Epidemic Diarrhea at Crater Lake from Enterotoxigenic Escherichia coli A Large Waterborne Outbreak MARK L. ROSENBERG, M.D.; JEFFREY P. KOPLAN, M.D.; I. KAYE WACHSMUTH, Ph.D.; JOY G. WELLS, M.S.; EUGENE J. GANGAROSA, M.D.; RICHARD L. GUERRANT, M.D.; and DAVID A. SACK, M.D.; Atlanta, Georgia; Charlottesville, Virginia; and Baltimore, Maryland
In June and July 1975, gastrointestinal illness occurred in more than 200 staff members and 2000 visitors to an American national park. It was characterized by prolonged diarrhea, cramps, nausea, and vomiting, lasted a median duration of 8 days, and was significantly associated with consumption of park water (P < 0.001), which had been contaminated by raw sewage. Enterotoxigenic Escherichia coli serotype 06:K15:H16 was isolated from 20 of 49 ill park residents and from the park's water supply, but not from 71 residents who had never been ill or had been well for at least 4 days. No other bacterial, viral, or parasitic pathogens were isolated from ill or well persons. This outbreak is the first waterborne epidemic of diarrheal illness shown to be due to enterotoxigenic E. coli, and this study documents one mode of transmission of this pathogen. This investigation also suggests the relative insensitivity of current methods for identifying persons infected with this organism, either by the culturing of randomly selected isolates or by measuring serologic responses.
ENTEROTOXIGENIC Escherichia coli, a recently recognized enteric pathogen with a worldwide distribution, is a cause of acute watery diarrhea, especially in developing countries, and is the most common cause of acute diarrhea in travelers (1-6). The disease caused by this organism varies considerably in severity, ranging from mild diarrhea to a seriously dehydrating cholera-like illness. Little is known, however, of the mode of transmission of the disease. Although enterotoxigenic E. coli have been isolated from food and water sources, consumption of such sources has not been directly related to human disease (4, 7, 8 ) . In the summer of 1975, an outbreak of diarrheal disease caused illness in more than 2000 staff and visitors who • From the Bureau of Epidemiology, Center for Disease Control, Public Health Service, U.S. Department of Health, Education, and Welfare, Atlanta, Georgia; Department of Internal Medicine, University of Virginia School of Medicine, Charlottesville, Virginia; and the Department of Medicine, Baltimore City Hospitals, Johns Hopkins University School of Medicine, Baltimore, Maryland.
714
drank contaminated water at an American national park. Using recently developed laboratory methods to test E. coli for enterotoxin production, we showed that enterotoxigenic E. coli had caused this waterborne common-source outbreak. Epidemiologic Investigation Crater Lake National Park is a 250-square-mile rectangular tract in southwestern Oregon, whose 2000-foot-deep lake attracts almost 3000 tourists a day during the summer (Figure 1). Reports of widespread gastrointestinal illness among resident park staff and their families in late June 1975 led to the initiation of this investigation on 7 July 1975. Epidemiologic and clinical information regarding gastrointestinal symptoms was obtained from three groups of people: [1] 320 ( 9 6 % ) of the 335 resident park personnel and family members completed questionnaires from 7 to 11 July; [2] 107 ( 1 2 % ) of the 900 visitors who had stayed overnight at the park from 4 to 9 July responded to an initial telephone survey on 9 and 10 July; [3] 5273 ( 6 3 % ) visitors who had stayed overnight at the park between 10 June and 11 July responded to a short questionnaire (without questions about specific symptoms) mailed out in September 1975. Illness (defined as diarrhea or vomiting after visits to the park) was reported by 90% (288/320) of the resident park personnel, 64% (68/107) of the visitors contacted by telephone, and 44% (2310/ 5273) of the visitors surveyed by mail. Illness was significantly associated with drinking park water (P < 0.001) among all groups except those visitors contacted by phone who had arrived between 7 and 9 July when water chlorine levels were being closely monitored (Table 1). Among park personnel the median duration of illness increased as daily water consumption increased. A majority of those ill experienced diarrhea, cramps, nausea, and vomiting (Table 2 ) . The median duration or illness for park personnel was 8 days. Although 72% of the personnel missed work during part of their illness, most employees, including food handlers, worked at least part of Annals of Internal Medicine 86:714-718, 1977
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19522/ by a Scott Memorial Library User on 08/02/2017
Table 1. Association between Illness and Consumption of Park Water
Group
Park personnel Visitors telephoned July 4-6 July 7-9 Visitors surveyed
Drank
Did Not Drink*
Significancet
111
Well
Attack Rate
111
Well
Attack Rate
269
17
94.1%
5
13
27.8%
P < 0.00001
42 18 1859
0 34 1652
100.0% 34.6% 52.9%
0 2 23
2 2 565
0.0% 50.0% 3.9%
P < 0.001 NS P < 0.00001
* Defined as drinking less than one glass of water per day. t By Fisher's 2-tail test.
the time they were ill. The median duration of illness reported through the visitors' mail survey was 6 days, and 38% of these ill persons were ill for 10 or more days. Water for the park came from a shallow spring and was chlorinated before distribution; however, there was no systematic monitoring of chlorine levels in the distribution system. Three of four water samples sent to the county health department in May 1975 for routine bacterologic analysis had been found to have coliform contamination ( > 1 per 50 ml) and had been judged unacceptable. When twice-daily monitoring of chlorine levels began on 7 July, it showed that water going to some parts of the park had no detectable chlorine, even though adequate levels were present in water at other areas. On the evening of 10 July, snow that had covered a manhole melted and revealed sewage overflow through the manhole, which was located about 2000 feet uphill from the springs that supplied the park's water. Direct contamination of the water supply with sewage was shown by fluorescein dye tests, and the park was officially closed the next morning. The park reopened on 1 August, and a second questionnaire was completed by 221 (62%) park personnel on 14 August, 2 weeks later. Fifteen people (6.8%) reported recurrences in August after having been well for at least 1 week; the median interval between the two episodes of illness was 22 days. While the park was closed, 109 employees had left the grounds to spend at least 2 days with family or friends who had not visited the park. The rate of subsequent illness (diarrhea or vomiting) in persons visited by ill personnel (31/335) was significantly greater (P 0.015) than that in persons visited by well personnel (5/160).
