To evaluate the effects of regular flushing, water from fifty emergency eyewash and shower stations was cultured for the presence of potentially pathogenic protozoa, heterotrophic bacteria, and Legionella species. This study also provided the opportunity to evaluate a commercially available molecular assay for the direct detection of Legionella sp in environmental samples. The Perkin Elmer Legionella EnviroAmp polymerase chain reaction (PCR) kit and culture on buffered charcoal yeast extract agar were used to detect Legionella species in water samples. Chemical and physical parameters of station water measured included: pH, hardness, alkalinity, turbidity, conductivity, total chlorine and assimilable organic carbon. Protozoal isolates were identified by classical identification methods, and isolates from the stations were identified as Hartmannella sp, Vexillifera sp, Vahlkampfia sp, Acanthamoeba sp, and Vanella sp. Heterotrophic plate counts ranged from 102 to 106 CFU ml−1 and acridine orange total counts ranged from 103 to 106 cells ml−1 after regular flushing. PCR and gene probe analysis showed that 89% of the stations
(eyewash and shower) were positive for Legionella species by PCR, while 6% of the samples were culture positive.

These results indicate that routine flushing alone is not sufficient to control microbial contamination and disinfection must also be included in a routine maintenance program. In addition, regular maintenance, disinfection, and monitoring of emergency eyewash and shower stations is important in preventing potential secondary microbial infections by either direct inoculation or aerosol transmission.

Keywords: emergency shower; eyewash stations; PCR; Legionella; protozoa

Introduction
Emergency eyewash and shower stations are relied upon to provide a cleansing stream of water in personal injuries resulting from chemical spills, burns or other accidents.
Since many of these emergency showers and eyewash stations go unused for months to years, a significant microbial load can accumulate in the pipes and hoses in the form of
biofilms [19]. Primary organisms of concern include protozoa and a number of opportunistic bacterial pathogens including Legionella and Pseudomonas sp. The primary route of transmission for Legionella is through the inhalation of contaminated aerosols, although wound infections
have also been reported [6,8,14]. One study found that 90% of the L. pneumophila cells which were recovered from  showerhead and faucet aerosols were in the 1–5 mm range, in the size range to penetrate the human lower respiratory
tract and initiate infection [4]. Similarly, emergency showers can produce aerosolized water droplets in a similar particle size that may represent not only a route of transmission but also a potential health risk to the end user.
Certain species of free-living amoebae belonging to the genera Acanthamoeba and Naegleria are well established ‘opportunistic pathogens’ capable of causing disease and death in humans [17]. Another species of protozoa, Hartmannella is involved in a complex interaction with Legionella, by providing an ideal intracellular multiplication and amplification niche. In addition, these protoza provide growth-supporting factors and shield the internalized Legionella from unfavorable environmental conditions
(ie disinfection) [13,20,25].
One month prior to the initiation of this study, regular flushing of the stations, as suggested by the American National Standards Institute (Z358.1–1981) and [23] was
implemented to evaluate the effects of regular flushing on microbiological water quality. This study was undertaken to determine if immediate implementation of a maintenance program might preclude the need for costly repairs or disinfectant applications. Prior to the initiation of this study, all 50 stations were subjected to regular flushing (bi-monthly for several minutes). Flushing consisted of turning the units on for a period of 3–5 min to flush the stagnant water from
the supply line.

Materials and methods

Sample collection Sample volumes of 100–250 ml were collected in sterile polypropylene sample bottles directly from the opening of the eyewash station or emergency shower. Samples for Assimilable Organic Carbon (AOC) were collected in 1-L acid-washed and baked glass bottles, and AOC calculations were determined as previously described [10]. A total of 50 emergency station water samples were collected, 25 eyewash stations and 25 emergency showers. In addition, a single floor drain sample was also collected to serve as a control. Parallel tap water samples were also collected as controls. Samples were collected from two states and from eight different buildings (which all contained copper plumbing systems). The eight buildings represented three

different municipal water supplies, all three similar in terms of water quality. The plumbing serving the emergency showers and eyewash stations in the eight buildings sampled could be broken down into three age groups, ,2 years, ,12 years, and .40 years. Only stationary eyewash and shower stations were examined. The eyewash units included in the study were all twin eyewash head units and the shower units were overhead (singlehead) drench type showers supplied by a vertical water supply line (in house copper plumbing). Both the eyewash and shower stations were constructed of stainless steel.

