Multidrug-resistant Acinetobacter baumannii has emerged as a significant clinical problem worldwide and colistin is being used increasingly as "salvage" therapy. MICs of colistin against A. baumannii indicate its significant activity. However, resistance to colistin in A. baumannii has been reported recently. Clonotypes of 16 clinical A. baumannii isolates and ATCC 19606 were determined by pulsed-field gel electrophoresis (PFGE), and colistin MICs were measured. The time-kill kinetics of colistin against A. baumannii ATCC 19606 and clinical isolate 6 were investigated, and population analysis profiles (PAPs) were conducted. Resistance development was investigated by serial passaging with or without exposure to colistin. Five different PFGE banding patterns were found in the clinical isolates. MICs of colistin against all isolates were within 0.25 to 2 g/ml. Colistin showed early concentration-dependent killing, but bacterial regrowth was observed at 24 h. PAPs revealed that heteroresistance to colistin occurred in 15 of the 16 clinical isolates. Subpopulations (<0.1% from inocula of 10 8 to 10 9 CFU/ml) of ATCC 19606, and most clinical isolates grew in the presence of colistin 3 to 10 g/ml. Four successive passages of ATCC 19606 in broth containing colistin (up to 200 g/ml) substantially increased the proportion of the resistant subpopulations able to grow in the presence of colistin at 10 g/ml from 0.000023 to 100%; even after 16 passages in colistin-free broth, the proportion only decreased to 2.1%. This represents the first demonstration of heterogeneous colistin-resistant A. baumannii in "colistinsusceptible" clinical isolates. Our findings give a strong warning that colistin-resistant A. baumannii may be observed more frequently due to potential suboptimal dosage regimens recommended in the product information of some products of colistin methanesulfonate.

Summary Klebsiella pneumoniae colonization is a significant risk factor for infection in ICU, with approximately half of K. pneumoniae infections resulting from patients’ own microbiota. Screening for colonization on admission could limit risk of infection in the colonized patient and others.

OR DECADES, INFECTIVE ENDOCARditis (IE) caused by Staphylococcus aureus has been viewed primarily as a community-acquired disease, especially associated with injection drug use. [1][2][3][4][5][6][7] In contrast, patients with nosocomial or intravascular catheterassociated S aureus bacteremia were considered to be at low risk for IE. 5,6,[8][9][10][11] S aureus IE is relatively infrequent at any individual institution, and observations of its characteristics were based primarily upon relatively small samples, 1,3,6,9,[12][13][14] single-center experiences, 5,6,8,9,[13][14][15][16] or retrospectively identified patients. 2,7,8,15,16 Patient characteristics, treatment practices, and outcomes in these single-center studiesoftendifferedconsiderably.Moreover, because no large, prospectively col-See also pp 3022 and 3061.

In this large, prospective, multinational cohort, more than one half of all cases of non-HACEK gram-negative bacillus endocarditis were associated with health care contact. Non-HACEK gram-negative bacillus endocarditis is not primarily a disease of injection drug users.

No abstract