Israel Medical Association Journal

Background: Skin colonization of microorganisms in blood cultures (BCs) are generally considered clinically non-significant and can be the source of a true infection, particularly in immunosuppressed patients.

Objectives: To study the epidemiology and risk factors for bacteremia caused by contaminants.

Methods: This retrospective, descriptive study is based on adult BCs collected (2004–2013) and categorized as positive (True bacteremia [TrueB] or contamination) or negative. Clinical, demographic, and laboratory characteristics of BCs positive for the six most common potential contaminant pathogens (PCPs) that can cause TrueB and contamination (Coagulase-negative Staphylococcus [CoNS], Streptococcus viridans, Propionibacterium acnes, Corynebacterium spp., Bacillus spp., Clostridium spp.) were assessed. Ninety-two TrueB were identified vs. 196 contaminations (1:2 ratio).

Results: From 74,014 BCs, PCPs were found in 3735 samples, of which 3643 (97.5%) were contaminations and 92 (2.5%) were TrueB. The overall rate of BC contamination decreased during the study period from 6.7% to 3.8%. CoNS was the most common PCP. Bacillus spp. were only contaminants. Clostridium spp. and Streptococcus viridans were more often TrueB. In a multivariate model, predictors of TrueB included high creatinine levels, Streptococcus viridans in BC, and multiple positive BCs. A single culture of CoNS was strongly predictive of contamination.

Conclusions: Ten years of data on BCs, focusing on six PCPs, demonstrates a significant, yet insufficient reduction in the rate of contamination. High creatinine level, isolation of Streptococcus viridans, and multiple positive BCs were predictors of TrueB, while growth of CoNS was strongly predictive of contamination. This model could assist in diagnostic and therapeutic decision making.

Background: The role of routine active surveillance cultures (ASCs) in predicting consequent blood stream infections is unclear.

Objectives: To determine prospectively whether routine screening ASCs obtained on admission to the intensive care unit (ICU) can predict the causative agent of subsequent bloodstream infections.

Methods: We prospectively studied a cohort of 100 mechanically ventilated patients admitted consecutively to a 16-bed ICU. On admission, ASCs were obtained from four sites: skin cultures (swabs) from the axillary region, rectal swabs, nasal swabs, and deep tracheal aspirates. Thereafter, cultures were obtained from all four sites daily for the next 5 days of the ICU stay.

Results: Of the 100 recruited patients 31 (31%) had culture-proven bacteremia; the median time to development of bacteremia was 5 days (range 1–18). Patients with bacteremia had a longer median ICU stay than patients without bacteremia: 14 days (range 2–45) vs. 5 days (1–41) (P < 0.001). ICU and 28 day mortality were similar in patients with and without bacteremia. Most ASCs grew multiple organisms. However, there was no association between pathogens growing on ASCs and eventual development of bacteremia.

Conclusions: ASCs obtained on ICU admission did not identify the causative agents of most subsequent bacteremia events. Therefore, bloodstream infections could not be related to ASCs.

Background: The epidemiology of bacteremic febrile neutropenia differs between locations and constitutes the basis for selection of empiric antibiotic therapy for febrile neutropenia.

Objectives: To describe the epidemiology of bacteremia among patients with neutropenia in a single center in Israel.

Methods: We conducted a prospective data collection on all patients with neutropenia (< 500/mm³) and clinically significant bacteremia or fungemia during the period 1988–2004.

Results: Among adults (462 episodes) the most common bloodstream isolate was Esherichia coli. Gram-negative bacteria predominated throughout the study period and the ratio between Gram-negative and Gram-positive bacteremia increased from 1.7 to 2.3 throughout the study period. Among children (752 episodes), the ratio between Gram-negative and Gram-positive bacteremia reversed from 1.2 to 0.7, due to increasing prevalence of coagulase-negative staphylcoccal bacteremia. Both among adults and children, the length of hospital stay prior to bacteremia had a major impact on the pathogens causing bacteremia and their antibiotic susceptibilities. The prevalence of E. coli decreased with time in hospital, while the rates of Pseudomonas aeruginosa, Klebsiella pneumoniae, Enterobacter spp., Acinetobacter spp., Enterococcus spp. and Candida spp. increased. Resistance to broad-spectrum empiric monotherapy in our center was observed in > 40% of Gram-negative bacteria when bacteremia was acquired after 14 days in hospital.
Conclusions: Improved infection-control measures for neutropenic cancer patients in our center are needed. Empiric antibiotic treatment should be tailored to patients’ risk for multidrug-resistant organisms. Individual hospitals should monitor infection epidemiology among cancer patients to guide empiric antibiotic treatment