Spatial Epidemiology of Salmonellosis in Florida, 2009–2018

Abstract

Non-typhoidal Salmonella enterica infections cause a high disease burden in the United States with an estimated 1.2 million illnesses annually. The state of Florida consistently has a relatively high incidence compared to other states in the United States. Nevertheless, studies regarding the epidemiology of nontyphoidal salmonellosis and its spatial and temporal patterns in Florida were rarely reported. We examined the spatial and temporal patterns of 62,947 salmonellosis cases reported to FL Health Charts between 2009 and 2018. Dominant serotypes circulating in Florida were also explored using whole genome sequencing (WGS) based serotype-prediction for 2,507 Salmonella isolates sequenced by the Florida Department of Health during 2017 and 2018. The representativeness of laboratory-sequenced isolates for reported cases was determined by regression modeling. The annual incidence rate of salmonellosis decreased from 36.0 per 100,000 population in 2009 to 27.8 per 100,000 in 2016, and gradually increased in 2017 and 2018. Increased use of culture-independent testing did not fully explain this increase. The highest incidence rate was observed in children, contributing 40.9% of total reported cases during this period. A seasonal pattern was observed with the incidence peaking in September and October, later than the national average pattern. Over these 10 years, the Northeast and Northwest regions of the state had higher reported incidence rates, while reported rates in the Southeast and South were gradually increasing over time. Serotypes were predicted based on WGS data in the EnteroBase platform. The top-five most prevalent serotypes in Florida during 2017–2018 were Enteritidis, Newport, Javiana, Sandiego and Braenderup. The highest percentage of isolates was from children under 5 years of age (41.4%), and stool (84.7%) was the major source of samples. A zero-inflated negative binomial regression model showed that the reported case number was a strong predictor for the number of lab-sequenced isolates in individual counties, and the geospatial distribution of sequenced isolates was not biased by other factors such as age group. The spatial and temporal patterns identified in this study along with the prevalence of different serotypes will be helpful for the development of efficient prevention and control strategies for salmonellosis in Florida.

Molecular Epidemiology of Salmonellosis in Florida, USA, 2017–2018

“…We present a retrospective analysis of whole genome sequencing data of Salmonella enterica isolates from Florida in the years 2017-2018. Our earlier work has shown that the sequenced isolates in 2017-18 were a relatively unbiased sample of confirmed clinical cases (4). The population structure of Salmonella is Florida was found to be highly diverse with 115 serotypes observed in a set of 1,632 sporadic clinical isolates obtained over this 2-year period.…”

“…Metadata Collection and ManagementAn integrated and deidentified metadata set from BPHL Bionumerics and FDOH Merlin database was made available by FDOH under a data sharing agreement, details are provided in Li et al(4). We selected isolates from metadata received from FDOH, with a date of sample collection by the clinical laboratory from January 2017 to December 2018 and downloaded sequence data from Enterobase (https://enterobase.warwick.ac.uk) using PNUSA numbers specified in the metadata (https://dataverse.…”

“…harvard.edu/dataverse/Salmonella_FL). After data cleaning [as detailed in Li et al(4), involved removal of isolates with missing data or duplicated in the database] a curated set of 1,709 isolates was available for further analysis. We additionally included WGS data from Enterobase of 230 isolates collected from non-clinical sources (including soil/water, food, poultry, animal sources) in Florida and 1,317 S. Enteritidis isolates collected from poultry sources from other states during same time period, using identifiers from NCBI.…”

Geographical distribution and space–time clustering of human illnesses with major Salmonella serotypes in Florida, USA, 2017–2018

“…Serotype prediction was performed in the EnteroBase online server (https://enterobase.warwick.ac.uk/) [19] as described in [6]. We included the top five Salmonella serotypes in Florida, plus two additional serotypes, Typhimurium and I 4, [5],12:i:-(a monophasic variant of Typhimurium), that are on the list of top five serotypes across the whole USA (Table 1).…”

“…Some serotypes are mainly transmitted by food, such as the serotypes Saintpaul, Heidelberg and Berta, but others such as Bareilly and Mississippi appear to be less linked to food pathways [8]. The five most common serotypes identified through the FoodNet surveillance system in the USA are Enteritidis, Typhimurium, Newport, Javiana and I 4, [5],12:i:-, a monophasic variant of Typhimurium [9], while the major serotypes occurring in Florida are Enteritidis, Newport, Javiana, Sandiego and Braenderup [6]. In our previous work, we conducted a phylogenetic analysis of Salmonella isolates from five dominant serotypes in Florida and identified the geographical distribution of clusters of genetically related isolates using core genome multi-locus sequence typing (cgMLST), a subtyping approach that provides high resolution to distinguish isolates within the same serotype [10].…”