Pulmonary embolism: emboli and fibrinolysis inhibition in isolated canine lungs MARY
I. TOWNSLEY,
Department
of Physiology,
SCOTT
A. BARMAN,
College of Medicine,
TOWNSLEY, MARY I., SCOTTA. BARMAN, AND AUBREY E. TAYLOR. Pulmonary embolism: emboli and fibrinolysis inhibition in isolated canine lungs. Am. J. Physiol. 258 (Heart Circ. Physiol. 27): H754-H758, 1990.-The effect of fibrinolysis inhibition with tranexamic acid on pulmonary microvascular permeability during glass bead embolization was investigated in the isolated lung. Lung lobes from nonheparinized dogs were treated in vivo with the equivalent of 0.6 g/kg 100 pm glass bead emboli alone, emboli after tranexamic acid, tranexamic acid alone, or the bead vehicle alone. After 40-50 min, the lobes were isolated for ex vivo perfusion with heparinized autologous blood. There were no changes in any parameter over the 120min perfusion period. Blood flow at 120 min was decreased after both emboli alone and emboli with tranexamic acid, reflecting an increase in vascular resistance compared with the Tween or tranexamic acid controls. Furthermore, tranexamic acid increased the ratio of pre- to postcapillary resistance in embolized lobes compared with that after emboli alone or in the Tween or tranexamic acid controls. The isogravimetric capillary pressure and the osmotic reflection coefficient were not significantly decreased by tranexamic acid compared with those after emboli alone; however, it did result in an increase in the capillary filtration coefficient compared with that after emboli alone or in the control groups. We conclude that although fibrinolysis inhibition does not clearly exacerbate the lung injury seen after emboli, the tranexamic acid-induced changes in hemodynamics would tend to accelerate edema formation.
capillary filtration coefficient; osmotic reflection coefficient; osmotic transient; isogravimetric capillary pressure; tranexamic acid
THEROLEOFHUMORALFACTORS intheedemaformation resulting from embolization of the lung has long been debated. Malik and van der Zee (7) determined that heparin pretreatment before glass bead embolization of the canine lung prevented the edema formation seen when coagulation and fibrinolysis were allowed to proceed normally. Subsequently, these investigators showed that inhibition of fibrinolysis delayed the resolution of emboli-induced edema formation in the canine lung (6). In contrast, after complete inhibition of fibrinolysis in the sheep lung the increase in the lymphatic protein clearance normally seen as a result of thrombin embolization was attenuated. This finding suggests that some slow, continual generation of fibrin degradation products is responsible for the emboli-induced changes in pulmonary microvascular protein and fluid exchange (4). Based on these findings, Malik et al. and others (6, 7, 14) have proposed that fibrinolysis with the resultant production H754
0363-6135/90 $1.50 Copyright
0
AND
University
AUBREY
E. TAYLOR
of South Alabama,
Mobile, Alabama 36688
of fibrin degradation products is a necessary condition for the development of emboli-induced microvascular injury. However, whether fibrinolysis results in lung injury would still seem to be an open question, since edema formation and changes in lymphatic protein clearance may develop from either a change in the transcapillary Starling forces (hemodynamic edema) or from a direct microvascular injury. It is pertinent in this regard that parameters that specifically describe the permeability of the lung microvasculature, such as the capillary filtration coefficient (&,,) or the osmotic reflection coefficient (ad), were not measured in any of these studies. Thus the role of fibrinolysis is difficult to determine. We have recently shown that glass bead embolization (0.2-0.6 g/kg) in the absence of a fibrinolysis inhibitor results in an increase in pulmonary capillary pressure and microvascular permeability in the intact canine lung lymphatic model (19). ad was significantly reduced to 0.55 overall compared with that in normal lungs where fld = 0.62 (10). Although this study demonstrated that inhibition of fibrinolysis was not necessary for injury to occur, no conclusions could be made regarding the role of fibrinolysis inhibition in modulating the pulmonary injury following emboli. As we encountered difficulties in maintaining dogs with lung lymph catheters throughout the entire protocol with 0.6 g/kg dose of beads (19), we chose to examine the problem of fibrinolysis inhibition in the isolated canine lung lobe. Glass beads, at a dose equivalent to 0.6 g/kg, were administered to the canine lower left lobe in situ in the presence and absence of a fibrinolysis inhibitor, tranexamic acid. The dose of tranexamic acid administered in vivo was exactly that used previously (6). Lobes were subsequently isolated for ex vivo perfusion to measure several permeability and hemodynamic parameters, including &C and ad. METHODS
Twenty-four mongrel dogs (20.7 t 1.0 kg, means t SE) were anesthetized with 30 mg/kg pentobarbital sodium iv, intubated, and ventilated with room air (15 ml/ kg at 12 breaths/min). The left carotid artery and jugular vein were cannulated with polyethylene catheters for blood withdrawal or drug infusion, respectively. A surgical plane of anesthesia was maintained with supplemental intravenous boluses of pentobarbital. The left lung was exposed via an incision along the fifth intercostal space. The branch of the left pulmonary artery to the
1990 the American
Physiological
Society
Downloaded from www.physiology.org/journal/ajpheart at Macquarie Univ (137.111.162.020) on February 15, 2019.
EMBOLI
AND
FIBRINOLYSIS
INHIBITION
left upper lobe was cannulated in a retrograde manner with a small polyethylene catheter. The catheter was fashioned with a right angle 1 cm from the tip of the catheter and inserted such that this l-cm tip was parallel (orthograde) to the direction of flow in the left pulmonary artery. The branch feeding the left middle lobe was either clamped or ligated. Four groups of lungs were studied. One group received the equivalent of 0.6 g/kg 100 pm glass beads (range 77125 pm, Ferro) via the upper lobar catheter (n = 8). Beads were prepared in 5 ml of saline with 1 drop Tween 80 to prevent clumping. Since lower left lobe weight averages 25% of the entire lung weight (18), 25% of the calculated 0.6 g/kg dose was infused into the lower left lobe. In six lobes of this group, siliconized glass beads were used, and in the remaining two lobes, noncoated beads were infused. As there were no differences observed in any parameter with respect to presence or absence of silicon, these lungs were treated as a single group. The second group of lobes (n = 6) received a 50 mg/kg iv bolus of tranexamic acid followed by its infusion (0.83 mg* kg-’ . min-‘) for the next 55 min (6). The equivalent of 0.6 g/kg beads was injected 15 min after initiating the tranexamic acid infusion. Control lobes received either tranexamic acid alone at the same dose (n = 5) or the saline-Tween vehicle alone (n = 5). Immediately after bead infusion, the left upper and middle lobes were ligated at the hilar margin and excised. Loose ligatures were placed around the left pulmonary artery and lobar bronchus; flow to the left lower lobe was maintained. Heparin (10,000 U iv) was administered 30 min after bead infusion. Ten minutes later, 600-800 ml of blood were removed via the carotid catheter to prime the perfusion system. The pulmonary artery was then ligated, and the left lower lobe with the attached left atria1 appendage was then rapidly excised and weighed. Plastic cannulas were secured in the lobar artery and vein and the lobar bronchus, and perfusion was begun within lo-20 min of excision (50-60 min after bead infusion). The isolated lung perfusion system has previously been described in detail (1,13,18). Briefly, the lower left lobes (53.7 t 3.8 g) were suspended from a counterbalanced force transducer and perfused with warmed (38°C) and oxygenated (95% 02-5% COs) autologous blood at constant perfusion pressures. Pulmonary arterial (Pa) and venous (PJ pressures were measured with pressure transducers via thin catheters positioned at the orifices of the inflow and outflow cannulas, respectively, and referenced to the top of the lung hilus. Blood flow (Q) was measured with a flow-through electromagnetic flow probe positioned in the venous outflow line. P,, P,, airway pressure, and lung weight were continuously recorded on a Beckman polygraph. The sensitivity of the weighing system was adjusted to provide 1 cm of pen deflection per gram weight change. Measurements. Lobes were inflated with compressed air to a constant airway pressure of 3 cmH20 and perfused under Zone III conditions. P, was set at 4-5 cmHg0 and P, adjusted to yield an isogravimetric state, i.e., the lobe was neither gaining nor losing weight. Isogravimetric capillary pressure (Pc,i) was measured via the double
