Korean J Med.  1998 Jul;55(1):11-20.

Comparative Analysis of Risk Factors and Severity of Illness Scores for Predicting Mortality in Sepsis Patients Treated in Medical Intensive Care Unit

• Shin TR
• Cheon SH
• Chang JH

Abstract

OBJECTIVES
Despite the advances in medical practice and technology, a steady increase in the incidence of sepsis has been reported and the mortality has been persistently high ranging from 20 to 50%, especially in medical intensive care unit. Therefore, objective assessment of severity of illness based on risk estimates derived from severity of illness score has been important. Thus this study was designed for identification of risk factors and comparison of severity of illness scores: Acute physiology and chronic health evaluation II(APACHE II), Acute physiology and chronic health evaluation III (APACHE III), Multiorgan failure scoring system(MOF), Simplified acute physiology score(SAPS), Septic severity score(SSS).
METHODS
A retrospective clinical study was done on 97 patients with sepsis treated in medical intensive care unit, who had been admitted to Ewha Womans University Mokdong Hospital from September 1994 to May 1997. Multiple factors which may influence mortality and severity of illness score were evaluated by univariate and logistic regression analysis.
RESULTS
1) Of the 97 patients, 52 were male and 45 were female. The mean age was 62 years and the overall mortality was 59.8%. 2) On severity of illness, sepsis was 21cases, severe sepsis 36 and septic shock 40. On primary source of infection, pulmonary origin was 39cases, gastrointestinal origin 23 and urinary tract origin 13. The infection caused by documented organism was 53cases. 3) Based on the univariate analysis, age, sex, respiratory rate, number of organ failure, admission source, ventilatory support, hematocrit, WBC, prothrombin time, Na and K, pH, PaO2/FiO2, D(A-a)O2 and the following severity of illness scores :APACHE II, APACHE III, MOF, SAPS, and SSS for predicting mortality risk were all significant factors discriminating between survivors and nonsurvivors. 4) In distribution curve of APACHE II, APACHE III, MOF, SAPS, and SSS, the cut-off values in predictive risk of 50% were 21, 88, 4, 15, and 29. By logistic regression analysis after substituting their cut off value, APACHE II and APACHE III were significant independent predictors influencing survival with odd ratio 3.82 in APACHE II and 4.62 in APACHE III. In predictive risk of 50%, sensitivity was 55%, specificity 85%, accuracy 85% and relative risk 6.77 in APACHE II. And sensitivity was 47%, specificity 97%, accuracy 88% and relative risk 7.62 in APACHE III.
CONCLUSION
APACHE II and APACHE III were significant predictor influencing mortality in sepsis patients treated in medical intensive care unit and APACHE III was superior to that of its predecessor. In near future, prospective and multicentral studies are required to confirm these results.

MeSH Terms

APACHE
Female
Hematocrit
Humans
Hydrogen-Ion Concentration
Incidence
Intensive Care Units*
Critical Care*
Logistic Models
Male
Mortality*
Physiology
Prothrombin Time
Respiratory Rate
Retrospective Studies
Risk Factors*
Sensitivity and Specificity
Sepsis*
Shock, Septic
Survivors
Urinary Tract