Comparison of two methods for detecting microbial contamination in intravenous fluids

Abstract

An aliquot sampling method and a qualitative membrane filltration method of microbial detection in intravenous infusion solutions were compared. One-liter plastic bags of total parenteral nutrition solutions were contaminated at approximately 1000, 100, 10, and 1 organisms per liter using clinically isolated strains of Escheichia coli, Pseudomonas aeruginosa, Staphylococcus epidermidis, and Candida albicans. Ten-milliliter aliquots of the solutions were injected into blood culture bottles to test the aliquot method. All remaining fluid was filtered through the Addi-Chek system to test the filtration method. Samples were incubated at 30 degrees C for 10 days and inspected daily for turbidity. The aliquot sampling method consistently detected each of the four organisms at levels of 100 organisms/liter and above. The membrane filtration method consistently detected all levels of contamination (as few as four organisms/liter). The aliquot sampling method is inexpensive and easy to use but may fail to detect some contaminated solutions. The qualitative membrane filtration system will detect all levels of contamination but is more costly in both time and money, and its reliability has not been assessed objectively.