Measurement of chemical and physical parameters Sample pH was recorded using a Corning pH meter (model 125) (Corning, NY, USA). Chlorine concentrations were determined by using the N,N-diethyl-p-phenylenediamine (DPD) titration method (American Public Health Association, Standard Methods for the Examination of Water and Wastewater 1985). Hardness, alkalinity, turbidity, and conductivity were all measured according to the Standard Methods for the Examination of Water and Wastewater [1].

Protozoal identification

Samples were concentrated by centrifugation (1500 × g) for 5 min to a final volume of 2 ml of which 100 ml was analyzed. Minimal agar plates (15 g agar (Difco Laboratories, Detroit, MI, USA), 0.1 g each of malt extract and yeast extract) streaked with Klebsiella pneumoniae (ATCC 27889) were inoculated with 100 ml of the concentrate in duplicate and incubated at room temperature (21–22°C). In addition, a liquid medium, ATCC Medium 802 (rye grass cerophyll in distilled water) supplemented with 106 cells ml−1 K. pneumoniae was also inoculated with 100 ml of the sample concentrate. Samples were examined with a Leitz inverted microscope over a 2-week period. Amoebae were identified to the genus level based on morphological features, measurement of living trophozoites and cysts. Ciliates, flagellates, yeasts, fungi, and nematode eggs were also isolated. One fungal isolate was identified to the species level. Bacterial enumeration

The heterotrophic plate count was determined by membrane filtration on R2A agar (Difco Laboratories) with incubation at 28°C for 7 days and by the pour plate method at 35°C for 48 h. Acridine Orange Direct Counts (AODC) were performed according to the method of Hobbie et al [11]. Pseudomonas species were identified by biochemical tests as previously described [19]. Assimilable Organic Carbon (AOC) measurements were made according to the method of van der Kooij et al [24]. PCR analysis for Legionella species The EnviroAmp Legionella PCR kit obtained from PerkinElmer (Foster City, CA, USA) was used for the molecular detection of Legionella species as per manufacturer’s instructions. Emergency shower and eyewash samples were analyzed directly, however a few samples contained inhibitors of PCR and required dilution (1 : 10 or 1 : 100) or treatment with bovine serum albumin (0.2% BSA). The BSA served to stabilize the AmpliTaq polymerase and may also act to bind inhibitors. Water samples were analyzed for Legionella species by filtering a minimum of 10 ml of each water sample. PCR was used to amplify specific sequences from a conserved region of Legionella’s 5S rRNA [16] and from the mip (macrophage infectivity potentiator) gene specific for the L. pneumophila species [7]. Biotinylated amplification products 5S rRNA and/or mip, are detected following hybridization on nylon membranes to complementary immobilized probes with detection by streptavidin-horseradish peroxidase conjugate (via a reverse dot blot technique). The assay is able to detect approximately 10 Legionella cells ml−1 in the original sample, however, sensitivity can be enhanced by increasing the volume that is filtered. The sensitivity of the PCR assay was confirmed through seeding studies using Legionella ATCC type strains. Genus and species identification is achieved by interpreting the nylon membranes; two blue dots of equal intensity next to the ‘L’ and the ‘P’ symbols indicates the presence of L. pneumophila. Each strip has a positive and negative control, the intensity of the positive control equals 103 cells ml−1 .

In these experiments, when the ‘L’ or ‘P’ dot was lighter than the internal positive control the sample was considered to contain less than 103 cells ml−1 in the original sample. When the ‘L’ or ‘P’ dot was darker than the internal positive control the sample was considered to contain greater than 103 cells ml−1 in the original sample. Based on the intensity of the dots, the assay is semiquantitative. The assay is able to detect 25 Legionella species.

Legionella culture and microbial analyses All water samples were analyzed for Legionella by culture. One hundred-milliliter samples were filtered through 0.2- mm polyethersulfone-based membrane filters (Supor filters, Gelman, Ann Arbor, MI, USA), placed in 3 ml of filtersterilized deionized water and vortexed for 60 s to dislodge any trapped bacteria. Following concentration, samples were subjected to acid treatment to eliminate nonLegionella organisms [5]. After acid treatment, 0.1 ml was spread plated in duplicate onto selective and non-selective media. Legionella samples were cultured on buffered charcoal yeast extract agar amended with alpha-ketoglutarate (BCYEa) alone or amended with glycine (3 g L−1 ) and the following antibiotics: vancomycin (5 mg L−1 ), polymyxin B (100 IU ml−1 ), and cyclohexamide (80 mg L−1 ) (BCYEaGVPC). Plates were incubated from 5–15 days at 37°C with 2.5% CO2 in a humid environment. A sequential culturing method [21] was also employed in an attempt to improve the recovery efficiency of Legionella species. Colonies exhibiting Legionella morphology were transferred to BCYEa with and without cysteine; colonies which were unable to grow without cysteine were further tested by immunofluorescent staining.

