Abstract 1016
Poster Session IV, Tuesday, 5/4 (poster 124)
The majority of cases of human salmonellosis in Arkansas are due to S. newport (SN) and S. typhimurium (ST). Most infections are thought to be sporadic cases of human disease, but acute outbreaks with either of these serotypes would be difficult to identify using standard epidemiologic tools. Recently, it has been demonstrated that pulsed-field gel electrophoresis (PFGE) is useful in identifying genetically identical non-typhoidal Salmonella isolates. This technique could be useful in distinguishing outbreak-related from non-outbreak-related isolates and would be helpful in guiding a more focused epidemiological investigation. The purpose of this study was to apply both standard and molecular epidemiological methods in investigating human salmonellosis cases in our State. All isolates of SN or ST submitted to the Arkansas Department of Health (ADH) from 7/1/97-6/30/98 were gathered and underwent PFGE. Each pattern was recorded, compared and banked by computer. Patients were contacted and completed a questionnaire. Statistical calculations were made using the Wilcoxon rank-sum test, Fisher's exact test, and Pearson's chi-squared. During the study period, 191 isolates of SN and ST were reported. This represented 54% of all isolates reported to the ADH. Twenty-four isolates were excluded (20-repeat isolates, 4-environmental isolates) leaving 167 isolates (84-SN, 83-ST) from separate individuals. Forty-three individuals did not complete the questionnaire leaving 124 subjects (65-SN, 59-ST). Patients infected with SN were younger (11.1 years vs 17.6 years; p<0.05), more likely to be white (90% vs 66%; p<0.01), and have a history or pet ownership (74% vs 53%; p<0.02). Eighteen patients infected with SN could be grouped into 7 clusters of identical PFGE patterns. Standard epidemiological information linked patients in 3 of the clusters while patients comprising the remaining clusters were identified only with PFGE. For those infected with ST, 15 patients could be grouped into 6 clusters. Only patients in one of these clusters were linked with standard epidemiological tools. Patients in the 5 remaining clusters were only linked with the use of PFGE. Although the majority of all clusters could be related in time or location, no common risk factor could be identified in any. These data demonstrate that approximately 20% of individuals infected with SN or ST in Arkansas can be linked to at least one other individual. Only one-third of these individuals however, were linked using standard epidemiological tools. PFGE greatly enhanced the ability to identify patients infected with identical isolates. The use of PFGE confirms that only 20% of the two most common Salmonella serotypes in our State can be linked to another individual and reiterates the difficulty in attempting to find a common source of infection for all cases. Any attempt to decrease human infections with Salmonella must have interventions at many levels to impact human disease.
