Search

Objective: To evaluate the association between nosocomial infections (NI) and the mortality of Intensive Care Unit (ICU) patients, adjusting for the effect on mortality of other predictive variables., Methods: Prospective study on 944 concurrent patients admitted for at least 24 hours in the ICU of a tertiary level hospital between February and November of 1994. The association between NI (diagnosed using CDC criteria) and mortality was studied using multivariable logistic regression., Results: The cummulative incidence of mortality in the ICU was 11.2% (CI95% = 9.9-12.5). This incidence was significantly higher in infected patients with a crude mortality relative risk of 2.2 (CI95% = 1.5-3.1). In the multivariable analysis, the effect of NI (global, pneumonias, of the urinary tract and bacteriemias) on adjusted mortality depended on the patient's Acute Pysiology and Chronic Health Evaluation II (APACHE II) score. With low APACHE II scores, NI was associated with an increased mortality risk. Conversely, with higher APACHE II scores, the relevance of NI as a determinant of mortality decreased and prognosis was mainly associated with the patient's severity of illness., Conclusions: The association between NI and mortality, adjusting for other prognostic factors for mortality, is confirmed.

The interregional variation coefficient (VC) has been considered as an accurate measure of the dispersion of regional infant, neonatal, postneonatal and perinatal mortality rates. Thus, trends of annual VC have been analyzed, for each rate, from 1940 to 1986, to identify the evolution in time of the regional differences with respect to these mortality rates. None of the four mortality rates showed a decreasing trend in their respective VCs. This may indicate that interregional differences do not change along time. The postneonatal mortality rate has been shown to have the highest VCs during the study period (ranging from 23 to 40%), with an independent evolution with respect to neonatal mortality rate, probably because the factors that influence both rates are clearly different.