IN
ISOLATED
H755
LUNGS
vascular occlusion technique (17). If the lobe was not exactly isogravimetric, this technique measured the prevailing pulmonary capillary pressure (PC). Under these conditions, P,,i could be estimated (13) by relating the ongoing rate of weight gain (AW/At) to the capillary filtration coefficient (Kf,,) Estimated
AWlAt
P,i, = P, - 7
f,c
P,i can be used as a qualitative index of permeability when plasma protein concentration remains stable, since P,i is related to the osmotic reflection coefficient for total proteins (Q): P,i = Pt + c&,-?rJ, where Pt is tissue hydrostatic pressure, and ?rp and 7rt are the plasma and tissue colloid osmotic pressures, respectively. Total vascular resistance (RT) was calculated as (P, - PV)/&. P, was used to partition RT into pre- and postcapillary resistance segments; the ratio (RJR,) is reported here. &, may also be used as an index of microvascular permeability in the Zone III isolated lung (1, 12), assuming that exchange surface area remains constant (18). &, was calculated following a step increase in both P, and P, of approximately 5-8 cmHa0. This results in an initial rapid weight gain due to a shift of blood into the lungs followed by a slower rate of weight gain, which can be attributed to transcapillary filtration. P, was measured just before and at the end of the step increase in pressure. The rate of weight gain after the increase in pressure was plotted on semilogarithmic paper as a function of time, and the slow phase of the weight transient was extrapolated to time 0. This initial filtration rate at t= 0 was divided by the induced pressure increment (AP,) to yield Kf,,. Finally, &-Jwas measured with the osmotic transient technique (13). The extrapolated AW/At at time 0 following a step decrease in ?rpwas used to calculate &-J
(2) Plasma protein concentrations (C,) were measured by refractometry before and after dilution of the circulating blood with warmed saline. The refractometer was calibrated with pooled dog plasma. rp was calculated using the equation derived by Navar and Navar (8) for dog plasma proteins. Experimental protocol. Isogravimetric conditions were established fairly quickly after perfusion of the lung lobes, and initial measurements were begun within 30 min of perfusion. Hemodynamic parameters, Kf,, and Pc,i, were measured every 30 min for 120 min. When an isogravimetric state was reached after the fourth KfC measurement, C, and P, were measured and the blood rapidly diluted with warmed saline to produce the osmotic transient. Care was taken to adjust vascular pressures if necessary to keep them constant during this maneuver. Lung weight was monitored for lo-15 min after dilution. After the osmotic transient analysis, Kf,, measurements were repeated to ensure that saline dilution had induced no changes. Statistical analyses. All data are means 2 SE. The effects of perfusion time within each group as well as differences between treatment groups were statistically
Downloaded from www.physiology.org/journal/ajpheart at Macquarie Univ (137.111.162.020) on February 15, 2019.
H756
EMBOLI
AND
FIBRINOLYSIS
INHIBITION
IN
evaluated using an analysis of variance with post hoc tests to identify specific differences (15).
ISOLATED
LUNGS 0 m
30 min 120 mfn
RESULTS
There were no effects of perfusion time in any group for any parameter, thus only the data collected after 30 i . 0.2 and 120 min of perfusion are presented. The stability of .the isolated lobes with perfusion time implies that interE group differences are the result of changes that were T initiated in vivo. 0.1 Embolization resulted in a fall in the isogravimetric & in these experiments, as shown in Table 1. Q was markedly decreased and P, increased (P < 0.05) in the emboli and emboli plus tranexamic acid groups compared with Twoan TEA Em boli those in either of the control groups. RT tended to Controls + TEA increase as a result to 41.2t 11.1and 72.7t 31.4cmHzOS FIG. 1. Capillary filtration coefficients (&,) as a function of time 1-l *min. 100 g at 120 min after emboli and emboli plus after perfusion after emboli in absence or presence of fibrinolysis tranexamic acid, respectively. However, these changes in inhibitor, tranexamic acid (TEA). Control lobes were treated with saline-Tween vehicle alone (Tween) or tranexamic acid alone. * P < RT following emboli were only statistically significant 0.05 vs. Tween control group; + P < 0.05 vs. tranexamic acid control when fibrinolysis was inhibited. RJR” in the emboli group; and @P < 0.05 vs. response after emboli alone. group (3.27 & 0.50) was increased at 30 min compared with-that in the tranexamic acid or Tween control-lobes [ 30 min (1.27 & 0.36 and 1.24 & 0.13, respectively). After 120 min m 120 min of perfusion, this difference in R,/Rv was not significant. The only hemodynamic effect exacerbated as a result of fibrinolysis inhibition was the emboli-induced increase in RJR”. RJR” nearly doubled in the embolized lobes with tranexamic acid (6.08 * 1.24, P < 0.05 vs. emboli alone), reflecting the larger arterial or precapillary resistance in this group. .The effects of fibrinolysis inhibition on pulmonary 5 microvascular permeability after emboli are shown in on a, Figs. 1 and 2. After 120 min of perfusion, KfC in the emboli group (0.18 t 0.02 ml Smin-l cmH20D1 *‘lo0 g-‘) was not different from that in the Tween or tranexamic acid controls (0.17 & 0.01 and 0.16 t 0.01 mlmin-la Twoon TEA Emboli Em boll cmHzO-’ 100 g-l, respectively). After tranexamic acid, Controls +TEA however, embolization resulted in a significant increase FIG. 2. Isogravimetric capillary pressure (Pqi) as a function of time in & to 0.29 & 0.03 ml. min-’ •rnHzO-‘~ 100 g-l at 120 after perfusion. * P < 0.05 vs. Tween control group and + P < 0.05 vs. min (P < 0.05). In addition, after 120 min of perfusion, tranexamic acid (TEA) control group. l
TABLE 1. Pulmonary hemodynamics in absence or presence of fibrinolysis inhibition with tranexamic acid in isolated embolized canine lung lobes Group Time, min
R
30 120 30
P”
120
Q
30 120
Tween
Tranexamic
Tranexamic acid plus emboli
acid
Emboli
152~1.2 15.fkk1.3 4.3kO.3 4.2t0.3 0.87kO.07
16.7kO.7 17.3zk1.3 5.3t0.4* 4.8kO.4 0.87zkO.11
22.1,+1.7"~
24.5+1.8*t
21.4f1.6*? 4.2*O.lt
24.0+1.5*t 4.5+O.lt
4.0&0.2-f 0.54+0.07*-t
4.620.2 0.46+0.06*"f
0.91kO.07 12.4kl.l 12.7k1.5
0.85t0.13 14.MO.9 15.521.7
0.54&0.06*t 0.41+0.07*t 30 42.1A10.2 53.1-e15.9*t RT 120 41.2tll.l 72.7a31.4”t 30 1.24kO.13 1.27k0.36 3.27+0.50*7 5.90+1.07*t$ RafRv 120 1.21kO.05 1.17t0.27 2.88k0.45 6.08+1.24*?$ Data are means k SE at 30 and 120 min after initiation of in vitro perfusion. Pa, lobar arterial pressure (cmHzO); P,, lobar venous pressure (cmHnO); Q, blood flow (1. min-’ 100 g-‘); RT, total vascular resistance (cmHa0 1-l *min. 100 g); and RJR,, the ratio of pre- to postcapillary resistance. * P < 0.05 vs. Tween control; t P c 0.05 vs. tranexamic acid control; $ P < 0.05 vs. emboh. l
l
P,,i was significantly decreased (P c 0.05) in both the emboli (8.6 t 0.2 cmHsO) and the emboli plus tranexamic acid group (7.7 $- 1.1 cmHeO) as shown in Fig. 2. ad was 0.51 t 0.04 and 0.41 -c-0.03 after emboli alone or emboli plus tranexamic acid, respectively. Although both values are less (P < 0.05) than the ad measured in normal canine lobes (0.65 -$-0.02) when this osmotic transient technique (13) is used, only the cd measured after emboli plus tranexamic acid was significantly different from that in the tranexamic acid controls (0.63 t 0.02, n = 4). DISCUSSION
The results of this study support our earlier finding in intact dogs that, despite the high fibrinolytic activity in dogs, fibrinolysis inhibition is not necessarily a prerequisite for pulmonary VaSCdar injury after emboli. &J for total protein averaged 0.50 after 0.6 g/kg emboli in three intact dogs (19), and in the current study using the same dose of beads in the isolated lung, c&Javeraged 0.51. Although this average od did not differ significantly from that in the tranexamic acid control lobes, ad was still
Downloaded from www.physiology.org/journal/ajpheart at Macquarie Univ (137.111.162.020) on February 15, 2019.
