![[ Visit SHEA online ]](images/banner.gif)
Background: Selecting appropriate empiric antimicrobial therapy for allogeneic hematopoietic stem cell transplant recipients (allo-HSCT) is challenging due to high illness acuity and nosocomial pathogen risk. Clinicians must balance risks and benefits of broad empiric coverage.
Objective: The objective of this study was to determine the percentage of blood stream isolates in allo-HSCT recipients resistant to the commonly prescribed anti-pseudomonal beta-lactam/vancomycin regimen.
Methods: Allo-HSCT recipients transplanted between March 1, 2002 and June 30, 2009 were retrospectively reviewed for blood stream infection (BSI) occurring within 1 year of stem cell infusion. Patients with a single Bacillus, coagulase negative Staphylococcus [CONS], Propionobacterium, or Micrococcus culture were not considered to have a BSI. Gram negative (GN) resistance was defined as non-susceptibility to ceftazidime, cefepime, piperacillin/tazobactam, or imipenim-cilastin. Gram positive (GP) resistance was defined as non-susceptibility to vancomycin. Significance testing was performed using the Chi-Square Test (Stata Corp, College Station, TX).
Results: 378 patients underwent 409 transplants during the study. There were 411 BSIs in 205 transplants (GN = 137, GP = 263, fungal = 11); 204 transplants were BSI-free. Risk factors for any BSI were unrelated donor (p = 0.018) and acute GVHD (p = 0.001). The most common GN pathogens were Pseudomonas aeruginosa (n = 37) and Escherichia coli (n = 28); resistance was present in 24 GN BSIs (5.8% of all BSIs). Including organisms frequently resistant to anti-pseudomonal beta-lactams (such as Stenotrophomonas) increases this number to 39 (9.5% of all BSIs). The most common GP pathogens were CONS (n = 111), Enterococcus (n = 73), Staphylococcus aureus (n = 32), and Streptococcus species (n = 29); resistance was present in 50 BSIs (all due to vancomycin resistant enterococcus [VRE], 12.2% of all BSIs). 42 of the 50 patients with a VRE BSI had a VRE rectal swab performed prior to BSI onset; 38 were positive. Risk factors for resistant GN BSI were unrelated donor and chronic GVHD. 47 deaths occurred within 7 days of BSI and 119 deaths occurred within 30 days of BSI. Presence of resistance did not influence 7 or 30 day survival.
Conclusions: Empiric coverage with an anti-pseudomonal beta-lactam and vancomycin may miss almost 20% of BSI isolates in allo-HSCT recipients. Broadening of antimicrobials in patients with a positive gram stain until species and susceptibility results are known is prudent in those who are clinically deteriorating or who are known to be colonized with resistant organisms. However, since resistance did not affect short-term survival, this practice cannot be universally recommended.