Bacterial Removal from Inexpensive Portable ... - Wiley Online Library

Bacterial Removal from Inexpensive Portable Water Treatment Systems for Travelers O. Schlosser, C. Robert, C. Bourderioux, M. Rey, and M.R. de Roubin Background:There are many personal portable water treatment systems for travelers on the market, including chemical agents, iodine resin purifiers and filters. However, information on the real efficacy of these systems in the field is often lacking. We have therefore estimated the capabilities of several inexpensive personal portable water treatment systems for travelers to remove bacteria in various situations of water quality, using stressed indigenous strains of bacteria. Methods: Four chemical agents (Drinkwell chlorine,® Hydroclonazone,® Aquatabs,® 2% iodine in ethanol), two iodine resin purifiers (the straw PentaPure Outdoor M1-E,® the PentaPure Traveler ® purifying and filtration system) and four filters (the flexible bottle Pres2Pure,® the hand-pump filters Mini Ceramic,® First Need Deluxe® and WalkAbout ®) were evaluated in triplicate using both turbid and clear water at 25°C. Bacteria were counted by conventional culturing techniques, colorimetric and fluorescent assays of coliforms and Escherichia coli enzyme activities (Colilert ®/Quantitray ® method), and viable but not culturable bacteria were assessed quantitatively by 5-cyano-2,3-dilotyl-tetrazolium staining. Results: The best systems were the three hand-pump filters, Mini Ceramic,® First Need Deluxe,® and WalkAbout.® All had a submicron filtration element that completely removed 3 log (99.9%) or more of viable bacteria, and no coliforms or E. coli were detected in the effluent. The PentaPure Traveler ® removed more than 99.3% of the viable bacteria. The only chemical agents that gave a bacterial inactivation of over 2 log in clear water were the Drinkwell chlorine,® the Aquatabs,® and the 2% iodine in ethanol. The three other devices, Hydroclonazone,® Outdoor M1-E,® and Pres2Pure,® performed poorly, as coliforms and E. coli were detected in the treated water by the Colilert ® method. The chemical agents and the iodine resin straw performed poorly on raw river water; coliforms and E. coli were detected in the treated water. Conclusion: These data demonstrate the differences between the systems tested. The effectiveness of other devices on the market should also be tested, so as to help travelers and hikers select the most appropriate portable water treatment system.

In most industrialized countries, waterborne diseases are at a relatively low level, even if outbreaks still occur as reported by national surveillance systems that exist in the USA,1 in the United Kingdom2 or in Finland.3 The pathogens that have been targeted as being the most important in the West are the human enteric viruses and the protozoans Giardia lamblia and Cryptosporidium parvum, because they are resistant to the disinfection with chlorine. Bacterial pathogens are sensitive to disinfection, and cholera and typhoid fever are good examples of water-

borne disease controlled in industrialized countries by sanitation and water treatment during the 20th century. In many parts of the world, the water is not safe to drink. Poorer areas lack sewage collection and water treatment facilities, the maintenance of the systems is bad, and unprotected water sources are heavily polluted. Waterborne bacterial disease outbreaks involved pathogens such as Vibrio cholerae, Shigella, Salmonella typhi, Campylobacter jejuni, Escherichia coli (enterotoxigenic E. coli, E. coli O157:H7), or Yersinia, and the impact of water supply and water quality on health have been investigated in numerous studies in the developing world.4 Travelers from industrialized countries are not immune to many infectious fecal organisms, so they are at risk from waterborne infection when they go to developing countries. Hikers and campers in wilderness areas are also exposed to waterborne disease if they drink raw water from rivers and lakes. Giardiasis accounts for most cases of diarrhea in backpackers,5–7 but Campylobacter enteritis infection has also been reported among hikers who drank untreated surface water,8 and among campers who drank untreated water from wells.9 Therefore, safe water is a priority for travelers. They are usually advised to drink bottled water, especially carbonated water.10 Boiled water (a rolling boil) is probably the safest.11 This concern about waterborne disease