negative colonies were selected from MacConkey's agar (Difco Laboratories; Detroit, Michigan) plates and tested for toxigenicity. Production of E. coli heat-labile toxin was determined by the mouse adrenal tumor cell (11, 12) and Chinese hamster ovary (13) tissue culture techniques, production of heat-stable toxin by the infant-mouse assay (14). The toxigenic isolates were serotyped for O and H antigens (9) and found to be all the same serotype, 06:HI6. The remaining 87 swabs received were examined for
Laboratory Investigation
Rectal swabs or stool swab specimens were obtained from 120 of the park staff and resident family members who filled out clinical information questionnaires. Swabs were immediately placed in Cary-Blair transport medium (Baltimore Biological Laboratories; Cockeysville, Maryland) and shipped on ice to laboratories at the Center for Disease Control. The first 33 fecal cultures were screened for Salmonella, Shigella, and pathogenic vibrios by previously described methods (9, 10) and for Yersinia enterocolitica by culturing on Salmonella-Shigella agar (Difco Laboratories; Detroit, Michigan) incubated at 22 °C for 48 h. All were negative for these bacteria. From each of these 33 fecal cultures, five lactose-positive and two lactose-
Figure 1 . Crater Lake National Park. Rosenberg et al. • Escherichia coli Diarrhea
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19522/ by a Scott Memorial Library User on 08/02/2017
715
Table 2. Symptoms and Severity of Gastrointestinal Illness among III Park Personnel and Visitors, 13 June-11 July 1975
Symptom
All 111 Park Personnel " ~~ (/! = 288)
Diarrhea Cramps Nausea Vomiting Chills Headache Muscle ache Fever Bloody diarrhea Median duration of illness Median interval between arrival and onset of illness Changed daily pattern of activities Missed work Consulted physician Hospitalized
93% 77% 72% 68% 48% 47% 45% 41% 0.3%
Park Visitors Personnel with Contacted Enterotoxigenic by Escherichia coli Telephone („ = 20) (/i - 68) 100% 91% 95% 79% 75% 88% 80% 79% 45% 54% 80% 53% 60% 47% 50% 37% 0% 2%
8 days
11 days
6 days*
18 days
20 days
30 hours
90% 72% 10% 0%
94% 75% 13% 0%
85% 24% 27% 0%
* Most ill visitors were still ill when interviewed initially; median duration of illness among 2310 ill visitors who later replied to mail questionnaire was 6 days.
E. coli only. Five lactose-positive and two lactose-negative isolates from each specimen were screened with E. coli 06 antiserum. All 06-positive E. coli (and a sample of the 06negative isolates) were tested for toxin production in the adrenal cell, Chinese hamster ovary, and infant mouse assays. Enterotoxigenic E. coli that produced both heat-stable and heat-labile toxins was isolated from 20 (16.7% ) of the 120 people examined (Figure 2 ) . All enterotoxigenic E. coli were serotype 06:H16. A 34% sample (335/985) of 06-negative isolates was negative for toxin producers. Escherichia coli 06 organisms that were not enterotoxigenic were isolated from 15 (12.5%) people. Four enterotoxigenic isolates were tested, and all were found to possess K 15 antigen (9). Among the 120 persons from whom a rectal culture had been obtained, there was a strong association between having a culture positive for enterotoxigenic E. coli and having had diarrhea at the time of culturing (P < 0.000001 Fisher's 2-tail test). Positive cultures were obtained from 17 of the 40 persons who had diarrhea when cultured, and from three of the 80 persons who did not have diarrhea when cultured. The three positive cultures in this second group came from nine persons who had diarrhea 1 to 3 days previously; no positive cultures were obtained from 71 persons who had not been ill or who had been free of diarrhea for 4 or more days. One hundred and twelve E. coli isolates from 16 ill persons were tested for invasiveness by the Sereny test (16), and none were positive. No ova or parasites were found in fresh stool specimens from 15 acutely ill persons, which were examined by the Oregon Health Division laboratory. Frozen stool specimens from six acutely ill individuals 716
were examined for enteric viruses in four viral isolation systems: primary monkey kidney cells, rhabdomyosarcoma cells of human origin, human embryonic lung fibroblast cells, and newborn mice. These specimens were also examined for reoviruslike and parvoviruslike gastroenteritis agents by direct electronmicroscopy and immunoelectronmicroscopy (16, 17). All viral studies were negative. Two litres of raw water taken from the park springs on 10 July were examined and found negative for Salmonella, Shigella, pathogenic vibrios, and Y. enterocolitica. Twenty lactose-positive and five lactose-negative colonies were selected from each sample: additional isolates were selected from water samples collected at various points in the distribution system. Seventeen of 396 water isolates were 06 positive, and 14 of these were enterotoxigenic in all three assays: the enterotoxigenic isolates were all serotype 06:H16. Those 06-positive isolates that were not enterotoxigenic were negative in H I 6 antiserum. Additionally, 113 water isolates that were 06 negative were tested in the adrenal cell tissue culture assay; all were negative. All enterotoxigenic isolates from both patients and water were multiply sensitive when tested by the Kirby-Bauer method (18) against ampicillin, carbenicillin, cepholethin, chloramphenicol, colistin, gentamicin, nalidixic acid, neomycin, nitrofurantoin, streptomycin, sulfathiazole, tetracycline, and trimethoprim-sulfamethoxazole. We measured antibodies to heat-labile E. coli toxin in 14 paired serum samples using the Chinese hamster ovary cell assay (19), and the adrenal cell assay* (20). Both assays identified five persons whose serum showed a greater than four-fold increase between acute and convalescent sera; two of these five persons had a positive culture for enterotoxigenic E. coli (Table 3). The four other culture-positive individuals did not show a significant rise in antibody titer in either assay. Discussion
Waterborne outbreaks of gastrointestinal illness continue to be a public health problem in the United States. Ninetynine outbreaks involving 16 950 cases of waterborne illness were reported to the Center for Disease Control from 1971 to 1974 (21-23). For 47% of those outbreaks, and 4 8 % of the cases, no causal agent could be identified. The outbreak described here is the first documentation of entero* Richard Guerrant and I. Kaye Wachsmuth independently tested the samples in the Chinese hamster ovary cell assay; David Sack tested them in the adrenal cell assay.
Figure 2. Identification of enterotoxigenic Escherichia coli isolates. (*) Number in parentheses represents persons from whose fecal culture the isolates were selected; some persons submitted more than one specimen. (**) 335 of the 985 06-negative isolates were tested for enterotoxin production.
