Prevalence and Outcome of Bronchiolitis Obliterans Syndrome After Lung Transplantation

Transcription

1 Prevalence and Outcome of Bronchiolitis Obliterans Syndrome After Lung Transplantation Sudhir Sundaresan, MD, Elbert P. Trulock, MD, Thallachallour Mohanakumar, PhD, Joel D. Cooper, MD, G. Alexander Patterson, MD, and The Washington University Lung Transplant Group Divisions of Cardiothoracic Surgery and General Surgery, Department of Surgery, and Division of Respiratory and Critical Care, Department of Medicine, Washington University School of Medicine, Barnes Hospital, St. Louis, Missouri Background. Bronchiolitis obliterans syndrome (BOS) is the main cause of late morbidity and mortality in lung transplantation. This study was designed to accurately determine the prevalence of this syndrome of chronic lung allograft dysfunction (which is presumed to be due to chronic rejection). Methods. A retrospective analysis was done of 212 consecutive lung transplantations performed at Barnes Hospital between July 1988 and March 1994 to characterize the prevalence and course of BOS. One hundred eighty-seven transplant recipients survived at least 3 months after transplantation, putting them at risk for BOS. Recipients free of BOS (group I) were distinguished from those with BOS (group II) based on the presence of declining spirometry (forced expiratory volume in 1 second persistently less than 80% of previous baseline) or histologic obliterative bronchiolitis in group II. Results. There were 110 transplantations in group I (59%) and 77 in group II (41%). At follow-up, BOS was detected using the following criteria: declining forced expiratory volume in 1 second alone, 40 of 77 (52%); positive histologic results alone, 7 of 77 (9.1%); and both, 30 of 77 (38.9%). Declining spirometry was the most common initial sign of BOS onset (57 of 77, 74%). There were no differences between groups with respect to age, sex, indication for transplantation, or type of transplantation performed. The mortality rate was significantly higher with BOS (group II, 22 of 77 [28.6%] versus group I, 8 of 110 [7.3%]; p = 0.001) and was not related to either the type of transplantation performed or the indication for transplantation. Follow-up of group II (mean 35.1 months; range, 7.1 to 63.7 months) showed a delay until BOS onset ( months); when BOS was fatal death ensued within months of its onset. Comparison of the first and last quartiles of recipients in this series (QTR1 versus QTR4, 53 patients in each) demonstrated a higher prevalence of BOS in QTR1 (24 with BOS of 43 at risk [55.8%] versus QTR4, 5 with BOS of 52 at risk [9.6%]; p < 0.001) and a worse BOS functional score in QTR1 ( versus QTR4, ; p = O.OO7). Conclusions. (1) Bronchiolitis obliterans syndrome is truly a clinical syndrome, not simply a pathologic entity; (2) BOS displays considerable latency in onset and progression; (3) lung transplant recipients must therefore be followed up for a sufficient interval to determine the actual prevalence and mortality rate of BOS; and (4) the prevalence and mortality rates of BOS are higher than previously appreciated, exceeding 50% and 40%, respectively. (Ann Thorac Surg 1995;60:1341-7) ung transplantation is now an established rnodality L in the treatment of selected patients with end-stage lung disease. A number of lung transplantation programs, including our own [1l, have reported operative mortality rates of less than 10%. Despite this improvement in early survival, the long-term outlook for lung allograft recipients is clouded by a late-onset syndrome of chronic lung allograft dysfunction. This syndrome, believed (but not proved) to be a manifestation of chronic aliograft rejection, has emerged as the main source of late mortality and morbidity for these patients. In the context of lung transplantation, this entity was Presented at the Thirty-first Annual Meeting of The Society of Thoracic Surgeons, Palm Springs, CA, Jan 30-Feb 1, Address reprint requests to Dr Sundaresan, Division of Cardiothoracic Surgery, Washington University School of Medicine, Suite 3107 Queeny Tower, One Barnes Hospital Plaza, St. Louis, MO first described in 1984 by the Stanford group [2]. They observed a process characterized by declining forced expiratory volume in 1 second (FEV1) and histologic features of obliterative bronchiolitis (OB) in a significant number of their heart-lung allograft recipients. Based on the early clinical experience in heart-lung transplantation, OB was believed to affect 10% to 54% of long-term survivors of this procedure [3]. As experience with isolated lung transplantation accumulated, OB was also encountered in recipients of single and bilateral lung transplants. Early follow-up reports suggested that OB might occur somewhat less frequently in isolated lung transplantation, perhaps in about 20% of long-term survivors [4]. We performed this retrospective analysis of our own experience with the following goals:(1) to determine the prevalence of this syndrome in our lung transplant recipient population (for this purpose, we defined 1995 by The Society of Thoracic Surgeons /95/$9.50 SSDI (95)

