Salmonella in broiler breeder flocks in Israel during 2008-2018: Prevalence, risk factors and repeated detections Authors: Anat Wiseman, Hadas Poria Daly, Ehud Elnekave Published: 2026 ABSTRACT Non-typhoidal Salmonella infections in broiler breeder flocks pose a major economic challenge to poultry production and increased risk to public health. This study aimed to determine the prevalence and risk factors for infections in commercial broiler breeder flocks in Israel, and the risk for repeated detections.Data comprised a census of 4,100 flocks from 142 farms, with 64,217 test results derived from national surveillance program records collected during 2008-2018. Prevalence was calculated at the test, flock, and farm level. Univariable and multivariable mixed-effect logistic regression models were used to assess risk factors for detection at the test level, accounting for clustering within flocks and farms. Repeated detections were analyzed separately at the flock and farm levels using descriptive statistics, and logistic regression models. In addition, the prevalence and incidence of dominant serotypes infecting broiler breeder flocks and humans, respectively, were compared.On average, 6.7% of tests were Salmonella-positive annually, and 29.6% of flocks and 59.6% of farms tested positive at least once per year. Overall, 91 serotypes were identified, with five dominant serotypes (Infantis, Hadar, Kentucky, Muenchen and Virchow) accounting for 67% of isolates. The risk for Salmonella detection was higher in the fall season, in younger flocks (≤45 weeks), when sampled by the Israeli Veterinary Services, in the Center region (vs. South), and in the Ross breed. Repeated detections (at least two positive detections) were found in ∼70% of the positive flocks, with an average of 13% of subsequent tests positive after initial detection; older age at first detection (>45 weeks), and Ross breed (vs. other) increased odds of repeated detection. Comparison of serotypes between broiler breeder flocks and humans revealed substantial overlap in dominant serotypes, yet not with serotypes Enteritidis and Typhimurium.Overall, our findings highlight the complex epidemiology of salmonellosis in broiler breeder flocks. The high prevalence and persistence of Salmonella in these flocks, despite continuous control efforts, emphasize the remaining gaps in knowledge for mitigating this important pathogen and protecting the food chain. Introduction Non-typhoidal salmonellosis caused by Salmonella enterica subsp. enterica is one of the most important food-borne zoonoses worldwide, responsible for significant morbidity, mortality, and economic losses across human and animal populations (Antunes et al., 2016). More than 2,600 Salmonella serotypes have been identified to date (Cheng, et al., 2019); however, vast majority of clinical infections are caused by a relatively small number of predominant serotypes (Centers for Disease Control and Prevention (CDC), 2018). Among these, S. Enteritidis and S. Typhimurium are major zoonotic serotypes worldwide and are commonly prioritized in surveillance programs due to their substantial contribution to human infections. According to the World Health Organization (WHO) data, from 2001 to 2007, these two serotypes accounted for about 60% of human infections combined (Hendriksen, et al., 2011). In the United States, during 1996–2011, an average annual incidence of 13.1/100,000 population was reported, and the five most common serotypes, including S. Typhimurium and S. Enteritidis, accounted for 62% of infections (Boore, et al., 2015). In the European Union (EU) between 2014 and 2018, the average Salmonella incidence rate was 20.4/100,000 population and S. Enteritidis and S. Typhimurium were also the most common serotypes (European Food Safety Authority and European Centre for Disease Prevention and Control (EFSA and ECDC), 2019). While various hosts were associated with Salmonella, poultry and their products, particularly broiler meat and table eggs, were recognized as major sources of Salmonella infections in humans (European Food Safety Authority and European Centre for Disease Prevention and Control (EFSA and ECDC), 2023). Therefore, given their potential to vertically and horizontally disseminate infection through the production chain, control programs targeting breeder and layer flocks play a central role in national food safety strategies (;EFSA Panel on Biological Hazards (BIOHAZ), et al. 2019). In Israel, salmonellosis is a notifiable disease, and clinical cases in humans are monitored by the Ministry of Health (MOH) and the Israel Sentinel Laboratory-Based Surveillance Network (ISLBSN) that was established in 1997 by the Israel Center for Disease Control (ICDC) (Bassal, et al., 2023, 2012). Between 2008 and 2018, Israel experienced marked fluctuations in Salmonella incidence of human clinical cases, ranging from 16.1 to 39.1/100,000 population (age-adjusted). While S. Enteritidis remained the predominant serotype throughout the study period, it was intermittently surpassed by S. Infantis, which emerged in 2009 and gradually declined following the introduction of mandatory vaccination against S. Infantis in poultry in 2011. It was later replaced by S. Muenchen, which increased markedly from 2018 onward (Bassal, et al., 2023). Approximately 240 million broilers are raised annually in Israel. Heavy grandparent breeders are imported as day-old chicks, while parent stocks (i.e. broiler breeders) and broilers are produced locally (The Israeli Egg and Poultry Board (EPB), 2018). Broiler breeder flocks are managed under intensive production