ABSTRACT

Salmonella continues to rank as one of the most costly foodborne pathogens, and more illnesses are now associated with the consumption of fresh produce. The U.S. Department of Agriculture Microbiological Data Program (MDP) sampled select commodities of fresh fruit and vegetables and tested them for Salmonella, pathogenic Escherichia coli, and Listeria. The Salmonella strains isolated were further characterized by serotype, antimicrobial resistance, and pulsed-field gel electrophoresis profile. This article summarizes the Salmonella data collected by the MDP between 2002 and 2012. The results show that the rates of Salmonella prevalence ranged from absent to 0.34% in cilantro. A total of 152 isolates consisting of over 50 different serotypes were isolated from the various produce types, and the top five were Salmonella enterica serotype Cubana, S. enterica subspecies arizonae (subsp. IIIa) and diarizonae (subsp. IIIb), and S. enterica serotypes Newport, Javiana, and Infantis. Among these, Salmonella serotypes Newport and Javiana are also listed among the top five Salmonella serotypes that caused most foodborne outbreaks. Other serotypes that are frequent causes of infection, such as S. enterica serotypes Typhimurium and Enteritidis, were also found in fresh produce but were not prevalent. About 25% of the MDP samples were imported produce, including 65% of green onions, 44% of tomatoes, 42% of hot peppers, and 41% of cantaloupes. However, imported produce did not show higher numbers of Salmonella-positive samples, and in some products, like cilantro, all of the Salmonella isolates were from domestic samples. About 6.5% of the Salmonella isolates were resistant to the antimicrobial compounds tested, but no single commodity or serotype was found to be the most common carrier of resistant strains or of resistance. The pulsed-field gel electrophoresis profiles of the produce isolates showed similarities with Salmonella isolates from meat samples and from outbreaks, but there were also profile diversities among the strains within some serotypes, like Salmonella Newport.

According to statistics on foodborne illness, Salmonella almost always ranks at the top in the number of cases, hospital visits, premature death, and loss of productivity (53). Most Salmonella infections are caused by poultry products; however, a Centers for Disease Control and Prevention (CDC) study showed that various types of fresh produce have increasingly been implicated and that 46% of the illnesses can be attributed to them (37). In accordance with this, a recent source attribution study estimated that fruit and vegetables were implicated in about 50% of Salmonella illnesses (15).

The prevalence of Salmonella in meat and poultry products is well established, and these data have been useful for developing hazard analysis and critical control point guidelines for slaughter houses and processing plants (8, 42, 54, 57). In contrast, there are only a few reports on the presence of Salmonella in fresh produce, and most of these were limited studies that focused on pre- and postharvest practices for processing of ready-to-eat products (1, 3, 28, 29, 51). This lack of information has limited our understanding of the prevalence of Salmonella on the various fresh produce types, and little is known about the characteristics of Salmonella strains that are present in produce, such as serotypes, antimicrobial resistance, pulsed-field gel electrophoresis (PFGE) profiles, and whether their presence in produce is epidemiologically linked to outbreaks. The Microbiological Data Program (MDP) was established in 2001 and administered by the U.S. Department of Agriculture (USDA) Agricultural Marketing Service for the purpose of monitoring foodborne pathogens in fresh produce consumed in the United States. Over a period of 11 years, from 2002 to 2012 when the program was defunded, the MDP collected an average of 12,000 fresh produce samples annually from distribution centers across the United States and tested them for the presence of Salmonella, enterotoxigenic Escherichia coli (ETEC), Shiga toxin–producing E. coli (STEC), including O157:H7, and Listeria monocytogenes. The commodities selected for sampling were commonly consumed and frequently implicated in outbreaks as determined by the statistics compiled by the CDC. The statistical framework developed by the MDP for national-level sampling, program operations, and annual reports can still be accessed at www.ams.usda.gov/mdp (52). The MDP provided one of the largest publically available databases on the presence of pathogens in fresh produce, and these data were used to determine the prevalence of STEC (18) and ETEC (19) strains in fresh produce and helped to identify produce types that are more commonly associated with these bacteria. This report summarizes the MDP survey data on Salmonella for the 11-year period, including prevalence rates and characterization data, such as the serotypes, antimicrobial resistance profiles, and PFGE patterns of Salmonella organisms isolated from various types of fresh produce.

MATERIALS AND METHODS

Fresh produce commodities.

The commodities collected and tested by the MDP from 2002 to 2012 included cantaloupes, celery, cilantro, green onions, hot peppers, lettuce, parsley, spinach, alfalfa sprouts, and a variety of tomatoes. In some years, bagged ready-to-eat lettuce and spinach samples were also included. The spinach samples included both baby and mature plants, and lettuce included both romaine and iceberg, but for some of these, the numbers sampled were low and so no distinction was made as to variety. Most of the alfalfa sprouts and cherry tomato or grape tomato samples were in clam-shell packaging. Samples were collected randomly on a year-around basis from over 600 distribution centers and terminal markets located in 11 states within the continental United States. However, the broad distribution and the fluid movement of commodities from the participating distribution centers extended the product coverage to 13 additional states, including Alaska and Hawaii. The statistical framework for sampling was based on the populations of the participating states and the probability proportional to the distribution volume of the randomly selected collection sites, i.e., the amount of produce that moved through the sites. Therefore, the MDP data collected over several years may actually reflect national trends encompassing differences in geography and seasonality, in contrast to surveys that were done with smaller numbers of samples and within limited time frames (1, 28, 29). The MDP samples included both domestic and imported products, and the lists of commodities collected by year, states where samples were collected, and states of origin can be found online at www.ams.usda.gov/mdp (52).

Methods and characterization.

