CLINICAL

AND

TRANSLATIONAL RESEARCH

Consensus Conference on a Composite Endpoint for Clinical Trials on Immunosuppressive Drugs in Lung Transplantation Caroline Monchaud,1,2 Benoıˆt Marin,3 Marc Estenne,4 Pierre-Marie Preux,3 Pierre Marquet,1,2,5 and the eDelphi-Lung Transplant Group(*) Background. In lung transplantation, diverse clinical events may impact patient outcome. In clinical trials comparing intervention strategies, single primary endpoints require large populations, or long study durations, whereas composite endpoints (CEPs) do not take into account the respective impact of their components on patient survival. The objective of this study was to propose consensus recommendations on endpoints for clinical trials on immunosuppressants in lung transplantation. Methods. The consensus process was managed through the Internet using the Delphi method. Forty experts were invited by the pilot group with the help of the International Society for Heart and Lung Transplantation and The Transplantation Society. In the first round, a questionnaire was made available to the experts to complete, and the responses were analyzed. In each next round, a new questionnaire was developed from the previous responses and sent to the panel members. Results. Consensus between 17 experts was achieved after five rounds. Two score-type CEPs were defined for immunosuppressive drug efficacy (7 items) and for toxicity (15 items). Death related to graft loss or immunosuppressive drug toxicity was attributed a maximum weight of 100. The weights of the items included in the efficacy and toxicity CEPs ranged between 10 and 80 and between 25 and 70, respectively. The CEP scores are calculated by adding the weights of all the items composing them, without exceeding 90 as long as the patient is alive. Conclusion. This consensus conference proposed two score-type CEPs including relevant endpoints. After validation, they should allow clinical trials with higher statistical power, improving the evaluation of the interventions tested. Keywords: Consensus conference, Delphi, immunosuppressants, Composite endpoints, Lung transplantation. (Transplantation 2014;98: 1331Y1338)

The authors declare no funding or conflicts of interest. 1 INSERM UMR-S850, CHU Limoges, Universite´ de Limoges, France. 2 Service de Pharmacologie, toxicologie et pharmacovigilance, CHU Limoges, France. 3 Unite´ Fonctionnelle de Recherche Clinique et de Biostatistiques, CHU Limoges, Limoges, France. 4 Service de Pneumologie, Hoˆpital Universitaire Erasme, Universite´ Libre de Bruxelles, Bruxelles, Belgium. 5 Address correspondence to: Pierre Marquet, M.D., Ph.D., INSERM UMR-S850, Faculty of Medicine, University of Limoges, 2 rue du Dr Marcland, 87025, Limoges Cedex, France. E-mail: [email protected] C.M. participated in research design, writing of the article, performance of the research, and data analysis. B.M. participated in research design, writing of the article, and data analysis. M.E. participated in research design and in the writing of the article. P.-M.P. participated in research design. P.M. participated in research design and in the writing of the article. Supplemental digital content (SDC) is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.transplantjournal.com). Received 23 December 2013. Revision requested 31 January 2014. Accepted 3 April 2014. Copyright * 2014 by Lippincott Williams & Wilkins ISSN: 0041-1337/14/9812-1331 DOI: 10.1097/TP.0000000000000235

Transplantation

& Volume 98, Number 12, December 27, 2014

he primary objective of most clinical trials on immunosuppressants in lung transplantation is to evaluate the benefit-to-risk ratio of a therapeutic strategy. To address this, composite endpoints (CEP) are increasingly used. The events impacting patient outcome are diverse and happen at various times after transplantation. When only mortality or transplantation-related morbidity is taken into account, clinical trials comparing intervention strategies require large populations or long trial durations. Combining surrogate endpoint in a CEP present several advantages: (i) increasing the numbers of events allows reducing the number of subjects required; (ii) events concerning other systems than the graft may be important in the natural history of the disease; (iii) enlarging patient health evaluation improves the assessment of the interventions tested. Besides, it can help in avoiding wrong interpretation of competitive risks, when a serious or precocious event reduces the incidence of another event. Unfortunately, most CEPs present drawbacks. They often make no distinction between efficacy and toxicity, combining events such as acute rejection episodes (AREs) and side effects. Second, there is often no weighing of the CEP components to

