Keywords

Introduction

Nuts are important agricultural commodities and have been a significant part of the human diet for a purported 780,000 years (Goren-Inbar et al. 2002). Nuts are defined botanically as “a hard, indehiscent, one-seeded pericarp generally resulting from a compound ovary, as the chestnut or acorn” or filbert (Rosengarten 1984). However, the word “nut” is commonly used in a much broader sense to include drupes (almonds, pecans, pistachios, and walnuts), legumes (peanuts), and seeds (Brazil nuts, cashews, flax, sunflower seeds, sesame seeds, and pine nuts), which have a similar composition and structure to botanical nuts, but are not nuts in the strictest sense (Rosengarten 1984) (Table 1). Though botanically diverse, nuts generally have shared characteristics, namely, a relatively hard, inedible outer shell with a softer, edible inner kernel (also called the meat) that is referred to as the nut. Drupes also have a fleshy outer coating, often called the hull, which is removed in processing. The hull corresponds to the flesh of other drupes, like nectarines and peaches, and the nut corresponds to the stone of these fruits. The hull may be discarded as waste or composted (e.g., walnuts and pistachios) or may be used as animal feed (e.g., almonds); shells may also be used in a wide range of products or purposes from animal bedding to biofuel. Peanuts are legumes, like peas and other beans, in which the nut consists of a pod made of a single folded carpel surrounding two seeds. The pod of a peanut corresponds to the shell of a true nut. Seeds, in the sense used here, represent a variety of different plant seeds, which are consumed as nuts but do not fall into a single group.

Table 1 Classification of common nuts by category and botanical family
Full size table

Nuts are grown in a wide variety of climates and in virtually every region of the world. The top three producing countries in 2011 for various nuts are summarized in Table 2. The dynamics of production have changed rapidly in the past decade with a trend of increased nut production worldwide. Since nuts are a diverse group of agricultural products encompassing a variety of assorted plants, they have different requirements for growth. Almonds are grown most extensively in moderate climates, predominantly in the Central Valley of California in the USA, with other significant production in Australia, Spain, Iran, and several countries on the Mediterranean coast. Walnuts also are adapted to Mediterranean-like climate zones (California Walnut Board 2012) and are grown widely in China, Iran, and the USA. About 99 % of the US walnut supply is produced in orchards within the Central Valley of California. Other nuts are adapted to grow in more tropical climates: macadamia nuts are grown in Australia and Hawaii, Brazil nuts are grown in Bolivia and Brazil, and cashews are grown in Africa (Nigeria, Ivory Coast), India, Asia (Vietnam, Indonesia), and Brazil. Hazelnuts are widely grown in Turkey, Europe, Caucasus, the USA, and Iran. The top two producing countries for pistachios are Iran and the USA. Differences in climate, local production practices, harvest conditions, and postharvest handling create unique challenges for microbial contamination of nuts grown in each distinct region of the world.

Table 2 Top nut-producing countries in 2011 (FAO, 2012)
Full size table

Many nuts are mechanically harvested; others are harvested by hand or collected from the ground after the ripened product naturally drops. After harvest, all nuts go through various initial postharvest mechanical sorting steps that facilitate removal of debris such as sticks, rocks, leaves, and loose dirt. Most tree nuts have an outer hull that is removed. Hulling may occur before or after drying, depending on the nut. For example, almonds grown in California are shaken to the ground where they dry to a kernel moisture level of <7 %; drying facilitates removal of the hull using a series of sheer rollers, and the equipment can be configured such that the shells are removed at the same time. Other tree nuts (e.g., pistachios and walnuts) are more typically hulled shortly after harvest (before drying), usually via mechanical abrasion combined with water sprays. In-shell nuts are subsequently dried with forced, heated air or, in some countries, by sun exposure and ambient conditions.

Nuts are sold both in-the-shell and shelled (kernel only), with the exception of cashews, which are always shelled because the double shell surrounding the kernel contains toxic compounds (Menninger 1977). Nuts can be consumed out of hand as a snack, but they are also extensively used as food ingredients and in baked goods, confectionary products, and snack foods. Shelling is usually a dry mechanical process, but in some countries, nuts are still shelled by hand. Pecans are dried in-shell or are conditioned by soaking in hot water before shelling to make the kernels pliable and prevent breakage; the nutmeat is then dried (Beuchat and Mann 2011a). Nuts may be stored at ambient or under cool (<15 °C) dry conditions for a year or more before quality begins to degrade. Almonds are often stored after shelling, but most nuts are stored in-the-shell and are then shelled on an as-needed basis.