Prior to 2003, the participating MDP laboratories used the VIDAS (bioMérieux, St. Louis, MO) system to serologically screen produce samples enriched in lactose broth for Salmonella. Beginning in 2004, the MDP implemented the BAX PCR (DuPont, Wilmington, DE) system to screen enrichment samples for Salmonella spp. Briefly, fresh fruit and vegetable samples were culture enriched in universal preenrichment broth. Genomic DNA from each sample was extracted and purified using the Promega Maxwell system (Promega Corporation, Madison, WI) and used as the template in a PCR assay. All PCR-positive samples were culture confirmed as follows: universal preenrichment broth cultures from which PCR-positive samples were obtained were subcultured into tetrathionate broth and Rappaport-Vassiliadis broth or SDIX RapidCheck SELECT Salmonella medium (Strategic Diagnostics, Inc., Newark, DE). After overnight incubation at 42°C, aliquots were plated or streaked onto several selective agar media, including brilliant green sulfa, chromogenic, xylose lactose Tergitol, xylose lysine deoxycholate, Hektoen enteric, and bismuth sulfite agars. Colonies were also tested on triple sugar iron agar, lysine iron agar, and urea agar slants. The presumptive isolates were biochemically identified using the VITEK-2 system and serologically confirmed to be Salmonella with the VIDAS syatem (bioMérieux). All Salmonella isolates were further characterized for their specific serotypes using the U.S. Food and Drug Administration Bacteriological Analytical Manual protocol (55) and tested for antimicrobial resistance and PFGE profile. Antimicrobial resistance was tested using the National Antimicrobial Resistance Monitoring System gram-negative panel and the Sensititre system (Thermo Scientific, Oakwood Village, OH). All Salmonella isolates were subjected to PFGE by the participating state agriculture or public health laboratories that were PFGE certified by the CDC, and the data were uploaded to the CDC PulseNet database (40). Details of these procedures and annual MDP progress updates and summaries, including details of analytical testing methods used, results, and data management, can be found in the online MDP resource at www.ams.usda.gov/mdp (52).

RESULTS AND DISCUSSION

Salmonella prevalence in produce types.

The Salmonella prevalence in meat and poultry ranged from 1% to 10% depending on the type of meat or produce (8, 42, 46, 57). Several U.S. studies (28, 29, 51) showed that Salmonella prevalence is usually low in fresh produce (<1%), and the MDP data are consistent with these findings. In total, the microbiology laboratories of the various state agriculture departments and the USDA Agricultural Marketing Service federal facility that participated in the MDP isolated 152 Salmonella strains from fresh produce samples collected from 2002 to 2012. Prior to using PCR in 2004 to screen produce samples, four commodities were screened for Salmonella using the serological VIDAS assay. To eliminate any data variability due to differences in sensitivities and methods, the VIDAS and PCR data are presented and discussed separately. The data in Table 1 show the numbers and percentages of detection by VIDAS, and the prevalence percentage is calculated based on the numbers that were culture confirmed from the enrichment samples. All four products had many presumptive VIDAS-positive samples, but most could not be culture confirmed, so only a few Salmonella isolates were obtained from these, with most being from lettuce and cantaloupe (0.03%) (Table 1). Compared to the PCR data for these same commodities (Table 2), the prevalence percentage obtained with VIDAS were slightly lower. For instance, the prevalences for cantaloupes were 0.03% and 0.09% for VIDAS and PCR, respectively. These variations may be due to differences in the methods or the sensitivities and specificities of the assays (17, 58), but factors such as different samples, the number of samples tested, the source of the samples, and seasonality most likely all affected the results (2).

TABLE 1.

Salmonella prevalence in fresh produce tested between 2001 and 2003 using the serological VIDAS assay

Salmonella prevalence in fresh produce tested between 2001 and 2003 using the serological VIDAS assay
Salmonella prevalence in fresh produce tested between 2001 and 2003 using the serological VIDAS assay
TABLE 2.

Salmonella prevalence in fresh produce tested between 2004 and 2012 using BAX PCR

Salmonella prevalence in fresh produce tested between 2004 and 2012 using BAX PCR
Salmonella prevalence in fresh produce tested between 2004 and 2012 using BAX PCR

Based on PCR data, the highest Salmonella prevalences were observed in cilantro (0.34%), parsley and spinach (0.29%), hot peppers (0.26%), and sprouts (0.25%) (Table 2). These fresh produce types have contributed to many of the Salmonella infections for which produce has been implicated (15, 45). Parsley and cilantro have also been found to be frequent carriers of ETEC (19), and STEC were also commonly found in cilantro and spinach (18). With STEC, however, there appeared to be a close association with spinach, as over half (70 of 132) of the STEC isolated by the MDP were from spinach (18). No such link was observed for Salmonella, as large numbers of Salmonella were isolated from a variety of produce types (Table 2). Interestingly, tomatoes, which have caused a number of Salmonella outbreaks, were found to have low Salmonella prevalence (0.02%). The tomato samples tested by the MDP were retail quality and collected from distribution centers, so they had been washed and coated with mineral oil. It is possible that these postharvest handling practices may have reduced contamination levels (52). Microbial surveys of fresh produce in Canada and Mexico also showed low levels of Salmonella contamination in tomatoes (4, 7, 12). Many studies have reported that Salmonella can internalize in tomatoes, but the process is influenced by several factors, including the variety of tomato, Salmonella serotype, colonization sites on the fruit, and storage temperature (5, 25, 60, 61). Internalized pathogens would be difficult to detect and may be a contributing factor to the low prevalence observed.

The prevalences of Salmonella calculated for the various produce types are also dependent on the effectiveness of culture confirmation. Many VIDAS-positive samples could not be confirmed, and likewise, the numbers of PCR positives ranged from 0 in celery and organic lettuce to 0.61% for sprouts (Table 2), but the actual numbers of culture-confirmed samples were much lower. The inability to culture confirm presumptive positives is not unusual, as the Salmonella strain may have died off by the time of confirmation, or perhaps the high levels or types of normal flora in the samples may have interfered with confirmation. With some products, however, there may be intrinsic factors that could be affecting culture confirmation. For example, tomatoes had high rates of PCR-positive results (0.33%), but culture confirmation proved to be difficult, resulting in low prevalence (0.02%). Among the tomato varieties tested, confirmation was most successful with cherry tomatoes (3 of 7, 43%) and least successful with round tomatoes (1 of 59, 2%). It is uncertain whether there may have been differences in microflora or other factors associated with each variety that interfered with the confirmation procedures.

Potential sources for pathogen contamination of produce have been identified in the field (49) and at preharvest (38) and processing (26, 34) stages. These included soil, manure or fertilizer application, water from irrigation or nonirrigation sources, wildlife, and processing and handling practices. Contact of the under surfaces of lower leaves with soil or contaminated irrigation water is a possible risk factor (38). If it is a real one, it is likely that produce grown closer to the ground should be more susceptible to contamination. The MDP data are somewhat in agreement with this hypothesis, as products like spinach, cilantro, and parsley had the highest prevalence rates (0.29 to 0.34%). However, hot peppers (0.26%), which are grown on vines and off the ground, had Salmonella prevalence similar to that of spinach (0.29%), while lettuce, which is also grown in the soil, did not have high prevalence (0.05%) (Table 2). The fact that the outer leaves of lettuce are often removed and discarded during processing may have contributed to lower prevalence in this product.