Suspect colonies were placed in 5 ml of filter-sterilized water and 20 ml of each sample was added to one well of an eight-well toxoplasmosis slide (Bellco Biotechnology, Vineland, NJ, USA), air dried, and heat fixed. Next, 20-ml aliquots of a Poly II conjugate were added to each well. The conjugate contained antisera against: L. pneumophila serogroups 1–7 and L. bozemaneii, L. dumoffii, L. jordanis, L. micdadei, L. gormanii, and L. longbeachae serogroups 1 and 2 (Scimedix, Denville, NJ, USA). Following incu

bation at 37°C for 30 min, slides were washed in phosphate-buffered saline (pH 7.6), rinsed in distilled water and air dried. Slides were prepared for microscopy and analyzed under an epifluorescent microscope (Olympus Model BH2, Lake Success, NY, USA).
Results and discussion
Water quality measurements Table 1 presents the results and basic statistics of the chemical and physical parameters of the water collected from emergency showers and eyewash stations. Total chlorine levels were undetectable for many samples, with average values of 0.10 ppm for eyewash stations and 0.01 ppm for emergency showers. It is not surprising that chlorine residuals were not detected in many of the stations, since the water often remains in the pipes for extended periods. If the stations are flushed for 5 min, chlorine levels similar to the tap water controls are detected. Mean turbidity measurements were higher in emergency showers (1.92 NTU) than in eyewash stations (0.43 NTU). Higher shower turbidities may be due to the force at which water is expelled, possibly causing debris from the biofilm to slough off, whereas the eyewash samples are expelled upward, against gravity. Visible turbidity (debris and rust) was observed in 15/50 (30%) of all stations. The mean total chlorine level of the tap water controls was 0.9 ppm and the total bacterial count ranged from negative (no growth) to 1.3 × 104 CFU ml−1 on R2A media. Heterotrophic plate counts (HPC) for the eyewash stations were slightly higher by the membrane filtration method than those of emergency showers, however, AODC measurements for both were similar as seen in Table 1. Seven samples were subjected to Assimilable Organic Carbon (AOC) analyses. AOC are carbon compounds which are easily assimilable (utilizable as food source) by waterborne bacteria and may be responsible for bacterial regrowth or increased multiplication of indigenous microorganisms in response to increased nutrient availability. It has been suggested that an AOC concentration of less than 10 mg L−1 would be sufficient to maintain biological stability [24]. Mean AOC values of 45 mg L−1 (eyewash stations) and 31 mg L−1 (showers) were detected in this study, however one eyewash sample had an AOC value of 126.34 mg L−1 . Increased AOC values did not always correspond to the highest HPC counts, since the indigenous microorganisms may have been unable to utilize some of the recalcitrant forms of carbon present.

Protozoal and other microorganism identification
Of the fifty samples collected, twenty-six were examined for the presence of protozoa, fungi, yeasts, cilitates, flagellates, and nematodes. Table 2 presents the number of times a sample was positive for various protozoa. By far the most frequent isolate of eyewash stations was Hartmannella species, the same was true for emergency showers. Several samples were positive for four different genera of protozoa. Seventy-six percent of the 26 eyewash samples examined were positive for at least one species of protozoa. Examination of a sample from the furnace room floor drain was found to be positive for multiple protozoa (see Table 2). Two positive for a filamentous

fungi identified as Fusarium moniliforme. Two other eyewash samples were positive for yeast, while one was positive for nematode eggs and only the floor drain sample was positive for ciliates. In contrast to another study [23], where a mixture of Hartmannella and Acanthamoebae were primary isolates of eyewash stations, our study only found Acanthamoebae in two out of 25 (8.0%) stations. In previous studies [3,19] Hartmannella was the most prevalent protozoa identified, as it was in this study 11/25 (44.0%). Tap water control samples were negative for protozoa, Legionella (by PCR and culture), yeast, and fungi. Legionellae are able to survive and multiply within the vacuoles of certain species of protozoa (trophozite form) until the protozoal cell bursts, however if adverse environmental conditions force the protozoa to encyst, the Legionella are shielded until conditions for excystation exist [2,9,25]. In the cyst state protozoa and associated Legionella are extremely resistant to temperature extremes, desiccation and disinfection. For example, it has been demonstrated that L. pneumophila within Acanthamoeba polyphaga can survive exposure to 50 mg L−1 of free chlorine [13]. Given the association of protozoa in biofilms combined with the ability to encyst and resist disinfection, eradication efforts are extremely difficult.

