INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung transplantation is now an acceptable therapy for patients with end-stage lung disease (1, 2). As this therapeutic modality has gained greater acceptability, posttransplant lymphoproliferative disease (PTLD) has become a major cause of morbidity in pediatric lung transplant recipients. PTLD is a heterogeneous clinical entity manifested by an abnormal expansion of lymphoid cells, with clinical manifestations ranging from benign polyclonal histology to lymphoma (3).

Although PTLD is a serious complication of transplantation, few clinical strategies have been developed to identify the subgroup of patients at greatest risk for it. Previously implicated risk factors include primary Epstein-Barr virus (EBV) infections (4, 5), cytomegalovirus (CMV) infections (6); specific immunosuppressive agents such as cyclosporine A (7) and FK506 (8) and the use of antilymphocyte preparations such as antithymocyte globulin (ATG) and OKT3 (9). Despite recognition of specific risk factors associated with PTLD, few studies have quantified the magnitude of these risks so that surveillance and interventions for them can be targeted. We conducted a retrospective cohort analysis to identify the risk factors associated with PTLD in pediatric lung transplant recipients.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Population

One hundred and forty-one first-time lung transplants were performed in children at the St. Louis Children's Hospital of Washington University School of Medicine from July 1990 to December 1997. Thirteen patients who died in the first 30 d after transplantation were excluded from analysis. All patients were monitored in the same manner for rejection and PTLD. None of the 13 patients with CF who were later found to have PTLD initially had clinical or pathologic evidence of PTLD. Pediatric patients who were recipients of primary bilateral lung transplants were identified, and their medical records and laboratory data were reviewed.

Transplantation Procedure

Patients underwent a comprehensive multidisciplinary evaluation as previously described (2). Preoperative assessment included testing for EBV (IgG antibody to EBV viral capsid antigen; IgM antibody; EBV early antigen; nuclear antigen) and CMV serologies (IgG antibody to CMV). Published criteria (14) were used to guide decisions about the timing of transplant listing and subsequent surgery. Bilateral sequential lung transplantation was done with standard techniques (15, 16).

Immunosuppression and Surveillance for Graft Rejection

All patients, except one, underwent triple immunosuppression with cyclosporin A (CSA), azathioprine, and corticosteroids as previously described (15). To detect rejection and infection, flexible fiberoptic bronchoscopy with transbronchial biopsy and bronchoalveolar lavage were done at least four times in the first 3 mo after transplantation (at 2 wk, 1 mo, 2 mo, and 3 mo, respectively). Surveillance bronchoscopy was done subsequently at 3-mo intervals until 1 yr after transplantation, and every 6 to 12 mo thereafter. One child had been treated with FK506 after a liver transplant done elsewhere 18 mo before lung transplantation at our center, and administration of this immunosuppressive agent was continued in place of CSA. Eleven patients who underwent lung transplantation between October 1990 and July 1994 received induction immunotherapy with antilymphocyte globulin (ALG; Minnesota antilymphocyte globulin, University of Minnesota, St. Paul, MN) when this was available, and with ATG (ATGAM; Upjohn, Inc., Kalamazoo, MI) thereafter. Induction therapy was subsequently discontinued, and ATG and muromonab-CD3 (orthoclone OKT3; Ortho Biotech, Inc., Raritan, NJ) were used only for episodes of acute allograft rejection refractory to intravenous steroids or for active bronchiolitis obliterans (BO). Nineteen children received ATG therapy and nine received OKT3 therapy for these indications. Of these 19 children, eight received both ATG and OKT3 therapy.

Acute rejection (grade A2 or greater) (3) was treated with 10.0 mg/kg/d of intravenous methylprednisolone for a 3-d period for the first two or three episodes of rejection; enhanced immunotherapy with ATG was then used, if tolerated for 10 to 14 d. If acute rejection persisted despite therapy with ATG, treatment with OKT3 was instituted for 10 to 14 d. Patients with BO (17) suggested by biopsy, or with bronchiolitis obliterans syndrome (BOS) (18), were treated in the same manner as patients with steroid-refractory acute rejection.

Confirmation of PTLD

Children with clinical or radiographic findings suggesting PTLD underwent biopsy to confirm this diagnosis. Biopsy materials obtained from all 16 children with PTLD were examined by the study pathologist (M.K.) and classified according to established criteria (19).

