We studied 113 patients treated with intravenous amikacin to determine the value of determining serial trough and peak amikacin levels in plasma for predicting nephrotoxicity. Thirteen patients (11.5%) developed renal toxicity, with significant increases from 48 to 96 h in both peak and trough amikacin levels (6.7 ± 4.7 [standard deviation] days before the serum creatinine rose). The nontoxicity group had no change or even showed decrements iit amikacin levels in plasma. A higher nephrotoxicity risk was seen in patients with increments greater than 1 ,g/ml between 48 and 96 h, with odds ratios of 16.4 for trough, 8 for peak, and 7.2 for both levels. We suggest that an increment of at least 1 ,ug/ml in amikacin levels in plasma from 48 to 96 h may predict the appearance of renal toxicity.The aminoglycosides are commonly used in the treatment of severe gram-negative-bacillus infections (13). Unfortunately, nephrotoxicity is a common complication that requires discontinuation of therapy (14). Clinical studies have identified several risk factors, such as the underlying conditions of the patient or complications owing to the aminoglycoside itself (8,11,15,18,19).The aminoglycosides have a narrow therapeutic margin; patients given similar doses showed a wide disparity in levels in serum (12). The serum immunoassay has been accepted as a useful tool for enhancing drug efficacy, but its value in foretelling renal toxicity is controversial. This prospective study was designed to estimate the value of serial trough and peak levels in predicting renal toxicity in patients treated with intravenous amikacin.All hospitalized patients treated with intravenous amikacin for a minimum of 48 h, alone or in combination with other antibiotics, were included in this 3-month prospective study (April to June 1988). Patients with shock (systolic blood pressure less than 90 mm Hg and signs of hypoperfusion for more than 6 h); acute renal failure; hepatorenal syndrome (chronic liver disease plus oliguria, azotemia, and urinary sodium concentration lower than 10 meq/liter); end-stage renal disease under dialysis; and renal, liver, or bone marrow transplantation treated with cyclosporine were excluded from the study.Amikacin has been the only arninoglycoside used in our hospital since 1981 (17), and it is usually administered intravenously (5 mg/kg [body weight], three times per day) in a 30-min period; the dose was corrected according to the creatinine clearance (CLCR) calculated with the formula described by Cockcroft and Gault (male CLCR = 140 -age x weight/72 x serum creatinine; female CLCR = 0.9 x male CLCR) (5).The following data were obtained for each patient: age, sex, weight, underlying disease, control serum creatinine, * Corresponding author. trough and peak amikacin levels in plasma, calculated and given amnikacin in milligrams per day, and other antibiotics used. The serum creatinine was determined two times per week until cessation of therapy. The trough and peak amikacin levels in plasma were determined at 48 and 96 h after ...