Measured versus estimated creatinine clearance in patients with low serum creatinine values

To examine, in patients with a serum creatinine (SCr) < 85 mumol/L, whether the use of actual SCr versus the use of an SCr set at 85 mumol/L and applied to the Cockcroft-Gault method for calculation of creatinine clearance (Clcr), resulted in a similar ability to explain variance in prediction of Clcr.

Included in the study were patients with stable renal function receiving total parenteral nutrition, who underwent a 24-hour urine collection and had an SCr < 85 mumol/L. Clcr was calculated (CalcClcr) using the Cockcroft-Gault method, and actual SCr, and an SCr value set at 85 mumol/L (AdjCalcClcr). Calculated values were compared with 24-hour measured Clcr (MeasClcr).

Inpatient, acute-care hospital.

33 patients having 33 urine-collection periods. All patients were receiving total parenteral nutrition as their only form of nutritional supplement. Patients who had liver disease, trauma, or burns, or who were receiving certain pharmacologic agents, were excluded from the study.

MeasClcr based on the 24-hour urine collection, CalcClcr computed using the actual measured SCr, and SCr adjusted to 85 mumol/L.

A higher proportion of the variance of measured versus calculated Clcr was explained by the using the actual measured SCr value (r2 = 0.68) than SCr adjusted to 85 mumol/L (r2 = 0.32). When analyzed by gender, adjusting the SCr to 85 mumol/L was somewhat better at explaining the variance of measured versus calculated Clcr, but still explained less than 50 percent of the variance. However, the use of the actual SCr explained more of the variance of measured versus calculated Clcr than the use of the SCr value adjusted to 85 mumol/L.

In patients with an SCr < 85 mumol/L, the actual measured SCr should be used when calculating Clcr (regardless of the patient's gender), by the Cockcroft-Gault method.