahraeus effect, are longer than red blood cell transit times (1) and overestimate the amount of time available for uptake of O₂ by red blood cells crossing the pulmonary capillaries. In previous work, we found that plasma transit times for subpleural capillaries in the dog lung were ~1.4 times longer than red blood cell transit times (16). Because the ratio of red cell velocity to plasma velocity through capillary-size tubes is constant over a wide range of velocities (1) and because the diameter of subpleural capillaries in the rat (6.6 ± 1.6 µm) is similar to the diameter of subpleural capillaries in the dog (7.7 ± 1.7 µm) (21), the plasma transit times we measured under the various conditions in the present study can be converted to their approximate red cell transit time equivalent by dividing by 1.4.

The transit times we measured were not exclusively capillary in origin because they were computed from indicator dilution curves measured over arterioles and venules. To estimate the part of the transit time measurements that was due to the time the indicator spent outside the capillary bed, we measured the length of time it took for the leading edge of the indicator bolus to pass from the arteriolar measurement window to the end of the arteriole and to pass from the beginning of the venule to the venular measurement window. At a baseline flow rate of 25 ± 4 ml · kg¹ · min¹, it took the indicator ~0.27 s to travel this distance in the arteriole and another ~0.27 s in the venule, a total of ~0.54 s spent outside the capillaries. This estimate is a maximum because capillaries come off these vessels at right angles, such that only capillaries fed or drained at the very end of the arteriole or venule (i.e., farthest from the measurement window) would contain an error at this upper bound. We estimate the average time spent outside the capillaries to be one-half this value or ~0.27 s, which is 7% of the mean transit time at baseline flow (Table 1). At the maximum flow rate of 69 ± 7 ml · kg¹ · min¹, the arteriolar phase of the curve was ~0.07 s, and the venular phase was also ~0.07 s. This gives a maximum error of ~0.14 s and an average error of 0.07 s, which is 3% of the mean transit time measurement at a flow rate of 69 ml · kg¹ · min¹ (Table 1). When Ppv was raised at baseline flow, the arteriolar phase of the curve was ~0.29 s, and the venular phase was also ~0.29 s. This gives a maximum error of ~0.58 s and an average error of ~0.29 s, which is 4% of the mean transit time measurement made after increasing Ppv (Table 1). We conclude that the time spent in the feeding arteriole and draining venule was small compared with the time spent in the capillaries, which means that our measurements pertain essentially to the capillaries alone.

Although we observed a plateau in the plot of capillary recruitment vs. microvascular pressure (Fig. 5), the shape of the recruitment curve is determined by the degree of alveolar distension (8), which, in turn, is a result of the inflating pressure used. In a previous study, Godbey et al. (8) found that capillary recruitment in the dog lung rose rapidly to a plateau at low microvascular pressures when Paw was low and alveolar size was small. As Paw increased and the alveoli distended, recruitment became linearly related to microvascular pressure. These results explained the pattern of capillary recruitment in species such as the dog with lungs that were large enough to have a vertical gradient of alveolar size in which the largest and, therefore, most distended alveoli were at the top of the lung and the least distended alveoli were located at the bottom, as demonstrated by Glazier et al. (7). The results of the Godbey et al. (8) experiment show that the recruitment pattern of highly distended alveoli at the top of the lung is linear, but at the bottom of the lung, where the alveoli are less distended, capillaries recruit suddenly and completely over a narrow range of microvascular pressure. In contrast, significant vertical differences in alveolar size are unlikely in the rat lung. Therefore, the average alveolar diameter of ~60 µm for the intact adult rat (18) is likely to be representative of the rat lung as a whole. Our recruitment measurements were made with the lungs inflated to an Paw of 3 mmHg, which resulted in an average alveolar diameter of ~80 µm. Because this is larger than the average alveolar diameter of the intact rat and because vertical differences in alveolar size are unlikely, the alveoli in our preparation were probably more distended than those in the intact rat. Our recruitment measurements, therefore, were likely to be more linear than the average alveolus in the intact rat. Nevertheless, we still observed a plateau in the recruitment curve, indicating that recruitment of capillaries was nearing completion at blood flow rates and microvascular pressures that were less than basal.

Although subpleural capillary networks are less dense than interior networks (10, 14) and the diameters of subpleural capillaries are somewhat larger than those of interior capillaries (21), recent work has shown interior and subpleural recruitment patterns to be similar. Short et al. (21) found that the decreased density of subpleural capillaries in the rat lung resulted in recruitment measurements that were about one-half the magnitude of those for interior capillaries. However, changes in recruitment of subpleural capillaries paralleled changes in recruitment of interior capillaries over a range of pressures that spanned low zone 3 to high zone 1. Therefore, the absolute values of the recruitment measurements we report are less than those for interior capillaries, but the shape of the recruitment vs. microvascular pressure curve should be the same. Although it is less clear how morphological differences between subpleural and interior capillaries affect transit time, a study in dogs showed that subpleural plasma capillary transit times measured by dye dilution were similar to red blood cell capillary transit times for the whole lung measured by carbon monoxide diffusing capacity (5).

In summary, we have reported the distribution of capillary transit times and capillary recruitment in the rat. Similar to previous studies in the dog, increasing flow caused capillary recruitment and a decreased capillary transit time. The decrease in transit time was limited by an increased capillary volume due to recruitment and an increase in the homogeneity of the transit time distribution. Unlike in the dog, we found that the recruitable reserve in the rat was utilized at flow rates and pressures that were less than basal for the intact animal. This suggests that a limited reserve in species with high resting metabolic rates may contribute to their inability to raise O₂ consumption manyfold above basal values.

ACKNOWLEDGEMENTS

The authors thank Dr. Solbert Permutt for many insightful discussions regarding this research. A. J. Peterson, W. A. Baumgartner, Jr., and T. M. Wagner provided helpful criticism of the manuscript.

FOOTNOTES

Address for reprint requests: R. G. Presson, Jr., Riley Children's Hospital, 702 Barnhill Dr., Rm. 2001, Indianapolis, IN 46202-5200.

Received 4 October 1996; accepted in final form 18 April 1997.

REFERENCES