A new assay was developed to determine the concentration of amphotericin B in body fluids. The Bactec radiometric system was used to measure CO2 production by a test strain, Candida albicans MDA 448, in the presence of amphotericin B. After 5-h incubation, drug concentrations as low as 0.2 ,ug/ml could be detected. The results are comparable to those of the commonly used agar diffusion assay with Paecilomyces varioti. Asay bottles. Bactec 6A culture bottles (Johnston Laboratories, Cockeysville, Md.) containing TSB were used in the study, since TSB does not affect the assay for amphotericin B (6). However, only 10 ml of the 6A medium containing a total of 0.5 pLCi of 14C substrate was used in each bottle; the remaining 20 ml was dispensed in 10-ml aliquots into empty sterile bottles provided by Johnston Laboratories. Each 6A bottle was iWected with either 1 ml of pooled serum containing amphotericin B or 1 ml of pooled serum without added amphotericin B. The assay was also performed with spinal fluid instead of serum.Inoculum and incubation. Each 6A bottle containing 10 ml of TSB and 1 ml of serum was inoculated with 1.0 ml of 1% cell suspension. Inoculated bottles were incubated on a rotary shaker (150 rpm) at 3700 for 5 h.Radiometric determination of C02 production.After a 5-h incubation, production of radioactive C02 was measured with a Bactec 301 radiometric detection system with an expanded growth index (GI) scale. In this system, a GI reading of 100 is equivalent to the production of 0.025 ,uCi of 14CO2. Calculation of amphotericin B concentration.The GI for each drug concentration was recorded and compared to the GI obtained with the drug-free control. Percent changes in C02 (AC02) were calculated by the formula: AC02 (%) = GI control -GI test/GI control x 100. The resulting values were used to construct a standard curve (see Fig. 2