Search

A recently marketed radioimmunoassay (RIA) for vancomycin (Monitor Science Corp.) was evaluated in vitro and in vivo. The RIA and bioassay results on 15 spiked serum samples found no significant difference between the two assays, although RIA results were significantly less variable. The coefficient of variation (RIA) was 5.5%. Vancomycin showed no significant degradation in plasma at 4°C and —20°C for up to 36 days. Five dialysis patients had a total of 62 serum levels determined. When vancomycin 1 g was given weekly to two anuric adults, marked accumulation occurred. This finding was consistent with vancomycin's long terminal half-life (19.8 and 17.8 days) obtained from a new computer subroutine which obtains least square estimators of model parameters from multiple dose data. Despite serum levels well above the MIC, two patients remained bacteremic. Peritoneal levels of vancomycin in two patients with suspected peritonitis equalled corresponding serum levels.

Previously published methods for determining an appropriate aminoglycoside dosage have been adapted for use in the handheld programmable calculator (Texas Instrument-59). The purpose of this study was to evaluate the accuracy of calculator-generated serum level predictions compared with measured values. Twenty-four adult patients receiving intravenous aminoglycosides (gentamicin, netilmicin, or tobramycin) had urine and serum collected for creatinine clearance calculations and had blood obtained for peak and trough aminoglycoside concentration determinations. All but one patient had normal renal function. A statistically significant relationship between predicted and measured creatinine clearance values was found. There was also generally good agreement between the calculator-predicted aminoglycoside levels and the measured levels, although some patients exhibited unexplained deviations. Error -in the estimation of creatinine clearance and adjustments for lean body weight did not improve prediction accuracy. We conclude that the hand-held calculator was able to predict measured aminoglycoside levels for most patients with acceptable precision. However, the calculator method offers no real advantage over other methods (except speed in performing the calculation), and may give the user a false sense of security.

We conducted a retrospective chart review to estimate the potential for drug interactions in subjects infected with the human immunodeficiency virus‐1 when a protease inhibitor was added to existing therapy. Medical records of 50 patients in each of three immunologic strata (CD4 cell counts/μl < 100, 100–199, 200–500) were randomly selected from records of all patients receiving care at the clinic; 114 records were evaluable. The probabilities of one interaction or more were 31%, 42%, and 77% of patients if treated with indinavir, saquinavir, and ritonavir, respectively, across all CD4 groups; when the CD4 count was below 100 cells/μl, the probabilities were 55%, 63%, and 93%. Many of these interactions, however, resulted from administration of rifabutin, a drug likely to decrease in importance as less toxic alternatives become more widely administered. The potential for drug interactions is high when starting protease inhibitor therapy, especially in patients with low CD4 cell counts who receive ritonavir. Concurrent therapy must be evaluated before treatment, as many agents are either contraindicated or require dosage modification.

Cefamandole nafate (CM) and cephalothin sodium (CP) were administered as prophylaxis in a randomized, prospective study to 30 consecutive patients undergoing prosthetic cardiac valve insertion. A single dose of 20 mg/kg was given intramuscularly during anesthesia induction, and serial plasma antibiotic concentrations, atrial muscle and cardiac valve tissue antibiotic levels, plasma bactericidal activity against pathogenic staphylococci, and infectious complications were determined and compared for the two drugs. Both antibiotics produced high plasma levels (>20 μg/ml 30 min after injection) which fell less than 25% during the period of cardiopulmonary bypass. However, CM levels were significantly higher at most time periods (P<0.05) than CP levels. CP levels were undetectable in atrial muscle from 14 of 15 patients and in valves from 10 of 15 patients. In contrast, CM bioactivity was found in all tissues. Differences in tissue antibiotic concentration could not be accounted for by differences in plasma concentrations or by CP tissue binding and were assumed to be caused by differences in penetration. Plasma bactericidal activity against staphylococci was equal for the two drugs (median titer, 1:16). No infections were seen in either group. CM appeared to be an effective and perhaps preferable prophylactic antibiotic for use during cardiac surgery.