We studied 201 critically ill patients during 267 courses of gentamicin (139 courses) or tobramycin (128 courses) therapy. Clinical and pharmacokinetic data were obtained on 240 of 267 courses (120 courses each of gentamicin and tobramycin). Two judgments of nephrotoxicity and its cause were made independently in this study, using a clinical and a pharmacokinetic definition of nephrotoxicity. The two sets of criteria were generally in good agreement, as all but 10 of 41 patients who were judged nephrotoxic by pharmacokinetic criteria were independently judged nephrotoxic by the clinical definition. Groups of patients judged nontoxic did not differ from groups judged nephrotoxic in age, sex, weight, initial creatinine clearance, total dose given, duration of treatment, initial aminoglycoside trough serum levels, number of dosage adjustments, concurrent use of furosemide, or concurrent cephalosporins. Prior aminoglycosides (usually gentamicin) had been used more frequently in the nontoxic group (P < 0.05). Two major conclusions of this study are at variance with those of previous investigators: (i) we found no clinical parameters of value in predicting nephrotoxicity in critically ill patients; and (ii) aminoglycoside serum concentrations, once in the therapeutic range, were of limited value in prevention of aminoglycoside nephrotoxicity in our patients.We and others have previously shown that the various aminoglycosides differ in their intrinsic nephrotoxic potential in animals (1, 9, 11, 13) and humans (10,17,18,19). In such studies, both nontoxic and nephrotoxic patients must have precise control of the serum concentrations before the population can be analyzed for the factors predisposing to nephrotoxicity. We have treated 201 critically ill patients with suspected or proven infections, with aminoglycosides administered in regimens that were closely monitored by measured blood levels (17). Dosing adjustments were made as required to maintain the recommended concentration range. The judgment of nephrotoxicity was made independently by both clinical and pharmacokinetic criteria. Gentamicin was more frequently (24%) nephrotoxic than was tobramycin (12%). In this analysis, we compared all nontoxic patients with all nephrotoxic patients so as to use our large groups of nephrotoxic patients to define the characteristics of those who experienced renal damage when appropriately dosed with aminoglycosides.
METHODSPatients. We gave 201 acutely ill patients 267 courses of aminoglycoside therapy. Pharmacokinetic and clinical data were available on 240 courses, which were used for analysis. The patients were older adults in critical care units, with serious infections complicating either major medical or surgical disorders. Most infections were intra-abdominal or pneumonic. Patients were entered into this study at the time of request for dosing and pharmacokinetic monitoring.Dosing and sample collection. Aminoglycosides were administered by intravenous infusion over 30 to 60 min. Dosing rates were determined initially by n...