SEARCH

SEARCH BY CITATION

Keywords:

  • Healthcare-associated infection;
  • hospital water supply;
  • infection control;
  • POU water filter;
  • waterborne pathogens

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Hospital water supplies often contain waterborne pathogens, which can become a reservoir for healthcare-associated infections (HAIs). We surveyed the extent of waterborne pathogen contamination in the water supply of a Liver Transplant Unit. The efficacy of point-of-use (POU) water filters was evaluated by comparative analysis in routine clinical use. Our baseline environmental surveillance showed that Legionella spp. (28%, 38/136), Pseudomonas aeruginosa (8%, 11/136), Mycobacterium spp. (87%, 118/136) and filamentous fungi (50%, 68/136) were isolated from the tap water of the Liver Transplant Unit. 28.9% of Legionella spp.-positive water samples (n = 38) showed high-level Legionella contamination (≥10CFU/L). After installation of the POU water filter, none of these pathogens were found in the POU filtered water samples. Furthermore, colonizations/infections with Gram-negative bacteria determined from patient specimens were reduced by 47% during this period, even if only 27% (3/11) of the distal sites were installed with POU water filters. In conclusion, the presence of waterborne pathogens was common in the water supply of our Liver Transplant Unit. POU water filters effectively eradicated these pathogens from the water supply. Concomitantly, healthcare-associated colonization/infections declined after the POU filters were installed, indicating their potential benefit in reducing waterborne HAIs.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Hospital water supplies have served as reservoirs for waterborne pathogens such as Legionella spp., Pseudomonas aeruginosa, Stenotrophomonas maltophilia, Acinetobacter spp., Mycobacterium spp. and fungi [1-5]. The degree of the colonization in water supplies has been correlated with the incidence of healthcare-associated infections (HAIs) [6, 7]. Forty-two per cent of ICU patients with Pseudomonas aeruginosa harbored isolates with identical genotypes to those found in the taps [8]. Water supplies were recognized as one of the most important and controllable, and yet the most overlooked, sources of HAIs [1, 2].

Despite water treatment with chlorination, domestic water supplies may still be contaminated by low concentrations of various microorganisms [9]. Although most of the microorganisms are not harmful to the general public, some opportunistic pathogens pose threats to hospitalized patients. In China, the waterborne pathogen contaminations of water supplies have often been overlooked. In fact, the European Working Group for Legionella Infections (EWGLI) reported in 2009 that China was one of the top 15 countries implicated in cases of travel-associated Legionnaires' disease [10]. In a study of eight hospital water supplies in Shanghai [11], 43.0% (83/193) of water samples were positive for Legionella spp., and 63 water samples exceeded the concentration of 103 CFU/L. So, we sought to determine if waterborne pathogens were present in the water supply of our hospital, especially in the Liver Transplantation Unit (LTU), where the patients are most susceptible to opportunistic infections. Furthermore, could removal of these waterborne pathogens reduce the incidence rate of hospital-acquired infections in the LTU? Thus, we performed an infection control intervention by: (i) investigating the baseline frequency of waterborne pathogens in the water supply of the LTU, and (ii) evaluating the efficacy of point-of-use (POU) water filters in removing waterborne pathogens. To our knowledge, this is the first environmental surveillance of waterborne pathogens in a hospital water supply in China.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Study site

This study was performed in an 18-bed LTU of a university-affiliated general hospital with 1600 beds in Shanghai, China. The Unit consists of nine patient rooms (two patient beds and one sink/tap in each room), one nurses' station and one doctor's office. The hospital receives its water from a municipal water treatment plant without additional on-site disinfection.

Study design

Cold tap water samples were collected between 2009 and 2011 (June, September and October in 2009, January, July, August, September, October and November in 2010, and March in 2011) from each tap outlet in sterile containers with 0.01% w/v sodium thiosulphate.

