Common Causes of Nosocomial Infections and Their Susceptibility Patterns in Two Hospitals in Addis Ababa

Abstract

Background: Nosocomial infection is defined as an infection, which is not present or incubating on admission and is usually not manifested in the first 72 hours of hospitalization. Infectious diseases including hospital acquired infections account for 25% of mortality worldwide and 45% in low-income countries

Objective: To assess the susceptibility pattern of the bacterial causes of common nosocomial infections and their susceptibility pattern in surgical patients.

Methods: A one-year prospective survey of nosocomial infections of surgical patients admitted in 1999-2000 at Tikur Anbesa

Hospital (TAH) and Saint Paul Hospital (SPH) was performed. The samples were collected and processed following standard microbiological techniques as part of the routine clinical management of the patient. Antibiotic sensitivity testing was done on pure culture isolates employing the disc-diffusion method for the commonly used antibiotics

Results: A total of 133 patients in Tikur Anbesa (69) and Saint Paul (64) were found to have nosocomial infection among surgical patients admitted in the study period making the prevalence 6.19% in TAH and 5.74% in SPH. Wound infection was the leading nosocomial infection 36(52.1%) in TAH and 54(79%) in SPH followed by urinary tract infection 26 (38%) in TAH and 10(15.9%0) in SPH. The most frequently isolated organisms were Escherichia coli 16(25.4%) in TAH and 29(46.0%) in SPH, followed by Klebsiella 22(34%) in TAH, 13 (25.4%) in SPH and Proteus species 5(6.8). The percentage of resistant strains to Gentamycin was found to be 56% in samples taken from Hospital acquired infection at TAH. The resistance level is even higher for the penicillin group such as, crystalline penicillin and ampicillin which is (95%) of isolates from nosocomial infection

Conclusions: Most isolates from hospital-acquired infection were found to be resistant to the most frequentl used antibiotics in the setting where the study was conducted.