Epidemiology of wilderness-acquired diarrhea: implications for prevention and treatment

Journal of Wilderness Medicine 3, 241-249 (1992) ORIGINAL ARTICLE Epidemiology of wilderness-acquired diarrhea: implications for prevention and treatment STEVEN C. ZELL, MD, FACP* Division ofgeneral Internal Medicine, University ofnevada School ofmedicine, Nevada, USA The popularity of backcountry travel has brought to public attention concerns over the safety of wilderness water consumption. Epidemiologic studies are limited, but place the incidence of wilderness-acquired diarrhea (WAD) in the range of 3-5%. Although the United States Forest Service and National Park Service focus efforts on disinfecting water to prevent giardiasis, the pathogen may only account for a small proportion of diarrheal cases in backcountry travelers. Stool specimen analysis from symptomatic recreationalists suggests bacteria and viruses to be more frequent isolates. A reasonable estimate of etiology can be made based upon the expedition length. On short trips (less than 7-10 days), symptomatic diarrhea will be due to viral or bacterial pathogens and can be treated empirically with agents proven of benefit for management of traveler's diarrhea. For longer trips, illness attributable to Giardia lamblia complicates empiric therapy. Backcountry travelers should filter water to remove protozoan cysts. This should be followed by halogenation to disinfect against bacterial and viral pathogens. Key words: camping, diarrhea, Giardia Introduction In the past decade, backcountry wilderness travel has become a popular pastime. Heightened public awareness of health issues unique to wilderness travel has evolved, and backcountry enthusiasts share a common concern over the safety of wilderness water consumption [1]. Currently, warnings fostered by the United States Forest Service (USFS) and National Park Service (NPS) are creating hydrophobia in backcountry travelers [2]. Misplaced concerns about the safety of wilderness water have even led to anecdotes of dehydration in recreationalists at high altitude, requiring emergent medical evaluation. The USFS and NPS recommendations to disinfect all wilderness water are measures directed at preventing acquisition of giardiasis. Other agents, such as bacteria and viruses, playa significant role in water-borne diarrheal illness [3]. Finally, recent literature suggests that manufacturer's recommendations for water disinfection devices and chemical treatments may not be completely effective, promoting unfounded confidence in backcountry travelers [4,5]. The identification of Giardia lamblia cysts in national parks and forests [6] has fostered recommendations from a vast community of non-medical authorities [7]. Wilderness medicine, an exciting new discipline, has yet to address thoroughly the issue Address for correspondence: Division of Internal Medicine, Department of Medicine, Washoe Medical Centre, 77 Pringle Way, Reno NY 89520, USA 0953-9859 1992 Chapman & Hall

242 Zell of wilderness acquired diarrhea (WAD), as has been done for traveler's diarrhea [8]. However, information exists in both the medical literature and governmental agency publications to allow sensible recommendations for directing future studies regarding prevention and treatment of WAD. In this paper, the author reviews the epidemiology of diarrheal illness in backcountry travelers. The role of Giardia lamblia in WAD is highlighted and information regarding the parasite's level of contamination is related to its risk of establishing infection. Medical literature investigating the efficacy of various manufacturers' water disinfection methods is cited. Finally, alternatives to the traditional practice of water disinfection are proposed and the rationale for directing study into such new approaches is presented. Epidemiology Unfortunately, little data exist regarding the prevalence of WAD in backcountry travelers. Most epidemiologic investigations of water-borne diarrheal illness relate to surface water supplies intended for municipal usage [3]. Information on backcountry travelers acquiring WAD is limited, since most studies rely upon observation and fail to adhere to the principles basic to epidemiologic research [1,2,9-11]. Problems precluding an accurate determination of WAD prevalence include inability to follow-up longitudinally travelers from a particular designated wilderness [9], reliance on anecdotal information and personal observations [2], and failure to confirm etiologic agents by stool analysis [1]. Making accurate conclusions regarding the role of specific pathogens in WAD would require considerable effort. Implicating an infectious agent would necessitate baseline prevalence data on backcountry travelers compared to acquisition rates after exposure to a defined wilderness area. An unbiased longitudinal follow-up of backcountry travelers would