Shelled nuts may be further processed, including sorting for size and quality. The kernel pellicle (skin) may be removed by a dry process (e.g., dry blanching of peanuts) or by application of hot water or steam (e.g., wet blanching of almonds). Kernels can be used whole or transformed into many different forms: halved, chopped (into various sizes), sliced, slivered, or ground (into nut meals, flours, or pastes). Nuts can be roasted (e.g., by hot air or in oil) and some roasting methods may include introduction of salt or other flavorings. In some cases, treatments may be applied to specifically reduce microbial loads without changing sensory properties.

Nuts or seeds can be ground into fine particles to yield a paste-like consistency. These products are often called “butters.” Nut pastes may be made solely of nuts or may have other ingredients added such as salt, sugar or other seasonings, and/or hydrogenated vegetable oils to prevent separation. Palm oil has also been recently added to some peanut butters, marketed as “no-stir,” based on consumer demand for products free of hydrogenated oils. Peanut butter is one of the most common and easily recognizable nut pastes and accounted for about 60 % (~2.5 billion pounds [1.1 million metric tons]) of peanut use in the USA in 2010 (ERS 2012). Also available are butters made from almonds, cashews, hazelnuts, macadamias, pistachios, sunflower seeds, and sesame seeds (tahini) (Mangels 2001) or flavored spreads that combine nut butters with other ingredients such as chocolate (which are especially popular in Europe). Nuts ground into pastes with sweeteners may also be used in confectionary products such as marzipan (a mixture of ground almonds, sugar or honey, and flavoring) or halva (made by mixing tahini with sugar and other ingredients).

Worldwide nut production has expanded rapidly in recent years. In 2012, global production of peanuts and tree nuts was 36.5 and 3.4 million metric tons (80 and 7.5 billion pounds), respectively, an increase of ca. 12 % and 20 % from 2009, respectively (International Nut and Dried Fruit Council (INC) 2012). Over the past decade, there has also been an increase in reported outbreaks of foodborne illness linked to the consumption of nuts and nut products. It is crucial for the nut industry to implement food safety programs that are adequate to handle current production levels and anticipated increases in production. This chapter will discuss (1) how nuts may become contaminated with foodborne pathogenic bacteria, (2) the survival of these organisms in the product and the nut production and processing environments, and (3) strategies for their control.

Foodborne Pathogens

Nuts tend to be high in fat and low in moisture (Beuchat 1978). With the exception of seasonal specialty products, such as fresh undried almonds or pistachios, nuts are typically dried to a water activity (a w) below 0.70 (target between 0.50 and 0.65) (Beuchat 1978; Kader and Thompson 2002), which is lower than the minimum required for bacteria and most fungi to grow. For this reason, nuts and nut products have long been considered low risks for microbial food safety. However, foodborne pathogens can survive for long periods of time at low population levels in low-a w foods; these products, including nuts and nut-based foods, are increasingly recognized for their contribution to outbreaks of foodborne illness (Beuchat et al. 2013a; Podolak et al. 2010; Scott et al. 2009).

Outbreaks of Foodborne Illness

Salmonella accounted for 18 of the 23 (78 %) reported outbreaks that have been associated with consumption of nuts and nut butters (Table 3). Other outbreaks have been linked to E. coli O157:H7 gastroenteritis (in-shell hazelnuts [Miller et al. 2012] and walnut kernels) and, in very unusual cases, Clostridium botulinum intoxication (peanut butter and canned peanuts). As with other low-a w foods, outbreaks linked to nuts tend to be spread over many months and over wide geographic areas and as such they are challenging to investigate. It is possible that some nut-associated outbreaks, especially those involving strains with common serotypes or fingerprints, have gone unrecognized.

Table 3 Outbreaks of foodborne illness associated with the consumption of nuts and nut pastesa
Full size table

Consumption of raw almonds was associated with North American outbreaks in 2000/2001 and 2003/2004. A total of 168 cases reported in the USA and Canada from October 2000 to July 2001 were epidemiologically linked to the consumption of raw California almonds that were sold in bulk (Isaacs et al. 2005). The outbreak was identified, in part, by association with a very rare strain of Salmonella Enteritidis, phage type (PT) 30. Ultimately, the same microbe was isolated from case patients and almond samples collected from homes, retail outlets, distributors, and warehouses implicated in the outbreak. Traceback investigations led to a processing facility where the nuts were packed and to the hulling and shelling facility where the implicated lots of almonds were handled. The same strain of Salmonella Enteritidis PT 30 was isolated from environmental swabs collected at both the huller and processing facility (Isaacs et al. 2005). This was significant because at the time of the investigation the huller had not been in operation for several months. Likewise, drag swabs of the implicated orchards collected nearly 9 months after the outbreak-associated almonds had been harvested were positive for Salmonella Enteritidis PT 30 (Isaacs et al. 2005). This strain continued to be isolated from the implicated orchards for an additional 5 years (Uesugi et al. 2007). Although many potential sources were investigated, the ultimate origin of the orchard contamination was never determined. However, it was concluded that the almonds most likely acquired Salmonella in the orchard during harvest and that this contamination was spread during postharvest handling.

