LOCATION OF FOOD CONSUMPTION AND TRAVELERS' DIARRHEA

* AMERICAN JOURNAL or EPIDIMIOLOOY Vol. 6, No 4 Copyright O 977 by The Johns Hopkins University School of Hygiene and Public Health Printed in USA j LOCATION OF FOOD CONSUMPTION AND TRAVELERS' DIARRHEA W. S. TJOA, H. L. DuPONT, P. SULLIVAN, L. K. PICKERING, A. H. HOLGUIN, J OLARTE, D. G. EVANS AND D J EVANS, JE "> Tjoa, W. S., H. L. DuPont (U. of Texas Medical School, Houston, TX 773), P. Sullivan, L. K. Pickering, A. H. Holguln, J. Olarte, D. G. Evans and D. J. * Evans, Jr. Location of food consumption and travelers' diarrhea. Am J Epidemiol 6:6-66, 977. Dally food histories for one month were obtained In summer, 975, from students attending a Mexican university to determine the Influence of food, consumption on the development and etiology of diarrhea. In newly-arrived students from the U.S. who ate half or more of their meals in the school K cafeteria and public restaurants there were significant increases in diarrhea (p <.5), shlgella infection (p <.5) and toxlgenic E. coll Infection (p <.5) compared to the students eating a comparable number of meals in private homes. In the summer U.S. students there was also an association of diarrhea and eating from street vendors (p <.5). In full-time U.S. students V who had lived in Mexico a year or longer as well as in Latin American students a relationship between location of meals and occurrence of enteric disease " was not apparent. High numbers of enteric bacteria were recovered from food from the school's cafeteria, public restaurants, street vendors and small grocery stores. Shlgella were isolated from cooked and uncooked hamburger patties from the school cafeteria. Four shigella carriers were found among kitchen personnel at the school. This study demonstrates that food serves as a major vehicle through which travelers' diarrhea occurs. t diarrhea; disease outbreaks; travel Diarrhea of travelers usually develops Microbial pathogens have been focused within 4 days of arrival in a foreign coun- upon as etiologic agents of travelers' diartry. It is worldwide in distribution but ap- rhea. Recent investigations have shown pears to be more prevalent in tropical and evidence of enterotoxigenic Escherichia sub-tropical countries with poor sanitation coli as the single most important cause of r and hygiene. The condition is character- travelers' diarrhea (-3). While tourists ized by acute onset of watery diarrhea with the absence of blood and mucus in quest8: Dr. H. L. DuPont, 64 West Cullen Street, stools and often with nausea, vomiting, Houston, TX 773.) abdominal cramps, malaise, chills or fe- ^ ^PP ^,!* f * *T * e National instii, i. tutes of Health (AI 699-) and a grant from the ver. The disease generally lasts for two to Norwich Pharmacal Company, Division of Mortonr three days and in nearly all cases it is self- Norwich Products, Inc. limited Presented at the Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, IL October 9, 976. Received for publication December 7, 976, and The authors acknowledge the support of the facin final form March,977. ulty, staff and students at Universidad de las Amer- 4 Abbreviations; EMB agar, eosin methylene blue icas, and the cooperation of Drs. M. Rendon, H. agar; LT-E. coli, heat labile enterotoxic Escherichia Kirkpatrick, E. Simmen, E. Rice-Wray, P. Carrillo, / coli; S-S agar, Salmonella-Shigella agar. and C. Moctezuma. G. Haynes and E. Galindo From the University of Texas Medical School at worked on the study. Dr. B. P. Hsi helped with the Houston and The University of Texas School of Pub- statistical analysis and Dr. C. S. Hacker assisted lie Health, Houston, TX and Hospital Infantil de with the study design. K. Boyle helped to prepare Mexico, Mexico City, D. F., Mexico. (Reprint re- the manuscript. 6

