AMERICAN JOURNAL OF EPIDJMIOLOGY Copyright 198 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved Vol. Ill, No. 4 Printed in USA. HEAT-LABILE ENTEROTOXIGENIC ESCHERICHIA COLI INDUCED DIARRHEA ABOARD A MIAMI-BASED CRUISE SHIP 1 ROBERT M. LUMISH, ROBERT W. RYDER, 1 DANIEL C. ANDERSON, JOY G. WELLS AND NANCY D. PUHR Lumlsh, R. M., R. W. Ryder (Enteric Diseases Branch, Bacterial Diseases Division, CDC, Atlanta, GA 3333), D. C. Anderson, J. G. Wells, and N. D. Puhr. Heat-labile enterotoxigenic Escherichia coll induced diarrhea aboard a Miami-based cruise ship. Am J Epidemiol 111:432-436, 198. Beginning In late December, 1975, almost one-third of the passengers on two successive cruises of a Miami-based cruise ship noted the onset of diarrheal Illness while on board. A single serotype of Escherichia coll that produced heat-labile enterotoxin without producing heat-stable enterotoxln was recovered from the stool of most ill passengers cultured. Epldemiologic investigation could not specifically define the mode of spread. The clinical picture presented was similar to the illness caused by enterotoxigenic E. coll that produce only heat-stable enterotoxin or that produce heat-stable and heat-labile enterotoxins. ' disease outbreaks; enterotoxins; Escherichia coll infections; gastroenteritis; ships; travel Figures obtained from the Port of Miami and the Harbor of Port Everglades, Miami, Florida, indicate that in 1975 roughly one-half million people used southern Florida cruise ships to travel in the Caribbean. As with travelers to foreign lands, some passengers developed gastrointestinal illness. During a recent investigation of an outbreak of diarrhea aboard a Miami-based cruise ship, we isolated a strain of enterotoxigenic Escherichia coli (serotype 25:K98:non- Received for publication February 5, 1979, and in final form October 25, 1979. Abbreviations: ETEC, enterotoxigenic Escherichia coli; LT, heat-labile enterotoxin; NM, non-motile; ST, heat-stable enterotoxin. 1 From the Field Services Division and the Bacterial Diseases Division, Bureau of Epidemiology, CDC, Atlanta, GA 3333. 1 Reprint requests to Dr. Ryder, Enteric Diseases Branch, Bacterial Diseases Division, CDC, Atlanta, GA 3333. The authors thank John Yashuk and Ross Cox, of the Miami Quarantine Station, and Eugene J. Gangarosa, M.D., Chief, Enteric Diseases Branch and Deputy Director, Bacterial Diseases Division, Bureau of Epidemiology, CDC, for their help in the investigation. motile (NM)) that produced only heatlabile enterotoxin (LT). We describe our findings in this report. MATERIALS AND METHODS The ship on which the outbreak occurred carried 9 passengers and 386 crew and made biweekly round trip cruises from Miami to Nassau in the Bahamas. During a three-day cruise in late December, 1975, because of notification that 11 per cent of passengers had experienced a diarrheal illness defined as greater than or equal to three loose or watery bowel movements per 24 hours, we conducted a limited investigation of the ship. This investigation included a sanitation inspection, a questionnaire survey of 17 passengers, and an examination of rectal swabs for enteric pathogens from 16 ill and 1 well passengers and from six ill and eight well crew. Because of continued illness on the next cruise, we conducted a more extensive investigation while the ship was in transit from Nassau to Miami. The sanitation in- 432
SHIPBOARD ENTEROTOXIGENIC E. COLI INDUCED DIARRHEA 433 spection was continued, questionnaires were distributed to all 94 passengers and 386 crew, and rectal swabs were obtained from 19 ill and five well passengers and 11 ill and six well crew. Thirty-one food samples and 41 environmental samples were taken from food preparation areas. Acute sera (at the time of the cruise) and convalescent sera (three to five weeks later) were collected from eight ill passengers. The rectal swabs were immediately placed in refrigerated Cary-Blair transport medium and sent to the Center for Disease Control, where they were examined for Shigella (1), Salmonella (2), and pathogenic Vibrios (3). Specimens were tested for Yersinia enterocolitica by plating to SS agar incubating at 22 C for 48 hours and further speciating typical colonies. Specimens were also streaked on MacConkey agar and incubated for 24 hours at C. From each MacConkey plate, five lactose positive colonies typical of E. coli and two lactose negative, colonies were selected. The isolates were tested for production of LT using the Y-l adrenal cell tissue culture system (4), for heat-stable enterotoxin (ST) production using the infant mouse assay (5), and for enteroinvasiveness using the Sereny test (6). E. coli organisms isolated from food and environmental specimens were examined for the production of LT. Enterotoxigenic E. coli (ETEC) were serotyped using 138 E. coli O antisera and 21 pools (2). Routine biochemical characterization (2) and standardized antibiotic susceptibility (Kirby-Bauer) (7) were determined on all of the ETEC isolates. One of the epidemic isolates was serotyped using K antisera. Acute and convalescent serum antibody titers to LT were measured using the Y-l adrenal cell assay (8). RESULTS Epidemiologic Extensive diarrheal illness was documented on both cruises. On the first cruise 64 (41 per cent) of the 156 passengers returning questionnaires reported diarrheal illness. On the second cruise 829 (92 per cent) of the 94 passengers returned questionnaires and 259 (31.2 per cent) reported diarrhea. On this same cruise 339 (88 per cent) of the 386 crew returned questionnaires and 26 (7.7 per cent) reported having had diarrheal illness at some time during the past week. Illness in passengers began 12 hours after they boarded and peaked in 36-48 hours (figure 