INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: lung transplantation; organ preservation solutions; prognosis; reperfusion injury

Lung transplantation has gained widespread acceptance as a therapeutic option in a selected number of patients with end-stage lung diseases refractory to other forms of medical treatment (1, 2). However, overall mortality after lung transplantation is still significant. According to the International Registry, the 1-yr actuarial survival is approximately 70%, with most of the deaths occurring within 30 d after transplant (3). The main cause of death during the first postoperative month is ascribed to nonspecific graft failure (2, 4). Early graft failure after lung transplantation has been given various names such as reimplantation edema, reperfusion edema, primary graft failure, or allograft dysfunction. The diagnosis rests on the occurrence of noncardiogenic pulmonary edema with gas exchange impairment associated, in the severe cases, with hemodynamic failure (7). The predominant mechanism underlying early graft failure is now considered to be ischemia/reperfusion (I/R) injury and, for that reason, the quality of lung preservation seems to be a key determinant of initial graft function. In clinical transplantation, most centers now use flush perfusion of the donor lung with a preservation solution. However, the composition of the latter differs markedly among centers (8). Much work has been performed on the prevention of lung I/R injury, mainly by comparing the effects of different lung preservation solutions. Despite a considerable amount of data about experimental lung transplantation, few clinical data are available. The aim of this retrospective study was to assess the influence on early graft dysfunction of (1) different preservation solutions and (2) extracellular- and intracellular-type solutions.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

We retrospectively reviewed the medical records of all patients who underwent lung transplantation in two centers (Hospital Beaujon [Clichy, France] and Hospital Foch [Suresnes, France]) between 1988 and 1999 and in whom donor lung was preserved by the flush perfusion technique.

Transplantation Procedure

Donor selection was based on widely accepted guidelines (9). The single flush technique was performed with four preservation solutions (Table 1). Two of them were of intracellular type (modified Euro-Collins and University of Wisconsin [UW]) whereas two solutions were of extracellular type (Cambridge and Celsior). The composition of these different preservation solutions has already been described (10). The technique consisted of hypothermic (4° C) flush perfusion of the main pulmonary artery solution for approximately 4 min to a total volume of 50 ml/kg. Thereafter, the graft was excised and stored in a sterile plastic bag filled with saline at 4° C. 

In most of the cases (90%), prostaglandin E₁ or prostaglandin I₂ was administered into the superior vena cava before aortic clamping, the dosage being gradually increased according to blood pressure tolerance.

Single lung transplantation was performed by a classic technique (14), whereas bilateral lung transplantation was a bilateral sequential operation (15). Cardiopulmonary bypass was used intraoperatively in eight patients (Table 1).

Immunosuppression

Intravenous methylprednisolone and azathioprine were administered preoperatively. Postoperative initial immunosuppression consisted of cyclosporine, azathioprine, and corticosteroids.

Recipient Analysis

On the basis of the different types of preservation, the four groups of patients were compared. Moreover, the patients were classified according to the extra- or intracellular composition of the preservation solution. We retrospectively compared the early postoperative course of the different groups with respect to I/R injury-related complications by analyzing the medical records of all patients. In particular, we compared the incidence of reimplantation edema, the duration of mechanical ventilation, the Pa_O2/fraction of inspired O₂ (FI_O2) ratio during the first 3 d after surgery, and the 1-mo survival in the different groups of patients. Pulmonary reimplantation edema was diagnosed when the three following criteria were met: (1) diffuse alveolar infiltrate involving the lung allograft developing within the first 3 d of transplantation; (2) Pa_O2/FI_O2 ratio < 200 mm Hg; and (3) no evidence of bacterial infection or rejection. Gas exchange during the early postoperative period was assessed by retrieving in the medical records of each patient the worst Pa_O2/FI_O2 ratio recorded within the first 3 d after lung transplantation.

