Abstract 4868


Bacteremia is a major cause of morbidity and mortality in patients with hematological disorders during chemotherapy-induced neutropenia. Initial antimicrobial treatment is chosen empirically. Increasing antimicrobial resistance is a global obstacle for improving outcome for hematological patients. The aim of this study was to compare temporal trends in species distribution and antimicrobial susceptibility in bacteremic patients hospitalized in a hematology ward at Karolinska University Hospital, Stockholm, Sweden, over a 21-year period.

Materials and methods:

A total of 794 clinically significant isolates, 667 clinical episodes of bacteremia in 463 patients were identified during the last 7 years and the results were compared with those published from the same ward during the preceding 14 years. Isolates and episodes were identified from the laboratory information system, which has been in use during the entire 21-year period. For coagulase-negative staphylococci (CoNS), Corynebacterium spp., Propionibacterium spp. and Bacillus spp. bacteremia was considered significant only if the isolates were detected in two or more blood specimens. No major changes in patient selection were introduced during these 21 years and quinolone prophylaxis was used only to a very limited extent.


Any shift between Gram-negative and Gram-positive bacteria has not been noted at any time and the proportion of Gram-positive isolates has remained at a level of 53–55%. In the last 7-year period the male patient predominance increased from 58% to 62%, and the median age increased from 58 to 62 years. The 7- and 30-day crude mortality rates were 5.3% and 14.8%, respectively, compared with 6.3% and 16% in the preceding 7-year period (p>0.05). Polymicrobial bacteremia, defined as growth of more than one organism during the same bacteremic episode within 24 hours or 72 hours, was seen in 13.7% and 14.8% of patients respectively, compared to 11% (24 hours) in the preceding 7-year period (p>0.05). Polymicrobial bacteremia was seen in 25% and 28% respectively of patients who died within 7 days. Crude mortality at day 7 in patients with polymicrobial bacteremia was significantly higher than for patients with monomicrobial bacteremia (p=<0.0001).

As in the previous periods the dominating pathogens were Escherichia coli (18%), CoNS (15%), viridans streptococci (14%) and Klebsiella spp. (10%). A significant increase of Enterobacter spp. (from 3.1 to 5.4%) and decrease of Staphylococcus aureus (from 10 to 6.9%) was seen compared with the middle 7-year period. Enterococcus faecium increased during the middle 7-year period and has remained frequent and stable at 8%.

In E. coli resistance to piperacillin-tazobactam (TZP) was 9.2%, ciprofloxacin (CIP) 8.5% and ceftazidime (CAZ) 2.1%, whereas no resistance was observed to imipenem-cilastatin (IPM). In Klebsiella spp. resistance to TZP was 9.0%, CIP 1.3% and CAZ 1.3%, whereas no resistance was observed to IPM. Among Enterobacter spp. resistance to TZP was 26%, CAZ 26%, IPM 2.3% and CIP 4.7%. In P. aeruginosa resistance to IPM was 12%, TZP 4.8% and CIP 7.1%, whereas no CAZ-resistant strains were found. Among CoNS resistance to isoxazolyl-penicillin was 71% in the last period, compared to 58% in the middle period. Only one isolate of E. faecium (cultured in the last 7-year period) was resistant to vancomycin. Two methicillin-resistant S. aureus were detected in the last 7-year period. Resistance to benzyl-penicillin among viridans streptococci was 20% in the last 7-year period and 12% in the middle 7-year period. Only the increase in CIP resistance in E. coli during the last 7-year period was statistically significant (p= 0.02).


We observed a stable distribution of species and a low rate of antimicrobial resistance during a period of 21 years. In the last 7-year period a significant increase in CIP-resistance was observed in E. coli. This finding can possibly be related to a very limited use of antibacterial prophylaxis in our centre, but also reflects the general low level of antimicrobial resistance in Sweden. Crude mortality was unchanged during the study period. Mortality was higher in patients with poly- th.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.