Aminoglycosides constitute one of the oldest classes of antimicrobials. Despite their toxicity, mainly nephrotoxicity and ototoxicity, aminoglycosides are valuable in current clinical practice, since they retain good activity against multidrug-resistant Gram-negative pathogens, such as Pseudomonas aeruginosa and Acinetobacter spp. Time-kill studies have shown a concentration-dependent and partially concentration-dependent bacterial killing against Gram-negative and Gram-positive bacteria, respectively. Pharmacodynamic data gathered over recent decades show that the administration of aminoglycosides by an extended-interval dosing scheme takes advantage of the maximum potential of these agents, with the goal of achieving an area under the concentration-time curve (AUC) of 100 mg · h/L over 24 hours and a peak plasma drug concentration (C(max)) to minimum inhibitory concentration (MIC) ratio of 8-10. Several clinical conditions that are common in seriously ill patients result in expansion of the extracellular space and can lead to a lower than desirable C(max) with the usual loading dose. Extended-interval dosing schemes allow adequate time to decrease bacterial adaptive resistance, a phenomenon characterized by slow concentration-independent killing. Adaptive resistance is minimized by the complete clearance of the drug before the subsequent dose, thus favouring the extended-interval dosing schemes. The efficacy of these schemes is also safeguarded by the observed post-antibiotic sub-MIC effect and post-antibiotic leukocyte enhancement, which inhibit bacterial regrowth when the serum aminoglycoside levels fall below the MIC of the pathogen. In everyday clinical practice, aminoglycosides are usually used empirically to treat severe sepsis and septic shock while awaiting the results of antimicrobial susceptibility testing. The European Committee on Antimicrobial Susceptibility Testing acknowledges the regimen-dependent nature of clinical breakpoints for aminoglycosides, i.e. of MIC values that classify bacterial isolates into sensitive or resistant, and bases its recommendations on extended-interval dosing. To a large extent, the lack of correlation between in vitro antimicrobial susceptibility testing and clinical outcome is derived from the fact that the available clinical breakpoints for aminoglycosides are set based on mean pharmacokinetic parameters obtained in healthy volunteers and not sick patients. The nephrotoxicity associated with once- versus multiple-daily administration of aminoglycosides has been assessed in numerous prospective randomized trials and by several meta-analyses. The once-daily dosing schedule provides a longer time of administration until the threshold for nephrotoxicity is met. Regarding ototoxicity, no dosing regimen appears to be less ototoxic than another. Inactivation of aminoglycosides inside the bacterial pathogens occurs by diverse modifying enzymes and by operation of multidrug efflux systems, making both of these potential targets for inhibition. In summary, despite their use for several decades, the ideal method of administration and the preferred dosing schemes of aminoglycosides for most of their therapeutic indications need further refinement. Individualized pharmacodynamic monitoring has the potential of minimizing the toxicity and the clinical failures of these agents in critically ill patients.