Methods

Our anesthesia technologists spiked IV bags in the anesthesia workroom using our standard protocol. The technician removed each IV administration set (Codan US, Santa Ana, CA) from its sterile packaging using clean gloves, without masks or gowns. The plastic outer covering of the IV bags (Hospira, Lake Forest, IL) was removed, and the bags were hung on IV poles. Each IV bag administration port cover was removed, and the administration set spike was inserted into the IV bag in the usual manner. The end cap of the IV administration set was removed, and the roller clamp was opened. When fluid was seen emerging from the distal tip of the IV set, after all air bubbles had been removed, the fluid was stopped with the roller clamp and the end cap of the IV tubing was replaced. To avoid affecting standard procedure, the technologists were not informed about the study prior to bag spiking.

In the first part of the study, 25 one-liter bags of normal saline (NS) were spiked and placed in storage at room temperature in the anesthesia workroom. At 1 hour, 24 hours, 48 hours, 5 days, and 9 days, 5 bags were randomly selected and sampled for microbial contamination. At each time interval, 20-mL samples were collected after thorough mixing by repeatedly squeezing each bag immediately before sampling. All samples were taken from the rubber stopper on the IV bag after it was cleaned with alcohol. The stopper was allowed to dry completely before a sample was taken. Each 20-mL sample was divided equally and placed into 2 BACTEC Plus Aerobic/F blood culture vials (Becton Dickinson, Sparks, MD). The sample vials were transported to the clinical microbiology laboratory and incubated in the BACTEC FX blood culture system (Becton Dickinson). The vials were monitored for 5 days for growth of bacteria and fungi. Each culture vial contained a fluorescent chemical sensor to detect increases in carbon dioxide produced by the growth of microorganisms, including bacteria and fungi. Each vial was monitored every 10 minutes for an increase in fluorescence. A positive reading would indicate the presence of microorganisms. To confirm microbial growth in positive vials detected by the BACTEC system, smears were obtained and Gram stained. A subculture on media for isolation (ie, blood agar and chocolate agar) was used to identify any microbes detected by the BACTEC system. In the second part of the study, we added an additional 3 groups. Group 2 consisted of 5 bags of 5% dextrose in lactated Ringer’s (D5LR), a chemically complex solution more vulnerable to microbial contamination and sustained growth, for comparison to NS group 1. We also wanted to have positive controls to verify that both NS and D5LR solutions could support readily detectable bacterial growth following low-level contamination (0.1 CFU/mL). Group 3 consisted of 5 bags of NS contaminated with E. coli, and group 4 consisted of 5 bags of D5LR contaminated with E. coli. Both groups 3 and 4 were inoculated with E. coli at time zero. Escherichia coli was selected as the positive control based on microbiology recommendation. Sampling protocol for part 2 was identical to the one used in part 1 described above. The contaminated bags were stored at room temperature in a private locked office in the hospital to eliminate any chance of cross-contamination.