systems that include controlled feeding, with feed allocation restricted and adjusted according to body weight, flock uniformity, egg production, and egg mass. Such feeding practices may influence Salmonella persistence and shedding by affecting gut physiology, the intestinal microbiota, and bird stress (Durant, et al., 1999; Joat, et al., 2021; Wilson, et al., 2018). A small percentage (∼5%) of the broiler breeder flocks may undergo a molting program, during which feed and lighting are managed to temporarily suppress production and subsequently stimulate renewed reproductive activity. When induced by substantial feed restriction or withdrawal, the associated physiological and immunological changes have been shown to increase susceptibility to S. Enteritidis, intestinal colonization, and fecal shedding (Holt, 2003). A national Salmonella surveillance and control program for breeding flocks was initiated in 1994. This program includes routine bacteriological testing, improvements in housing infrastructure and hygiene, application of on-site biosecurity measures, vaccination, and mandatory culling of flocks positive for either S. Enteritidis or S. Typhimurium. While these efforts led to a marked decline in the prevalence of both S. Enteritidis and S. Typhimurium in all poultry breeding flocks, reducing the annual incidence in broiler breeding flocks to less than 0.3% since 2004, the overall Salmonella infection remained high (29.6% annually on average), with some flocks and farms being repeatedly infected (The Israeli Egg and Poultry Board (EPB), 2018). In this study, we aimed to determine the prevalence and risk factors of Salmonella infections in broiler breeder flocks in Israel between 2008 and 2018. Moreover, given the persistence of Salmonella infections, we aimed to quantify the percentage of repetitive Salmonella detections and potential risk factors for such infections in flocks and farms overtime. Materials and methods Study population The study population includes a census of all commercial broiler breeder flocks in Israel between 2008 and 2018. Flock was defined according to the EU Regulation No., 2160/2003, Article 2.3(b) (Regulation (EC) No 2160/2003, 2003), as poultry kept within the same house (airspace). In Israel, each holding (farm) may maintain multiple flocks of a single age, following an “all-in, all-out” management policy. The flock census information, including the details of the farm (farm license number, village name and region) and the flock (flock number, hatching date and breed), was digitally recorded and continuously updated in a database of the Poultry Health Laboratories of the Egg and Poultry Board (EPB). For the purpose of this study, breed information was categorized to “Ross”, “Cobb” and “Others” (i.e. breed types that had less than 1000 samplings or those that were kept with other breeds on the same flock). Salmonella surveillance Israel has adopted the guidelines of the EU surveillance program for poultry breeding flocks, as outlined in Commission Regulation (EU) No. 200/2010 (Commission Regulation (EU) No 200/2010, 2010), with one modification: environmental sampling consisted of five pairs of drag swabs per house (flock), instead of using boot swabs, feces, or dust samples. In addition, between 2008 and April 2017, sampling was conducted both by farmers, who collected samples twice a month starting at 25 ± 2 weeks of age, and inspectors from the Israeli Veterinary Services (IVS), who sampled at 26 and 57 weeks of age (and additionally at 74 and 94 weeks in the case of induced molting). Beginning in April 2017, the sampling policy was revised, and since then, breeding flocks have been sampled solely by farmers, twice a month throughout the laying period (The Israeli Egg and Poultry Board (EPB), 2019). Salmonella detection and classification Regardless of the sampler’s identity (farmer or IVS inspector), all samples were analyzed at the EPB Poultry Health Laboratories according to ISO 6579:2002/Amendment 1:2007, Annex D: Detection of Salmonella spp. in animal feces and in environmental samples from the primary production stage. Five pairs of drag swabs were pooled into two containers, each containing 225 ml of buffered peptone water for pre-enrichment (Hylabs, Rehovot, Israel). After incubation, 1 ml from each container was pooled and mixed and three drops were inoculated onto a Modified semi-solid Rappaport-Vassiliadis (MSRV) plate. Salmonella-suspected growth on the MSRV plate, was streaked for isolation on selective media: Xylose lysine deoxycholate (XLD) and brilliant green (BG) agar (Hylabs, Rehovot, Israel). Preliminary serogrouping (B, C, D, E, G, or polyvalent) was performed on three colonies from each selective medium, following growth on nutrient agar (Hylabs, Rehovot, Israel) and an autoagglutination test (Becton, Dickinson and Company, Franklin Lakes, NJ, USA or Statens Serum Institute, Copenhagen, Denmark). Final serotyping of Salmonella isolates (one isolate for serogroup per test) was conducted at the National Salmonella reference laboratory (the Public Health Laboratories, MOH, Israel), according to the Kauffmann–White scheme (Grimont and Weill, 2007). Data analysis Salmonella diagnostic test recordings, including the date of sampling and age of the flock, sampler identity, test result, and serogroup and serotype identification, were obtained from the EPB database. Data was obtained as a single Excel file, and further analyses were conducted using R (R Core Team, 2022) and Microsoft Excel®. The initial database included 76,902 rows, yet only 64,275 rows were included in the final analysis after