INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pulmonary transplantation has proved successful in the treatment of selected patients with end-stage pulmonary disease. Improved preservation techniques have largely contributed to ameliorate the success of lung transplantation. A 1-yr patient survival rate greater than 90% for single lung transplantation and greater than 85% for bilateral lung transplantation has been achieved (1). However, an initial graft dysfunction frequently increases postoperative morbidity and can be as high as 46% (2). Consequently, the quality of preservation is an important factor to consider, because it attenuates lung ischemia/ reperfusion injuries and ensures a better morphologic, biochemical, and functional integrity of the lung.

Since 1984, the modified Euro-Collins (EC) solution has been commonly used for pulmonary preservation. This crystalloid solution is an intracellular-type solution containing a high concentration of potassium and a low concentration of sodium, and was initially developed for kidney preservation (3). Several experimental studies showed the advantages of using extracellular colloid-based solutions for pulmonary preservation (4). Indeed, low-potassium dextran (LPD), an extracellular solution containing a phosphate buffer system and dextran 40, has been successfully used in various animal models (5, 8). Another extracellular colloid solution (modified blood or Wallwork solution) has also been used with excellent clinical results (12).

The quality of preservation is commonly assessed by lung function parameters, lung water content, and histopathologic changes. However, lung ischemia/reperfusion injuries may induce subclinical changes of microvascular permeability. Thus, the aim of the study was to compare, besides the lung function parameters, the microvascular permeability of lungs preserved with two extracellular colloid solutions, leukocyte-depleted blood (BL) and low-potassium dextran, with the currently used Euro-Collins solution, in an ex vivo porcine lung reperfusion model.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Preparation

Fasted domestic pigs of both sexes (n = 22, 36 ± 3 kg) were premedicated with intramuscular ketamine (10 mg/kg), midazolam (0.05 mg/ kg), and atropine (0.02 mg/kg). Anesthesia was induced with halothane and a catheter was placed in an ear vein. After endotracheal intubation (orotracheal tube No. 7.0), anesthesia was maintained with halothane (1-2%), intravenous fentanyl (2-5 mg/kg in repeated doses), and pancuronium (0.4 mg/kg in a single dose). The lungs were mechanically ventilated (Servo Ventilator 900; Siemens-Elema, Solna, Sweden) with a tidal volume of 15 ml/kg and the respiratory rate was adjusted to obtain an end-tidal CO₂ concentration of between 4 and 5 vol%. An inspired oxygen concentration of 100% was used and a positive end-expiratory pressure (PEEP) of 5 cm H₂O was applied. The right external jugular vein was cannulated for fluid administration. The animals were heparinized by systemic administration of 10,000 IU of heparin (Liquemin; Roche Pharma, Basel, Switzerland) after completion of the preliminary dissection. During anesthesia, 1,500 ml of blood was sampled from the animal and replaced with a dextran 70 solution (Macrodex 6%, in saline solution; Braun Medical AG, Emmenbrücke, Switzerland). The sampled blood was preserved at 4° C in bags containing citrate-phosphate-dextrose (Terumo Europe N.V., Leuven, Belgium) and used to prepare the ex vivo circuit perfusate and the leukocyte-depleted blood preservation solution. The experimental protocol conformed to the Guiding Principles in the Care and Use of Animals as approved by the Council of the American Physiological Society, and was reviewed by the Ethics Committee for Animal Research and the Veterinary Office of our institution.

Harvesting and Preservation of the Lungs

After median sternotomy and opening of the pleura and pericardium, the superior vena cava, pulmonary artery, and ascending aorta were dissected. A purse-string suture was placed anteriorly on the main pulmonary artery. After administration of 500 µg of prostaglandin E₁ (PGE₁) into the main pulmonary artery, both lungs were flushed with 60 ml/kg of cold (4° C) preservation solution from a height of 30 cm, over an average of 5 min. The ascending aorta was clamped and sectioned to decompress the left side of the heart. Superior and inferior venae cavae were then ligated. The heart was extracted, leaving in situ the posterior wall of the left atrium. After transection of the triangular ligaments and the trachea, both lungs were extracted in a single block. Before clamping the trachea, the lungs were inflated with 100% oxygen at a pressure of 30 cm H₂O to avoid atelectasis.

