INTRODUCTION: Sepsis is the most common noncoronary cause of mortality in intensive care units (ICUs). This study compared different systems for predicting outcomes in a population of critically ill patients with sepsis originating from different infection sites, including intra-abdominal and pulmonary infections. METHODS: This post hoc analysis of an accumulated database enrolled 161 heterogeneous critically ill patients diagnosed as severe sepsis and septic shock patients admitted to medical ICUs from June 2005 to May 2007. Demographic characteristics, clinical and laboratory variables, comorbidities and infection source were prospectively recorded on the first day of ICU admission. Patient evaluations included acute physiology and chronic health evaluation (APACHE) II, APACHE III, sequential organ failure assessment scores, organ system failure and risk of renal failure, injury to kidney, failure of kidney function, loss of kidney function and end-stage renal failure (RIFLE) classification. RESULTS: Regarding the different originating sites of severe sepsis, intra-abdominal infections and pulmonary infections had the highest mortality rates (83.3% and 48.5%, respectively; P < 0.001). The APACHE III was the best mortality predictor for the overall sepsis population [areas under the receiver operating characteristic curve (AUROC) 0.800], whereas RIFLE classification was the best predictor in those with intra-abdominal infection (AUROC 0.856). The AUROC analyses verified that RIFLE classification had significantly (P < 0.05) better discriminatory power for predicting hospital mortality in patients with intra-abdominal infections than in those with pulmonary infections (AUROC 0.545). CONCLUSIONS: This investigation confirms that different infection sites have different outcomes. In terms of mortality prediction, outcome scoring systems are significantly more accurate in patients with intra-abdominal infections than in those with pulmonary infections.
INTRODUCTION:Sepsis is the most common noncoronary cause of mortality in intensive care units (ICUs). This study compared different systems for predicting outcomes in a population of critically illpatients with sepsis originating from different infection sites, including intra-abdominal and pulmonary infections. METHODS: This post hoc analysis of an accumulated database enrolled 161 heterogeneous critically illpatients diagnosed as severe sepsis and septic shockpatients admitted to medical ICUs from June 2005 to May 2007. Demographic characteristics, clinical and laboratory variables, comorbidities and infection source were prospectively recorded on the first day of ICU admission. Patient evaluations included acute physiology and chronic health evaluation (APACHE) II, APACHE III, sequential organ failure assessment scores, organ system failure and risk of renal failure, injury to kidney, failure of kidney function, loss of kidney function and end-stage renal failure (RIFLE) classification. RESULTS: Regarding the different originating sites of severe sepsis, intra-abdominal infections and pulmonary infections had the highest mortality rates (83.3% and 48.5%, respectively; P < 0.001). The APACHE III was the best mortality predictor for the overall sepsis population [areas under the receiver operating characteristic curve (AUROC) 0.800], whereas RIFLE classification was the best predictor in those with intra-abdominal infection (AUROC 0.856). The AUROC analyses verified that RIFLE classification had significantly (P < 0.05) better discriminatory power for predicting hospital mortality in patients with intra-abdominal infections than in those with pulmonary infections (AUROC 0.545). CONCLUSIONS: This investigation confirms that different infection sites have different outcomes. In terms of mortality prediction, outcome scoring systems are significantly more accurate in patients with intra-abdominal infections than in those with pulmonary infections.
Authors: Sami Safadi; Musab S Hommos; Felicity T Enders; John C Lieske; Kianoush B Kashani Journal: Mayo Clin Proc Date: 2020-01-31 Impact factor: 7.616