Analysis of celery using the VIDAS system found 1 of ~4,900 samples tested to contain Salmonella, while analysis of celery by PCR showed the absence of Salmonella in 1,100 samples. Perhaps PCR testing of larger numbers of celery samples might have found some Salmonella positives, but celery testing was discontinued after 2004 to accommodate the testing of other, high-risk commodities, such as sprouts, cilantro, and green onions (51, 52). These limited sampling data, however, show that Salmonella can be present in celery, albeit at very low prevalence.

Salmonella was not detected by PCR in organically grown lettuce, but only 1,159 samples were tested, so the sample size was limited. It should also be cautioned that these data do not imply that the absence of Salmonella is in any way related to organic agricultural practices. Many studies in the United States have compared the microbiological quality of organically and conventionally grown preharvest fresh produce samples or spring salad mixes, and they did not find Salmonella in either type of samples (32, 39). Similar studies from other countries also showed no Salmonella contamination in lettuce samples, either grown conventionally or organically (31, 35).

The numbers of Salmonella PCR positives were slightly higher in bagged lettuce samples than in whole heads of lettuce (0.55% versus 0.36%). The same was noted between bagged and bunched spinach samples (0.58% versus 0.4%). It has been shown that postharvest handling of leafy produce, such as fresh-cut spinach, lettuce, and cilantro, exposes cut wound areas to which the pathogen can bind if present in contaminated wash water (10, 21, 22). Therefore, it is possible that the slightly higher prevalence of Salmonella in bagged products may be due to adherence or other factors during the processing (10, 22, 56).

Salmonella prevalence in imported produce.

In a 2008 outbreak of Salmonella Saintpaul in the United States, jalapeño peppers imported from Mexico were implicated (6). A follow-up study on the microbial quality of serrano and jalapeño peppers in Mexico found that 1 of 40 jalapeño pepper samples tested was contaminated with Salmonella (12). The MDP did not specifically target imported produce, nor did it use the country of origin as a sampling parameter, but the randomly collected samples included many imported commodities. By country of origin, 74% of the MDP samples were domestic products, including about 1% of the bagged lettuce samples that had mixed origins and were labelled as “United States and Canada” or “United States and Mexico” (Table 3). All alfalfa sprout samples were domestic, and so were most of the leafy greens, as imported lettuce and spinach accounted for only 3 to 4% of the samples of these produce types tested (Table 4). About 3% of the samples were of unknown origin, but 23% of the MDP samples were imported and the country of origin was identified (Table 3). Proportionally, 64% of the imported produce came from Mexico, followed by Guatemala, Canada, Honduras, Costa Rica, and others (Table 3). About 10 to 15% of the cilantro and parsley samples tested were imported, but over 40% of the cantaloupe and hot pepper and 65% of the green onion samples tested were imported (Table 4). Domestic produce comprised 74% of the samples tested by the MDP, and more Salmonella isolates were obtained from domestic produce (123 of 152, 81%) than from imported products (22 of 152, 14%), as expected. Similarly, a majority of imported produce samples were from Mexico, and they accounted for 20 of 22 Salmonella isolates obtained from imported products (Table 3), with 11 strains from hot peppers (5 serrano and 6 jalapeño), 1 each from green onions and cherry tomatoes, and the rest from other produce types.

TABLE 3.

Produce samples tested by MDP between 2002 and 2012, broken down by country of origin, percentage, and number of Salmonella isolates obtained

Produce samples tested by MDP between 2002 and 2012, broken down by country of origin, percentage, and number of Salmonella isolates obtained
Produce samples tested by MDP between 2002 and 2012, broken down by country of origin, percentage, and number of Salmonella isolates obtained
TABLE 4.

Percentages of commodity types tested by MDP between 2002 and 2012 that were imported and number of Salmonella isolates obtained from each commodity type

Percentages of commodity types tested by MDP between 2002 and 2012 that were imported and number of Salmonella isolates obtained from each commodity type
Percentages of commodity types tested by MDP between 2002 and 2012 that were imported and number of Salmonella isolates obtained from each commodity type

The MDP data did not reveal any significant differences in the presence of Salmonella among imported samples, as low Salmonella prevalence was noted in both imported and domestic products. A Canadian study on produce imported from various sources, including from the United States, found no Salmonella in the imported samples (3). Similarly, another study evaluated the microbial quality of domestic and Mexican produce, and no Salmonella was detected in any of the samples (29). The presence of pathogens in fresh produce is highly unpredictable and varies depending on many factors aside from the place of origin. For example, 20% (31 of 152) of the MDP Salmonella isolates were from cilantro samples (Table 2), but while 15% of the samples were imported, all of the Salmonella isolates were obtained from domestic cilantro samples (Table 4).

Salmonella serotypes found in fresh produce.

The MDP survey found at least 51 different Salmonella serotypes associated with various produce commodities, but 6 of 152 strains (4%) were untypeable (Table 5). Characterization of ETEC and STEC isolates from fresh produce showed that 50 to 60% of the E. coli isolates were untypeable or had partial serotypes (18, 19), so comparatively, the level of untypeable Salmonella strains in produce is low. Both cilantro and sprouts had over 30 Salmonella isolations, and each accounted for 20% of the total Salmonella isolates (Table 1), but the cilantro strains showed more diversity, with 18 serotypes, compared to 13 in sprouts (Table 5). Perhaps because cilantro is cultivated in soil, it is exposed to the environment and, therefore, can be expected to have a more diversified flora, whereas sprouts are grown hydroponically, so the sources of contamination are most likely from seeds and water. Aside from cultivation practices, variations in serotypes present in different commodities may also be due to differences in plant characteristics which may affect Salmonella adherence to the plant surfaces (9, 24, 30). Also, it is not certain whether the culture medium used in selective enrichment might have created a bias toward enriching certain serotypes (23, 58).

TABLE 5.

