|Home | About | Journals | Submit | Contact Us | Français|
Rationale: Primary graft dysfunction (PGD) is the main cause of early morbidity and mortality after lung transplantation. Previous studies have yielded conflicting results for PGD risk factors.
Objectives: We sought to identify donor, recipient, and perioperative risk factors for PGD.
Methods: We performed a 10-center prospective cohort study enrolled between March 2002 and December 2010 (the Lung Transplant Outcomes Group). The primary outcome was International Society for Heart and Lung Transplantation grade 3 PGD at 48 or 72 hours post-transplant. The association of potential risk factors with PGD was analyzed using multivariable conditional logistic regression.
Measurements and Main Results: A total of 1,255 patients from 10 centers were enrolled; 211 subjects (16.8%) developed grade 3 PGD. In multivariable models, independent risk factors for PGD were any history of donor smoking (odds ratio [OR], 1.8; 95% confidence interval [CI], 1.2–2.6; P = 0.002); FiO2 during allograft reperfusion (OR, 1.1 per 10% increase in FiO2; 95% CI, 1.0–1.2; P = 0.01); single lung transplant (OR, 2; 95% CI, 1.2–3.3; P = 0.008); use of cardiopulmonary bypass (OR, 3.4; 95% CI, 2.2–5.3; P < 0.001); overweight (OR, 1.8; 95% CI, 1.2–2.7; P = 0.01) and obese (OR, 2.3; 95% CI, 1.3–3.9; P = 0.004) recipient body mass index; preoperative sarcoidosis (OR, 2.5; 95% CI, 1.1–5.6; P = 0.03) or pulmonary arterial hypertension (OR, 3.5; 95% CI, 1.6–7.7; P = 0.002); and mean pulmonary artery pressure (OR, 1.3 per 10 mm Hg increase; 95% CI, 1.1–1.5; P < 0.001). PGD was significantly associated with 90-day (relative risk, 4.8; absolute risk increase, 18%; P < 0.001) and 1-year (relative risk, 3; absolute risk increase, 23%; P < 0.001) mortality.
Conclusions: We identified grade 3 PGD risk factors, several of which are potentially modifiable and should be prioritized for future research aimed at preventative strategies.
Clinical trial registered with www.clinicaltrials.gov (NCT 00552357).
Primary graft dysfunction (PGD) is a form of acute lung injury occurring after lung transplantation and is the major cause of early post–lung transplant morbidity and mortality. Previous studies of PGD clinical risk factors have produced conflicting results, possibly because of small sample sizes, inconsistencies in PGD phenotype, and inability to control for multiple confounding variables.
We performed a multicenter, prospective cohort study of 1,255 lung transplant recipients across 10 US transplant centers. We identified receipt of an organ from a donor with any smoking history, elevated FiO2 during allograft reperfusion, preoperative sarcoidosis or pulmonary arterial hypertension, use of cardiopulmonary bypass, single lung transplant, large-volume blood product transfusion, elevated pulmonary arterial pressures, and overweight or obese recipient body habitus as risk factors for grade 3 PGD. Several of these risk factors are potentially modifiable, and thus may suggest preventative strategies, whereas other risk factors should be prioritized for future mechanistic research efforts.
Primary graft dysfunction (PGD) is a form of acute lung injury (ALI) that occurs within the first few days after allograft reperfusion in lung transplant recipients. The incidence of PGD is 10–30% and is the major cause of mortality within the first post-transplant year (1, 2). PGD leads to increased duration of mechanical ventilation and intensive care unit length of stay, poor functional outcomes, and increased risk of bronchiolitis obliterans syndrome (3, 4). Investigations that specifically evaluate PGD risk factors have the potential to profoundly affect future outcomes in patients undergoing lung transplantation.
Previous studies of PGD risk factors have produced conflicting results. Some explanations for these variances include small sample sizes; inconsistencies in PGD phenotype; inability to control for multiple confounding variables; and frequent use of retrospective, single center, or administrative data sets that lack rigorous PGD definitions (5, 6).
