Aminoglycoside clearance depends on kidney function, but the Australian Therapeutic Guidelines for antibiotics (version 14, 2010) recommend initial dosing based on weight without consideration of kidney function. Other guidelines that modify dosing based on kidney function estimates often use the Cockroft-Gault equation, but the role of the estimated glomerular filtration rate equations for this purpose is unclear. To determine the performance of current guideline dosing in achieving target area-under-the-curve and examine the relative precision of the estimated glomerular filtration rate equations compared with traditional Cockroft-Gault creatinine clearance in predicting aminoglycoside clearance. We analysed 496 aminoglycoside treatment episodes involving 1377 infusions in adult patients. Conformity with antibiotic guideline dosing was achieved if the discrepancy between prescribed and recommended dose was less than 15%. Aminoglycoside clearance was determined from linear regression using a one compartment model with the Aminoglycoside Levels and Daily Dose Indicator programme. We assessed the precision of the Cockroft-Gault, Modification of Diet in renal Disease Study and Chronic Kidney Disease-Epidemiology Collaboration (CKD-EPI) equations in predicting aminoglycoside clearance by correlation and linear regression. Conformity with guideline dosing was not associated with achieving target area-under-the-curve. The CKD-EPI estimated glomerular filtration rate adjusted for body surface area showed the highest correlation (gentamicin, r = 0.66; tobramycin, r = 0.82) and best predictive model for aminoglycoside clearance. Current guideline dosing may be suboptimal for achieving target area-under-the-curve. The CKD-EPI equation adjusted for patient body surface area best predicts aminoglycoside clearance, and could be evaluated as a covariate in determining initial aminoglycoside dosing.