O. Schlosser, MD: Department of Medical Services, Vivendi, Paris, France; C. Robert, BSc, and M.R. de Roubin, PhD: Department of Microbiology, Anjou Recherche, Vivendi Water Central Laboratory, Saint-Maurice; C. Bourderioux, MD: Travel Clinic, Hopital Bretonneau, Tours; M. Rey, MD: Société de Médecine des Voyages, Paris. This paper was presented at the Sixth Conference of the International Society of Travel Medicine, Montreal, Canada, June 6–10, 1999. Reprint requests: Dr Olivier Schlosser, Vivendi, 42 avenue de Friedland, 75380 Paris Cedex 08, France. J Travel Med 2001; 8:12–18. 12

S ch l o s s e r e t a l . , B a c t e r i a l R e m o v a l f r o m I n e x p e n s i v e Po r t a b l e Wa t e r Tr e a t m e n t S y s t e m s

has led to the development of portable water treatment systems for travelers; they may be chemical agents, iodine resin purifiers, or filters. Sport and mountain goods stores, pharmacies and commercial web sites sell devices at various prices and of differing quality. Information about the real efficacy of these systems in the field is often lacking. We have, therefore, estimated the capabilities of several inexpensive portable personal water treatment systems for travelers to remove indigenous bacteria from water of varying quality. We have tested only the removal of bacteria, without regard to viruses or parasites.

Methods Systems

Four chemical agents, two iodine resin purifiers, and four filters were evaluated. These devices were selected from the products available on the French market on the basis of their low purchase price, under US $140. Each device tested was new. The four chemical agents were (Table 1): three chlorine compounds (Drinkwell chlorine,® Aquatabs,® and Hydroclonazone®), and 2% iodine in ethanol. The dose and the contact time of the chlorine compounds were those in the manufacturers’ instructions. The 2% iodine in ethanol doses and contact times were those recommended by the CDC.12 Two iodine resin purifiers were tested. One was the PentaPure® straw Outdoor M1-E (WTC/PentaPure Inc., West St. Paul, USA), which has a prefilter, an iodine resin element, and an activated charcoal bed. The capacity is 475 liters or 9 months of use whichever comes first. The other was the PentaPure® Traveler purification and filtration system, which is fastened to the faucet. This system has a prefilter, an iodine resin, an activated charcoal bed, and a 1 µm filter. The maximum flow

13

rate is 1 L/min and the capacity is 1000 liters or 12 months of use, whichever comes first. The four portable water filters were a flexible bottle (Pres2Pure®),and three hand-pump filters (Mini Ceramic,® First Need Deluxe,® and WalkAbout®) (Table 2). Each system was operated according to the manufacturer’s instructions. System Testing

The water used for tests was taken from the river Marne, in a Paris suburb. Two types of samples were collected, raw river water (turbid surface water), and settled, sand-filtered water from a water treatment plant (clear water). The samples were collected in 5 liter polypropylene containers sterilized by gamma irradiation. The samples were transported to the laboratory in an ice box and brought to 25°C (the temperature used for the trials) in the laboratory.Trials were run as soon as the water reached 25°C. The chemical disinfectants were tested using sterile flasks filled with 2 liters of well homogenized sample (volume required to make all the measurements after disinfection). The disinfectant was added, correctly mixed, and left in contact with the sample for the recommended time. The product was then neutralized with sodium thiosulfate and the sample analysed. Water samples were aspirated through the iodine resin purifiers using a low speed pump equivalent to buccal aspiration or tap water flow, and collected in sterile flasks for analysis. Care was taken not to recover the first part of the aspiration. Residual iodine was neutralized with sodium thiosulfate. Water samples treated with portable filtration systems were collected for analysis in sterile flasks after discarding the first part of the pump output. Samples of raw river water were evaluated with iodine resin purifiers and filters last, to minimize system contamination and clogging. Systems which gave poor

Table 1 Characteristics of Chemical Agents Name Drinkwell chlorine® Aquatabs® Hydroclonazone® 2% Iodine in ethanol

Manufacturer MS Water GmbH Elsau, Switzerland Medentech Ltd., Wexford, Ireland Promedica La Chaussée Saint-Victor, France Local pharmacy

*As recommended by the manufacturer † In clear water ‡ In turbid water § As recommended by the CDC tab = tablet; d = drops.