1). Abdominal cramps, diarrhea and nausea characterized the illness (table 1). The median duration of illness for passengers and crew was two days, although 57 per cent of ill passengers on the second cruise were still ill when they debarked from the ship and were lost to follow-up. Increased consumption of ship's water was associated with a higher risk of developing illness in passengers but not crew (table 2). Analysis of food preference questionnaires completed by passengers on the second cruise revealed an association between developing illness and consumption of crabmeat cocktail (p <.1) and prime rib (p <.1) at dinner on the first night. Consumption of food at the midnight buffet the same evening was also associated with illness (p <.1), as was shrimp cocktail served at dinner the second night (p <.1). However, crosstable analysis of the three implicated food items demonstrated that prime rib and shrimp cocktail were not correlated with illness, and that illness was associated with ingestion of crabmeat cocktail (table 3). No association between age, sex, cabin location or dining table and illness could be demonstrated. Laboratory None of the food or environmental samples contained ETEC organisms. A total of 365.E. coli isolates from 8 persons cultured were serotyped and tested for heat-
434 LUMISH, RYDER, ANDERSON, WELLS AND PUHH 5-4- 35- a. g 3 u S 23- CRUISE 1 116 PASSENGERS 3AMPLE - - CRUISE 2 629 PASSENGERS SAMPLED r 13 1 2«27 ' 1 26 29 29 3 31 DEC OEC JAN. 1973 1973 1976 FIGURE 1. Ill passengers, by time of onset of diarrhea, on a Miami-based cruise ship. labile enterotoxin production within three weeks after collection. The results (table 4) demonstrate the significant relationship between isolation of E. coli serotype 25:NM producing only LT (the epidemic strain) and illness in passengers on the first cruise and crew on both cruises. Because we could not demonstrate a significant relationship between isolation of the epidemic strain and illness among passengers on the second cruise, we looked closely at the percentage of E. coli isolates producing LT in well TABLK l 1 and ill passengers. On the first cruise 55 (69 per cent) of 8 isolates from ill passengers were serotype 25:NM and produced LT; 11 (39 per cent) of 28 isolates from well passengers were serotype 25:NM and produced LT (p <.1). On the second cruise 77 (81 per cent) of 95 isolates from ill passengers were serotype 25:NM and produced LT; five ( per cent) of 25 isolates from well passengers were serotype 25:NM and produced LT (p <.1). All of the isolates of the epidemic strain tested for antibiotic sus- Symptoms associattd with diarrhea in passengers and crew culture-positive for heat-labile enterotoxigenic Escherichia coli 25:non-motile during two successive cruises of a Miami-based cruise ship Symptom Abdominal cramps Nausea Vomiting Headache Fever (subjective) Bloody diarrhea Passengers on first cruise* surveyed {N 156» Had symptom % 136 126 61 94 51 87 81 39 6 33 December 26-29, 1975. t December 29, 1975-January 2, 1976. t No crew questionnaires were distributed on the first cruise. Passengers and crew on second cruiset surveyed UV - 1168) Had symptom * 891 575 25 47 258 8 2 76 49 21 4 22.7
SHIPBOARD ENTEROTOXIGEN1C E. COU INDUCED DIARRHEA 435 TABLE 2 Water consumption by passengers* on second cruiset of a Miami-based cruise ship Average no. of glasses consumed per day 1-2 3-4 5+ No. ill 18 78 21 Total 146 344 199 57 % iu 25 31 39 * Excludes 83 passengers whose water consumption is unknown. Chi-square due to linear trend (p =.33). t December 29, 1975-January 2, 1976. ceptibility were uniformly resistant to tetracycline and sulfathiazole, but susceptible to nitrofurantoin, chloramphenicol, streptomycin, cephalothin, gentamicin, naladixic acid, ampicillin, colistin, neomycin, and kanamycin. None of the food or environmental samples contained the epidemic strain. In addition to the epidemic strain isolates, nontoxigenic E. coli serotype 25:NM organisms resistant to sulfathiazole and tetracycline were recovered from three patients, two of whom were simultaneously harboring the epidemic strain. Only two of eight paired sera obtained from 11 culture-positive passengers demonstrated a fourfold rise in antibody titer tolt. DISCUSSION Strains of ETEC which produce only LT have previously been isolated from individual patients with diarrhea (9). Although strains of ETEC that produce LT and ST or ST alone have been recovered in previous outbreak investigations (1, 11), the importance of ETEC that produce only LT as significant agents of human enteric disease has not been as well documented. Because of the association between this organism and illness, we believe that it was solely responsible for two successive shipboard outbreaks of diarrhea. That a greater proportion of an ill person's flora when compared to a well person's flora consisted of this strain of ETEC emphasizes the pathogenic role TABLE 3 Cross-table analysis for three food items passengers consumed on December 29 3, 1975, aboard a Miami-based cruise ship Ate prime rib* Did not eat prime rib Ate shrimp cocktailt Did not eat shrimp cocktail m 17 23 118 *p <.1. t p <.1 (Mantel-Haenszel test). Ate crabmeat cocktail Well Total %I11 181 43 32 288 66 318 52 35 38 Did not eat crabmeat cocktail Ul Well Total % Ul 48 22 25 45 157 91 64 184 5 113 89 229 23 28 TABLE 4 Number of persons positive for Escherichia coli 25 :non-motile producing heat-labile enterotoxin who sailed on a Miami-based cruise ship Cruise 111 Passengers Well Fisher's exact test 111 Crew Well Fisher's exact test 1st 2nd Total 12/16*(81)t 16/19(84) 29/35(83) 4/1(4) 2/5(4) 6/15(4) p <.5 p >.5 p <.1 5/6(83) 9/1(9) 14/16(88) 178(12) /6() 1/14(7) p <.5 p <.1 p <.1 * Number tested. t Percentage positive.
