To determine whether the total energy expenditure (TEE) determined on a single day reliably estimates TEE on subsequent days of a patient's ICU stay, day-to-day variability of TEE and its relation to clinical condition were assessed. TEE was measured by indirect calorimetry in 60 mechanically ventilated critically ill surgical patients during a 2 to 7-day period. Clinical condition was scored by daily determination of the simplified acute physiology score (SAPS). Day-to-day variation of TEE was significantly (p less than .0001) influenced by body temperature, but not by SAPS or day of measurement. In the individual patient, day-to-day variability of TEE may be large. When TEE on day 1 is used to estimate TEE on subsequent days, errors of up to 31% of actual TEE occurred. Addition of a temperature correction reduced the maximum error to 25%. If a balance between caloric supply and demand is intended, it is advised to adapt the caloric supply to the result of daily TEE measurement by indirect calorimetry.