Cytovene

Study objectives: To determine the causes of death in patients dying within 30 days after lung transplantation at the University of Florida, to assess the importance of several diagnostic modalities for determining the causes of their decline, and to construct an algorithm for the evaluation of patients with severe respiratory compromise occurring early after lung transplantation.

Design: Retrospective review of medical records and pathology slides from all patients dying within 30 days after lung transplantation, and biopsy specimen diagnoses from all lung allograft recipients at the University of Florida.

Patients: Nine deaths occurred during the first 30 days after transplantation among 117 patients undergoing 123 isolated lung transplantation operations.

Results: Infections accounted for the greatest number of deaths (bacterial pneumonia, four patients; catheter-related bacteremia, one patient). Persistent pneumonia confirmed by biopsy specimen was usually accompanied by histologic manifestations of acute cellular rejection and was associated with poor patient outcome (ie, death or subsequent development of bronchiolitis obliterans syndrome). In two patients, antibody-mediated rejection either was the immediate cause of death (hyperacute rejection, one patient) or preceded a fatal case of pneumonia (accelerated antibody-mediated rejection, one patient). Other causes of death included hypoxicischemic encephalopathy secondary to an intraoperative cardiac arrest (one patient), pulmonary venous thrombosis with bacterial colonization of the thrombotic material (one patient), and ischemic reperfusion injury (one patient). In most patients, more than one type of diagnostic technique was needed to ascertain the cause of the catastrophic decline.

Conclusions: The causes of early posttransplant death in our patient group included infections, antibody-mediated rejection, hypoxic-ischemic encephalopathy secondary to cardiac arrest, pulmonary venous thrombosis, and ischemic reperfusion injury. Because these processes often demonstrate overlapping clinical and morphologic features requiring multiple diagnostic techniques for resolution, a systematic multimodality approach to diagnosis is advantageous for determining the causes of decline in individual patients and for estimating the incidences of the different causes of early graft and patient loss in the lung transplant population. (CHEST 2001; 120:225-232)

Key words: antibody-mediated rejection; death; early hyperacute rejection; lung transplantation; mortality; primary allograft failure

Abbreviations: ECMO = extracorporeal membrane oxygenation; HLA = human leukocyte antigen

Although the rate of patient survival beyond the first 30 days after lung transplantation has improved, early postoperative death remains an important complication for many lung transplant recipients. Survival curves for lung allograft recipients show their sharpest declines during the first 30 days after transplantation.[1] The 30-day mortality after lung transplantation is currently approximately 10%,[1] and the majority of deaths in this early period are caused by infections and primary allograft failure.[2] Among patients who die because of infections, Gram-negative pneumonia and sepsis are cited as the most frequent direct causes of death.[3-6]

The clinical syndrome of primary graft failure consists of progressive hypoxemia, decreased lung compliance, and diffuse interstitial and alveolar infiltrates developing rapidly after transplantation, visible on a chest radiograph.[7,8] Although ischemic reperfusion injury probably accounts for many examples of this syndrome, antibody-mediated rejection, acute cellular rejection, pneumonia, aspiration, volume overload, and venous anastomotic obstruction can display similar physiologic and radiographic abnormalities. Systematic appraisal for these potential causes of catastrophic decline would ideally employ a diagnostic algorithm geared to assess for each consideration. In particular, underdiagnosis of antibody-mediated rejection is likely unless a sensitive lymphocyte cross-match can be obtained. Our earlier analysis of patients with severe early pulmonary graft dysfunction revealed that low-level preformed antibodies directed against class II, and perhaps against class I, human leukocyte antigen (HLA) antigens were a risk factor for severe early pulmonary graft dysfunction with the potential for progression to death.[9]

In this study, we sought to determine the causes of early deaths in lung transplant recipients at the University of Florida (Gainesville, FL). Clinical, radiologic, microbiological, immunologic, and morphologic information varied in importance for individual patients, with most final diagnoses requiring the input of at least two types of data. Our results suggest that the use of a multimodality algorithm for the assessment of early graft dysfunction may accelerate the determination of the etiology of dysfunction in individual patients and, ultimately, enhance the promptness and potential effectiveness of therapeutic interventions.