June 1977 • Annals of Internal Medicine • Volume 86 • Number 6
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19522/ by a Scott Memorial Library User on 08/02/2017
toxigenic E. coli as the pathogen in a waterborne outbreak. The number of waterborne outbreaks for which this organism is actually responsible is not known. Several different types of evidence implicated contaminated drinking water as the primary vehicle of transmission in this outbreak. Epidemiologic data showed a strong association between water consumption and illness among resident park personnel and visitors, and the same pathogenic organism—a single epidemic strain—was isolated from both ill persons and the drinking water. In addition, the flourescein dye demonstration of direct sewage contamination was corroborated by bacteriologic evidence of coliform contamination. The evidence also suggests that the E. coli, which produced both heat-labile and heat-stable toxins and was isolated from 20 ill persons, was the actual cause of their illness. [1] No other bacterial, parasitic, or viral pathogens were found in specimens from these or other ill persons. [2] The enterotoxigenic epidemic strain was also isolated from the epidemiologically implicated vehicle of transmission. [3] There was a strong association between having a positive culture and having diarrhea at the time of culture. [4] A greater than four-fold rise in titer of antibody against E. coli enterotoxin was confirmed in five ill persons in two independent assay systems. This same strain probably caused the many additional cases reported in both staff and visitors. Similarities in symptoms reported by culture-positive staff and other staff and visitors suggested that they shared the same illness. Of particular interest is the long median duration of illness reported by all groups in this study. This long duration of illness has been reported by other observers in association with enterotoxigenic E. coli. DuPont and associates (24) noted that a volunteer who was fed 1 X 1010 organisms developed a severe illness that lasted 19 days; Gorbach and colleagues (2) reported that 2 1 % of the cases of E. coli diarrhea in travelers lasted 5 to 15 days; and Merson and coworkers (1) reported that 3 1 % of the documented E. coli diarrheas among their study subjects lasted more than 9 days. This is in contrast, however, to the relatively short illness described in persons with enterotoxigenic E. coli diarrhea in cholera-endemic areas of India and Bangladesh (4, 6 ) , where repeated exposure to enterotoxigenic E. coli may be more common. In most persons there appears to be a relatively short duration of excretion of this organism after recovery from diarrhea. Ryder and associates (4) noted that cases continued to shed enterotoxigenic E. coli for an average of 3.6 days after diarrhea ceased; and Gorbach and colleagues (2) found that in 82% of cases, enterotoxigenic E. coli could not be recovered from the stool 5 days after diarrhea ceased. This could explain the absence of culture-positive individuals among those we cultured 4 or more days after their diarrhea had ceased, and, in turn, may reflect the insensitivity of our current culture methods. The insensitivity of testing only 5 to 10 E. coli isolates randomly chosen from nonselective media is suggested by our finding that [1] only 42.5% (17/40) of ill patients cultured acutely had positive cultures, and [2] three persons with stool cultures negative for enterotoxigenic E. coli had a greater than
Table 3. Antitoxin Titer Changes and Enterotoxigenic Escherichia coli (ETEC) Culture Results from 14 III People from Crater Lake National Park
Patient No.
Change in Titer (Convalescent/Acute) Adrenal Cell
1 2 3 4 5 6 7 8 9 10 11 12 13 14
50.0* 38.8* 31.2* 15.5* 8.0* 2.4 2.0 2.0 1.0 1.0 1.0 0.6 0.8 0.5
Isolation of ETEC
Chinese hamster ovary cell 33.3* 35.0* 11.7* 22.6* 11.0* 2.0 1.3 0.5 0.7 0.5 0.3 1.0 0.7 0.3
+ + + + +
+
* Greater than four-fold increase.
four-fold rise in titer of antibody to heat-labile toxin. Serum antitoxin immunity also appears to be a relatively insensitive diagnostic test in North American subjects with naturally acquired disease. Our findings that most ill persons, even those with positive cultures, did not develop a four-fold rise in titer is consistent with previous studies of naturally acquired disease ( 1 , 20). The rise in titer observed in cholera-endemic areas may reflect prior exposure to the antigen (4) or a larger antigenic challenge initiating the illness. Although sewage probably contaminated the drinking water with multiple organisms, only a single pathogenic strain was recovered from ill persons. It is likely that the water supply had become increasingly contaminated with this strain as more ill park residents added infectious fecal material to the water by way of the obstructed sewer line. We may have failed to detect other pathogens that could have caused sporadic cases of illness. We initially examined only the first 33 specimens for enteric bacterial pathogens, identified an epidemic strain, and screened subsequent specimens for this epidemic strain. Serotype screening after identification of an epidemic strain by pathogenicity testing proved a useful technique for examining large numbers of isolates during an outbreak. It is important to distinguish this practice from that of trying to establish the pathogenicity of an isolate solely on the basis of whether it belongs to one of the "classical enteropathogenic" serotypes. Because toxin-production is a plasmid-borne trait, organisms with these designated sertoypes may or may not carry that trait (25). Conversely, an organism such as the epidemic strain in this outbreak may be a toxin-producer without being one of the classical serotypes. Previously, this same enterotoxigenic serotype ( 0 6 : H I 6 ) had been isolated from persons with gastrointestinal illness in Vietnam (24), Mexico ( 1 ) , Bangladesh (26), Kenya (26), Japan (26), and Arizona ( 7 ) . The widespread finding in many different places of this serotype coupled with enterotoxigenicity suggests that some properties associated Rosenberg et al. • Escherichia coli Diarrhea
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19522/ by a Scott Memorial Library User on 08/02/2017
717
with this particular strain may facilitate its ability to harbor and hold the plasmid for toxin production. However, our finding of three nontoxigenic 0 6 : H I 6 isolates—two of them from individuals who also harbored the toxigenic strain— suggests that this serotype can also lose its plasmid for toxin production. This is the first report of an outbreak of disease caused by enterotoxigenic E. coli in which both the organism and mode of transmission have been clearly identified. In 1968 an "enteropathogenic" serotype of E. coli was epidemiologically associated with a waterborne outbreak of gastrointestinal illness, but the strain from that outbreak was not demonstrated to be pathogenic by either toxin production or tissue invasion ( 2 7 ) . And in those studies where enterotoxigenic E. coli have been clearly identified (1-7), the mode of transmission has not been defined. If the number of enterotoxigenic E. coli organisms necessary to infect exposed individuals is high, as DuPont and associates (24) have suggested, one would expect transmission to occur most commonly through contaminated common vehicles, and only infrequently through person-to-person contact where the number of organisms transferred might be less than an infecting dose. This is true for cholera (28), and, as with cholera, good water and food hygiene would be expected to decrease the incidence of this disease. This suggestion is compatible with the recent observation that sporadic cases of diarrhea due to enterotoxigenic E. coli are only infrequently encountered in American cities (29, 3 0 ) . However, as this outbreak shows, enterotoxigenic E. coli organisms are present in this country and represent a public health threat particularly in situations where water quality and food sanitation are inadequate. ACKNOWLEDGMENTS: The authors thank the following persons who participated or assisted in the investigation of this outbreak: Jeff Adams, G. Edward Barnes, David Bussen, Gunther Craun, John Fritz, John A. Googins, Monroe Holmes, Michael Jensen, Katherine Jones, Bill Mullen, Mary Murphy, Michael O'Neal, Joseph P. Schock, Richard Sims, Bruce Stubblefield, Jack Stump, Willard S. Titus, James P. Wiggins, Gina Wright, and Kay Yarabinec. They also thank Robert A. Pollard for statistical assistance; Milford A. Hatch for virological studies on stools and sera; Betty R. Davis for serotyping assistance; and the Gerontology Research Center, National Institute on Aging, for facilities extended under its Guest Scientist Program. Received 1 November 1976; revision accepted 24 March 1977.
