Candida species are the fourth most frequent cause of nosocomial bloodstream infections, and 25%-50% occur in critical care units. During an 18-month prospective study period, all patients admitted for > or = 72 hours to the surgical (SICUs) or neonatal intensive care units (NICUs) at each of the participant institutions were followed daily. Among 4,276 patients admitted to the seven SICUs in six centers, there were 42 nosocomial bloodstream infections due to Candida species (9.8/1,000 admissions; 0.99/1,000 patient-days). Of 2,847 babies admitted to the six NICUs, 35 acquired a nosocomial bloodstream infection due to Candida species (12.3/1,000 admissions; 0.64/1,000 patient-days). The following were the most commonly isolated Candida species causing bloodstream infections in the SICU: Candida albicans, 48%; Candida glabrata, 24%; Candida tropicalis, 19%; Candida parapsilosis, 7%; Candida species not otherwise specified, 2%. In the NICU the distribution was as follows: C. albicans, 63%; C. glabrata, 6%; C. parapsilosis, 29%; other, 3%. Of the patients, 30%-50% developed incidental stool colonization, 23% of SICU patients developed incidental urine colonization, and one-third of SICU health care workers' hands were positive for Candida species.

Specific strains of L. pneumophila can colonize hospital water supplies and cause nosocomial infections over long periods of time.

Objective. To validate the nosocomial infections surveillance system, establish its impact in morbi-mortality. Material and methods. Surveillance of every single patient admited during a one month period was done by one of us (DMG). Each posibile case was discussed with two other hospital epidemiologists (SPLR, MSRF). This intensive surveillance was compared against the routinely surveillance performed by the nurses. We included all hospitalized patients between 11 th July and 12 th of August according to CDC (Atlanta, GA) nosocomial infections definitions. Patients were followed everyday and information about age, gender, underlying diagnosis, microorganisms responsible for nosocomial infections, hospital length of stay and mortality. Results. During the study period 429 were admitted, 45 developed a nosocomial infection (cases) and 384 did not (controls). The incidence of nosocomial infections was 10.48 cases/ 100 discharges. The sensitivity and specificity of the surveillance system was 95.3 and 98.7%, respectively. Mortality in infected was 11.11% and in non infected was 2.4%. The average length of stay was 20 and 11days for cases and non infected respectively (p< 0.01). Urinary tract infections were the most common NI (42%), secondary bacteremia (14-ˇ%), pneumonia (11.11%) and deep surgical site infection (9.25%). The surgical wound infection rates were: 1.3%, 1.9% and 1.9% for clean, clean-contaminated and contaminated wounds. Patients with rapidly fatal diseases had an increased frequency of infections. The microorganisms most commonly isolated were Escherichia coli (28%), Staphylococcus aureus (11.11%), and Pseudomonas aeruginosa (8.6%). The level of antibiotic resistance was in average of 43% for those antibiotics tested. Conclusions. The sensitivity and specifi- ResumenObjetivo. Validar el programa de vigilancia de infecciones nosocomiales y conocer la morbilidad y la mortalidad. Material y métodos. Un médico especialmente capacitado, realizó vigilancia intensiva de todos los pacientes admitidos en el hospital. Los casos de infección fueron discutidos con otros dos médicos y el resultado se comparó con la vigilancia rutinaria. Se incluyó a todos los pacientes hospitalizados del 11 de julio al 12 de agosto de 1995, que no tenían un proceso infeccioso activo o que no manifestaban un periodo de incubación a su ingreso. Se siguieron diariamente y se registraron datos de: edad, sexo y padecimiento de ingreso. Se recabó información sobre tratamiento antimicrobiano, microrganismo aislado y susceptibilidad. Se evaluó el estado clínico final y se estimó el tiempo de estancia hospitalaria. Resultados. De 429 pacientes, 45 desarrollaron infección nosocomial (casos) y 384 no lo hicieron (con-