Predictors of Acute Rejection After Lung Transplantation

Transcription

1 Predictors of Acute Rejection After Lung Transplantation Abeel A. Mangi, MD, David P. Mason, MD, Edward R. Nowicki, MD, MS, Lillian H. Batizy, MS, Sudish C. Murthy, MD, PhD, Diane J. Pidwell, PhD, Robin K. Avery, MD, Kenneth R. McCurry, MD, Gösta B. Pettersson, MD, PhD, and Eugene H. Blackstone, MD Department of Thoracic and Cardiovascular Surgery, Heart and Vascular Institute; Department of Quantitative Health Sciences, Research Institute; Department of Pulmonary, Allergy, and Critical Care Medicine, Medicine Institute; and Infectious Disease, Medicine Institute, Cleveland Clinic, Cleveland, Ohio Background. Acute rejection (AR) after lung transplantation (LTx) impacts survival and quality of life. The objective of this study, therefore, was to identify risk factors for AR after LTx, focusing on donor- and recipientspecific factors, operative variables, and immunologic issues, including pretransplant panel-reactive antibody (PRA) levels, and donor recipient human leukocyte antigen (HLA) mismatch. Methods. From March 1996 to November 2007, 481 adults undergoing LTx had 3237 serial transbronchial biopsy specimens that were evaluated for perivascular rejection (grade A0 to A4). Longitudinal analysis was used to characterize the prevalence of rejection grade and influence of donor, recipient, technical, and immunologic variables. Results. AR was highest (54% > A1) in the first 2 months after LTx, decreased at 6 months (16% > A1), then remained steady. Prevalence of AR at any time was dominated by donor-specific factors of young age (p < ), blunt trauma (p 0.008), and nonblack race (p 0.012) and by recipient class II PRA exceeding 10% (p 0.005). AR within 2 months was associated with HLA mismatch at the DR locus (p ) and use of non-o blood-group donors (p 0.008). AR at 4 years and longer after LTx was associated with HLA mismatch at the B locus (p 0.01). Conclusions. Only a few recipient and operative factors were identified for AR after LTx. Moderately sensitized recipients identified by class II PRA exceeding 10% and those with HLA mismatches at the B and DR loci appear to be more susceptible to AR; however, such immunologic variations appear to be well controlled with current donor selection and immunosuppression protocols. The impact of donor-specific variables on AR is surprisingly strong and warrants closer inspection. (Ann Thorac Surg 2011;91: ) 2011 by The Society of Thoracic Surgeons Acute rejection (AR) after lung transplantation (LTx) accounts for important morbidity and occasional mortality [1, 2]. It is associated with long-term allograft injury and bronchiolitis obliterans syndrome (BOS), the most important cause of late mortality after LTx [3, 4]. Yet despite its clinical importance, risk factors for AR remain poorly described [5]. The objective of this study was to identify risk factors for biopsy-proven AR monitored longitudinally after LTx, focusing on donor- and recipient-specific factors, technical variables surrounding LTx, and a broad range of immunologic variables. Patients and Methods Patients From January 1, 1996, to September 10, 2007, 511 patients underwent primary LTx at Cleveland Clinic. Routine Accepted for publication Jan 21, Address correspondence to Dr Mason, Cleveland Clinic, 9500 Euclid Ave, Desk J4-1, Cleveland, OH 44195; masond2@ccf.org. surveillance bronchoscopic biopsy specimens were obtained at 3 and 6 weeks, and at 3, 6, 9, and 12 months after LTx and as symptoms dictated. The study excluded 13 patients without biopsy specimens to test for histologic vascular rejection during follow-up, 6 who were younger than 18 years old, and 11 who underwent heart lung transplantation, leaving a cohort of 481 (Table 1). Patient, transplant, and follow-up data were extracted from the Unified Transplant Database, approved for use in research by the Institutional Review Board, with patient consent waived. Immunologic variables, including panel-reactive antibody (PRA), cross-match, and human leukocyte antigen (HLA) data, were obtained from the Institutional Review Board approved Allogen Laboratory database. Follow-Up Anniversary follow-up as of January 8, 2008, was used for analyses. Thirteen patients were lost to follow-up when they transferred