In August 1991, surveillance of food poisoning in Riyadh, Saudi Arabia, detected twice in one week a single broasted (fried) chicken restaurant as a suspected source of gastroenteritis. Four other broasted chicken restuarants were also reported during August compared to only 1 during the previous 12 months. We defined a case as gastroenteritis identified either through food poisoning reports, health center or emergency room log books, salmonella surveillance, or interviews of affected families in a person who ate food from a broasted chicken restaurant during the 2 days before onset of symptoms. We obtained food histories for all members of each family with a case. Food specific attack rates and relative risks for each food item from each restaurant were computed and tested for statistical significance using Fisher's exact test. We detected 92 cases and 19 well family contacts (attack rate = 82.9 %) with exposure to 5 restaurants. Among the 19 families the median attack rate was 100%. We found no single case families. Salmonella sp. was isolated from the stool of 32 cases from one restaurant. For this restaurant cases were associated with eating broasted chicken (RR = 2.82, 95% CI = 0.87 to 9.16). Several women in these families reported that the chicken was not completely cooked near the bone. Cases (40) from other restaurants (4) were also associated with chicken but we lacked enough well family contacts (3 only) for valid statistical conclusions. Broasting involves cooking in pressurized oil at 150 degrees C for 8 minutes. The selection of families rather than individual customers and reports of incompletely cooked chicken by these families suggests that the broasting machine might have been overloaded. The high attack rates also point to heavy contamination. Municipal authorities responsible for investigation of the restaurants and the source of chicken were advised.
Conclusion: