LAUSR

Haemostatic complications are commonly observed in patients undergoing blood or marrow transplantation (BMT), and are related to thrombocytopenia and immunological complications. In allogeneic transplant recipients, the cumulative incidence of bleeding was higher than venous thromboembolism (VTE) (30% vs. 11 8%) at 14 years after transplant. Development of bleeding was associated with inferior survival but not thrombosis, though the median follow-up in this study was only 20 months (Labrador et al, 2013). A comprehensive analysis of sociodemographics, therapeutic exposures and comorbidities, and their impact on the risk of mortality in long-term BMT survivors with a history of VTE has not been performed. We hypothesized that a longer follow-up is needed to study this effect. We addressed this gap in knowledge using the resources offered by the Blood or Marrow Transplant Survivor Study (BMTSS). Eligible participants included patients who received BMT at City of Hope or the University of Minnesota between 1 January 1974 and 31 December 1998, survived ≥2 years after transplantation, were alive and aged ≥18 years at study participation. BMT survivors were approached to complete a BMTSS survey between 2000 and 2004. The BMTSS survey covered sociodemographics, access to and use of medical care, diagnosis of physical health conditions by a healthcare provider with age at diagnosis and medication use. Institutional transplant databases were used to obtain information regarding primary diagnosis, transplant preparative regimens, stem cell source and graft type. The BMTSS questionnaire identified patients with a history of VTE diagnosed by their health care providers. Patients were categorized into those with and without a past history of VTE, irrespective of the time of onset with respect to BMT. National Death Index (https://www.cdc.gov/nchs/ndi/in dex.htm, accessed 31 December 2015) and/or medical records provided information regarding the date and cause of death until 31 December 2015. Additional information from Accurint databases (www.accurint.com, accessed 31 December 2016) was used to extend the vital status information up to 31 December 2016. This enabled a near-complete ascertainment of follow-up of this cohort and enhanced accuracy of mortality status. All patients were assigned a primary and, if present, a secondary cause of death. In the analyses of relapse-related mortality (RRM) and non-relapse-related mortality (NRM), all patients with primary disease as a primary or secondary cause of death were assigned relapse-related cause of death. A total of 1,022 BMT survivors completed the survey. Demographic and clinical characteristics of these patients are summarized in Table SI. Median [interquartile range (IQR)] age at BMT was 34 9 (25 1–45 0) years, and median (IQR) length of follow-up after completion of the survey was 15 1 (12 6–16 0) years. Seventy-one patients (6 9%) had a history of VTE in this cohort. Subsequent to completion of the survey, 309 (30 3%) deaths were observed. The overall survival (OS) among patients with and without VTE was 77 5% vs. 91 1% at 5 years and 50 4% vs. 72 3% at 15 years after survey completion, P < 0 0001 (Fig 1A; Table SII). Cox regression analysis, adjusting for age, sex, race/ethnicity, income, insurance, education, co-morbidities, use of hormonal therapy, primary diagnosis, conditioning regimen, stem cell source, transplant type, history of graft-versus-host disease, and relapse of primary cancer or development of subsequent neoplasm, showed that patients with a history of VTE were at a 59% higher risk of subsequent all-cause mortality compared to those without [hazard ratio (HR) = 1 59, 95% confidence interval (CI): 1 09–2 31, P = 0 015] (Table I). Cumulative incidence of RRM was similar among those with and without VTE (5 6% vs. 5 4% at 15 years, P = 0 9) (Fig 1B, Table SII). On the other hand, the cumulative incidence of NRM was higher among those with VTE (35 4% vs. 17 9% at 15 years, P = 0 0003) (Fig 1C, Table SII). Proportional sub-distribution hazards model (Fine-Gray) for competing risks was used for determining the association between VTE and subsequent RRM and NRM. The adjusted hazard of death from RRM was not significantly different between those with and without VTE (HR = 0 69, 95% CI, 0 2–1 8, P = 0 54). In contrast, patients with VTE were more likely to die from NRM (HR = 1 68, 95% CI: 1 1–2 5, P = 0 018). In analysis restricted to patients with VTE and adjusted for socio-demographics, a history of diabetes was associated with a significantly higher hazard of all-cause late mortality (HR = 3 28, 95% CI: 1 31– 8 21, P = 0 011). All analyses were performed with SAS software version 9.4 (SAS Institute Inc., Cary, North Carolina, USA). In the general population, individuals with VTE experience inferior survival compared to age-, sexand racematched individuals without VTE; the mortality risk remains elevated up to 30 years (Sogaard et al, 2014). Older age, male sex, lower body mass index, confinement to a hospital/nursing home, congestive heart failure, chronic lung disease, neurological disease and active cancer are independent predictors Correspondence