to outside institutions. All patients had at least 1 year of follow-up, with mean follow-up among survivors of years by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg MANGI ET AL 2011;91: PREDICTORS OF ACUTE REJECTION AFTER LTx Table 1. and Characteristics (N 481) a Data available. Characteristic No. a No. (%) or Mean SD Demographics Age at transplant, y Body mass index, kg/m Education level 407 Posthigh school 151 (37) Primary insurance 463 Private 333 (72) Diagnoses Antitrypsin deficiency or COPD 229 (48) Usual interstitial pneumonitis 96 (20) Bronchiectasis or cystic fibrosis 88 (18) PAH or Eisenmenger syndrome 17 (3.5) Other 51 (11) Comorbidities Lung allocation score at listing Creatinine, mg/dl Hemodynamics Mean PAP, mm Hg Serology: Blood type A (40) AB (4.2) B (11) O (44) recipient EBV IgG (59) Pre-op presensitization/induction 481 Plasmapheresis at transplant 6 (1.2) ATG 21 (4.4) Demographics Age at death (y) Body mass index, kg/m Comorbidities History of smoking (51) Serology: blood type A (32) AB (1.3) B (11) O (56) Cause of death Cerebral hemorrhage (21) Stroke (27) Head trauma (39) recipient mismatch Cytomegalovirus (18) EBV (6.0) Transplant details Double-lung transplant (52) Years from 1/1996 to transplant Maximum ischemic time, min GENERAL THORACIC ATG antithymocyte globulin; COPD chronic obstructive pulmonary disease; EBV Epstein-Barr virus; IgG immunoglobulin G; PAH pulmonary arterial hypertension; PAP pulmonary artery pressure.

3 1756 MANGI ET AL Ann Thorac Surg PREDICTORS OF ACUTE REJECTION AFTER LTx 2011;91: Preoperative Immunologic Testing Class I and II antibody specificities were determined by several techniques throughout this period, including complement-dependent T- and B-cell assays, enzymelinked immunosorbent assay, and various plastic flow beads coated with HLA antigens, including single antigen beads most recently. Class I (HLA-A, HLA-B, and HLA-C) and class II (HLA-DR) PRA are reported as percentages, together with specificity. Antibody specificity indicates presence of anti-hla antibodies to particular antigens that may result in a positive cross-match with donor cells carrying the corresponding antigen. Class I and II PRA data were available for all patients; 391 (81%) used flow cytometry. These tests were repeated approximately bimonthly as patients awaited transplantation. Sensitized recipients underwent a virtual cross-match when potential donors were identified; that is, the potential recipient s PRA data were analyzed for donor-specific antibody, and the likelihood of a positive flow cytometric or lymphocytotoxic cross-match was estimated [6]. Perioperative or postoperative donor-specific cross-matches were routinely performed for all transplants. HLA antigens at the A, B, and DR loci were available for both recipient and donor in 417 patients. These were compared at each locus, and the number of donor antigens not in common with the recipient (mismatches) was determined and expressed as the total number of mismatches across all 3 loci (Table 2). A few patients had no HLA mismatches at any of the 3 loci, but none had zero total mismatches for both class I and class II. Table 2. Human Leukocyte Antigen Mismatch Mismatches, No. No. a Patients No. (%) Locus A (9.1) (39) (52) B (1.8) (31) (67) DR (4.5) (43) (53) Total mismatches: A B DR (0) 1 1 (0.24) 2 17 (4.1) 3 52 (12) (26) (35) 6 91 (22) a Data available. Immunosuppression Immunosuppression used a three-drug regimen of calcineurin inhibitor (cyclosporine up to August 2001; tacrolimus thereafter) with selective use of lymphocyte-depleting induction therapy, intravenous immunoglobulin, and plasmapheresis [7]. Induction therapy, typically antithymocyte globulin (ATG), was used infrequently for patients in whom donor-specific antibody was present, positive flow cytometry T or B cross-match was predicted, or the recipient s clinical status suggested urgent transplantation. These decisions were made on a case-by-case basis, factoring in organ availability and the patient s ability to tolerate induction therapy. Plasmapheresis use depended on antibody strength. Biopsy Histology After transplant, patients underwent 3232 transbronchial biopsies, 92% performed in the first 2 years (Fig 1). Histologic vascular rejection was graded according to