Three taps located in one patient room, the nurses' station and the doctor's office were installed with 0.2 μm POU filters (AQ14F1S, Pall Corp., Port Washington, NY, USA) for removal of the waterborne pathogens (Fig. 1). A pre-filtration fixture (pore size, 1.2 μm) was also installed for capturing particulate debris to extend the life of the POU filter. Filters were changed every 2 weeks according to the manufacturer's instructions from July to November 2010 (18 weeks), and water samples were collected and cultured every 3–4 days. The unfiltered tap water sample served as the control, while the water filtered through the pre-filter alone served as the pre-filtered water control. We picked the doctor's office and nurses' station for installation so that all medical staff had access to filtered (pathogen-free) water before and between patients' care.

image

Figure 1. Tap installed with POU water filter and pre-filtration fixture.

Download figure to PowerPoint

The incidence of Gram-negative bacteria colonization/infection in the LTU was also monitored. We analyzed patient-related data for the same 4-month period before the installation of the water filters (from July to November 2009) and a corresponding 4-month period after outlets had been equipped with filters (from July to November 2010). Patient data were retrieved from the hospital surveillance system. Microbiological cultures from patients were performed only when clinically indicated. No additional control measures were carried out during this period.

Microbiological analysis

Total heterotrophic plate count (HPC) bacteria, cultured on R2A agar (Oxoid, Basingstoke, UK) at 25°C for 14 days, were enumerated by the standard pour plate method [12]. Legionella spp. was monitored using GVPC selected agar (Oxoid) according to ISO 11 731 [13]. Colonies morphologically consistent with Legionella spp. were identified by the latex agglutination test (Oxoid). For Pseudomonas aeruginosa, filamentous fungi and Mycobacterium spp. detection, water samples were filtered (pore size of 0.45 μm, Millipore, USA) and the filter membrane was placed on Cetrimide agar plates, Sabouraud dextrose agar plates containing 25 mg/L penicillin and 400 mg/L chloramphenicol (Oxoid) and Middlebrook 7H10 plates (BD, Franklin Lakes, NJ, USA), and incubated at 35°C for 48 h, 30°C for 28 days and 35°C for up to 8 weeks, respectively.

Statistical analyses

An ANOVA (SPSS ver. 15.0) was used to analyze the bacterial counts in POU-filtered, pre-filtered and unfiltered control samples. Comparison of the incidence of Gram-negative bacterium infection/colonization in the post-filtration period with that in the pre-filtration period was carried out by use of the chi-squared test (SPSS ver. 15.0). The correlation coefficient of temperature and the number of positive water samples were calculated by use of two-tailed Spearman's analysis.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Baseline of waterborne pathogens in the LTU

A total of 136 cold water samples were enumerated for the targeted pathogens. Legionella spp, Pseudomonas aeruginosa, Mycobacterium spp. and filamentous fungi were detected in 38 (27.9%), 11 (8.1%), 118 (86.8%) and 68 (50.0%) water samples, respectively. HPC bacteria were detected in all the water samples, with a mean concentration of 1.1 × 10CFU/L. Among the isolated Legionella, 29 of these (76.3% of positive samples and 21.3% of the total) were identified as L. pneumophila. Furthermore, 28.9% of Legionella spp.-positive samples were detected with high-level contamination (≥10CFU/L). More than 18% (7/38) of the samples were positive for both L. pneumophila and Legionella of other species (Table 1).

Table 1. Characteristics of pathogen contamination in the cold water samples without filter installation (n = 136)
ParametersLegionella spp. Legionella pneumophila Pseudomonas aeruginosa Mycobacterium spp. Filamentous fungi HPC bacteriaa
  1. a

    Only positive samples were included.

  2. b

    HPC, heterotrophic plate count.

Positive samples, No. (%)38 (27.9)29 (21.3)11 (8.1)118 (86.8)68 (50.0)136 (100.0)
Samples with >10CFU/L, No. (%)11 (8.1)8 (5.9)019 (13.9)0136 (100.0)
Geometric mean count (CFU/L a, Mean (Range))2.9 × 103 (50–5.8 × 104)3.4 × 103 (100–2.0 × 104)70.0 (5–3.6 × 102)5.9 × 102 (2–5.0 × 103)41.5 (10–62)1.1 × 107 (1.0 × 104 –3.4 × 108)

Mycobacterium spp. was isolated from almost all water samples throughout the study. However, if we examine the data without considering Mycobacterium spp., sampling sites positive for target pathogens were higher in the hot season (from June to October), averaging 5.7 sites positive/month (40/7), compared with the cold season (from November to March), averaging three sites positive/month (9/3), which is almost a two-fold increase. Some pathogens seemed to persist in some outlets for a long time; for example, filamentous fungi were isolated from tap water of room one during the entire study period. We also found that the positive rate of Legionella spp. correlated with temperature fluctuations of tap water (correlation coefficient = 0.907; p 0.000), which suggested that cold water temperature below 20°C might be considered protective against Legionella contamination (Fig. 2).

image

Figure 2. Relationship between contamination rate of tap water and temperature (Legionella spp.).