T

www.transplantjournal.com

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

1331

1332

www.transplantjournal.com

Transplantation

& Volume 98, Number 12, December 27, 2014

TABLE 1. Efficacy and toxicity endpoints discussed during the first three rounds of the consensus conference, and final decision Round 1, n=24 experts Round 2, n=23 experts Round 3, n=21 experts Decision Efficacy endpoints Death Death related to graft loss Treated ARE Biopsy-proven ARE Humoral rejection Anti-HLA antibody Donor specific anti-HLA antibody BOS Biopsy-proven BOS Graft loss (with successful retransplantation) Toxicity endpoints Death in relation with the IS toxicity Discontinuation of IS because of toxicity or side effects Significantly altered blood count Leucopenia Anemia CMV infection Pulmonary CMV disease Extrapulmonary CMV disease CMV asymptomatic viremia Primary EBV infection BK virus infection Influenza infection RSV infection Other viral infection Bacterial infection Burkholderia infection Pseudomonas aeruginosa infection Staphylococcus aureus pulmonary infection Staphylococcus aureus extrapulmonary infection Other bacterial pulmonary infection Other bacterial extrapulmonary infection Fungal Infection Pulmonary aspergillosis Pulmonary candidosis Other fungal pulmonary infection Extrapulmonary aspergillosis Extrapulmonary candidosis Other fungal extrapulmonary infection Parasitic infection Pneumocystis infection Toxoplasmosis Other parasitic infection Cancer Donor-derived cancer Recipient-derived solid organ cancer Skin cancer PTLD Renal dysfunction

A+

A+ Uj A?

No new rating A+ A? A++ A? Uj A+ A+ Uj A+

No new rating A+ No new rating A+ No new rating No new rating No new rating A? No new rating

accepted accepted rejected accepted

A+

A+ A+

No new rating No new rating

accepted accepted

Uj A? Uj A? A+ A? Uj Uj Uj Uj Uj Uj A? Uj Uj Uj U? Uj U? A+ A+ Uj A? A+ Uj Uj Uj A? Ij I? A+ Uj A+ A? A+ No new rating

No new rating A? U? A? No new rating Uj Uj Uj I? Uj Uj Uj No new rating Uj Uj U? Uj Uj Uj No new rating No new rating Uj A+ No new rating Uj Uj Uj A? I? I? No new rating Uj No new rating A? No new rating

A? A+ Uj

Uj

A?

Uj Uj

A+

A+

A+ A++

accepted accepted accepted accepted

rejected rejected rejected accepted rejected rejected rejected rejected rejected rejected rejected rejected rejected rejected rejected rejected rejected accepted rejected accepted accepted rejected rejected rejected rejected rejected rejected rejected accepted rejected accepted

(Continued on next page)

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Monchaud et al.

* 2014 Lippincott Williams & Wilkins

1333

TABLE 1. (Continued) Round 1, n=24 experts Round 2, n=23 experts Round 3, n=21 experts Decision ARI, without grading ARI, with grading (AKIN classification) CKD, without grading CKD, with grading (KDIGO classification) Chronic dialysis Renal transplantation Other adverse events, specific to IS Hemolytic uremic syndrome Neurologic toxicity Gastroparesis, loss of appetite, nausea Diarrhea New onset, insulin-requiring diabetes New onset or worsening of hyperlipidemia New onset or worsening of osteoporosis New onset or worsening of arterial hypertension

A+

Uj Uj A? A? A+ A+ A+ A? A+ A? A? A+ A? A? A?

No No No No

No

Uj Uj A? A? new rating new rating new rating A? new rating A? A? new rating A? A? U-

rejected rejected rejected rejected accepted accepted rejected accepted rejected Rejected Accepted Rejected Rejected Rejected

Explanation of ‘A,’ ‘U,’ ‘I,’ and ‘++,’ ‘+,’ ‘?,’ and ‘j’: see Table 2; PTLD, posttransplant lymphoproliferative disorder; ARI, acute renal injury; AKIN, acute kidney injury network; CKD, chronic kidney disease; KDIGO, kidney disease: improving global outcomes; ARE, acute rejection episode; EBV, Epstein-Barr virus; CMV, cytomegalovirus; BOS, bronchiolitis obliterans syndrome; IS, immunosuppressive drugs.

compensate for their relative severity. Finally, the variety of CEPs used (1, 2) render comparisons across trials difficult. The objective of this study conducted under the auspices of the International Society for Heart and Lung Transplantation (ISHLT) and The Transplantation Society (TTS) was to propose a CEP for clinical trials on immunosuppressants in lung transplantation, based on a consensus between international experts in this field.