Escherichia coli isolated from patients with severe choleralike disease. J Infect Dis 123:378-385, 1971 7. SACK RB, HIRSCHHORN N, BROWNLEE I, et al: Enterotoxigenic
Escherichia-coli-associated diarrheal disease in Apache children. N Engl J Med 292:1041-1045, 1975 8. SACK RB,
SACK DA,
MEHLMAN
IJ, et
al:
Enterotoxigenic
Escherichia coli isolated from food sources. J Infect Dis 135: 313-317, 1977 9. EDWARDS PR, EWING WH: Identification of Enterobacteriaceae,
3rd ed. Minneapolis, Burgess Publishing Company, 1972, pp. 67-107 10. FEELEY JC, BALOWS A: Vibrio, in Manual of Clinical Microbiology, 2nd ed., edited by LENNETTE EH, SPAULDING EH,
TRUANT JP. Washington, D.C., American Society for Microbiology, 1974, pp. 238-245 11. DONTA ST, MOON HW, WHIPP SC: Detection of heat-labile
Escherichia coli enterotoxin with the use of adrenal cells in tissue culture. Science 183:334-336, 1974 12. SACK DA, SACK RB: Test for enterotoxigenic Escherichia coli using Y l adrenal cells in miniculture. Infect Immunol 11:334336, 1975 13. GUERRANT RL, BRUNTON LL, SCHNAITMAN TC, et al: Cyclic
adenosine monophosphate and alteration of Chinese hamster ovary cell morphology: a rapid, sensitive in vitro assay for the enterotoxins of Vibro cholerae and Escherichia coli. Infect Immunol 10:320-327, 1974 14. MORRIS GK, MERSON MH, SACK DA, et al: Laboratory in-
vestigation of diarrhea in travelers to Mexico: evaluation of methods for detecting enterotoxigenic Escherichia coli. J Clin Microbiol 3:486-495, 1976 15. SERENY B: Experimental Shigella Keratoconjunctivitis. Acta Microbiol Acad Sci Hung 2:293-296, 1955 16. MIDDLETON PJ, SZYMANSKI MT, ABBOTT GD, et al: Orbivirus
acute gastroenteritis of infancy. Lancet 1:1241-1244, 1974 17. KAPIKIAN AZ, WYATT RG, DOLIN R, et al: Visualization by
immune electron microscopy of a 27-nm particle associated with acute infectious nonbacterial gastroenteritis. J Virol 10: 1075-1081, 1972 18. BAUER AW, KIRBY WMM, SHERRIS JC, et al: Antibiotic sus-
ceptibility testing by a standardized single disk method. Am J Clin Pathol 45:493-496, 1966 19. GUERRANT RL, BRUNTON LL: Characterization of the Chinese
hamster ovary cell assay for the enterotoxin of Vibrio cholerae and Escherichia coli and for antitoxin: differential inhibition by gangliosides, specific antisera, and toxoid. J Infect Dis May 1977, in press 20. SACK RB, HIRSCHBORN N, WOODWARD WE, et al: Antibodies to
heat-labile Escherichia coli enterotoxin in Apaches in White River, Arizona. Infect Immunol 12:1475-1477, 1975 21. MERSON MH, BARKER WH JR, CRAUN GF, et al: Outbreaks of
waterborne disease in the United States, 1971-1972. / Infect Dis 129:614-615, 1974 22. HUGHES JM, MERSON MH, CRAUN GF, et al: Outbreaks of
waterborne disease in the United States, 1973. J Infect Dis 132:336-339, 1975 23. HORWITZ MA, HUGHES JM, CRAUN GF: Outbreaks of water-
borne disease in the United States, 1974. J Infect Dis 133:588593, 1976 24. DUPONT HL, FORMAL SB, HORNICK RB, et al: Pathogenesis
• Requests for reprints should be addressed to Enteric Diseases Branch, Bureau of Epidemiology, Center for Disease Control; Atlanta, GA 30333.
of Escherichia coli diarrhea. N Engl J Med 285:1-9, 1971 25. SMITH HW, HALLS A: The transmissible nature of the genetic factor in Escherichia coli that controls enterotoxin production. / Gen Microbiol 52:319-334, 1968
References
26. 0RSKOV F, 0RSKOV I, EVANS DJ JR, et al: Special Escherichia
1. MERSON MH, MORRIS GK, SACK DA, et al: Travelers' diarrhea
in Mexico. A prospective study of physicians and family members attending a congress. N Engl J Med 294:1299-1305, 1976
coli serotypes among enterotoxigenic strains from diarrhoea in adults and children. Med Microbiol Immunol 162:73-80, 1976 27. SCHROEDER SA, CALDWELL JR, VERNON TM, et al: A water-
2. GORBACH SL, KEAN BH, EVANS DG, et al: Travelers' diarrhea
borne outbreak of gastroenteritis in adults associated with Escherichia coli. Lancet 1:737-740, 1968
and toxigenic Escherichia coli. N Engl J Med 292:933-936, 1975
28. GANGAROSA EJ, MOSLEY WH: Epidemiology and surveillance
3. SHORE EG, DEAN AG, HOLIK KJ, et al: Enterotoxin-producing
Escherichia coli and diarrheal disease in audit travelers: a prospective study. J Infect Dis 129:577-582, 1974
of cholera, in Cholera, edited by BARUA D, BURROWS W. Philadelphia, W.B. Saunders Co., 1974, pp. 381-403 (pp. 393395)
4. RYDER RW, SACK DA, KAPIKIAN AZ, et al: Enterotoxigenic
29. ECHEVERRIA P, BLACKLOW NR, SMITH DH: Role of heat-labile
Escherichia coli and reovirus-like agent in rural Bangladesh. Lancet 1:659-662, 1976
toxigenic Escherichia coli and reovirus-like agent in diarrhoea in Boston children. Lancet 2:1113-1116, 1975
5. GUERRANT RL, MOORE RA, KIRSCHENFELD PM, et al: Role of
30. KAPIKIAN AZ, KIM HW, WYATT RG, et al: Human reovirus-
toxigenic and invasive bacteria in acute diarrhea of childhood. N Engl J Med 293:567-573, 1975
like agent as the major pathogen associated with "winter" gastroenteritis in hospitalized infants and young children. N Engl J Med 294:965-972, 1976
6. SACK RB, GORBACH SL, BANWELL JG, et al: Enterotoxigenic
718
June 1977 • Annals of Internal Medicine • Volume 86 • Number 6
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19522/ by a Scott Memorial Library User on 08/02/2017
In June and July 1975, gastrointestinal illness occurred in more than 200 staff members and 2000 visitors to an American national park. It was characterized by prolonged diarrhea, cramps, nausea, and vomiting, lasted a median duration of 8 days, and was significantly associated with consumption of park water (P < 0.001), which had been contaminated by raw sewage. Enterotoxigenic Escherichia coli serotype 06:K15:H16 was isolated from 20 of 49 ill park residents and from the park's water supply, but not from 71 residents who had never been ill or had been well for at least 4 days. No other bacterial, viral, or parasitic pathogens were isolated from ill or well persons. This outbreak is the first waterborne epidemic of diarrheal illness shown to be due to enterotoxigenic E. coli, and this study documents one mode of transmission of this pathogen. This investigation also suggests the relative insensitivity of current methods for identifying persons infected with this organism, either by the culturing of randomly selected isolates or by measuring serologic responses.