MATERIALS AND METHODS

Between March 1994 and August 1999, 117 patients underwent 123 isolated lung transplant operations (single lung, 81 operations; bilateral, 42 operations) at the University of Florida. Eight lung transplant recipients underwent retransplant operations, six with lungs only, and two with combined heart-lung grafts. One hundred three patients were adults (age range, 18 to 64 years), and 14 were children and adolescents (age range, 4 to 17 years). Recipients included 67 men (57.3%) and 50 women (42.7%). Before transplantation, screening for preformed T-cell IgG anti-HLA class I antibodies was performed either by flow cytometry using a 10-cell panel or with latex beads coated with purified class I and class II HLA antigens (One Lambda; Canoga Park, CA).

All patients were treated with triple immunosuppression consisting of cyclosporine A (Sandimmune or Neoral; Novartis Pharmaceuticals; East Hanover, NJ) or taerolimus (Prograf; Fujisawa Healthcare; Deerfield, IL), prednisone (Roxane Laboratories; Columbus, OH), and azathioprine (Imuran; Glaxo Wellcome; Research Triangle Park, NC) or mycophenolate mofetil (CellCept; Roche Pharmaceuticals; Nutley, NJ). Induction with OKT3 (Orthoclone OKT3; Ortho Biotech; Raritan, NJ) was performed only with the first 38 lung transplant recipients. After hospital discharge, the therapy of several pediatric patients was changed from cyclosporine A to tacrolimus. Pediatric patients undergoing transplantation in 1999 and later also received an interleukin 2 receptor antagonist, either daclizumab (Zenapax; Roche Pharmaceuticals) or basiliximab (Simulect; Novartis Pharmaceuticals) as part of the induction. Prophylaxis against bacterial pneumonia consisted of at least 5 days' treatment with IV clindamycin (Cleocin; Pharmacia & Upjohn; Peapack, NJ) and ceftazidime (Ceptaz or Fortaz; Glaxo Wellcome) or cefepime (Maxipime; Dura Pharmaceuticals; San Diego, CA). In addition, antibiotics given to several patients were adjusted as appropriate in accordance with the results of their preoperative sputum cultures. IV ganciclovir (Cytovene; Roche Pharmaceuticals) was administered as antiviral prophylaxis when either the donor or the recipient was seropositive for cytomegalovirus. If ganciclovir was not administered and patients had serologic evidence of herpes simplex virus infection, oral acyclovir (Zovirax; Glaxo Wellcome) prophylaxis treatment was administered.

A flow cytometry cross-match was performed in all patients within the first 24 h after transplantation, and a cross-match by cytotoxicity was performed in selected patients. Details about these procedures can be found in a previous report.[9] Bronchial washings were routinely obtained from the allograft at the time of reperfusion, and in the first 48 h after reperfusion, and were sent for bacterial, viral, mycobacterial, and fungal cultures. Additional bronchial washing fluid, blood, pleural fluid, peritoneal dialysis fluid, and allograft tissue were obtained and sent for studies, when indicated, based on the clinical findings. For patients who died, surgical pathology and autopsy materials were retrieved from the files of the University of Florida and were reviewed by one of the authors (D.S.Z.). Autopsy slides from patients 1, 5, and 9 also were reviewed by a second author (W.H.D.).

RESULTS

There were nine deaths in the first 30 days after the 123 lung transplant operations performed at the University of Florida. For each patient who died during the first 30 days posttransplantation, Table 1 lists demographic information, type of lung allograft, indication for transplantation, and posttransplant survival interval. The number of male patients (6 of 9 [66.7%]) exceeded the number of female patients, and bilateral transplant operations are disproportionately represented in this early mortality group (6 of 9 patients [66.7%] compared to all patients who received lung transplants (42 of 123 patients [34.1%]. All four patients who received transplants for pulmonary veno-occlusive disease died because of infections within the first 30 days.

The suspected causes of death, contributory antemortem and postmortem microbiology and immunology test results, and morphologic findings are presented in Table 2. An assessment of the significance of these results for determination of the final diagnosis is illustrated in Table 3. Among the nine patients who died during the first 30 days posttransplantation, five died because of a bacterial infection (bacterial pneumonia, four patients; catheter-related bacteremia and septic shock, one patient). In one of these patients, the pneumonia clearly was donor transmitted, as indicated by the histology and culture results from a portion of donor lung submitted at the time of transplantation. In the other three patients, the pneumonia was more likely ventilator associated and acquired after transplantation. In an additional patient, bacterial colonization of thrombotic material at a pulmonary venous anastomosis site contributed to the patient's death. During the first 24 h after transplantation, this last patient developed pulmonary edema (right worse than left) that did not respond to diuresis over the first 4 days. Transesophageal echocardiography showed flow in the pulmonary veins, but visualization of the inferior pulmonary veins was suboptimal, especially the right lower pulmonary vein. The next 4 days were marked by fevers, worsening bilateral infiltrates, and shock. BAL and pleural cultures yielded Pseudomonas aeruginosa. Autopsy showed bilateral pulmonary venous thrombi at the venous anastomoses, with colonization by bacteria.