EMBOLI
AND
FIBRINOLYSIS
INHIBITION
~0.53 in five of these eight lobes. In the emboli group pretreated with tranexamic acid gd was significantly decreased compared with that in the tranexamic acid controls. Although the postemboli cd tended to be further decreased by fibrinolysis inhibition, this trend was not significant. Despite the tendency for &J to be decreased, & after emboli alone was not significantly different from that measured in the control lobes, and the decrease in P,,i was minimal. In contrast, when fibrinolysis was inhibited, the decrease in bd was accompanied by a decrease in P,,i and an increase in &. However, it is important to remember that the effect of a fall in cd on P, i and KfC may be obscured to some extent by simultaneous increases in Pt or decreases in perfused surface area, respectively. Since lobes were embolized in vivo and left intact for a further 40-50 min, it is not unreasonable to expect some degree of edema formation even at the beginning of lobar perfusion. In the normally hydrated lung, small changes in interstitial volume lead to relatively large changes in P, (16), i.e., the lung interstitial compliance is low at this point. Increases in tissue fluid volume and P, would tend to offset the effect of decreases in bd on P,,i. In support of this argument, on excision, the embolized lobes were stiff and mottled in appearance, although no frank airway fluid was observed. Variability in the estimates of P, i and cd after emboli is likely related to a combination of emboli-induced factors including differences in Pt, flow heterogeneity, and the relation of P,i to all potential filtering areas, whereas the osmotic transient samples only those areas with flow. Unlike the measure of Pc,i, Kf,C is not affected by edema formation. Recently, Parker et al. (9) showed that KfC estimated by the zero-time extrapolation technique is insensitive to changes in interstitial fluid volume. In that study, even severe edema formation where the lobe weight increased to more than 150% of the basal weight did not alter &,,. Kf,,, however, is directly related to the amount of perfused exchange area. Emboli of the size used in this study (average 100 pm diameter) likely block some arterial inputs to capillary networks, reducing the amount of perfused exchange area. In that regard, Malik and van der Zee (7) found that with 100 pm glass beads, some areas of the lung were completely unstained by india ink, indicating lack of perfusion. In addition, Zelter et al. (20) found that a 0.3 g/kg dose of 200 pm glass beads decreased the permeability-surface area product for the small molecule urea. Townsley et al. (18) found that the increase in total vascular resistance in constant pressure perfused lobes is related in a predictable way to the amount of perfused lung mass and &. If all of the increase in RT in the present study resulted from vascular obstruction, we can estimate the amount of exchange area lost after embolization. Although these unperfused segments are likely exposed to some increase in vascular pressure during the & analysis, since both P, and P, are increased, the contribution of these segments to total lobe transvascular filtration may be brief because of rapid readjustment and equilibration of transvascular Starling forces under no or low flow conditions. RT increased by an average of 166% in the emboli group, compared with that in the tranexamic acid controls, which suggests that
IN
ISOLATED
LUNGS
H757
embolization resulted in a 55% loss of exchange area. This effect was exacerbated by fibrinolysis inhibition. From the average increase in RT in the emboli plus tranexamic acid group (369%), one can predict that microvascular exchange area was decreased by 73%. This analysis implies that since the measured J& is referenced to 100 g of initial lobe weight, these values in embolized lobes underestimate the actual KfC of the perfused microvasculature by 55 and 73%, respectively (18). Thus, in the embolized lobes, the effect of a small increase in microvascular permeability on & may be obscured by a concurrent decrease in exchange area. Malik and van der Zee (7) also found that lo-15% of the lung mass had a mottled appearance after emboli and india ink. This suggests that in some areas of the lung, perfusion to a segment is heterogeneous, and embolization of one arteriole may not totally impair perfusion of the capillary network downstream. This is supportable on anatomic grounds, as Reid showed that 25% of the total intralobar intra-acinar arterial supply consists of supernumerary arteries that provide a “backdoor” supply to the capillary bed, which may provide a source of collateral flow in embolic states (11). In addition, some capillary beds branch directly from arterioles that are 2100 pm in diameter (3). Finally, maintenance of a large effective capillary filtration surface area may be aided by a greater effective upstream or extra-alveolar filtration surface area as suggested by our earlier lymphatic analysis (19). The hypothesis that inhibition of fibrinolysis with delayed production of fibrin degradation products increases microvascular permeability and leads to greater edema formation has been proposed (5). This may potentially be important in our model, as fibrinolytic activity in the dog lung is high (14). However, fibrinolysis inhibition with tranexamic acid did not result in a clear-cut, significant exacerbation of embolic injury in these isolated canine lung lobes. Although & was increased in embolized lobes treated with tranexamic acid, and there was a tendency for both the mean osmotic transient gd and P,,i to be lower in the tranexamic acid group, these latter trends were not significant. Because of the high fibrinolytic activity in dog lungs (14), it remained possible that fibrinolysis inhibition with tranexamic acid was insufficient in these experiments and that higher doses of tranexamic acid would be necessary to show a clear effect of fibrinolysis inhibition. To evaluate this possibility, we performed three additional experiments where the dose of tranexamic acid (both the bolus injection and infusion rate) was doubled. In these lobes, the & (0.27 t 0.06 ml. mine1 cmHnO-’ JO0 g-l), RT (63.1 t 24.5 cmHBO l 1-1gmin. 100 g), and P,,i (7.5 t 1.6 cmHzO) were no different from those in the low-dose tranexamic acid group (using an unpaired t analysis) after 30 min of perfusion. Despite these initial similarities to the results obtained with low-dose tranexamic acid, & (0.16 t ml. min-l . cmH20-’ .lOO g-‘) and RT (47.8 t i5.4 cmHzOo 1-l *min. 100 g) in these lobes tended to return toward control values after 2 h of perfusion. Thus a greater degree of fibrinolysis inhibition did not exacerbate the embolic injury. In fact, these limited results would suggest that the hemodynamic and permeability changes l
Downloaded from www.physiology.org/journal/ajpheart at Macquarie Univ (137.111.162.020) on February 15, 2019.
