A commercially available radioimmunoassay kit was modified to enable us to measure, in triplicate, the amikacin concentration in 1 ,ld of perilymph fluid. Amikacin levels in plasma and perilymph were measured in guinea pigs after continuous intravenous infusion at four different dosing rates. After a 4-h infusion, a good linear correlation was found between the amikacin concentration in plasma and the dosing rate. Likewise, a significant linear relationship was found between concentrations of amikacin In perilymph and plasma (y = 0.21x + 2.56; r = 0.67; n = 45) after 6 h of infusion. These results suggest nonsaturation kinetics at the concentrations used.The specific toxic effect of aminoglycoside antibiotics (AGs) on the inner ear is related to the fact that they penetrate the inner ear fluid compartments and cause hair cell damage (5,8). Although the rate of entry of AGs into the perilymph space is relatively slow (1, 2, 6, 11), drug levels are prolonged owing to their slow elimination from the inner ear. Different authors have reported widely variable rates of elimination (3 to 15 h) for different AGs (10). As discussed by Manuel et al. (6), intervals between AG doses of less than four half-lives will lead to a notable accumulation in the perilymnph until an equilibrium condition is established. Some authors have suggested the existence of a toxic threshold in the inner ear, meaning that beyond a certain dose the kinetics of AGs are modified and that the concentration increases suddenly in the perilymph (9). Others found a good linear correlation between the dose and perilymph concentrations (3, 4, 6, 10). Pharmacokinetic differences among AGs, as well as differences in the animal species investigated, dose, route of drug administration, time interval between doses, time of sampling, and the analytical technique used, may explain the conflicting conclusions.Amikacin is a semisynthetic derivative of kanamycin and is often used clinically for infections that are resistant to gentamicin or tobramycin. tion of the percentage of relative binding versus the log of the standard concentrations. Logit, probit, and arcsin transformations were tried, and the latter was found to give the best linearity. The rest of the procedure was done according to the protocol provided with the kit. The reproducibility of the assay was assessed by determining the intraassay and interassay coefficients of variation for triplicate determination. Intraassay coefficients of variation ranged from 7.1% for the lowest standard concentrations (5 ng/ml) (n = 10) to 1.8% for the highest standard concentration (320 ng/ml) (n = 10). Two piasma samples with amikacin concentrations of 26.5 and 111.0 ng/ml were measured in triplicate in 12 different assays. Coefficients of variation between assays were found to be 7.2 and 3.7%, respectively.The recovery of drug in 1-plA samples was tested for known concentrations of 4, 9.6, 12.5, and 32 pig/ml with dilution factors Of 1/90, 1/120, 1/250, and 1/500, respectively. The technique described below for per...