An analysis of hospital-acquired bacteraemia in intensive care unit patients in a university hospital in Kuwait.

An analysis of hospital-acquired bacteraemia among ICU patients was carried out over a two-year period in order to determine the incidence, associated mortality rate and susceptibility pattern of causative pathogens. There was a high incidence of bacteraemia, occurring in 127 (18.4%) of 692 patients. Mortality attributable to nosocomial bacteraemia was 52% of the total 79 deaths from all causes. The highest mortality rate (58.5%) occurred in patients with fungal infections, whilst death from Gram-negative bacteraemia was only 17%. Over 98% of patients had underlying disease. Nearly half (46.8%) of 267 organisms isolated were Gram-positive. In comparison, Gram-negative bacteria accounted for 36.6% and the rest (17.6%) were fungi (mainly Candida albicans). The majority of the bactereamic episodes were monomicrobial (90.2%). Coagulase-negative staphylococci (CNS) were the commonest pathogens isolated, representing 32.6% of all organisms. Inducible beta-lactamase producing organism (Enterobacter spp. 9.7%, Serratia marcescens 6.7%, Klebsiella pneumoniae 6% and Pseudomonas aeruginosa 6%) formed the bulk of Gram-negative bacteria. In contrast, Escherichia coli (7.5%) and K. pneumoniae (4%) were the commonest Gram-negative bacteria from hospital-acquired bacteraemia in the general hospital population. The majority (80%) of CNS were resistant to methicillin (MRSE) but susceptible to vancomycin; they were relatively resistant to erythromycin, clindamycin and beta-lactams antibiotics. Whilst Gram-negative organisms were relatively susceptible to imipenem (85%), ciprofloxacin (88%) and amikacin (87%), they had unacceptably low levels of susceptibility to cefuroxime (59.3%), cefotaxime (71%), ceftazidime (60.9%), and piperacillin (51.1%). This study shows that hospital-acquired bacteraemia in ICU patients carries a poor prognosis. Information regarding the infective agents and their susceptibility in the ICU setting is valuable for the selection of empirical therapy before culture and susceptibility results are known.