Published ahead of print on December 7, 2006, doi:10.1164/rccm.200608-1079OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200608-1079OC
Impact of Immediate Primary Lung Allograft Dysfunction on Bronchiolitis Obliterans Syndrome1 Division of Pulmonary and Critical Care and 2 Division of Cardiothoracic Surgery, Washington University School of Medicine, St. Louis, Missouri Correspondence and requests for reprints should be addressed to Ramsey R. Hachem, M.D., Washington University School of Medicine, Division of Pulmonary and Critical Care, 660 S. Euclid Avenue, Campus Box 8052, St. Louis, MO 63110. E-mail: rhachem{at}im.wustl.edu
Rationale: Primary graft dysfunction is a common complication after lung transplantation and a significant risk factor for short- and long-term mortality. Objective: We examined the impact of primary graft dysfunction on bronchiolitis obliterans syndrome. Methods: We performed a retrospective cohort study of 334 adult lung transplant recipients at our program and graded the severity of primary graft dysfunction according to the International Society for Heart and Lung Transplantation definition. We evaluated the impact of primary graft dysfunction on acute rejection, lymphocytic bronchitis, and bronchiolitis obliterans syndrome stage 1, using univariable and multivariable Cox proportional hazards models. Main Results: Among the 334 recipients, 65 did not have primary graft dysfunction (grade 0), 130 had grade 1, 69 had grade 2, and 70 had grade 3. In the univariable analysis, all grades of primary graft dysfunction were associated with a significantly increased risk of bronchiolitis obliterans syndrome stage 1 (grade 1: relative risk [RR] = 1.73; grade 2: RR = 2.13; and grade 3: RR = 2.53, compared with grade 0). The multivariable model demonstrated that the increased risk of bronchiolitis obliterans syndrome associated with primary graft dysfunction was independent of acute rejection, lymphocytic bronchitis, and community-acquired respiratory viral infections. However, there was no association between primary graft dysfunction and acute rejection or lymphocytic bronchitis. Conclusions: Primary graft dysfunction is associated with an increased risk of bronchiolitis obliterans syndrome independent of acute rejection, lymphocytic bronchitis, and community-acquired respiratory viral infections, and this risk is directly related to the severity of primary graft dysfunction.
Key Words: lung transplantation • primary graft dysfunction • bronchiolitis obliterans syndrome
Primary graft dysfunction (PGD) is a common early complication after lung transplantation (1–3). It represents acute lung injury with variable severity ranging from mild radiographic pulmonary infiltrates to severe hypoxemic respiratory failure requiring intensive support. Because of this broad spectrum of severity and the absence of a standard definition, the incidence and outcomes of PGD have varied widely in the literature (3–7). Thus, the International Society for Heart and Lung Transplantation (ISHLT) proposed a standard definition and grading criteria for PGD based on the chest X-ray appearance and the ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO2/FIO2) (8). According to this definition, grade 0 is characterized by a PaO2/FIO2 ratio greater than 300 mm Hg and a clear chest X-ray, grade 1 by a PaO2/FIO2 ratio greater than 300 mm Hg and radiographic infiltrates consistent with pulmonary edema, grade 2 by a PaO2/FIO2 ratio of 200–300 mm Hg and pulmonary infiltrates, and grade 3 by a PaO2/FIO2 ratio less than 200 mm Hg and pulmonary infiltrates (8). Exclusion of other potential causes of graft dysfunction, such as hyperacute rejection, venous anastomotic complications, cardiogenic pulmonary edema, and pneumonia, is implicit in this definition. This ISHLT definition was validated in two single-center retrospective studies. These studies demonstrated that recipients who developed PGD grade 3 had significantly higher short- and long-term mortality compared with those who did not have PGD (9, 10). Similarly, in a retrospective review of a United Network for Organ Sharing and ISHLT registry, recipients who developed PGD had significantly worse survival 30 days after transplantation than those who did not have PGD (3). Furthermore, after excluding recipients who died in the first year after transplantation, those who developed PGD had significantly worse long-term survival than those without PGD, suggesting that PGD continues to hamper survival beyond the early postoperative recovery period (3). Because bronchiolitis obliterans syndrome (BOS) is a leading cause of death beyond the first year after transplantation (11–13), we hypothesized that survivors of PGD are at increased risk for the development of BOS and sought to investigate this relationship in this study.