require a high degree of subject compliance. Individuals without a diarrheal illness might not chose to participate in longitudinal follow-up, leaving a disproportionate number of travelers suffering a gastrointestinal illness to report spuriously elevated rates of WAD. Non-anecdotal information would be critical, with collection and testing of travelers' stool specimens. Finally, isolation of infectious agents from backcountry waters that correlate with stool specimen cultures obtained in symptomatic persons [12,13] would strengthen a causative association. Unfortunately, there has yet to be a study rigorously applying the above to WAD. In addition, the medical community has been reluctant to make giardiasis a reportable disease. Prior to 1980, only eight states listed it as reportable, but by 1987, 38 states had done so [14]. Even physicians contribute to this problem by empirically treating WAD cases for giardiasis without demanding laboratory confirmation or reporting cases to communicable disease authorities [2]. What evidence exists that backcountry travelers acquire WAD, specifically giardiasis? Barbour and colleagues reported a laboratory-documented outbreak of giardiasis among Brigham Young University (BYU) students traveling in the Uintas Mountains of Utah. Seventy-nine percent (22 of 28) of symptomatic campers' stools contained Giardia lamblia cysts [10]. This observation prompted a larger prospective study of 160 additional BYU students who received a follow-up trip questionnaire and underwent stool specimen analysis. Within this group, the prevalence of Giardia cysts was found to be 14.4%. However, a retrospective survey of a more heterogeneous group of 203 backcountry travelers was completed to determine acquisition of a diarrheal illness

Wilderness diarrhea - prevention and treatment 243 compatible with giardiasis. Approximately 4.5% experienced an illness satisfying a case definition of giardiasis, but only two-thirds of those contacted responded. Assuming the latter one-third of backcountry travelers were not compelled to do so because of a lack of illness, the adjusted incidence rate of WAD in this large heterogeneous group may rest between 3-4.5%. Information regarding WAD from governmental agencies such as the NPS and USFS exists but is limited in scope. Cowdin studied the behavior of recreationalists in Rocky Mountain National Park to survey hikers regarding water consumption habits [1]. A follow-up telephone survey of backcountry travelers was also completed to ascertain their incidence of diarrheal illness. Of 556 persons surveyed, only 111 (20%) drank water from a natural source, as most were day hikers and carried their own water. Follow-up of individuals consuming wilderness water revealed that only four persons experienced post-trip diarrheal illness, for an attack rate of 3.0%. The implication was made that cases were due to giardiasis, in light of the cysts being recovered at various lakes in the watershed [15]. However, laboratory confirmation of Giardia lamblia infection from travelers was not obtained, limiting conclusions to be drawn from this study regarding the parasite. In addition, despite extensive filtration per site (100 gallons), cyst recovery numbers were only in the single digit range (0-4 cysts per site) [15]. Regardless of the pathogen, this study's incidence rate for WAD of 3% is in line with a prior study in the medical literature [10]. With limited studies on WAD, epidemiologic surveillance systems serve as a valuable source of information. Disease transmission patterns from non-wilderness outbreaks can be identified and applied to describe characteristics of water at risk for consumption. The state of Vermont has provided invaluable insight into giardiasis, as its population has the highest annual incidence in the United States (45.9 cases per 100000). A recently published three-year study helps to identify characteristics of ground water associated with an increased relative risk of cyst acquisition in case contacts. These features include: (1) an increasing rate of infection in association with higher elevation; (2) a higher incidence of cyst acquisition from untreated ground water; and (3) a clustering of cases during summer months when outdoor activities bring more people into contact with water [16]. Information regarding the relationship of Giardia lamblia to other etiologic agents causing WAD is of interest. A study from the Grand Teton National Park encouraged all persons with an acute gastrointestinal illness to attend a free clinic, where stool specimens were processed for Salmonella, Shigella, Campylobacter and protozoan cysts. Of 178 persons experiencing a diarrheal illness, Campylobacter jejuni was the most common isolate, accounting for 23% of cases, whereas Giardia lamblia was detected in only 8% of individuals. Salmonella and Shigella were not isolated. Of