Introduction Septic shock remains a major problem in our intensive care units (ICUs) in terms of both morbidity and mor- tality. Despite promising results of new immunomodula- tory therapies in animals, clinical trials of such com- pounds have so far been negative [1]. Treatment is thus based on the control of infection and haemodynamic stabilization. Control of infection relies on identifica- tion of the infecting organism, administration of ade- quate doses of appropriate antibiotics and surgical drainage or removal of the source where possible. In some patients, infection is documented clinically, but microbiological data are negative. As many as 50 % of patients with severe sepsis may also not have document- ed bacteraemia [2]. In other cases, although presenta- tion is typical and other causes of shock can be exclud- ed, making it likely that there is an underlying infectious cause, the infectious source cannot be identified. In the present study we asked whether patients with presumed sepsis but without clinically identified infec- tion (whom we termed septic shock-like), have a differ- ent outcome than patients who present with a more ex- plicit form of septic shock. We reviewed patients treated in the ICU over a period of 3 years, to define the rela- tive proportions of these two groups of patients and identify differences, particularly in mortality, between them. Patients and methods We reviewed data from patients treated for septic shock in a 31- bed ICU over a 3-year period (March 1993±March 1996). In this unit, circulatory shock is defined as arterial hypotension with a sys- tolic blood pressure below 90 mmHg (or a reduction by 50 % in a previously hypertensive patient) resistant to fluid challenge and re- quiring vasopressor support, with signs of tissue hypoperfusion, such as oliguria and alterations in mental status, and with hyperlac- tataemia (blood lactate concentration > 2 mEq/l). A diagnosis of septic shock is dependent on three factors: (1) signs of sepsis in- cluding fever or hypothermia and an altered white blood cell count, (2) a high cardiac output ( > 3.5 l/m 2 ) and/or a low systemic W. J. Reyes S. Brimioulle J.-L. Vincent Septic shock without documented infection: an uncommon entity with a high mortality Received: 27 January 1999 Accepted: 18 August 1999 W.J. Reyes ´ S. Brimioulle ´ J.-L. Vincent ( ) ) Dept of Intensive Care, Erasme University Hospital, Free University of Brussels, Route de Lennik 808, B-1070 Brussels, Belgium e-mail: jlvincen@ulb.ac.be Tel.: + 32-2-5553380 Fax: + 32-2-5 55 45 55 Abstract Objectives: To determine whether patients with clinically identified infection have the same outcome as patients with apparent sepsis but no identified infectious source. Design: Retrospective analysis of patient data. Patients: All patients treated with septic shock in a 31-bed intensive care unit (ICU) over a 3-year peri- od. Results: Data from 227 patients were analysed. Eighty-seven percent had a clinically identified source of in- fection. ICU mortality was higher in septic shock patients without a clini- cally identified source of infection than in those with an identified source of infection (86 % versus 66 %, p < 0.05). Conclusions: A small number of pa- tients presenting with septic shock have no clinically identified infec- tion. These patients have a higher mortality rate than patients in whom an infection is identified. Key words Bacteraemia ´ Infection ´ Organ failure ´ Morbidity ´ Mortality ´ Outcome ´ Severe sepsis Intensive Care Med (1999) 25: 1267±1270 Ó Springer-Verlag 1999 ORIGINAL