6 TJOA ET AL commonly believe the condition is due to contaminated drinking water, there is little evidence to incriminate water in the common form of "turista". Food, with its capacity for supporting bacterial growth, was felt by us to be a more logical source. The present investigation was designed to correlate locations of food consumption with the development of enteric infection among a student population in Mexico. MATERIALS AND METHODS Two hundred eighty students attending summer school at the University of the Americas in Cholula, Puebla, Mexico, in the summer of 975 were prospectively enrolled in the study. They were comparable in age and health status. Two hundred sixty completed the 3-day study, giving a dropout rate of 7. per cent. This student population consisted of 49 newly-arrived U.S. summer students, 3 full-time U.S. students who had been at the school one year or longer, and 79 Latin American students from Venezuela or Mexico. The newly-arrived U.S. students were enrolled in the study within 4 hours of arrival at the school. A stool sample was collected on the day of arrival or as soon as their bowel habits permitted. A stool sample was collected from all students every three days for 3 days and each student was evaluated daily for the occurrence of diarrhea. Rectal swabs were not taken. Daily food history sheets containing information regarding type of food eaten and the date and place where each meal was eaten, including food eaten from a street vendor, were recorded by each subject. In the study, diarrhea was defined as greater than twice the usual number of unformed bowel movements within a hour period plus the presence of either fever, nausea, vomiting, abdominal cramps, or malaise. Students reporting illness were objectively evaluated for diarrhea; unformed stools had to be observed by the field team. When diarrhea was diagnosed the location of meals consumed the preceding week were identified. Students were grouped according to the general location of 5 per cent or more of the meals that week. This per cent was selected to avoid bias in data analysis. Laboratory procedures The stool specimens were collected in clean cartons provided by the field team. Stools were cultured for enteric pathogens. For the isolation of enteric pathogens, a stool was cultured on eosin methylene blue (EMB) agar, sodium tetradecyl sulfate (Tergitol 7), and Salmonella-Shigella agar (S-S). Shigella colonies were identified by routine techniques of Edwards and Ewing (4). Five colonies of Escherichia coli were transferred to nutrient agar stabs for transport back to Houston for studies of toxigenicity as recently described (3). Heat labile enterotoxigenic E. coli (UT-E. coli) infection was diagnosed by finding LT-E. coli in stool during illness and being absent on baseline cultures or by documenting fourfold serum antibody rises to the toxin (5). Food was collected in styrofoam specimen cups from the various sources, including restaurants, small stores, open markets, street vendors and the school cafeteria. Samples from private homes were not obtained in the present study. Ten grams of each sample were weighed on a Chaus Triple Beam Balance. All samples were homogenized with 9 ml of buffer solution in a sterile Waring blender for min and processed according to a published method (6). Dilutions of : and : were made from the original homogenate. A sample ( ml) of each of the -fold and -fold homogenate dilutions was mixed with 9 ml of EMB agar and plated immediately in disposable petri dishes.. ml of the samples (:, :, :,) were plated on EMB agar. All plates were incubated at 33 C for 48 hr. The enteric bacteria were enumerated and identified according to Edwards and Ewing (4). Stools were obtained from of the

FOOD CONSUMPTION AND TRAVELERS' DIARRHEA 63 kitchen personnel in the school cafeteria for isolation of enteric pathogens as a single survey. Data were analyzed by the chi-square test of independence. No adjustments were made for comparing groups of different size or composition. RESULTS Diarrhea development was associated with eating meals at the various local restaurants and the school cafeteria (see table ). Diarrhea occurred in 3 of the 9 U.S. summer students (45 per cent) who ate more than 5 per cent of their meals in the school cafeteria and/or public restaurants, compared to three of the U.S. summer students (5 per cent) who ate most of their meals in a private home (p <.5). Shigella infection was significantly greater (p <.5) in U.S. summer students eating in the cafeteria or restaurants (3 per cent) when compared to those eating in private homes (5 per cent). In contrast, there was no correlation between location of food consumption and occurrence of diarrhea or shigella isolation among full-time U.S. students or Latin Americans (Venezuelans and Mexicans). Table shows the occurrence of toxigenic