H758
EMBOLI
AND FIBRINOLYSIS
INHIBITION
seen in the low-dose group may have been attenuated with complete inhibition of fibrinolysis. A similar finding was obtained in the sheep lung by Johnson et al. (4), who showed that pretreatment with tranexamic acid prevented the thrombin-induced increases in both the concentration of circulating fibrin degradation products and the pulmonary lymphatic protein clearance. Although fibrinolysis inhibition did not clearly exacerbate the embolic injury, one can predict that embolization after tranexamic acid would promote more edema formation or higher lymph flows in the intact lung than would emboli alone. If pulmonary blood flow were maintained constant at 0.84 1 min-’ . 100 g-l (the average flow in all control lobes at 30 min), mimicking the in vivo situation for the lung, one can predict the constant flow P, from the mean RT observed after emboli or emboli plus tranexamic acid. Initially, P, would be 39.6 t 9.2 and 49.1 & 14.7 cmHnO, respectively. After 120 min of perfusion at control flows, the predicted P, would not be different in the emboli group (38.9 t 14.7 cmHa0) but would tend to increase if fibrinolysis was inhibited (65.9 t 29.1 cmHnO). These figures are not unreasonable, since in vivo P, ranged from 24 to 46 cmHn0 30 min after bead infusion (0.6 g/kg) in the absence of fibrinolysis inhibition (19). The predicted constant flow PCs do not differ in the two groups (11.9 t 1.2 vs. 11.0 t 1.6 cmHz0 at 30 min) because of the much higher arterial resistance after tranexamic acid. Although mean RT observed in the two groups were not statistically different and the predicted PCs were similar., the resultant differences in the predicted P, may well result in significantly different degrees of edema formation in vivo. Recently, Ehrhart et al. (2) reported that embolized canine lung lobes perfused at high P, became more edematous than those perfused at low P,, despite similar PCs. The site of fluid filtration that contributes to such edema formation is likely to be close to or upstream of the emboli based on two lines of evidence. First, the traditional microvascular Starling forces do not appear to be different in these two conditions, either in the current study or in that of Ehrhart et al. (2). Furthermore, our lymphatic pore analysis after emboli suggested an increased filtration from normal low permeability segments exposed to the high pressures upstream from the emboli (19). In conclusion, this in vitro assessment of microvascular permeability substantiates our finding in the intact canine lung (19) that fibrinolysis inhibition is not necessarily a prerequisite for embolic injury. ad in isolated lung lobes was not significantly different in the presence or absence of fibrinolysis inhibition. Although & was not significantly altered by emboli alone, the accompanying increase in pulmonary vascular resistance suggests that changes in microvascular permeability are likely accompanied by a reduction in microvascular exchange area, an effect that was exacerbated by inhibition of fibrinolysis with tranexamic acid. KfC after emboli and treatment with tranexamic acid was significantly increased, and RJR” was nearly doubled. Our data predict that although fibrinolysis inhibition does not clearly exacerbate emboli-induced injury in the canine lung,
IN ISOLATED
LUNGS
associated changes in pulmonary hemodynamics may tend to accelerate the rate of transvascular fluid filtration and thus edema formation. The authors thank Sandy Worley for secretarial assistance and Penny Cook and Bonnie White for preparing the illustrations. This work was supported by National Heart, Lung, and Blood Institute Grants HL-06901, HL-25549, and HL-39045. Address for reprint requests: M. I. Townsley, Dept. of Physiology, MSB 3024, College of Medicine, University of South Alabama, Mobile, AL 36688. Received 29 June 1989; accepted in final form 17 October 1989. REFERENCES
l
1. DRAKE, R., K. A. GAAR, AND A. E. TAYLOR. Estimation of the filtration coefficient of pulmonary exchange vessels. Am. J. Physiol. 234 (Heart Circ. Physiol. 3): H266-H274, 1978. 2. EHRHART, I. C., J. E. HALL, AND W. F. HOFMAN. Vascular pressure effects on lung edema formation after glass bead embolism. J. Appl. Physiol. 62: 1989-1996, 1987. 3. HORSFIELD, K. Morphometry of the small pulmonary arteries in man. Circ. Res. 42: 593-597, 1978. 4. JOHNSON, A., M. V. TAHAMONT, AND A. B. MALIK. Thrombininduced lung vascular injury. Roles of fibrinogen and fibrinolysis. Am. Rev. Resp. Dis. 128: 38-44, 1983. 5. MALIK, A. B. Pulmonary microembolism. Physiol. Rev. 63: lll51207,1983. 6. MALIK, A. B., B. C. LEE, H. VAN DER ZEE, AND A. JOHNSON. The role of fibrin in the genesis of pulmonary edema after embolization in dogs. Circ. Res. 45: 120-125, 1979. 7. MALIK, A. B., AND H. VAN DER ZEE. Mechanism of pulmonary edema induced by microembolization in dogs. Circ. Res. 42: 72-79, 1978.
8. NAVAR, P. D., AND L. G. NAVAR. Relationship between colloid osmotic pressure and plasma protein concentration in the dog. Am. J. Physiol. 233 (Heart Circ. Physiol. 2): H295-H298, 1977. 9. PARKER, J. C., M. I. TOWNSLEY, AND J. T. CARTLEDGE. Lung edema increases transvascular filtration rate but not filtration coefficient. J. Appl. Physiol. 66: 1553-1560, 1989. 10. PARKER, J. C., R. E. PARKER, D. N. GRANGER, AND A. E. TAYLOR. Vascular permeability and transvascular fluid and protein transport in the dog lung. Circ. Res. 48: 549-561, 1981. 11. REID, L. Structural and functional reappraisal of the pulmonary artery system. In: Scientific Basis of Medicine. Annual Reviews, 1968, p. 289-307. 12. RICHARDSON, P. D. I., D. N. GRANGER, AND A. E. TAYLOR. Capillary filtration coefficient: the technique and its application to the small intestine. Cardiovasc. Res. 13: 547-561, 1979. 13. RIPPE, B., M. TOWNSLEY, J. C. PARKER, AND A. E. TAYLOR. Osmotic reflection coefficient for total plasma protein in lung microvessels. J. Appt. Physiol. 58: 436-442, 1985. 14. SALDEEN, T. The microembolism syndrome. Microvasc. Res. 11: 227-259,1976. 15. STEEL, R. G. D., AND J. H. TORRIE. Principles and Procedures of Statistics (2nd ed.). New York: McGraw-Hill, 1980. 16. TAYLOR, A. E., AND J. C. PARKER. Pulmonary interstitial spaces and lymphatics. In: Handbook of Physiology. The Respiratory System. Circulation and Nonrespiratory Functions. Bethesda, MD: Am. Physiol. Sot., 1985, sect. 3, vol. I, chapt. 4, p. 167-230. 17. TOWNSLEY, M. I., R. J. KORTHUIS, B. RIPPE, J. C. PARKER, AND A. E. TAYLOR. Validation of double vascular occlusion method for P,,i in lung and skeletal muscle. J. Appl. Physiol. 61: 127-132,1986. 18. TOWNSLEY, M. I., J. C. PARKER, R. J. KORTHUIS, AND A. E. TAYLOR. Alterations in hemodynamics and Kf,, during lung mass resection. J. Appl. Physiol. 63: 2460-2466, 1987. 19. TOWNSLEY, M. I., J. C. PARKER, G. L. LONGENECKER, M. L. PERRY, R. M. PITT, AND A. E. TAYLOR. Pulmonary embolism: analysis of endothelial pore sizes in canine lung. Am. J. Physiol. 255 (Heart Circ. Physiol. 24): H1075-H1083, 1988. AND J. F. MURRAY. 20. ZELTER, M., A. LIPAVSKY, J. M. HOEFFEL, Effect of lung injuries on [ 14C]urea permeability-surface area product in dogs. J. Appl. Physiol. 56: 1512-1520, 1984.
Downloaded from www.physiology.org/journal/ajpheart at Macquarie Univ (137.111.162.020) on February 15, 2019.
