Comparison of plate counts, Petrifilm, dipslides, and adenosine triphosphate bioluminescence for monitoring bacteria in cooling-tower waters.

Effective bacterial control in cooling-tower systems requires accurate and timely methods to count bacteria. Plate-count methods are difficult to implement on-site, because they are time- and labor-intensive and require sterile techniques. Several field-applicable methods (dipslides, Petrifilm, and adenosine triphosphate [ATP] bioluminescence) were compared with the plate count for two sample matrices--phosphate-buffered saline solution containing a pure culture of Pseudomonas fluorescens and cooling-tower water containing an undefined mixed bacterial culture. For the pure culture, (1) counts determined on nutrient agar and plate-count agar (PCA) media and expressed as colony-forming units (CFU) per milliliter were equivalent to those on R2A medium (p = 1.0 and p = 1.0, respectively); (2) Petrifilm counts were not significantly different from R2A plate counts (p = 0.99); (3) the dipslide counts were up to 2 log units higher than R2A plate counts, but this discrepancy was not statistically significant (p = 0.06); and (4) a discernable correlation (r2 = 0.67) existed between ATP readings and plate counts. For cooling-tower water samples (n = 62), (1) bacterial counts using R2A medium were higher (but not significant; p = 0.63) than nutrient agar and significantly higher than tryptone-glucose yeast extract (TGE; p = 0.03) and PCA (p < 0.001); (2) Petrifilm counts were significantly lower than nutrient agar or R2A (p = 0.02 and p < 0.001, respectively), but not statistically different from TGE, PCA, and dipslides (p = 0.55, p = 0.69, and p = 0.91, respectively); (3) the dipslide method yielded bacteria counts 1 to 3 log units lower than nutrient agar and R2A (p < 0.001), but was not significantly different from Petrifilm (p = 0.91), PCA (p = 1.00) or TGE (p = 0.07); (4) the differences between dipslides and the other methods became greater with a 6-day incubation time; and (5) the correlation between ATP readings and plate counts varied from system to system, was poor (r2 values ranged from < 0.01 to 0.47), and the ATP method was not sufficiently sensitive to measure counts below approximately 10(4) CFU/mL.