Shelf-life of 151 pasteurized milk samples was recorded and correlation coefficients calculated using various microbiological factors: standard plate count (SPC), psychrotrophic plate count (PPC), coliform count (CC), and the impedance response detection time (DT) with incubation at both 21 and 32°C. These data were obtained for milk samples on the day of pasteurization as well as 4 and 8 d thereafter. Various treatments (media, dilution factors, temperature and sample volume) were compared. Of the SPC, PPC, CC and DT taken on the day of pasteurization, only the DT achieved a significant correlation with shelf-life. A correlation coefficient of 0.55 was obtained for one treatment applied to 61 samples and correlation coefficients of 0.28 to 0.32 were obtained for several other treatments applied to the entire 151 samples. Values as large as these could occur by chance in uncorrelated data with p<0.0005. Thus, of the total 61 samples, 80% were correctly classified by the impedance detection time test. It is concluded that for prediction of shelf-life on the day of pasteurization, the impedance method is superior to the SPC and the PPC. In addition, the impedance method is more rapid, i.e., 14 h vs.2 d for the SPC and 10 d for the PPC.
The electrical impedance of media is altered with chemical changes brought about by microbial metabolism and growth. Time required to bring about readily detectable change (detection time — DT) is a function of the initial levels of microorganisms in the sample. DTs were compared to Standard Plate Counts for 407 milk samples — homogenized, low fat, skim and raw. Using the criterion that a sample of pasteurized milk with a DT of 7 h or less was indicative of a plate count of 10,000/ml or greater, 323 of 380 samples were correctly classified. For raw milk, the DT was 10 h to resolve samples into greater or less than 10,000 organisms per ml. Results of a preliminary study on estimation of psychrotrophs in pasteurized milk showed that impedance monitoring at 21 C provided a 22-h screen correctly classifying 88% of the samples into categories of more than or less than 1,000 organisms per ml. Better agreement (91%) in a shorter time (13.7 h) was obtained with a screen for 10,000 organisms. Finally, for the first 22 samples analyzed, keeping quality data on pasteurized milk have correlated better with post-pasteurization impedance measurements than with either post-pasteurization total counts or psychrotrophic counts.