Active Chemical 25 mg/mL Sodium hypochlorite 3.5 mg Sodium dichloroisocyanurate

Dose

Contact Time

3 d/L* 1 tab/L*

60 min* 30 min*

12.2 mg Tosylchloramide (chloramine) 1 tab/L* 2% Iodine in ethanol 5 d/L†§ 10 d/L‡§

60 min* 30 min§ 30 min§

14

J o u r n a l o f Tr a v e l M e d i c i n e , Vo l u m e 8 , N u m b e r 1

Table 2 Characteristics of Filters (Manufacturer’s Data)

Name

Manufacturer

Operating Mode

Filter Type

Pres2Pure®

CrystalPure® USA

Flexible bottle

2 µm Porous plastic impregnated with powdered activated charcoal and other absorbant media

Mini Ceramic® Katadyn Produkt AG Wallisellen, Switzerland Hand pump First Need General Ecology, Inc. Hand pump Deluxe® Exton, USA and gravity WalkAbout® SweetWater® Longmont, USA Hand pump

Pumping Rate (L/min)

Capacity (L)

NA

750

Prefilter, 0.2 µm ceramic Ag-impregnated Prefilter, 0.4 µm structured matrix, electrokinetic action

0.5

7000

1.7

400

Prefilter, 0.2 µm labyrinth depth

0.7

380

NA = not applicable.

bacterial inactivation or removal when tested with sandfiltered water were not tested with raw river water. The chemical and bacterial parameters of the water samples were tested before and after treatment. The potentiometric method with a glass electrode (AFNOR standard T 90-008) was used to measure pH. Turbidity was determined according to AFNOR standard T 90033. Total organic carbon (TOC) was determined according to European standard NF EN 1484. Bacterial inactivation or removal was evaluated three times for each system, at 25°C. Bacteria were counted by three methods:

yellow and fluorescent blue wells were then counted and cell counts obtained from a table. 3. Viable but not culturable bacteria. Viable bacteria (bacteria with respiratory capacity) were counted after filtering the samples through a black polycarbonate membrane with 0.2 µm pores. The bacteria were then labelled at 20°C for 2 hours by incubation in 2 mM CTC (5-cyano-2,3-ditolyl-tetrazolium). The reagent acts as a terminal respiratory electron acceptor, producing crystals which fluoresce red under light. Fluorescent bacteria were counted under an epifluorescence microscope.

1. Conventional culturing techniques. Coliforms and thermotolerant coliforms and cells were counted with a membrane filtration method (AFNOR standard NF T 90-414), and counts of Enterococci used AFNOR standard NF T 90-416. Undiluted or diluted samples were filtered through a 0.45 µm pore membrane deposited on a selective agar medium. Two membranes were used for each sample; one was incubated for 24 to 48 h at 37°C to detect coliform bacteria, the other was incubated at 44°C for 24 to 48 h to detect thermotolerant coliform bacteria. Typical colonies were then counted and confirmed by reseeding. 2. Colilert ® /Quantitray ® method. This enzymatic method counts coliform bacteria and Escherichia coli by measuring the ß-D-galactosidase activity in all these bacteria, and the ß-glucuronidase activity in E. coli. The Colilert® reagent contains ONPG (orthonitrophenyl-galatoside), which gives a yellow color in response to the coliform bacteria enzymes, and MUG (methyl-umbelliferyl-glucuronide), which fluoresces blue (365nm) in the presence of E. coli. The Colilert® reagent was added to the concentrated or diluted sample (100 mL) and homogenized. One Quantitray® was filled, sealed and incubated for 24 h at 37°C. The