During the first 30 days after transplantation, biopsy specimens from nine patients (7.7% of total) showed acute bronchopneumonia (Table 4). Persistence of the pneumonia despite antibiotic therapy led to repeated biopsies in five of these patients. In this group of five patients, two died from the pneumonia. Two others developed bronchiolitis obliterans syndrome, a complication that eventually led to the death of one of these two patients. Interestingly, all patients who died because of the pneumonia or developed bronchiolitis obliterans syndrome had coexisting grade A2 acute cellular rejection and grade B1 lymphocytic bronchitis in the same biopsy specimen as the acute bronchopneumonia.

Antibody-mediated rejection led to or contributed to death in two transplant recipients. Hyperacute rejection was diagnosed in one patient (patient 5) based on the immediate onset of pulmonary infiltrates and severe hypoxemia after the transplantation procedure, with a positive lymphocyte cross-match. Postmortem examination of the lung allograft revealed the presence of a diffuse small-vessel vasculitis with alveolar hemorrhage and diffuse alveolar damage.[9] In a second patient (patient 7), although the results of a pretransplant lymphocyte cross-match were negative, a second cross-match performed 6 days after transplantation had positive results. The antibodies were specific for HLA class I, as demonstrated by their absorption with platelets, and the test result was confirmed by a repeat experiment. The following results were obtained: normal serum, 4.6 median fluorescence units; pretransplant serum, 4.0 median fluorescence units; day 6 serum, 42.2 median fluorescence units (ratio with normal serum, 9.17); and day 6 serum after platelet absorption for 15 min, 18.6 median fluorescence units (ratio with normal serum, 4.04). Like patient 5, patient 7 manifested small-vessel vasculitis and diffuse alveolar damage (Fig 1) with the additional finding of focal parenchymal necrosis (Fig 2) in a specimen from a biopsy performed 2 days after transplantation. The morphologic findings and cross-match results fit best with accelerated antibody-mediated rejection because of a recall response.

[ILLUSTRATIONS OMITTED]

Interestingly, reperfusion injury was not a frequent cause of death in our series. Reperfusion injury with refractory fight ventricular failure and hypotension accounted for one death. Last, one recipient had withdrawal of life support after confirmation of irreversible severe hypoxic brain injury related to an intraoperative cardiac arrest.

Most of these patients (six of nine patients) received a trial of nitric oxide, but only one patient (patient 8) responded by significantly improving oxygenation. Patient 6 was maintained on a regimen of extracorporeal membrane oxygenation (ECMO) from the first day after transplantation until the second lung transplant operation 4 days later, an example of a successful use of ECMO as a bridge to retransplantation. ECMO was planned for patient 5, but the patient suffered a cardiopulmonary arrest during the insertion of the catheters for ECMO and could not be resuscitated. For the other patients, the use of ECMO would probably not have altered the outcome of their complications. In the patient whose death was caused by reperfusion injury (patient 9), the lungs at autopsy demonstrated extensive airway remodeling with granulation tissue and fibrosis. This patient was considered a candidate for ECMO, but because he initially showed improvement with pressure-controlled inverse ratio ventilation, ECMO was not pursued. When it became evident that his condition was no longer improving, the patient was no longer a candidate for retransplantation. Patients 2 and 4 were not treated with nitric oxide or ECMO and died of bacterial pneumonia and septic shock. It is unlikely that improving oxygenation by either of these modalities would have changed the fatal course of these infections. Nor would these treatments have benefited patient 1, whose death was caused by hypoxic-ischemic encephalopathy.

Laboratory data and morphologic findings were instrumental for determining the causes of respiratory decline in all patients. Antemortem allograft biopsy specimens were crucial for determining the causes of respiratory decline in three patients, and autopsies were crucial for four others. Likewise, cultures were crucial in six patients, contributory in two patients, and noncontributory in only one patient. Immunologic testing (ie, lymphocyte cross-matching against the donor) was least frequently useful but was crucial for ascertaining the causes of respiratory decline in two patients. Chest radiographic findings of patchy or diffuse bilateral alveolar and interstitial infiltrates (Fig 3) indicated the presence of, but not the nature of, existing abnormalities.