ENTEROTOXIGENIC Escherichia coli, a recently recognized enteric pathogen with a worldwide distribution, is a cause of acute watery diarrhea, especially in developing countries, and is the most common cause of acute diarrhea in travelers (1-6). The disease caused by this organism varies considerably in severity, ranging from mild diarrhea to a seriously dehydrating cholera-like illness. Little is known, however, of the mode of transmission of the disease. Although enterotoxigenic E. coli have been isolated from food and water sources, consumption of such sources has not been directly related to human disease (4, 7, 8 ) . In the summer of 1975, an outbreak of diarrheal disease caused illness in more than 2000 staff and visitors who • From the Bureau of Epidemiology, Center for Disease Control, Public Health Service, U.S. Department of Health, Education, and Welfare, Atlanta, Georgia; Department of Internal Medicine, University of Virginia School of Medicine, Charlottesville, Virginia; and the Department of Medicine, Baltimore City Hospitals, Johns Hopkins University School of Medicine, Baltimore, Maryland.
714
drank contaminated water at an American national park. Using recently developed laboratory methods to test E. coli for enterotoxin production, we showed that enterotoxigenic E. coli had caused this waterborne common-source outbreak. Epidemiologic Investigation Crater Lake National Park is a 250-square-mile rectangular tract in southwestern Oregon, whose 2000-foot-deep lake attracts almost 3000 tourists a day during the summer (Figure 1). Reports of widespread gastrointestinal illness among resident park staff and their families in late June 1975 led to the initiation of this investigation on 7 July 1975. Epidemiologic and clinical information regarding gastrointestinal symptoms was obtained from three groups of people: [1] 320 ( 9 6 % ) of the 335 resident park personnel and family members completed questionnaires from 7 to 11 July; [2] 107 ( 1 2 % ) of the 900 visitors who had stayed overnight at the park from 4 to 9 July responded to an initial telephone survey on 9 and 10 July; [3] 5273 ( 6 3 % ) visitors who had stayed overnight at the park between 10 June and 11 July responded to a short questionnaire (without questions about specific symptoms) mailed out in September 1975. Illness (defined as diarrhea or vomiting after visits to the park) was reported by 90% (288/320) of the resident park personnel, 64% (68/107) of the visitors contacted by telephone, and 44% (2310/ 5273) of the visitors surveyed by mail. Illness was significantly associated with drinking park water (P < 0.001) among all groups except those visitors contacted by phone who had arrived between 7 and 9 July when water chlorine levels were being closely monitored (Table 1). Among park personnel the median duration of illness increased as daily water consumption increased. A majority of those ill experienced diarrhea, cramps, nausea, and vomiting (Table 2 ) . The median duration or illness for park personnel was 8 days. Although 72% of the personnel missed work during part of their illness, most employees, including food handlers, worked at least part of Annals of Internal Medicine 86:714-718, 1977
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19522/ by a Scott Memorial Library User on 08/02/2017
Table 1. Association between Illness and Consumption of Park Water
Group
Park personnel Visitors telephoned July 4-6 July 7-9 Visitors surveyed
Drank
Did Not Drink*
Significancet
111
Well
Attack Rate
111
Well
Attack Rate
269
17
94.1%
5
13
27.8%
P < 0.00001
42 18 1859
0 34 1652
100.0% 34.6% 52.9%
0 2 23
2 2 565
0.0% 50.0% 3.9%
P < 0.001 NS P < 0.00001
* Defined as drinking less than one glass of water per day. t By Fisher's 2-tail test.
the time they were ill. The median duration of illness reported through the visitors' mail survey was 6 days, and 38% of these ill persons were ill for 10 or more days. Water for the park came from a shallow spring and was chlorinated before distribution; however, there was no systematic monitoring of chlorine levels in the distribution system. Three of four water samples sent to the county health department in May 1975 for routine bacterologic analysis had been found to have coliform contamination ( > 1 per 50 ml) and had been judged unacceptable. When twice-daily monitoring of chlorine levels began on 7 July, it showed that water going to some parts of the park had no detectable chlorine, even though adequate levels were present in water at other areas. On the evening of 10 July, snow that had covered a manhole melted and revealed sewage overflow through the manhole, which was located about 2000 feet uphill from the springs that supplied the park's water. Direct contamination of the water supply with sewage was shown by fluorescein dye tests, and the park was officially closed the next morning. The park reopened on 1 August, and a second questionnaire was completed by 221 (62%) park personnel on 14 August, 2 weeks later. Fifteen people (6.8%) reported recurrences in August after having been well for at least 1 week; the median interval between the two episodes of illness was 22 days. While the park was closed, 109 employees had left the grounds to spend at least 2 days with family or friends who had not visited the park. The rate of subsequent illness (diarrhea or vomiting) in persons visited by ill personnel (31/335) was significantly greater (P 0.015) than that in persons visited by well personnel (5/160).