International Society for Heart and Lung Transplantation (ISHLT) guidelines: A0, no rejection; A1, minimal rejection; A2, mild rejection; A3, moderate rejection; A4, severe rejection [8]. No biopsy specimen was graded severe (A4). For analysis, A2 and A3 rejections were coalesced because only 16 biopsy specimens were graded A3. This resulted in three rejection grades for analysis: 0 none, 1 minimal, 2 mild/moderate. The B categorization of rejection was excluded because of concerns about grading reliability. Cytomegalovirus Prevention Between 1996 and 2005, patients at high risk for cytomegalovirus (CMV) infection (donor positive [D ]), recipient negative [R ]), underwent prophylaxis with CMV hyperimmune globulin, given at 0, 2, 4, 6, 8, 12, and 16 weeks, in addition to ganciclovir or valganciclovir. Current CMV prevention is a combination of prophylaxis and preemptive therapy. Valganciclovir at 900 mg/day (or renal-adjusted equivalent) is administered to any patient at risk for CMV (D /R, D /R, and D /R ), with quantitative CMV polymerase chain reaction performed every 1 to 2 weeks during the first year after transplant. Data Analysis Biopsy specimen category was analyzed longitudinally using a nonlinear cumulative logit mixed model for repeated measures that resolved a number of temporal components and their shaping parameters to generate the temporal pattern of average grade [9]. Each component was independently modulated by a time function with common random intercept., donor, recipient donor mismatch, and transplant variables (Table 1, Appendix) were used to create several potential longitudinal models: donor risk factors only, recipient plus donor factors, and recipient, donor, recipient donor mismatch, and transplant factors. Variables were initially screened using ordinary multivariable linear regression (SAS PROC REG) with the assumption of independence of observations and liberal entry (p 0.10) and stay (p 0.12) criteria. Resulting candidates were retained in the nonlinear mixed model if p 0.1.

4 Ann Thorac Surg MANGI ET AL 2011;91: PREDICTORS OF ACUTE REJECTION AFTER LTx 1757 Fig 1. Histogram shows the number of biopsies (solid bars) and number of patients (white bars) across follow-up time. GENERAL THORACIC When evaluating the association of pre-transplant PRA with posttransplant histologic rejection, a model of class I and II PRA sensitivity and specificity was used to form four groups: (1) sensitivity 10% or less and no specificity to donor-specific antigens (reference group for the analysis); (2) sensitivity exceeding 10% and specificity; (3) sensitivity exceeding 10% and no specificity; and (4) sensitivity 10% or less and specificity. The last three groups were then combined for analysis with other pretransplant recipient and donor risk factors for rejection identified in the first portion of the analysis. When HLA mismatch was analyzed, the number of mismatches at each locus was combined with pretransplant risk factors. Fig 2. Predicted average prevalence of acute rejection grade after lung transplantation. Symbols represent crude estimates based on averaging independent values within 10 time frames of an equal number of biopsies (these estimates do not account for repeated measures, as do solid curves). Green line no rejection (A0); blue line minimal rejection (A1); black line mild/moderate rejection (A2 A3). Data Presentation Continuous variables are summarized as mean standard deviation and as equivalent 15th, 50th (median), and 85th percentiles. Categoric data are summarized using frequencies and percentages. The bootstrap percentile method was used to obtain 68% confidence limits (equivalent to 1 SD) for longitudinal estimates [10]. Results Overall Temporal Trend of AR In the first 6 months after transplant, the predicted average prevalence of grade A0 rejection increased from 49% to 84%, grade A1 decreased from 23% to 10%, and grade A2/A3 decreased from 27% to 6% (Fig 2). After 6 months, the predicted prevalence of all rejection grades remained constant. Impact of - and -Specific Variables on AR The prevalence of AR throughout the study was dominated by donor-specific variables (Table 3), specifically, younger age (Fig 3A) and death by head trauma (Fig 3B). Rejection grade early after LTx was higher in recipients whose donors were non-o blood type and whose donors had a larger body surface area. Patients with CMV mismatch (D /R ) had lower rejection grade early after LTx. Rejection in the intermediate peaking phase correlated with recipient-specific and procedure factors, such as transplant date. Impact of Immunologic Variables on AR PRA LEVELS. Of 391 patients with PRA data available by solid-phase assay, 136 (34%) had class I PRA exceeding 10%, and 64 (16%) class II PRA exceeding 10%. Elevation in class I PRA was associated with no adverse effect on prevalence of rejection. Elevation in class II PRA exceed-