I. TOWNSLEY,
Department
of Physiology,
SCOTT
A. BARMAN,
College of Medicine,
TOWNSLEY, MARY I., SCOTTA. BARMAN, AND AUBREY E. TAYLOR. Pulmonary embolism: emboli and fibrinolysis inhibition in isolated canine lungs. Am. J. Physiol. 258 (Heart Circ. Physiol. 27): H754-H758, 1990.-The effect of fibrinolysis inhibition with tranexamic acid on pulmonary microvascular permeability during glass bead embolization was investigated in the isolated lung. Lung lobes from nonheparinized dogs were treated in vivo with the equivalent of 0.6 g/kg 100 pm glass bead emboli alone, emboli after tranexamic acid, tranexamic acid alone, or the bead vehicle alone. After 40-50 min, the lobes were isolated for ex vivo perfusion with heparinized autologous blood. There were no changes in any parameter over the 120min perfusion period. Blood flow at 120 min was decreased after both emboli alone and emboli with tranexamic acid, reflecting an increase in vascular resistance compared with the Tween or tranexamic acid controls. Furthermore, tranexamic acid increased the ratio of pre- to postcapillary resistance in embolized lobes compared with that after emboli alone or in the Tween or tranexamic acid controls. The isogravimetric capillary pressure and the osmotic reflection coefficient were not significantly decreased by tranexamic acid compared with those after emboli alone; however, it did result in an increase in the capillary filtration coefficient compared with that after emboli alone or in the control groups. We conclude that although fibrinolysis inhibition does not clearly exacerbate the lung injury seen after emboli, the tranexamic acid-induced changes in hemodynamics would tend to accelerate edema formation.
capillary filtration coefficient; osmotic reflection coefficient; osmotic transient; isogravimetric capillary pressure; tranexamic acid
THEROLEOFHUMORALFACTORS intheedemaformation resulting from embolization of the lung has long been debated. Malik and van der Zee (7) determined that heparin pretreatment before glass bead embolization of the canine lung prevented the edema formation seen when coagulation and fibrinolysis were allowed to proceed normally. Subsequently, these investigators showed that inhibition of fibrinolysis delayed the resolution of emboli-induced edema formation in the canine lung (6). In contrast, after complete inhibition of fibrinolysis in the sheep lung the increase in the lymphatic protein clearance normally seen as a result of thrombin embolization was attenuated. This finding suggests that some slow, continual generation of fibrin degradation products is responsible for the emboli-induced changes in pulmonary microvascular protein and fluid exchange (4). Based on these findings, Malik et al. and others (6, 7, 14) have proposed that fibrinolysis with the resultant production H754
0363-6135/90 $1.50 Copyright
0
AND
University
AUBREY
E. TAYLOR
of South Alabama,
Mobile, Alabama 36688
of fibrin degradation products is a necessary condition for the development of emboli-induced microvascular injury. However, whether fibrinolysis results in lung injury would still seem to be an open question, since edema formation and changes in lymphatic protein clearance may develop from either a change in the transcapillary Starling forces (hemodynamic edema) or from a direct microvascular injury. It is pertinent in this regard that parameters that specifically describe the permeability of the lung microvasculature, such as the capillary filtration coefficient (&,,) or the osmotic reflection coefficient (ad), were not measured in any of these studies. Thus the role of fibrinolysis is difficult to determine. We have recently shown that glass bead embolization (0.2-0.6 g/kg) in the absence of a fibrinolysis inhibitor results in an increase in pulmonary capillary pressure and microvascular permeability in the intact canine lung lymphatic model (19). ad was significantly reduced to 0.55 overall compared with that in normal lungs where fld = 0.62 (10). Although this study demonstrated that inhibition of fibrinolysis was not necessary for injury to occur, no conclusions could be made regarding the role of fibrinolysis inhibition in modulating the pulmonary injury following emboli. As we encountered difficulties in maintaining dogs with lung lymph catheters throughout the entire protocol with 0.6 g/kg dose of beads (19), we chose to examine the problem of fibrinolysis inhibition in the isolated canine lung lobe. Glass beads, at a dose equivalent to 0.6 g/kg, were administered to the canine lower left lobe in situ in the presence and absence of a fibrinolysis inhibitor, tranexamic acid. The dose of tranexamic acid administered in vivo was exactly that used previously (6). Lobes were subsequently isolated for ex vivo perfusion to measure several permeability and hemodynamic parameters, including &C and ad. METHODS
Twenty-four mongrel dogs (20.7 t 1.0 kg, means t SE) were anesthetized with 30 mg/kg pentobarbital sodium iv, intubated, and ventilated with room air (15 ml/ kg at 12 breaths/min). The left carotid artery and jugular vein were cannulated with polyethylene catheters for blood withdrawal or drug infusion, respectively. A surgical plane of anesthesia was maintained with supplemental intravenous boluses of pentobarbital. The left lung was exposed via an incision along the fifth intercostal space. The branch of the left pulmonary artery to the
1990 the American
Physiological
Society
Downloaded from www.physiology.org/journal/ajpheart at Macquarie Univ (137.111.162.020) on February 15, 2019.
EMBOLI
AND
FIBRINOLYSIS
INHIBITION
left upper lobe was cannulated in a retrograde manner with a small polyethylene catheter. The catheter was fashioned with a right angle 1 cm from the tip of the catheter and inserted such that this l-cm tip was parallel (orthograde) to the direction of flow in the left pulmonary artery. The branch feeding the left middle lobe was either clamped or ligated. Four groups of lungs were studied. One group received the equivalent of 0.6 g/kg 100 pm glass beads (range 77125 pm, Ferro) via the upper lobar catheter (n = 8). Beads were prepared in 5 ml of saline with 1 drop Tween 80 to prevent clumping. Since lower left lobe weight averages 25% of the entire lung weight (18), 25% of the calculated 0.6 g/kg dose was infused into the lower left lobe. In six lobes of this group, siliconized glass beads were used, and in the remaining two lobes, noncoated beads were infused. As there were no differences observed in any parameter with respect to presence or absence of silicon, these lungs were treated as a single group. The second group of lobes (n = 6) received a 50 mg/kg iv bolus of tranexamic acid followed by its infusion (0.83 mg* kg-’ . min-‘) for the next 55 min (6). The equivalent of 0.6 g/kg beads was injected 15 min after initiating the tranexamic acid infusion. Control lobes received either tranexamic acid alone at the same dose (n = 5) or the saline-Tween vehicle alone (n = 5). Immediately after bead infusion, the left upper and middle lobes were ligated at the hilar margin and excised. Loose ligatures were placed around the left pulmonary artery and lobar bronchus; flow to the left lower lobe was maintained. Heparin (10,000 U iv) was administered 30 min after bead infusion. Ten minutes later, 600-800 ml of blood were removed via the carotid catheter to prime the perfusion system. The pulmonary artery was then ligated, and the left lower lobe with the attached left atria1 appendage was then rapidly excised and weighed. Plastic cannulas were secured in the lobar artery and vein and the lobar bronchus, and perfusion was begun within lo-20 min of excision (50-60 min after bead infusion). The isolated lung perfusion system has previously been described in detail (1,13,18). Briefly, the lower left lobes (53.7 t 3.8 g) were suspended from a counterbalanced force transducer and perfused with warmed (38°C) and oxygenated (95% 02-5% COs) autologous blood at constant perfusion pressures. Pulmonary arterial (Pa) and venous (PJ pressures were measured with pressure transducers via thin catheters positioned at the orifices of the inflow and outflow cannulas, respectively, and referenced to the top of the lung hilus. Blood flow (Q) was measured with a flow-through electromagnetic flow probe positioned in the venous outflow line. P,, P,, airway pressure, and lung weight were continuously recorded on a Beckman polygraph. The sensitivity of the weighing system was adjusted to provide 1 cm of pen deflection per gram weight change. Measurements. Lobes were inflated with compressed air to a constant airway pressure of 3 cmH20 and perfused under Zone III conditions. P, was set at 4-5 cmHg0 and P, adjusted to yield an isogravimetric state, i.e., the lobe was neither gaining nor losing weight. Isogravimetric capillary pressure (Pc,i) was measured via the double