The lungs were separated and prepared ex vivo. Polyvinyl chloride catheters were placed into the trachea for the right lung (interior diameter [i.d.], 9 mm) and into the left mainstem bronchus for the left lung (i.d., 8 mm). Another catheter (i.d., 4 mm) was placed into the pulmonary artery of each lung. The pulmonary veins were largely opened. The trachea or the mainstem bronchus was clamped with the lung in midinflation.

Experimental Groups

The animals were randomly assigned to three groups. The EC lungs (n = 8) were flushed and preserved with modified Euro-Collins solution, the BL lungs (n = 7) with a leukocyte-depleted blood solution, and the LPD lungs (n = 7) with low-potassium dextran solution. The composition of each preservation solution is described in Table 1. One of the lungs, alternatively left or right, was reperfused immediately without preservation and was considered a nonischemic matched control in each animal (control lungs). The contralateral lung was stored at 4° C for 8 h, immersed in the flush solution in a plastic bag (preserved lungs).

Reperfusion Technique

The lung was suspended by the tracheal (or bronchial) cannula within a warmed (37° C) and humidified Plexiglas chamber. The lung was ventilated (Servo Ventilator 900 C; Siemens-Elema) with an inspired oxygen concentration of 100%, a tidal volume of 1 ml/g lung weight, a respiratory rate of 10 breaths/min, and a PEEP of 10 cm H₂O.

The circuit was primed with heparinized autologous blood and dextran 70 solution (Macrodex 6% in glucose 5%; Braun Medical AG) (hematocrit 15-18%). The Na⁺, K⁺, and Ca²⁺ concentrations and the pH of the perfusate were maintained in physiologic ranges. A pediatric membrane oxygenator (Lilliput D901; Dideco, Mirandola, Italy) upstream of the pulmonary artery, gassed with N₂ and CO₂, was used to obtain normal values of PO₂ (35-40 mm Hg) and PCO₂ (40-50 mm Hg) in the pulmonary artery.

The lung was then perfused via the pulmonary artery cannula from an arterial reservoir. Effluent blood was drained from the pulmonary veins by gravity in a venous reservoir. The first 100 ml perfused through the pulmonary vessels was eliminated from the perfusion circuit to avoid high levels of K⁺ in the perfusate after reperfusion of the EC lungs. A peristaltic pump (Ismatec pump; Glattburg, Zurich, Switzerland) was used to recirculate the perfusate from the venous to the arterial reservoir. The level in the arterial reservoir was maintained precisely by adjustment of the pump speed, in order to keep the pulmonary arterial perfusion pressure constant at 25-28 mm Hg. The circuit tubing was water jacketed to maintain the perfusate temperature between 36 and 38° C.

Functional Assessment

After a 20-min stabilization period, gas exchange, pulmonary vascular resistance, and pulmonary mechanics were assessed. Calibrated pressure transducers (model 156-PC 06-GW2; Honeywell, Zurich, Switzerland) were used to measure pulmonary arterial pressure (Ppa) and airway pressures (PI_max [peak inspiratory pressure], Pplat [plateau pressure], and PEEP). Pulmonary artery blood flow (PABF) was assessed by a transit-time flow meter (T101CDS; Transonic Systems, Ithaca, NY). The inspiratory flow in the airways (aw, I) and the tidal volume (VT) were measured by a pneumotachograph (model 17212; Gould Godart, Bilthoven, The Netherlands). These measurements were continuously recorded during the 90-min reperfusion period by an informatic system (DaqSys application; Centre Médical Universitaire, Geneva, Switzerland).

Every 20 min (designated throughout from T₂₅ to T₈₅), pulmonary vascular resistance (RL), inspiratory airway resistance (Raw, I), and dynamic lung compliance (CL, dyn) were calculated according to the following formula: RL (dyn · s/cm⁵) = 79.9 × [Ppa (mm Hg)/PABF (ml/min)] × 1,000; pulmonary veins being largely opened, pulmonary venous pressure was assumed to be zero; PABF was measured at end expiration. Raw, I (cm H₂O · s/L) = {[Pplat (cm H₂O)  PEEP (cm H₂O)]/aw, I (ml/s)} × 1,000; inspiratory flow (aw, I) was measured at midtidal volume. CL, dyn (ml/cm H₂O) = VT (ml)/[PI_max (cm H₂O)  PEEP (cm H₂O)].