2 1342 SUNDARESAN ET AL Ann Thorac Surg BOS IN LUNG TRANSPLANTATION 1995;60: this entity as "bronchiolitis obliterans syndrome" [BOS] and used the corresponding definitions and conventions of the 1993 report of the International Society for Heart and Lung Transplantation [5]); (2) to characterize the clinical features and progression of BOS; and (3) to estimate the mortality rate of BOS given the limited therapeutic options currently available. Patients and Methods General We retrospectively analyzed 212 consecutive lung transplantations performed on 208 patients in the Washington University Lung Transplant Program between July 1, 1988, and March 1, Technical aspects of the donor [6] and recipient [7-9] operative procedures, as well as details regarding immunosuppression protocols [10] and posttransplant follow-up and management [11], have been reported elsewhere. Follow-up was completed on July 1, 1994 (the date of beginning the analysis), or at the time of the recipient's death. Follow-up was complete in all recipients and ranged from 4 to 72 months. Definition of Bronchiolitis Obliterans Syndrome The following conventions of the 1993 International Society for Heart and Lung Transplantation report [5] were adopted. (1) Recipients were considered evaluable ("at risk" for developing BOS) only if they survived at least 3 months after their transplantation [5]. Twenty-one of 208 recipients (10%) died within 3 months of their procedure and were excluded. The remaining 187 patients at risk for BOS formed the population for this analysis. (2) The diagnosis of BOS (see below) rests on functional data (from pulmonary function testing) and pathologic data (from bronchoscopy with transbronchial lung biopsy) obtained from the recipient periodically during follow-up. Although we do not adhere to a rigid protocol, a general schedule for these surveillance procedures in our program is as follows. Pulmonary function testing relies on the FEV 1 [5]. Spirometry is obtained weekly for the first 3 months posttransplantation, then monthly for the remainder of the first year, and then every 2 to 3 months thereafter [10]. A baseline FEV 1 value was documented for all evaluable recipients. This baseline value was the average of the two highest consecutive FEV1 measurements obtained 3 to 6 weeks apart after transplantation and was used to determine the fractional change in FEV 1 later in follow-up. Transbronchial lung biopsy is obtained at 2 to 3 weeks after transplantation; at approximately 3, 6, and 12 months; and then annually after 1 year [10]. Bronchoscopy with transbronchial lung biopsy is also performed whenever clinically indicated, particularly in the first several months posttransplantation. The main value of transbronchial lung biopsy is to identify a specific treatable lesion when clinical or spirometric decline is encountered during follow-up [5]. However, the presence of OB [12] is sought and is used as another criterion for the diagnosis of BOS (see below). (3) Recipients were determined to be free of BOS (and Table 1. Numeric Staging System for Bronchiolitis Obliterans Syndrome FEV 1 BOS Stage Severity (% of Baseline) Subcategory a 0 No abnormality >80% a, b 1 Mild 66-80% a, b 2 Moderate 51-65% a, b 3 Severe -<50% a, b = Absence of pathologic evidence of bronchiolitis obliterans; b = presence of bronchiolitis obliterans pathologically; BOS bronchiolitis obliterans syndrome; FEV 1 - forced expiratory volume in I second. placed in group I) if their FEV 1 remained at 80% or more of their baseline value and there was no histologic evidence of OB (BOS stage 0a; see below). (4) Recipients were defined as having BOS (and placed in group II) if there was a sustained and significant fractional decline in their FEV 1 after 3 months from the date of the transplantation. (The decline in FEV1 was considered sustained and significant when the average of two FEV 1 values, obtained at least I month apart, was less than 80% of the previously established baseline value.) Recipients were also considered positive for BOS if OB was documented histologically. (5) The conventions for BOS numeric stages (a means of quantifying the severity of BOS) and subcategory designations [5] were used, as summarized in Table 1. For each numeric stage, the subcategory designation "a" signifies no histologic evidence of OB, whereas the designation "'b'" signifies histologic presence of OB. Therefore, according to this convention, all patients in group I (free of BOS) are denoted as being in stage oa. Characterization of Clinical Features and Progression of BOS The mean time interval until the onset of BOS was determined. To evaluate the time course of this syndrome, several indices were compared between the four consecutive quartiles (n = 53 patients in each) making up this series. The main difference between these quartiles was simply the available duration of recipient follow-up, which was longer in earlier quartiles and shorter in the more recent quartiles. The indices compared included the number of recipients at risk for BOS, the number (and percentage) of those at risk in whom BOS ultimately developed, the time until onset of BOS, and the mean BOS stage. Clinical variable analysis included age, sex, diagnosis (indication for transplantation), and type of transplantation performed (single lung transplantation [SLT] versus bilateral lung transplantation [BLT]). Distributions were determined for groups I and II to look for a possible association between any of these variables and the development of BOS. Absolute mortality was compared between group I and group II. Also analyzed were the mean interval (in months) until BOS was fatal in group II recipients and the distribution of deaths according to the type of transplantation performed (SLT versus BLT) and according to the recipient's underlying diagnosis in the two groups.