Diversity of Salmonella serotypes found in produce tested by the MDP from 2002 to 2012

Diversity of Salmonella serotypes found in produce tested by the MDP from 2002 to 2012
Diversity of Salmonella serotypes found in produce tested by the MDP from 2002 to 2012

The most common serotypes found by MDP in produce were S. enterica serotype Cubana and S. enterica subspecies arizonae (IIIa), neither of which is common in retail meats or from human sources (13, 33, 36, 43, 50), and they are not among the top 5 Salmonella serotypes that cause foodborne illness (Table 6) (27, 41). All of the Salmonella Cubana isolates were from sprouts, and most of the S. enterica subsp. arizonae strains came from spinach. Among the other common produce serotypes found, S. enterica serotypes Newport, Javiana, and Infantis are also common in retail meat and human samples, and they are also implicated in salmonellosis infections that require hospitalization (37). Salmonella Newport was isolated from cilantro, cantaloupe, lettuce, hot peppers, and spinach, while Salmonella Javiana was isolated from cantaloupe, green onions, and hot peppers. Clonal analysis of Salmonella Newport strains isolated from animals showed intraserotype genetic differences, and these strains grouped into distinct clades (11, 27, 36). It would have been interesting to see whether the Salmonella Newport isolates from produce showed similar diversities. In the MDP study, Salmonella Infantis was mostly isolated from cilantro, and it is also common in retail meats and human samples, but it is not among the top serotypes that cause infections (Table 6). Other serotypes identified in produce, such as S. enterica serotypes Oranienburg, Anatum, Poona, Montevideo, and Rubislaw, were found in at least three produce types (Table 5), and some of these have had a history of causing outbreaks associated with produce. For example, S. enterica serotypes Javiana, Poona, Muenchen, Mbandaka, Senftenberg, and Litchfield accounted for more than 50% of the outbreaks associated with food plants (27). Only Salmonella Litchfield was not isolated from produce by MDP. Salmonella Poona and Javiana are often associated with infections via reptile or amphibian contact (27). Their presence in fresh produce suggests that these Salmonella serotypes, which have environmental, amphibian, or reptile reservoirs, may also be transmitted by fresh produce (48). A single S. enterica serotype Paratyphi B was also isolated, from a spinach sample.

TABLE 6.

Salmonella serotypes most commonly found by MDP in produce samples, the top five serotypes that cause foodborne outbreaks, and the common serotypes found in retail meat and in humans

Salmonella serotypes most commonly found by MDP in produce samples, the top five serotypes that cause foodborne outbreaks, and the common serotypes found in retail meat and in humans
Salmonella serotypes most commonly found by MDP in produce samples, the top five serotypes that cause foodborne outbreaks, and the common serotypes found in retail meat and in humans

The most common Salmonella serotypes that cause foodborne illness are S. enterica serotypes Typhimurium and Enteritidis, and both are commonly found in meat and eggs (13, 16). These serotypes were found in a few samples of cantaloupes, cilantro, green onions, celery, and lettuce, but they were not very prevalent (Table 5). Still, all these produce types have been implicated in past outbreaks with both of these serotypes, so the MDP data are consistent with the idea that these serotypes can be found in fresh produce (33, 52). Subtyping or genomic studies of these Salmonella serotypes that are common in produce may reveal their clonal origins and how they may be epidemiologically linked to serotypes from meat products or associated with outbreaks, as well as shed some insight on sources of contamination (36, 50).

Antimicrobial resistance of Salmonella strains from produce.

The CDC reported that 100,000 of the 1.2 million nontyphoidal Salmonella infections per year in the United States were caused by drug-resistant strains and raised the threat level to “serious” for resistance to cephalosporin and fluoroquinolone classes of antibiotics (14). Consistent with those findings, about 30% of the Salmonella isolates obtained from human stool and blood samples between 1996 and 2007 showed resistance to one or more antibiotics, and the most common serotypes were Salmonella Enteritidis, Typhimurium, and Heidelberg (16). Antibiotic resistance is also prevalent among Salmonella isolates from other sources. A study from New York State showed that 35.6% of the Salmonella isolates from humans and cattle exhibited antimicrobial resistance (47). Another study showed that 84% of the Salmonella isolates obtained from retail meat in the Washington, DC, area were resistant to at least one antibiotic, and the most prevalent serotype was S. enterica serotype Agona (59). Animal husbandry and meat production practices were thought to have contributed to the increased resistance in Salmonella strains from meat samples (20, 59).

Antibiotics are seldom used in crop production, but runoff from cattle farms may contaminate irrigation water, disseminating antibiotic-resistant bacteria onto the crops. For example, 16% of the STEC and ETEC strains isolated from produce by MDP between 2004 and 2012 had resistance to one or more antibiotics (52). Little is known about the prevalence of antimicrobial-resistant Salmonella strains in produce. One study examined environmental samples in various California produce-growing regions and found that only 1 of 55 bacterial strains (1.8%) was resistant to antibiotics, and none of the strains was Salmonella (24). The MDP study showed that 6.5% (10 of 152) of Salmonella isolates from produce showed resistance to antimicrobial compounds (Table 7). Except for lettuce samples, which harbored three different Salmonella serotypes resistant to different antibiotics, no particular produce type had a higher number of resistant strains. Among the three commodities from which most Salmonella were isolated, none of the spinach, either bunched or bagged samples, had any antibiotic-resistant Salmonella, and cilantro and sprouts both had one resistant strain each. Also, no Salmonella serotype was most commonly associated with antimicrobial resistance, and only one strain each of S. enterica serotypes Havana, Kentucky, and Tucson carried resistance to more than one antibiotic (Table 7). Hence, compared to Salmonella isolates from humans and other sources, the prevalence of antibiotic-resistant Salmonella in produce is fairly low.

TABLE 7.

Antimicrobial resistance of Salmonella serotypes found in produce

Antimicrobial resistance of Salmonella serotypes found in produce
Antimicrobial resistance of Salmonella serotypes found in produce

PFGE profiles of Salmonella strains isolated from fresh produce.