In 2005, the International Society for Heart and Lung Transplantation (ISHLT) standardized the PGD definition to facilitate research on risk factors associated with the development of this syndrome (7). Subsequent studies have demonstrated the construct validity of this definition with clinical outcomes and biologic markers of ALI severity (8, 9). In this study, we aimed to identify donor, recipient, and perioperative risk factors for PGD using the ISHLT definition in a large, multicenter, prospective cohort study design.
The Lung Transplant Outcomes Group (LTOG) is a US National Institutes of Health sponsored, multicenter, prospective cohort study designed to evaluate risk factors for PGD. Details of subgroups in the LTOG cohort have previously been described (10–13). We included patients aged 18–80 years undergoing single or bilateral lung transplantation at 10 US transplant centers between March 2002 and December 2010 (see Table E1 in the online supplement). Clinical parameters were collected prospectively. Additional information was verified from the US United Network for Organ Sharing. The institutional review boards at each center approved this study.
PGD was graded according to ISHLT criteria, which is based on PaO2/FiO2 ratio and the presence of diffuse parenchymal infiltrates in the allograft on chest radiograph. Chest radiographs were interpreted independently by two physicians masked to the clinical variables, with adjudication of conflicts by a third reviewer (PGD grade classification kappa = 0.95) (7). The primary outcome was the presence of grade 3 PGD (PaO2/FiO2 ratio < 200) at 48 or 72 hours after transplantation, previously demonstrated to have construct validity for long-term outcomes and concurrent lung injury markers (3, 8). We performed a sensitivity analysis using grade 3 PGD occurring at any point within 72 hours of transplantation as a secondary outcome (3).
Potential risk factors for grade 3 PGD previously identified in the literature or with hypothetical clinical or biologic plausibility were selected for analysis a priori (5–7, 14–20). Details of covariate definitions are included in the online supplement.
Candidate risk factors were cross-classified for evidence of collinearity and zero cell counts. Recipient body mass index (BMI) was included as a categorical variable in multivariable modeling because of its observed nonlinearity. Transplant center was evaluated as a fixed effect using conditional logistic regression. A limited number of hypothesis-driven interaction terms were evaluated using multiplicative conditional logistic regression. A parsimonious final model was developed by eliminating factors that were not confounders based on a less than 20% change in odds ratio (OR). Ischemic time was forced into the final multivariable model. A preoperative diagnosis of pulmonary arterial hypertension was evaluated in a multivariable model without mean pulmonary artery pressure (mPAP) and bypass use given the strong collinearity with these variables. A secondary analysis evaluating risks within bilateral lung transplant (BLT) and single lung transplant (SLT) recipient groups individually was also performed. We approached the problem of missing data using multiple imputation. Analyses proceeded by use of 10 imputed datasets, and confidence intervals (CIs) for point estimates of the ORs were determined using the “mim” command in STATA 11.2 software (STATA Corp., College Station, TX). Postestimation marginalized standardized risks for grade 3 PGD were calculated based on the final logistic regression model for selected categorical variables. Individual data elements had varying degrees of missing data, ranging from 0–46% (see Table E2). STATA 11.2 was used for all analyses; GraphPad Prism 5 (GraphPad Software, La Jolla, CA) was used for generating graphs.
There were 2,011 lung and heart-lung transplants performed at study centers during the study period. Of these, 1,255 patients were enrolled in the cohort study (Figure 1). There were no significant differences in sex or age, but there was more chronic obstructive pulmonary disease, less cystic fibrosis, and more SLT in the enrolled group (see Table E3). A total of 211 subjects (16.8%; 95% CI, 14.7–18.9) met criteria for grade 3 PGD, and 386 subjects (30.8%; 95% CI, 28.2–33.3) met the secondary PGD definition of grade 3 PGD at any time during the first 72 hours after transplantation.