Statistical Analysis

Results are expressed as the mean ± standard deviation for quantitative variables. Comparisons of continuous and categorical variables were made by using the unpaired Student t test and Fisher exact test, respectively. Multiple comparisons between continuous variables were made by analysis of variance with the Newman-Keuls test as the post hoc test.

To avoid bias linked to potential confounding variables, we used multiple logistic regression analysis to assess the effect of the different preservation solutions and of their intra- or extracellular type on edema formation and 1-mo mortality. In a first step, we used univariate analysis to identify potential confounding variables associated with edema occurrence and early death. All significant or nearly significant variables (p < 0.1) were included in the multiple logistic regression analysis. All p values are two tailed and p values of less than 0.05 were considered to indicate statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

During the study period, 234 patients underwent a single or double lung transplantation. Because patients with severe pulmonary hypertension may experience perioperative complications mimicking those related to I/R injury in the case of single or double lung transplantation, we excluded from the analysis all recipients transplanted for this indication (n = 11). Since January 1998, all recipients (n = 11) were systematically given inhaled nitric oxide and pentoxifylline during graft reperfusion at Hospital Beaujon in order to avoid complications related to I/R injury (16). Given the potential benefit offered by this strategy, these patients were also excluded from analysis. Forty-two patients were also excluded for the following reasons: donor lung preserved by topical cooling (n = 36), donor lung preserved with Bretschneider solution, given the low number of preservation procedures performed with this solution (n = 6). Thus, 64 patients were excluded from analysis. All the remaining patients (n = 170) were eligible for analysis. The type of transplantation procedure was a single (n = 124) or bilateral (n = 46) lung transplantation. The mean age of the 170 patients was 45.2 ± 13.4 yr. The other characteristics of these patients are described in Table 1. The 4 groups of patients were as follows (Table 1): Euro-Collins (n = 61), UW (n = 24), Cambridge (n = 64), and Celsior (n = 21). The four groups were comparable in terms of recipient age, underlying recipient disease, donor Pa_O2/FI_O2 ratio, and cause of brain death in the donor. The graft ischemic times for the different groups were not comparable (Table 1). In the Celsior group, the mean graft ischemic time was significantly longer than in the other groups. When the patients were classified according to the extra- or intracellular composition of the preservation solution (extracellular type, n = 85; intracellular type, n = 85), graft ischemic time was significantly shorter with intracellular solutions than with extracellular solutions (240 ± 93 vs. 286 ± 111 min, p < 0.005). With regard to the type of transplantation procedure, the four groups were different, double lung procedures being more frequently performed when extracellular solutions were used.

Reimplantation edema occurred in 48% of all patients, and the 1-mo survival rate was 84%. From 1988 to 1999, the trends in the incidence of edema and 1-mo survival remained constant (Figure 1). No significant difference in the incidence of edema was observed between the four groups despite a tendency to a lower incidence in the Cambridge and Celsior groups (Table 2). Similarly, the incidence of edema with extracellular and intracellular solutions was similar (Table 3). The worst Pa_O2/FI_O2 ratio within the first three postoperative days was significantly lower in the Euro-Collins group than in the Celsior group (Table 2), and was significantly higher (225 ± 131 vs. 173 ± 105 mm Hg, p < 0.05) with the extracellular solutions than with the intracellular solutions (Table 3). The duration of mechanical ventilation and the 1-mo survival rate did not differ significantly between the different groups and between the extra- and intracellular solutions (Tables 2 and 3).

View larger version (14K): [in this window] [in a new window]   Figure 1.   To test the presence of a learning effect in our series, the 170 patients of the study, who were transplanted from 1988 to 1999, are divided into four consecutive cohorts. Each cohort is composed of 42 or 43 consecutive patients. The 1-mo survival and the incidence of reimplantation edema in these four cohorts are represented. Over time, we did not observe a higher 1-mo survival or a lower incidence of edema.