3 Ann Thorac Surg SUNDARESAN ET AL ;60: BOS IN LUNG TRANSPLANTATION Table 2. Clinical Data by Quartile for Lung Transplant Recipients With and Without Bronchiolitis Obliterans Syndrome Patient Group/Follow-Up Transplantations At Risk Group I (BOS-) (too) Performed for BOS No. (%) No. (%) Group II (BOS+) Time to Onset (mo) Mean BOS Stage Overall/ /187 (59%) 77/187 (41%) 16.1 _ _ Quartile 1/ (44.2%) 24/43 a (55.8%) 17.5 _+ 2.1 b 2.21 _ c Quartile 2/ /45 (33.3%) 30/45 (66.7%) 19.1 _ ± 0.19 Quartile 3/ (61.7%) 18/47 (38.3%) 11.5 _ _ Quartile 4/ /52 (90.4%) 5/52 a (9.6%) 5.6 _+ 0.9 b 0, c a p < b p < c p = BOS = bronchiolitis obliterans syndrome. Data Analysis Whenever applicable, data are presented as mean +_ standard error of the mean. Statistical comparison of values between the groups was performed using the unpaired t test, Fisher's exact test, or the Mann-Whitney nonparametric test, when appropriate. Values were considered significantly different at p values less than Results Prevalence and Criteria of Bronchiolitis Obliterans Syndrome Of 208 lung transplant recipients, 187 recipients (90%) survived at least 3 months after transplantation, putting them at risk for BOS. One hundred ten recipients at risk remained free of BOS (stage 0a, 59%), whereas 77 recipients at risk experienced BOS (41%) (Table 2). The criteria signifying the onset of BOS were declining spirometry in 57 of 77 (74%), a biopsy specimen positive for OB in 14 of 77 (18.2%), and both simultaneously in 6 of 77 (7.8%). However, at completion of follow-up, only 40 of 77 recipients with BOS (52%) had evidence of declining spirometry alone, 30 of 77 (38.9%) had evidence of both declining spirometry and histologic OB, and only 7 of 77 (9.1%) had a biopsy specimen positive for OB as the sole criterion of BOS. Clinical Features and Progression of Bronchiolitis Obliterans Syndrome The mean time until onset of BOS was 16.1 _+ 1.2 months after transplantation (Table 2). Because this suggested a degree of latency, a quartile analysis was performed, as summarized in Table 2. In this analysis, the prevalence, the mean interval until onset of BOS, and the mean BOS stage were compared between the four consecutive quartiles of patients making up this series. The principal reason for performing this comparison was the difference in the durations of recipient follow-up between these quartiles; for example, quartile 1 (QTR1, the first of the four consecutive quartiles) offered a range of 46 to 72 months of follow-up, whereas quartile 4 (QTR4, the last quartile) offered a range of only 4 to 17 months of follow-up. In progressing from QTR1 to QTR4, the following were noted: (1) a decrease in the prevalence of BOS (QTR1, 24 with BOS of 43 at risk [55.8%[ versus QTR4, 5 with BOS of 52 at risk [9.6%]; p < 0.001), (2) a shorter duration until onset of BOS (QTR1, months versus QTR4, months; p < 0.001), and (3) a lower mean BOS stage (QTR1, versus QTR4, ; p = 0.007). Clinical indices were compared between groups and II. Mean age (group I, 43.6 ~ 1.1 years versus group II, years; p = 0.34) and sex distribution (group I, males 52% versus group II, males 38%; p ) did not differ between the groups. Figure 1A, which shows the distribution of the indications for transplantation in groups I and II, demonstrates no difference between groups with respect to the underlying diagnosis (p = 0.360). There was a nearly identical distribution of the various types of procedures in the two groups: group I, 55 SLT (50%), 54 BLT (49.1%), and one en bloc double lung A in n "6 E o= #_ GROUP I (BOS-) i Ski DIAGNOSIS [] COPD [] ATDef CF IPF [] PVD "Other" p = GROUP U (BOS ) Type of Transplant B Fig 1. (A) Comparison of distribution of diagnoses in the two lung transplant recipient groups. There was no difference between groups (p 0.360). (ATDef - antitrypsin deficiency emphysema; BOS - bronchiolitis obliterans syndrome; CF = cystic fibrosis; COPD = chronic obstructive pulmonary disease; IPF = idiopathic pulmonary fibrosis; "Other" = miscellaneous indications; PVD = pulmonary vascular disease.) (B) Distribution of BOS stages according to type of transplantation performed. There was no difference between patients undergoing single lung transplantation (SLT) or bilateral lung transplantation (BLT) (p ). I //////////////J IIIIIIMIII BLT BOS STAGE DOA lob ~t 12 []3