note, 69% of cases had no identifiable pathogen, raising the question as to the role of viral infection in WAD [9]. Such data challenge the prevailing preoccupation with Giardia lamblia as the sole agent of concern in wilderness waters and provide evidence that bacteria and viruses may play a significant role in producing WAD. Collaborative data regarding outbreaks of water-borne diarrheal disease from the Centers for Disease Control and the Environmental Protection Agency validate the latter observations. Data from 1986-1988 regarding water-borne illness implicate Giardia lamblia as the most common pathogen, but always in association with community water supplies lacking adequate chlorination or filtration [3]. However, in reported outbreaks

244 Zell related specifically to recreational water usage, excluding Pseudomonas dermatitis from hot tubs and spas, most cases of water-borne diarrheal illness were due to viruses or were of unknown etiology. Shigella sonnei was the bacterium most frequently isolated as an etiologic agent in recreational water outbreaks; Giardia lamblia only accounted for two of the 26 reported episodes [3]. In summary, epidemiologic evidence suggests that WAD may be an infrequent occurrence. The attack rate for backcountry travelers is probably less than 10%, with scant data placing it in the 3-5% range. Emphasis on Giardia lamblia as a pathogen by the USFS and NPS is understandable, but other etiologic agents, such as bacteria and viruses, may playa greater role in causing WAD. Giardia lamblia: distribution and relation to infection Recent information quantifying the degree of Giardia cyst contamination in backcountry lakes and streams sheds light upon the parasite's role in WAD [17]. Albeit limited, prior research investigating the minimum innoculum of cysts necessary to establish infection in humans allows quantitative description of the acquisition risk imposed by Giardia upon backcountry travelers. In the 1950s, Rendtdorff [18] studied the innoculum of Giardia lamblia cysts necessary to establish infection in federal prison volunteers. Known numbers of cysts were placed along with a small amount of saline into gelatin capsules ingested by subjects. Stool specimens from prisoners were examined for up to 165 days after exposure and infection defined as ability to identify Giardia lamblia cysts in the subject's stool, regardless of symptomatology. No person exposed to one cyst became infected, whereas 36.4% of 22 men receiving from 10-25 cysts established infection. All subjects receiving 100 cysts or more developed infection. Of note, no individual developed symptomatic giardiasis and there was a tendency for infections to disappear spontaneously over time [18]. Research from the United States Geological Survey Office in Sacramento provides Table 1. Sampling for Giardia lamblia at selected sites, Sierra Nevada Mountains [17]. Site location Gallons filtered Number ofcysts Desolation Wilderness Susie Lake Inlet (L) 100 0 Susie Lake Outlet (H) 100 41 Pyramid Creek (H) 125 0 Ralston Lake Inlet (L) 100 2 Yosemite National Park Tuolumne River (H) 100 2 Lyell Fork (H) 100 2 McLure Creek (L) 100 0 Inyo National Forest Lone Pine Creek (H) 100 1 Guitar Lake Inlet (L) 60 0 Forrester Lake Outlet (H) 79 1 (H) Designates High Use Area (L) Designates Low Use Area

Wilderness diarrhea - prevention and treatment 245 insight into the relative contamination of wilderness waters in the forests of the western United States. Suk and colleagues [19] extensively sampled water for Giardia cysts at 69 sites in the Sierra Nevada mountains, including Yosemite and Sequoia National Parks. High use areas were identified by the USFS and NPS based upon visitor usage, water drainage characteristics and proximity of latrine sites to water courses. Giardia cysts were detected in 44.9% (22 of 49) of samples collected from sites designated as high use and in only 17.2% (5 of 29) samples obtained from areas oflow recreational use [19]. These data suggest that intense human usage may play a significant role in Giardia contamination of backcountry surface water. Table 1 gives a typical representation of attempts to isolate Giardia cysts in the backcountry lakes and streams. As mentioned before, cyst concentrations per 100 gallons of filtrate are usually low, in the single digit range [17]. However, laboratory analysis of the filtration method utilized in such studies [20] suggests that only 30% of Giardia cysts are recoverable. Utilizing the highest level of Giardia contamination at 40 cysts per 100 gallons [19], this could realistically be extrapolated to a concentration of 1.2 cysts per gallon of filtrate. In terms of importance to the backcountry traveler, this would translate into a cyst concentration of 0.3 cyst per quart, or a 1 in 33 chance of acquiring a minimum innoculum (10 cysts) from a canteen full of water. Water