Assessment of renal function and relating this parameter to amino-glycoside clearance is important for an appropriate individualization of dosage regimens in patients with impaired renal function. However, it has been suggested that in cystic fibrosis (CF), creatinine clearance (CrCl) is not a good predictor of tobramycin clearance because of a lack of correlation. The aim of this study was to investigate the correlation between the tobramycin clearance and the measured CrCl. In addition, because most pharmacokinetic computer models use an a priori relationship between CrCl and tobramycin elimination rate constant [k(el)], regression analysis of k(el) on CrCl was performed. Eighteen CF patients (12 men, 6 women, ages 21-55 years) were treated with intravenous tobramycin. Blood and urine samples were collected for tobramycin analysis and determination of the CrCl. For each patient, CrCl was also estimated using the formulas of Cockcroft and Gault (C/G), Jelliffe I (J I), and Jelliffe II (J II). Predictive performance of these formulas was evaluated using mean error and mean squared error as reflections of bias and precision. Tobramycin total body clearance as well as renal clearance correlated significantly with CrCl (r = 0.52; p = 0.02 and r = 0.78; p = 0.02, respectively). Regression of the tobramycin k(el) versus CrCl gave the following equation: k(el) = 0.135 + 0.00134*CrCl/1.73 m2 (r = 0.64; p = 0.004). The 95% confidence interval of intercept and slope of the regression line were 0.019-0.251 and 0.00049-0.00219, respectively. The formulas of C/G, J I, and J II all overpredicted CrCl. Bias was 19, 24, and 8 ml/min, and precision was 37, 42, and 33 ml/min, respectively, for the C/G, J I, and J II formulas. In our CF population the J II formula gave the best estimation of the CrCl but calculated estimates deviated -25(-)+62% from measured values.