 Legionella species detection by PCR and culture

 The use of the EnviroAmp allowed the detection of Legionella species at concentrations as low as 10 cells ml−1 . The PCR is semiquantitative, and samples yielding a positive at the same intensity as the internal positive control contain approximately 1000 Legionella cells ml−1 . Subsequently darker positives contain greater than 1000 cells and faint positives are ‘graded’ less than 1000 cells or 10– 1000 cells ml−1 . Table 3 examines the concentration of Legionella species in eyewash and emergency shower samples respectively. It is interesting to note that 22% of eyewash samples and 30% of emergency showers are positive for Legionella species at concentrations of 10–1000 cells ml−1 . Legionella pneumophila was also more frequently detected in eyewash stations than in emergency showers and interestingly amoebae were also more frequently detected in eyewash stations (76.9%) than in shower (46.2%).

This increased detection may be due to the smaller pipe diameter and increased surface area of the eyewash stations. The number of eyewash stations positive for Legionella was slightly greater in this study (94.4%), than previously reported (87.5%) [19]. The previous studies utilized direct fluorescent antibody staining which may have missed Legionella present in low concentrations due to shielding of the organisms by debris. Another study of ground, surface and potable waters for the presence of Legionella showed similar results in that there was a much higher rate of positives when samples were analyzed by PCR in comparison with culture or DFA [15]. This may be due to the increased sensitivity and increased number of species (over DFA) provided by the PCR assay. Despite inclusion of the sequential culturing method, Legionella isolates were only recovered from three samples, one eyewash and two shower. Of the three positive samples, only one (showerhead) was recovered through repeated culturing. This is not surprising given the recent report of inhibition of Legionella species by bactericins secreted by HPC bacteria [22]. Legionellae are also known to exist in a viable but nonculturable state when exposed to unfavorable environmental conditions [12]. Non-culturable Legionella cells have also been shown to cause legionellosis. In one report by Miller et al [18] an outbreak of Pontiac fever at a health resort was traced to a contaminated hot tub. Legionella cells were only detected by PCR and DFA, culture results during the outbreak were uniformly negative. Interestingly, about 7 months after the outbreak the sand filters tested culture positive for the same serogroup of Legionella that was implicated in the outbreak [18]. In light of this information, even emergency stations which contain dead Legionella cells should be flushed and cleaned thoroughly, followed by routine monitoring employing culture and DFA or culture and PCR for the detection of Legionella. The results of this study clearly show that initiation of a routine maintenance program, consisting of bi-monthly flushing, is not sufficient to control microbial contamination of emergency stations. More aggressive disinfection methods including hyperchlorination and heat shock treatments should be investigated to prevent contaminated stations especially in older buildings. In addition, the use of large volume, sterile water, self-contained single-use units should be investigated in high risk areas (industrial chemical plants, university chemistry laboratories).

Acknowledgements

The authors would like to thank Mr Elmer Davis of the Mycology Department of the American Type Culture Collection, Rockville, MD for the examination and identification of the fungal isolates. The authors also thank Ken A Sakamoto of the Orange County Sanitation District, Fountain Valley, CA for excellent technical assistance

References
1 American Public Health Association. Standard Methods for the Examination of Water and Wastewater. 1992. APHA, AWWA, and WPCF, Washington, DC (18th edn).

2 Anand CM, AR Skinner, A Malic and JB Kurtz. 1983. Interaction of Legionella pneumophila and a free-living amoeba (Acanthamoeba palestinensis). J Hyg 91: 167–178.

3 Bier JW and TK Sawyer. 1990. Amoebae isolated from laboratory eyewash stations. Curr Microbiol 20: 349–350.

4 Bollin GE, JF Plouffe, MF Para and B Hackman. 1985. Aerosols containing Legionella pneumophila generated by shower heads and hotwater faucets. Appl Environ Microbiol 50: 1128–1131. 　

5 Bopp CA, JW Sumner, GK Morris and JG Wells. 1981. Isolation of Legionella spp from environmental water samples by low-pH treatment and use of selective medium. J Clin Microbiol 13: 714–719. 　