5 1758 MANGI ET AL Ann Thorac Surg PREDICTORS OF ACUTE REJECTION AFTER LTx 2011;91: Table 3.,, and Transplant Risk Factors for Higher Biopsy Grade Factor Estimate SE p Value Overall History of diabetes Age a Mechanism of death blunt injury Race other than black Early constant phase Forced vital capacity (higher) b Non-O blood type Body surface area (greater) c recipient mismatch No CMV mismatch Intermediate peaking phase Date of operation (more recent) d Mean systolic pressure (lower) e EBV IgG Despite the central importance of AR, its risk factors remain incompletely identified. This study assessed the combination of donor, recipient, and transplant variables, as well as immunologic matching details, to identify risk factors for AR, with the ultimate goal of minimizing rejection risk and tailoring immunosuppression. Principal Findings OVVERALL TEMPORAL TREND OF AR. Histologically proven AR occurred most frequently in the first months after LTx, then rapidly declined. This is consistent with other reports of the temporal pattern of rejection in which AR peaked early after transplant and returned to a low but continuous level [7, 16]. This pattern suggests that allograft immunogenicity is highest early after transplant and that strategies for preventing AR should be targeted around this time. IMPACT OF DONOR- AND RECIPIENT-SPECIFIC VARIABLES ON AR. Surprisingly, when considering donor, recipient, and transplant variables, AR was most associated with donor variables. Nonimmunologic donor factors have been noted to impact AR, although studies are few and mechanisms largely hypothetical [17, 18]. a Exp (age/37), exponential transformation. b Exp (forced vital capacity/ 2.17), exponential transformation. c 1/body surface area, inverse transformation. d Log (date of operation), natural logarithmic transformation. e (Mean systolic pressure/89) 2, squared transformation. CMV cytomegalovirus; EBV Epstein-Barr virus; IgG immunoglobulin G. ing 10% without antigenic specificity for donor HLA antigens strongly predicted lower rejection (Fig 4); this remained highly predictive throughout the study by multivariable analysis (Table 4). HLA MISMATCH. Mismatch at HLA-A appeared to have no association with rejection (Table 4). Mismatch at HLA-B exerted a powerful effect on rejection in the late rising phase (Fig 5A). Mismatch at the DR locus exerted a strong effect on rejection immediately after transplant, but this diminished over time (Fig 5B). There was little independent association between the total number of HLA mismatches and histologic rejection by multivariable analysis. A Head trauma No Yes Comment Acute rejection remains a major morbidity after LTx and can range from mild to severe [11]. Although some patients with rejection episodes can be treated as outpatients, other AR episodes require hospitalization, return to the intensive care unit, intubation, and occasionally even cause death [5, 12 14]. Even a single episode of AR is a risk factor for developing BOS, the disease process associated with progressive and inevitable decline in lung function and cause of death of most patients [4]. There is mounting evidence that BOS is affected by both alloimmune and nonalloimmune mechanisms [15]. B Fig 3. Prevalence of biopsy grade A0. Symbols represent crude estimates based on averaging independent values within 10 time frames of equal number of biopsies (these estimates do not account for repeated measures, as do solid curves). Green line no rejection (A0); blue line minimal rejection (A1); black line mild/moderate rejection (A2 A3). (A) One year after transplant according to donor age. Graph is a nomogram for Table 3. (B) According to cause of donor death.