4 1344 SUNDARESAN ET AL Ann Thorac Surg BOS IN LUNG TRANSPLANTATION 1995;60: transplantation (0.9%) versus group II, 39 SLT (50.7%), 37 BLT (48.1%), and 1 double lung transplantation (1.2%) (p = 0.962). These data are depicted in a slightly different manner in Figure 1B, which shows a comparable distribution of the four BOS stages along with the percentage of recipients free of BOS (stage 0a) in recipients of SLT and BLT (p = 0.406). Absolute mortality was significantly higher among recipients in whom BOS developed (group II, 22 of 77 [28.6%] versus group I, 8 of 110 [7.3%]; p < 0.001) (Fig 2A). In patients with BOS who died, fatality ensued within 11.5 _+ 2.4 months of the onset of BOS. In comparing groups I and II with respect to deaths, there was no significant difference between groups with regard to the type of transplantation performed (Fig 2B) or the underlying diagnosis (Fig 2C). Comment In their 1993 report, Cooper and colleagues [5] summarized their recommendations for a clinically applicable system for the staging of chronic lung allograft dysfunction. They proposed the use of the term "bronchiolitis obliterans syndrome" to refer to this entity and offered a number of pertinent guidelines. This group acknowledged that ongoing reappraisal of their formulation would be appropriate after sufficient experience was gained with a number of patients. In the current report, we have applied their conventions retrospectively to a large 6-year single-institution experience with lung transplantation and identified a substantial subset of patients in whom BOS developed. A number of our findings merit consideration. Our data suggest strongly that BOS is a clinical syndrome that is not always demonstrable as a pathologic entity. In our analysis, the initial feature heralding the onset of BOS was declining spirometry alone in 57 of 77 recipients at risk (74%); ultimately, 40 of 77 (52%) were still defined as having BOS solely on the basis of physiologic decline. The initial sign of BOS was histologic evidence of OB in only 14 of 77 (18.2%), and ultimately, only 7 of 77 (9.1%) were deemed as having BOS solely on this basis. These findings may explain the conflicting opinions expressed in early reports dealing with the prevalence of this entity [3]. Krarner and associates [13] reviewed their experience with 16 patients with advanced BOS (defined by physiologic decline) and showed that only seven of 42 specimens (15.2%) of transbronchial lung biopsies obtained from these patients were positive for OB. In addition to demonstrating the low sensitivity of transbronchial lung biopsy in detecting OB, their data lend support to the general belief that a comprehensive definition of this entity should include the use of physiologic indices, as proposed by Cooper and co-workers [5], and not only histologic documentation of OB. Open lung biopsy will increase the detection of histologic OB, but is obviously a more invasive alternative and not practical for routine patient surveillance. Open lung biopsy is better restricted to the evaluation of serious graft dysfunction after all other less invasive approaches have failed to distinguish between infection and rejection. In our experience, virtually all patients who died of this syndrome were found to have OB at autopsy even when it was not detected before death, providing the ultimate correlation between the observed functional decline and this pathologic "hallmark." One of the questions arising from this study is: What is c- O A = #_ " 50 25' 0 p < I f/////////////////////~" Group I (BOS-) Group II (BOS*) p = p = ' ' ' ' E " SLT BLT SLT BLT Group I (BOS-) Group II (BOS +) --- p=0.193 I 'ij p=0.803 COPD ATDef CF IPF PVD "Other" COPD ATDef CF IPF PVD "Other" Group I (BOS-) Group II (BOS ) STATUS r-lalive [] Dead STATUS I-IAlive [] Dead i ~ STATUS [] Alive [] Dead Fig 2. (A) Mortality rates for recipients without bronchiolitis obliterans syndrome (BOS) (group I) and those with BOS (group II) at the completion of follow-up. Absolute mortality was significantly higher in recipients in whom BOS developed (p < 0.001). (B) Comparison of mortality rates according to type of transplantation performed in the two groups. Mortality did not correlate with &e type of transplantation in group I (p = 0.978) or in group 1I (p ). (BLT = bilateral lung transplantation; SLT = single lung transplantation.) (C) Comparison of mortality rates according to diagnosis in the two groups. Mortality did not correlate with diagnosis in group I (p = 0.193) or in group II (p = 0.803). (ATDef = antitrypsin deficiency emphysema; CF cystic fibrosis; COPD = chronic obstructive pulmonary disease; IPF = idiopathic pulmonary fibrosis; "Other" - miscellaneous indications; PVD - pulmonary vascular disease.)

