Prevalence of bacterial pathogens in biofilms of drinking water distribution systems.

September SM, Els FA, Venter SN and Brozel VS. (2007) Journal of Water & Health, 5 (2); 219-227.

Water is normally disinfected prior to being distributed to the consumers and the microbial levels of the water are required to be within set limits when leaving the treatment plant. However, by the time the water reaches the consumers tap the quality may be very different from the quality at the time of treatment. This can often be due to recontamination after treatment as a result of regrowth of sub-lethally damaged bacteria or contamination from bacteria living in biofilms. This study was undertaken to determine the prevalence of Aeromonas , E. coli , Pseudomonas , Salmonella , Shigella and Vibrio in biofilms in drinking water distribution systems in large and small towns and home storage systems in South Africa .

Biofilm samples were collected from September 2001 to August 2002. There were 95 samples collected from two well-serviced urban areas (31 from Pretoria and 30 from Pietermaritzburg), a semi-urban developing community (Botshabelo five tap and 17 bucket samples) as well as other towns with small distribution networks (22 samples). The water from the two urban areas had been treated with chlorination and chloramination whereas the rest of the water was only chlorinated. Biofilm samples ( c . 1cm 2 ) were taken using a sterile swab from the inner surface of the service pipe and in the case of the buckets an equivalent sized area was swabbed. Water from the systems was tested for the presence of faecal coliforms. The heterotrophic culturable count was determined for both the water and biofilm phases of the samples. Biofilm samples were analysed to determine the number of potential pathogens ( Aeromonas , E. coli , Pseudomonas , Salmonella , Shigella and Vibrio ). These pathogens were quantified by the three-tube most probable number (MPN) method using enrichment broths and plating on selective agars.

The heterotrophic culturable counts in the water samples ranged from 1.0 x 10 -1 to 1.9 x 10 9 colony forming units (cfu) per ml and for the biofilms analysis of the same samples between 1.0 x 10 1 and more than 1.9 x 10 9 cfu cm -2 . Bacterial biofilms were present on the walls of all of the surfaces tested. Faecal coliforms were found in 7.7% of the tap samples and 23.5% of the buckets. These waters were not in compliance with the South African National Standard. Some sampling points had a variety of potential pathogens isolated at densities up to 10 4 cfu cm -2 . Pathogens isolated included Aeromonas , E. coli , Pseudomonas , Salmonella , Shigella and Vibrio. Samples from the larger distribution networks of Pretoria and Pietermaritzberg in general had much lower incidence of pathogens than those from the small towns and Botshabelo. The containers in Botshabelo in general harboured a higher number of pathogens than the tap except for Salmonella .

The partial sequences of the 16S rDNA genes of 74 randomly selected isolates were determined and revealed 5 Acinetobacter , 15 Aeromonas , 4 Enterobacter , 6 Klebsiella , 6 Pantoea and 34 Pseudomonas along with one or two representatives of various other genera. There were no putative Salmonella or Shigella confirmed however, indicating that none of these virulent pathogens could be detected in the drinking water-associated biofilms tested. The Pseudomonas isolates were all very similar to each other, and none of those isolated belonged to the nosocomial pathogens P. aeruginosa or P. mendocina . The selective culture media was found to be not as selective as reported when used for analysis of non-medical samples with a high incidence of false positives. Therefore, bacterial analyses of water based on selective isolation and culturing techniques should be interpreted cautiously. Water needs to be protected from the source to the tap using a comprehensive safety plan. Such a plan should address multi-barrier treatment and integrity of the water distribution systems to avoid pathogens entering the system.

Comment Exposure to the opportunistic pathogens detected in this study also occurs through other routes including food, and may be of public health significance.