Clinical characteristics are presented in Table 1. Of the 479 subjects receiving a lung from a donor with any previous smoking history, 101 (21%) developed grade 3 PGD, compared with 14% (110 of 776) receiving a lung from a lifelong nonsmoker. Nearly 62% (130 of 211) of patients with grade 3 PGD in the cohort received bypass during the transplant procedure, and 28% (130 of 467) of patients receiving bypass developed grade 3 PGD. Donor preoperative oxygenation, as determined by lowest PaO2 measured as part of an oxygen challenge before lung procurement (P = 0.1) or highest oxygen challenge PaO2 (P = 0.2), was not associated with grade 3 PGD (P = 0.1).
Conditional multivariable analyses are presented in Table 2. In the fully adjusted multivariable model, independent risk factors for grade 3 PGD included use of cardiopulmonary bypass, SLT, pulmonary hypertension, a preoperative diagnosis of sarcoidosis, higher BMI, large-volume PRBC transfusion, donor smoking history, and increased FiO2 during allograft reperfusion. Of the 1,255 transplant recipients, 479 subjects received an organ from a donor with a history of any prior cigarette use, whereas 198 received an organ from a donor with a history of more than 20 pack-years. Receipt of an organ from a donor with any prior cigarette use was significantly associated with grade 3 PGD (OR, 1.8; 95% CI, 1.2–2.6; P = 0.002), whereas receipt of a lung from a donor with a greater than 20 pack-year history had an attenuated association with grade 3 PGD (OR, 1.5; 95% CI, 1.0–2.4; P = 0.06). Because of difficulty in accurately collecting FiO2 at reperfusion, the reperfusion FiO2 was missing from 46% of all subjects. In a multivariable complete case analysis of 619 subjects with complete reperfusion FiO2 information, the association between reperfusion FiO2 and grade 3 PGD was significant, with a similar point estimate for the OR as in the fully imputed analysis (OR, 1.1; 95% CI, 1.0–1.3; P = 0.04).
Calculated standardized predicted risks of grade 3 PGD for significant individual risk factors are presented in Figure 2. The predicted risk of grade 3 PGD increased with increasing FiO2 during allograft reperfusion from 12% (95% CI, 7–16%) at FiO2 less than 0.4 to 18% (95% CI, 16–21%) at FiO2 greater than or equal to 0.4, an absolute risk increase (ARI) of 6%. Overweight recipients had an ARI of 7% for grade 3 PGD compared with normal weight, whereas obese recipients had an ARI of 11% for grade 3 PGD. Large-volume blood transfusion was associated with an ARI of 9%, donor smoking was associated with an ARI of 5%, and cardiopulmonary bypass was associated with an ARI of 15%.
As shown in Figure 3, there was significant variation in the incidence of grade 3 PGD across the 10 centers included in the cohort, ranging from 2–27%. There was no significant detected interaction of grade 3 PGD risk factors by center, and evaluation of individual significant risk factors within the four largest centers did not identify substantial variation in risk factor effect estimates across centers. Sensitivity analyses conducted using grade 3 PGD at any time point were consistent with analyses using the primary endpoint (see Table E4), although the association with mPAP (OR, 1.1; 95% CI, 1.0–1.3; P = 0.05) was attenuated. Additionally, total ischemic time (OR per hour, 1.2; 95% CI, 1.0–1.3; P = 0.005) and pretransplant diagnosis of idiopathic pulmonary fibrosis (OR, 1.5; 95% CI, 1.1–2.1; P = 0.02) demonstrated significant association with development of grade 3 PGD using this alternate outcome definition.
Because of concern that grade 3 PGD after an SLT may be misclassified secondary to the impact of the native lung on PaO2/FiO2 ratio, SLT and BLT were evaluated using separate analyses (see Table E5). Donor smoking history was significantly associated with grade 3 PGD in SLT recipients (OR, 2.0; 95% CI, 1.1–3.7; P = 0.03) and BLT recipients (OR, 1.7; 95% CI, 1.0–2.7; P = 0.03). Cardiopulmonary bypass use was also a significant risk factor among single (OR, 5.0; 95% CI, 2.2–11.6; P < 0.001) and bilateral (OR, 3.7; 95% CI, 2.1–6.3; P < 0.001) recipients. Increasing mPAP was only significantly associated with grade 3 PGD among BLT recipients (OR per 10 mm Hg increase, 1.3; 95% CI, 1.1–1.6; P = 0.001), although only four SLT recipients had a mPAP greater than 60 mm Hg.