5 Ann Thorac Surg SUNDARESAN ET AL ;60: BOS IN LUNG TRANSPLANTATION the significance of a biopsy specimen positive for OB in the recipient who has been clinically and physiologically stable previously? It would be useful to know whether this histologic change is a predictor of later functional decline, because the fibrosis of OB is irreversible once established. This situation applied to 14 recipients in this study. We found that 7 of these 14 (50%) later had physiologic decline, and 3 of these 7 died of BOS. However, it was intriguing that the other 7 remained in BOS stage 0b throughout their posttransplantation course and that 6 of these 7 recipients are alive without signs of physiologic decline 8 to 42 months (mean, 27.3 months) after transplantation. The data from this small subgroup, in the context of our entire analysis, might lead one to speculate that histologic OB can be identified in virtually all lung allografts given sufficiently meticulous tissue sampling, but that the true prevalence of clinical graft dysfunction is still only about 50% overall. The reason that only a subgroup of recipients manifests graft dysfunction remains an unanswered but important question. Our analysis showed that overall, in 41% of recipients at risk, BOS developed after a mean interval of about 16 months after transplantation. However, our quartile analysis yielded several noteworthy observations. First, the number of recipients at risk for BOS increased from QTR1 (43 at risk of 53 transplantations) to QTR4 (52 at risk of 53 transplantations). This chronologic improvement in early survival is due to a number of factors. First, en bloc double lung transplantation was used frequently in our early experience. This procedure is associated with an increased operative mortality rate [9] and was soon replaced by the bilateral sequential technique. Furthermore, there has been a general improvement in virtually all facets of intraoperative and perioperative patient care as a result of greater clinical experience. Second, a true appreciation of the prevalence, severity, and latency of BOS is derived from the early quartiles, which provided the longest duration of follow-up. Our data showed that the prevalence and the mean BOS numeric stage were both significantly higher in QTR1 than in QTR4. Thus, although our analysis of the BOS group overall suggests that the prevalence of this syndrome is 41%, it may actually exceed 50% (based on the 55.8% and 66.7% prevalences observed in QTR1 and QTR2, respectively, in comparison with the 9.6% prevalence seen in QTR4). Similarly, the adverse effect of BOS on graft function was more apparent with longer follow-up, with a mean BOS stage of in QTR1, compared with 0.80 _ in QTR4. Also, although the BOS group overall demonstrated a mean interval until onset of the syndrome of 16.1 months, this interval may actually be slightly longer (based on the and 19.1-month intervals seen in QTR1 and QTR2, respectively). (The short interval until BOS onset of only 5.6 months seen in QTR4 was likely the result of having a limited duration of follow-up of a small number of recipients with BOS in this most recent quartile.) In summary, our findings confirm the notion held by many that BOS is indeed a progressive, lateonset syndrome of graft dysfunction. Unfortunately, the data also suggest, in contrast to earlier reports [4], that BOS eventually affects a large percentage of long-term survivors, perhaps even the majority. There is considerable evidence from previous clinical [14-16] and experimental [17, 18] studies that BOS has an immunologic basis, leading to the widely held belief that it represents a chronic immune response directed against the lung allograft. Although the current study was not intended to be a detailed analysis of risk factors, our comparison of groups I and II showed that BOS affects all subgroups of lung allograft recipients without distinction for recipient age, sex, the type of transplantation performed, or the underlying diagnosis. In a previous study of the first 112 consecutive lung transplantations performed at our institution [19], we performed a more formal analysis of several clinical and immunologic variables in comparing 54 recipients free of BOS and 40 recipients with BOS, using the same BOS criteria as in the current study. That study revealed a significant correlation between the development of lymphocytotoxic anti-human lymphocyte antigen antibodies and the development of BOS [19], similar to the association demonstrated by previous investigators between the development of such antibodies and chronic cardiac allograft rejection [20]. That correlation [19] demonstrated further that lung transplant recipients in whom BOS develops give evidence of allosensitization, and may also imply a mechanistic role for these antibodies in the genesis of BOS. We are currently updating that analysis to determine further the significance of that finding. The true pathogenesis of BOS is unknown, but because it is believed to arise from an immune response to the allograft, we and other centers [21, 22] have treated it empirically using augmented immunosuppression protocols. This has consisted of bolus treatment with corticosteroids as well as antilymphocytic agents. Despite the use of these modalities, our analysis has shown that absolute mortality rates are significantly higher for lung transplant recipients with BOS and that this mortality does not correlate with the type of transplantation or the underlying diagnosis. Once again, the true mortality rate of BOS may be better reflected through the quartile analysis (data not shown); although the mortality in group II overall was 22 of 77 (28.6%), it was 10 of 24 (42%) in QTR1. Our study did not show a significant difference in mortality between QTR1 (10 of 24, 42%) and QTR4 (1 of 5, 20%; p = versus QTR1), but this was likely the result of the small number of patients in QTR4. In conclusion, our data suggest that BOS is indeed a clinical syndrome of lung allograft dysfunction that is frequently associated with histologic evidence of OB, that it displays considerable latency in its onset and progression, that it is far more prevalent among recipients of isolated lung transplants than has been proposed by previous investigators, and that despite all currently available treatment modalities, it carries a high mortality rate. It is the greatest source of late mortality and morbidity in lung transplant recipients and is the greatest obstacle to long-term survival in clinical lung transplantation today. We gratefully acknowledge the help of Richard B. Schuessler, PhD, with the statistical analysis.

