POOR CORRELATION BETWEEN PUBLISHED METHODS TO PREDICT CREATININE CLEARANCE AND MEASURED CREATININE CLEARANCE IN ASYMPTOMATIC HIV-INFECTED INDIVIDUALS

The purpose of this study was to evaluate the predictive ability of si x published creatinine clearance (C-Cr) equations in healthy human imm unodeficiency virus (HIV) infected individuals. A 24-h urine collectio n to determine C-cr was done on an out-patient basis in 18 subjects. P redicted C-Cr was compared with the measured values, and the predictiv e performance was assessed with percentage mean error (bias) and perce ntage root mean error (precision). Mean +/- standard deviation measure d C-Cr was 107 +/- 35 mL/min/l7.3 M-2. C-Cr determined using each of t he published equations correlated poorly with measured values. C-Cr de termined using Hull methods was significantly different from the measu red values. Though Cockcroft and Gault and Jelliffe methods had the lo west mean bias and greater precision, a significant range of differenc e from measured C-Cr was observed (12 to + 28%). All methods over esti mated the measured C-Cr in HIV-infected individuals. Until other appro aches are developed, a 21 h urine collection may be the best approach for assessing renal function in HIV-infected individuals, especially i n those receiving medications with nari ow therapeutic indices that ar e cleared by the kidney.