Study Design and Patient Population We conducted a retrospective cohort study to determine the risk of BOS among lung transplant recipients who developed PGD. Between January 1, 1998, and June 30, 2004, 336 adults underwent 337 lung transplantation procedures at our program. Three recipients developed hyperacute rejection and their postoperative courses were excluded from this study; among these, two died and one underwent retransplantation. The postoperative course for the recipient who underwent retransplantation was included in this cohort. Overall, 334 recipients were included in this study, and we reviewed their baseline characteristics, operative data, and postoperative outcomes. The general medical management and follow-up protocols of our program have been previously detailed (14), and a description is included in the online supplement. The study protocol was approved by the Washington University School of Medicine Institutional Review Board for human studies before data acquisition.
Variables
Statistical Analysis
Follow-up was completed through May 1, 2006. The mean duration of observation per patient was 3.7 ± 2.1 years and the study included 1,236 person-years of follow-up. Immediately after transplantation, 65 (19%) recipients did not have PGD (grade 0), 130 (39%) had PGD grade 1, 69 (21%) had PGD grade 2, and 70 (21%) had PGD grade 3. Recipients who developed PGD grade 3 were more likely to have a preoperative diagnosis of pulmonary hypertension or pulmonary fibrosis than those who did not have PGD and those who had PGD grade 1 (Table 1). In contrast, those who did not have PGD and those who had PGD grade 1 were more likely to have a preoperative diagnosis of emphysema or cystic fibrosis. In addition, recipients who required intraoperative cardiopulmonary bypass were more likely to have PGD grade 3. However, interpreting this association is difficult because recipients who develop PGD grade 3 intraoperatively are more likely to require cardiopulmonary bypass support. Other baseline characteristics including recipient and donor age, recipient and donor sex, mean ischemic time, the proportion of organs with ischemic times greater than 330 min (20), the proportion of marginal donors (as defined in the online supplement), and the operation performed were not significantly different between the four groups (Table 1). Baseline lung function is shown in the online supplement (see Table E1).
We excluded 14 recipients who died in the first 90 days after transplantation from the BOS analysis because they were not evaluable for BOS (18) and to avoid the potential confounding bias of multiple competing events (i.e., death and BOS). Among the remaining 320 recipients, 161 developed BOS, and the median time to the onset of BOS was 3.89 years. Recipients who did not have PGD (grade 0) had significantly higher freedom from BOS than those who developed any grade of PGD (Figure 1). We explored potential risk factors for BOS, using univariable Cox regression models, and identified an acute rejection grade A2 or higher, lymphocytic bronchitis grade B2 or higher, CARV infection, and all grades of PGD as significant risk factors for BOS (Table 2). The direct relationship between the severity of PGD and the relative risk of BOS is noteworthy. Recipients with PGD grade 1 had a relative risk of BOS of 1.73, those with grade 2 had a relative risk of 2.13, and those with grade 3 had a relative risk of 2.53 (Table 2). In addition, single-lung transplantation, cardiopulmonary bypass, and the preoperative diagnosis were associated with BOS, but these associations were not statistically significant (Table 2).