negative colonies were selected from MacConkey's agar (Difco Laboratories; Detroit, Michigan) plates and tested for toxigenicity. Production of E. coli heat-labile toxin was determined by the mouse adrenal tumor cell (11, 12) and Chinese hamster ovary (13) tissue culture techniques, production of heat-stable toxin by the infant-mouse assay (14). The toxigenic isolates were serotyped for O and H antigens (9) and found to be all the same serotype, 06:HI6. The remaining 87 swabs received were examined for
Laboratory Investigation
Rectal swabs or stool swab specimens were obtained from 120 of the park staff and resident family members who filled out clinical information questionnaires. Swabs were immediately placed in Cary-Blair transport medium (Baltimore Biological Laboratories; Cockeysville, Maryland) and shipped on ice to laboratories at the Center for Disease Control. The first 33 fecal cultures were screened for Salmonella, Shigella, and pathogenic vibrios by previously described methods (9, 10) and for Yersinia enterocolitica by culturing on Salmonella-Shigella agar (Difco Laboratories; Detroit, Michigan) incubated at 22 °C for 48 h. All were negative for these bacteria. From each of these 33 fecal cultures, five lactose-positive and two lactose-
Figure 1 . Crater Lake National Park. Rosenberg et al. • Escherichia coli Diarrhea
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19522/ by a Scott Memorial Library User on 08/02/2017
715
Table 2. Symptoms and Severity of Gastrointestinal Illness among III Park Personnel and Visitors, 13 June-11 July 1975
Symptom
All 111 Park Personnel " ~~ (/! = 288)
Diarrhea Cramps Nausea Vomiting Chills Headache Muscle ache Fever Bloody diarrhea Median duration of illness Median interval between arrival and onset of illness Changed daily pattern of activities Missed work Consulted physician Hospitalized
93% 77% 72% 68% 48% 47% 45% 41% 0.3%
Park Visitors Personnel with Contacted Enterotoxigenic by Escherichia coli Telephone („ = 20) (/i - 68) 100% 91% 95% 79% 75% 88% 80% 79% 45% 54% 80% 53% 60% 47% 50% 37% 0% 2%
8 days
11 days
6 days*
18 days
20 days
30 hours
90% 72% 10% 0%
94% 75% 13% 0%
85% 24% 27% 0%
* Most ill visitors were still ill when interviewed initially; median duration of illness among 2310 ill visitors who later replied to mail questionnaire was 6 days.
E. coli only. Five lactose-positive and two lactose-negative isolates from each specimen were screened with E. coli 06 antiserum. All 06-positive E. coli (and a sample of the 06negative isolates) were tested for toxin production in the adrenal cell, Chinese hamster ovary, and infant mouse assays. Enterotoxigenic E. coli that produced both heat-stable and heat-labile toxins was isolated from 20 (16.7% ) of the 120 people examined (Figure 2 ) . All enterotoxigenic E. coli were serotype 06:H16. A 34% sample (335/985) of 06-negative isolates was negative for toxin producers. Escherichia coli 06 organisms that were not enterotoxigenic were isolated from 15 (12.5%) people. Four enterotoxigenic isolates were tested, and all were found to possess K 15 antigen (9). Among the 120 persons from whom a rectal culture had been obtained, there was a strong association between having a culture positive for enterotoxigenic E. coli and having had diarrhea at the time of culturing (P < 0.000001 Fisher's 2-tail test). Positive cultures were obtained from 17 of the 40 persons who had diarrhea when cultured, and from three of the 80 persons who did not have diarrhea when cultured. The three positive cultures in this second group came from nine persons who had diarrhea 1 to 3 days previously; no positive cultures were obtained from 71 persons who had not been ill or who had been free of diarrhea for 4 or more days. One hundred and twelve E. coli isolates from 16 ill persons were tested for invasiveness by the Sereny test (16), and none were positive. No ova or parasites were found in fresh stool specimens from 15 acutely ill persons, which were examined by the Oregon Health Division laboratory. Frozen stool specimens from six acutely ill individuals 716
were examined for enteric viruses in four viral isolation systems: primary monkey kidney cells, rhabdomyosarcoma cells of human origin, human embryonic lung fibroblast cells, and newborn mice. These specimens were also examined for reoviruslike and parvoviruslike gastroenteritis agents by direct electronmicroscopy and immunoelectronmicroscopy (16, 17). All viral studies were negative. Two litres of raw water taken from the park springs on 10 July were examined and found negative for Salmonella, Shigella, pathogenic vibrios, and Y. enterocolitica. Twenty lactose-positive and five lactose-negative colonies were selected from each sample: additional isolates were selected from water samples collected at various points in the distribution system. Seventeen of 396 water isolates were 06 positive, and 14 of these were enterotoxigenic in all three assays: the enterotoxigenic isolates were all serotype 06:H16. Those 06-positive isolates that were not enterotoxigenic were negative in H I 6 antiserum. Additionally, 113 water isolates that were 06 negative were tested in the adrenal cell tissue culture assay; all were negative. All enterotoxigenic isolates from both patients and water were multiply sensitive when tested by the Kirby-Bauer method (18) against ampicillin, carbenicillin, cepholethin, chloramphenicol, colistin, gentamicin, nalidixic acid, neomycin, nitrofurantoin, streptomycin, sulfathiazole, tetracycline, and trimethoprim-sulfamethoxazole. We measured antibodies to heat-labile E. coli toxin in 14 paired serum samples using the Chinese hamster ovary cell assay (19), and the adrenal cell assay* (20). Both assays identified five persons whose serum showed a greater than four-fold increase between acute and convalescent sera; two of these five persons had a positive culture for enterotoxigenic E. coli (Table 3). The four other culture-positive individuals did not show a significant rise in antibody titer in either assay. Discussion
Waterborne outbreaks of gastrointestinal illness continue to be a public health problem in the United States. Ninetynine outbreaks involving 16 950 cases of waterborne illness were reported to the Center for Disease Control from 1971 to 1974 (21-23). For 47% of those outbreaks, and 4 8 % of the cases, no causal agent could be identified. The outbreak described here is the first documentation of entero* Richard Guerrant and I. Kaye Wachsmuth independently tested the samples in the Chinese hamster ovary cell assay; David Sack tested them in the adrenal cell assay.
Figure 2. Identification of enterotoxigenic Escherichia coli isolates. (*) Number in parentheses represents persons from whose fecal culture the isolates were selected; some persons submitted more than one specimen. (**) 335 of the 985 06-negative isolates were tested for enterotoxin production.