[ILLUSTRATION OMITTED]

DISCUSSION

Although recent data suggest a trend toward improved survival times in pediatric lung transplant recipients,[10] the life expectancy of adult lung allograft recipients has remained essentially unchanged since 1992.[1] In comparison with patients receiving transplants between 1988 and 1991, however, adults receiving transplants since 1992 had a substantially reduced risk of early postoperative death. Nevertheless, patient survival declines more rapidly in the early postoperative period than at any other time after transplantation. During the first 30 days after transplantation, nonspecific graft failure and infection account for the greatest numbers of deaths.[2]

In our population of lung transplant recipients, infection was the most common cause of early death. The high frequency of bacterial pneumonia and sepsis as causes of early deaths in our series parallels experiences at other centers. Husain et al[3] reported a fatal infection in 6 of 12 patients who died within the first 30 days after transplantation, most often Gram-negative bacterial pneumonia. The other six patients in this series died because of intraoperative and postoperative complications, including two cases classified as allograft failure that showed diffuse organizing alveolar damage in the donor lung (one case) and pulmonary edema (one case); the second patient had secondary pulmonary hypertension. Cagle et al[4] likewise reported a high frequency of Gram-negative pneumonia and diffuse alveolar damage in their patients who died early after transplantation, as well as perivascular infiltrates that suggested rejection in two individuals. In a large series by Bando et al,[11] 53 deaths occurred during the the first 100 days after transplantation. The deaths were caused by infection (25 patients), ARDS/diffuse alveolar damage (13 patients), intraoperative bleeding (6 patients), technical failure (5 patients), and other causes (4 patients).

In our patient group, acute bronchopneumonia was not an unusual finding in biopsy specimens obtained during the first 30 days after transplantation. However, persistence of the bronchopneumonia in multiple biopsy specimens boded poorly for patients, leading to death (two patients) or preceding the eventual development of bronchiolitis obliterans syndrome (two patients) in four of the five patients in whom it was observed. Interestingly, morphologic findings of acute cellular rejection accompanied the acute bronchopneumonia in these four patients. Although the number of cases is small, these data suggest that the combination of acute bronchopneumonia and acute cellular rejection may represent a risk factor for poor outcome.

In lung allograft recipients, antibody-mediated rejection is an uncommon cause of early death after transplantation. The four fatal cases of hyperacute rejections,[8,9,12,13] that have been reported share features that define a clinicopathologic syndrome probably representing the pulmonary analog of classic hyperacute rejection.[14] These characteristics include the following: (1) the onset of pulmonary edema and hemorrhage during or immediately after transplantation; (2) the presence of antibodies to class I or II HLA antigens; (3) histologic findings of interstitial neutrophilia and diffuse alveolar damage, sometimes accompanied by platelet/fibrin thrombi[13] and small-vessel vasculitis (ie, arterial or arteriolar fibrinoid necrosis)[9]; and (4) a rapidly fatal course. Although her immediate cause of death was bacterial pneumonia with septic shock, a second patient in our series also fits this clinicopathologic profile except that her anti-class I antibodies were first detected after transplantation, rather than before.

Given the high degree of clinical and radiographic overlap between ischemic reperfusion injury and hyperacute or accelerated antibody-mediated rejection, it is likely that cases of antibody-mediated rejection have been placed in the category of primary graft failure. Without a positive result for lymphocyte cross-matching, some examples may be classified as ischemic reperfusion injury based on the observation of diffuse alveolar damage in biopsy specimens. Neutrophil aggregation in alveolar capillaries is a feature of early diffuse alveolar damage. The histologic distinction between acute diffuse alveolar damage and hyperacute or accelerated antibody-mediated rejection rests on finding evidence of widespread endothelial injury (hemorrhage) associated with extensive and marked interstitial neutrophil infiltrates. Fibrinoid necrosis of vessels, prominent microthrombi, and parenchymal necrosis also weigh in favor of antibody-mediated rejection. Neutrophil aggregation primarily in alveolar sacs with lesser infiltration of the alveolar septa, however, fits better with a diagnosis of acute bronchopneumonia.

The clinical and histologic overlap between the different etiologies of early graft dysfunction suggests that a systematic approach to the differential diagnosis may be advantageous. Use of a multimodality algorithm (Fig 4) designed to assess for the major causes of early death may extend our ability to diagnose and manage patients with significant complications, and may also yield more accurate information regarding the frequencies of the different causes of early posttransplantation death. The prospective application of such an algorithm deserves further study.