To examine the impact of PGD on BOS in the context of other recognized risk factors for BOS and to test our a priori hypothesis, we constructed a multivariable Cox proportional hazards model that included an acute rejection grade A2 or higher, lymphocytic bronchitis grade B2 or higher, CARV infection, and PGD. All grades of PGD were significant risk factors for BOS independent of acute rejection grade A2 or higher, lymphocytic bronchitis grade B2 or higher, and CARV infection (Table 3). Of note, a direct relationship between the severity of PGD and the relative risk of BOS persisted in the multivariable model. Recipients with PGD grade 1 had a relative risk of BOS of 1.68, those with grade 2 had a relative risk of 2.04, and those with grade 3 had a relative risk of 2.61 (Table 3). In addition, acute rejection grade A2 or higher and lymphocytic bronchitis grade B2 or higher were significant risk factors for BOS independent of each other, PGD, and CARV infection. CARV infection was associated with an increased risk of BOS, but this did not reach statistical significance (relative risk [RR], 1.53; 95% confidence interval [CI], 0.9–2.6; p = 0.119) (Table 3). Furthermore, we forced other potential risk factors for BOS identified in the univariable models and variables that were significantly different at baseline among the four groups of PGD, including the preoperative diagnosis, cardiopulmonary bypass, and single-lung transplantation, into the multivariable model individually. None of these improved the –2 log likelihood ratio of the overall model and none were statistically significant. Importantly, these additional covariates did not alter the relative risks or statistical significances of the original covariates, and our final model is the original hypothesis-driven model. Lastly, because this cohort consisted of a disproportionate number of bilateral recipients, we repeated the analysis excluding single-lung recipients and the results were unchanged (see Table E2).
To examine the potential impact of PGD on acute rejection and lymphocytic bronchitis we performed univariable Cox proportional hazards model analysis with acute rejection and lymphocytic bronchitis as the dependent variables separately. No grade of PGD was a significant risk factor for acute rejection grade A2 or greater. Recipients with PGD grade 1 had a relative risk of acute rejection of 0.77 (95% CI, 0.5–1.2; p = 0.2), those with PGD grade 2 had a relative risk of 1.1 (95% CI, 0.7–1.7; p = 0.7), and those with PGD grade 3 had a relative risk of 0.7 (95% CI, 0.4–1.2; p = 0.15). Similarly, no grade of PGD was a significant risk factor for lymphocytic bronchitis grade B2 or higher. Those with PGD grade 1 had a relative risk of lymphocytic bronchitis of 0.95 (95% CI, 0.5–1.8; p = 0.9), those with PGD grade 2 had a relative risk of 0.8 (95% CI, 0.4–1.6; p = 0.5), and those with PGD grade 3 had a relative risk of 0.8 (95% CI, 0.4–1.7; p = 0.6). Eight recipients who had PGD grade 3 (11%) died within 90 days of transplantation compared with none of those who did not have PGD (grade 0), five who had PGD grade 1 (4%), and one who had PGD grade 2 (1%) (p = 0.004). We analyzed long-term survival conditional on survival to 90 days after transplantation, using univariable Cox proportional hazards models, and detected an increased risk of death for recipients who developed PGD grade 3 (RR, 1.92; 95% CI, 1.02–3.6; p = 0.04) (Table 4). PGD grades 1 and 2 were associated with trends to increased risks of death, but these were not statistically significant (Table 4). BOS (as a time-dependent variable) and single-lung transplantation were the only significant risk factors for death (RR, 5.29; 95% CI, 3.5–8.0; p < 0.0005, and RR, 3.45; 95% CI, 2.1–5.7; p < 0.0005, respectively) (Table 4). To evaluate the impact of PGD on long-term survival in the context of other risk factors, we entered PGD, single-lung transplantation, and BOS (as a time-dependent variable) into a multivariable Cox proportional hazards model. BOS and single-lung transplantation remained significant risk factors for death in the multivariable model (RR, 4.82; 95% CI, 3.2–7.3; p < 0.0005, and RR, 2.37; 95% CI, 1.4–4.0; p < 0.001, respectively) (Table 5), but PGD was no longer a significant risk factor (Table 5). Although single-lung transplantation was a significant risk factor for death, it should be noted that patients were not randomized to receive a single versus a bilateral transplant, and a selection bias may have contributed to the increased mortality among single-lung recipients.
We confirmed the proportional hazards assumption for all time-independent variables in the Cox proportional hazards models, using the log-minus-log method. We performed regression diagnostics for both multivariable models. After excluding four influential cases in the risk factors for BOS model, the relative risks, 95% confidence intervals, and p values for acute rejection grade A2 or higher, lymphocytic bronchitis grade B2 or higher, and all grades of PGD remained unchanged, but CARV infection became a statistically significant risk factor (RR, 2.40; 95% CI, 1.4–4.1; p = 0.001). Excluding three influential cases in the risk factors for death model did not change any of the covariate relative risks or p values. Finally, we excluded nonlinearity for all variables in the Cox proportional hazards models, using partial residual plots.