The PFGE profiles of all Salmonella strains isolated from produce by the MDP were uploaded to the CDC PulseNet database (40) and can be accessed using “MDP” as a prefix. All 14 Salmonella Cubana strains isolated by the MDP from alfalfa sprouts between 2008 and 2011 were compared with an isolate involved in a 2011 California sprout recall and a strain from a 2009 Canadian alert on contaminated sprouts (Fig. 1). The Salmonella Cubana strains shared about 90% similarity in their profiles (Fig. 1), but the profile of the MDP 2008 Wisconsin isolates was indistinguishable from that of the strain isolated from contaminated sprouts that resulted in the 2009 Canadian health hazard alert. Also, the profile of the MDP 2010 Salmonella Cubana strains isolated from New York and California was indistinguishable from that of a 2011 isolate from Texas, suggesting that these had a common seed source. Furthermore, the profile of a 2008 California isolate (CA 36845 MDP) was indistinguishable from that of the sprout isolate from the 2011 California recall, not only suggesting that the seed source was the same but also that the same Salmonella Cubana strain may have persisted in seeds for three years. Salmonella Newport has emerged as a common Salmonella serotype in meats (11, 27, 43), so the PFGE profile of the MDP produce isolates of Salmonella Newport were compared to those of isolates from ground beef or from cases of human illness. The profiles of a cluster of produce strains comprised of the MDP 2010 Texas isolates from cantaloupe and a 2012 Texas isolate from hot peppers were indistinguishable from those of a 2010 isolate from ground beef, a 2008 isolate from a meat slicer, and an isolate from a human stool sample (Fig. 2). Also, a 2010 MDP Salmonella Newport isolate from California lettuce was indistinguishable from a 2011 Ohio isolate from spinach, suggesting a common product source or, perhaps, that Salmonella Newport strains are genetically conserved. However, Salmonella Newport isolates from Texas hot peppers in 2010, Florida cilantro in 2010, and Washington hot peppers in 2011 all showed distinct PFGE profiles, indicative of genetic diversity among strains of this serotype. These results are consistent with the clade diversity that was reported (11, 27, 36) and show that Salmonella Newport strains are genetically diverse and that they may be cycled through beef and produce, perhaps due to the juxtaposition of the two agricultural operations and crop cultivation and animal husbandry practices (27, 36, 44).

FIGURE 1.

PFGE profiles of MDP Salmonella Cubana isolates from alfalfa sprouts compared with those of human isolates from an outbreak and product recall. * State where sample was collected.

FIGURE 1.

PFGE profiles of MDP Salmonella Cubana isolates from alfalfa sprouts compared with those of human isolates from an outbreak and product recall. * State where sample was collected.

FIGURE 2.

PFGE profiles of MDP Salmonella Newport strains isolated from various types of fresh produce compared with those of Salmonella Newport isolates from human and meat samples.

FIGURE 2.

PFGE profiles of MDP Salmonella Newport strains isolated from various types of fresh produce compared with those of Salmonella Newport isolates from human and meat samples.

The PFGE profiles of some produce isolates could be matched to those of food, environmental, and human isolates, but it is more difficult to match the PFGE profiles of survey isolates to that of an outbreak strain. For example, in response to the 2008 Salmonella Saintpaul outbreak with hot peppers in New Mexico, the MDP screened 225 hot pepper samples, but no Salmonella was found (52). In 2009, Salmonella Saintpaul was isolated from a Florida cilantro and a New York hot pepper sample, but these isolates only showed 70 to 75% similarity to the 2008 outbreak strain from New Mexico (Fig. 3) and so were not part of the outbreak. Marketed fresh produce tends to have broad areas of distribution, so it is not certain that targeting sample collection to the outbreak areas will find the outbreak strain. Even so, PFGE profiles of produce isolates can still provide useful epidemiological data. For example, the profiles of Salmonella Enteritidis isolates from a 2010 New York cilantro sample and a 2012 Colorado cantaloupe sample were indistinguishable from that of a 2010 strain that caused an outbreak with shell eggs (Fig. 4), indicating that Salmonella strains can cycle through human infections and animal and produce sources.

FIGURE 3.

PFGE profiles of MDP Salmonella Saintpaul isolates compared with that of a 2008 outbreak strain.

FIGURE 3.

PFGE profiles of MDP Salmonella Saintpaul isolates compared with that of a 2008 outbreak strain.

FIGURE 4.

PFGE profiles of MDP Salmonella Enteritidis isolates compared with that of a 2010 outbreak strain.

FIGURE 4.

PFGE profiles of MDP Salmonella Enteritidis isolates compared with that of a 2010 outbreak strain.

In conclusion, the MDP data showed that the prevalence of Salmonella in fresh produce marketed in the United States is fairly low, and most isolates did not exhibit antimicrobial resistance. While this may be indicative that fresh produce in the United States is of good microbiological quality, it does not correlate with the increases in Salmonella illnesses and outbreaks associated with produce. However, even at low prevalence, the presence of Salmonella in produce can still have a broad impact on public health, as fresh produce is produced in large quantities, widely distributed, and almost always consumed raw. Some Salmonella serotypes found in produce are also common in other foods, and some are among the top five Salmonella serotypes that cause foodborne illness. In some cases, produce strains also had PFGE profiles that were identical to those of strains that caused outbreaks, suggesting that they can cycle through humans and animal and food sources, including produce. The MDP generated one of the largest publically available databases on the presence of Salmonella and other pathogens in a variety of fresh produce. Although the MDP has since been discontinued, it is critical that such monitoring efforts continue elsewhere to further our knowledge on the microbiological quality of fresh produce, which has had and will continue to have a great impact on consumer health and safety.

ACKNOWLEDGMENTS

The authors thank all the state agriculture laboratories that participated in the MDP. Their efforts and hard work generated this large and useful database, which has contributed greatly to our knowledge on the prevalence of pathogens in fresh produce.