disinfection Although backcountry travelers acquire WAD at low rates, the decision to disinfect one's water source is an individual matter. However, new information regarding the efficacy and proper utilization of several water disinfection methods deserves mention. Epidemiologic evidence requires that efforts at water disinfection be directed towards enteric bacteria and viruses, as well as at protozoan cysts [21]. Available methods of water disinfection include: (1) halogenation, or the chemical disinfection of water using chlorine or iodine; (2) filtration to physically remove pathogens; and (3) pasteurization, the process of sterilization of water by heat. Although all are advertised to disinfect water, the methods have variable efficacy against the three classes of pathogens described above. By far, pasteurization is the most reliable method of water disinfection. Prolonged boiling is required for sterilization, but not for disinfection. Giardia is very sensitive to heat and rapidly inactivated within ten minutes in water at only 60-70 C [3,22]. Bacterial spores, resistant to heat, are fortunately not enteric pathogens. Thus, most bacteria in water arising from fecal contamination are killed within minutes at temperatures of 80-90 C [23]. Viruses are generally heat sensitive [24]. Halogenation has variable activity against the three classes of pathogens and is of reduced efficacy in cloudy or cold water [21]. Microorganisms vary in their susceptibility to chlorine or iodine. Bacteria are very sensitive, viruses intermediate and protozoan cysts can be relatively resistant. Nitrogenous waste react rapidly with halogens to reduce their disinfectant powers. Studies looking at manufacturer recommendations for various halogen preparations provide insights into their limitations. Ongerth and colleagues [4], suspended high concentrations of Giardia Lamblia cysts in cold spring-fed water and tested the ability of commercially available chlorine and iodine disinfectants to inactivate cysts. At concentrations and contact times (30 minutes) recommended by the manufacturers, no commercial halogenation method was able to produce 100% cyst in-

246 Zell activation. After an eight-hour contact time, chosen since it represents a likely overnight incubation period at a campsite, all of the iodine preparations were reliably able to achieve 100% cyst inactivation. Despite prolonged contact times, chlorine disinfectants did not achieve complete cyst inactivation. Other authors have echoed the same concerns over the inability of some halogen preparations to inactivate Giardia LambLia cysts reliably in cold and turbid water conditions [5,25,26], but have limited contact times to no more than 60 minutes. Ongerth's inability to inactivate Giardia cysts with chlorine, despite overnight contact times, may relate to the incredible concentration of cysts utilized (30000 ml- I ) in water samples. Caution should be recommended in the extrapolation of these data, as experimental conditions were not indicative of actual cyst concentrations in backcountry waters (30-40 cysts per 100 gallons filtrate). Filtration is an excellent method for cyst removal, but is variable for bacteria and unable to trap virus particles [21]. Recent testing of two commercially available filters, the First Need and Katadyn effectively removed Giardia cysts in both clear and turbid water. Ongerth and colleagures [4] demonstrated the ability of both filters to achieve 100% cyst removal from stock Giardia solutions prepared with spring run-off water. In addition, the filters continued to maintain their reliability after generating 90 liters of filtered water. Clinical recommendations Even though wilderness travel is seemingly associated with a low incidence of WAD, proper water disinfection is currently recommended to prevent acquisition of diarrheal illness in backcountry travelers. However, the cost effectiveness of this approach has not been addressed. Alternative approaches, such as empiric treatment of diarrheal illness as recommended for traveler's diarrhea (TD) [8], may be applicable towards WAD and worthwhile of study. Because the agents responsible for cases of WAD vary in incubation period, a reasonable estimation of etiology can be made based upon the prior length of expeditionary travel. During trips of short duration (less than 10 days), symptomatic gastrointestinal illness will be due to bacterial or viral pathogens and empiric treatment of diarrheal illness as utilized for TD is a sound alternative [27]. The rationale is founded upon two basic principles: (1) the self-limited nature of TD and (2) theoretical similarities between the etiologic agents responsible for cases of WAD and TD (bacteria and viruses) [9,29]. One should note that most cases of TD result from bacterial contamination of water and arise from infection with enterotoxigenic E. coli. This pathogen