6 Ann Thorac Surg MANGI ET AL 2011;91: PREDICTORS OF ACUTE REJECTION AFTER LTx The strong association between younger donor age and increased rejection is novel and surprising. Current donor guidelines recommend using younger donors [19], and most transplant physicians believe that younger donor lungs provide better long-term outcomes. Although studies have shown variability in immunogenicity related to donor age in kidney transplantation, our finding warrants further investigation in LTx [20, 21]. No study has related donor age to immunogenicity of lung allografts. Our finding of increased rejection when donors died of traumatic head injury is supported by other studies showing increased rejection in heart and lung recipients whose donors died this way [22, 23]. The proposed mechanism is increased intracranial pressure inducing cytokine release, leading to adhesion molecule mediated endothelial injury in the allograft, followed by AR [24]. The findings that non blood-type O donors and high donor body surface area predicted rejection after LTx are novel, but the mechanisms remain unclear. Acute rejection in the intermediate period after LTx more strongly correlated with recipient-specific and procedure-related factors, with the strongest association being more recent transplantation. The explanation for this is unclear, but may be due to changes in immunosuppressive regimens; more recently transplanted patients underwent immunosuppression with tacrolimus rather than cyclosporine as the calcineurin inhibitor. Although a recent meta-analysis reported that AR was less common in LTx patients using a tacrolimus-based immunosuppression regimen compared with cyclosporine [25], this was based on two small studies, one of which failed to reach statistical significance [26] and the other used routine induction therapy [27]. A change to tacrolimus-based immunosuppression may account for increased AR, although this explanation is hypothetical, unexpected, and requires further investigation. Our finding of decreased rejection in CMV D /R patients can be explained by our aggressive use of CMV Table 4. Effect of Immunologic Factors on Higher Biopsy Grade Adjusted for,, and Transplant Factors Factor 1759 Estimate SE p Value Overall phase History of diabetes Class-II flow PRA PRA 10% and no specificity Interval from PRA test to transplant Age (younger) a Mechanism of death blunt injury Race other than black Early constant phase Cross-match Negative class-i flow Non-O blood type Body surface area (greater) b recipient mismatch HLA mismatch Fewer mismatches at B locus c More mismatches at DR locus c No CMV mismatch Intermediate peaking phase Cross-match Positive class-i flow cross-match Date of operation (more recent) d Mean systolic pressure (lower) e EBV IgG Late rising phase recipient mismatch HLA mismatch More at B locus b Fewer at DR locus b GENERAL THORACIC a Exp (age/37), exponential transformation. b 1/body surface area, inverse transformation. c Log (number of mismatches), natural logarithmic transformation. d Log (date of operation), natural logarithmic transformation. e (Mean systolic pressure/89) 2, squared transformation. CMV cytomegalovirus; EBV Epstein-Barr virus; IgG immunoglobulin G; PRA panel-reactive antibody. Fig 4. Prevalence of biopsy grade A0 stratified by class II preoperative panel reactive antibody (PRA). Class II PRA. Green line PRA 10%, no specificity; blue line PRA 10%, no specificity. Depiction is as in Figure 2. hyperimmune globulin prophylaxis, which has been shown to decrease AR after cardiothoracic transplantation. IMPACT OF PRA LEVELS ON AR. Elevated PRAs can occur secondary to blood transfusion, pregnancy, previous transplantation, and occasionally, infection or immunization [28], and may influence outcomes after LTx. A large percentage of our patients had elevation of class I or II PRA. Patients with class II PRA exceeding 10% had decreased risk for AR. Two studies examined the effect of elevated PRA on post-ltx survival with the primary end point of survival [29, 30], and two evaluated survival and rejection [28, 31]. Hadjiliadis and colleagues [29] reported

