Objective: To determine the influence of antimicrobial therapy and of predisposing illness on the septicemia mortality rate.
Method: All blood-culture-positive episodes of septicemia in the Department of Medicine at the University Hospital in Frankfurt between 1989 and 1993 were entered on a database. Underlying illnesses were classified as immunocompromising diseases (hematological malignancies, AIDS and others), severe chronic and chronic illnesses and no predisposing illnesses. Therapy was judged on the basis of the in-vitro-susceptibility of the organism ('appropriate') and the interval (no. of days) between the onset of septicemia and start of appropriate treatment noted. For mortality all deaths within 28 days after the onset of septicemia were counted.
Results: Overall mortality due to septicemia was 18.1%, ranging from 9.4% (organ transplantation) to 50% (liver cirrhosis) according to the underlying illness. Mortality in patients receiving appropriate treatment (83.1%) was 16% as opposed to 28%, if no appropriate treatment was given (p<0.001). Comparison of appropriate treatment started within and after 48 hours revealed a reduction in mortality from 30.9% to 15.4% for early appropriate therapy in patients with hematological malignancies (p<0.002). For septicemia in patients with AIDS and chronic illnesses mortality was significantly higher (p<0.05) if treatment remained inappropriate (AIDS 28.6%, chronic illness 33.3%), but was similar when early and delayed appropriate therapy were compared (AIDS: 13% vs. 12.8%, chronic illness 11.8% vs. 11.1%).
Conclusion: First-line treatment regimens for septicemia in patients with hematological malignancies should include the greatest possible part of the spectrum of causative organisms. In contrast to that it may be acceptable to rely to some extent on a change of treatment, when treating septicemia in patients with chronic illnesses or AIDS. These considerations are of value in the debate on rising health care costs. Several other facts, such as the stable mortality rate of 8 - 12% in previously healthy patients and the range of mortality from 9.4 - 50%, if predisposing illnesses are present, indicate the existence of adverse factors influencing the outcome of septicemia in spite of appropriate therapy. These pathophysiological factors will have to be studied in detail in order to improve the prognosis of septicemia further.