LT-E. coli infection (which occurred throughout the month of study) and location of major food consumption the week preceding the infection. Toxigenic E. coli infection was related to eating food at the school cafeteria and public restaurants. A TABLE Occurrence of diarrhea and shigella isolation from stool, according to location of major food consumption during a 3-day period, m 6 students attending a Mexican university, summer, 975 Student status U.S Summer Full-time Latin American 7 5 98 In private homes 3(5) (6) 6() 9 (5) Greater than fifty per cent of meals Shipella isolation (%) (5) 4 (9) 7(4) 3 (6) In school cafeteria and/or public restaurants students 9 5 8 6 Diarrhea <%) 3 (45) () 3() 7(7) Shigella isolation (%) 9 (3) () (7) (9) TABLE Occurrence of toxigenic LT-E. coli infection, according to location of major food consumption during the preceding week, in 6 students attending a Mexican university, summer, 975 Student status U.S. Summer Full-time No students 7 In private homes Greater than fifty per cent of meals LT infection (*) 3() In school cafeteria and/or public restaurants students 9 5 LT infection (%) (34) Latin American 5 (4) 8 (4) 98 5(3) 6 (8)

64 TJOA ET AL. significantly greater number of U.S. summer students eating half of their meals in these establishments developed diarrhea due to.e. coli when compared to U.S. summer students eating in a home (p <.5). No apparent relation was noted between food consumption and LT-E. coli infection in full-time U.S. and Latin American students. A significantly (p <.5) greater number of U.S. summer students who ate food obtained from a street vendor at any time during the month of surveillance experi-' enced diarrhea when compared to U.S. summer students who denied eating food from street vendors (see table 3). Full-time U.S. and Latin American students failed to show the same relationship between eating from street vendors and occurrence of diarrhea. Infections due to shigella strains and toxigenic E. coli were not related to food consumption from food vendors in any student group. Qualitative study of the food and water samples from the school cafeteria, restaurants, and street vendors showed widespread contamination with enteric bacteria (table 4). Seven water samples were bacteriologically sterile while two ice blocks cultured from an open market stand were contaminated with Shigella C and other enteric organisms. Salmonella was recovered from a piece of watermelon bought from a street vendor. Shigella organisms were isolated from uncooked hamburger patties and cooked hamburger from the school cafeteria. Shigella was isolated from stools of four kitchen personnel employed in the school cafeteria. A separate quantitative study on food cultured from local restaurants, street vendors, and small stores showed high counts of Escherichia coli in the food (table 5). E. coli was recovered in 6 out of 67 samples from food served in restaurants (39 per cent), six out of ll samples from street vendors (55 per cent), and in two out of five samples bought from small grocery stores (4 per cent). The highest counts of E. coli were found in dairy products such as cheese, milk and butter. However, coliform contamination also was detected in cooked foods such as chicken enchiladas, beans with cheese, fried beef in sauce and a hot dog. DISCUSSION Various investigators have shown a high degree of probability of. developing diarrhea among U.S. travelers to Latin America. Evidence is accumulating to suggest that toxigenic E. coli is the single most important cause of this illness (-3). The reservoir and mode of spread are unknown. This study presents data that strongly associate food consumption with TABLE 3 Occurrence of diarrhea, shigella isolation from stool and LT-E. coli infection in 8 students attending a Mexican university in summer, 975, who ate.or did not eat from street vendors during a 3-day period* Student status U.S Summer Full-time students 77 Ate from street vendors Diarrhea (48) 7 () Shigella isolation (*) 5(4) 7 () ' LT-. cou 4(9) (3) students 33 66 Did not eat from street vendors Diarrhea 7() 8() Shigella isolation 6(8) (8) LT-E coli (*) 5(5) () r i Latin American 9 3() 3() (3) 54 6() 6() 3(6) 7 3(4) 5() 7(6) 53 (4) 4(6) * Data given for all 8 students prospectively enrolled in the study, including students who did not remain in the study for the full 3 days. 9(6)

