An evaluation of direct microscopical counts and endotoxin measurements as alternatives for total plate counts

Abstract

Acridine orange direct counts (AODCs) were compared with endotoxin (LPS) concentrations in 233 unchlorinated drinking water samples. Different fractions of LPS (total, free, and bound) were obtained by centrifuging the samples and determined using a chromogenic Limulus amoebocyte lysate (LAL) test kit. However correlations with AODC were not increased when the bound LPS fraction was used instead of total LPS. For the total number of samples the correlation between total LPS and AODC was relatively high ( r = 0.80), but with a residual standard deviation indicating considerable variations in the mean LPS content per cell from sample to sample. The obtained best-fitting regression line was significantly dependent on both water type and water processing plant; likewise, different borings within the same water supply seemed to influence the LPS/AODC ratio. It is concluded that the AODC method is more reliable than LPS measurements when determining the total number of viable bacteria present, and seems to represent a good alternative to the traditional plate count method. But based on an apparent relation between the LPS concentration and the judged mean cell size, the value of the LAL test in estimating biomass is indicated, and with an extensive and fairly homogenous sample material in, for instance, acute contamination situations, the ease and speed of the LAL test would favour this method