6 1346 SUNDARESAN ET AL Ann Thorac Surg BOS IN LUNG TRANSPLANTATION 1995;60: References 1. Cooper JD, Patterson GA, Trulock EP, and the Washington University Lung Transplant Group. Results of single and bilateral lung transplantation in 131 consecutive recipients. J Thorac Cardiovasc Surg 1994;107: Burke CM, Theodore J, Dawkins KD, et al. Post-transplant obliterative bronchiolitis and other late lung sequelae in human heart-lung transplantation. Chest 1984;86: Burke CM, Theodore J, Baldwin JC, et al. Twenty-eight cases of human heart-lung transplantation. Lancet 1986;1: Anzueto A, Levine SM, Bryan CL, et al. Obliterative bronchiolitis in single lung transplant recipients. Am Rev Respir Dis 1992;145:A Cooper JD, Billingham M, Egan T, et al. A working formulation for the standardization of nomenclature and clinical staging of chronic dysfunction in lung allografts. J Heart Lung Transplant 1993;12: Sundaresan S, Trachiotis GD, Aoe M, Patterson GA, Cooper JD. Donor lung procurement: assessment and operative technique. Ann Thorac Surg 1993;56: Cooper JD, Pearson FG, Patterson GA, et al. Technique of successful lung transplantation in humans. J Thorac Cardiovasc Surg 1987;93: Patterson GA, Cooper JD, Goldman B, et al. Technique of successful clinical double-lung transplantation. Ann Thorac Surg 1988;45: Pasque MK, Cooper JD, Kaiser LR, Haydock DA, Triantafillou A, Trulock EP. Improved technique for bilateral lung transplantation: rationale and initial clinical experience. Ann Thorac Surg 1990;49: Trulock EP. Management of lung transplant rejection. Chest 1993;103: Trulock EP, Ettinger NA, Brunt EM, Pasque MK, Kaiser LR, Cooper JD. The role of transbronchial lung biopsy in the treatment of lung transplant recipients: an analysis of 200 consecutive procedures. Chest 1992;102: Yousem SA, Berry GJ, Brunt EM, et al. A working formula- tion for the standardization of nomenclature in the diagnosis of heart and lung rejection: Lung Rejection Study Group. J Heart Lung Transplant 1990;9: Kramer MR, Stoehr C, Whang JL, et al. The diagnosis of obliterative bronchiolitis after heart-lung and lung transplantation: low yield of transbronchial lung biopsy. J Heart Lung Transplant 1993;12: Taylor PM, Rose ML, Yacoub MH. Expression of MHC antigens in normal human lungs and transplanted lungs with obliterative bronchiolitis. Transplantation 1989;48: Yousem SA, Curley JM, Dauber J, et al. HLA-class II antigen expression in human heart-lung allografts. Transplantation 1990;49: Milne DS, Gascoigne A, Wilkes J, et al. The immunohistopathology of obliterative bronchiolitis following lung transplantation. Transplantation 1992;54: Uyama T, Winter JB, Groen G, Wildevuur CR, Monden Y, Prop J. Late airway changes caused by chronic rejection in rat lung allografts. Transplantation 1992;54: A1-Dossari GA, Kshettry VR, Jessurun J, Bolman RM. Experimental large-animal model of obliterative bronchiolitis after lung transplantation. Ann Thorac Surg 1994;58: Sundaresan S, Mohanakumar T, Phelan D, et al. Development of cytotoxic antibodies post-lung transplantation correlates with the development of bronchiolitis obliterans syndrome. Surg Forum 1994;45: Suciu-Foca N, Reed E, Marboe C, et al. The role of anti-hla antibodies in heart transplantation. Transplantation 1991;51: Glanville AR, Baldwin JC, Burke CM, Theodore J, Robin ED. Obliterative bronchiolitis after heart-lung transplantation: apparent arrest by augmented immunosuppression. Ann Intern Med 1987;107: Paradis IL, Duncan SR, Dauber JH, et al. Effect of augmented immunosuppression on human chronic lung allograft rejection. Am Rev Respir Dis 1992;145:A705. DISCUSSION DR THOMAS M. EGAN (Chapel Hill, NC): I congratulate Dr Sundaresan for a very concise presentation, which represents an extensive audit of their program to establish the incidence and prevalence of posttransplantation bronchiolitis obliterans syndrome, or BOS. They have brought to our attention the alarming frequency of development of this complication, and they have also emphasized the heterogeneity of this clinical syndrome. Our data at the University of North Carolina are similar. Among 72 lung transplant recipients surviving longer than 3 months who are at risk the distribution of prevalence is 39 patients in whom BOS has not developed and 33 patients in whom BOS or biopsy-proven obliterative bronchiolitis has developed. There are 3 patients in grade 1, 8 in grade 2, and 15 in grade 3; 14 of the 33 patients with BOS have died. Rather than assess these data by quartile, we have calculated the actuarial freedom from BOS, and I believe that the current report would be strengthened by inclusion of this analysis. This depicts the time course to development of BOS, and it is clear that in most patients, BOS will begin to develop between 6 and 30 months. A good study raises additional questions, and this one is no exception. This study clearly begs the question, why do so many patients