REFERENCES

REFERENCES
1.
Abadias
,
M.
,
J.
Usall
,
M.
Anguera
,
C.
Solsona
, and
I.
Vinas
.
2008
.
Microbiological quality of fresh, minimally-processed fruit and vegetables, and sprouts from retail establishments
.
Int. J. Food Microbiol
.
123
:
121
129
.
2.
Ailes
,
E. C.
,
J. S.
Leon
,
L.-A.
Jaykus
,
L. M.
Johnston
,
H. A.
Clayton
,
S.
Blanding
,
D. G.
Kleinbaum
,
L. C.
Backer
, and
C. L.
Moe
.
2008
.
Microbial concentrations on fresh produce are affected by postharvest processing, importation, and season
.
J. Food Prot
.
71
:
2389
2397
.
3.
Allen
,
K. J.
,
J.
Kovacevic
,
A.
Cancarevic
,
J.
Wood
,
J.
Xu
,
B.
Gill
,
J. K.
Allen
, and
L. R.
Mesak
.
2013
.
Microbiological survey of imported produce available at retail across Canada
.
Int. J. Food Microbiol
.
162
:
135
142
.
4.
Arthur
,
L.
,
S.
Jones
,
M.
Fabriand
, and
J.
Odumeru
.
2007
.
Microbial survey of selected Ontario-grown fresh fruits and vegetables
.
J. Food Prot
.
70
:
2864
2867
.
5.
Barak
,
J. D.
,
L. C.
Kramer
, and
L.
Hao
.
2011
.
Colonization of tomato plants by Salmonella enterica is cultivar dependent, and type 1 trichomes are preferred colonization sites
.
Appl. Environ. Microbiol
.
77
:
498
504
.
6.
Behravesh
,
C. B.
,
R. K.
Mody
,
J.
Jungk
,
L.
Gaul
,
J. T.
Redd
,
S.
Chen
,
S.
Cosgrove
,
E.
Hedican
,
D.
Sweat
,
L.
Chávez-Hauser
,
S. L.
Snow
,
H.
Hanson
,
T. A.
Nguyen
,
S. V.
Sodha
,
A. L.
Boore
,
E.
Russo
,
M.
Mikoleit
,
L.
Theobald
,
P.
Gerner-Smidt
,
R. M.
Hoekstra
,
F. J.
Angulo
,
D. L.
Swerdlow
,
R. V.
Tauxe
,
P. M.
Griffin
,
I. T.
Williams
.
Salmonella Saintpaul Outbreak Investigation Team
.
2011
.
2008 outbreak of Salmonella Saintpaul infections associated with raw produce
.
N. Engl. J. Med
.
364
:
918
927
.
7.
Bohaychuk
,
V. M.
,
R. W.
Bradbury
,
R.
Dimock
,
M.
Fehr
,
G. E.
Gensler
,
R. K.
King
,
R.
Rieve
, and
P. R.
Barrios
.
2009
.
A microbiological survey of selected Alberta-grown fresh produce from farmers' markets in Alberta, Canada
.
J. Food Prot
.
72
:
415
420
.
8.
Bosilevac
,
J. M.
,
M. N.
Guerini
,
N.
Kalchayan
, and
M.
Koohmaraie
.
2009
.
Prevalence and characterization of salmonellae in commercial ground beef in the United States
.
Appl. Environ. Microbiol
.
75
:
1892
1900
.
9.
Brandl
,
M. T.
, and
R. E.
Mandrell
.
2002
.
Fitness of Salmonella enterica serovar Thompson in the cilantro phyllosphere
.
Appl. Environ. Microbiol
.
68
:
3614
3621
.
10.
Buchholz
,
A. L.
,
G. R.
Davidson
,
B. P.
Marks
,
E. C. D.
Todd
, and
E. T.
Ryser
.
2014
.
Tracking an Escherichia coli O157:H7 contaminated batch of leafy greens through a pilot scale fresh cut processing line
.
J. Food Prot
.
77
:
1487
1494
.
11.
Byrne
,
L.
,
I.
Fischer
,
T.
Peters
,
A.
Mather
,
N.
Thomson
,
B.
Rpsner
,
H.
Bernard
,
P.
McKeown
,
M.
Cormican
,
J.
Cowden
,
V.
Aiyedun
,
C.
Lane
.
International Outbreak Control Team
.
2014
.
A multi-country outbreak of Salmonella Newport gastroenteritis in Europe associated with watermelon from Brazil, confirmed by whole genome sequencing: October 2011 to January 2012
.
Euro Surveill
.
19
:
6
13
.
12.
Cardenas
,
C.
,
K.
Molina
,
N.
Heredia
, and
S.
Garcia
.
2013
.
Evaluation of microbial contamination of tomatoes and peppers at retail markets in Monterrey, Mexico
.
J. Food Prot
.
76
:
1475
1479
.
13.
Centers for Disease Control and Prevention
.
2011
.
National enteric disease surveillance: Salmonella annual report, 2011
. .
14.
Centers for Disease Control and Prevention
.
2013
.
Antibiotic resistance threats in the United States, 2013
.
Available at: http://www.cdc.gov/drugresistance/threat-report-2013/. Accessed September 2015
.
15.
Centers for Disease Control and Prevention
.
2015
.
Foodborne illness source attribution estimates for Salmonella, Escherichia coli O157 (E. coli O157), Listeria monocytogenes (Lm) and Campylobacter using outbreak surveillance data. Report from the Interagency Food Safety Analytics Collaboration (IFSAC) project
. .
16.
Crump
,
J. A.
,
F. M.
Medalla
,
K. W.
Joyce
,
A. L.
Krueger
,
M.
Hoekstra
,
J. M.
Whichard
,
E.
Barzilay
.
Emerging Infections Program NARMS Working Group
.
2011
.
Antimicrobial resistance among invasive nontyphoidal Salmonella enterica isolates in the United States: National Antimicrobial Resistance Monitoring System, 1996 to 2007
.
Antimicrob. Agents Chemother
.
55
:
1148
1154
.
17.
Eriksson
,
E.
, and
A.
Aspan
.
2007
.
Comparison of culture, ELISA, and PCR techniques for Salmonella detection in faecal samples for cattle, pig and poultry
.
BMC Vet. Res
.
3
:
1
19
.
18.
Feng
,
P. C. H.
, and
S.
Reddy
.
2013
.
Prevalence of Shiga toxin subtypes and selected other virulence factors among the Shiga-toxigenic Escherichia coli strains isolated from fresh produce
.
Appl. Environ. Microbiol
.
79
:
6917
6923
.
19.
Feng
,
P. C. H.
, and
S.
