Defining an optimal time of analysis for detecting bacterial contaminants in intravenous admixtures was studied. Three different intravenous solutions were inoculated with low-level numbers (10(1)) of Staphylococcus epidermidis and tested for sterility using the Ivex-2 Filterset method at six time intervals after inoculation: less than 1, 20, 40, 60, 120, and 240 minutes. Solutions used were 5% dextrose injection 1000 ml, 0.9% sodium chloride injection 1000 ml, and 5% dextrose injection 50 ml. At each interval, 10 solutions of each type were tested. An additional 20 controls were employed to monitor technique, facilities, and environmental conditions. Successful recovery of Staph. epidermis decreased significantly when sample processing was delayed for longer than 40-60 minutes after inoculation. Furthermore, the number of false negatives was greater for dextrose solutions than for sodium chloride solutions after this time period. Volume of admixtures had no effect on contamination detection. This study suggests that sterility testing should be completed within 40-60 minutes after preparation of intravenous admixtures.