On July 18, 2001, the New York City Department of Health and Mental Hygiene received a complaint of illness from a person who ate at a wedding celebration on July 14; S. Uganda was isolated from the stool of another wedding attendee. By early August, a distinct strain of S. Uganda had been isolated from 11 New York City residents with illness onsets occurring June 24–August 4 (Figure 1). All 11 case-patients were of Hispanic ethnicity, and 6 of 10 interviewed reported having eaten roast pork from a New York City restaurant in the 3 days before illness onset. Additionally, roast pork from that restaurant had been served at the wedding named in the initial consumer complaint. A sample of leftover roast pork from the wedding was positive for the same strain of S. Uganda as the one isolated from patients.

Preparation of the roast pork at the restaurant consisted of seasoning 30–40 lb of raw pork in buckets with salt, pepper, oregano, and garlic. The pork was subsequently cooked in an oven for 2 h at 500°F (260°C), then for 1 to 2 h more at 350°F (177°C). After being cooked, the pork was stored in a hot-holding unit at the front of the restaurant and cut into pieces upon request.

At the time of a sanitary inspection initiated by the consumer complaint, raw pork was held at inadequate temperatures at the restaurant, and thermometers were inadequately used during cooking and hot-holding. Potential sources of cross-contamination, surfaces and wiping cloths, were not properly sanitized. The same S. Uganda strain found in patients was isolated from a cooked pork sample collected from the restaurant on July 18. Raw pork sampled at that time was positive for Salmonella, but not S. Uganda. The restaurant was closed on August 6 and reopened on September 26, after food safety violations were corrected and pork samples were confirmed negative for Salmonella.

From August 22 to 23, 2001, the Chicago Department of Public Health received reports of four isolates of S. Uganda obtained from Chicago residents. By early September, 12 confirmed cases of S. Uganda infection had been identified in Chicago with stool collection dates from August 13 to September 6. All 12 patients were of Hispanic ethnicity, and each reported having eaten carnitas (fried pieces of pork) in the 3 days before illness onset. Carnitas were the only food item reported to have been eaten by all the patients. For 10 (83%) patients, the source of the carnitas was a Chicago grocery store. Of the two patients who did not identify that grocery, one had bought carnitas from a mobile lunch truck located outside of his workplace; the source of these carnitas could not be verified. The other patient was visiting from out of state and could not name the source of the carnitas.

Preparation of carnitas at the Chicago grocery consisted of cutting 90 lbs. of raw pork into 15 cm x 15 cm x 10 cm slabs, frying them in pork lard for 2.5 h, and adding salt and lemon juice 15 min before the end of frying. Cooked carnitas were then chopped, shredded into smaller pieces, and placed in a hot holding tray on top of the meat cooler. Using tongs, a customer would place carnitas into a Styrofoam container, then hand it to a butcher for weighing.

No Salmonella was isolated from samples of cooked carnitas, raw pork, the holding tray, or the serving tongs obtained from the Chicago grocery on August 24. S. Uganda was isolated from stool from 1 of 10 butchers. He reported having worked while having a mild gastrointestinal illness around August 15 and said he frequently ate carnitas from the Chicago grocery. The infected butcher was restricted from work until two consecutive stool specimens, collected 48 h apart, were negative for Salmonella. The workers and grocery manager were advised to avoid contact with food when sick.

The investigation of the Chicago grocery showed that processed meats were being stored at improper temperatures and that potential sources of cross-contamination were not controlled (i.e., cutting boards and utensils were inadequately sanitized, and the same scale was used for raw meats as well as cooked pork). Methods of preventing cross-contamination after cooking were stressed. Upon reinspection, the violations had been corrected.

Comparison of the outbreak isolates by PFGE showed a three-band difference (Figure 2). The sources of the pork implicated in the outbreaks were traced to two separate distributors; a more extensive traceback (i.e., to the packing-plant or farm source) was not conducted.