Secondary Treatment of Acute Graft-Versus-Host Disease: a Critical Review
Total Page：16
File Type：pdf, Size：1020Kb
REVIEW Secondary Treatment of Acute Graft-versus-Host Disease: A Critical Review Paul J. Martin,1,2 Yoshihiro Inamoto,1,2 Mary E. D. Flowers,1,2 Paul A. Carpenter1,3 Management of steroid-resistant or steroid-refractory acute graft-versus-host disease (aGVHD) poses one of the most vexing and difficult problems faced by transplantation physicians. In the current study, we used 10 criteria to evaluate 67 reports describing secondary treatment of patients with aGVHD. The goal of this ex- ercise was not only to provide a critical summary of the literature but also to offer suggestions that could improve future studies. Areas especially in need of improvement include the use of a consistent treatment regimen, the assessment of response at a consistent prespecified time point, consideration of concomitant treatment in assessing response, documentation that selection bias was minimized, and the use of methods that test a formal statistical hypothesis based on a contemporaneous or historical benchmark. Our results suggest that previous published reports collectively offer little guidance in discerning the most effective treat- ments for patients with steroid-resistant or steroid-refractory aGVHD. Adherence to the proposed criteria in future reports would enable meaningful comparisons across studies and thereby accelerate progress in evaluating new treatments for patients with aGVHD. Biol Blood Marrow Transplant 18: 982-988 (2012) Ó 2012 American Society for Blood and Marrow Transplantation KEY WORDS: Acute graft-versus-host disease, Treatment, Phase II clinical trials, Review INTRODUCTION cGVHD [4]. Less than 10% of the reports documented an effort to minimize patient selection bias, used a con- Allogeneic hematopoietic cell transplantation is sistent treatment regimen, or tested a formal statistical frequently complicated by acute and chronic graft- hypothesis that was based on a contemporaneous or versus-host disease (aGVHD and cGVHD) [1,2]. historical benchmark. Only 22% of the reports had Management of steroid-resistant or steroid-refractory a prespecified time for assessment of response, and aGVHD poses one of the most vexing and difficult only 35% to 45% of the reports provided adequate in- problems faced by transplantation physicians. Many formation regarding eligibility criteria, concomitant retrospective reviews and prospective studies have treatment, and overall survival. Based on these obser- been carried out to evaluate approaches for secondary vations, we concluded that previous studies offer little treatment of patients with aGVHD. To date, no con- guidance in discerning the most effective treatments sensus has been reached regarding the optimal choice for patients with steroid-resistant or steroid-refractory of agents for secondary treatment, and clinical manage- cGVHD. In the current study, we used the same list ment is generally approached through empirical trial of 10 criteria to evaluate reports describing secondary and error. Treatment choices are based on physician systemic treatment of patients with aGVHD. experience, ease of use, need for monitoring, risk of toxicity, and potential exacerbation of pre-existing comorbidity [3]. METHODS In a previous review, we used a list of 10 criteria to Literature Search evaluate reports describing secondary treatment of We searched the Medline (PubMed) database us- ing a broad search strategy to identify studies evaluat- 1 2 From the Division of Clinical Research; Department of Medicine; ing secondary treatment of aGVHD with reports and 3Pediatrics, Fred Hutchinson Cancer Research Center, University of Washington, Seattle, Washington. published between 1990 and 2011. The search was Financial disclosure: See Acknowledgments on page 986. conducted using the terms ‘‘Graft versus host disease’’ Correspondence and reprint requests: Paul J. Martin, MD, Fred and ‘‘Treatment’’ excluding ‘‘Review.’’ Relevant refer- Hutchinson Cancer Research Center, PO Box 19024, Seattle, ences in the publications identified were also reviewed. WA 98109-1024 (e-mail: [email protected]). Both retrospective and prospective studies were in- Received February 8, 2012; accepted April 10, 2012 Ó 2012 American Society for Blood and Marrow Transplantation cluded, but studies with cohorts containing fewer 1083-8791/$36.00 than 10 patients, case reports, and studies of agents doi:10.1016/j.bbmt.2012.04.006 that are not commercially available were excluded. 