data wrangling - mainly the removal of 7,713 rows from pullet flocks and 4,366 duplicated rows (detailed in the supplementary material). These included 64,217 unique test results (in 58 tests, more than one serotype was identified at the same sampling time). Salmonella prevalence Overall cumulative and annual prevalence of Salmonella-positive tests in Israel during the study period was calculated as the proportion of positive tests among all tests performed, cumulatively and by year, respectively. The prevalence at the flock and farm level were calculated similarly; for this purpose, repeated positive results were removed - thus, a positive farm or flock was only accounted for once in each calculation. In addition, the relative percentage of serogroups and serotypes at the test level were summarized by the five EPB geographic regions: South, Lowland and Mountain, Center, Valleys, and North. The location of all commercial broiler breeder poultry farms was obtained from the Ministry of Agriculture GIS database, and mapped using R ggplot2 package v4.0.1 (Wickham, 2016). Risk factor analysis Univariable and multivariable mixed-effect logistic regression analyses, with Salmonella isolation as the outcome variable, were performed at the test level (n = 64,217), using glmmTMB package v1.1.13 (Brooks, et al., 2017; McGillycuddy, et al., 2025) in R. All factors with a p-value <0.2 in the univariable models were included in the multivariable model. The independent variables tested were season, age, breed, sampler and geographic region. The flock, farm, and village identity variables (in a nested formation) were used as random effects to account for repeated sampling and spatial clustering. The absence of co-linearity between variables was assessed using Variance Inflation Factor (VIF) and direct test of association (chi-square and Cramér's V) using R packages DescTools v0.99.60 (Signorell, 2023) and vcd v1.4-13 (Meyer, et al., 2023). Survival analysis A survival analysis was conducted to estimate the time to first detection of Salmonella in the flocks. Each flock served as the unit of analysis with the age in weeks at sampling as the time variable. An event was defined as the earliest flock age in which a positive result was recorded. Otherwise, the censoring time corresponded to the age at the last negative sampling. Kaplan–Meier survival estimates were stratified by geographic area, and a right-censored survival object and survival curves were created using the survival v3.8-3 (Therneau, 2026) and survminer v0.5.2 (Kassambara, et al., 2026) packages in R. Repeated detection analysis at the flock level Data of flocks found positive to Salmonella at least once (n = 1,642) was filtered to only include flocks with at least nine sampling events following the first Salmonella detection (n = 1,270; hereafter: positive flocks). For each positive flock the number of positive detections out of the total number of sampling events following the first positive detection of Salmonella in the flock was calculated. We further calculated the average, median, Q1, Q3 and range of within flock repeated detections of Salmonella and of the ten most common serotypes in our data. Statistical differences between the ten serotypes were analyzed using a one-way variance analysis (ANOVA) test, followed by Tukey’s honestly significant difference (HSD) post hoc test for pairwise comparisons. The positive flock data was then categorized to compare between flocks without repeated detections (i.e. only single detection of Salmonella) and flocks with repeated detections (i.e. at least one detection of Salmonella infection following the first detection). The variables: (a) flock age at first detection (>45 weeks vs. ≤45 weeks), (b) Number of houses on the farm (≤2 vs. >2),(c) the geographic region according to the EPB districts (South, Lowland and Mountain, Valleys or North vs. Center), (d) breed (Ross, Cobb and Other), and (e) season at first detection were estimated as potential risk factors for repeated detections in a flock using univariable logistic regression models, and all factors with a p-value <0.2 in the univariable models were included in the multivariable model. Statistical analysis was performed using stats package in R (R Core Team, 2022) and a significance level of α= 0.05 was applied in all statistical models. Repeated detection analysis at the farm level Farms may include multiple poultry houses, of which only some may be active at each growing cycle. The term “growing cycle” defines the period from the placement of poultry on a farm until their depopulation under the “all-in, all-out” management policy used in Israel. Thus, for calculating the percentage of repeated detections at the farm level, we used the data of farms in which positive flocks were found. The average number of poultry houses on a growing cycle in a farm was calculated by dividing the sum of poultry houses in the farm in the entire study period by the number of the growing cycles in the farm during the entire study period (including both growing cycles that were found negative at all tests and those found positive in at least one test). In addition, the number of growing cycles in which the farm was infected was calculated and dichotomized into two categories: a single detection in the entire period, and more than one detection in the entire period. A univariable logistic regression model was fitted to estimate the association between the average number of poultry houses (>2 vs. ≤2) on a growing cycle in a farm and the detection of Salmonella in more than one growing cycle at the farm. Statistical analysis