have development of BOS? Perhaps the better query is why BOS does not develop in some patients. I have several questions for Dr Sundaresan. Have you performed any multivariate analyses to identify risk factors for the development of BOS? Although you have noted a higher risk of mortality among patients with BOS, how many of your BOS patients remained stable and what is the incidence of progression from grade 1 to 2, from 2 to 3, and so on? Could you elaborate on specifically when to treat patients whose FEV 1 begins to fall and when to resort to cytolytic therapy? Given that some patients progress and that many patients succumb before a grade 3 diagnosis is made, could you comment on the utility of the current BOS grading system? Do you have any suggestions to improve it based on this analysis? And finally, I take issue with the concluding statement in the manuscript that BOS poses the most serious problem in clinical lung transplantation today. Although it is the greatest source of late mortality after lung transplantation, we have had far more patients than 14 die before transplantation because of the critical shortage of suitable pulmonary donors. And, parenthetically, we could perhaps offer retransplantation to patients with severe BOS if we could solve the donor shortage. DR RICHARD J. NOVICK (London, Ontario, Canada): I also congratulate Dr Sundaresan on his presentation. The pulmonary retransplant registry was established in 1991 to document the outcome and predictors of survival after retransplantation. Approximately 55% of the 150 patients in the registry have had retransplantation for OB. At 1 year after retransplantation for OB, 66% of patients have normal pulmonary function, whereas 14% have recurrent, stage 3 (ie, severe) OB. At 2 years only half the patients have normal pulmonary function, whereas 33% have severe OB. Absolute FEV1 declined 11% and 27% at 1 and 2 years after retransplantation, respectively, compared with baseline values after retransplantation. The two questions I have are similar to those of Dr Egan. First,

7 Ann Thorac Surg SUNDARESAN ET AL ;60: BOS IN LUNG TRANSPLANTATION do you have actuarial data on the prevalence of BOS stages 1, 2, and 3 in your patients after primary lung transplantation, as we do after retransplantation? Second, in view of the fact that Dr Cooper's article last year did not demonstrate a correlation between the number of acute early rejection episodes and the development of OB, have you performed a multivariate analysis to determine which factors correlate with the subsequent development of OB after lung transplantation? DR DAVID J. SUGARBAKER (Boston, MA): I enjoyed your report very much. I would like to ask you if you could correlate the incidence of BOS with the number or severity of acute rejection episodes in the immediate postoperative period. There are some data from the solid-organ transplant groups showing that the number and severity of acute rejection episodes do correlate with long-term development of BOS. DR ROBERT J. KEENAN (Pittsburgh, PA): I enjoyed your presentation very much, and we too have made the same surprising finding that single lung and bilateral lung transplants seem to do just as well after the diagnosis of OB, which is not something that we would have intuitively thought. I wonder whether the early patients were somewhat discriminated against because BOS was only recently described, so that the people in the third and fourth quartiles of your analysis might have been treated more aggressively after this syndrome was recognized, as opposed to those in the first or the second quartile? The other issue I would address with you is whether the more aggressive treatment of this syndrome might have come at a cost of infectious complications, such as what we have seen at Pittsburgh, with a surprising frequency of fungal or opportunistic viral infections after giving aggressive cytolytie therapy. DR CHRISTOPHER G. A. McGREGOR (Rochester, MN): Again, I congratulate Dr Sundaresan on a fine presentation. Could I ask him to allude briefly to the current management strategy for this problem at St. Louis? It is clearly a depressing medium-term survival after lung transplantation, and I would be grateful for some guidance as to their current management strategy. I would briefly mention that we are now undertaking a protocol of total lymphoid irradiation in selected patients with OB to try to mitigate the progression of this disease. DR JOHN R. BENFIELD (Sacramento, CA): You say that the true pathogenesis of bronchiolitis obliterans is unknown. That is true. But I do not think that the true pathogenesis of bronchiolitis obliterans in conjunction with lung transplantation is unknown. It seems to me that it is chronic transplant rejection. Is that too sweeping a statement? What is wrong with changing the terminology from bronchiolitis obliterans to chronic transplant rejection syndrome? DR