June 1977 • Annals of Internal Medicine • Volume 86 • Number 6
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19522/ by a Scott Memorial Library User on 08/02/2017
toxigenic E. coli as the pathogen in a waterborne outbreak. The number of waterborne outbreaks for which this organism is actually responsible is not known. Several different types of evidence implicated contaminated drinking water as the primary vehicle of transmission in this outbreak. Epidemiologic data showed a strong association between water consumption and illness among resident park personnel and visitors, and the same pathogenic organism—a single epidemic strain—was isolated from both ill persons and the drinking water. In addition, the flourescein dye demonstration of direct sewage contamination was corroborated by bacteriologic evidence of coliform contamination. The evidence also suggests that the E. coli, which produced both heat-labile and heat-stable toxins and was isolated from 20 ill persons, was the actual cause of their illness. [1] No other bacterial, parasitic, or viral pathogens were found in specimens from these or other ill persons. [2] The enterotoxigenic epidemic strain was also isolated from the epidemiologically implicated vehicle of transmission. [3] There was a strong association between having a positive culture and having diarrhea at the time of culture. [4] A greater than four-fold rise in titer of antibody against E. coli enterotoxin was confirmed in five ill persons in two independent assay systems. This same strain probably caused the many additional cases reported in both staff and visitors. Similarities in symptoms reported by culture-positive staff and other staff and visitors suggested that they shared the same illness. Of particular interest is the long median duration of illness reported by all groups in this study. This long duration of illness has been reported by other observers in association with enterotoxigenic E. coli. DuPont and associates (24) noted that a volunteer who was fed 1 X 1010 organisms developed a severe illness that lasted 19 days; Gorbach and colleagues (2) reported that 2 1 % of the cases of E. coli diarrhea in travelers lasted 5 to 15 days; and Merson and coworkers (1) reported that 3 1 % of the documented E. coli diarrheas among their study subjects lasted more than 9 days. This is in contrast, however, to the relatively short illness described in persons with enterotoxigenic E. coli diarrhea in cholera-endemic areas of India and Bangladesh (4, 6 ) , where repeated exposure to enterotoxigenic E. coli may be more common. In most persons there appears to be a relatively short duration of excretion of this organism after recovery from diarrhea. Ryder and associates (4) noted that cases continued to shed enterotoxigenic E. coli for an average of 3.6 days after diarrhea ceased; and Gorbach and colleagues (2) found that in 82% of cases, enterotoxigenic E. coli could not be recovered from the stool 5 days after diarrhea ceased. This could explain the absence of culture-positive individuals among those we cultured 4 or more days after their diarrhea had ceased, and, in turn, may reflect the insensitivity of our current culture methods. The insensitivity of testing only 5 to 10 E. coli isolates randomly chosen from nonselective media is suggested by our finding that [1] only 42.5% (17/40) of ill patients cultured acutely had positive cultures, and [2] three persons with stool cultures negative for enterotoxigenic E. coli had a greater than
Table 3. Antitoxin Titer Changes and Enterotoxigenic Escherichia coli (ETEC) Culture Results from 14 III People from Crater Lake National Park
Patient No.
Change in Titer (Convalescent/Acute) Adrenal Cell
1 2 3 4 5 6 7 8 9 10 11 12 13 14
50.0* 38.8* 31.2* 15.5* 8.0* 2.4 2.0 2.0 1.0 1.0 1.0 0.6 0.8 0.5
Isolation of ETEC
Chinese hamster ovary cell 33.3* 35.0* 11.7* 22.6* 11.0* 2.0 1.3 0.5 0.7 0.5 0.3 1.0 0.7 0.3
+ + + + +
+
* Greater than four-fold increase.
four-fold rise in titer of antibody to heat-labile toxin. Serum antitoxin immunity also appears to be a relatively insensitive diagnostic test in North American subjects with naturally acquired disease. Our findings that most ill persons, even those with positive cultures, did not develop a four-fold rise in titer is consistent with previous studies of naturally acquired disease ( 1 , 20). The rise in titer observed in cholera-endemic areas may reflect prior exposure to the antigen (4) or a larger antigenic challenge initiating the illness. Although sewage probably contaminated the drinking water with multiple organisms, only a single pathogenic strain was recovered from ill persons. It is likely that the water supply had become increasingly contaminated with this strain as more ill park residents added infectious fecal material to the water by way of the obstructed sewer line. We may have failed to detect other pathogens that could have caused sporadic cases of illness. We initially examined only the first 33 specimens for enteric bacterial pathogens, identified an epidemic strain, and screened subsequent specimens for this epidemic strain. Serotype screening after identification of an epidemic strain by pathogenicity testing proved a useful technique for examining large numbers of isolates during an outbreak. It is important to distinguish this practice from that of trying to establish the pathogenicity of an isolate solely on the basis of whether it belongs to one of the "classical enteropathogenic" serotypes. Because toxin-production is a plasmid-borne trait, organisms with these designated sertoypes may or may not carry that trait (25). Conversely, an organism such as the epidemic strain in this outbreak may be a toxin-producer without being one of the classical serotypes. Previously, this same enterotoxigenic serotype ( 0 6 : H I 6 ) had been isolated from persons with gastrointestinal illness in Vietnam (24), Mexico ( 1 ) , Bangladesh (26), Kenya (26), Japan (26), and Arizona ( 7 ) . The widespread finding in many different places of this serotype coupled with enterotoxigenicity suggests that some properties associated Rosenberg et al. • Escherichia coli Diarrhea
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19522/ by a Scott Memorial Library User on 08/02/2017
717
with this particular strain may facilitate its ability to harbor and hold the plasmid for toxin production. However, our finding of three nontoxigenic 0 6 : H I 6 isolates—two of them from individuals who also harbored the toxigenic strain— suggests that this serotype can also lose its plasmid for toxin production. This is the first report of an outbreak of disease caused by enterotoxigenic E. coli in which both the organism and mode of transmission have been clearly identified. In 1968 an "enteropathogenic" serotype of E. coli was epidemiologically associated with a waterborne outbreak of gastrointestinal illness, but the strain from that outbreak was not demonstrated to be pathogenic by either toxin production or tissue invasion ( 2 7 ) . And in those studies where enterotoxigenic E. coli have been clearly identified (1-7), the mode of transmission has not been defined. If the number of enterotoxigenic E. coli organisms necessary to infect exposed individuals is high, as DuPont and associates (24) have suggested, one would expect transmission to occur most commonly through contaminated common vehicles, and only infrequently through person-to-person contact where the number of organisms transferred might be less than an infecting dose. This is true for cholera (28), and, as with cholera, good water and food hygiene would be expected to decrease the incidence of this disease. This suggestion is compatible with the recent observation that sporadic cases of diarrhea due to enterotoxigenic E. coli are only infrequently encountered in American cities (29, 3 0 ) . However, as this outbreak shows, enterotoxigenic E. coli organisms are present in this country and represent a public health threat particularly in situations where water quality and food sanitation are inadequate. ACKNOWLEDGMENTS: The authors thank the following persons who participated or assisted in the investigation of this outbreak: Jeff Adams, G. Edward Barnes, David Bussen, Gunther Craun, John Fritz, John A. Googins, Monroe Holmes, Michael Jensen, Katherine Jones, Bill Mullen, Mary Murphy, Michael O'Neal, Joseph P. Schock, Richard Sims, Bruce Stubblefield, Jack Stump, Willard S. Titus, James P. Wiggins, Gina Wright, and Kay Yarabinec. They also thank Robert A. Pollard for statistical assistance; Milford A. Hatch for virological studies on stools and sera; Betty R. Davis for serotyping assistance; and the Gerontology Research Center, National Institute on Aging, for facilities extended under its Guest Scientist Program. Received 1 November 1976; revision accepted 24 March 1977.
