Introduction. Incidence and mortality due to multiresistant (MR) bloodstream infections (BSI) are increasing in the last years among hematological patients (pts). Also rate of MR colonization seems to be relevant, but few data are available on the actual incidence of MR colonization and the probability of developing a BSI in hematological pts. In order to evaluate the incidence of MR bacterial colonization and the probability of developing MR BSI in hematological pts, we carried out a multicentric prospective observational study within the SEIFEM group.

Patients and Methods. Since March 1st 2015, all cancer patients were screened at least for rectal swab at admission in 11 Hematology Centres participating to SEIFEM; culture of other sites were performed if clinically indicated. Patients showing MR bacterial colonization were recorded in a database where the occurrence of any BSI was correlated with age, gender, type and phase of haematological disease, stem cell transplantation (SCT), presence of invasive devices, type of colonizing bacteria.

Results. During a 3-month period (March 1st -May 31st 2015), 75 pts with MR colonization were observed. Incidence was 9.8% among all admitted pts (75/764). Median age was 61y (range 20-81) and M/F ratio 46/29. Thirty-three pts were affected by acute leukemia (AL), 30 by lymphoma (Ly), 10 by myeloma (MM), 1 by myelofibrosis and 1 by aplastic anemia. Nineteen pts underwent SCT during the period of observation (11 autologous and 8 allogeneic, respectively). Vancomycin resistant enterococci (VRE) were responsible for colonization in 6 (8%) pts, extended spectrum β-lactamases producing (ESBL) enterobacteria in 29 (38.8%) and carbapenemase producing (CP) Gram-Negative Rods (GNR) in 49 (65.3%). Seven pts showed multiple colonizations. In 8 cases colonizing enterobacteria (6 E. coli, 1 E. cloacae, 1 P. mirabilis) were both ESBL and carbapenemase producers. Rectum was the most frequent site of colonization (70.7%). No risk factor for type of colonizing bacteria emerged except for previous intensive care unit stay for CP GNR colonization (p=0.02, chi-square test). Overall, 15 (20%) pts colonized with MR bacteria developed BSI by the same pathogen (MR related BSI) (1 VRE, 5 ESBL producing E. coli, 5 CP K. pneumoniae, 2 CP P. aeruginosa and 2 CP Acinetobacter spp). Therefore, rate of related BSI according to type of antibiotic resistance was 16.7% for VRE colonization, 17.2% for ESBL producing enterobacteria and 18.4% for CP GNR, respectively. Among CP enterobacteria, K. pneumoniae, but not other enterobacteria, was predictive of related BSI (25%). All but 2 related BSI occurred during neutropenia. At multivariate analysis, previous cephalosporin treatment was associated to related BSI (OR 0.25 [CI 0.07-0.91], p=0.03). Unrelated BSI were also observed in 15 pts (20%), including 3 MR BSI (2 CP P. aeruginosa and 1 CP K. pneumoniae). After a median follow-up of 61 days (range 30-116), 9 pts died (12%). Death was attributable to CP GNR related BSI in 3 pts. Multivariate analysis showed that CP GNR related BSI (OR 6.1, CI 1-36.2, p=0.04), together with a relapsed/refractory hematological disease (OR 10, CI 1.5-66.4, p=0.02) and presence of urinary catheter (OR 9.1, CI 1.3-62.1, p=0.02), was an independent predictor variable for death.

Conclusions. Incidence of MR colonizing bacteria, particularly CP GNR, is quite high among hematological cancer pts and is associated with related BSI in a considerable proportion of them. Unrelated MR BSI may also occur. Previous antibiotic treatment with cephalosporins is a risk factor for MR related BSI and CP related BSI are predictive of death. Empiric antibiotic treatment should be planned taking into account these results.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract

Sign in via your Institution