was performed using stats package in R (R Core Team, 2022) and a significance level of α= 0.05 was applied. Serotype prevalence in broiler breeding flocks and incidence in humans Data on Salmonella isolates from human cases was extracted from the annual reports of the Public Health Laboratories at the MOH, Israel (Israel Ministry of Health – Jerusalam Central Laboratories, 2008-2018). The cumulative incidence (2008-2018) of serotypes from humans was calculated from the yearly total number of isolates and annual incidences of the serotypes. We compared the incidence of top ten serotypes isolated from human clinical cases with the prevalence of those isolated from broiler breeders. Results The prevalence of Salmonella in broiler breeder flocks in Israel between 2008 and 2018 A total of 4,100 flocks from 490 poultry houses across 142 farms were sampled between 2008 and 2018, yielding 64,217 test results. The average yearly rate of positive tests was 6.7%, peaking in 2011 (13.4%) and reaching a low in 2018 (3.1%; Fig. 1 and Fig. S1). The average percentage of Salmonella detection also varied by month, with higher percentage of positive results during late fall and winter (September through January; Fig. S2). The average yearly rate of positive flocks was 29.6% (range: 16.8-50.9%) and the average yearly rate of positive farms (i.e. with at least one positive flock) was 59.6% (range: 41.0-85.1%).Fig. 1Annual number of total and positive Salmonella tests and the proportions of the five most frequently isolated serotypes from broiler breeder flocks in Israel between 2008 and 2018. Annual number of total and positive Salmonella tests and the proportions of the five most frequently isolated serotypes from broiler breeder flocks in Israel between 2008 and 2018. Total of 91 serotypes were identified - between only once (n = 32 serotypes) to 1,280 times (n=one serotype). The ten most common serotypes, accounting for 85% of all isolates, were S. Infantis (29.0%), S. Hadar (12.2%), S. Kentucky (11.9%), S. Muenchen (8.1%), S. Virchow (6.1%), S. Java (4.9%), S. Newport (4.1%), S. Anatum (3.7%), S. Mbandaka (2.6%) and S. Lexington (2.2%). Among positive flocks (n = 1,642), 68% had a single serotype identified (often repeatedly), 24% had two serotypes (usually at different sampling events), and only 8% had three or more serotypes. Serotype Muenchen has increased from 1.9% in 2008 to become the most prevalent serotype in 2018 (31.7%), gradually replacing Infantis, which decreased from a peak of 46.3% in 2011 to 6.7% in 2018 (Fig. 1). The prevalence of Salmonella and serogroups by geographic region The number of farms, flocks and tests varied across the EPB geographic regions: South (22 farms, 769 flocks, and 11,873 tests), Lowland and Mountain (26 farms, 940 flocks, and 15,038 tests), Center (43 farms, 1,158 flocks, and 17,971 tests), Valleys (44 farms, 983 flocks, and 15,542 tests), and North (7 farms, 250 flocks, and 3,793 tests); yet the number of tests per flock remained similar across regions, ranging between 15.17 and 15.99. The proportions of positive tests (5.0–7.5%), flocks (33.2–42.1%), and farms (81.8–100%) in these regions are shown in Fig. S3. The top ten serotypes varied across regions, yet S. Infantis was the most predominant serotype in all regions (Table S1). The south region had the highest proportion of serogroup E isolates (Fig. 2). This pattern was driven mainly by two farms with repeated isolations of serotypes Anatum and Lexington in successive flocks.Fig. 2The relative proportions of Salmonella serogroups across geographic regions in Israel. Total number of isolates = 4,411 (A), and the distribution of broiler breeder farms (n = 142) by region (B). The relative proportions of Salmonella serogroups across geographic regions in Israel. Total number of isolates = 4,411 (A), and the distribution of broiler breeder farms (n = 142) by region (B). The risk for flock infection During the entire study period, 91.5% (130/142) of farms and 40.0% (1,642/4,100) of flocks were positive for Salmonella (isolation of Salmonella in at least one test during 2008-2018). Flock age at final Salmonella testing ranged from 23.6 to 112 weeks (average age= 60.8 weeks, with standard deviation=12.5 weeks). Moreover, up to approximately 70 weeks of age, the survival probability of flocks to remain negative to Salmonella detection tests declines gradually from 100% to about 70% across all geographic regions (Fig. 3). The survival probability decreases more sharply, with variation across regions, beyond the age of 70 weeks.Fig. 3Kaplan–Meier survival curves for the probability of broiler breeder flocks to remain Salmonella-negative over time (in weeks), stratified by geographic region. Kaplan–Meier survival curves for the probability of broiler breeder flocks to remain Salmonella-negative over time (in weeks), stratified by geographic region. In the univariable analysis of risk factors for a positive test result, all independent variables examined (season, flock age, region, sampler, and breed) showed significant associations with Salmonella detection (p < 0.2; Table S2). Consequently, all these variables were included in the multivariable mixed-effects logistic regression model, where they remained significant at the α < 0.05 level. Specifically, Salmonella detection was associated with the fall season (compared with winter, summer, and spring), younger flock age (≤ 45 weeks), sampling conducted by an IVS inspector (vs. farmer), the central region (vs. south), and breed (Ross vs. Cobb/Other) (Table 1).Table 1Risk factors for Salmonella isolation (total test number = 64,217) were estimated using multivariable mixed-effect logistic regression models. The proportion of cases as a function of exposure to each risk factor (numerical and percentage), odds ratio (OR) with 95% confidence intervals (CI95%), and the significance level are presented.VariableCategoryPositive/n (%)OR (CI95%)p valueSeasonFall1,177/15,062 (7.8)Ref.Winter1,184/16,356 (7.2)0.81 (0.74-0.89)<0.001Spring935/15,729 (5.9)0.57 (0.51-0.63)<0.001Summer1,057/17,070 (6.2)0.66 (0.59-0.72)<0.001Age≤45 weeks2,739/34,103 (8.0)Ref.