Bacteremias caused by Escherichia coli in cancer patients - analysis of 65 episodes.

OBJECTIVES: The aims of this study were to evaluate risk factors, clinical presentation, outcome and antimicrobial susceptibility in patients with Escherichia coli bacteremia occurring over seven years in a single cancer hospital. METHODS: Sixty five episodes of bacteremia from E. coli appearing over seven years from 12,301 admissions in a single cancer institution were retrospectively analyzed. RESULTS: The proportion of bacteremia caused by E. coli among Gram-negative bacteremia was 20.8% (the second most common organism after Pseudomonas aeruginosa), and infection-associated mortality was 17%. The incidence in 1989-1995 varied from 14.3 to 24.7%. The most common risk factors were: solid tumors as the underlying disease (70.7%); central venous catheter insertion (32.3%); prior surgery (46.2%), and prior chemotherapy within 48 h (44.4%). Neutropenia and urinary catheters did not place patients at high risk in any of the subgroups. When we compared the two subgroups of 61 cases of bacteremia - monomicrobial and polymicrobial (when E. coli was isolated from blood culture with another microorganism) - we found that acute leukemia and breakthrough (recurrence while receiving antibiotics) bacteremia were more frequently associated with polymicrobial E. coli bacteremia. There was also a difference in infection-associated mortality: monomicrobial bacteremia due to E. coli only had a significantly lower mortality in comparison with polymicrobial E. coli bacteremia (8.9 vs 35.0%, respectively; P<0.03). CONCLUSION: The susceptibility of 115 E. coli strains isolated from 65 episodes of bacteremia was stable. Only two episodes caused by quinolone-resistant strains occurred, both in 1995, after six years of using ofloxacin for prophylaxis in neutropenic patients in our hospital. We found that 85.2-91.3% of all strains were susceptible to aminoglycosides, 97.8% to quinolones, and 90-100% to third generation cephalosporins and imipenems. The patients most commonly infected had solid tumors and the mortality was only 17%.

Candida glabrata fungemia in a tertiary cancer institution in Slovakia.

Because of controversial data on virulence and mortality, six cases of fungemia caused by Candida glabrata were reviewed in a single cancer institution within 8 years. Risk factors and outcome of C. glabrata, Candida albicans, and other non-albicans Candida spp. appearing within the same period under the same antibiotic policy at the same institution were compared. Among other non-albicans Candida spp. in 1990-1997 C. glabrata fungemias showed a decreasing tendency, from 9% to 4.5% in 1997. Analyzing the proportion of C. glabrata among blood cultures, among 170 positive blood cultures 12 were caused by C. glabrata (6.2% among all pathogens and 24% among non-albicans Candida spp.). C. glabrata among all fungemias was diagnosed as the fourth most common pathogen after C. albicans, C. krusei, and C. parapsilosis. Three of six C. glabrata fungemias were breakthrough. Two appeared during prophylaxis with itraconazole and one during fluconazole prophylaxis. Five of six received broadspectrum antibiotic therapy with third-generation cephalosporines, five of six had vascular catheter insertion, four of six were neutropenic, and two of six received amphotericin B therapy. One patient died before his blood cultures were reported to be positive. Overall mortality of C. glabrata fungemia was 16.7%. One patient died of underlying disease with fungemia. There were no significant differences in risk factors between C. glabrata and C. albicans. However, overall and crude mortality was lower in C. glabrata than in C. albicans (25.5% vs. 16.7%; P = 0.03). Attributable mortality was lower in comparison to C. albicans (0 vs. 15.7% in C. albicans; P = 0.001).

Bacteremia due to multiresistant gram-negative bacilli in neutropenic cancer patients: a case-controlled study.

The aim of this study was to assess whether multiresistant gram-negative bacteremias (MRGNB) were associated with specific risk factors for higher mortality than sensitive gram-negative bacteremias. Two groups of subjects: (51 cases and 102 controls) were matched for sex, age, underlying disease, and neutropenia. There were no significant differences in the incidence of cytotoxic chemotherapy administered, vascular catheter insertion, catheter as source of bacteremia, and etiology of bacteremia. The proportion of Klebsiella-Enterobacter, Pseudomonas aeruginosa, Acinetobacter spp., and Stenotrophomonas maltophilia was similar in the two groups. Prior surgery (21.6% vs 7.9%, P < 0.05) was significantly associated with sensitive gram-negative bacteremia. Previous prophylaxis with ofloxacin (45.1% vs 24.5%; P < 0.05) and prior therapy with broad-spectrum antibiotics (41.2% vs 27.5%; P < 0.05), such as first and second generation cephalosporins (19.6% vs 7.8%; P < 0.05), third generation cephalosporins (41.2% vs 13.7%; P < 0.01), aminoglycosides (39.2% vs 9.8%; P < 0.01), ofloxacin (11.8% vs 2.0%; P < 0.005), and imipenem (19.6% vs 2.0%; P < 0.001) were significantly more frequently observed among cases than controls. Cases (patients with bacteremia due to multiresistant gram-negative bacteremias) were also significantly more frequently infected with bacteria resistant to ceftazidime (68.6% vs 17.6%; P < 0.001), amikacin (52.9% vs 7.8%; P < 0.001), imipenem (50.1% vs 23.5%; P < 0.05), ciprofloxacin (32.1% vs 5.9%; P < 0.001), and piperacillin (41.2% vs 7.8%; P < 0.01). With regard to outcome, attributable mortality was similar (15.7% vs 13.8%; not significant) in the two groups; however, cure rates were lower among cases (patients infected with MRGNB) because crude mortality was higher in cases (35.3% vs 13.8%; P < 0.01) than in controls.