Download figure to PowerPoint

Control modality using POU filter

As the water samples were found to be highly contaminated by Legionella spp. (103–104 CFU/L), three POU water filters were installed. From July to November 2010, a total of 190 water samples were collected from these three tap outlets, of which 57 were unfiltered water, 43 were pre-filtered water and 90 were POU-filtered water. No significant difference was observed in Legionella isolation between pre-filtered and unfiltered water (Table 2). In contrast, all samples filtered by the POU water filter were culture-negative for any of these pathogens. The difference in isolation between POU-filtered and control water was significant at p < 0.05. It is noteworthy that one of 34 (2.9%) water samples tested positive for HPC bacteria after 3 days use of filters, four of eight (42.1%) water samples tested positive after 7 days use, and the positive rate increased to 69.2% after 14 days use. Retrograde contamination may occur during use over time.

Table 2. Growth of different pathogens and HPC bacteria in unfiltered, pre-filtered and POU-filtered water samples
 Total No. of samplesNo. of positive samples (%)Mean concentration of organisms in positive samples (CFU/L)
HPC bacteriaLegionellaMycobacteriumFilamentous fungi
Unfiltered water5757 (100)1.2 × 1081.0 × 1034.0 × 10216.5
Pre-filtered water4343 (100)3.7 × 1081.8 × 1037.0 × 10212.7
Filtered water9034 (37.7)3.4 × 104000
Filtered water after 3-day interval341 (2.9)0000
Filtered water after 7-day interval198 (42.1)1.3 × 104000
Filtered water after 10-day interval217 (33.3)5.4 × 104000
Filtered water after 14-day interval2618 (69.2)2.6 × 104000

The number of Gram-negative bacterium infection/colonization patients per 1000 patient-days of hospitalization in the post-filtration period (1.70 ± 0.95) was significantly lower than that in the pre-filtration period (3.20 ± 1.25; χ2 = 2.119, p 0.067). Gram-negative bacterium colonizations/infections were reduced by 46.9%.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Opportunistic waterborne pathogens can be introduced into a healthcare facility water distribution system. Despite water treatment and a chlorine disinfection process, treated water may still contain low concentrations of various microorganisms, such as Legionella, P. aeruginosa, non-tuberculous mycobacteria and fungi (e.g. Aspergillus). Pathogens can enter the water system of healthcare facilities and can colonize the water supply piping, hot water tanks, sinks, faucet aerators and shower heads. Hospital water distribution systems might be one of the most important sources of HAIs [1]. Thus, the World Health Organization (WHO) published its fourth edition of ‘Guidelines for Drinking-Water Quality' [9], which specifically stated the importance of disinfection of the water supply as a control measure to prevent healthcare-associated infections. However, as in healthcare facilities throughout the world, no mandate exists for Chinese healthcare facilities to survey for waterborne pathogens in the water supply of healthcare facilities.

We conducted this prospective surveillance in the absence of any recognized outbreak attributable to waterborne pathogens of Legionella spp., Pseudomonas aeruginosa, non-tuberculous mycobacteria and filamentous fungi. A high prevalence rate of waterborne pathogens was found in the water supply of the LTU. During the interventional control strategy using POU water filter, we found that POU filters completely eliminated these waterborne pathogens from the water supply over 14 days of use. Furthermore, the rate of Gram-negative bacterium infection/colonization patients per 1000 patient-days of hospitalization in the post-filtration period (1.70 ± 0.95) decreased significantly compared with the pre-filtration period (3.20 ± 1.25; χ2 = 2.119, p = 0.067), a 47% reduction! Our study demonstrated that POU water filters provided a barrier against various waterborne pathogens that can further reduce the rate of nosocomial infections. However, the retrograde contaminations may occur by either splash water from the water basin during use or by direct contact with contaminated hands and dirty clothes of staff or patients [14, 15]. In our surveillance, HPC bacteria were recovered from the filtered water after 1 week of use. The source of the HPC bacteria from filtered water remains unclear, and molecular typing may be useful to track the dissemination.