The impact of grade 3 PGD on unadjusted 90-day and 1-year mortality is presented in Table 3. The primary definition of grade 3 PGD at 48 or 72 hours after transplant was associated with a relative risk (RR) of 4.8 (95% CI, 3.3–7.0; P < 0.001) for death within 90 days of transplant compared with those without grade 3 PGD and an ARI of 18% (95% CI, 12–24). Grade 3 PGD was associated with a significantly increased 1-year mortality (RR, 3.0; 95% CI, 2.3–3.9; P < 0.001) compared with those without grade 3 PGD, and an ARI of 23% (95% CI, 15–30). Although the magnitude of the association between grade 3 PGD and mortality was attenuated when the alternate definition of any grade 3 PGD within 72 hours was used in the sensitivity analyses, the association remained significant at 90 days (RR, 3.5; 95% CI, 2.3–5.1; P < 0.001) and 1 year (RR, 2.5; 95% CI, 1.9–3.3; P < 0.001) (see Table E6).
In the first prospective, multicenter cohort study of donor, recipient, and perioperative risk factors for grade 3 PGD after lung transplantation, we have identified receipt of an organ from a donor with any smoking history, elevated FiO2 during reperfusion, preoperative sarcoidosis, independent of pulmonary pressures or pulmonary arterial hypertension, use of cardiopulmonary bypass, SLT, large-volume blood product transfusion, elevated pulmonary arterial pressures, and overweight or obese recipient body habitus as risk factors for grade 3 PGD. Several of these risk factors are potentially modifiable (e.g., FiO2 at reperfusion, obesity) and thus may suggest preventative strategies, whereas other risk factors should be prioritized for mechanistic research efforts (e.g., donor smoking status and bypass use). The results of this study may lead to prospective studies evaluating alterations in perioperative recipient management, donor-recipient matching, and potentially recipient selection.
Donor cigarette use emerged as a significant risk factor for grade 3 PGD, consonant with prior findings of mortality (21). The United Network for Organ Sharing defined donor smoking history of more than 20 pack-years fails to include active smokers with less than 20 pack-years of tobacco exposure and was not statistically significantly associated with grade 3 PGD, possibly because of the small number of high pack-year donors identified in the cohort. Defining donor smoking as any tobacco use includes low total pack-year, active smokers, who may in fact represent a higher-risk donor pool. Our findings are consistent with previous smaller studies suggesting increased risk of grade 3 PGD, higher alveolar-arterial oxygen gradients, and longer intensive care unit length of stay in recipients of lungs from previous smokers (20, 22). The mechanisms of this association are unclear, but cigarette exposure may result in increased oxidative injury and nicotine exacerbates reperfusion injury in experimental models (23). Because smoking status has recently been shown to increase the risk of ALI in trauma patients, it is plausible that tobacco smoke exposure in the donor lung might exacerbate lung injury that occurs at the time of allograft reperfusion (24). However, given the limited pool of available lung donors, it is not currently feasible to exclude patients who were previous smokers as potential lung donors. A recent Lancet study demonstrated that, although recipient survival was worse after receipt of a lung from a smoking donor compared with a nonsmoking donor, overall survival was significantly better than if the recipient continued on the wait list (21). However, given that current methods of determining donor smoking history from interview of surrogates may be prone to measurement bias, we believe that more accurate quantification of smoking exposure in donors and research into mechanisms of donor smoking on increasing grade 3 PGD risk are important priorities for future investigation (24, 25).
Increased FiO2 during allograft reperfusion was strongly associated with development of grade 3 PGD, independent of transplant type, bypass use, and pretransplant diagnosis. Cold ischemia of the allograft followed by reperfusion results in a significant oxidative burst (26), which may overwhelm cellular antioxidant pathways and lead to cellular necrosis and apoptosis, production of proinflammatory cytokines, and worsening edema and gas exchange in animal models (27). Although we attempted to determine the FiO2 for each subject before allograft reperfusion, we appreciate that FiO2 is a dynamic variable, which may have been confounded by patient needs during the surgical procedure. Despite the prospective nature of the study and the inclusion of reperfusion FiO2 on the case report forms, we were only able to obtain accurate information on this variable for 54% of the study subjects. However, variability in FiO2 used at reperfusion by center suggests that there is variation in practice-related preference, and not simply a direct result of response to intraoperative changes in physiology. Two centers with the lowest PGD incidence also had the lowest mean reperfusion FiO2. Although high reperfusion FiO2 secondary to poor functioning of the allograft at the time of reperfusion is not a modifiable PGD risk factor, intraoperative practice patterns and preferences may be modifiable. Future investigations evaluating interventions aimed at decreasing reperfusion FiO2, while also evaluating immediate allograft function at reperfusion, are warranted.
Tidal volume per kilogram of ideal body weight at reperfusion was not associated with the development of PGD. We were unable to assess the relationship between postoperative ventilator strategies and PGD. Although high tidal volume ventilation has been shown to be a risk factor for ALI, many subjects developed PGD before a time when postoperative ventilatory management would be predicted to affect the risk of ALI (28–30). Ventilator management decisions are made concurrently with the development of PGD, making it difficult to determine whether ventilation strategy is a risk factor for PGD or a response to altered oxygenation. A large percentage of patients are extubated early after transplant resulting in missing data on ventilator management in the postoperative period.
Our study confirms elevated BMI as a potential risk factor for grade 3 PGD as previously reported in a subset of this cohort study (12, 17). In addition to obesity, overweight recipient BMI is also significantly associated with grade 3 PGD risk. Future efforts aimed at understanding the mechanistic link of adiposity and grade 3 PGD are warranted. Although we were also able to establish an association between the use of cardiopulmonary bypass and subsequent development of grade 3 PGD, it was not possible to accurately differentiate planned use of cardiopulmonary bypass from emergent initiation intraoperatively because of deterioration in patient hemodynamics or oxygenation. Differentiating emergent initiation of bypass from planned bypass should be an area of future investigation because it may lead to important alterations in practice patterns. Additionally, although all centers used controlled reperfusion at the cessation of bypass, the exact technique for reperfusion likely varies by center, and we were unable to fully capture these practice variations. Likewise, the relationship of large-volume blood transfusion with grade 3 PGD is difficult to separate from confounding because of unmeasured procedural characteristics leading to transfusion requirements. Nonetheless, because blood product transfusion in-and-of-itself is associated with ALI in at-risk groups this finding may warrant further research into mechanisms of increased grade 3 PGD risk (31).
In our multivariable analysis, elevated mPAP remained a significant risk factor, independent of diagnosis and use of cardiopulmonary bypass. Potential mechanisms for the elevated grade 3 PGD risk seen with secondary pulmonary hypertension include endothelial shear stress, or circulating humoral factors associated with pulmonary hypertension (17). Future research into these underlying mechanisms may lead to improved preventative strategies.
PGD incidence varied across the 10 centers included in this study (Figure 3). The risk factors identified were also significantly associated within the four largest centers. Some of the differences in PGD incidence across center are explained by risk factor distribution within centers. There was no standardization of recipient criteria, surgical techniques, or perioperative management across the centers in this observational cohort. Intraoperative use of cardiopulmonary bypass ranged from 9–71% and reperfusion FiO2 ranged from 25–90% across centers. Some centers use cardiopulmonary bypass for all BLT procedures. We believe that further evaluation of individual practice paradigms at different transplant centers should be an area of future evaluation.
