Changes in microbial population in hospital water networks treated by monochloramine and hydrogen peroxide

Objectives: Legionella control still remains a critical issue in healthcare settings and any disinfection methods applied as control measure needs to be fine-tuned to obtain satisfactory results in individual hospitals over prolonged time periods. Aim of the study was to analyze the effect of new disinfectants on changes in microbial population and selection of waterborne pathogens. Methods: Two hospital setting were studied, an Emergency Department building, a 112-bed ward of a teaching hospital, where a monochloramine-based (MC) disinfection of the hot water network (2 mg/L) began in November 2010; a 136-bed general hospital where a hydrogen peroxide (HP) treatment of the hot water (25 mg/L) started in July 2013. Legionella was isolated in accordance with ISO-11731 and typed following the SBT protocol (ELDSNet). Negative samples were assayed for Viable But Nonculturable Cells (VBNC) by real-time PCR and ATP-bioluminescence analysis following immunomagnetic separation. Positive samples for both methods were subjected to “resuscitation” test. The presence of nontuberculous Mycobacteria (NTMs) was assessed by culture and confirmed with sequencing of hsp65 gene. Results: In the first setting, all samples taken from sites previously positive for L.pneumophila ST269 (mean count 7.2x103CFU/L) resulted negative during the 42-months monitoring, with the exception of three instances temporally linked to failure of the MC generator. Among the 112 negative water samples, 69 gave positive results at PCR analysis (mean 3.7x103GU/L) and high ATP values (mean 8.2x102pg/L), suggesting the presence of VBNC cells although only two samples resulted positive at “resuscitation” test (MC equal to 1.5 mg/L). Following the start of MC treatment, a significant rise in NTMs mean loads (from 1.9X101 to 1.4X103 CFU/L, p-value<0.05) was observed for MC concentration below 2 mg/L. The predominant species was M. gordonae (86%). NTMs were not isolated from water following the increase of MC to 3 mg/L. However, recently M. gordonae reappeared with 1,1x103CFU/L mean load. In the second hospital, L.pneumophila sg4, sg5, sg10 were isolated in all sites with a mean count of 9,9x103CFU/L; following the start of HP-treatment (25 mg/L) the percentage of positive sites was reduced by 50% and the mean load was 4x103CFU/L; however other Legionella species never cultured before appeared, including L.pneumophila sg1. Conclusion: A good efficacy in controlling Legionella colonisation was demonstrated for MC disinfection ( 2 mg/L), however the alteration of the water microbial population cannot be excluded. NTMs resulted difficult to control. On the other hand, HP seemed not to be able to eliminate all species of Legionella.