In addition, blood gases from the pulmonary effluent were analyzed (Radiometer ABL 505, Copenhagen, Denmark) 25, 40, 55, 70, and 80 min after the start of reperfusion to determine the oxygen partial pressure in "arterialized" blood (Pa_O2).

Wet Weight Gain and Protein Permeability

The lungs were weighed before and after reperfusion and the difference between these two values represented the wet weight gain during reperfusion.

At the end of the 90-min reperfusion period, a bronchoalveolar lavage was performed. A flexible fiberoptic bronchoscope was introduced into the bronchial tree and wedged in the lower lobe. Physiologic saline solution (three 20-ml volumes) was slowly infused and the effluent was gently aspirated. After sample centrifugation, total protein concentration (alveolar proteins) was measured by spectrophotometry (Bio-Rad protein assay; Bio-Rad Laboratories, Hercules, CA). Simultaneously, samples were taken from the perfusate at the end of the reperfusion to determine the total plasma protein concentration (arterial proteins) by the Biuret method (automated analysis, Hitachi system 717; Boehringer GmbH, Mannheim, Germany). The alveolar/arterial protein concentration ratio was then calculated.

To detect increased lung microvascular permeability to proteins, we used the double radionuclide method, involving the labeling of circulating transferrin (MW 76,000 D) and red blood cells. This technique, first described by Gorin and coworkers (13) and modified by Basran and colleagues (14), has been shown to be specific and sensitive for the determination of vascular permeability (15). Red blood cells of the perfusate were labeled with technetium-99m (^99mTc; physical half-life, 6 h) by the following method: 5 ml of a "cold" solution containing 11.9 mg of sodium pyrophosphate and 3.4 mg of stannous chloride pyrophosphate was added to the circulating perfusate after the start of lung reperfusion. After 10 min, a 5-ml perfusate sample was drawn to be labeled with 200 µCi of ^99mTc-pertechnetate (obtained from a sterile generator [Elumatic III, CIS, or Ultratechnekow FM; Mallinckrodt, St. Louis, MO]). After a 5-min incubation, the labeled sample was injected into the perfusion circuit. Transferrin of the perfusate was then labeled by the adjunction of 100 µCi of indium-111/chloride (¹¹¹In; physical half-life, 67.4 h) (carrier-free cyclotron produced isotope, obtained by proton irradiation of cadmium; Mallinckrodt) to the perfusion system. Two scintillation  detectors were placed into the Plexiglas chamber containing the perfused lung, one facing the upper lobe and the other facing the lower lobe. The detector consisted of a collimated sodium iodide crystal (Oakfield Instruments, Oxford, UK). Each probe was connected on a separate channel to a personal computer, and the radioactivity was counted on line in counts per second (Mediscint program; Oakfield Instruments) during a 60-min period after the injection of ¹¹¹In. A dual-channel analyzer enabled the ^99mTc and ¹¹¹In counts to be detected simultaneously and recorded separately. Corrections were made for background activity, spillover of ¹¹¹In into the ^99mTc window, and physical decay. In accordance with the literature, the transferrin leak index (TLI), representing the transvascular transferrin flux, is determined from the following ratio plotted against time: [¹¹¹In lung counts (counts/s)/¹¹¹In blood counts (counts/s)]/[^99mTc lung counts (counts/s)/^99mTc blood counts (counts/s)] (14). Because the perfusion circuit was a closed system, no biotransformation of labeled red blood cells or transferrin was expected. Consequently, we could assume that the perfusate activity was constant over the 60-min acquisition period and the following simplified ratio was therefore considered: [¹¹¹In lung counts (counts/ s)/^99mTc lung counts (counts/s)]. This ratio was calculated for each 10-s counting period after a 20-min stabilization period. The transferrin leak index was then calculated, using linear regression analysis, from the slope of the radioactivity ratio.

