Biological measures for the formulation of a hospital prognostic index

Open Access
Publisher Website
Google Scholar
Abstract
Factors that might predict sepsis and mortality were investigated in 282 hospitalized patients referred for metabolic and nutritional therapy. Routine assessment parameters, including delayed hypersensitivity response to recall skin antigens, serum albumin and transferrin, total lymphocyte count, and anthropometric measurements were evaluated with respect to ultimate outcome. A discriminant function (0.91 (albumin) − 1.00 delayed hypersensitivity − 1.44 (sepsis) + 0.98 (diagnosis) − 1.09) was developed having an overall predictive value for subsequent hospital mortality of 72%, a sensitivity of 74%, and a specificity of 66%. Serum albumin was found to be the best single indicator of concurrent sepsis and anergy, and predictor of mortality; an initial albumin <2.2 g/dl was associated with a greater than 75% chance of having concurrent anergy and sepsis and dying. An improvement in delayed hypersensitivity response was the most accurate predictor of an improved prognosis (overall predictive value 86%, sensitivity 93%, specificity 63%). Thus, the use of serum albumin, delayed hypersensitivity response to recall skin antigens, clinical status, i.e., septic or not septic, and presence or absence of cancer cachexia can serve to identify high risk patients in an objective fashion and evaluate the effectiveness of hospital therapy. This information may also assist the hospital team in allocation of resources.