Plumbing Africa September 2017 | Page 23

Health and sanitation << Continued from page 19 • POU filtration devices have been certified by NSF International for removal of protozoa, bacteria, and viruses in general, using surrogate microorganisms as challenge organisms during testing and evaluation. Characterisation of effectiveness against legionella Several case studies describe the effectiveness of POU membrane filtration devices for removal of legionella. • Casini et al. (2014) reported the efficacy of POU filtration installed in selected wards of an Italian hospital to further reduce legionella growth within the building hot water system after chloride dioxide disinfection. POU filters used in this study had a 0.2-µm nominal pore size and 30-day replacement rate. This integrated disinfection- filtration strategy, although expensive, significantly reduced legionella counts to less than 10 3 CFU/L and achieved a positive sample rate of less than 30%. • Baron et al. (2014b) evaluated a new faucet filter at five sinks in a cancer centre and found that legionellae were removed from all filtered samples for 12 weeks, exceeding the manufacturer's recommended maximum duration of use of 62 days. The filters contain a 30-µm pre-filtration layer, a 1-µm membrane, and a 0.2-µm membrane. • Marchesi et al. (2011) performed a 10-year review of multiple treatment methods to control for legionella at a hospital in Italy, including POU filtration, though information on the characteristics of the filters was not supplied. Filters were placed in high-risk units of the hospital only, where high levels of legionella contamination were identified, and were replaced every 30 days. No legionellae were detected at taps containing POU filters. • Daeschlein et al. (2007) evaluated a reusable POU filter for removing waterborne pathogens, including L. pneumophila, in a hospital’s transplant unit for eight weeks. Filters had three configurations: (1) hollow fibre of polyethersulfone with pore size 0.2µm and surface area of 800 cm 2 ; (2) hollow fibre of polyethersulfone with pore size 0.2µm, surface area of 1100 cm 2 , and inner encasement coated with nanosilver; and (3) same as (2) with metallic silver outlet. Filters were placed on 18 taps (12 taps, six showers) in the hospital’s transplant unit and each filter was monitored for pathogens at one, four and eight weeks, reprocessed and reused in three additional trials. Over the test period, no legionella or other pathogens were detected in any filter effluent. Because bacterial counts in filtered water exceeded the limit of >100 CFU/mL eight times, the following criteria were developed to prevent carry- over contamination from reuse of the filters: filters were cleaned with a strong chemical followed by flushing and thermal disinfection in a quality control-compliant washer- disinfector once a week, in addition to alcohol disinfection of the filter encasement. With this reprocessing, the authors determined that filters should be changed after four weeks in high-risk areas and after eight weeks in moderate-risk areas. www.plumbingafrica.co.za • • 21 A newer versi on of the filters described in the example by Daeschlein et al. (2007) was evaluated by Vonberg et al. (2008) at a hospital in Germany. The new version had a membrane surface coated with nanosilver. Fifteen taps in a thoracic surgery department were selected and sampled before adding filters. Filters were placed on those taps and sampled after one, two, three and four weeks of usage. Samples were analysed for the pathogens legionella and pseudomonas, in addition to the indicators enterococci and heterotrophic bacteria. Legionellae were detected in nearly half (48.3%) of taps before filters were added and only one sample (week 1) after filters were added (L. pneumophila serogroup 1, 4 CFU/mL); no pseudomonas were detected. The authors did not attempt to reprocess the filters as in the Daeschlein study and did see heterotrophs increase to >100 CFU/mL in some filters after one week of use. The authors concluded that incorporation of nanosilver in the filter’s membrane surface coating may prevent biofilm growth in this POU device and that use of these POU filters with weekly replacement in high-risk patient wards may be effective at preventing nosocomial legionellosis. Sheffer et al. (2005) evaluated POU filtration devices containing positively charged nylon membranes with a 0.2-µm nominal pore size. Filters were placed on four taps in the administration building at a hospital and monitored for legionella, heterotrophic bacteria, and mycobacteria, along with three taps without filters, every 2–3 days for 13 days, before and after a one- minute flush. Samples from taps with filters before flush were negative for legionella during the 13-day period, while mean concentration in taps without filters was 104.5 CFU/ mL. Mycobacterium gordonae was isolated from 10.3% of taps without filters before flushing, but no mycobacteria were isolated from taps with filters before flushing. Heterotrophs were significantly reduced at taps with filters. One post-flush sample from a tap with a filter was positive for legionella on day 10, with a concentration of 5 CFU/mL. No post-flush samples from taps with or without filters were positive for mycobacteria. The authors concluded that the POU filters used in this study effectively controlled legionella and mycobacteria through seven days of use. Molloy et al. (2008) evaluated three types of POU solid block activated carbon filters for removal of L. pneumophila in a laboratory-simulated domestic water system: (1) carbon containing copper, (2) carbon containing copper and silver, and (3) carbon without metals. Filters were challenged with tap water seeded with L. pneumophila multiple times and water was monitored under simulated domestic use for six weeks. Levels of legionella were reduced by all three filters by nearly 8 log (99.999999%), but they were detected in all filter effluents for the length of the study. The authors concluded that the organisms attached to the carbon blocks and sloughed off over time. PA September 2017 Volume 23 I Number 7