Tissue Preparation for Light Microscopy

At the end of the 90-min reperfusion procedure, samples of upper and lower lobes were fixed in Formol and embedded in paraffin. The sections were stained with hematoxylin-eosin and viewed by light microscopy.

Statistical Analysis

The results are expressed as means ± SEM. A two-way analysis of variance (ANOVA) was used to compare data at each time point between and within groups; when significance was obtained, the mean values were compared by the Fisher protected least square deviation PLSD test. Significance was considered for p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ischemic Time

Mean total ischemic time was similar in control (EC, 64 ± 7 min; BL, 55 ± 3 min; LPD, 52 ± 3 min; p = NS) and preserved lungs (EC, 482 ± 14 min; BL, 465 ± 18 min; LPD, 469 ± 4 min; p = NS).

Oxygenation

In control lungs, Pa_O2 was similar in each group after the initial stabilization period (T₂₅, Figure 1, top left). Pa_O2 then increased significantly in both BL and LPD control lungs, whereas Pa_O2 did not change significantly in EC control lungs. At T₅₅, T₇₀, and T₈₀, Pa_O2 was higher in BL and LPD control lungs than in EC control lungs. In preserved lungs (Figure 1, top right), oxygenation improved significantly in BL and LPD groups, whereas no significant increase was observed in the EC group. In addition, Pa_O2 values were significantly higher in LPD- than in EC-preserved lungs at the end of the reperfusion period.

View larger version (31K): [in this window] [in a new window]   Figure 1.   Oxygen partial pressure in "arterialized" blood (PaO2; top) and pulmonary vascular resistance (RL; bottom) during reperfusion of control (left) and preserved lungs (right) with Euro-Collins (open squares, EC lungs, n = 8), leukocyte-depleted blood (solid circles, BL lungs, n = 7), or low-potassium dextran solution (solid triangles, LPD lungs, n = 7). Data are expressed as means ± SEM. *p < 0.05 versus EC lungs. p < 0.05 versus LPD lungs. §p < 0.05 versus baseline value at T25.

Pulmonary Hemodynamics

RL remained stable in BL and LPD control lungs (Figure 1, bottom left). In the EC control lungs, RL was higher at T₂₅ and decreased over time without reaching significance. In preserved lungs (Figure 1, bottom right), RL values remained stable over time in all three groups, but were significantly and persistently lower in BL lungs compared with EC or LPD lungs.

Ventilatory Mechanics

In BL and LPD control lungs, Raw, I remained constant and CL, dyn increased significantly during reperfusion. In contrast, Raw, I increased significantly over time and CL, dyn remained stable in EC control lungs (Figure 2, left). A similar evolution over time was also observed in preserved lungs. Raw, I progressively increased and CL, dyn progressively decreased in EC lungs, while these variables did not change over time in the two other groups (Figure 2, right).

View larger version (26K): [in this window] [in a new window]   Figure 2.   Airway resistance (Raw, I; top) and dynamic lung compliance (CL, dyn; bottom) during reperfusion of control (left) and preserved lungs (right) with Euro-Collins (open squares, EC lungs, n = 8), leukocyte-depleted blood (solid circles, BL lungs, n = 7), or low-potassium dextran solution (solid triangles, LPD lungs, n = 7). Data are expressed as means ± SEM. §p < 0.05 versus baseline value at T25.

Wet Weight Gain

At the end of the 90-min reperfusion period, EC-preserved lungs presented a wet weight gain of about twice that of BL- and LPD-preserved lungs (Figure 3, top). This was also noticed under control conditions. There were no significant differences in wet weight gain between control and preserved BL or LPD lungs.

View larger version (24K): [in this window] [in a new window]   Figure 3.   Wet weight gain (top) and alveolar/arterial protein concentration ratio (bottom) after 90 min of reperfusion of control (open bars) and preserved lungs (hatched bars). Data are expressed as means ± SEM. EC = modified Euro-Collins solution; BL = leukocyte-depleted blood solution; LPD = low-potassium dextran solution. *p < 0.05 versus EC-preserved lungs. p < 0.05 versus control lungs of the same treatment group. §p < 0.05 versus EC control lungs.