has not been definitely implicated in WAD, because the Grand Teton National Park Study chose to study the role of other bacterial pathogens. Regardless of differences among the bacterial pathogens identified in WAD and TD, the concept of fecally contaminated water producing illness by a similar group of bacteria, the Gram negative coliforms, is shared by both and supports treatment of WAD with empiric antibiotic therapy. For backcountry trips of short duration (less than 10 days), water disinfection is not an absolute necessity and a diarrheal illness can be managed empirically. In the absence of invasive symptomatology (no fever or bloody stools), anti-motility agents may be taken, of which loperamide has been the best studied and of proven efficacy [29]. Invasive symptoms indicate bacterial mucosal invasion, and such cases can be managed with 3-5 days of antibiotics, induding trimethoprim/sulfamethoxazole [30], ciprofloxacin [31] or

Wilderness diarrhea - prevention and treatment 247 norfloxacin [33]. Is such an approach reasonable for backcountry travelers? Albeit empiric therapy may be effective against bacteria and viral pathogens, it does not eradicate Giardia lamblia infection. For trips of short duration, giardiasis may occur rarely and not be expected to produce symptoms during travel. However, the concern exists of establishing infection and experiencing gastrointestinal symptoms weeks to months later. Thus, filtration to remove this parasite is recommended. Empiric pharmacologic treatment, effective at ameliorating illness from bacteria and viruses, only allows the backcountry traveler the freedom from utilizing pasteurization or iodination to disinfect his water. In contrast, for prolonged trips water disinfection is recommended utilizing pasteurization or the combination of filtration followed by halogenation to eliminate infectious agents. Empiric pharmacologic therapy during prolonged trips (greater than 10 days) has practical limitations. Symptomatic giardiasis can occur after a ten-day incubation [18]. Because this protozoan parasite requires specific antimicrobial therapy and produces vague symptomatology, it is unreasonable to expect the lay public to distinguish this infection from invasive bacterial or viral infections that require selection of other antimicrobial agents. Future studies investigating the cost effectiveness of antimicrobial prophylaxis directed against bacterial pathogens along with filtration to remove protozoan cysts would be an area ofstudy interest if WAD incidence rates were found to be in a range near that for TD. Also, antimicrobial prophylaxis as utilized in TD [28] without efforts at water disinfection needs to be studied, if Giardia lamblia is truly responsible for only a small number of WAD cases. Person-to-person transmission of enteropathogens still remains a concern. For example, humans are a key reservoir for Shigella infection, in which transmission occurs by close personal contact [35]. Thus, proper hygiene at the campsite during food preparation is necessary to limit diarrheal illness not preventable by water disinfection. In the future, research must better define the role that Giardia lamblia has in producing WAD. A pooling of sources regarding the percentage of TD cases due to protozoan parasites range from 0-14% [26,28,33] which parallels the 8% isolation rate for giardiasis noted in symptomatic persons in the Grand Teton National Park Survey [9]. If Giardia lamblia is an uncommon isolate and the incidence of WAD infrequent, it may not be cost effective to recommend filtration for recreationalists involved in short duration trips. Assuming that the First Need Filter (retail price $40.00), can be shared by a group of four persons, for every 100 backpackers, a total of $1000.00 would be spent for such a device. On the contrary, if recreationalists focused on empiric treatment directed at bacteria or viruses and experienced giardiasis at an attack rate of 5%, dollar expenditures for a single physician visit, ova and parasite analysis and drug therapy would be roughly equivalent (five persons at $220.00 per individual, totalling $1100.00). However, this latter figure assumes that all cases of WAD will be attributable to giardiasis and that dollar expenditures would be lower if this were not the case. Finally, regardless of the low risk of acquiring WAD, utilizing the empiric pharmacologic therapy for cases of diarrheal illness is not sufficient without management directed towards fluid replacement. Emphasis should be placed upon rehydration as the key treatment modality irrespective of the causative agent. Backcountry travelers should not proceed into wilderness areas unprepared against WAD. Despite the concern over WAD, it can be prevented or adequately treated, and should not preclude backcountry enthusiasts from enjoying wilderness forests and national parks.