RESULTS This consensus conference was initially conducted between March 2010 and February 2011. Of the 40 invited experts (adult and pediatric thoracic and cardiovascular surgeons and pulmonologists from Europe, North America and Australia), 24 answered the first round of questions, 23 participated in the first two, 21 in the first three, and 20 (83%) in four rounds of questions. Seventeen experts

TABLE 2.

participated in an extra round conducted between August and October 2012. Consensus on the CEP(s) Consensus on the number, type, and method of construction and analysis of the CEPs was reached at round 2 (see Table S1, SDC, http://links.lww.com/TP/B7). Two score-type CEPs were retained: one for efficacy and one for toxicity. It was considered that endpoints composing each CEP should be weighted, and that the calculation of the scores would be based on the addition of the weights of the endpoints. Finally, the experts considered that the comparison of several groups could be based on survival analysis of the CEPs or comparison of percentages. Consensus on the Items Included in the CEPs General items were proposed at round 1 then more precisely defined at rounds 2 and 3, based on the experts’

Decision rules (rounds 1 to 3)

Rating of appropriateness Appropriate (A) Uncertain (U) Inappropriate (I) Grading of agreement between experts Strong (++) Relative (+) Undertermined (?) Disagreement (j)

Panel median within the three-point regions 7Y9 Panel median within the three-point regions 4Y6 Panel median within the three-point regions 1Y3 All the ratings (100%) within the three-point region of the median location Some ratings (but G1/3) outside the three-point region of the median location Q1/3 of the ratings outside the three-point region of the median location (but G1/3 in the other extreme region with regards to the median location) Q1/3 of the ratings in the other extreme three-point region with regards to the median location

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

1334

www.transplantjournal.com

Transplantation

comments. Experts were proposed a series of efficacy and toxicity outcomes: for efficacy, the propositions related to death, graft loss, graft rejection, and the presence of antiYhuman leukocyte antigen antibodies; for toxicity, the propositions related to nonspecific and specific toxicity (Table 1). Consensual endpoints were selected at round 3 according to the decision rules (Table 2), and the last round allowed a sharper

TABLE 3.

& Volume 98, Number 12, December 27, 2014

definition of the toxicity outcomes, resulting in a consensus on 7 efficacy and 15 toxicity endpoints (Table 3). Consensus on the Weights Attributed to the Items Included in the CEPs The final weights attributed to the endpoints included in the CEPs are presented in Table 3.

Items included in the final CEPs and corresponding weights (round 5) Min

Efficacy CEP Death related to graft loss Treated acute rejection episode Biopsy-proven acute rejection episode (10) Acute rejection A1 Acute rejection A2 Acute rejection A3 Acute rejection A4 Airway inflammation B1R Airway inflammation B2R Airway inflammation X Donor-specific anti-HLA antibodies BOS (11) BOS0-p BOS1 BOS2 BOS3 Humoral rejection Graft loss Toxicity CEP Death in relation with the IS toxicity Discontinuation of IS caused by toxicity or side effects G7 days 97 days Pulmonary CMV disease With pneumonia Without pneumonia Pulmonary aspergillosis Noninvasive Invasive Extrapulmonary aspergillosis Other fungal pulmonary infections (other than aspergillosis and candidosis) Recipient-derived solid organ cancer PTLD Chronic dialysis/kidney transplantation Headache Peripheral neuropathy Seizures Tremor Posterior leukoencephalopathy New onset, insulin-requiring diabetes