Protein Permeability

Alveolar/arterial protein ratio. The alveolar/arterial protein concentration ratio (Figure 3, bottom) was similar in the three control groups. In contrast, after 8-h preservation, the ratio was fourfold higher in EC-preserved lungs than in EC control lungs, whereas the ratio did not change in BL- or LPD-preserved lungs.

Transferrin leak index. Under control and preserved conditions, the TLI was 4- to 10-fold higher in EC than in BL or LPD lungs (Table 2). The increase in TLI was as important in control as in preserved EC lungs. In contrast, the TLI was low in the BL and LPD control groups, and did not increase after 8 h of preservation.

Morphologic Changes

Qualitative light microscopic examination of control lungs flushed with either EC, BL, or LPD solution showed a well preserved pulmonary structure (Figure 4, left). The alveolar walls were thin, with only mild peribronchovascular edema. The epithelium lining bronchi and bronchioles was intact. Small groups of polymorphonuclear leukocytes and platelets were observed in peribronchovascular capillaries. In contrast, lungs preserved for 8 h with EC solution showed major alterations (Figure 4, right). In many places, the alveolar walls were thickened and infiltrated with numerous inflammatory cells. Severe intra-alveolar hemorrhage was often observed. In lungs preserved with BL or LPD solutions changes were much less severe. Scanty polymorphonuclear leukocyte infiltrates, interstitial edema and dilated capillaries were found, but no intraalveolar hemorrhage was seen.

View larger version (105K): [in this window] [in a new window]   Figure 4.   Photomicrographs of control lungs (left) and preserved lungs (right) with modified Euro-Collins solution (EC; A and B), leukocyte-depleted blood solution (BL; C and D), or low-potassium dextran solution (LPD; E and F ), after the 90-min reperfusion period. Control lungs flushed with either EC, BL, or LPD solution showed a well-preserved structure (A, C, and E ): the alveolar walls were thin, with only mild peribronchovascular edema; small groups of polymorphonuclear leukocytes and platelets were seen in peribronchovascular capillaries. In contrast, lungs preserved for 8 h with EC solution showed major alterations (B): the alveolar walls were thickened and infiltrated with numerous inflammatory cells and severe intra-alveolar hemorrhage was often observed. In lungs preserved with BL or LPD solution changes were much less severe (D and F ): scanty polymorphonuclear leukocyte infiltrates, interstitial edema, and dilated capillaries were found, but no intra-alveolar hemorrhage was seen (hematoxylin-eosin stain; A: original magnification, ×100; B-F: original magnification, ×200).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of the present study clearly demonstrate, in this model of porcine isolated reperfused lung, that both BL and LPD flush solutions better preserve the endothelial-epithelial barrier compared with the EC solution. The quality of preservation, assessed by permeability indices and morphologic analysis, was similar for BL and LPD solutions. However, pulmonary vascular resistance was higher in LPD- than in BL-preserved lungs.

During lung transplantation, ischemia/reperfusion may induce an increase in microvascular permeability by injuring the wall of microvessels after activation of neutrophils and production of reactive oxygen species and proteases (16, 17), and by promoting the generation of mediators, such as leukotrienes, thromboxane, and endothelin 1 (18). In the present study, the pulmonary microvascular permeability for transferrin was increased in the EC control and preserved lungs, whereas the TLI was low in BL and LPD control lungs and remained low after preservation. Thus, in control lungs flushed with EC solution, after only short ischemia, the microvascular permeability to transferrin was already increased. Furthermore, in all control lungs, the alveolar/arterial protein ratio was moderately increased. After 8 h of preservation, the ratio remained low in BL and LPD lungs, but significantly increased in EC lungs. Thus, the increased alveolar protein content in EC-preserved lungs demonstrates that an alveolar membrane injury was associated with the endothelial cell injury. Because the alveolar epithelial barrier is more resistant to injury than is endothelium (21), these results confirm that injury was much more severe in EC-preserved lungs than in the other groups. Morphologic analysis of selected lung fragments confirmed these results. In EC-preserved lungs, rupture of alveolar septa and severe alveolar edema and hemorrhage were obvious, whereas BL- and LPD-preserved lungs showed a relatively well-preserved structure. Alveolar edema was only minimal in the three control groups.