7 1760 MANGI ET AL Ann Thorac Surg PREDICTORS OF ACUTE REJECTION AFTER LTx 2011;91: A B Mismatch decreased survival in patients with PRA exceeding 25% in a multicenter study of 656 patients; Shah and colleagues [30] demonstrated decreased survival in patients with more than 25% PRA using the United Network of Organ Sharing database of more than 10,000 patients. This worsened survival was not seen in the more modern cohort of patients they evaluated. Previous studies evaluating the effect of pretransplant PRA on LTx outcomes did not differentiate between class I and II. It is not possible to explain the mechanism for this apparent third-party antibody effect in patients with elevated-pretransplant PRA. IMPACT OF HLA MISMATCH ON AR. Our study demonstrated an association between mismatches at the HLA-B and DR loci and increased AR after LTx. The temporal pattern of rejection suggests an early impact of HLA-DR mismatch on AR and a late impact of HLA-B mismatch. These findings build on limited data evaluating the effect of HLA mismatches on outcomes after LTx. Our study defines the temporal trends during which each HLA mismatch carries importance and may account for inconsistency in studies focusing on the relative importance of each locus, although it must be noted that HLA-B mismatch was present in only 8 patients Mismatch Fig 5. Prevalence of biopsy grade A0 stratified by preoperative human leukocyte antigen mismatch. Symbols represent crude estimates based on averaging independent values within 10 time frames of equal number of biopsies (these estimates do not account for repeated measures, as do solid curves). Green line no rejection (A0); blue line minimal rejection (A1); black line mild/moderate rejection (A2 A3). (A) At B locus. (B) At DR locus. Wisser and colleagues [32] showed a trend toward worse survival with mismatch at the B and DR loci in a small cohort of 78 patients undergoing LTx. In another study of 102 LTx patients, van den Berg and colleagues [33] showed that HLA-DR mismatch was a risk factor for BOS and graft loss. Schulman and colleagues [34] demonstrated that HLA-DR and HLA-B loci were risk factors for highgrade rejection (A2 A3) after LTx in a cohort of 152 patients, whereas Quantz and colleagues [35] reported that mismatches at HLA-A and DR and total number of mismatches predicted a small increase in mortality after LTx in a multicenter study of 3549 patients. Our study demonstrates the importance of HLA-B and DR mismatching on outcomes after transplant. Although increased HLA matching between donor and recipient may not be practical in LTx because of time constraints, our findings suggest that patients mismatched with their donors at HLA-B and DR are at increased risk for AR and might benefit from closer surveillance or a tailored immunosuppression regimen. Strengths and Limitations This study s primary limitation is that it was a single-center experience. Only patients believed to have an adequately safe preoperative cross-match, based on subjective factors, underwent transplantation. Additionally, techniques for immunologic evaluation varied across time. Despite small numbers, the study s strengths are that it is the largest to date evaluating predictors of AR after LTx; includes multivariable analysis of a broad range of donor, recipient, transplant, and immunologic matching variables; and is supported by histologic confirmation of AR in a longitudinal manner with consistent pathologic analysis. Conclusions The key findings of our study suggest that multiple donor, recipient, and immunologic variables affect AR after LTx. However, their impact is relatively small and appears controlled using