Reddy
.
2014
.
The prevalence and diversity of enterotoxigenic Escherichia coli strains in fresh produce
.
J. Food Prot
.
77
:
820
823
.
20.
Foley
,
S. L.
, and
A. M.
Lynne
.
2007
.
Food animal-associated Salmonella challenges: pathogenicity and antimicrobial resistance
.
J. Anim. Sci
.
86
:
E173
E187
.
21.
Golberg
,
D.
,
Y.
Kroupitski
,
E.
Belausove
,
R.
Pinto
, and
S.
Sela
.
2011
.
Salmonella typhimurium internalization is variable in leafy vegetables and fresh herbs
.
Int. J. Food Microbiol
.
145
:
250
257
.
22.
Gomez-Lopez
,
V. M.
,
A.
Marin
,
A.
Allende
,
L. R.
Beuchat
, and
M. I.
Gil
.
2013
.
Postharvest handling conditions affect internalization of Salmonella in baby spinach during washing
.
J. Food Prot
.
76
:
1145
1151
.
23.
Gorski
,
L.
2012
.
Selective enrichment media bias the types of Salmonella strains isolated from mixed strain cultures and complex enrichment broths
.
PLoS One
7
:
1
8
.
24.
Gorski
,
L.
,
C.
Parker
,
A.
Liang
,
M.
Cooley
,
M.
Jay-Russell
,
A.
Gordus
,
E. R.
Atwill
, and
R. E.
Mandrell
.
2011
.
Prevalence, distribution, and diversity of Salmonella enterica in a major produce region of California
.
Appl. Environ. Microbiol
.
77
:
2734
2748
.
25.
Gu
,
G.
,
J.
Hu
,
J. M.
Cevallos-Cevallos
,
S. M.
Richardson
,
J. A.
Bartz
, and
A. H. C.
van Bruggen
.
2011
.
Internal colonization of Salmonella enterica serovar Typhimurium in tomato plants
.
PLoS One
6
:
1
10
.
26.
Hanning
,
I. B.
,
J. D.
Nuttand
, and
S. C.
Ricke
.
2009
.
Salmonellosis outbreaks in the United States due to fresh produce: sources and potential intervention measures
.
Foodborne Pathog. Dis
.
6
:
635
644
.
27.
Jackson
,
B. R.
,
P. M.
Griffin
,
D.
Cole
,
K. A.
Walsh
, and
S. J.
Chai
.
2013
.
Outbreak-associated Salmonella enterica serotypes and food commodities, United States, 1998–2008
.
Emer. Infect. Dis
.
19
:
1239
1243
.
28.
Johnston
,
L.
,
L.-A.
Jaykus
,
D.
Moll
,
M.
Martinez
,
J.
Ancisco
,
B.
Mora
, and
C. L.
Moe
.
2005
.
A field study of microbiological quality of fresh produce
.
J. Food Prot
.
68
:
1840
1847
.
29.
Johnston
,
L. M.
,
L.-A.
Jaykus
,
D.
Moll
,
J.
Anciso
,
B.
Mora
, and
C. L.
Moe
.
2006
.
A field study of the microbiological quality of fresh produce of domestic and Mexican origin
.
Int. J. Food Microbiol
.
112
:
83
95
.
30.
Lapidot
,
A.
, and
S.
Yaron
.
2009
.
Transfer of Salmonella enterica serovar Typhimurium from contaminated irrigation water to parsley is dependent on curli and cellulose, the biofilm matrix components
.
J. Food Prot
.
72
:
618
623
.
31.
Loncarevic
,
S.
,
G. S.
Johannessen
, and
L. M.
Rorvik
.
2005
.
Bacteriological quality of organically grown leaf lettuce in Norway
.
Lett. Appl. Microbiol
.
41
:
186
189
.
32.
Mukherjee
,
A.
,
D.
Speh
,
E.
Dyck
, and
F.
Diez-Gonzalez
.
2004
.
Preharvest evaluation of coliform, Escherichia coli, Salmonella, and Escherichia coli O157:H7 in organic and conventional produce grown by Minnesota farmers
.
J. Food Prot
.
67
:
894
900
.
33.
National Antimicrobial Resistance Monitoring System
.
2012
.
National Antimicrobial Resistance Monitoring System: 2010 Executive Report, p. 19
. .
34.
Olaimat
,
A.
, and
R. A.
Holley
.
2012
.
Factors influencing the microbial safety of fresh produce: a review
.
Food Microbiol
.
32
:
1
19
.
35.
Oliveira
,
M.
,
J.
Usall
,
I.
Vinas
,
M.
Anguera
,
F.
Gatius
, and
M.
Abadias
.
2010
.
Microbial quality of fresh lettuce from organic and conventional production
.
Food Microbiol
.
78
:
795
799
.
36.
Olsen
,
S. J.
,
R.
Bishop
,
F.
Brenner
,
T.
Roels
,
N.
Bean
,
R. V.
Tauxe
, and
L.
Slutsker
.
2001
.
The changing epidemiology of Salmonella: trends in serotype isolated from humans in the United States, 1987–1997
.
J. Infect. Dis
.
183
:
753
761
.
37.
Painter
,
J. A.
,
R. M.
Hoekstra
,
T.
Ayers
,
R. V.
Tauxe
,
C. R.
Braden
,
F.
Angulo
, and
P. M.
Griffin
.
2013
.
Attribution of foodborne illnesses, hospitalizations, and deaths to food commodities by using outbreak data, United States, 1998–2008
.
Emerg. Infect. Dis
.
19
:
407
415
.
38.
Park
,
S.
,
B.
Szonyi
,
R.
Gautam
,
K.
Nightingale
,
J.
Anciso
, and
R.
Ivanek
.
2012
.
Risk factors for microbial contamination in fruits and vegetables at the preharvest level: a systematic review
.
J. Food Prot
.
75
:
2055
2081
.
39.
Phillips
,
C. A.
, and
M. A.
Harrison
.
2005
.
Comparison of the microflora on organically and conventionally grown spring mix from a California processor
.
J. Food Prot
.
68
:
1143
1146
.
40.
Ribot
,
E. M.
,
M. A.
Fair
,
R.
Gautom
,
D. N.
Cameron
,
S. B.
Hunter
,
B.
Swaminathan
, and
T. J.
Barrett
.
2006
.
Standardization of pulsed-field gel electrophoresis protocols for the subtyping of Escherichia coli O157:H7, Salmonella, and Shigella for PulseNet
.
Foodborne Pathog. Dis
.
3
:
59
67
.
41.
Robinson
,
S.