248 References 1. Cowdin, N.P. Conceptions and behaviors of recreationalists regarding water quality in Rocky Mountain National Park. Proceedings of the National Wilderness Research Conference, Fort Collins, Colorado 1985; 241-4. 2. Welch, T.R and Parker, M.J. Hydrophobia in the Adirondacks. Adirondac 1988; June, 16-18. 3. Levine, W.e. and Craun, G.F. Water-borne disease outbreaks, 1986-1988. MMWR 1990; 39, 1-14. 4. Ongerth, J.E., Johnson, RL., Macdonald, S.C., Frost, F. and Stibbs, H.H. Backcountry water treatment to prevent giardiasis. Am J Public Health 1989; 79, 1633-7. 5. Jarroll, E.L., Bingham, AK. and Meyer, E.A Inability of an iodination method to destroy Giardia cysts completely in cold water. West J Med 1980; 132, 567-9. 6. Starr, R. Giardiasis confirmed at Olympic National Park. Park Science 1981; 2, 10. 7. Knotts, D.M. Purifying water in the wild. Sierra 1983; July-August, 57-9. 8. Dupont, H.L., Ericsson, e.d. and Johnson, P.e. Chemotherapy and chemoprophylaxis of traveler's diarrhea. Ann Intern Med 1985; 102, 260-1. 9. Taylor, D.N., McDermott, U.T., Little, J.R, Well, J.G. and Blaser, M.J. Campylobacter enteritis from untreated water in the Rocky Mountains. Ann Int Med 1983; 99, 38-40. 10. Barbour, AG., Nichols, C.R and Fukushima, T. An outbreak of giardiasis in a group of campers. Am J Trop Med Hyg 1976; 25, 384-9. 11. Wright, R.A, Spencer, H.e., Brodsky, RE. and Vernon, T. Giardiasis in Colorado: an epidemiologic study. Am J EpidemioI1977; 105, 330-6. 12. Kent, G.P., Greenspan, J.R, Herndon, J.L. et al. Epidemic giardiasis caused by a contaminated public water supply. Am J Public Health 1988; 78, 139-43. 13. Weniger, B.G., Blaser, M.J. and Gedrose, J. An outbreak of water-borne giardiasis associated with heavy water run-off due to warm weather and volcanic ashfall. Am J Public Health 1983; 73,868-72. 14. Jokubowski, W. Purple burbs and the filtration of drinking water supplies (editorial). Am J Public Health 1988; 78, 123-5. 15. Kunkle, S., Cowdin, N., Wilson, J., Grondin, J., Ricketts, T. and Flora, M. Field Survey of Giardia in streams and wildlife of the Glacier Gorge and Loch Vale Basins, Rocky Mountain National Park. Fort Collins, Colorado: National Park Service, 1985. Natural Resources Report Series, 83-5. 16. Birkhead, G. and Vogt, L. Epidemiologic surveillance for endemic Giardia lamblia infection in Vermont: the roles of water-borne and person-to-person transmission. Am J Epidemiol1989; 129,162-8. 17. Suk, T.J., Sorenson, S. and Dileanis, P.D. Map showing the number of Giardia cysts in water samples from 69 stream sites in the Sierra Nevada Mountains, California. Sacramento, CA: United States Geological Survey, Division of Water Resources, 1986. USGS Open File Request 86-404. 