Several characteristics previously reported as risk factors for grade 3 PGD were not identified as significant risk factors in our study. None of the previously identified donor variables, including sex, race, age, or mode of death were significantly associated with grade 3 PGD (14, 18, 20). Although we did not specifically evaluate a “marginal donor status” definition, our findings may indicate that standard donor variables do not increase grade 3 PGD risk, and that more sophisticated methods to evaluate subclinical lung injury in donors are warranted. Although donor PaO2 was not identified as a risk factor for PGD, low donor PaO2 often eliminates a potential organ from use for transplant, thus limiting the range of PaO2 available for analysis. Differences in our results compared with prior publications may be caused by the more severe phenotype of PGD used as the primary outcome and the prospective collection of covariates in our study.
Although PGD presents as a spectrum of disease severity, we chose a more severe phenotype based on prior research (8). This PGD definition was very strongly associated with increased risk of 90-day and 1-year mortality after transplant, demonstrating the significant impact that grade 3 PGD has on clinical outcomes in the first year after lung transplantation, and providing further validity for the ISHLT definition. Furthermore, sensitivity analyses using an alternate, less severe, PGD definition yielded similar results.
There are several limitations to this study. There is the potential that unmeasured confounding or bias secondary to missing data limited our results. In particular, we were unable to assess the effects of induction therapy because the practice was completely uniform within centers during the study time period. Likewise, although we used multiple imputation to account for missing data, some of the covariates had large percentages of missing data, especially reperfusion FiO2, which may have led to inflated variances caused by uncertainties of imputation. There is the potential for selection bias because not all transplant recipients from each site were enrolled in the cohort (see Table E1). However, most sites enrolled most of their patients, and although there were some differences in baseline variables between enrolled and nonenrolled patients, no identified risk factor was more prevalent in the nonenrolled population (see Table E3). Additionally, although we imposed strict criteria for PGD, there remains the potential for misclassification bias. We attempted to minimize this possibility, however, by independently reading radiographs and using a standard definition (32, 33). Although the ISHLT PGD criteria were first published online June 4, 2005, patients were enrolled prospectively in LTOG starting in 2002. One hundred twenty-one patients were enrolled before the publication of the PGD guidelines; PGD grades based on the ISHLT guidelines were retrospectively assigned to these patients. Exclusion of these subjects did not change the results. Given the long enrollment period for this study, there is potential for bias based on changes in clinical practice over time. Although patients were first enrolled in 2002 at a single site, 1,158 of the 1,255 patients (92%) were enrolled from June 2005 through December 2010, narrowing the enrollment period for most of the cohort. When evaluating these patients alone, there were no differences in the risk factors identified or their ORs. Additionally, the most recent ISHLT report includes 2004–2010 as a single era when presenting survival analyses (34). There were no differences in the results when transplant year was included as a potential confounder of the relationship between our identified risk factors and grade 3 PGD.
In summary, we identified risk factors for the development of grade 3 PGD after lung transplantation, and demonstrated the high attributable mortality of grade 3 PGD in the modern era of lung transplantation. These findings provide new knowledge to suggest mechanistic studies, including further evaluation of the relationship between donor smoking and PGD, and serve as the basis for evaluating interventions targeting potentially modifiable risk factors, such as body habitus and reperfusion FiO2. Our findings can be used to develop predictive models for PGD that may allow for risk factor modification, more objective donor-recipient matching algorithms, and lead to a more detailed understanding of the incremental risk associated with these factors.