Statistical Analysis

Given the observation that most of the patients with PTLD had CF, we focused our risk analysis on the cohort of patients with CF. For the analysis we used univariate logistic regression, with PTLD as the dependent variable. The independent variables were results of CMV serology in the transplant recipient at the time of transplantation (positive, negative), and the recipient's EBV serology status at the time of transplantation (positive, negative), age at time of transplantation, number of acute rejections (fewer than two episodes in the first 3 mo after transplantation, versus two or more episodes in the 3 mo after transplantation), induction therapy with ATG or ALG (yes or no), and use of ATG or OKT3 therapy for steroid-resistant rejection (yes or no). A risk factor was considered statistically significant if the odds ratio (OR) for its occurrence was greater than 1.0 and the 95% confidence interval (CI) for the factor did not include 1.

We also determined whether the frequency of PTLD was greater during the beginning of the transplant program than at later periods. Specifically, we used the chi-square test of association to determine whether there was a relationship between the frequency of PTLD before and after the periods of 1993 to 1994 and 1994 to 1995.

To determine whether any clinical parameter might explain why patients with CF had a greater risk for PTLD, we performed a univariate logistic regression analysis for the entire cohort, using the same covariates as named earlier, with the exception of the underlying diagnosis (diagnosis versus exclusion of CF). Statistical analyses were done with the microcomputer-based SPSS System for Windows software package, version 9.0 (SPSS, Inc., Chicago, IL).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Of 128 first-time lung transplant recipients, 61 had CF. Among the entire cohort, 16 patients developed PTLD, yielding a frequency of the disease of 13% (16 of 128) patients among primary graft recipients. Thirteen of the 16 patients with PTLD had CF, and the three remaining patients had other conditions associated with end-stage lung disease. A diagnosis of CF was associated with an OR of 5.8 (95% CI: 0.6 to 21.4) Patients with CF had the highest frequency of PTLD, at 21% (13 of 61 patients) when compared with patients who had lung transplantation for other reasons 4% (three of 67 patients).

We could not identify any previously known risk factor for PTLD that distinguished the patients with CF from the patients without CF. The frequency of an EBV seronegative status of the graft recipient before transplantation, the most commonly associated risk factor for PTLD (4, 5), was approximately the same among patients with and without CF, at 51% (21 of 60 patients) and 48% (20 of 53 patients), respectively.

The frequency of two or more rejection episodes in the first 3 mo after transplantation was higher in the CF group than in the non-CF group, at 44% (27 of 61 patients) versus 18% (12 of 67 patients), respectively. The level of immunosuppression was comparable in both groups. The frequency of administration of ATG during induction or at any time was similar for patients with and those without CF, at 16% (10 of 61 patients) versus 18% (12 of 67 patients), and 34% (21 of 61 patients) versus 22% (15 of 67 patients), respectively. The frequency of administration of OKT3 at any time was higher in patients with CF than in those without CF, at 15% (nine of 61 patients) and 8% (five of 67 patients), respectively.

The median age at transplantation was significantly different for patients with CF and those without CF, at 13.8 yr and 5.0 yr, respectively (p < 0.05). Although this was statistically significant, age at the time of transplantation has not been previously identified as a risk factor for PTLD in pediatric lung transplant recipients.

The only two risk factors statistically associated with PTLD in the entire cohort were diagnosis of CF and two or more acute rejection episodes in the first 3 mo after transplantation, (p < 0.05 in both instances). In a multivariate logistic regression equation including underlying diagnosis and frequency of rejection episodes, both variables remained statistically significant (data not shown).

Risk of PTLD Developing in Patients with CF

Since 85% (13 of 16) of the patients with PTLD had CF, we subsequently focused on the covariates potentially associated with PTLD in the CF population. In this cohort of patients with CF, 18 (30%) were male and 42 (70%) were female. The median age of the patients with PTLD was similar to that of those without PTLD, at 12.0 yr and 14.4 yr, respectively. In the patients with PTLD, the median time from transplantation to diagnosis of PTLD was 4.2 mo (range: 1.8 to 71.4 mo). In five patients, the diagnosis was made more than 1 yr after transplantation, at 23.7, 24.3, 33.1, 53.4, and 71.4 mo, respectively.

The frequency of rejection in the first 3 mo after transplantation was the only significant risk factor for the development of PTLD in the CF population. Patients who had at least two rejection episodes during the first 90 d after lung transplantation were at significant risk of developing PTLD (OR: 11.0; 95% CI: 2.2 to 55.7) (Table 2). The chance of confirming acute graft rejection via bronchoscopy with biopsy was uniform across all patients, since every patient underwent prescheduled bronchoscopy, as described in METHODS.