>45 weeks1,614/30,114 (5.4)0.46 (0.42-0.5)<0.001SamplerFarmer3,716/58,195 (6.4)Ref.IVS inspector637/6,022 (10.6)1.74 (1.57-1.92)<0.001EPB regionCenter1,358/17,971 (7.6)Ref.South853/11,873 (7.2)0.56 (0.34-0.92)0.021Valleys1,122/15,542 (7.2)0.98 (0.65-1.47)0.91Lowland and Mt828/15,038 (5.5)0.67 (0.42-1.04)0.077North192/3,793 (5.1)0.54 (0.25-1.15)0.111BreedRoss3,522/45,205 (7.8)Ref.Cobb633/12,697 (5)0.6 (0.48-0.77)<0.001Other198/6,315 (3.1)0.34 (0.26-0.45)<0.001 Table 1Risk factors for Salmonella isolation (total test number = 64,217) were estimated using multivariable mixed-effect logistic regression models. The proportion of cases as a function of exposure to each risk factor (numerical and percentage), odds ratio (OR) with 95% confidence intervals (CI95%), and the significance level are presented.VariableCategoryPositive/n (%)OR (CI95%)p valueSeasonFall1,177/15,062 (7.8)Ref.Winter1,184/16,356 (7.2)0.81 (0.74-0.89)<0.001Spring935/15,729 (5.9)0.57 (0.51-0.63)<0.001Summer1,057/17,070 (6.2)0.66 (0.59-0.72)<0.001Age≤45 weeks2,739/34,103 (8.0)Ref.>45 weeks1,614/30,114 (5.4)0.46 (0.42-0.5)<0.001SamplerFarmer3,716/58,195 (6.4)Ref.IVS inspector637/6,022 (10.6)1.74 (1.57-1.92)<0.001EPB regionCenter1,358/17,971 (7.6)Ref.South853/11,873 (7.2)0.56 (0.34-0.92)0.021Valleys1,122/15,542 (7.2)0.98 (0.65-1.47)0.91Lowland and Mt828/15,038 (5.5)0.67 (0.42-1.04)0.077North192/3,793 (5.1)0.54 (0.25-1.15)0.111BreedRoss3,522/45,205 (7.8)Ref.Cobb633/12,697 (5)0.6 (0.48-0.77)<0.001Other198/6,315 (3.1)0.34 (0.26-0.45)<0.001 Risk factors for Salmonella isolation (total test number = 64,217) were estimated using multivariable mixed-effect logistic regression models. The proportion of cases as a function of exposure to each risk factor (numerical and percentage), odds ratio (OR) with 95% confidence intervals (CI95%), and the significance level are presented. Repeated detections in flocks Salmonella was detected at least once in 1,270 flocks that were tested at least 9 times following the first detection of Salmonella in the flock. Flocks were housed in 122 farms. In 388 flocks (30.6%) Salmonella was detected only once, while in 882 flocks (69.4%) Salmonella was detected at least once following the first detection (Fig. 4). On average repeated detections of Salmonella in positive flock occurred in 13.1% of the tests, ranging between 0 and 100% of the tests (Fig. 5 and Table S3). The average percentage of repeated detections in positive flocks varied across the ten common serotypes (Fig. 5) with significant differences found between the serotype pairs: Kentucky-Anatum, Lexington-Anatum, Newport-Anatum, Kentucky-Hadar, Muenchen-Kentucky, Virchow-Kentucky, Muenchen-Lexington, Virchow-Lexington, Newport-Muenchen, and Virchow-Newport (Tukey’s HSD, Table S4).Fig. 4Distribution of broiler breeder flocks by the percentage of repeated detections of Salmonella within a flock (in flocks that were tested at least nine times following first detection).Fig. 5Distribution of repeated Salmonella detection percentages. Boxplots represent the percentage of repeated detections within flocks for all Salmonella isolates and for the ten most common serotypes (in flocks that were tested at least nine times following first detection). The center line indicates the median, the solid points indicate the mean, boxes represent the interquartile range (IQR), and whiskers extend to 1.5 × IQR. Individual hollow dots denote outliers. Distribution of broiler breeder flocks by the percentage of repeated detections of Salmonella within a flock (in flocks that were tested at least nine times following first detection). Distribution of repeated Salmonella detection percentages. Boxplots represent the percentage of repeated detections within flocks for all Salmonella isolates and for the ten most common serotypes (in flocks that were tested at least nine times following first detection). The center line indicates the median, the solid points indicate the mean, boxes represent the interquartile range (IQR), and whiskers extend to 1.5 × IQR. Individual hollow dots denote outliers. The age at the first Salmonella detection in a positive flock, number of houses in the farm, EPB geographic region, breed, and season at first detection were evaluated as potential risk factors for repeated detections (Table 2). Older age of the flock at the first Salmonella detection (>45 weeks) and the flock breed (Ross vs. other) were found significantly associated with increased risk for repeated detections of Salmonella in the flock in the multivariable model (Table 3).Table 2Risk factors for repeated detections of Salmonella in flocks that were found positive for Salmonella at least once were estimated using univariable logistic models. The proportion of cases as a function of exposure to each risk factor (numerical and percentage), odds ratio (OR) with 95% confidence intervals (CI95%), and the significance level are presented.VariableCategoryPositive/n (%)OR (CI95%)p valueAge at first detection≤45 weeks790/1,157 (68.28)Ref.>45 weeks92/113 (81.42)2.04 (1.27-3.40)<0.01Number of houses on the farm>2747/1,086 (68.78)Ref.