436 LUMISH, RYDER, ANDERSON, WELLS AND PUHR these organisms played. The finding of the epidemic strain in six of 15 well passengers suggests that either they had subclinical infections or they were still within the incubation period for diarrhea and became ill subsequent to debarking. Our ability to demonstrate a fourfold rise in antitoxin antibody in only two of eight culture-positive ill passengers confirms work done by other investigators, demonstrating the frequent lack of antitoxin antibody response in persons living in the United States (8, 9, 12). Because a few persons were noted to carry both LT producing and nonenterotoxin producing serologically identical non-motile E. coli, we conclude that loss of gene(s) for LT enterotoxin production may have occurred in these strains. Although the isolated strain of E. coli producing LT was thought to be the cause of these outbreaks, the mode of transmission remains unclear. That 4 per cent of the well passengers harbored the epidemic strain suggests that the majority of passengers were exposed to the vehicle(s) of transmission. Since many persons harboring the epidemic strain were classified as "well" on these very short cruises when their questionnaires were tabulated, it is difficult to clearly define the mode of transmission. The positive association between drinking water and illness suggests that water was a possible source of spread. The finding of residual free chlorine in the distribution system at the time of the inspections makes water a less likely vehicle, but one cannot rule out contamination of the water supply earlier in the cruise. Alternatively, the crabmeat cocktail may also have facilitated transmission of the epidemic strain even though one-third of all ill passengers claimed not to have eaten this item. Despite our inability to define the mode of spread of illness, this investigation demonstrates that ETEC that produces only LT can cause epidemic diarrheal disease and confirms that the clinical picture produced by different types of ETEC is similar. REFERENCES 1. Morris GK, Koehler JA, Gangarosa EJ, et al: Comparison of media for direct isolation and transport of shigella from fecal specimens. Appl Microbiol 19:434-4, 197 2. Edwards PR, Ewing WH: Identification of Enterobacteriaceae. Third edition. Minneapolis, MN, Burgess Publishing Company, 1972 3. Feeley JC, Balows A: Vibrio. In Lennette EH, Spaulding EH, Truant JP (eds): Manual of Clinical Microbiology. Second edition. Washington DC, American Society for Microbiology, 1974, pp 238-245 4. Sack DA, Sack RB: Test for enterotoxigenio scherichia coli using Yl adrenal cells in miniculture. Infect Immun 11:334-336, 1975 5. Dean AG, Ching Y, Williams RG, et al: Test for Escherichia coli enterotoxin using infant mice: Application in a study of diarrhea in children in Honolulu. J Infect Dis 125:47-411, 1972 6. Sereny B: Experimental Shigella keratoconjunctivitis. Acta Microbiol Acad Sci Hung 2:293-296, 1955 7. Bauer AW, Kirby WMM, Sherris JC, et al: Antibiotic susceptibility testing by a standardized single disk method. Am J Clin Pathol 45:493-496, 1966 8. Sack RB, Hirschbom N, Woodward WE, et al: Antibodies to heat-labile Escherichia coli enterotoxin in Apaches in Whiteriver, Arizona. Infect Immun 12:1475-1477, 1975 9. 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 294:1299-135, 1976 1. Rosenberg ML, Koplan JP, Wachsmuth IK, et al: Epidemic diarrhea at Crater Lake from enterotoxigenic Escherichia coli. A large waterborne outbreak. Ann Intern Med 86:714-718, 1977 11. Ryder RW, Wachsmuth IK, Buxton AE, et al: Infantile diarrhea produced by heat-stable enterotoxigenic Escherichia coli. N Engl J Med 295:849-853, 1976 12. Wachsmuth IK, Wells JG, Ryder RW:, scherichia coli heat-labile enterotoxin: Comparison of antitoxin assays and serum antitoxin levels. Infect Immun 18:348-351, 1977