Distribution of Salmonella in pistachios and development of effective sampling strategies.

Nuts and other low-moisture foods have generally been considered low-risks for foodborne illness because they are consumed in a dry state where water activity (available moisture) is too low to support microbial growth. However, it is increasingly recognized that many foodborne pathogens can cause illness at very low concentrations, such that microbial growth is not required. In the past decade, outbreaks associated with consumption of raw almonds and in-shell hazelnuts have been documented in the U.S. In 2009 there was a recall of pistachios when Salmonella was isolated from commercial products. With the exception of almonds, very little is known about the prevalence and levels of Salmonella in tree nuts and nothing is known about overall distribution of the organism within contaminated lots. These data are important to develop robust Quantitative microbial risk assessment (QMRA) and for developing scientifically-sound product sampling schemes for verification of food safety plans. This proposal will evaluate the prevalence, levels and distribution of Salmonella in U.S. pistachios. Salmonella isolates will be characterized as a means of providing insight into potential routes of contamination. These data will be used to determine appropriate sampling schemes for the pistachio industry and will be used to update and improve a QMRA that is currently in development.

Nuts and other low-moisture foods have generally been considered low-risks for foodborne illness because they are consumed in a dry state where water activity (available moisture) is too low to support microbial growth. However, it is increasingly recognized that many foodborne pathogens can cause illness at very low concentrations, such that microbial growth is not required. In the past decade, outbreaks associated with consumption of raw almonds and peanut butter have been documented in the U.S. In 2009 there was a large recall of pistachios when Salmonella was isolated from commercial products. With the exception of almonds, very little is known about the prevalence, levels and distribution of Salmonella in nut production and processing environments. These data are important to develop robust Quantitative microbial risk assessment (QMRA) and for developing scientifically-sound product sampling schemes for verification of food safety plans. This proposal will evaluate the prevalence, levels and distribution of Salmonella in U.S. pistachios. Working with industry collaborators we will test 2,000 samples of in-shell pistachios per year for the presence of Salmonella in 100-g samples. Salmonella isolates will be characterized as a means of providing insight into potential routes of contamination. These data will be used to determine appropriate sampling schemes for the pistachio industry and will be used to update and improve a Quantitative Microbial Risk Assessment (QMRA) that is currently under development.

1. To determine the prevalence, concentration and distribution of Salmonella in U.S. pistachios for 2 crop years.

2. To develop cost-effective sampling strategies that could be used by the pistachio industry to evaluate the microbial status of raw pistachios.

The overall weighted prevalence of Salmonella in raw California inshell pistachios in 100-g samples determined for nearly 4,000 samples collected over three harvests (2010, 2011, and 2012) was 0.6%. This finding is similar to the prevalence of Salmonella observed for other tree nuts: e.g. almonds at 1% prevalence (Lambertini et al., 2012), and inshell pecans at 0.8% prevalence (Danyluk personal communication). The overall levels of Salmonella determined for positive samples (average 0.8 MPN/100 g or 0.008 MPN/g) were also similar to those reported in other nuts and other low moisture foods. However, there was a significant difference in the overall prevalence of Salmonella in sinker (0.4%) and floater pistachios (2%). The average level of Salmonella in sinkers (0.2 MPN/100 g) was also significantly lower than that of floaters (1 MPN/100 g). In addition, more than one positive sample was not typically identified from the 12 samples collected from a single sinker silo. The sinker stream makes up roughly 85% of the total volume by weight of the annual crop (Figure 1). The majority of sinkers (90%) are inshell product with small amounts of kernel (3%), shell (5%), and inedible kernel (2%). In contrast, the 15% floater stream is largely inedible kernel (50%) with relatively small amounts of inshell (10%) and kernel (15%). The data suggest that the higher prevalence and levels of Salmonella in floater pistachios is driven by a subset of silos that are much more contaminated than the rest. Data from 2012 suggest that this contamination is strongly associated with pistachios that are “second shake.” For these floater silos the prevalence of Salmonella among the samples analyzed ranged from 14 to 100% (average 48% positive 100-g samples from 11 silos). The level of Salmonella in these silos was the same as the calculated overall average (0.008 MPN/g). These silos each contain 1 to 1.5 million pounds of pistachios and even though only 25% of the weight is edible product, the amounts are significant. It is not known if the prevalence of Salmonella from these silos would be similar after the product was sorted and shelled. While most pistachios are treated by one or more processes that have been validated to reduce Salmonella, large volumes of pistachios, even when contaminated at low levels, pose a risk of contaminating the both the equipment and facility in which they are handled. This increases the risk of a postprocessing recontamination event. To maintain optimum quality, pistachio processors target short times between shaking nuts from the tree to the time the hull is removed. Based on data from our earlier CPS-funded study Salmonella can grow on harvested pistachios under temperatures and humidities that can be achieved in harvest trailers that are held for several hours. Increases in levels of Salmonella are significant after 6 hours. Salmonella can also multiply in hulled pistachios when there are delays between hulling and drying. Because the prevalence of Salmonella was lower in sinker pistachios it is likely that the cause of the higher prevalence in a number of floater silos occurred after the float tank. The association with second shake pistachios may be related to the condition of the hulls that adhere to a greater portion of hulled floaters but also suggests the possibility of delays between hulling and drying. The isolation of a narrow range of Salmonella serovars and PFGE patterns over the 3 years of this study suggests that several strains of Salmonella may have established resident and persistent populations at one or more of the pistachio handlers that participated in this study. It is possible that several silos are contaminated with these Salmonella. Cleaning and sanitizing silos is challenging and it is possible that the reoccurring contamination of floaters is due to an introduction of Salmonella after the pistachios are dried and as they are loaded into these silos. However, this explanation does not explain the strong association of positive samples with second shake pistachios and especially with floaters. Pistachios were sampled through silo sample ports using a sample trier. The sampling was not supervised by UC Davis but was undertaken by a contract laboratory. Although the triers are sprayed with a 70% ethanol solution between each sample, the construction of each triers are complex making sanitation a challenge. It is possible that the sanitation step was inadequate and resulted in some cross-contamination among samples. This might explain why some of the floater silos had a higher prevalence of Salmonella but would not fully explain why the positives were consistently associated with floater samples in each of the 3 years since the same triers were used to collect both floater and sinker samples. Although all of the pistachios from this survey have already been processed and distributed, it is strongly recommended that the participants in this study closely evaluate their floater stream for future crops. Particular attention should be given to those silos that contain second shake or mixed pistachios. Analyzing a single 250 to 375-g sample of pistachios per silo by enrichment (assuming subsamples are taken from throughout the silo and well mixed) should give a reasonable likelihood of finding Salmonella if present at the levels observed in 2012. If positives are found, these facilities should handle pistachios from these silos with caution and they should consider performing a root cause analysis with the goal of developing an action plan to 1) reduce prevalence of Salmonella in floater pistachios and 2) eradicate potentially resident populations of Salmonella. The overall results from sinker pistachios provide substantial data demonstrating that production of pistachios with low prevalence and levels of Salmonella is possible. By focusing on identifying the root cases for contamination of second shake floater pistachios the U.S. pistachio industry should be able to implement targeted mitigation strategies that will further reduce the overall prevalence of Salmonella in this commodity.