19 August 2013
.
The big five: most common Salmonella strains in foodborne illness outbreaks
.
Food Saf. News
. .
42.
Rose
,
B. N.
,
W. E.
Hill
,
R.
Umholtz
,
G. M.
Ransom
, and
W. O.
James
.
2002
.
Testing for Salmonella in raw meat and poultry products collected at federally inspected establishments in the United States, 1998 through 2000
.
J. Food Prot
.
65
:
937
947
.
43.
Sarwari
,
A. R.
,
L.
Magder
,
P.
Levine
,
A.
McNamara
,
S.
Knower
,
G.
Armstrong
,
R.
Etzel
,
J.
Hollingsworth
, and
J. G.
Morris
, Jr
.
2001
.
Serotype distribution of Salmonella isolates from food animals after slaughter differs from that of isolates found in humans
.
J. Infect. Dis
.
183
:
1295
1299
.
44.
Schneider
,
J. L.
,
P. L.
White
,
J.
Weis
,
D.
Norton
,
J.
Lidgard
,
L. H.
Gould
,
B.
Yee
,
D. J.
Vugia
, and
J.
Mohle-Boetani
.
2011
.
Multistate outbreak of multi-drug resistant Salmonella Newport infections associated with ground beef, October to December 2007
.
J. Food Prot
.
74
:
1315
1319
.
45.
Sivapalasingam
,
S.
,
C.
Friedman
,
L.
Cohen
, and
R. V.
Tauxe
.
2004
.
Fresh produce: a growing cause of outbreaks of foodborne illness in the United States, 1973 through 1997
.
J. Food Prot
.
67
:
2342
2352
.
46.
Sofos
,
J. H.
,
S. L.
Kochevar
,
G.
Bellinger
,
D.
Buege
,
D.
Hancock
,
S.
Ingham
,
J. B.
Morgan
,
J. O.
Reagan
, and
G. C.
Smith
.
1999
.
Sources and extent of microbiological contamination of beef carcasses in seven United States slaughtering plants
.
J. Food Prot
.
62
:
140
145
.
47.
Soyer
,
Y.
,
J.
Richards
,
K.
Hoelzer
,
L. D.
Warnick
,
E.
Fortis
,
P.
McDonough
,
N. B.
Dumas
,
Y. T.
Grohn
, and
M.
Weidman
.
2013
.
Antimicrobial drug resistance patterns among cattle-and human-associated Salmonella strains
.
J. Food Prot
.
76
:
1676
1688
.
48.
Srikantiah
,
P.
,
J.
Lay
,
S.
Hand
,
J.
Crump
,
J.
Campbell
, and
M.
Van Duyne
.
2004
.
Salmonella enterica serotype Javiana infections associated with amphibian contact, Mississippi, 2001
.
Epidemiol. Infect
.
132
:
273
281
.
49.
Strawn
,
L. K.
,
Y. T.
Grohn
,
S.
Warchocki
,
R.
Worobo
,
E. A.
Bihn
, and
M.
Weidmann
.
2013
.
Risk factors associated with Salmonella and Listeria monocytogenes contamination of produce fields
.
Appl. Environ. Microbiol
.
79
:
7618
7627
.
50.
Tauxe
,
R. V.
,
M. P.
Doyle
,
T.
Kuchenmuller
,
J.
Schlundt
, and
C. E.
Stein
.
2010
.
Evolving public health approaches to the global challenge of foodborne infections
.
Int. J. Food Microbiol
.
139
:
S16
S28
.
51.
Thunberg
,
R. L.
,
T.
Tran
,
R.
Bennett
,
R.
Matthews
, and
N.
Belay
.
2002
.
Microbial evaluation of selected fresh produce obtained at retail markets
.
J. Food Prot
.
65
:
677
682
.
52.
U.S. Department of Agriculture, Agriculture Marketing Service
.
2015
.
Microbiological data program
.
Available at: http://www.ams.usda.gov/datasets/mdp. Accessed 22 September 2015
.
53.
U.S. Department of Agriculture, Economic Research Service
.
2014
.
Cost estimates of foodborne illnesses
. .
54.
U.S. Department of Agriculture, Food Safety and Inspection Service
.
2015
.
Microbiology/Baseline Data
. .
55.
U.S. Food and Drug Administration
.
2014
.
Salmonella, chap. 5
.
In
Bacteriological analytical manual online
. .
56.
Vandamm
,
J. P.
,
D.
Li
,
L. J.
Harris
,
D. W.
Schaffner
, and
M. D.
Danyluk
.
2013
.
Fate of Escherichia coli O157:H7, Listeria monocytogenes, and Salmonella on fresh-cut celery
.
Food Microbiol
.
34
:
151
157
.
57.
Vipham
,
J. L.
,
M.
Brashears
,
G. H.
Loneragan
,
A.
Echeverry
,
J. C.
Brooks
,
W. E.
Chaney
, and
M. F.
Miller
.
2012
.
Salmonella and Campylobacter baseline in retail ground beef and whole-muscle cuts purchased during 2010 in the United States
.
J. Food Prot
.
75
:
2110
2115
.
58.
Wang
,
H.
,
V. S.
Gill
,
C.-M.
Cheng
,
N.
Gonzalez-Escalona
,
B. M.
Cathryn
,
J.
Zheng
,
R.
Bell
,
A. P.
Jacobson
, and
T. S.
Hammack
.
2015
.
Evaluation of rapid methods for detection and isolation of Salmonella in naturally contaminated pine nuts using different preenrichment media
.
Food Microbiol
.
46
:
58
65
.
59.
White
,
D. G.
,
S.
Zhao
,
R.
Sudler
,
S.
Ayers
,
S.
Freidman
,
S.
Chen
,
P. F.
McDermott
,
S.
McDermott
,
B. S.
David
,
D.
Wagner
, and
J.
Meng
.
2001
.
The isolation of antibiotic resistant Salmonella from retail ground meats
.
New Engl. J. Med
.
345
:
1147
1154
.
60.
Xi
,
X.
,
Y.
Luo
,
Y.
Yang
,
B.
Vinyard
,
K.
Schneider
, and
J.
Meng
.
2012
.
Effects of tomato variety, temperature differential, and post-stem removal time on internalization of Salmonella enterica serovar Thompson in tomatoes
.
J. Food Prot
.
72
:
297
303
.
61.
Zheng
,
J.
,
S.
Allard
,
S.
Reynolds
,
P.
Millner
,
G.
Arce
,
R. J.
Blodgett
, and
E. W.
Brown
.
2013
.
Colonization and internalization of Salmonella enterica in tomato plants
.
Appl. Environ. Microbiol
.
79
:
2494
2502
.