Estimating mortality risk in preoperative patients using immunologic, nutritional, and acute-phase response variables

We measured the delayed type hypersensitivity (DTH) skin test response, along with additional variables of host immunocompetence in 245 preoperative patients to determine which variables are associated with septic-related deaths following operation. Of the 14 deaths (5.7%), 12 were related to sepsis and in 2 sepsis was contributory. The DTH response (p less than 0.00001), age (p less than 0.0002), serum albumin (p less than 0.003), hemoglobin (p less than 0.02), and total hemolytic complement (p less than 0.03), were significantly different between those who died and those who lived. By logistic regression analysis, only the DTH skin test response (log likelihood = 41.7, improvement X2 = 6.24, p less than 0.012) and the serum albumin (log likelihood = 44.8, improvement X2 = 17.7, p less than 0.001) were significantly and independently associated with the deaths. The resultant probability of mortality calculation equation was tested in a separate validation group of 519 patients (mortality = 5%) and yielded a good predictive capability as assessed by (1) X2 = 0.08 between observed and expected deaths, NS; (2) Goodman-Kruskall G statistic = 0.673) Receiver-Operating-Characteristic (ROC) curve analysis with an area under the ROC curve, Az = 0.79 +/- 0.05. We conclude that a reduced immune response (DTH skin test anergy) plus a nutritional deficit and/or acute-phase response change are both associated with increased septic-related deaths in elective surgical patients.