IN
ISOLATED
H755
LUNGS
vascular occlusion technique (17). If the lobe was not exactly isogravimetric, this technique measured the prevailing pulmonary capillary pressure (PC). Under these conditions, P,,i could be estimated (13) by relating the ongoing rate of weight gain (AW/At) to the capillary filtration coefficient (Kf,,) Estimated
AWlAt
P,i, = P, - 7
f,c
P,i can be used as a qualitative index of permeability when plasma protein concentration remains stable, since P,i is related to the osmotic reflection coefficient for total proteins (Q): P,i = Pt + c&,-?rJ, where Pt is tissue hydrostatic pressure, and ?rp and 7rt are the plasma and tissue colloid osmotic pressures, respectively. Total vascular resistance (RT) was calculated as (P, - PV)/&. P, was used to partition RT into pre- and postcapillary resistance segments; the ratio (RJR,) is reported here. &, may also be used as an index of microvascular permeability in the Zone III isolated lung (1, 12), assuming that exchange surface area remains constant (18). &, was calculated following a step increase in both P, and P, of approximately 5-8 cmHa0. This results in an initial rapid weight gain due to a shift of blood into the lungs followed by a slower rate of weight gain, which can be attributed to transcapillary filtration. P, was measured just before and at the end of the step increase in pressure. The rate of weight gain after the increase in pressure was plotted on semilogarithmic paper as a function of time, and the slow phase of the weight transient was extrapolated to time 0. This initial filtration rate at t= 0 was divided by the induced pressure increment (AP,) to yield Kf,,. Finally, &-Jwas measured with the osmotic transient technique (13). The extrapolated AW/At at time 0 following a step decrease in ?rpwas used to calculate &-J
(2) Plasma protein concentrations (C,) were measured by refractometry before and after dilution of the circulating blood with warmed saline. The refractometer was calibrated with pooled dog plasma. rp was calculated using the equation derived by Navar and Navar (8) for dog plasma proteins. Experimental protocol. Isogravimetric conditions were established fairly quickly after perfusion of the lung lobes, and initial measurements were begun within 30 min of perfusion. Hemodynamic parameters, Kf,, and Pc,i, were measured every 30 min for 120 min. When an isogravimetric state was reached after the fourth KfC measurement, C, and P, were measured and the blood rapidly diluted with warmed saline to produce the osmotic transient. Care was taken to adjust vascular pressures if necessary to keep them constant during this maneuver. Lung weight was monitored for lo-15 min after dilution. After the osmotic transient analysis, Kf,, measurements were repeated to ensure that saline dilution had induced no changes. Statistical analyses. All data are means 2 SE. The effects of perfusion time within each group as well as differences between treatment groups were statistically
Downloaded from www.physiology.org/journal/ajpheart at Macquarie Univ (137.111.162.020) on February 15, 2019.
H756
EMBOLI
AND
FIBRINOLYSIS
INHIBITION
IN
evaluated using an analysis of variance with post hoc tests to identify specific differences (15).
ISOLATED
LUNGS 0 m
30 min 120 mfn
RESULTS
There were no effects of perfusion time in any group for any parameter, thus only the data collected after 30 i . 0.2 and 120 min of perfusion are presented. The stability of .the isolated lobes with perfusion time implies that interE group differences are the result of changes that were T initiated in vivo. 0.1 Embolization resulted in a fall in the isogravimetric & in these experiments, as shown in Table 1. Q was markedly decreased and P, increased (P < 0.05) in the emboli and emboli plus tranexamic acid groups compared with Twoan TEA Em boli those in either of the control groups. RT tended to Controls + TEA increase as a result to 41.2t 11.1and 72.7t 31.4cmHzOS FIG. 1. Capillary filtration coefficients (&,) as a function of time 1-l *min. 100 g at 120 min after emboli and emboli plus after perfusion after emboli in absence or presence of fibrinolysis tranexamic acid, respectively. However, these changes in inhibitor, tranexamic acid (TEA). Control lobes were treated with saline-Tween vehicle alone (Tween) or tranexamic acid alone. * P < RT following emboli were only statistically significant 0.05 vs. Tween control group; + P < 0.05 vs. tranexamic acid control when fibrinolysis was inhibited. RJR” in the emboli group; and @P < 0.05 vs. response after emboli alone. group (3.27 & 0.50) was increased at 30 min compared with-that in the tranexamic acid or Tween control-lobes [ 30 min (1.27 & 0.36 and 1.24 & 0.13, respectively). After 120 min m 120 min of perfusion, this difference in R,/Rv was not significant. The only hemodynamic effect exacerbated as a result of fibrinolysis inhibition was the emboli-induced increase in RJR”. RJR” nearly doubled in the embolized lobes with tranexamic acid (6.08 * 1.24, P < 0.05 vs. emboli alone), reflecting the larger arterial or precapillary resistance in this group. .The effects of fibrinolysis inhibition on pulmonary 5 microvascular permeability after emboli are shown in on a, Figs. 1 and 2. After 120 min of perfusion, KfC in the emboli group (0.18 t 0.02 ml Smin-l cmH20D1 *‘lo0 g-‘) was not different from that in the Tween or tranexamic acid controls (0.17 & 0.01 and 0.16 t 0.01 mlmin-la Twoon TEA Emboli Em boll cmHzO-’ 100 g-l, respectively). After tranexamic acid, Controls +TEA however, embolization resulted in a significant increase FIG. 2. Isogravimetric capillary pressure (Pqi) as a function of time in & to 0.29 & 0.03 ml. min-’ •rnHzO-‘~ 100 g-l at 120 after perfusion. * P < 0.05 vs. Tween control group and + P < 0.05 vs. min (P < 0.05). In addition, after 120 min of perfusion, tranexamic acid (TEA) control group. l
TABLE 1. Pulmonary hemodynamics in absence or presence of fibrinolysis inhibition with tranexamic acid in isolated embolized canine lung lobes Group Time, min
R
30 120 30
P”
120
Q
30 120
Tween
Tranexamic
Tranexamic acid plus emboli
acid
Emboli
152~1.2 15.fkk1.3 4.3kO.3 4.2t0.3 0.87kO.07
16.7kO.7 17.3zk1.3 5.3t0.4* 4.8kO.4 0.87zkO.11
22.1,+1.7"~
24.5+1.8*t
21.4f1.6*? 4.2*O.lt
24.0+1.5*t 4.5+O.lt
4.0&0.2-f 0.54+0.07*-t
4.620.2 0.46+0.06*"f
0.91kO.07 12.4kl.l 12.7k1.5
0.85t0.13 14.MO.9 15.521.7
0.54&0.06*t 0.41+0.07*t 30 42.1A10.2 53.1-e15.9*t RT 120 41.2tll.l 72.7a31.4”t 30 1.24kO.13 1.27k0.36 3.27+0.50*7 5.90+1.07*t$ RafRv 120 1.21kO.05 1.17t0.27 2.88k0.45 6.08+1.24*?$ Data are means k SE at 30 and 120 min after initiation of in vitro perfusion. Pa, lobar arterial pressure (cmHzO); P,, lobar venous pressure (cmHnO); Q, blood flow (1. min-’ 100 g-‘); RT, total vascular resistance (cmHa0 1-l *min. 100 g); and RJR,, the ratio of pre- to postcapillary resistance. * P < 0.05 vs. Tween control; t P c 0.05 vs. tranexamic acid control; $ P < 0.05 vs. emboh. l
l
P,,i was significantly decreased (P c 0.05) in both the emboli (8.6 t 0.2 cmHsO) and the emboli plus tranexamic acid group (7.7 $- 1.1 cmHeO) as shown in Fig. 2. ad was 0.51 t 0.04 and 0.41 -c-0.03 after emboli alone or emboli plus tranexamic acid, respectively. Although both values are less (P < 0.05) than the ad measured in normal canine lobes (0.65 -$-0.02) when this osmotic transient technique (13) is used, only the cd measured after emboli plus tranexamic acid was significantly different from that in the tranexamic acid controls (0.63 t 0.02, n = 4). DISCUSSION
The results of this study support our earlier finding in intact dogs that, despite the high fibrinolytic activity in dogs, fibrinolysis inhibition is not necessarily a prerequisite for pulmonary VaSCdar injury after emboli. &J for total protein averaged 0.50 after 0.6 g/kg emboli in three intact dogs (19), and in the current study using the same dose of beads in the isolated lung, c&Javeraged 0.51. Although this average od did not differ significantly from that in the tranexamic acid control lobes, ad was still
Downloaded from www.physiology.org/journal/ajpheart at Macquarie Univ (137.111.162.020) on February 15, 2019.