Results The chemical and bacterial characteristics of the water samples are shown in Table 3. Raw river water samples were moderately (10.4 nephelometric turbidity unit [NTU]) to quite turbid (52.3 NTU) after a high rainfall. Sand-filtered water was clear and consequently had a lower turbidity and less organic matter. Nevertheless, the coliform concentration was significant and E. coli was present. Thus sand-filtered water could not be considered to be safe drinking water. Tests on Clear Water

Conventional culturing techniques showed total coliforms and thermotolerant coliforms when clear water samples were treated with Hydroclonazone® or with the bottle Pres2Pure.® These results were confirmed by the detection of coliforms and/or E. coli using the Colilert®/Quantitray® method (Table 4). These two systems were therefore not tested with turbid water. No Enterococci were detected in any treated water sample. The inactivation or removal of viable bacteria (CTCstaining method) is shown in Table 5. The best results were obtained with the Mini Ceramic® filter, with a 3.25


15

Table 3 Chemical and Bacterial Characteristics of Water Samples

Turbidity (NTU) TOC (mg C/L) pH Total coliforms (/ 100 mL) Thermotolerant coliforms (/ 100 mL) Enterococci (/ 100 mL) E. coli, Colilert®/Quantitray® ( / 100 mL) Viable bacteria, CTC (/ 100 mL)

Raw River Water

Sand-Filtered Water

10.4–52.3 2.0–3.3 8.10–8.15 19,000–159,000 1,800–13,900 390–4,350 130–3,600 5.2 106–10.4 106

0.15–0.55 1.7–2.0 7.80–8.00 37–160 3–17 0–1 0–20 1.38 105–1.45 105

NTU = nephelometric turbidity unit; TOC = Total Organic Carbon.

log (99.94%) removal of viable bacteria in each experiment. The First Need Deluxe® removed more than 2.5 log (99.7%) from each clear water sample. The WalkAbout® gave a 3 log removal in the first test, but the device released viable bacteria after this first test, due to system contamination. It was therefore cleaned, disinfected with bleach and thoroughly rinsed before the next test. The first two tests with the Outdoor M1-E® indicated a removal of less than 0.5 log, with considerable interfering bacteria in the effluent that prevented coliform counting. The system was disinfected with bleach, and the third test gave a removal of 1.32 log, which is the only value recorded. Tests on Turbid Water

The hand pump filters decreased the turbidity of raw river water greatly (> 95%), down to 0.15–0.30 NTU with the Mini Ceramic® and the WalkAbout,® and a little less with the First Need Deluxe,® which gave 0.90–2.15 NTU from 42.70 NTU for raw water samples. The Outdoor M1-E® iodine purifier straw only decreased turbidity by 38.4%, and the turbidity of the effluent from 52.30 NTU raw water was still 38.60 NTU. Conventional culturing techniques showed total coliforms and thermotolerant coliforms when turbid

Table 4 Colilert®/Quantitray® Results Coliforms and/or E. coli in Treated Water Drinkwell chlorine® Hydroclonazone® Aquatabs® 2% Iodine in ethanol Outdoor M1-E® Traveler® Pres2Pure® Mini Ceramic® First Need Deluxe® WalkAbout®