In this study, we evaluated the impact of PGD after lung transplantation on BOS. Our results indicate that recipients who develop PGD have an increased risk of BOS independent of acute rejection, lymphocytic bronchitis, and CARV infections. In addition, we detected a consistent direct relationship between the severity of PGD and the risk of BOS; recipients who developed PGD grade 3 and survived beyond 90 days after transplantation had the highest risk of BOS. This relationship further supports the hypothesis that the acute lung injury of PGD may predispose to BOS months or years later. Furthermore, we found no significant associations between PGD and acute rejection or lymphocytic bronchitis. Lastly, our data suggest that although PGD may affect long-term survival beyond the first 90 days after transplantation, this effect does not appear to be independent of BOS. In the context of previous studies evaluating the impact of PGD on BOS, our findings corroborate those of Fiser and colleagues, who found that early graft dysfunction, defined by radiographic and gas exchange abnormalities in the first 24 hours after transplantation, was an independent predictor of the development of BOS and its progression (21). In contrast, Fisher and colleagues investigated this potential relationship, using a definition of early graft dysfunction based on the histologic finding of diffuse alveolar damage approximately 7 days after transplantation, and found no association with BOS (22). In addition, although recipients who had histologic evidence of diffuse alveolar damage had a high 30-day mortality rate, there was no difference in long-term survival between those who had diffuse alveolar damage and those who did not, when deaths in the first 30 days were excluded (22). The most obvious explanation for these conflicting results is the different definitions of PGD. It is possible that some of the recipients, described by Fisher and colleagues, who did not have diffuse alveolar damage 7 days after transplantation may have had PGD according to the ISHLT definition (8). Indeed, the median PaO2/FIO2 ratio 24 hours after transplantation for this group was 315 mm Hg with a range of 53 to 750 mm Hg (22). Hence, a subgroup of these recipients may have had PGD and would have had an increased risk of BOS (according to our findings), thus attenuating the apparent difference in the risk of BOS between those who had diffuse alveolar damage and those who did not. The ISHLT Working Group on PGD sought to alleviate problems with different classifications of PGD by standardizing the definition so that independent groups can study this complication and interpret the findings of others appropriately (8). Primary graft dysfunction represents the end result of multiple insults that begin with donor brain death, possible aspiration- and ventilator-associated lung injury, cold ischemia, and finally reperfusion at the time of implantation. Nonetheless, it is believed that ischemia–reperfusion injury is the major insult and that the other complications further intensify this injury (1, 3, 9). The potential mechanisms by which ischemia–reperfusion lung injury may lead to BOS have yet to be elucidated. However, this injury has been associated with chronic allograft dysfunction after kidney (23–25) and heart transplantation (26, 27). It is proposed that ischemia–reperfusion injury may amplify the immunogenicity of the allograft by increasing the expression of major histocompatibility complex molecules (28, 29) and recruiting recipient antigen-presenting cells into the allograft (30). Indeed, delayed graft function after kidney transplantation has been recognized as a risk factor for acute rejection in many (23, 24) but not all studies (31). However, we did not detect any relationship between PGD and acute rejection or lymphocytic bronchitis in this study. Likewise, Elidemir and colleagues investigated this relationship in a cohort of pediatric lung recipients and found no association between PGD and acute rejection (32). Although the relationship between acute rejection and BOS is well established, it is not clear that acute rejection causes BOS. Acute rejection may only be a marker for the donor-specific alloreactivity that results in BOS (14). Alternatively, acute rejection may be one of multiple injuries that induce a stereotypic graft response that results in BOS (33). Indeed, lung allografts are susceptible to multiple insults, including acute rejection, infection, ischemia–reperfusion injury, and silent aspiration of gastric contents, and all of these injuries have been linked to BOS. Recent evidence proposes a key role for autoimmunity in the development of BOS after lung transplantation (34–37). Collagen V, a sequestered but immunogenic self-protein, is exposed during lung injury and can stimulate the expansion of autoreactive collagen V–specific T cells, which may then