In patients with CF, we did not find any other covariate than frequency of rejection to be associated with PTLD. Notably, recipient EBV or CMV status prior to transplantation, induction therapy with ATG, and the use of ALG, ATG, or OKT3 for steroid-resistant acute rejection after transplantation were not found to be significant risk factors for the development of PTLD (Table 2). FK506 was used in only one patient with PTLD, and could not be analyzed as a risk factor. The presence of nasal polyposis with chronic sinusitis confirmed by physical examination and computed tomographic scanning of the sinuses, age at time of transplantation, and secular trends were not statistically significant risk factors for PTLD (data not shown).

Clinical Outcome of PTLD in Patients with CF

The location of PTLD in 10 of the 13 cases of the disease (77%) included the primary graft. PTLD outside of the graft occurred in two patients, in the paranasal sinuses, as well as in one child with both pulmonary and central nervous system (CNS) involvement and in three children with disease involving the liver.

The mortality rate associated with PTLD in children with CF was high. Six of the 13 children (46%) in the cohort of CF patients who developed PTLD died as a direct result of this complication. Three further children were managed with transiently reduced immunosuppression and had successful resolution of their PTLD. At the time of this writing, these children were well, with good graft function. Three children underwent retransplantation after developing BOS (18), presumably as a result of a transient reduction in immunotherapy. These three patients are well, with good graft function. The last of the 13 patients has progressive disease not amenable to further therapy, and is receiving supportive care.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

For many children with end-stage chronic obstructive pulmonary disease, lung transplantation has become a viable therapeutic option. As the survival after lung transplantation has increased, the importance of comorbid conditions following transplantation, such as PTLD, has gained new significance. The treatment of PTLD is varied, and in part reflects the lack of knowledge about the risk factors and etiology of this heterogenous entity. To identify risk factors for PTLD, we used a retrospective cohort analysis in a homogeneous group of lung transplant recipients.

Two findings in our cohort analysis were important and unexpected. First, most pediatric lung transplant recipients with PTLD had CF (13 of 16 patients). Second, patients with CF and multiple rejection episodes had an 11-fold greater risk of developing PTLD than did patients with CF who had one or no rejection episodes in the first 90 d after transplantation. Neither finding was explained by greater frequency of the known risk factors for PTLD: EBV seronegative status at the time of transplantation (4, 5) and augmentation of immunosuppression (9, 13).

Why are pediatric patients with CF, and particularly those with recurrent graft rejection, at greater risk of PTLD? We speculate that patients with CF have a known genetic defect that may predispose to PTLD in the setting of an immunocompromised state. We do not know whether the predisposition results from an increased or altered response to inflammation, or whether immunosuppressive regimens are equally effective in patients with and without CF. Recent evidence suggests that patients with mutations in the cystic fibrosis transmembrane regulator gene have enhanced immune activation. Bronchoalveolar lavage fluid specimens from infants with CF, taken before their first pulmonary infections, showed increased concentrations of inflammatory mediators (20), and another study suggested that IL-10, an antiinflammatory cytokine, was reduced in CF lung (21).

A further factor contributing to the increased risk of PTLD in patients with CF may be their altered metabolism of CSA. The pharmacokinetic profile of CSA is different in patients with and those without CF. Patients with CF have higher CSA clearance, higher apparent oral clearance, a shorter mean residence time, and more erratic absorption of the drug than those without CF (22). They, may also have an altered response to CSA. Whether these differences collectively or individually contribute to an increased risk of PTLD is unknown, but warrants further investigation.

Despite primary EBV seroconversion being an important risk factor for PTLD, we did not find that EBV seropositivity was a risk factor for PTLD. This observation was not surprising, given that only 50% of our patients with CF had EBV-positive serology at the time of lung transplantation, and that primary EBV infection is a known risk factor for PTLD (4, 5). A limitation of our study in this regard was the inability to assess patients who were EBV seropositive at the time of lung transplantation for subsequent reactivation of the virus; however, EBV serology was not done uniformly after lung transplantation.

In the large pediatric lung transplant cohort examined in our study, we were able to systematically assess a full range of risk factors for PTLD among lung transplant recipients. The greatest risk factor for PTLD was a diagnosis of CF, and in this subgroup of patients the greatest risk factor for PTLD was two or more rejection episodes in the first months after transplantation. Our findings suggest that the etiologic basis for PTLD should be expanded to include multiple episodes of rejection rather than increased immunosuppression associated with rejection episodes. Additional basic scientific investigation and detailed, prospective clinical studies are needed to further understand why pediatric patients with CF are at such high risk for developing PTLD, and the contribution of multiple rejection episodes to the risk of PTLD.