Clear Water

Turbid Water

No Yes No No No No Yes No No No

Yes NT Yes Yes Yes NA NT No No No

NA = not applicable; NT = not tested

water samples were treated with chemicals, with the straw Outdoor M1-E ® and, one time, with the WalkAbout® filter (2 total coliforms/100 mL in the effluent from a sample containing 57,000 total coliforms/100 mL). The effluent after treatment with Aquatabs,®= 2% iodine-ethanol and the straw Outdoor M1-E,® still contained Enterococci. The chemical agents and the Outdoor M1-E® straw did not inactivate the coliforms and E. coli in turbid water, as detected by the Colilert®/Quantitray® method (E. coli up to 12.4/100 mL in the effluent from the Outdoor M1-E® straw). But neither coliforms nor E. coli were detected in the effluents from the three hand pump filters (see Table 4). This performance was confirmed by the removal of more than 3 log (99.9%) of viable bacteria using the CTC-staining method (see Table 5). There was a clear relationship between the turbidity of the untreated water and the performance of the chemical agents and the straw Outdoor M1-E,® the more turbid the water, the fewer bacteria were inactivated (Table 6). Discussion Chemical agents, iodine resin purifiers and filters were tested with clear and turbid water, to determine their performance with different qualities of water, that might be encountered when staying in hotels or trekking in wilderness areas. River water was used because it contains indigenous strains of bacteria rather than strains from laboratory collections. Many of the bacteria in aquatic systems can enter a starvation state when stressed by nutritional deficiency or other environmental factors (pH, temperature, oxidants, etc.).13 These stressed bacteria may be biologically active (viable) but will not be detected on growth medium. They are viable but not culturable, but retain their ability to infect humans, animals, or plants. Stressed bacteria are smaller than normal (usually forming small spheres) and resistant to a wide range of antibacterial substances, including disinfecting agents. Thus, stressed indigenous bacteria may be 10 to

16


Table 5 Inactivation or Removal of Viable Bacteria (CTC Method) Clear Water

Turbid Water

System

Concentration in Treated Water (/100 mL)

% Reduction

Log Reduction

Concentration in Treated Water (/100 mL)

% Reduction

Drinkwell chlorine® Hydroclonazone® Aquatabs® 2% Iodine in ethanol Outdoor M1-E® Traveler® Pres2Pure® Mini Ceramic® First Need Deluxe® WalkAbout®

234–1,350 8,750–41,200 234–2,340 234–1,670 6,560 625–859 625–18,900 78–78.1 78–469 156–547

99.1–99.8 69.7–94.0 98.3–99.8 98.9–99.8 95.2 99.3–99.6 86.5–99.6 > 99.9 99.7- > 99.9 99.6–99.9

2.0–2.8 0.5–1.2 1.8–2.8 1.9–2.8 1.3 2.2–2.4 0.9–2.4 3.2–3.3 2.6–3.3 2.4–3.0

39.4 103–46.4 103 NT 96.8 103–814 103 87.5 103–232 103 176 103–3,889 103 NA NT 104–112 103 208–1,460 937–10.4 103

99.2–99.6 90.9–99.0 97.4–99.1 56.6–98.1 97.8 –> 99.99 > 99.9 –> 99.99 ≥ 99.9

Log Reduction 2.1–2.4 1.0–2.0 1.6–2.1 0.4–1.7 1.7–4.9 3.9–4.7 2.9–3.8

NA = not applicable; NT = not tested

100 times more resistant than bacteria from laboratory collections.14 The indigenous viable bacteria may be counted by direct epifluorescent microscopy after staining them with CTC fluorochrome.15 In addition to counting culturable bacteria by conventional techniques (which may lead to overestimation of the efficacy of the systems, particularly chemical agents), we therefore used CTC staining to count viable bacteria, and so obtain a better assessment of the efficacy of the systems. Because we enumerated viable indigenous bacteria without the culture step, the results of this study appear to indicate a lower performance of the treatment systems than most published data based on conventional culture techniques to enumerate laboratory grown bacteria16,17 or indigenous bacteria.18,19 It is not possible to examine water for the presence of all possible pathogens. The use of indicator organisms is universally accepted for assessing the hygienic quality of water and the efficacy of disinfection. Thermotolerant coliforms are commonly used indicators of fecal pollution for routine purposes. Among this group, E. coli is regarded as the indicator of choice for assessing fecal contamination. Total coliforms are indicators of the treatment efficacy and allow detection of regrowth phenomena in the distribution network.