propagate graft injury (34–37). Ischemia–reperfusion injury may also expose collagen V and trigger an autoimmune response by expanding autoreactive T cells, which can propagate the BOS lesion (34). On the other hand, ischemia–reperfusion injury may up-regulate cell surface adhesion molecules necessary for leukocyte infiltration into the graft (38, 39), and trigger the release of proinflammatory cytokines to create an inflammatory milieu (40–42) that results in cell death and a "danger signal" that activates the recipient's alloimmune response (43, 44). Finally, PGD may lead to BOS via nonimmune pathways if the recovery results in an exuberant but disordered repair process. This study has inherent limitations related to data availability and the retrospective design. However, we performed a detailed analysis of a large cohort of consecutive recipients over 6.5 years and included complete outcomes data with a mean follow-up of more than 3 years. In addition, our management protocols did not change substantially over this time period. Nonetheless, PGD may have been misclassified. We excluded the postoperative courses of three transplant procedures from this study because of hyperacute rejection. Furthermore, we graded PGD on the basis of chest X-ray and arterial blood gas results obtained immediately after transplantation on arrival to the intensive care unit to minimize the impact of other potential complications such as pneumonia and volume overload on the operational definition of PGD. We did not analyze the impact of PGD grades at later time points after transplantation on BOS because, in a retrospective study such as this, the immediate PGD grade is perhaps most accurate because other complications are likely to emerge in the ensuing 12 to 72 hours that may affect PGD grading. However, because we did not evaluate PGD grades at later time points, we cannot assess the impact of changes in PGD grade over time on the subsequent risk of BOS from our data. Furthermore, the lack of assessment of PGD grades at later time points could have biased our results if changes in PGD grade over time affect the risk of BOS. In addition, the small sample sizes of the PGD grade 2 and 3 groups could have overinflated the relative risk of BOS, but we made only one variable time dependent in the multivariable models specifically to avoid risk inflation. Finally, we have not validated our models using an independent cohort because of the limited sample size. Nonetheless, a bias toward a specific cohort would be necessary for any of these limitations to significantly alter our results, and none is readily apparent. In conclusion, our data demonstrate that PGD is a significant risk factor for BOS independent of other recognized risk factors and that there is a direct relationship between the severity of PGD and the risk of BOS. This association has important clinical implications and suggests that the immunosuppressive regimen should be optimized and that lung function should be monitored closely for patients who develop PGD. In addition, although previous clinical studies evaluating preventive and therapeutic interventions for PGD have generally been disappointing, future treatments that might mitigate the severity of PGD may reduce the subsequent risk of BOS. Clearly, additional studies investigating the molecular mechanisms of PGD and its relationship with BOS are necessary.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200608-1079OC on December 7, 2006 Conflict of Interest Statement: S.A.D. has no financial relationship with a commercial entity that has an interest in the subject of this manuscript. R.D.Y. has no financial relationship with a commercial entity that has an interest in the subject of this manuscript. B.F.M. has no financial relationship with a commercial entity that has an interest in the subject of this manuscript. M.M.C. has no financial relationship with a commercial entity that has an interest in the subject of this manuscript. M.J.W. has no financial relationship with a commercial entity that has an interest in the subject of this manuscript. A.A.A. has no financial relationship with a commercial entity that has an interest in the subject of this manuscript. G.A.P. has no financial relationship with a commercial entity that has an interest in the subject of this manuscript. E.P.T. has received a $270,000 research grant from Astellas Pharma, Inc., to conduct a randomized controlled trial comparing the efficacy of tacrolimus with cyclosporin A after lung transplantations. R.R.H. has received a $2,000 honorarium from Astellas Pharma, Inc.; in addition, R.R.H. has received a research grant of $270,000 from Astellas Pharma, Inc., between September 2002 and January 2006 to conduct a clinical trial comparing the efficacy of tacrolimus with cyclosporin A after lung transplantation.
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