Due to the poor quality of the supplied tap water in the LTU, we installed pre-filtration fixtures on three taps upstream of POU filters in order to remove particulate debris. Before the POU filter study, laboratory and field tests were conducted for evaluating the performance of pre-filtration of various materials and styles and removal ratings, and 1.2-μm pore size was chosen for the pre-filtration fixture (data not shown). The pre-filtration fixture would not prohibit the waterborne pathogens from tap water. However, the concentration of various pathogens after pre-filtration was surprisingly higher than that in unfiltered water in some samples. The reason may be the growth of pathogens within the pre-filtration media because of higher nutrient content from the trapped debris in the water.

Although only 27% of distal sites (3/11) were equipped with POU filters, the incidence of Gram-negative bacterium colonization/infection decreased significantly, possibly because of the use of filtered tap water for perineal washing of patients, the bed environment and the hands of nursing personnel. In this study, it remains unclear whether such filters contribute to the reduction of non-tuberculosis mycobacterium and filamentous fungi infections in high-risk patients. Therefore, more research is needed to evaluate the efficacy and cost-effectiveness of POU filters in preventing specific colonization/infection of hospitalized patients. Many studies have focused primarily upon recognized outbreaks of Legionella spp. and P. aeruginosa [16, 17]. However, other opportunistic waterborne pathogens may also cause nosocomial infections, outbreaks or sporadic infections [18, 19]. There is controversy over whether it is economical to invest medical resources in preventing opportunistic waterborne pathogen-associated nosocomial infections, especially because the disposable POU water filters have a limited effective life and could be very expensive. We propose a modest approach whereby removal of waterborne pathogens is targeted towards areas of highest risk of nosocomial infections, such as our transplant unit. In such settings the costs are likely to be justifiable. We should remove waterborne pathogens from transplantation units because these patients are at the highest risk of nosocomial. If an organ transplant patient dies from nosocomial infection, a valuable organ is also being destroyed. Given such a high potential cost associated with nosocomial infections in transplant patients, POU water filters may be a viable economical option [20, 21]. Instead of treating the entire hospital water supply with systematic chemical disinfection (e.g. chlorination), POU filters can be easily installed at a few sites for prevention of infection [22-24]. Furthermore, based on our data, we suggest installing POU water filters only in the hot season (June to October) in countries with limited medical resources.

In conclusion, hospital water supplies were highly contaminated by various waterborne pathogens. Using POU filters appeared to be one of the most simple and cost-effective methods to reduce the risk of waterborne pathogen-associated infections in hospitals.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

We gratefully thank the nurses and medical staff of the Liver Transplantation ICU, who cooperated fully with this study. We also thank Professor Yuanlin Song for his valuable suggestions and the review of manuscript.