≤2135/184 (73.37)0.80 (0.56-1.13)0.21EPB regionCenter262/377 (69.5)Ref.Lowland and Mt172/266 (64.66)0.80 (0.58-1.12)0.2North41/61 (67.21)0.90 (0.51-1.63)0.72South178/250 (71.20)1.09 (0.77-1.54)0.65Valleys229/316 (72.47)1.16 (0.83-1.61)0.39Season at first detectionFall250/365 (68.49)Ref.Winter226/327 (69.11)1.03 (0.75- 1.42)0.86Spring206/295 (69.83)1.06 (0.76- 1.49)0.71Summer200/283 (70.67)1.11 (0.79- 1.56)0.55BreedRoss722/1,018 (70.92)Ref.Cobb119/181 (65.75)0.79 (0.57-1.11)0.16Other41/71 (57.75)0.56 (0.34-0.92)0.02Table 3Risk factors for repeated detections of Salmonella in flocks that were found positive for Salmonella at least once were estimated using a multivariable logistic model. The proportion of cases as a function of exposure to each risk factor (numerical and percentage), odds ratio (OR) with 95% confidence intervals (CI95%), and the significance level are presented.VariableCategoryPositive/n (%)OR (CI95%)p valueAge at first detection≤45 weeks790/1,157 (68.28)Ref.>45 weeks92/113 (81.42)1.99 (1.24-3.33)<0.01BreedRoss722/1,018 (70.92)Ref.Cobb119/181 (65.75)0.80 (0.57-1.11)0.19Other41/71 (57.75)0.58 (0.35-0.95)0.03 Table 2Risk factors for repeated detections of Salmonella in flocks that were found positive for Salmonella at least once were estimated using univariable logistic models. The proportion of cases as a function of exposure to each risk factor (numerical and percentage), odds ratio (OR) with 95% confidence intervals (CI95%), and the significance level are presented.VariableCategoryPositive/n (%)OR (CI95%)p valueAge at first detection≤45 weeks790/1,157 (68.28)Ref.>45 weeks92/113 (81.42)2.04 (1.27-3.40)<0.01Number of houses on the farm>2747/1,086 (68.78)Ref.≤2135/184 (73.37)0.80 (0.56-1.13)0.21EPB regionCenter262/377 (69.5)Ref.Lowland and Mt172/266 (64.66)0.80 (0.58-1.12)0.2North41/61 (67.21)0.90 (0.51-1.63)0.72South178/250 (71.20)1.09 (0.77-1.54)0.65Valleys229/316 (72.47)1.16 (0.83-1.61)0.39Season at first detectionFall250/365 (68.49)Ref.Winter226/327 (69.11)1.03 (0.75- 1.42)0.86Spring206/295 (69.83)1.06 (0.76- 1.49)0.71Summer200/283 (70.67)1.11 (0.79- 1.56)0.55BreedRoss722/1,018 (70.92)Ref.Cobb119/181 (65.75)0.79 (0.57-1.11)0.16Other41/71 (57.75)0.56 (0.34-0.92)0.02 Risk factors for repeated detections of Salmonella in flocks that were found positive for Salmonella at least once were estimated using univariable logistic models. The proportion of cases as a function of exposure to each risk factor (numerical and percentage), odds ratio (OR) with 95% confidence intervals (CI95%), and the significance level are presented. Table 3Risk factors for repeated detections of Salmonella in flocks that were found positive for Salmonella at least once were estimated using a multivariable logistic model. The proportion of cases as a function of exposure to each risk factor (numerical and percentage), odds ratio (OR) with 95% confidence intervals (CI95%), and the significance level are presented.VariableCategoryPositive/n (%)OR (CI95%)p valueAge at first detection≤45 weeks790/1,157 (68.28)Ref.>45 weeks92/113 (81.42)1.99 (1.24-3.33)<0.01BreedRoss722/1,018 (70.92)Ref.Cobb119/181 (65.75)0.80 (0.57-1.11)0.19Other41/71 (57.75)0.58 (0.35-0.95)0.03 Risk factors for repeated detections of Salmonella in flocks that were found positive for Salmonella at least once were estimated using a multivariable logistic model. The proportion of cases as a function of exposure to each risk factor (numerical and percentage), odds ratio (OR) with 95% confidence intervals (CI95%), and the significance level are presented. Repeated detections in farms Farms could include flocks of the same age housed in multiple poultry houses. These were managed by the farmer as a single unit throughout the growing cycle, i.e. from placement to depopulation. One hundred and twenty-two farms had Salmonella-positive flocks that were tested at least nine times following the first detection during a growing cycle. Salmonella was detected in more than one growing cycle on 24/29 (82.76%) farms with an average of up to two houses occupied per cycle compared with 85/93 (91.40%) farms with an average of more than two houses occupied per cycle. No significant difference was found between the two groups. Cumulative serotype incidence in humans and prevalence in broiler breeders A comparison of the ten most common serotypes in clinical isolates from humans with those prevalent in broiler breeders revealed several shared serotypes in both populations. Serotype Infantis predominated both in humans (21.8%) and in broiler breeders (29.0%). Serotypes Hadar, Kentucky, Muenchen, and Virchow were also common to both. However, serotypes Enteritidis and Typhimurium that were predominant in humans (23.4% and 5.1% respectively), were rare in broiler breeders (0.1% each) (Fig. 6). Conversely, serotypes Java, Anatum, Newport, Mbandaka, and Lexington ranked among the top ten in poultry, yet were found in less than 1% of human clinical cases throughout the period.Fig. 6The incidence of top ten serovars of Salmonella in human morbidity and their prevalence in broiler-breeder flocks. The incidence of top ten serovars of Salmonella in human morbidity and their prevalence in broiler-breeder flocks. Discussion During the period between 2008 and 2018, almost 30% of the broiler breeder flocks and 60% of the boiler breeder farms in Israel were found to be infected with Salmonella each year. Infections were caused by 91 different serotypes, of which the ten top leading serotypes contributed almost 85% of the positive test results. The flock-level prevalence found in this study is consistent with findings from broiler breeder