Escherichia coli isolated from patients with severe choleralike disease. J Infect Dis 123:378-385, 1971 7. SACK RB, HIRSCHHORN N, BROWNLEE I, et al: Enterotoxigenic
Escherichia-coli-associated diarrheal disease in Apache children. N Engl J Med 292:1041-1045, 1975 8. SACK RB,
SACK DA,
MEHLMAN
IJ, et
al:
Enterotoxigenic
Escherichia coli isolated from food sources. J Infect Dis 135: 313-317, 1977 9. EDWARDS PR, EWING WH: Identification of Enterobacteriaceae,
3rd ed. Minneapolis, Burgess Publishing Company, 1972, pp. 67-107 10. FEELEY JC, BALOWS A: Vibrio, in Manual of Clinical Microbiology, 2nd ed., edited by LENNETTE EH, SPAULDING EH,
TRUANT JP. Washington, D.C., American Society for Microbiology, 1974, pp. 238-245 11. DONTA ST, MOON HW, WHIPP SC: Detection of heat-labile
Escherichia coli enterotoxin with the use of adrenal cells in tissue culture. Science 183:334-336, 1974 12. SACK DA, SACK RB: Test for enterotoxigenic Escherichia coli using Y l adrenal cells in miniculture. Infect Immunol 11:334336, 1975 13. GUERRANT RL, BRUNTON LL, SCHNAITMAN TC, et al: Cyclic
adenosine monophosphate and alteration of Chinese hamster ovary cell morphology: a rapid, sensitive in vitro assay for the enterotoxins of Vibro cholerae and Escherichia coli. Infect Immunol 10:320-327, 1974 14. MORRIS GK, MERSON MH, SACK DA, et al: Laboratory in-
vestigation of diarrhea in travelers to Mexico: evaluation of methods for detecting enterotoxigenic Escherichia coli. J Clin Microbiol 3:486-495, 1976 15. SERENY B: Experimental Shigella Keratoconjunctivitis. Acta Microbiol Acad Sci Hung 2:293-296, 1955 16. MIDDLETON PJ, SZYMANSKI MT, ABBOTT GD, et al: Orbivirus
acute gastroenteritis of infancy. Lancet 1:1241-1244, 1974 17. KAPIKIAN AZ, WYATT RG, DOLIN R, et al: Visualization by
immune electron microscopy of a 27-nm particle associated with acute infectious nonbacterial gastroenteritis. J Virol 10: 1075-1081, 1972 18. BAUER AW, KIRBY WMM, SHERRIS JC, et al: Antibiotic sus-
ceptibility testing by a standardized single disk method. Am J Clin Pathol 45:493-496, 1966 19. GUERRANT RL, BRUNTON LL: Characterization of the Chinese
hamster ovary cell assay for the enterotoxin of Vibrio cholerae and Escherichia coli and for antitoxin: differential inhibition by gangliosides, specific antisera, and toxoid. J Infect Dis May 1977, in press 20. SACK RB, HIRSCHBORN N, WOODWARD WE, et al: Antibodies to
heat-labile Escherichia coli enterotoxin in Apaches in White River, Arizona. Infect Immunol 12:1475-1477, 1975 21. MERSON MH, BARKER WH JR, CRAUN GF, et al: Outbreaks of
waterborne disease in the United States, 1971-1972. / Infect Dis 129:614-615, 1974 22. HUGHES JM, MERSON MH, CRAUN GF, et al: Outbreaks of
waterborne disease in the United States, 1973. J Infect Dis 132:336-339, 1975 23. HORWITZ MA, HUGHES JM, CRAUN GF: Outbreaks of water-
borne disease in the United States, 1974. J Infect Dis 133:588593, 1976 24. DUPONT HL, FORMAL SB, HORNICK RB, et al: Pathogenesis
• Requests for reprints should be addressed to Enteric Diseases Branch, Bureau of Epidemiology, Center for Disease Control; Atlanta, GA 30333.
of Escherichia coli diarrhea. N Engl J Med 285:1-9, 1971 25. SMITH HW, HALLS A: The transmissible nature of the genetic factor in Escherichia coli that controls enterotoxin production. / Gen Microbiol 52:319-334, 1968
References
26. 0RSKOV F, 0RSKOV I, EVANS DJ JR, et al: Special Escherichia
1. MERSON MH, MORRIS GK, SACK DA, et al: Travelers' diarrhea
in Mexico. A prospective study of physicians and family members attending a congress. N Engl J Med 294:1299-1305, 1976
coli serotypes among enterotoxigenic strains from diarrhoea in adults and children. Med Microbiol Immunol 162:73-80, 1976 27. SCHROEDER SA, CALDWELL JR, VERNON TM, et al: A water-
2. GORBACH SL, KEAN BH, EVANS DG, et al: Travelers' diarrhea
borne outbreak of gastroenteritis in adults associated with Escherichia coli. Lancet 1:737-740, 1968
and toxigenic Escherichia coli. N Engl J Med 292:933-936, 1975
28. GANGAROSA EJ, MOSLEY WH: Epidemiology and surveillance
3. SHORE EG, DEAN AG, HOLIK KJ, et al: Enterotoxin-producing
Escherichia coli and diarrheal disease in audit travelers: a prospective study. J Infect Dis 129:577-582, 1974
of cholera, in Cholera, edited by BARUA D, BURROWS W. Philadelphia, W.B. Saunders Co., 1974, pp. 381-403 (pp. 393395)
4. RYDER RW, SACK DA, KAPIKIAN AZ, et al: Enterotoxigenic
29. ECHEVERRIA P, BLACKLOW NR, SMITH DH: Role of heat-labile
Escherichia coli and reovirus-like agent in rural Bangladesh. Lancet 1:659-662, 1976
toxigenic Escherichia coli and reovirus-like agent in diarrhoea in Boston children. Lancet 2:1113-1116, 1975
5. GUERRANT RL, MOORE RA, KIRSCHENFELD PM, et al: Role of
30. KAPIKIAN AZ, KIM HW, WYATT RG, et al: Human reovirus-
toxigenic and invasive bacteria in acute diarrhea of childhood. N Engl J Med 293:567-573, 1975
like agent as the major pathogen associated with "winter" gastroenteritis in hospitalized infants and young children. N Engl J Med 294:965-972, 1976
6. SACK RB, GORBACH SL, BANWELL JG, et al: Enterotoxigenic
718
June 1977 • Annals of Internal Medicine • Volume 86 • Number 6
Downloaded From: https://annals.org/pdfaccess.ashx?url=/data/journals/aim/19522/ by a Scott Memorial Library User on 08/02/2017