current immunosuppressive strategies. Some of these variables may be modifiable and warrant further investigation. This study was supported in part by the Peter and Elizabeth C. Tower and Family Endowed Chair in Cardiothoracic Research, James and Sharon Kennedy, the Slosburg Family Charitable Trust, and Stephen and Saundra Spencer (GBP), and the Kenneth Gee and Paula Shaw, PhD, Chair in Heart Research (EHB). We thank Lynn Klingman for analytic advice. References 1. Hodge G, Hodge S, Chambers D, Reynolds PN, Holmes M. Acute lung transplant rejection is associated with localized increase in T-cell IFNgamma and TNFalpha proinflammatory cytokines in the airways. Transplantation 2007;84: De Vito Dabbs A, Hoffman LA, Iacono AT, Zullo TG, McCurry KR, Dauber JH. Are symptom reports useful for differentiating between acute rejection and pulmonary infection after lung transplantation? Heart Lung 2004;33:

8 Ann Thorac Surg MANGI ET AL 2011;91: PREDICTORS OF ACUTE REJECTION AFTER LTx 3. Hachem RR, Khalifah AP, Chakinala MM, et al. The significance of a single episode of minimal acute rejection after lung transplantation. Transplantation 2005;80: Burton CM, Iversen M, Carlsen J, et al. Acute cellular rejection is a risk factor for bronchiolitis obliterans syndrome independent of post-transplant baseline FEV1. J Heart Lung Transplant 2009;28: Knoop C, Estenne M. Acute and chronic rejection after lung transplantation. Semin Respir Crit Care Med 2006; 27: Appel JZ 3rd, Hartwig MG, Cantu E 3rd, Palmer SM, Reinsmoen NL, Davis RD. Role of flow cytometry to define unacceptable HLA antigens in lung transplant recipients with HLA-specific antibodies. Transplantation 2006;81: Mason DP, Little SG, Nowicki ER, et al. Temporal pattern of transfusion and its relation to rejection after lung transplantation. J Heart Lung Transplant 2009;28: Stewart S, Fishbein MC, Snell GI, et al. Revision of the 1996 working formulation for the standardization of nomenclature in the diagnosis of lung rejection. J Heart Lung Transplant 2007;26: Mason DP, Rajeswaran J, Murthy SC, et al. Spirometry after transplantation: how much better are two lungs than one? Ann Thorac Surg 2008;85: Efron B, Tibshirani RJ. An introduction to the bootstrap. New York: Chapman and Hall/CRC; DeVito Dabbs A, Hoffman LA, Iacono AT, et al. Pattern and predictors of early rejection after lung transplantation. Am J Crit Care 2003;12: Bando K, Paradis IL, Komatsu K, et al. Analysis of timedependent risks for infection, rejection, and death after pulmonary transplantation. J Thorac Cardiovasc Surg 1995; 109: Hopkins PM, Aboyoun CL, Chhajed PN, et al. Association of minimal rejection in lung transplant recipients with obliterative bronchiolitis. Am J Respir Crit Care Med 2004;170: Burton CM, Iversen M, Scheike T, Carlsen J, Andersen CB. Minimal acute cellular rejection remains prevalent up to 2 years after lung transplantation: a retrospective analysis of 2697 transbronchial biopsies. Transplantation 2008;85: Weigt SS, Wallace WD, Derhovanessian A, et al. Chronic allograft rejection: epidemiology, diagnosis, pathogenesis, and treatment. Semin Respir Crit Care Med 2010;31: Hopkins PM, Aboyoun CL, Chhajed PN, et al. Prospective analysis of 1,235 transbronchial lung biopsies in lung transplant recipients. J Heart Lung Transplant 2002;21: Belperio JA, Weigt SS, Fishbein MC, Lynch JP, 3rd. Chronic lung allograft rejection: mechanisms and therapy. Proc Am Thorac Soc 2009;6: Sharples LD, McNeil K, Stewart S, Wallwork J. Risk factors for bronchiolitis obliterans: a systematic review of recent publications. J Heart Lung Transplant 2002;21: Appendix Variables Considered in Multivariable Model Sundaresan S, Trachiotis GD, Aoe M, Patterson GA, Cooper JD. lung procurement: assessment and operative technique. Ann Thorac Surg 1993;56: de Fijter JW, Mallat MJ, Doxiadis, II, et al. Increased immunogenicity and cause of graft loss of old