Using univariate analysis, among the different variables studied (recipient age, sex, indication for lung transplantation, use of cardiopulmonary bypass, graft ischemic time, transplantation procedure, transplantation year, donor gas exchange, and transplantation center), only graft ischemic time was found to be associated with edema formation and 1-mo mortality. In the multivariate model, compared with Euro-Collins solution, the use of Cambridge solution, and the extracellular type of preservation fluid, were associated with a lower incidence of reimplantation edema (odds ratio [OD] = 0.44, p = 0.04 and OR = 0.43, p = 0.02, respectively) whereas the graft ischemic time was associated with an increased incidence of edema (OR = 1.3, p = 0.02) (Table 4). Using Celsior solution, despite a similar odds ratio compared with Cambridge, statistical significance was not reached. With regard to 1-mo mortality, only graft ischemic time was significantly associated with higher mortality (OR = 1.3, p = 0.02) (Table 5).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study shows that flush perfusion of donor lung with a preservation solution of extracellular composition is associated with a reduced incidence of pulmonary reimplantation edema after clinical lung transplantation, without demonstrable effects on 1-mo mortality. With regard to edema occurrence, despite similar associated odds ratios, statistical significance was observed in the Cambridge group but not in the Celsior group. In our opinion, it does not infer that Celsior is inferior to Cambridge in the prevention of early graft dysfunction. This discrepancy could be explained by the lower number of patients in the Celsior group. The lack of demonstrable effect of the composition of the preservation fluid on 1-mo mortality could be explained by the fact that incidence of mortality is much less frequent than incidence of edema. Thus, the number of patients required in order to detect statistical significance should be higher. Our data also underline the role of graft ischemic time in the occurrence of edema and in 1-mo mortality. Because graft ischemic time was not evenly distributed between groups, the control of this confounding factor by mean of multivariate analysis was mandatory. The difference in ischemic time between the groups could be ascribed to the higher proportion of double lung procedures in the extracellular group.

Many complications may occur during the early postoperative period after lung transplantation, accounting for the 10- 20% perioperative mortality still observed in most centers (1, 2). These complications are mainly related to infection, hemorrhage, acute rejection, bronchial ischemia, or I/R injury. In particular, the complications related to I/R injury are reimplantation edema and less frequently early hemodynamic failure, both complications being responsible for substantial morbidity or mortality (4). Thus, many centers worldwide have concentrated efforts on prevention strategies of lung I/R injury, mainly by use of experimental models. These models include lung transplantation in various animals such as rabbits, rats, dogs, or pigs and also models of isolated-perfused lungs. From such studies, I/R injury was demonstrated to be secondary to an influx of activated polymorphonuclear neutrophils (PMNs) within the lungs with subsequent release of cytotoxic mediators including reactive oxygen species (17).

The influence of graft ischemic time on reimplantation edema has already been demonstrated in a rat model of lung transplantation (18), where edema formation increased in proportion to the duration of graft ischemia. Controversial results have been observed in clinical transplantation. A negative influence of graft ischemic time on early graft function was found by several groups including ours (6, 19, 20), but others have not reported similar results (4, 5, 21, 22). The discrepancy between previous results and our data may be explained by the high statistical power of our study. The deleterious effect of a long graft ischemia on survival was pointed out by Snell and coworkers (23) and another study by Novick and colleagues showed that the interaction of graft ischemic time and donor age was a significant predictor of 1-yr mortality whereas ischemic time per se was not a risk factor (24). The fact that in our study we considered the 1-mo mortality instead of the 1-yr mortality may explain the discrepancy with the work of Novick and colleagues.