Moreover, in EC-preserved lungs, airway resistance increased and dynamic pulmonary compliance decreased over time. In contrast, these variables remained stable in the two other preserved groups. Interstitial edema in the bronchial wall may explain this effect, because the water content, assessed by a high wet weight gain, increased in EC-preserved lungs. To a lesser extent, the increase in airway resistance over time with a wet weight gain also suggests the development of edema in bronchial wall of EC control lungs. Such an evolution was not observed in BL and LPD control lungs. As suggested by morphologic analysis and measurement of alveolar protein content, the edema was essentially limited to the interstitial space in EC control lungs, whereas it was also located in the alveolar space in EC-preserved lungs.

Lung ischemia/reperfusion alters gas exchange by inducing abnormalities in ventilation/perfusion distribution and by limiting alveolar-capillary diffusion. At the end of the reperfusion, Pa_O2 was lower in EC control lungs than in BL and LPD control lungs. After preservation, these differences were paradoxically less prominent, because at the end of the reperfusion, Pa_O2 was only slightly lower in EC lungs than in LPD lungs and similar to that in BL lungs. Although Pa_O2 is generally regarded as a sensitive index of the quality of lung preservation, oxygenation is strongly influenced by the presence of macro- or microatelectasis. In our model of ex vivo-isolated perfused lung, atelectasis is more difficult to overcome during positive pressure ventilation, because the absence of thoracic cage cannot counteract the elastic recoil of the lungs. We thus believe that this limitation renders the Pa_O2 value less sensitive in our experimental model than permeability indices.

A rise in RL is a common feature of lung transplantation. In vivo, it is partly due to the effects of surgical anastomosis and to the denervation of the transplanted organ (22). Ischemia/ reperfusion contributes largely to the increase in RL by different mechanisms such as perivascular edema formation, presence of microthrombi, reduction of potent endothelium-dependent relaxing factors such as prostacyclin and nitric oxide, as well as the generation of vasoconstrictors such as thromboxane and endothelin 1 (18, 20, 23, 24). In the present study, RL was lower in lungs preserved with the leukocyte-depleted blood solution than in lungs preserved with the two other solutions, suggesting a better preservation. However, because the quality of preservation assessed by permeability indices was similar in both BL and LPD lungs, unknown factors present in the blood solution may favorably interact with the control of pulmonary vascular tone and explain the low RL of BL-preserved lungs. Further studies modifying the composition of the blood solution should address the possible mechanisms underlying this finding.

The deleterious effect of EC compared with the two other solutions was not only demonstrated after 8 h of ischemic preservation, but was also observed after a single flush through the pulmonary artery. The high potassium concentration in EC solution is likely to exert a direct or indirect toxic effect on parenchymal cells during ischemic conditions. Indeed, viability of type II alveolar cells is less well maintained with high-potassium solutions than with low-potassium solutions (25). Moreover, potassium-induced cell membrane depolarization generates hydrogen peroxide and other oxidants (26). In addition, during the pulmonary artery flush, the vasoconstriction induced by a high potassium concentration may reduce the homogeneity of the distribution of the preservation solution, and induce endothelial cell injury via increased shear stress as previously suggested (7, 27, 28). Although the vasodilator effect of prostaglandins may reduce pulmonary injury (29), the injection of PGE₁ into the pulmonary artery before the flush of the EC solution was insufficient to prevent the pathophysiologic changes observed.