donor kidneys. J Am Soc Nephrol 2001;12: Martins PN, Pratschke J, Pascher A, et al. Age and immune response in organ transplantation. Transplantation 2005;79: Singhal AK, Sheng X, Drakos SG, Stehlik J. Impact of donor cause of death on transplant outcomes: UNOS registry analysis. Transplant Proc 2009;41: Ciccone AM, Stewart KC, Meyers BF, et al. Does donor cause of death affect the outcome of lung transplantation? J Thorac Cardiovasc Surg 2002;123: Faropoulos K, Apostolakis E. Brain death and its influence on the lungs of the donor: how is it prevented? Transplant Proc 2009;41: Fan Y, Xiao YB, Weng YG. Tacrolimus versus cyclosporine for adult lung transplant recipients: a meta-analysis. Transplant Proc 2009;41: Keenan RJ, Konishi H, Kawai A, et al. Clinical trial of tacrolimus versus cyclosporine in lung transplantation. Ann Thorac Surg 1995;60: Treede H, Klepetko W, Reichenspurner H, et al. Tacrolimus versus cyclosporine after lung transplantation: a prospective, open, randomized two-center trial comparing two different immunosuppressive protocols. J Heart Lung Transplant 2001;20: Lau CL, Palmer SM, Posther KE, et al. Influence of panelreactive antibodies on posttransplant outcomes in lung transplant recipients. Ann Thorac Surg 2000;69: Hadjiliadis D, Chaparro C, Reinsmoen NL, et al. Pretransplant panel reactive antibody in lung transplant recipients is associated with significantly worse post-transplant survival in a multicenter study. J Heart Lung Transplant 2005;24:S Shah AS, Nwakanma L, Simpkins C, Williams J, Chang DC, Conte JV. Pretransplant panel reactive antibodies in human lung transplantation: an analysis of over 10,000 patients. Ann Thorac Surg 2008;85: Gammie JS, Pham SM, Colson YL, et al. Influence of panelreactive antibody on survival and rejection after lung transplantation. J Heart Lung Transplant 1997;16: Wisser W, Wekerle T, Zlabinger G, et al. Influence of human leukocyte antigen matching on long-term outcome after lung transplantation. J Heart Lung Transplant 1996;15: van den Berg JW, Hepkema BG, Geertsma A, et al. Long-term outcome of lung transplantation is predicted by the number of HLA-DR mismatches. Transplantation 2001;71: Schulman LL, Weinberg AD, McGregor C, Galantowicz ME, Suciu-Foca NM, Itescu S. Mismatches at the HLA-DR and HLA-B loci are risk factors for acute rejection after lung transplantation. Am J Respir Crit Care Med 1998;157: Quantz MA, Bennett LE, Meyer DM, Novick RJ. Does human leukocyte antigen matching influence the outcome of lung transplantation? An analysis of 3,549 lung transplantations. J Heart Lung Transplant 2000;19: GENERAL THORACIC Demographics: Comorbidity: Pulmonary function: Hemodynamics: Serology/immunology: Female, Caucasian, African American, age, body surface area, height, weight, weight/height Diabetes, previous lung operation, previous thoracic operation FEV1% (percent of predicted), NHANES normalized; FVC% (percent of predicted), NHANES normalized; FEV 1 %/FVC%; total vital capacity: forced vital capacity Systolic blood pressure, diastolic blood pressure, mean blood pressure Rh, CMV IgG, EBV IgG Continued

9 1762 MANGI ET AL Ann Thorac Surg PREDICTORS OF ACUTE REJECTION AFTER LTx 2011;91: Appendix. Continued Demographics: Comorbidity: Serology/immunology: Mechanism of death: - mismatch: Transplant Procedure: Female, Caucasian, African American, body surface area, height, weight, weight/height Hypertension, creatinine Rh, CMV IgG, EBV IgG Blunt injury, gunshot wound recipient CMV mismatch, donor recipient Rh mismatch Use of induction therapy, right lung transplant only, left lung transplant only, cardiopulmonary bypass time (min), years from listing to transplant, years from referral to transplant CMV cytomegalovirus; EBV Epstein-Barr virus; FEV 1 forced expiratory volume in 1 second; FVC forced vital capacity; IgG immunoglobulin G; NHANES National Health and Nutrition Examination Survey.