FOOD CONSUMPTION AND TRAVELERS' DIARRHEA 65 TABLB 4 Food and water cultures from Cholula and Puebla (town and city adjacent to school m Mexico attended by 8 students prospectwely enrolled in study) Location Open market Descnption of food Marinated Beef Chile Pepper Ice Blocks () Bacteria isolated Shigella C, Klebsiella, Eschenchia coli Meat (butcher's) shop Sausage (cooked) Klebsiella Street vendors (4) Field (growing wild) Restaurant # Restaurant # School cafeteria School clinic and dormitories Source of food Restaurants Street vendors Small stores Tortilla with sauce Fruit Watermelon Custard Mushroom Tap water Pork steak Salad Hahiburger (cooked) Raw beef patties Chopped meat Raw liver Catsup and salad dressing Environmental surfaces Bottled water Tap water Bottled water Tap water Salmonella species Shigella B Enterobacter Klebsiella, Pseudomonas aeruginosa Shigella B Shigella B, Paracolon, Klebsiella sp Klebsiella Paracolon, Citrobacter, Klebsiella Klebsiella Klebsiella, Pseudomonas, Paracolon, Enterobacter TABLE 5 Escherichia coli recovered from food cultured from various sources sampled 67 5 E coli recovered 6 6 % 39 55 4 Colony counts per gm IP IIP 8 4 6» 83 34 4 8 3 the diarrhea of U.S. travelers to Mexico. There was a relationship among newlyarrived U.S. students between eating food at a school cafeteria and public restaurants and developing enteric disease (including diarrhea, shigellosis and toxigenic E. coli infection). The students who prepared their own food in a private home experienced less enteric disease. Eating food purchased from street vendors was associated with the development of diarrhea but not shigellosis nor LT-E. coli infection among these newly-arrived students.

66 TJOA ET AL. Diarrhea, shigellosis and UT-E. coli infection were not related to location of food ingestion for those U.S. students who had remained in Mexico a year or longer or for the Latin American students. However, enteric diseases in general were far less common in these students, who presumably had been exposed chronically to the infectious agents. Food was found to be fecally contaminated in the school cafeteria and in restaurants, further supporting our contention that food is a major vehicle for "Turista". The presence of high counts of E. coli in our food sampling indicates contamination from human or animal sources. Poor sanitary practices by the food handlers probably caused the contamination of food and subsequent growth of the microorganisms. Presence of these indicator bacteria is considered as evidence of inadequate heat treatment, contaminated equipment, and/or post heat treatment contamination (7). A recent investigation suggested that salads containing raw vegetables were implicated as a vehicle of toxigenic E. coli infection among U.S. travelers to Mexico (). Studies must now be undertaken to document the toxigenicity of fecal and enteric organisms isolated from food, determine the reasons for such widespread contamination and, hopefully, implement means for improving food sanitation. In the meantime, the traveler to Latin America is advised to carefully choose the place where he eats his meals. Food from street vendors should be considered as one of the vehicles of transmission for enteropathogens in diarrheal illness. REFERENCES Gorbach SL, KeanBH, Evans DG, et al: Travelers' diarrhea and toxigeruc Eschenchia coli. N Engl J Med 9.933-936, 975. Mereon MH, Moms GK, Sack DA, et al: Travelers' diarrhea in Mexico. N Engl J Med 94:99-35, 976 3. DuPont HL, Olarte J, Evans DG, et al: Comparative susceptibility of Latin American and U S. students to enteric pathogens. N Engl J Med 95:5-5, 976 4. Edwards PR, Ewing WH: Identification of Enterobacteriaceae Minneapolis, MN, Burgess Publishing Co., 97 5 Ruiz-Palacios G, Evans DJ, Jr, Evans DG, et al. Humoral antibodies to E coli enterotojrin and stool isolation of toxigenic E. coli in students at a Mexican university. Abstract presented at Intersrientific Conference on Antimicrobial Agents and Chemotherapy, October 9, 976 6 Direccion General De Investigation En Salud Pubhca: Technicas Para El Muestreo Y Anlisis Microbwlogico De Alimentos Mexico City, Direccion General De Investigation En Salud Publica, 975 7. Davis JG: Microbiological standards for foods. Part. Lab Practice 8:839, 969 i