EMBOLI
AND
FIBRINOLYSIS
INHIBITION
~0.53 in five of these eight lobes. In the emboli group pretreated with tranexamic acid gd was significantly decreased compared with that in the tranexamic acid controls. Although the postemboli cd tended to be further decreased by fibrinolysis inhibition, this trend was not significant. Despite the tendency for &J to be decreased, & after emboli alone was not significantly different from that measured in the control lobes, and the decrease in P,,i was minimal. In contrast, when fibrinolysis was inhibited, the decrease in bd was accompanied by a decrease in P,,i and an increase in &. However, it is important to remember that the effect of a fall in cd on P, i and KfC may be obscured to some extent by simultaneous increases in Pt or decreases in perfused surface area, respectively. Since lobes were embolized in vivo and left intact for a further 40-50 min, it is not unreasonable to expect some degree of edema formation even at the beginning of lobar perfusion. In the normally hydrated lung, small changes in interstitial volume lead to relatively large changes in P, (16), i.e., the lung interstitial compliance is low at this point. Increases in tissue fluid volume and P, would tend to offset the effect of decreases in bd on P,,i. In support of this argument, on excision, the embolized lobes were stiff and mottled in appearance, although no frank airway fluid was observed. Variability in the estimates of P, i and cd after emboli is likely related to a combination of emboli-induced factors including differences in Pt, flow heterogeneity, and the relation of P,i to all potential filtering areas, whereas the osmotic transient samples only those areas with flow. Unlike the measure of Pc,i, Kf,C is not affected by edema formation. Recently, Parker et al. (9) showed that KfC estimated by the zero-time extrapolation technique is insensitive to changes in interstitial fluid volume. In that study, even severe edema formation where the lobe weight increased to more than 150% of the basal weight did not alter &,,. Kf,,, however, is directly related to the amount of perfused exchange area. Emboli of the size used in this study (average 100 pm diameter) likely block some arterial inputs to capillary networks, reducing the amount of perfused exchange area. In that regard, Malik and van der Zee (7) found that with 100 pm glass beads, some areas of the lung were completely unstained by india ink, indicating lack of perfusion. In addition, Zelter et al. (20) found that a 0.3 g/kg dose of 200 pm glass beads decreased the permeability-surface area product for the small molecule urea. Townsley et al. (18) found that the increase in total vascular resistance in constant pressure perfused lobes is related in a predictable way to the amount of perfused lung mass and &. If all of the increase in RT in the present study resulted from vascular obstruction, we can estimate the amount of exchange area lost after embolization. Although these unperfused segments are likely exposed to some increase in vascular pressure during the & analysis, since both P, and P, are increased, the contribution of these segments to total lobe transvascular filtration may be brief because of rapid readjustment and equilibration of transvascular Starling forces under no or low flow conditions. RT increased by an average of 166% in the emboli group, compared with that in the tranexamic acid controls, which suggests that
IN
ISOLATED
LUNGS
H757
embolization resulted in a 55% loss of exchange area. This effect was exacerbated by fibrinolysis inhibition. From the average increase in RT in the emboli plus tranexamic acid group (369%), one can predict that microvascular exchange area was decreased by 73%. This analysis implies that since the measured J& is referenced to 100 g of initial lobe weight, these values in embolized lobes underestimate the actual KfC of the perfused microvasculature by 55 and 73%, respectively (18). Thus, in the embolized lobes, the effect of a small increase in microvascular permeability on & may be obscured by a concurrent decrease in exchange area. Malik and van der Zee (7) also found that lo-15% of the lung mass had a mottled appearance after emboli and india ink. This suggests that in some areas of the lung, perfusion to a segment is heterogeneous, and embolization of one arteriole may not totally impair perfusion of the capillary network downstream. This is supportable on anatomic grounds, as Reid showed that 25% of the total intralobar intra-acinar arterial supply consists of supernumerary arteries that provide a “backdoor” supply to the capillary bed, which may provide a source of collateral flow in embolic states (11). In addition, some capillary beds branch directly from arterioles that are 2100 pm in diameter (3). Finally, maintenance of a large effective capillary filtration surface area may be aided by a greater effective upstream or extra-alveolar filtration surface area as suggested by our earlier lymphatic analysis (19). The hypothesis that inhibition of fibrinolysis with delayed production of fibrin degradation products increases microvascular permeability and leads to greater edema formation has been proposed (5). This may potentially be important in our model, as fibrinolytic activity in the dog lung is high (14). However, fibrinolysis inhibition with tranexamic acid did not result in a clear-cut, significant exacerbation of embolic injury in these isolated canine lung lobes. Although & was increased in embolized lobes treated with tranexamic acid, and there was a tendency for both the mean osmotic transient gd and P,,i to be lower in the tranexamic acid group, these latter trends were not significant. Because of the high fibrinolytic activity in dog lungs (14), it remained possible that fibrinolysis inhibition with tranexamic acid was insufficient in these experiments and that higher doses of tranexamic acid would be necessary to show a clear effect of fibrinolysis inhibition. To evaluate this possibility, we performed three additional experiments where the dose of tranexamic acid (both the bolus injection and infusion rate) was doubled. In these lobes, the & (0.27 t 0.06 ml. mine1 cmHnO-’ JO0 g-l), RT (63.1 t 24.5 cmHBO l 1-1gmin. 100 g), and P,,i (7.5 t 1.6 cmHzO) were no different from those in the low-dose tranexamic acid group (using an unpaired t analysis) after 30 min of perfusion. Despite these initial similarities to the results obtained with low-dose tranexamic acid, & (0.16 t ml. min-l . cmH20-’ .lOO g-‘) and RT (47.8 t i5.4 cmHzOo 1-l *min. 100 g) in these lobes tended to return toward control values after 2 h of perfusion. Thus a greater degree of fibrinolysis inhibition did not exacerbate the embolic injury. In fact, these limited results would suggest that the hemodynamic and permeability changes l
Downloaded from www.physiology.org/journal/ajpheart at Macquarie Univ (137.111.162.020) on February 15, 2019.