The European Community regulation (98/83/EU directive) states that safe drinking water must not contain any E. coli or Enterococcus in 100 mL, and the WHO guidelines for drinking water quality recommend no E. coli and no thermotolerant coliforms in 100 mL of drinking water samples.20 Because enteric viruses, protozoan cysts and helminths eggs are known to be more resistant to disinfection than E. coli and Enterococcus, the absence of the latter organisms will not necessarily indicate freedom from the former. Similarly, because viruses are much smaller than bacterial indicators, the absence of the latter in effluent from submicron filtration devices will not necessarily indicate the absence of viruses. Further research is needed to assess the removal of viruses and parasites with the tested treatment systems. This study demonstrates the differences in efficacy of the systems. The best systems were the three hand-pump filters, Mini Ceramic,® First Need Deluxe,® and WalkAbout,® which removed 3 log (99.9%) or more bacteria, with no coliforms or E. coli detected in the effluent. The efficacy of a filtration device depends on the porosity of the cartridge and on the quality of the overall device, particularly the internal seal. All three devices had submicron

Table 6 Water Turbidity and Removal of Viable Bacteria (CTC Method) Turbidity Log Removal (CTC) Drinkwell chlorine® Aquatabs® 2 % Iodine in ethanol Outdoor M1-E® NTU = nephelometric turbidity unit.

0.5 NTU

10.4 NTU

16.4 NTU

52.3 NTU

2.8 2.8 2.8 1.3

2.4 2.0 2.1 1.7

2.1 1.7 1.8 1.0

2.3 1.0 1.6 0.4


filtration elements (0.2 to 0.4 µm, according to the manufacturers). This should remove most bacteria, as the majority are larger than 0.5 µm. More bacteria were removed from the raw water samples than from the sand-filtered water samples, probably because of adsorption of the bacteria to the particles in the turbid water. The 2 µm porous plastic filter of the bottle Pres2Pure® did not remove coliforms from the clear water samples, as detected by conventional culturing, and E. coli was still present in the effluent. The iodine resin straw, Outdoor M1-E,® did not remove more than 1.75 log of viable bacteria, and coliforms and E. coli were still present in effluent from raw river water. This poly-iodine resin purifier is a contact demand-release disinfecting system. Bacteria in the water are attracted to the resin by electrostatic force and contact with the resin bead causes iodine to be released and inactivate the bacteria. Most micro-organisms in turbid water are adsorbed on the particles. Thus, some bacteria do not contact the iodine resin, and are not killed. Despite the prefilter, the turbidity was only slightly decreased, which is consistent with the bacterial results. Cloudy water should be prefiltered through two coffee filters or a clean cloth to decrease turbidity before using iodine resin purifiers which do not include a filtration element. As a rule, water should be clarified before pumping through filters and systems, including filtration and iodine resin purification, to maximize the performance and avoid premature clogging of the system. Most devices are fitted with a prefilter, but some of them are designed to remove only very large particles. Others are finer, but clog more quickly. The faucet fastening system, PentaPure® Traveler, removed coliforms and E. coli from clear water, but it did not remove 3 log of viable bacteria, according to the CTC method, unlike the hand-pump filters. As expected, the chemical agents performed very poorly with raw river water because the active agent was consumed by the organic and inorganic matter and by the suspended particles. It should be possible to increase the efficacy of the chemical agents by using a larger dose and/or a longer contact time, because the efficacy of the chemical agents depends on the product Concentration Time. In practice, the concentration is limited by the toxicity and the bad taste of the chemical and its by-products, and the contact time is limited by the impatience of the traveler. The ability of chemicals to inactivate bacteria varies inversely as the water turbidity. The best solution in this case is to remove particles and organic matter by filtration (e.g., coffee filters) before disinfection. Drinkwell chlorine,® Aquatabs,® and 2% iodine in ethanol performed the same as the PentaPure® Traveler

