The diagnosis of infection in the intensive care unit is confounded by
the presence of non-infectious causes of leucocytosis. Unless such ca
uses are recognised, time and effort will be spent on unnecessary inve
stigations and treatments. In a prospective study we have shown that t
he transfusion of blood frequently (45/50 patients) causes an acute le
ucocytosis in such patients. This effect was not seen in 8 patients wh
o received plasma. Blood transfusion should be added to the list of no
n-infectious causes of leucocytosis in the critically ill.