H758
EMBOLI
AND FIBRINOLYSIS
INHIBITION
seen in the low-dose group may have been attenuated with complete inhibition of fibrinolysis. A similar finding was obtained in the sheep lung by Johnson et al. (4), who showed that pretreatment with tranexamic acid prevented the thrombin-induced increases in both the concentration of circulating fibrin degradation products and the pulmonary lymphatic protein clearance. Although fibrinolysis inhibition did not clearly exacerbate the embolic injury, one can predict that embolization after tranexamic acid would promote more edema formation or higher lymph flows in the intact lung than would emboli alone. If pulmonary blood flow were maintained constant at 0.84 1 min-’ . 100 g-l (the average flow in all control lobes at 30 min), mimicking the in vivo situation for the lung, one can predict the constant flow P, from the mean RT observed after emboli or emboli plus tranexamic acid. Initially, P, would be 39.6 t 9.2 and 49.1 & 14.7 cmHnO, respectively. After 120 min of perfusion at control flows, the predicted P, would not be different in the emboli group (38.9 t 14.7 cmHa0) but would tend to increase if fibrinolysis was inhibited (65.9 t 29.1 cmHnO). These figures are not unreasonable, since in vivo P, ranged from 24 to 46 cmHn0 30 min after bead infusion (0.6 g/kg) in the absence of fibrinolysis inhibition (19). The predicted constant flow PCs do not differ in the two groups (11.9 t 1.2 vs. 11.0 t 1.6 cmHz0 at 30 min) because of the much higher arterial resistance after tranexamic acid. Although mean RT observed in the two groups were not statistically different and the predicted PCs were similar., the resultant differences in the predicted P, may well result in significantly different degrees of edema formation in vivo. Recently, Ehrhart et al. (2) reported that embolized canine lung lobes perfused at high P, became more edematous than those perfused at low P,, despite similar PCs. The site of fluid filtration that contributes to such edema formation is likely to be close to or upstream of the emboli based on two lines of evidence. First, the traditional microvascular Starling forces do not appear to be different in these two conditions, either in the current study or in that of Ehrhart et al. (2). Furthermore, our lymphatic pore analysis after emboli suggested an increased filtration from normal low permeability segments exposed to the high pressures upstream from the emboli (19). In conclusion, this in vitro assessment of microvascular permeability substantiates our finding in the intact canine lung (19) that fibrinolysis inhibition is not necessarily a prerequisite for embolic injury. ad in isolated lung lobes was not significantly different in the presence or absence of fibrinolysis inhibition. Although & was not significantly altered by emboli alone, the accompanying increase in pulmonary vascular resistance suggests that changes in microvascular permeability are likely accompanied by a reduction in microvascular exchange area, an effect that was exacerbated by inhibition of fibrinolysis with tranexamic acid. KfC after emboli and treatment with tranexamic acid was significantly increased, and RJR” was nearly doubled. Our data predict that although fibrinolysis inhibition does not clearly exacerbate emboli-induced injury in the canine lung,
IN ISOLATED
LUNGS
associated changes in pulmonary hemodynamics may tend to accelerate the rate of transvascular fluid filtration and thus edema formation. The authors thank Sandy Worley for secretarial assistance and Penny Cook and Bonnie White for preparing the illustrations. This work was supported by National Heart, Lung, and Blood Institute Grants HL-06901, HL-25549, and HL-39045. Address for reprint requests: M. I. Townsley, Dept. of Physiology, MSB 3024, College of Medicine, University of South Alabama, Mobile, AL 36688. Received 29 June 1989; accepted in final form 17 October 1989. REFERENCES
l
1. DRAKE, R., K. A. GAAR, AND A. E. TAYLOR. Estimation of the filtration coefficient of pulmonary exchange vessels. Am. J. Physiol. 234 (Heart Circ. Physiol. 3): H266-H274, 1978. 2. EHRHART, I. C., J. E. HALL, AND W. F. HOFMAN. Vascular pressure effects on lung edema formation after glass bead embolism. J. Appl. Physiol. 62: 1989-1996, 1987. 3. HORSFIELD, K. Morphometry of the small pulmonary arteries in man. Circ. Res. 42: 593-597, 1978. 4. JOHNSON, A., M. V. TAHAMONT, AND A. B. MALIK. Thrombininduced lung vascular injury. Roles of fibrinogen and fibrinolysis. Am. Rev. Resp. Dis. 128: 38-44, 1983. 5. MALIK, A. B. Pulmonary microembolism. Physiol. Rev. 63: lll51207,1983. 6. MALIK, A. B., B. C. LEE, H. VAN DER ZEE, AND A. JOHNSON. The role of fibrin in the genesis of pulmonary edema after embolization in dogs. Circ. Res. 45: 120-125, 1979. 7. MALIK, A. B., AND H. VAN DER ZEE. Mechanism of pulmonary edema induced by microembolization in dogs. Circ. Res. 42: 72-79, 1978.
8. NAVAR, P. D., AND L. G. NAVAR. Relationship between colloid osmotic pressure and plasma protein concentration in the dog. Am. J. Physiol. 233 (Heart Circ. Physiol. 2): H295-H298, 1977. 9. PARKER, J. C., M. I. TOWNSLEY, AND J. T. CARTLEDGE. Lung edema increases transvascular filtration rate but not filtration coefficient. J. Appl. Physiol. 66: 1553-1560, 1989. 10. PARKER, J. C., R. E. PARKER, D. N. GRANGER, AND A. E. TAYLOR. Vascular permeability and transvascular fluid and protein transport in the dog lung. Circ. Res. 48: 549-561, 1981. 11. REID, L. Structural and functional reappraisal of the pulmonary artery system. In: Scientific Basis of Medicine. Annual Reviews, 1968, p. 289-307. 12. RICHARDSON, P. D. I., D. N. GRANGER, AND A. E. TAYLOR. Capillary filtration coefficient: the technique and its application to the small intestine. Cardiovasc. Res. 13: 547-561, 1979. 13. RIPPE, B., M. TOWNSLEY, J. C. PARKER, AND A. E. TAYLOR. Osmotic reflection coefficient for total plasma protein in lung microvessels. J. Appt. Physiol. 58: 436-442, 1985. 14. SALDEEN, T. The microembolism syndrome. Microvasc. Res. 11: 227-259,1976. 15. STEEL, R. G. D., AND J. H. TORRIE. Principles and Procedures of Statistics (2nd ed.). New York: McGraw-Hill, 1980. 16. TAYLOR, A. E., AND J. C. PARKER. Pulmonary interstitial spaces and lymphatics. In: Handbook of Physiology. The Respiratory System. Circulation and Nonrespiratory Functions. Bethesda, MD: Am. Physiol. Sot., 1985, sect. 3, vol. I, chapt. 4, p. 167-230. 17. TOWNSLEY, M. I., R. J. KORTHUIS, B. RIPPE, J. C. PARKER, AND A. E. TAYLOR. Validation of double vascular occlusion method for P,,i in lung and skeletal muscle. J. Appl. Physiol. 61: 127-132,1986. 18. TOWNSLEY, M. I., J. C. PARKER, R. J. KORTHUIS, AND A. E. TAYLOR. Alterations in hemodynamics and Kf,, during lung mass resection. J. Appl. Physiol. 63: 2460-2466, 1987. 19. TOWNSLEY, M. I., J. C. PARKER, G. L. LONGENECKER, M. L. PERRY, R. M. PITT, AND A. E. TAYLOR. Pulmonary embolism: analysis of endothelial pore sizes in canine lung. Am. J. Physiol. 255 (Heart Circ. Physiol. 24): H1075-H1083, 1988. AND J. F. MURRAY. 20. ZELTER, M., A. LIPAVSKY, J. M. HOEFFEL, Effect of lung injuries on [ 14C]urea permeability-surface area product in dogs. J. Appl. Physiol. 56: 1512-1520, 1984.
Downloaded from www.physiology.org/journal/ajpheart at Macquarie Univ (137.111.162.020) on February 15, 2019.