Q1

20

30

0 20 20 30 10 15 0 15

10 20 38.5 50 10 25 6.25 25

10 20 30 50 25 80

16.25 30 50 60 40 80

Median 100 40

Q3

Max

40

50

10 30 40 50 10 30 10 30

20 30 50 57.5 18.75 40 10 37.5

70 80 90 90 80 90 20 50

20 40 50 65 45 80

25 48.75 58 70 50 90

60 60 70 80 80 90

100

20 30

35 40

40 50

40 50

80 80

20 0

31.25 20

40 30

47.5 40

70 60

0 25 20 8

26.25 31.25 40 28.75

30 45 48 40

40 65 50 62.5

70 80 80 80

20 25 40 0 0 0 0 0 0

50 50 60 11.25 20 30 16.25 35.5 22.25

55 60 70 25 30 36 25 40 30

60 60 70 33.75 40 40 30 50 35

80 80 80 50 70 80 60 80 50

PTLD, posttransplant lymphoproliferative disorder; CMV, cytomegalovirus; CEP, composite endpoints; BOS, bronchiolitis obliterans syndrome.

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

* 2014 Lippincott Williams & Wilkins

DISCUSSION During this consensus conference, two CEPs for clinical trials on immunosuppressants in lung transplantation were defined: one efficacy CEP and one toxicity CEP (Table 3). Strengths of the study To our knowledge, this is the first international consensus conference on the definition of endpoints for clinical trials in transplantation. The usual rules of the Delphi technique were applied, including a structured flow of information and controlled feedback to the participants (3, 4). Dominance phenomena were avoided by keeping the anonymity of the participants during the whole process. Although the accepted number of participants in such a procedure is usually between 10 and 15, here, 24 experts (60% of the 40 contacted experts) accepted to participate, reflecting a good adherence to the project. Seventeen (71%) completed the entire consensus conference, indicating the motivation of the panel members. Testing new immunosuppressive drugs raises the question of the benefits-risk balance. To our knowledge, none of the endpoints used so far in clinical trials have addressed these questions, often combining in a single CEP efficacy and toxicity criteria, of binary type (0/1 no matter the event). Among the advantages of the CEPs proposed in this study is that they clearly differentiate efficacy and toxicity, are based on weighted events, and are reported as numerical values, thus allowing evaluation of the benefit:risk ratio, contrary to most other endpoints. Contradictory Debates Among the Experts In clinical trials, CEPs can be used as primary endpoints, allowing a global comparison of immunosuppressive strategies, and their components can be used as secondary endpoints. Therefore, the objective of this work was to provide reliable CEPs, accepted by the scientific community, including relevant endpoints, that is, significantly impacting morbidity or mortality. Because the key factors predicting success after lung transplantation are still unknown, it is not surprising that some endpoints were the subject of debates among experts. Although graft loss was accepted as an efficacy endpoint, it may be the result of a variety of alloimmune-dependent (inappropriate immunosuppression) and alloimmune-independent processes (airway anastomotic complications, graft infection, pulmonary emboli). Similarly, biopsy-proven AREs were consensually included in the efficacy CEP despite the high interobserver variability in pathologic scoring (5), and their association with confounding factors such as viral infections. Many experts considered grade A1-2 ARE a minor complication, whereas some experts considered these A grades a serious problem impacting outcome, grade A2 ARE being uniformly viewed as an important indicator of immunosuppressive treatment failure. A3 and A4 ARE were in majority considered as a significant complication, associated with an increased risk of graft failure and death, whereas some experts argued that there is no proven relationship between A3 ARE and outcome. Airway inflammation (B grades) was often not considered a valid measure (poor sensitivity and specificity, high variability, unclear significance in outcome),

Monchaud et al.

