Two Related, Multi-Strain Salmonella Outbreaks in the Arkansas Prison System – August, 2012

BACKGROUND:  Salmonella is the most common cause of bacterial foodborne illness in the United States and in Arkansas. In August, 2012, the Arkansas Department of Health (ADH) investigated two large outbreaks of gastrointestinal illness in the Arkansas prison system. ADH learned of a large number of inmates experiencing illness in Prison A on August 6. On August 21, ADH identified a second outbreak in Prison B.

METHODS:  A case-control study was conducted at each of the two prisons. Controls were matched to cases by barrack. Epidemiologists interviewed convenience samples of inmates and staff to collect food history data and symptom histories. Probable cases were defined as persons with self-reported diarrhea. Confirmed cases had laboratory evidence of Salmonella in stool. Epidemic curves were constructed and food history data were analyzed. Environmental Health Specialists inspected the prison kitchens and dining areas. Food samples and stool samples from ill inmates and all kitchen workers were collected for laboratory analysis, including Salmonella culture, serotyping, and pulsed-field gel electrophoresis (PFGE).

RESULTS:  

Analysis from Prison A suggested that chicken salad was the food item responsible for illness (mOR=7.5). Interviews and video evidence revealed that the cooked chicken remained at room temperature for approximately 16 hours before consumption. The epidemic curve suggested point-source contamination.

Analysis of Prison B data revealed multiple food items associated with illness. Again, chicken salad was associated (mOR=4.0), followed by breakfast franks (mOR=2.5), but no food item explained the preponderance of cases. Delayed reporting of the outbreak contributed to delays in excluding 99 inmate cases who worked in the kitchen while ill. Further, the epidemic curve for Prison B suggested mixed modes of transmission, including person-to-person or contamination of multiple food items, possibly by infected kitchen workers.

In total, 597 probable/confirmed cases were identified in the two prisons. Multiple species of Salmonella were isolated from inmate stool samples (10 serotypes with 14 distinct PFGE patterns). Notably, five of these patterns, including three never before reported in PulseNet, were identified in both prisons, definitively linking the two outbreaks.

CONCLUSIONS:  

These outbreaks were caused by multiple strains of Salmonella. Five indistinguishable strains by PFGE were present at both prisons. Distinctly different modes of transmission were observed in each prison despite the temporal proximity of the outbreaks. Mishandling of food and delayed exclusion of ill inmate kitchen workers likely contributed to the outbreaks. Although not definitively identified, a contaminated food item or product distributed to both prisons was likely involved.