Transparency Declaration

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Financial support was received from the Special Fund for Health-scientific Research in the Public Interest: Research and application for the prevention and control of nosocomial infections coursed by multi-drug resistant bacteria (No. 201002021). All authors report no conflicts of interest relevant to this article.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References
  • 1
    Anaissie EJ, Penzak SR, Dignani MC. The hospital water supply as a source of nosocomial infections: a plea for action. Arch Intern Med 2002; 162: 14831492.
  • 2
    Cervia JS, Ortolano GA, Canonica FP. Hospital tap water a reservoir of risk for health care-associated infection. Infect Dis Clin Pract 2008; 16: 349353.
  • 3
    Kusnetsov J, Torvinen E, Perola O et al. Colonization of hospital water systems by legionellae, mycobacteria and other heterotrophic bacteria potentially hazardous to risk group patients. APMIS 2003; 111: 546556.
  • 4
    Merlani GM, Francioli P. Established and emerging waterborne nosocomial infections. Curr Opin Infect Dis 2003; 16: 343347.
  • 5
    Wang JL, Chen ML, Lin YE et al. Association between contaminated faucets and colonization or infection by non-fermenting gram-negative bacteria in intensive care units in Taiwan. J Clin Microbiol 2009; 47: 32263230.
  • 6
    Kool JL, Bergmire-Sweat D, Butler JC. Hospital characteristics associated with colonization of water systems by Legionella and risk of nosocomial Legionnaires' disease: a cohort study of 15 hospitals. Infect Control Hosp Epidemiol 1999; 20: 798805.
  • 7
    Lin YE, Stout JE, Yu VL. Prevention of hospital-acquired legionellosis. Curr Opin Infect Dis 2011; 24: 350356.
  • 8
    Cuttelod M, Senn L, Terletskiy V et al. Molecular epidemiology of Pseudomonas aeruginosa in intensive care units over a 10-year period (1998-2007). Clin Microbiol Infect 2011; 17: 5762.
  • 9
    World Health Organization. Guidelines for drinking-water quality, 4th edn. Geneva (Switzerland): World Health Organization, 2011.
  • 10
    Joseph CA, Ricketts KD, Yadav R, European Working Group for Legionella Infections et al. Travel-associated Legionnaires' disease in Europe in 2009. Euro Surveill 2010; 15: pii=19683.
  • 11
    Tao LL, Hu BJ, Zhou ZY et al. Contamination of Legionella spp. in water distribution systems in 8 hospitals of Shanghai, China. Chin J Nosocomiol 2010; 20: 17101712.
  • 12
    Colony count by the pour plate method. National Standard Method W4. 2007. Washington DC: Health protection agency, 2007.
  • 13
    International Organization for Standardization. Water quality-Detection and enumeration of Legionella - Part 2: Direct membrane filtration method for waters with low bacterial counts. ISO 11731-2: 2004.
  • 14
    Sheffer PJ, Stout JE, Wagener MM et al. Efficacy of new point-of-use water filter for preventing exposure to Legionella and waterborne bacteria. Am J Infect Control 2005; 33: S20S25.
  • 15
    Vonberg RP, Sohr D, Bruderek J. Impact of a silver layer on the membrane of tap water filters on the microbiological quality of filtered water. BMC Infect Dis 2008; 8: 133.
  • 16
    Warris A, Onken A, Gaustad P et al. Point-of-use filtration method for the prevention of fungal contamination of hospital water. J Hosp Infect 2010; 76: 5659.
  • 17
    Williams MM, Chen TH, Keane T. Point-of-use membrane filtration and hyperchlorination to prevent patient exposure to rapidly growing mycobacteria in the potable water supply of a skilled nursing facility. Infect Control Hosp Epidemiol 2011; 32: 837844.
  • 18
    Hussein Z, Landt O, Wirths B et al. Detection of non-tuberculous mycobacteria in hospital water by culture and molecular methods. Int J Med Microbiol 2009; 299: 281290.
  • 19
    Hageskal G, Kristensen R, Fristad RF et al. Emerging pathogen Aspergillus calidoustus colonizes water distribution systems. Med Mycol 2011; 49: 588593.
  • 20
    DaeschleinG , Krüger WH, Selepko C. Hygienic safety of reusable tap water filters (Germlyser) with an operating time of 4 or 8 weeks in a haematological oncology transplantation unit. BMC Infect Dis 2007; 23: 45.
  • 21
    Cervia JS, Farber B, Armellino D et al. Point-of-use water filltration reduces healthcare associated infections in bone marrow transplant recipients. Transpl Infect Dis 2010; 12: 238241.
  • 22
    Holmes C, Cervia JS, Ortolano GA et al. Preventive efficacy and cost-effectiveness of point-of-use water filtration in a subacute care unit. Am J Infect Control 2010; 38: 6971.
  • 23
    Trautmann M, Halder S, Hoegel J et al. Point-of-use water filtration reduces endemic Pseudomonas aeruginosa infections on a surgical intensive care unit. Am J Infect Control 2008; 36: 421429.
  • 24
    Marchesi I, Marchegiano P, Bargellini A et al. Effectiveness of different methods to control legionella in the water supply: ten-year experience in an Italian university hospital. J Hosp Infect 2011; 77: 4751.