ABSTRACT
Consumers increasingly use bottled water and home water treatment systems to avoid direct tap water. According to the International Bottled Water Association (IBWA), an industry trade group, 5 billion gallons of bottled water were consumed by North Americans in 2001. The principal aim of this study was to assess the microbial quality of in-house and imported bottled water for human consumption, by measurement and comparison of the concentration of bacterial endotoxin and standard cultivable methods of indicator microorganisms, specifically, heterotrophic and fecal coliform plate counts. A total of 21 brands of commercial bottled water, consisting of 10 imported and 11 in-house brands, selected at random from 96 brands that are consumed in Puerto Rico, were tested at three different time intervals. The Standard Limulus Amebocyte Lysate test, gel clot method, was used to measure the endotoxin concentrations. The minimum endotoxin concentration in 63 water samples was less than 0.0625 EU/mL, while the maximum was 32 EU/mL. The minimum bacterial count showed no growth, while the maximum was 7,500 CFU/mL. Bacterial isolates like P. fluorescens, Corynebacterium sp. J-K, S. paucimobilis, P. versicularis, A. baumannii, P. chlororaphis, F. indologenes, A. faecalis and P. cepacia were identified. Repeated measures analysis of variance demonstrated that endotoxin concentration did not change over time, while there was a statistically significant (p < 0.05) decrease in bacterial count over time. In addition, multiple linear regression analysis demonstrated that a unit change in the concentration of endotoxin across time was associated with a significant (p < 0.05) reduction in the bacteriological cell count. This analysis evidenced a significant time effect in the average log bacteriological cell count. Although bacterial growth was not detected in some water samples, endotoxin was present. Measurement of Gram-negative bacterial endotoxins is one of the methods that have...