populations in North America and Australia (Murray, et al., 2023; Siceloff, et al., 2025; Sivaramalingam, et al., 2013; Willson and Chousalkar, 2023), but much higher than in the EU, which reported a prevalence of 2.04% of Salmonella in chicken breeding flocks in 2018 (European Food Safety Authority and European Centre for Disease Prevention and Control (EFSA and ECDC), 2019). The peak in Salmonella prevalence observed in 2011 coincided with a marked surge of the Infantis serotype in Israel. The emergence of this serotype was suggested to be the result of high adaptation to poultry environments, following the acquisition of a megaplasmid named pESI that harbors, among others, multiple antimicrobial resistance genes (Aviv, et al., 2014). This serotype spread globally (Guzinski, et al., 2023; Mattock, et al., 2024; Piña-Iturbe, et al., 2026) and recent reports indicate that S. Infantis prevalence in the poultry industry has increased in many countries worldwide (Kim and Lee, 2025; Marder, et al., 2018). Across Europe, S. Infantis emerged as the most prevalent serotype in fowl (Gallus gallus) in 2018, representing 36.7% of all serotyped isolates and being widespread across most EU member states, particularly in broilers (36.5% of all serotyped isolates) and broiler meat (56.7%) (European Food Safety Authority and European Centre for Disease Prevention and Control (EFSA and ECDC), 2019). Following the introduction of mandatory vaccination against S. Infantis in Israel to all breeding flocks starting in September 2011 (Israeli Veterinary Services (IVS), 2011), S. Infantis was gradually replaced by the serotype Muenchen. The latter, which was the dominant serotype in broiler breeder flocks by the end of 2018, was shown to acquire the pESI megaplasmid (Cohen, et al., 2022) and has been spreading rapidly across poultry flocks in the meat sector, both vertically and horizontally, and potentially contributing to its dramatic increase in incidence in humans (Perry, et al., 2025). Seasonality, sampler identity, flock age, breed and region were found to be significantly associated with Salmonella infection. The probability of a positive test was higher during the fall, aligning with previous reports of increased environmental Salmonella loads under autumn conditions in various poultry sectors (European Food Safety Authority (EFSA), 2007; Murray, et al., 2023; Shachaf, et al., 2026). Sampler identity was likewise found to be a significant factor in table-egg layer farms in Israel, UK and France (Arnold, et al., 2023; Huneau-Salaün, et al., 2025; Shachaf, et al., 2026). The reasons suggested for the differences between samplers on those studies were risk-based sampling (targeted, rather than random); and higher diagnostic sensitivity of official sampling compared with operator sampling, attributable to differences in sample type, number of samples, timing, and technique. In the Israeli broiler breeder control program, where both farmers and IVS inspectors used the same sample type and number, and most IVS sampling was routine, the observed difference may reflect superior sampling technique by inspectors and/or that sampling was timed to coincide with periods of known stress (transport, onset of lay, post-molt, late lay), during which transient increases in shedding can enhance detection (Andreatti Filho, et al., 2019; Holt and Porter, 1993). Young flock age (≤45 weeks) was also associated with a higher likelihood of a positive result. We selected 45 weeks as the age category threshold because it matches the average testing age in our dataset and aligns with the flocks age at mid-production. The younger age category included the point-of-lay in which the susceptibility for Salmonella infection is increased (Johnston, et al., 2012). Similarly, Siceloff et al. (2025) reported that Salmonella prevalence peaked in flocks aged 28–35 weeks. Two heavy poultry breeds dominate the broiler meat industry in Israel: Ross and Cobb. Ross, the more prevalent breed, was associated with a higher risk of Salmonella detection. Previous studies have shown clear differences in Salmonella susceptibility among poultry breeds, with commercial broiler breeds generally being more susceptible than native or heritage breeds (Mir, et al., 2025; Zhang, et al., 2024). Moreover, within commercial broiler lines, susceptibility also varies, and some lines appear more prone to intestinal colonization and systemic dissemination of S. Enteritidis than others (Bolder, et al., 2002; Schokker, et al., 2012). Salmonella detection rates in the South region of Israel were significantly lower than those observed in the Central region (OR = 0.56, p = 0.02). In addition, the South had higher proportion of serogroup E isolates (specifically serotypes Anatum and Lexington) compared to other regions. Variation in the prevalence of Salmonella and specific serotypes across geographical regions is well recognized and may reflect ecological, climatic, and environmental differences, as well as the integrated structure of the chicken meat industry (Huneau-Salaün, et al., 2025; Siceloff, et al., 2022; Willson and Chousalkar, 2023). Moreover, geographical clustering of specific serotypes may also reflect contamination persistence in housing systems and local management practices (Davies and Wray, 1996). Repeated detection of Salmonella in the same flock was common; however, the average proportion of repeated positive tests within an infected flock was relatively low (13.09%). The sensitivity of environmental sampling depends mainly on the prevalence of infection in the flock and the sampling method (samples type and