A retrospective evaluation of 350 axillary and femoral artery cannulations in ICU patients revealed a 6.3-day mean duration of catheterization. Minor incidents such as hematoma (3.7%), catheter malfunction (5.1%), and local inflammation (2.5%) were noticed. The catheter was removed in 25 patients in whom it was suspected of causing sepsis. Six septic patients had the same organism in catheter and arterial blood cultures (with a negative venous blood culture), probably denoting an infected, catheter-linked thrombus. In 241 cases the catheter was removed at the end of the monitoring period. In only 21 of these cases did an organism grow in the catheter culture, but in 18 of these cases, the same organism was found in previous blood, sputum, and/or urine cultures. Thus, in only six patients (of 266 survivors) could arterial cannulation be held responsible for generalized sepsis. However, because no bacteriologic study was performed on catheters removed after the patient's death, this number could have been higher. Percutaneous axillary and femoral artery cannulation is highly recommended for extended monitoring because of its low incidence of minor complications, and no evident danger of tissue ischemia.