[ILLUSTRATION OMITTED]

In conclusion, our results confirm previous reports showing a high incidence of fatal bacterial pneumonia in patients dying early after lung transplantation. Of note, two of our patients developed severe antibody-mediated rejection (hyperacute rejection, one patient; accelerated antibody-mediated rejection, one patient) that was the immediate cause of death in one patient and that led to a septic death in the other patient. Although hyperacute and accelerated antibody-mediated rejection are considered to be rare complications of lung transplantation, the actual frequencies of these complications are unknown. We believe that additional studies using sensitive methodologies are needed to determine the incidences of these phenomena and other causes of early death after transplantation.

REFERENCES

[1] Hosenpud JD, Bennett LE, Keck BM, et al. The Registry of the International Society for Heart and Lung Transplantation: sixteenth official report: 1999. J Heart Lung Transplant 1999; 18:611-626

[2] Hosenpud JD, Bennett LE, Keck BM, et al. The Registry of the International Society for Heart and Lung Transplantation: fourteenth official report: 1997. J Heart Lung Transplant 1997; 16:691-712

[3] Husain AN, Siddiqui MT, Reddy VB, et al. Postmortem findings in lung transplant recipients. Mod Pathol 1996; 9:752-761

[4] Cagle PT, Truong LD, Holland VA, et al. Factors contributing to mortality in lung transplant recipients: an autopsy study. Mod Pathol 1989; 2:85-89

[5] Montoya A, Mawulawde K, Houck J, et al. Survival and functional outcome after single and bilateral lung transplantation: Loyola Lung Transplant Team. Surgery 1994; 116: 712-718

[6] Deusch E, End A, Grimm M, et al. Early bacterial infections in lung transplant recipients. Chest 1993; 104:1412-1416

[7] Nizami I, Frost AE. Clinical diagnosis of transplant-related problems. In: Cagle PT, ed. Diagnostic pulmonary pathology. New York, NY: Marcel Dekker, 2000; 485-499

[8] Christie JD, Bavaria JE, Palevsky HI, et al. Primary graft failure after lung transplantation. Chest 1998; 114:51-60

[9] Scornik JC, Zander DS, Baz MA, et al. Susceptibility of lung transplants to preformed donor-specific HLA antibodies as detected by flow cytometry. Transplantation 1999; 68:1542-1546

[10] Boucek MM, Faro A, Novick RJ, et al. The Registry of the International Society of Heart and Lung Transplantation: Third Official Pediatric Report: 1999; J Heart Lung Transplant 1999; 18:1151-1172

[11] Bando K, Paradis IL, Komatsu K, et al. Analysis of time-dependent risks for infection, rejection, and death after pulmonary transplantation. Thorac Cardiovasc Surg 1995; 109:49-57

[12] Frost AE, Jammal CT, Cagle PT. Hyperacute rejection after lung transplantation. Chest 1996; 110:559-562

[13] Choi JK, Kearns J, Palevsky HI, et al. Hyperacute rejection of a pulmonary allograft: immediate clinical and pathologic findings. Am J Respir Crit Care Med 1999; 160:1015-1018

[14] Kissmeyer-Nielsen F, Olsen S, Petersen VP, et al. Hyperacute rejection of kidney allografts, associated with pre-existing humoral antibodies against donor cells. Lancet 1966; 2:662-665

[15] Yousem SA, Berry GJ, Cagle PT, et al. Revision of the 1990 working formulation for the classification of pulmonary allograft rejection: Lung Rejection Study Group. J Heart Lung Transplant 1996; 15:1-15

[16] Cooper JD, Billingham M, Egan T, et al. A working formulation for the standardization of nomenclature and for clinical staging of chronic dysfunction in lung allografts: International Society for Heart and Lung Transplantation. J Heart Lung Transplant 1993; 12:713-716

(*) From the Departments of Pathology (Drs. Zander, Donnelly, and Scornik), Medicine (Dr. Baz), Pediatrics (Drs. Visner and Faro), and Surgery (Dr. Staples), University of Florida College of Medicine, Gainesville, FL.

Manuscript received July 27, 2000; revision accepted November 3, 2000.

Correspondence to: Dani S. Zander, MD, Department of Pathology, Immunology, and Laboratory Medicine, University of Florida College of Medicine, Box 100275, Gainesville, FL 32610; e-mail: zander@pathology.ufl.edu

COPYRIGHT 2001 American College of Chest Physicians
COPYRIGHT 2001 Gale Group