1335

whereas some experts reported it as a risk factor for bronchiolitis obliterans syndrome (BOS). Bronchiolitis obliterans syndrome was considered a relevant efficacy endpoint, even though it could be induced by nonimmunologic factors, such as gastrointestinal reflux or viral infection. Although the majority considered BOS0-p displayed a low predictive value with regards to the patient outcome, some experts argued that it was the earliest sign of BOS, thus being a good predictor of more severe BOS stages. Most experts considered BOS1 displayed a high positive predictive value for death despite the relatively good preservation of lung function and quality of life. Finally, most experts considered that BOS2 and BOS3 displayed a high positive predictive value for death in the short term. Donor specific anti-HLA antibodies and humoral rejection were retained as efficacy endpoints at round 3. Even though donor specific anti-HLA antibodies represent the first step to chronic allograft dysfunction according to most experts, their predictive value for long-term outcome is still controversial. Similarly, humoral rejection was rejected by some experts because a standard definition is lacking, with a variable incidence between centers. The experts who retained this endpoint argued that it was potentially signing ineffective immunosuppressive therapy and that its impact on long-term morbidity and mortality in other solid organs transplantation was well documented. A consensual definition of humoral rejection in lung transplantation is needed to better estimate its predictive value on patient survival. Limitations of the Study Definition of the Endpoints The questionnaire was adjusted between rounds 1 and 2, taking into account the experts’ comments. For instance, ‘‘death,’’ initially proposed as a single endpoint, was split into ‘‘death-related to graft loss’’ (efficacy endpoint) and ‘‘death in relation with the immunosuppressive drug toxicity’’ (toxicity endpoint). The proposed endpoints remained identical until round 4, and consensus was initially considered reached after round 4. A supplementary round was requested by independent representatives of TTS and the ISHLT to whom the results had been submitted to address some remaining issues, such as the wide variety of conditions included in ‘‘neurologic toxicity’’ or the lack of precision of ‘‘pulmonary aspergillosis’’ and ‘‘pulmonary cytomegalovirus disease,’’ deserving to be more detailed, based on their potential impact on patient outcome. Some of the proposed endpoints will remain hard to interpret. For instance, the experts should have been consulted on the appropriateness to define AREs according to their response to the treatment (responsive, recurrent, persistent, and refractory). However, this would have necessitated too many extra rounds and was excluded by the pilot group, provided that these elements are included in future studies based on statistical modeling of cohort data. Also, BOS was proposed with reference to the definition proposed in 2002 by the ISHLT (6), although we acknowledge that this definition does not fully account for the complex clinical spectrum of chronic allograft dysfunction that has been increasingly recognized in recent years (7). Finally, after extensive discussion, the pilot group decided to exclude endpoints related to

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

1336

www.transplantjournal.com

treatment acceptability despite its potential impact on treatment compliance because there was no validated questionnaire to address this issue in lung transplantation. Method of Construction of the CEP A calculation of the scores based on an algebraic sum of endpoints’ weights was consensually admitted. However, this may not be optimal for the calculation of the efficacy score because: (i) the risk of occurrence of an ARE in a single patient probably changes with time, conditioned by the history of previous AREs; (ii) the relative impact of two AREs on patient survival may not be simply equal to twice the impact of a single ARE; (iii) the risk of chronic rejection depends on the history of AREs (6); the integration of the weight of BOS in the score should thus be conditioned by the history of AREs; and (iv) BOS is an evolving syndrome. Simply adding the weights of successive levels of BOS may thus not be appropriate, and it may have been more pertinent to include in the score only the weight of the highest BOS stage reached by the patient. This being said, the method of calculation of the scores was consensually admitted (Table S1, SDC, http://links.lww.com/TP/B7) and proposing several methods of calculation to take into account all the situations would have led to a complicated questionnaire, not adapted to the Delphi method. Duration of Postoperative Period Used to Compute CEPs In the first four rounds, the definition of the follow-up period for CEP computation was omitted. However, using score-type CEPs allows making valid comparisons between groups provided they are performed at a given time point after transplantation and evaluated on their ability to predict survival. Furthermore, defining the period of observation may influence the weights attributed to the endpoints included in the CEPs because their prevalence and clinical relevance may vary over time. Consequently, in the fifth round, the target period for CEP computation was defined as the first postoperative year and the prediction target as patient survival at 5 years after transplantation. Weighing of Endpoints One must keep in mind that the components of a CEP must be homogeneous in consequence on patient outcome: hence the interest of weighing, to balance the relative influence of the each endpoint on survival. In this consensus conference, most experts considered weighing appropriate; the pilot group decided to give the maximum weight of 100 to death, and to cap the score of living patients to 90. This constraint was imposed to reduce the impact that a single patient with multiple events, eventually conducting to death, may have on the global or mean score. Table 3 shows that the experts attributed a wide range of weights and that high weights were attributed to several endpoints, despite warnings made by the pilot group at rounds 4 and 5. Consequently, the proposed CEPs might have a reduced ability to discriminate among the immunosuppressive strategies evaluated. One example of this difficulty is given by ARE: because the experts decided that the CEP should be calculated by summing the weights of each endpoint, three or more

