Published online by Cambridge University Press: 01 November 1999
Enumeration of bacteria in raw milk is of public health and economic importance. Among the proposed rapid methods for assessment of bacterial contamination in raw milk, ATP bioluminescence has proved to be one of the most promising (Griffiths, 1991). Several companies produce ATP bioluminescence reagent kits and equipment for analysing raw milk samples for total bacterial count (Sutherland et al. 1994; Reybroeck & Schram, 1995). The principle of ATP bioluminescent bacterial assay is based on the following assumptions (Olsen, 1991). All living organisms contain ATP, ATP is neither associated with dead cells nor absorbed on to surfaces, colloids and so on, and there is a fairly constant ratio of ATP to biomass/number of cells for all microbial taxa independent of metabolic activity or environmental conditions. Of these assumptions, only the first seems to be indisputable. It is not the number of bacterial cells, but rather the colony forming unit (cfu) that is the denomination used when assessing the microbial quality of milk. For Gram-negative rods of the genus Enterobacteriaceae, a cfu is usually derived from a single cell. However, Gram-positive cocci (staphylococci and streptococci) grow in bunches and chains respectively (Gregg, 1991), and estimation of cell numbers may not give good agreement with the colony counts.
Several approaches have been investigated to increase the sensitivity of the bioluminescent method (Pahuski et al. 1991; Sutherland et al. 1994; Reybroeck & Schram, 1995; Froundjian et al. 1999). Although the detection limit achieved by these modifications (104 cfu/ml) may be sufficient for practical use (Bautista et al. 1992; Reybroeck & Schram 1995), the accuracy of the analysis was not significantly improved. The reported values for accuracy of the estimate for cfu/ml in raw milk (Syx) by the bioluminescent method were in the range 0·27–0·87 log units (Bautista et al. 1992; Reybroeck & Schram, 1995). The purpose of the present study was to determine the reasons for the lack of accuracy of the bacterial ATP assay in raw milk.