SUNDARESAN: I thank Dr Benfield and all the other discussants for their remarks. If I can just discuss Dr Benfield's questions first, I think there is an overwhelming amount of evidence that suggests that this syndrome, BOS, is the result of an immune-mediated allograft injury. However, a lot of that evidence is indirect in a variety of ways. And although the pathology of this entity resembles that seen in other solid-organ allografts, it has not been definitively proved that this is an immune-mediated phenomenon. I think, however, that most of us do treat it as such. I agree with Dr Egan's remark that it is a very intriguing question, not only why so many patients get BOS but also, in looking at those who have survived for a long period without it, why those patients do not get it. Second, he asked, as did Dr Novick, I believe, whether we have done a multivariate analysis. In a fairly detailed analysis of the first 112 consecutive trans- plantations done at Barnes, we compared a group of 54 patients without BOS to a group of 40 with BOS using the same criteria as in this study. We looked at a number of clinical factors, including the ones that I showed here today, as well as graft ischemic time and use of cardiopulmonary bypass for the transplantation. However, the focus of that study was to see whether we could find any immunologic indices that correlated with the development of BOS. So we looked at human lymphocyte antigen tissue typing (that is, A, B, and DR antigen mismatches), the incidence of positive donor-specific cross match, as well as the development of lymphocytotoxic anti-human lymphocyte antigen antibodies by the panel reactive antibody method. We did not do a multivariate analysis in that study, but after systematic analysis of all those variables, the development of these anti-human lymphocyte antigen antibodies was the only factor to emerge as correlating with the development of BOS. I think Dr Egan also wanted to know about the progression of patients once they are diagnosed with BOS. For example, if we identify 20 patients who are in BOS stage 1 today, what is their fate in 3 months or 6 months, or even 12 months from now? That is actually the subject of a study just being completed by my colleagues. I do not have precise data to show you, but I can just summarize the overall result by saying that virtually all patients identified with BOS do, with time, move to worse stages. It is a distinct minority that actually move to lower stages of BOS over time. Another question Dr Egan asked was when to treat a patient when the FEV 1 declines. This question coincides with Dr McGregor's question regarding management of BOS, so I will deal with these together. In our program, when a patient is found to manifest signs of declining spirometry, the patient undergoes fiberoptic bronchoscopy, mainly to rule out other specific causes of their dysfunction. Once anastomotic and septic problems have been ruled out, these patients receive treatment with bolus administration of steroids as well as a course of cytolytic therapy, and then they are followed up very carefully with pulmonary function tests at approximately 2-week intervals. Doctor Egan asked about ways that we could try to improve the staging system. The main comment I have is that the current staging system has improved our understanding of this whole phenomenon, mainly because it deals not only with pathologically proven disease but also with the aspect of physiologic decline. I think this actually is a strong point in favor of the current system that is being used. Doctor Egan took exception to my last statement that BOS is the most pressing problem of clinical lung transplantation. I certainly have to agree that the shortage of donors, making it such a scarce resource, is clearly right up there. Doctor Novick asked me about actuarial data for patients in stages 1, 2, and 3. I think I have partly dealt with that issue already. In general, I would agree that an actuarial approach to our data would be very useful. The main impact of our presentation here is that we can state an absolute prevalence of this problem given a sufficient period of follow-up, which in our experience was about 2.5 years. Doctor Keenan asked whether the early patients in our series were perhaps disadvantaged with respect to their treatment. I probably would have to agree, because at that time it was not truly appreciated that their graft dysfunction was the consequence of this entity. He also asked about whether the aggressive augmented immunosuppression protocol we use is associated with infectious complications, and I think that that is indeed the case in our experience. In fact, data from our lung transplant registry show that the most common causes of late death in lung allograft recipients are not only this entity, BOS, but also the septic events that ensue as a complication of augmented immunosuppression therapy.