number) (Arnold, et al., 2014). Therefore, the low proportion of repeated detections in positive farms observed in this study may indicate limited sensitivity of the detection method and, consequently, of current Salmonella control programs. Moreover, our findings underscore the challenge of achieving true eradication once contamination has been established. For most Salmonella serotypes, control measures generally consist of cleaning and disinfection of the poultry house at the end of the growing cycle, yet the difficulty of eradicating Salmonella infection is also evident when drastic interventions such as flock culling are applied, as in the case of serotypes Typhimurium and Enteritidis. A recent study showed that nearly identical S. Typhimurium isolates, differing by only 5 SNPs, were recovered from poultry flocks raised on the same table-eggs layer farm even two years apart (Perry, et al., 2024). Salmonella can survive in the environment of positive farms for extended periods in soil, puddles, poultry litter, and dust (Lublin and Farnoushi, 2023; Newton, et al., 2021; Pal, et al., 2021). In addition, insects, rodents, reptiles, and wild birds found in and around the farm can act as biological or mechanical reservoirs of Salmonella and serve as sources of reinfection for successive flocks (Barua, et al., 2023; Guyomard-Rabenirina, et al., 2019; Skov, et al., 2008). In our dataset, first detection at an older age (>45 weeks) was one of the variables associated with increased odds of repeated detections. This may reflect cumulative environmental contamination and waning vaccine-induced immunity, both of which can contribute to elevated Salmonella shedding in older flocks (Menanteau, et al., 2018; van de Reep, et al., 2018). At the same time, the higher risk of positive test results observed in younger flocks (≤45 weeks) may indicate that the association between age and Salmonella detection is not linear. Rather, it could be consistent with a bimodal pattern, with one peak in early production, potentially linked to point-of-lay immunosuppression and another in late production, possibly reflecting environmental accumulation and declining immunity. From a public health standpoint, the significant overlap in dominant serotypes (Infantis, Hadar, Kentucky, Muenchen and Virchow) between broiler breeders and human cases reinforces the epidemiological link between poultry and human salmonellosis. Although the transmission chain in broilers is less direct than in layers, involving several intermediate production stages, the resemblance in serotype profiles is nonetheless pronounced. In contrast, while serotypes Enteritidis and Typhimurium remain leading causes of human salmonellosis, their minimal presence in broiler breeder populations suggests infection sources other than chicken meat. Both serotypes are regarded as generalists found in a broad range of hosts and sources (Feasey, et al., 2012). Notably, contaminated table eggs are considered as the primary source of human S. Enteritidis infections globally (EFSA Panel on Biological Hazards (BIOHAZ), 2014). The importance of table eggs as a source of infection for both serotypes in Israel was recently demonstrated by the presence of S. Enteritidis (2.98%) and S. Typhimurium (2.09%) in producing commercial layer flocks (Shachaf, et al., 2026). Conclusions The study results provide insight into long-term epidemiological trends, regional and serotype variation, and risk factors for infection within the broiler breeder sector. Repeated detections within farms over time emphasize the difficulty in controlling Salmonella, and the necessity of improving current control measures, or their implementation, to reduce Salmonella prevalence and enhance public health and food safety. Ethical statement Data analyzed in this study are from isolates collected as part of a mandatory routine surveillance program using environmental sampling, conducted according to ethical standards and in compliance with Israeli law. Author contributions Anat Wiseman (AW): contributed substantially to the conceptualization and design, as well as for the analyses and interpretation of the data (Requirement 1). AW was responsible (together with the other authors) for writing the original manuscript draft and critically reviewing and editing it (Requirement 2). Additionally, AW reviewed and gave final approval of the exact version to be published (Requirement 3) and agree to be personally and publicly accountable for the accuracy and integrity of all aspects of the work (Requirement 4).Hadas Poria Daly (HPD): contributed substantially for the analyses and interpretation of the data (Requirement 1). HPD was actively involved (together with the other authors) in writing the original draft of the manuscript (Requirement 2). Additionally, HPD reviewed and gave final approval of the exact version to be published (Requirement 3) and agree to be personally and publicly accountable for the accuracy and integrity of all aspects of the work (Requirement 4).Ehud Elnekave (EE): contributed substantially to the conceptualization and design, as well as for the analyses and interpretation of the data (Requirement 1). EE was responsible (together with the other authors) for writing the original manuscript draft and critically reviewing and editing it (Requirement 2). Additionally, EE reviewed and gave final approval of the exact version to be published (Requirement 3) and agree to be personally and publicly accountable for the accuracy and integrity of all aspects of the work (Requirement 4). Disclosures The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.