18. Rendtorff, R The experimental transmission of human intestinal protozoan parasites. Am J Hyg 1954; 59, 209-20. 19. Suk, T.J., Sorenson, S.K. and Dileanis, P.D. The relationship between human presence and occurrence of Giardia cysts in streams in the Sierra Nevada Mountains, California. J Fresh EcoI1987;4,71-75. 20. Sorenson, S.K., Riggs, J.L. and Dileanis, P.D. Isolation and detection of Giardia cysts from water using direct immunogluorescence. Water Resour Bull 1986; 22, 843-5. 21. Backer, H. Field water disinfection (editorial). lama 1988; 259, 3185. 22. Jarrol, E.L., Hoff, J.e. and Meyer, E.A Resistance of cysts to disinfection agents. In: Erlandsen, S.L. and Myers, E.A eds. Giardia and Giardiasis: Biology, Pathogenesis and Epidemiology. New York: Plenum Press, 1984; 311-28. 23. Joslyn, L. Sterilization by heat. In: Block, S.S. ed. Disinfection, Sterilization and Preservation. Zell

Wilderness diarrhea - prevention and treatment 249 Philadelphia: Lea and Febiger, 1983. 24. Scheid, R, Deinhardt, F., Frosner, G. et al Inactivation of Hepatitis A or B viruses and risk of iatrogenic transmission. In: Third International Symposium on Viral Hepatitis. Philadelphia; Franklin Institute Press, 1981: 627-8. 25. Jarroli, E.L., Bingham, AK. and Meyer, E.A Effect of chlorine on Giardia lamblia cyst viability. Applied Environ Micro 1981; 41, 483-7. 26. Zemlyn, S., Wilson, W.W. and Heliweg, P.A. A caution on iodine water purification. West J Med 1981; 135,166-7. 27. Dupont, H.L., Ericsson, C.D. and Johnson, P.C. Chemotherapy and chemoprophylaxis of traveler's diarrhea (editorial). Ann Intern Med 1985; 102, 260-1. 28. Traveler's diarrhea: National Institutes of Health Consensus Development Conference statement. JAMA 1985; 253, 2700-4. 29. Johnson, P.C., Ericsson, C.D., Dupont, H.L., Morgan, D.R, Bitsura, J.M. Comparison of loperamide with bismuth subsalicylate for the treatment of acute traveler's diarrhea. JAMA 1986; 255, 757-60. 30. Dupont, H.L., Galindo, E., Evans, D.G., Cabada, F.J., Sullivan, P. and Evans, DJ. Prevention of traveler's diarrhea with trimethoprim-sulfamethoxazole and trimethoprim alone. Gastroentero11983; 84, 75-80. 31. Ericsson, C.D., Johnson, P.C., Dupont, H.L., Morgan, D.R., Bitsura, J.M. and de la Cabada, F.J. Ciprofioxacin or trimethoprim-sulfamethoxazole as initial therapy for traveler's diarrhea. Ann Int Med 1987; 106, 216-220. 32. Scott, D.A, Haberberger, RL. and Thornton, S.A Norfloxacin for the prophylans of traveler's diarrhea in U.S. military personnel. Am J Trop Med Hyg 1990; 42,160-4. 33. Taylor, D.N., Houston, R, Shlim, D.R and Hyams, K.C. Etiology of diarrhea among travelers and foreign residents in Nepal. JAMA 1988; 268, 1245-8. 34. Carpenter, CJ. Shigeliosis. In: Wyngaarden, J.B. and Smith, L.H. eds. Cecil Textbook of Medicine. Philadelphia: W.B. Saunders, 1985: 1595-7.