This article is more than 5 years old. Information may no longer be current.
Algorithm may improve reporting of mortality after allogeneic HSCT
Click Here to Manage Email Alerts
The review and adjudication of cause-specific deaths within 1 year after allogeneic hematopoietic stem cell transplantation has led to the development of definitions that will help transplant centers consistently and accurately report cause of death, according to study results.
“We need to make sure that patients who experienced similar clinical events after a transplant were consistently defined regardless of where they were treated,” Theresa Hahn, PhD, professor of oncology and member of the Blood and Marrow Transplantation program at Roswell Park Cancer Institute, said in a press release. “This is a difficult topic to consider, but we can’t make progress to improve transplant outcomes without these discussions.”
Researchers plan to conduct a genome-wide association study of cause-specific mortality after allogeneic HSCT to better match patients with unrelated donors. In preparation for this study, Hahn and colleagues sought to reduce endpoint misclassification and subsequent overestimation or underestimation of genetic effects on mortality through consensus panels that reviewed and adjudicated cause-specific deaths.
The panel calculated discordance rates between their review and causes of death reported by individual transplant centers, identified factors leading to inconsistency in these determinations and established definitions of cause-specific death after unrelated-donor allogeneic HSCT.
The panel evaluated data from two cohorts comprising 3,532 patients (cohort 1, n = 2,609; cohort 2, n = 923) with acute lymphoblastic leukemia, acute myeloid leukemia or myelodysplastic syndrome who underwent allogeneic HSCT at one of 151 U.S. transplant centers between 2000 and 2011. Of these patients, 1,484 died within 1 year of HSCT.
Patients in cohort 2, or the validation cohort, tended to be older (64% aged ≥ 40), had higher BMI and were more likely to have advanced disease compared with cohort 1, the training cohort.
The transplant centers classified deaths as disease-related or treatment-related, which the adjudication panel either confirmed or questioned.
For cohort 1, the consensus panel reached an overall concordance rate of 87.5%. The adjudicators confirmed 463 of 465 cases defined as disease-related by transplant centers, resulting in a concordance rate of 99.6%. Further, they confirmed 80% (521 of 651 cases) of deaths classified as treatment-related. They reclassified the remaining 130 deaths as disease-related.
The consensus panel reached a similar overall concordance rate for cohort 2 (89%). The adjudicators agreed with 98.8% of deaths categorized as disease-related (162 of 164 cases) and 81% of deaths categorized as treatment-related (165 of 204 cases).
Events leading to cause-specific death included disease, graft-versus-host disease (GVHD), infection, organ failure and other. When considering these five cause-specific death categories, the adjudicators reached a 73% agreement rate with the transplant center for cohort 1 and 70% agreement rate for cohort 2. In both cohorts, the adjudicators confirmed a significant majority of deaths the transplant centers attributed to GVHD (cohort 1, 79.5%; cohort 2, 90.5%) and disease (cohort 1, 99.6%; cohort 2, 98.8%) deaths.
Most of the discordance occurred among deaths transplant centers attributed to organ failure (concordance rates for cohort 1, 47.8%; cohort 2, 27.6%) or other (concordance rates for cohort 1, 33.6%; cohort 2, 31.6%).
Factors significantly influencing concordance rates included year of transplant and disease status. The adjudicators had a lower agreement rate with transplant centers when considering patients treated between 2000 and 2003 compared with patients treated between 2009 and 2011.
Further, the adjudicators relied more heavily on disease status — or whether a patient achieved remission — to determine cause-specific deaths than transplant centers.
In a sensitivity analysis of early deaths before day +100 HSCT, the researchers observed the greatest discordance between the adjudicators and transplant centers among patients with early disease risk myelodysplastic syndrome who died within +30 days of HSCT (18%). Agreement was highest for patients with early disease status AML or ALL who died before day +30 post HSCT (81%), as well as patients with early or intermediate disease ALL or AML who died before day +100 post HSCT (73% to 74%).
Based on these data, all patients with myelodysplastic syndrome who die within 100 days of HSCT and all patients whose cause of death is reported as “other” or “organ failure” should be prioritized for future studies needing adjudication of cause-specific death, according to the researchers.
Hahn and colleagues concluded that standard predefined criteria for adjudicating cause-specific death would result in consistent application to similar clinical scenarios and greater understanding of cause-specific death categories.
“It is important that we accurately define outcomes in these types of genomic studies as precisely as possible,” Lara Sucheston-Campbell, PhD, associate professor of oncology in the department of cancer prevention and control at Roswell Park Cancer Institute, said in a press release. “Our work is a critical first step toward the ultimate goal of finding a better match for patients receiving an unrelated-donor blood or marrow transplant." – by Cameron Kelsall
Disclosure: The researchers report no relevant financial disclosures.
Perspective
Back to Top
Edward Copelan, MD, FACP
Approximately 90% of articles appearing in subspecialty journals describe observational research. The validity and general applicability of the results of such studies wholly depend on the accuracy and completeness of the data. Precision in defining cause of death is fundamental to many studies of patients with cancer. Misattribution has contributed to errors in estimating temporal trends in deaths due to prostate cancer, in evaluation of screening for various cancers and in calculation of cancer death rates.Studies using observational databases have been particularly influential in the field of hematopoietic stem cell transplantation (HSCT). The fundamental differentiation between disease-related and transplant-related death helps determine the safety and effectiveness of new procedures, and in the context of this article, to investigate genetic factors that contribute to cause-specific mortality following unrelated HSCT. The report cites significant disagreement between institutional and central designation of cause of death: 20% of deaths reported by institutions as transplant related were wrongly reported — the authors are more tactful in their wording. Such imprecision in assignation of cause of death has been previously identified and reported by adjudication committees in clinical trials of HSCT. The persistent misguided designation by some centers of the last event in a cascade of events leading to death as the primary cause, despite the previous reports of this problem and the provision of the CDC and WHO definitions to transplant centers, is a sobering concern. This failure to follow established guidelines sabotages the credibility of all studies using similarly collected datasets. Individual centers and registries must take full responsibility for the validity of the data they generate and address these serious shortcomings.
Click Here to Manage Email Alerts