Associations between glycemic status and graft-versus-host disease, infection, and mortality among hematopoietic stem cell transplant recipients

Background. Patients undergoing allogeneic hematopoietic cell transplantation (HCT) for hematological malignancies are at risk for non-relapse mortality, primarily through the development of acute grades 2-4 graft-versus-host disease (aGvHD) and infections. A possible contributor to these unfavorable outcomes is blood glucose (BG) dysregulation, termed malglycemia. The purpose of this study was to evaluate associations between malglycemia, particularly hyperglycemia, and increased risk for aGvHD, infection, and mortality.;Methods. A retrospective cohort study was conducted using descriptive evaluation, Kaplan-Meier survival estimates, and Cox regression analyses to assess the associations between glycemic status and aGvHD, infection, and mortality throughout the first 99 days post HCT. Patient age, type and severity of underlying malignancy, and human leukocyte antigen (HLA) donor matching were described in relation to glycemic status and adjusted in the Cox regression model. Infection and aGvHD were also controlled for the outcome of mortality, and aGvHD was controlled for infection incidence. BG was evaluated as a mean in 5-day time blocks as a predictor for an event within the subsequent 5-days.;Results. Evaluation of 66,062 BG measurements from 1,175 HCT recipients from the Fred Hutchinson Cancer Research Center, between 2001 and 2005, revealed that 53.4% of all BG measurements were hyperglycemic (BG > 126 mg/dl). Within the first 99 days post transplant, 175 patients died, of which 59.9% had aGvHD and 86.6% had infections. With and without adjusting for all covariates, there was an almost 3-fold increased risk for mortality for every 100 mg/dl increase in BG within 5 days prior to the event (Hazard Ratio [HR] 2.84, p < .0001). BG was also associated with an increased risk for infection (HR 1.21; p = .021), except when adjusted for HLA-donor matching (HR 1.17; p = .085). BG was not associated with an increased risk for aGvHD (HR .91; p = .188).;Conclusions and implications. These analyses showed positive associations between malglycemia, particularly hyperglycemia, and mortality and infection. Further studies are needed for better glycemic monitoring and management. The generalizability of these findings to other acutely ill patients with cancer remains to be determined.