Several studies have recently investigated the effects of adding components of the prostaglandin/adenosine 3',5'-cyclic monophosphate (cAMP) and nitric oxide (NO)/cyclic guanosine monophosphate (cGMP) pathways to EC solution. In clinical practice, the vasodilatory prostaglandins PGE₁ and PGI₂ (prostacyclin) added to EC, University of Wisconsin, or Wallwork solutions, are widely used. Although PGE₁ and PGI₂ were initially used to enhance distribution of flush solution and accelerate cooling of the lung, it has not been proven that their benefit is due to pulmonary vasodilation during the flush. PGE₁ and PGI₂ may attenuate ischemia/reperfusion injuries by other mechanisms such as inhibition of leukocyte and platelet aggregation and adhesion, decrease of lysosomal enzyme and superoxide anion production by neutrophils, and reduction of the effect of vasoactive mediators on microvascular permeability (23, 30). However, the beneficial effect of the addition of PGE₁ to EC solution remains controversial. Whereas some investigations demonstrate an improvement of the quality of preservation (31), others do not find any benefit (30). There is more evidence in the literature that the addition of PGI₂ or iloprost, a prostacyclin analog, to EC solution improves the quality of preservation (30, 32). Increased microvascular permeability after lung ischemia/reperfusion has been associated with decreased cAMP levels. Agents increasing intracellular cAMP, such as isoproterenol and forskolin through adenylyl cyclase activation, or rolipram and pentoxifylline through phosphodiesterase inhibition, have beneficial effects on ischemia/reperfusion lung microvascular injuries (33). Addition of dibutyryl-cAMP to University of Wisconsin solution has been shown to decrease microvascular permeability induced by ischemia/reperfusion in an isolated rat lung model (34). Canine lung allograft function is improved by adding pentoxifylline to EC solution (35).

Decreased NO availability after reperfusion may contribute to ischemia/reperfusion injuries, because NO inhibits platelet aggregation and neutrophil adhesion by increasing cGMP in effector cells. Several studies have investigated the NO/ cGMP pathway in lung preservation. Interventions in the NO pathway by supplementing EC solution with a cGMP analog (24), or with the NO donors glyceryl trinitrate (36) and nitroprusside (37), improve the quality of preservation. In addition, a beneficial effect resulted from the administration of L-arginine, the nitrogen donor for NO synthesis, during reperfusion of lungs preserved with EC solution (38).

In our opinion, these studies indicate that EC solution is by far not optimal, since numerous pharmacological additives improve significantly the quality of preservation. In addition, in the present study, we demonstrate that EC solution has a deleterious effect per se, because lung injuries were observed not only after 8 h of ischemic preservation, but also after a single flush.

The reasons explaining the superiority of colloidal extracellular solutions over crystalloid intracellular solutions are largely unknown. Several experimental works have demonstrated the superiority of LPD over EC solution (5, 11, 39). As indicated by Keshavjee and coworkers (4), the beneficial effect of LPD solution is attributed to both the presence of dextran and the low potassium concentration. Dextran, by improving the flow in the microcirculation, may reduce the distribution inhomogeneity of the preservation solution during the initial flush (40). Moreover, via its oncotic properties, dextran may contribute to diminish intracellular and interstitial edema formation. Sakamaki and coworkers (39) indicate that the beneficial effect of LPD over EC may be due to a reduced membrane lipid peroxidation after ischemia/reperfusion.

The modified blood solution described by Wallwork is also an extracellular-type low-potassium solution, containing colloids that maintain oncotic pressure (12). The quality of preservation with this solution is demonstrated in an experimental study (8) and in clinical practice (12). Mechanisms for optimal preservation might be partly attributed to the presence of buffer systems and metabolic substrates in the blood. In addition, the presence of inflammation inhibitors such as inhibitors of the complement activation or various cytokine inhibitors may attenuate ischemia/reperfusion injury (41, 42). It should be emphasized that leukocyte depletion may be an important factor for the quality of preservation obtained with the blood solution in our study. Indeed, polymorphonuclear cells are involved in ischemia/reperfusion injury (17), and the addition of whole blood to LPD solution does not improve (6) and may actually diminish the quality of preservation (43).

In conclusion, our results demonstrate the superiority of BL and LPD solutions over EC solution. The quality of preservation, assessed by permeability indices and morphologic analysis, was excellent and similar for BL and LPD solutions. However, pulmonary vascular resistance was higher in LPD- than in BL-preserved lungs. The use of EC solution was clearly deleterious, not only in preserved lungs after 8 h of ischemic preservation, but also in control lungs flushed with the preservation solution. The low cost and the simple preparation of LPD solution could favor its use in clinical practice. However, the decrease in pulmonary vascular resistance with BL solution may be clinically relevant.