Participating Centers and Investigators in the Lung Transplant Outcomes Group: University of Pennsylvania (coordinating site): Jason Christie, M.D., M.S. (PI), Steven M. Kawut, M.D., M.S., Alberto Pocchetino, M.D., Y. Joseph Woo, M.D., Ejigayehu Demissie, M.S.N., Robert M. Kotloff, M.D., Vivek N. Ahya, M.D., James Lee, M.D., M.S., Denis Hadjiliadis, M.D., M.H.S., Melanie Rushefski, B.S., Richard Aplenc, M.D., Clifford Deutschman, M.D., M.S., Benjamin Kohl, M.D., Edward Cantu, M.D., Joshua M. Diamond, M.D., M.S., Rupal J. Shah, M.D., and Laurel Kalman. Columbia University: David Lederer, M.D., M.S. (PI), Selim Arcasoy, M.D., Joshua Sonett, M.D., Jessie Wilt, M.D., Frank D'Ovidio, M.D., Lori Shah, M.D., Hilary Robbins, M.D., Matthew Bacchetta, M.D., Nilani Ravichandran, N.P., Genevieve Reilly, N.P., Jeffrey Okun, M.D., Debbie Rybak, B.A., Michael Koeckert, B.A., Robert Sorabella, B.A., Nisha Ann Philip, M.B.B.S., Nadine Al-Naamani, M.D., Matthew LaVelle, B.S., Megan Larkin, M.P.H., and Shefali Sanyal, B.S. Vanderbilt University: Lorraine Ware, M.D. (PI), Aaron Milstone, M.D. (PI), Jean Barnes, R.N., Stephanie Logan, R.N., Carla Ramsey, R.N., Thelma Walden, and Shaquita Claybrooks, R.N. Stanford University: Ann Weinacker, M.D. (PI), Susan Spencer Jacobs, M.S.N., Val Scott, M.S.N., and Tal Alfasi, M.S. University of Alabama, Birmingham: Keith Wille, M.D. (PI), and Necole Harris, R.N. Johns Hopkins University: Jonathan Orens, M.D. (PI), Ashish Shah, M.D., John McDyer, M.D., Christian Merlo, M.D., M.P.H., Matthew Pipeling, M.D., Reda Girgis, M.D., Karen Oakjones, R.N., and April Thurman. University of Michigan: Vibha Lama, M.D., M.S. (PI), Fernando Martinez, M.D., M.S., Emily Galopin, Douglas R. Armstrong R.N., M.S., and Mary Maliarik, B.S. Duke University: Scott M. Palmer, M.D., M.H.S. (PI), David Zaas, M.D., M.B.A., R. Duane Davis, M.D., Ashley Finlen-Copeland, M.S.W., Jessica Martissa, and William A. Davis. University of Chicago: Sangeeta Bhorade, M.D. (PI), and Mark Lockwood, R.N., M.S.N. University of Pittsburgh: Maria Crespo, M.D. (PI), Joseph Pilewski, M.D., Christian Bermudez, M.D., and Kathleen Hanze. Indiana University: David S. Wilkes, M.D., David Wilson Roe, M.D., Thomas Wozniak, M.D., Ronda L. McNamee, R.N., Kim A. Fox, R.N., Danyel F. Gooch, R.N., and Tonya Isaacs, R.N.
Supported by NIH grants R01 HL087115, R01 HL081619, and R01 HL096845. None of the authors have any financial relationship with a biotechnology or pharmaceutical manufacturer that has an interest in the subject matter or materials discussed in the submitted manuscript.
Author Contributions: Conception and design, J.M.D., J.C.L., S.M.K., R.J.S., S.M.P., L.B.W., A.R.L., S.L.B., and J.D.C. Acquisition of data, D.J.L., J.C.L., E.C., V.N.L., S.M.B., M.C., E.D., J.S., K.W., J.O., A.W., D.S.W., S.A., P.D.S., L.B.W., S.M.P., and J.D.C. Analysis and interpretation of data, J.M.D., S.M.K., S.M.P., A.R.L., S.L.B., L.B.W., and J.D.C. Drafting or revising the manuscript for important intellectual content, J.M.D., S.M.K., D.J.L., J.C.L., E.C., R.J.S., B.A.K., V.N.L., A.R.L., S.L.B., S.M.B., M.C., E.D., J.S., K.W., J.O., A.S.S., A.W., D.S.W., S.A., P.D.S., L.B.W., S.M.P., and J.D.C. Final approval of the version to be published, J.M.D., S.M.K., D.J.L., J.C.L., E.C., B.A.K., R.J.S., V.N.L., M.C., E.D., J.S., K.W., J.O., A.S.S., A.W., D.S.W., S.A., P.D.S., A.R.L., S.L.B., S.M.B., L.B.W., S.M.P., and J.D.C.
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.201210-1865OC on January 10, 2013