Risk Factors for Infections in Recipients of Hematopoietic Cell Transplantation in Relation to Donor Source

Infection is a major source of morbidity and mortality in patients undergoing allogeneic hematopoietic cell transplantation (HCT).  Umbilical cord blood (UCB) recipients appear to be at increased risk, and strategies to better identify and mitigate infection are needed. To investigate patterns and risk factors for infections by graft source (UCB vs. bone marrow (BM) or peripheral blood stem cells [PBSC]), we conducted a single center retrospective study of 308 consecutive adult allogeneic HCT recipients transplanted between 2006 - 2011 (BM 156, PBSC 103, UCB 49 patients).  All documented and presumed infections during the first year post-HCT were reviewed by a single reviewer.  Among BM/PBSC recipients, 53% received HCT using an unrelated donor.  Median age for UCB and BM/PBSC groups was 50 and 51 years.  Diagnosis and disease status, HCT-CI scores, performance status, and conditioning regimen intensity were comparable among the two groups.  Median time to neutrophil recovery (>500/µL) was 27 days in UCB and 13 days in BM/PBSC recipients.  The median length of transplant hospitalization was 46 and 30 days, respectively.  Bacterial infections were more common for CB recipients (94% vs. 63%, P < 0.001), without significant differences in viral, fungal, or C. difficile infections.  The day of first infection was notably earlier in the UCB recipients (median 6 vs. 30 days, P < 0.001), and >1 infection (bacterial, viral or fungal) occurred in 65% of UCB vs. 34% of BM/PBSC recipients.  Incidences of acute and chronic GVHD were similar between groups, however CB recipients were less likely to have GVHD prior to diagnosis of bacterial (11% vs. 35%, P < 0.001) or viral infection (36% vs. 59%, P = 0.03).  On multivariate analysis, UCB donor source remained a significant risk factor prognostic for all infection types (Table).  Bacterial infection was associated with increased mortality (HR 2.92, P < 0.001), as was fungal infection (HR 2.58, P < 0.001), but not viral infection/reactivation.  In conclusion, UCB recipients have increased risk of infections compared to BM/PBSC.  Infections occurred earlier and before engraftment in UCB recipients, suggesting the role of factors other than prolonged neutropenia in their pathogenesis.  The association of donor source and post-HCT infections is likely multifactorial.

Table: Donor source and infection risk (multivariate analysis)

	Hazard ratio for UCB vs. BM/PBSC (95% CI)	P-Value
Any infection	3.2 (2.3-4.4)	<0.001
Bacterial	3.9 (2.8-5.5)	<0.001
Viral	1.8 (1.1-2.7)	0.01
Fungal	2.7 (1.1-6.5)	0.03