Creatinine clearance predictions in acutely ill patients.

In acutely ill patients, predicted creatinine clearance (Clcr) values [obtained using the Siersbaek-Nielsen nomogram (SNN) and ideal body weight (IBW)] were compared with actual Clcr based on measured urine creatinine concentrations. Timed urine collections were obtained from 118 patients (423 collections from 68 men, 400 collections from 50 women) before, during, and after aminoglycoside therapy. Patients were in intensive care units for management of acute exacerbations of sepsis, pneumonia, or abscess; most had other complicating conditions and were chronically ill. Urine was collected for 8 to 24 hours (70% for 24 hours) from Foley-catheterized patients. Patients were divided into low (less than 15 mg/kg/day), normal (15-25 mg/kg/day), and elevated (greater than 25 mg/kg/day) urinary creatinine excretion groups. Actual body weight (ABW) was used in some patient subgroups to explore differences between using IBW and ABW. SNN most accurately predicted Clcr in the 20% of the urine collections characterized by normal urinary creatinine excretion (Ucr). Most study patients excreted significantly less creatinine than the age-matched population tested in developing the SNN. In acutely ill patients with low Ucr, SNN overpredicted Clcr by 10-20 ml/min. SNN also overpredicted Clcr values in obese patients, but use of IBW rather than ABW improved the correlation between measured and predicted values in this subgroup. The SNN nomogram is applicable to critically ill patients if adjustments are made in predicted values. The characteristics of patients with low Ucr and the mechanism responsible for the decreased Ucr deserve further study.