Incidence, risk factors, and outcome of bloodstream infections during the pre-engraftment phase in 521 allogeneic hematopoietic stem cell transplantations

BackgroundBloodstream infection (BSI) after allogeneic hematopoietic stem cell transplantation (HSCT) is a well-known complication during the pre-engraftment phase. Knowledge of trends in etiology and antibiotic susceptibility of BSI is important as the time to effective antibiotic treatment is closely associated with survival in bacteremic patients with septic shock. MethodsBSI during the pre-engraftment phase was studied retrospectively in 521 patients undergoing HSCT at our center in 2001-2008. Incidence, risk factors, outcome, and microbiology findings were investigated and compared with BSI in a cohort transplanted during 1975-1996. ResultsThe incidence of at least 1 episode of BSI was 21%, the total attributable mortality of BSI was 3.3%, and crude mortality at day 120 after transplantation was 21%. The rate of gram-positive and gram-negative BSI was 80% and 13%, respectively. Gram-negative BSI was more frequent both in 2001-2004 and in 2005-2008 compared with 1986-1996 (P=0.023 for 2001-2004, P=0.001 for 2005-2008), with fluoroquinolone-resistant Escherichia coli as the predominant finding. BSI with viridans streptococci and E.coli occurred significantly earlier after HSCT than BSI with Enterococcus species, with median time of 4, 8, and 11days, respectively (P<0.01 both for viridians streptococci vs. Enterococcus species, and E.coli vs. Enterococcus species). Risk factors for BSI in multivariate analysis were transplantation from unrelated donor and cord blood as stem cell source, whereas peripheral blood as stem cell source was protective. ConclusionsDespite low attributable mortality of BSI, crude mortality at day 120 after transplantation was 21%, indicating an association between BSI and other risk factors for death. The risk of gram-negative BSI increased over time in parallel with an increased rate of quinolone resistance. However, the incidence and attributable mortality of gram-negative BSI remained low. Thus, prophylaxis with ciprofloxacin is still deemed appropriate, but continued assessments of the risk and benefits of fluoroquinolone prophylaxis must be performed.