The correlation coefficient between the rapid enzymatic and the ovemight microbiological assays for 211 urine and serum specimens was 0.96. The 95% confidence limits yielded a correlation coefficient between 0.92 and 0.98. Both methods tended to underestimate the amount of a gentamicin added to urine. When only serum samples were considered, the predicted value obtained from the linear regression analysis of either method was within 0.57 ,ug/ml 99% of the time. This high degree of positive correlation will permit safe rapid adjustment of individualized patient gentamicin dosages.Since gentamicin is excreted primarily through the kidneys, hazardous concentrations can accumulate in patients with renal failure (2). Therefore, there are many indications for the determination of blood gentamicin concentration. Rapid assays have been developed to allow close monitoring of individual patients (1,3,4,10). These include spore-inoculated agar plates (7), radioimmunoassay (1, 3), enzymatic (8), and aminoglycoside inhibition of the metabolic activity of microorganisms (4). The reference assay systems have been the ovemight Staphylococcus aureus and the Bacillus subtilis spore methods. Before one of the rapid assays can be routinely used for patient management, its accuracy and precision relative to the reference methods must be known. This report details the comparison of the overnight assay technique (5) (S. aureus) with the enzymatic method (8).
MATERIALS AND METHODSSpecimen. Sixty-nine samples were mock unknowns constructed by the addition of a known quantity of gentamicin to serum from hospitalized patients not receiving gentamicin. Fifty-two of these samples were derived from patients not receiving any antimicrobial agents; of these, 23 were from patients with abnormal renal function (blood urea nitrogen greater than 50 mg%) or abnormal hepatic function (bilirubin greater than 3 mg% and serum glutamic oxalacetic transaminase greater than 100 IU), or both. The remaining 17 samples contained chloramphenicol, ampicillin, nafcillin, oxacillin, or neomycin alone or in combination. One hundred and twenty-three serum specimens were obtained from patients receiving parenteral gentamicin and usually a semisynthetic (type I penicillinase-resistant) penicillin. Nineteen urine samples were constructed by the addition of known amounts of gentamicin to the urine of healthy adult male volunteers. All specimens were frozen at -20 C after collection until assay.Microbiological assay, Gentamicin was assayed in triplicate by the method previously described (5). Sterile 6.35-mm paper disks were dipped in the specimen, shaken to remove the excess, and placed on agar plates previously seeded with S. aureus negative forms (ATCC 6538 P). After overnight incubation (usually 18 h) at 37 C, the zone diameter of the samples were compared with that of standards prepared in pooled normal human serum using gentamicin of known potency. All samples were preincubated with type [I ,-lactamase (Whatman Associates).Enzymatic assay. The enzymatic assay was...