Transplantation

& Volume 98, Number 12, December 27, 2014

grade A2 or B2R AREs are sufficient to reach the maximal score of 90. A fifth round was organized to better define the weights retained at round 4. Nevertheless, and despite the warnings made by the pilot group, the results obtained at round 5 remained identical, indicating that convergence was reached. Validation of the CEPs The CEPs proposed by this consensus conference are mostly based on informed opinion, and their usability in clinical trials will only be demonstrated when the actual link of the scores with survival is shown. A retrospective ‘‘Benefits risks Evaluation after Lung Transplantation’’ study on 121 lung transplant patients followed up for 5 years after transplantation is currently being conducted to formally validate these CEPs. Its primary objective is to validate the CEPs proposed by this consensus conference on their ability to predict the 5-year survival of lung transplant recipients. Depending on the results of this study, the consensus conference might be reactivated to adjust the composition of the CEPs. In conclusion, this Delphi consensus conference proposed two relevant score-type CEPs for clinical trials on immunosuppressants in lung transplantation, for efficacy and for toxicity. The final CEPs proposed by this conference represent the subjective opinions of an expert panel, nonetheless composed by international experts in lung transplantation. The ISHLT registry (8) could be used by the experts as a tool to select and ponder the endpoints to provide evidence-based judgments. Once validated and despite their limitations, some of which can hardly be overcome, these CEPs should allow conducting more powerful clinical trials, largely improving the evaluation of efficacy, toxicity and benefit:risk ratio of interventions tested.

MATERIALS AND METHODS This consensus conference was conducted following the Delphi method, developed by RAND Corporation (3). In this process, iterative questionnaires of close-ended questions are designed by the study’s organizers (pilot group), and administered to a group of geographically scattered participants who never meet, to define a consensus (4, 9). The main characteristics of the Delphi method are: (i) anonymity of the participants (to avoid leader effect); (ii) administration of iterative questionnaires with controlled feedbacks; (iii) quantified analysis of the group responses. This consensus conference was conducted through a dedicated website, with the following steps (Fig. 1):

Definition of the Objective The objective, precisely defined by the pilot group (four members), was to define a CEP for clinical trials on immunosuppressants in lung transplantation.

ParticipantsVExpert Panel Selection The experts were identified by the pilot group with the help of the ISHLT and TTS on the basis of their recognized knowledge of lung transplantation and as representatives of a variety of end-users. A minimum number of 10 to 15 participants are usually considered necessary for the recommendations of a consensus conference to be valid, 40 experts were originally contacted.

Elaboration and Administration of the Questionnaires The questions were formulated to solicit close-ended answers. In rounds 1 to 3, the participants responded using a rating scale ranging from 1 (‘‘total disagreement’’) to 9 (‘‘total agreement’’). In rounds 3 to 5, they were asked

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Monchaud et al.

* 2014 Lippincott Williams & Wilkins

FIGURE 1.

1337

Schematic description of the Delphi consensus conference.

to attribute weights to the outcomes retained, on a scale ranging from 0 to 90, possibly using the data of the ISHLT registry (8), to base their judgment on evidence. The questionnaires consisted of six sets of questions, concerning the following:

& The type of CEP needed for clinical trials on immunosuppressants in lung transplantation (Table S1, SDC, http://links.lww.com/TP/B7).

& The potential merits of a single CEP including efficacy and toxicity & &

vs. 2 CEPs, distinguishing efficacy and toxicity (Table S1, SDC, http://links.lww.com/TP/B7). The relevant efficacy and toxicity criteria to include in the CEP(s). In case of a score-type CEP, the weights attributed to the items included in the CEP(s).

& T h e m e t h o d o f co n s t r u c t i o n o f t h e C E P ( Ta b l e S 1 , S D C , http://links.lww.com/TP/B7).

& The statistical analysis of the CEP (Table S1, SDC, http:// links.lww.com/TP/B7).

The same questionnaire served as the main thread throughout the consensus conference. It was enriched at each round by the results and comments generated at the preceding round.