6 Breiman RF, BS Fields, GN Sanden, L Volmer, A Meier and JS Spika. 1990. Association of shower use with Legionnaires’ disease. JAMA 263: 2924–2926.

　7 Engleberg NC, C Carter, DR Weber, NP Cianciotto and BI Eisenstein. 1989. DNA sequences of mip, a Legionella pneumophila gene associated with macrophage infectivity. Infect Immun 57: 1263–1270. 　

8 Fallon RJ and TJ Robotham. 1990. Microbiological investigations into an outbreak of Pontiac fever due to Legionella micdadei associated with use of a whirlpool. J Clin Path 45: 47. 　

9 Fields BS, GS Sanden, JM Barbaree, W Morrill, RM Wadowsky, EH White and JC Feeley. 1989. Intracellular multiplication of Legionella in amoebae isolated from hot water tanks. Curr Microbiol 18: 131–137.

　10 Hacker PA, C Paszko-Kolva, MH Stewart, RL Wolfe and EG Means. 1993. Production and removal of assimilable organic carbon under pilot plant conditions through the use of ozone and PEROXONE. Ozone Sci & Eng 16: 197–212.

　11 Hobbie JE, RJ Daley and S Jasper. 1977. Use of nucleopore filters for counting bacteria by fluorescence microscopy. Appl Environ Microbiol 33: 1225–1228. 　

12 Hussong D, RR Colwell, M O’Brien, AD Weiss, AD Pearson, RM Weiner and WD Burge. 1987. Viable Legionella pneumophila not detectable by culture on agar media. Biotechnology 5: 947–950. 　

13 Kilvington S and J Price. 1990. Survival of Legionella pneumophila within cysts of Acanthamoeba polyphaga following chlorine exposure. J Appl Bacteriol 68: 519.

　14 Lowry PW, RJ Blankenship, W Gridley, NJ Troup and LS Tompkins. 1991. A cluster of Legionella sternal wound infections due to postoperative topical exposure to contaminated tap water. N Engl J Med 324: 109–113.

15 Lye D, GS Fout, SR Crout, R Danielson, C Thio and C Paszko-Kolva. 1997. Survey of ground, surface, and potable waters for the presence of Legionella species by EnviroAmp PCR Legionella kit, culture, and immunofluorescent staining. Water Res 31: 287–293.

　16 MacDonell MT and RR Colwell. 1987. The nucleotide sequence of the 5S rRNA from Legionella pneumophila. Nucl Acids Res 15:1335. 　

17 Martinez AJ. 1985. Free-living amoebas: natural history, prevention, diagnosis, pathology and treatment of the disease. CRC Press Boca Raton Fl MRC Symposium. Thorax 32: 653–663. 　

18 Miller LA, L Beebe, JC Butler, W Martin, R Benson, RE Hoffman and BS Fields. 1993. Use of polymerase chain reaction in an epidemiological investigation of Pontiac fever. JID 168: 769–772.

　19 Paszko-Kolva C, H Yamamoto, M Shahamat, TK Sawyer, G Morris and RR Colwell. 1991. Isolation of amoebae and Pseudomonas and Legionella spp from eyewash stations. Appl Environ Microbiol 57: 163–167.

　20 Rowbotham TJ. 1980. Preliminary report on the pathogenicity of Legionella pneumophila for freshwater and soil amoebae. J Clin Path 33: 1179–1193. 　

21 Shahamat M, C Paszko-Kolva, J Keiser and RR Colwell. 1991. Sequential culturing method improves recovery of Legionella spp from contaminated environmental samples. Zbl Bakt 275: 312–319.

　22 Toze S, LI Sly, IC MacRae and JA Fuerst. 1990. Inhibition of growth of Legionella species by heterotrophic plate count bacteria isolated from chorinated drinking water. Curr Microbiol 21: 139–143.

　23 Tyndall RL, MM Lyle and KS Ironside. 1987. The presence of freeliving amoebae in portable and stationary eyewash stations. Am Ind Hyg Assoc J 48: 933–934. 　

24 Van der Kooij D, A Visser and WAM Hijnen. 1982. Determining the concentration of easily assimilable organic carbon in drinking water. J Am Water Works Assoc 74: 540–545.

　25 Wadowsky RM, LJ Butler, MK Cook, SM Verma, MA Paul, BS Fields, G Keleti, JL Sykora and RB Yee. 1988. Growth-supporting activity for Legionella pneumophila in tap water cultures and possible implication of hartmannellid amoebae as growth factors. Appl Environ Microbiol 54: 2677–2682.