17

with clear water samples, but these chemical agents all required a certain contact time. Hydroclonazone® did not inactivate coliforms and E. coli in clear water with a contact time of 60 min. The manufacturer’s recommended contact time is 1–2 hours. We did not test Hydroclonazone® with a 2 hour contact time, for the reason given above. The bactericidal activity of chemical agents, especially chlorine compounds, varies greatly with the pH, organic content, and temperature of the water to be treated. The pH, TOC, and turbidity of the water samples were determined and experiments were performed at 25°C. Chemicals would have inactivated fewer bacteria in cold water samples. Hence, contact times should be increased with cold water. Iodine resin purifiers and filters should be effective for the designed operational life of the equipment. The devices we tested were new and experiments were only performed at 0% of their lifetime. The good results with the three hand-pump filters should therefore be confirmed after varying intervals. In other respects, devices with a cleanable filter element can be contaminated while cleaning, and an internal seal can be damaged when removing and replacing the filter cartridge. Great care must be taken when following the manufacturer’s instructions. Acknowledgments The authors thank Vivendi Water that supported the laboratory water samples analysis, and the French Société de Médecine des Voyages which supported the purchase of the tested products. The authors declare no interest with the tested products that might be construed as a potential conflict of interest. References 1. CDC Surveillance Summaries. Surveillance of waterborne disease outbreaks—United States 1995–1996. MMWR Surveillance Summaries 1998; 47(SS-5):1–34. 2. PHLS Communicable Disease Surveillance Center. Surveillance of waterborne disease and water quality. Commun Dis Rep 1999; 9:305–308. 3. Lahti K, Hiisvirta L. Causes of waterborne outbreaks in community water systems in Finland: 1980–1992. Wat Sci Tech 1995; 31:33–36. 4. Hunter PR. Waterborne disease—Epidemiology and ecology. Chichester: John Wiley and Sons, 1997. 5. Barbour AG, Nichols CR, Fukushima T. An outbreak of giardiasis in a group of campers. Am J Trop Med Hyg 1976; 25:384–389. 6. Wright RA, Spencer HC, Brodsky RE, Vernon TM. Giardiasis in Colorado: an epidemiologic study. Am J Epidemiol 1977; 105:330–336.

18


7. Kettlewell JS, Bettiol SS, Davies N, et al. Epidemiology of giardiasis in Tasmania: a potential risk to residents and visitors. J Travel Med 1998; 5:127–130. 8. Taylor DN, McDermott KT, Little JR, et al. Campylobacter enteritis from untreated water in the Rocky Mountains. Ann Intern Med 1983; 99:38–40. 9. Stehr-Green JK, Nicholls C, McEwan S, et al. Waterborne outbreak of Campylobacter jejuni in Christchurch: the importance of a combined epidemiologic and microbiologic investigation. N Z Med J 1991; 104:356–358. 10. Harris JR. Are bottled beverages safe for travelers ? Am J Public Health 1982; 72:787–788. 11. Backer HD. Effect of heat on the sterilization of artificially contaminated water [Editorial]. J Travel Med 1996; 3:1–4. 12. Treatment of water. In: Centers for Disease Control and Prevention. Health Information for International Travel 1999–2000, Atlanta, GA: DHHS, 165–166. 13. Jones JG. The micobiological quality of water: the nature of the problem. J Water SRT-Aqua 1997; 46:346–352. 14. Joret JC, Mennecart V, Robert C, et al. Inactivation of indigenous bacteria in water by ozone and chlorine. Wat Sci Tech 1997; 35:81–86.

Venice, Italy. Submitted by Charles D. Ericsson, MD.

15. Cervantes P, Mennecart V, Robert C, et al. Persistence of viable but not-culturable bacteria during the production and distribution of drinking water. In: Sutcliffe DW, ed. The microbiological quality of water. Ambleside UK: Fresh Biological Association, 1997:55–62. 16. Naranjo JE, Chaidez Cl, Quinonez M, et al. Evaluation of a portable water purification system for the removal of enteric pathogens. Wat Sci Tech 1997; 35:55–58. 17. Gerba CP, Naranjo JE. Microbiological water purification without the use of chemical disinfection. Wilderness Environ Med 2000; 11:12–16. 18. Galvan M, de Victorica J. Assessment of a water filtration device for household use in rural communities in Mexico. Wat Sci Tech 1997; 35:65–69. 19. Baylac P, Sere O, Wanegue C, et al. Comparison of the disinfecting power of chloramines T and sodium dichloroisocyanurate on river water. Rec Med Vet 1996; 173:391–399. 20. Microbiological aspects. In: Guidelines for drinking-water quality—2nd Ed. Vol 1. Geneva: World Health Organization, 1994:8–29.