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

1338

www.transplantjournal.com

Processing of the Results Initially, it was decided that up to four rounds would be organized to reach convergent responses. A fifth round was nonetheless requested by independent representatives of TTS and the ISHLT to address the issues remaining after the fourth round. The results of each round were analyzed by the pilot group. At round ‘‘n,’’ the experts were provided an individualized table presenting the results of round ‘‘n-1,’’ including their personal rating, the median opinion, and the dispersion of opinions in the expert panel, translated in appropriateness of the proposal and level of agreement between the experts (Table 2), to compare their own opinion at round ‘‘n-1’’ with regards to that of the panel. With these results, the experts were asked to express their opinion again at round ‘‘n,’’ possibly maintaining or modifying their response. If it differed strongly from the group median at round ‘‘n-1,’’ they were asked to justify their opinion at round ‘‘n.’’ The potential disagreements were interpreted using the comments collected from the experts. The participants were also invited to make free comments and propose events which they believed deserved to be included in a CEP. Only items classified as A++ at round 1 and A++ or A+ at rounds 2 and 3 were included in the final CEP. Starting from round 3, the experts were asked to attribute a weight to the items retained for the CEP. For each item, it was a priori decided that the median weight attributed in the last round would be considered as the consensual weight. To have concordant weights between the experts, a homogeneous frame for weight assignment was proposed, and one rule was set by the pilot group: (i) death was attributed the maximum weight of 100 points, being the ultimate failure of the procedure; (ii) CEPs were capped to 90 points, as long as the patient is alive, to discriminate living from dead patients, and avoid taking the risk that a living patient is attributed a score higher than 100.

ACKNOWLEDGMENTS The authors gratefully thank Carrie Andrews and Lori West for their invaluable help in the selection of the experts and coordination of this project, on behalf of The Transplantation Society and of The International Society for Heart and Lung Transplantation. The authors also thank deeply Fre´de´ric

Transplantation

& Volume 98, Number 12, December 27, 2014

Bousson for his excellent technical assistance and the development of the website dedicated to this consensus conference.

REFERENCES 1.

2. 3.

4. 5. 6.

7. 8.

9.

10. 11.

Hachem RR, Yusen RD, Chakinala MM, et al. A randomized controlled trial of tacrolimus versus cyclosporine after lung transplantation. J Heart Lung Transplant 2007; 26: 1012. Snell GI, Valentine VG, Vitulo P, et al. Everolimus versus azathioprine in maintenance lung transplant recipients: an international, randomized, double-blind clinical trial. Am J Transplant 2006; 6: 169. Linstone HA, Turoff M. The Delphi Method: Techniques and Applications. Ann Arbor, MI: Addison-Wesley Publishing Company Advanced Book Program; 1975. Mullen PM. Delphi: myths and reality. J Health Organ Manag 2003; 17: 37. Arcasoy SM, Berry G, Marboe CC, et al. Pathologic interpretation of transbronchial biopsy for acute rejection of lung allograft is highly variable. Am J Transplant 2011; 11: 320. Estenne M, Maurer JR, Boehler A, et al. Bronchiolitis obliterans syndrome 2001: an update of the diagnostic criteria. J Heart Lung Transplant 2002; 21: 297. Knoop C, Estenne M. Chronic allograft dysfunction. Clin Chest Med 2011; 32: 311. Christie JD, Edwards LB, Kucheryavaya AY, et al. The Registry of the International Society for Heart and Lung Transplantation: twentyseventh official adult lung and heart-lung transplant reportY2010. J Heart Lung Transplant 2010; 29: 1104. Bourree F, Michel P, Salmi LR. Consensus methods: review of original methods and their main alternatives used in public health. Rev Epidemiol Sante Publique 2008; 56: 415. Stewart S, Fishbein MC, Snell GI, et al. Revision of the 1996 working formulation for the standardization of nomenclature in the diagnosis of lung rejection. J Heart Lung Transplant 2007; 26: 1229. Cooper JD, Billingham M, Egan T, et al. A working formulation for the standardization of nomenclature and for clinical staging of chronic dysfunction in lung allografts. International Society for Heart and Lung Transplantation. J Heart Lung Transplant 1993; 12: 713.

Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Consensus conference on a composite endpoint for clinical trials on immunosuppressive drugs in lung transplantation.

In lung transplantation, diverse clinical events may impact patient outcome. In clinical trials comparing intervention strategies, single primary endp...
809KB Sizes 0 Downloads 5 Views