982 Biol Blood Marrow Transplant 18:982-988, 2012 Secondary Treatment of aGVHD 983 Review Criteria therapy for standard or high-risk disease and for Each report was evaluated according to the follow- secondary therapy than for subsequent therapies. ing 10 criteria. For third-line or subsequent therapy, such consis- tency is feasible only if prior treatment with agents 1. Adequately defined eligibility criteria. Inclusion other than glucocorticoids is discontinued. and exclusion criteria should specify affected sites, 4. Objective criteria for organ response. Categor- severity of manifestations, and prior treatment ical criteria should be defined for complete used to define the cohort. Exclusion criteria should response, partial response, stable disease, and indicate whether factors such as the presence of worsening for each organ or site affected by infection, inability to tolerate the study treatment, GVHD, because conclusions of the study are presence of persistent malignancy, or low perfor- based on response rates. Definitions require for- mance score were used to define the cohort. Studies mal assessment at baseline and at the comparison intended to evaluate treatment of ‘‘steroid-refrac- time point. Scales for staging organ severity and tory’’ GVHD should indicate the glucocorticoid overall GVHD grade can be used for this purpose. dose and duration of treatment used to define the 5. Unambiguous criteria for overall response. cohort. Eligibility criteria are typically more pre- The definition of overall response is distinct cisely defined for prospective studies than for retro- from the criteria for organ response. Overall re- spective studies. Data from retrospective studies sponses are often defined according to the overall describing all patients who received the study treat- pattern of organ responses. At a minimum, overall ment of interest are difficult to interpret unless partial response indicates improvement in at least additional selection criteria are applied to improve one organ. The category assigned for patients with homogeneity within the study cohort. improvement in one organ but deterioration in an- 2. Documented minimization of bias in the selec- other organ should be clearly stated. tion of patients. Readers should be given enough 6. Prespecified time for assessment of response. information to determine whether the characteris- To facilitate comparisons between studies, at least tics of the patients included in a study are represen- one specified time point should be used for assess- tative of the more general population of patients ment of response, and the data for this assessment with steroid-resistant or steroid-refractory should be shown. Additional information can be aGVHD. Risk factors that could affect outcomes shown as a time-to-initial partial or complete re- should be delineated. Ideally, either a historical sponse or as time-to-initial complete response, or contemporaneous cohort should be identified but this approach cannot account for subsequent for comparison, and any differences in the preva- loss of response. The number of patients who lence of risk factors between the study cohort and died, had recurrent malignancy, or had a treatment the comparison cohort should be noted. The use change before the assessment time point should of randomization to define cohorts helps to ensure be specified, and the results should clearly indicate the absence of bias, but this procedure does not en- whether these patients were excluded from consid- sure that the study cohort is representative of the eration in the assessment of response or whether more general population of patients with the indi- they were included as nonresponders. Tabulation cation of interest. Enrollment of all consecutive of results according to ‘‘best response’’ or ‘‘last patients who meet eligibility criteria could provide value carried forward’’ is not appropriate because a representative cohort but would raise ethical these categories do not reflect clinical benefit at questions about informed consent. a specific time point. 3. Consistent treatment regimen. The study treat- 7. Concomitant treatment taken into account. ment of interest should be administered in a consis- New systemic treatment for patients with GVHD tent manner in dose, schedule, and duration of added after enrollment but before the assessment administration. Differences in dose, schedule, or time point because of inadequately controlled dis- duration of administration can be addressed by ease manifestations should be categorized as non- stratified analysis of each specific subgroup. As response. Even in studies that use ‘‘best response’’ much as possible, concomitant treatment with as the endpoint, the text should state whether re- immunosuppressive agents other than glucocorti- sponse was evaluated before any new systemic coids should also be administered in a consistent treatment was added. Changes in glucocorticoid manner in order to facilitate the interpretation of dose should be described, but a temporary small in- results. Such consistency greatly improves the abil- crease in glucocorticoid dose during a