Using experimental models as outlined above, research on the prevention of I/R injury has mainly consisted of testing the benefit of several agents administrated at the time of reperfusion and of comparing the effects of different preservation solutions during harvesting (2, 25). In the studies comparing the effects of different preservation solutions on lung I/R injury, quality of preservation was assessed by lung function parameters (gas exchange, ventilatory mechanics), lung water content, and histopathological changes (10, 12, 26). More recently, assessment of lung microvascular permeability by an isotopic method (11) or by evaluation of lipid peroxidation (31) was used. The extracellular-type solutions were mainly low-potassium dextran (LPD) (11, 26, 28), Celsior (10, 12), or less frequently blood-based solutions (including the Cambridge solution) (10, 11, 26, 29, 30) whereas intracellular-type solutions consisted most often of Euro-Collins and UW (11, 12, 26, 31). Overall, in these animal studies, extracellular-type solutions achieved a better preservation than Euro-Collins or UW (11, 12, 26, 31).

In clinical lung transplantation, to the best of our knowledge, only two retrospective studies compared the effects of different preservation solutions on I/R injury. Hardesty and coworkers (32) compared the early postoperative results of lung transplantations in which preservation was based on flush perfusion of two intracellular solutions (Euro-Collins and UW solutions). The UW group consisted of 70 grafts whereas the Euro-Collins group consisted of 30 grafts. The authors found that reperfusion injury identified by radiograph on Day 1 was more severe in the Euro-Collins group but failed to detect any difference between the two groups in terms of gas exchange and early mortality. More recently, Muller and coworkers retrospectively compared the effect of flush-perfused LPD and Euro-Collins solutions on I/R injury in 80 lung transplantation recipients (33). Despite a significantly longer duration of ischemia, incidence and severity of reperfusion injury as well as the 30-d mortality were lower in the LPD group without achieving statistical significance. The only parameter associated with statistical significance was the alveolar-arterial oxygen gradient at admission to the intensive care unit, which was in favor of the LPD group. This may have resulted from the low number of patients accounting for the absence of a clear difference in favor of the extracellular-type LPD solution. We have previously compared the early postoperative course according to the type of preservation, but failed to detect a difference (34). For that reason, we decided to pool data from two centers.

The reasons for the apparent superiority of the extracellular-type preservation solutions in both experimental and clinical studies are not well understood. High potassium content of intracellular-type solutions may have deleterious effects in lung preservation. During pulmonary artery flush, vasoconstriction induced by high potassium content may alter homogeneity of distribution of the preservation solution (35). In addition, high potassium content may have direct toxic effects on Type II alveolar cells (36) or affect chemotaxis of PMNs, which are the cornerstone cells in I/R injury (37).

There are limitations to our study. First, the design of the study is retrospective. Thus, introduction of a bias is not excluded. Because the different solutions have been used in a sequential way, the superior results observed with extracellular-type solutions may reflect a training curve. However, the use of intracellular-type solution was not limited to the beginning of our transplantation programs (UW solution was used up to 1998), whereas the first introduction of extracellular-type solutions was January 1992. When the 1-mo survival and the incidence of reimplantation edema were analyzed in both centers, a learning curve effect was not observed (Figure 1). Moreover, transplantation year was linked neither with occurrence of reimplantation edema nor with 1-mo survival in the univariate analysis. This finding does not support the role of a learning effect in our study. A prospective, randomized, comparative study comparing the effects of different preservation solutions on the outcome of lung transplantation would have been more appropriate. Given the scarcity of lung donors in France, such a study would have been difficult to carry out, even on a multicenter basis. Second, even if the number of patients included in the study is important, it is likely that we lacked statistical power to detect the differences between the four groups.

In conclusion, a significant negative relation between graft ischemic time and both edema occurrence and 1-mo survival was demonstrated. Furthermore, our data indicate that despite the absence of a demonstrable effect on 1-mo mortality, extracellular-type solutions were associated with better lung preservation than intracellular-type solutions in clinical lung transplantation. These results are in agreement with most of the experimental studies. Among the extracellular-type solutions, the superiority of one solution over the other is not clear. Further studies are needed to determine which of these extracellular-type solutions achieves the best lung preservation.