To describe the duration of fever, fever management, and outcomes among intensive care patients with fever and known or suspected infection. Prospective observational trial in three tertiary intensive care units over 6 weeks in 2010. Adult patients were screened for eligibility and inclusion if they had a fever of ≥38.0°C and known or suspected infection being treated with antimicrobials; those with neurological injury or elective surgery within 72 hours were excluded. Mean and peak daily temperatures were recorded and the use of antipyretics and other cooling measures were recorded over the first 7 days. Mortality, ICU-free survival, ventilator-free survival and renal replacement therapy-free survival were determined at Day 28. 51/565 patients (9.0%) were included. The mean daily peak temperature and the proportion of patients with a documented temperature of ≥38.0°C decreased over the first 3 days after first documented fever. Thereafter, the proportion of patients who had daily peak temperatures ≥38.0°C remained about 20%. Paracetamol was administered to 58%-70% of patients per day. Physical cooling was used at least once for 12% of patients. Mean ICU-free survival to Day 28 in eligible patients was 16.0 (SD, 9.2) days. The mortality rate of eligible patients was more than double that of ineligible patients (8/51 [16%] v 36/514 [7%]; P = 0.05). We have described the typical time course of fever in an easily identified cohort of patients with known or suspected infection and have determined that these patients have significant morbidity and mortality. This information is vital to the design of interventional studies for the treatment of fever in ICU.