The 2003/2004 outbreak linked to almonds was associated with equally rare Salmonella Enteritidis PT 9c, with 47 cases reported in the USA and Canada from September 2003 to April 2004. Raw almond kernels recovered from a consumer’s house and samples collected at the almond processor were negative for Salmonella; however, the outbreak strain was isolated from one environmental sample collected at the processor and from three samples obtained at two huller-shellers that supplied almonds to the primary implicated processor (Keady et al. 2004). The source of the Salmonella was not identified.

Peanut butter was first linked to an outbreak of salmonellosis in 1996 in Australia (Scheil et al. 1998). Fifteen cases were identified; the outbreak strain was isolated in peanut butter from consumer households and from unopened jars collected at retail outlets and the processor. The source was ultimately determined to be contaminated roasted peanuts received from a peanut roasting facility.

A decade later, two large outbreaks (715 and 714 confirmed cases) in Canada and the USA (2006/2007 and 2008/2009) were linked to consumption of peanut butter; nine deaths were associated with the 2008/2009 outbreak (Cavallaro et al. 2011; Sheth et al. 2011). Both outbreaks were widespread and prolonged, with cases reported in 48 and 46 states over 12 and 7 months, respectively. The peanut butter associated with the 2008/2009 outbreak was sold in bulk for use in institutions and as an ingredient in multiple foods. The company had routinely tested the final product for Salmonella and had occasionally shipped lots of product that originally tested positive, although were negative in a second test. As a consequence, several thousand different products containing peanut butter were recalled (FDA 2009b). Despite the widespread media attention given to the 2008/2009 outbreak, a third US peanut butter outbreak occurred in 2012 (CDC 2012). This company also shipped product after repeatedly isolating Salmonella from nut butters and the production environment over a 3-year period (FDA 2012), leading to recall of hundreds of nut butters and nut products and temporary (from November 2012 to May 2013) suspension of the facility’s registration (FDA 2013).

Clostridium botulinum is not considered an issue in low-aW foods because the organism cannot grow and produce toxin below an aW of 0.93 (Baird-Parker and Freame 1967). However, three unusual outbreaks of botulism in nut products have been reported (Chou et al. 1988; O’Mahony et al. 1990; Sheppard et al. 2012). Canned peanuts processed in an unlicensed facility were implicated in a botulism outbreak in Taiwan in 1986 among workers who ate in their factory cafeteria (Chou et al. 1988). The dried, shelled peanuts were boiled, placed into glass jars with the cooking liquid, and the jars were steamed for about an hour. This unvalidated process was insufficient to eliminate C. botulinum spores, and the high moisture levels in the product were sufficient to support growth and toxin production. Achieving an equilibrium aW of ≤0.94 but not ≥0.96 in peanut spreads prevents C. botulinum toxin production (Clavero et al. 2000).

A large outbreak (27 cases and one death) of botulism in June 1989 was linked to consumption of hazelnut yogurt in the UK. The toxin was detected in opened and unopened containers of yogurt and in a single sealed but swollen can of the low-acid hazelnut preserve used to flavor the yogurt (O’Mahony et al. 1990). The preserve processor had prepared the product from a mixture of roasted hazelnuts, water, starch, and other ingredients; the bulk mixture was briefly heated before being pumped into metal cans, which were then sealed and processed in boiling water for 20 min. The pH of the majority of preserve available at the processor was between 5.0 and 5.5. While most of the preserve lots were formulated with sugar, for some lots, including those received by the yogurt processor associated with the most cases, the sugar was replaced with aspartame. The processor had noted some blown cans among the stored aspartame-sweetened product. The cause of the outbreak was likely insufficient thermal treatment, possibly coupled with a higher water activity in the aspartame-sweetened product (Brett 1999), thus allowing C. botulinum to survive, grow, and produce toxin during several months of ambient storage before the preserve was added to the yogurt.

Under very different circumstances, immune-compromised adult patients experienced intestinal toxemia after ingestion of peanut butter containing C. botulinum spores (Sheppard et al. 2012). As with infant botulism, spores of C. botulinum can grow in the intestinal tract of persons with Crohn’s disease or other intestinal complications (Sobel 2005).

Recalls of Nuts

Nuts and nut products are often recalled for undeclared allergens, presence of foreign material, and elevated levels of aflatoxin. Nuts (almonds, hazelnuts, macadamias, peanuts, pine nuts, pistachios, and walnuts) and nut pastes (peanut butter, cashew butter, and tahini) have been associated with a number of Class I recalls in the USA and Canada due to isolation of Salmonella. Between 2004 and 2011, nuts, seeds, and their products were the predominant low-aW food category implicated in recalls and market withdrawals in the USA and Canada associated with Salmonella (Beuchat et al. 2013a). To a lesser extent, nuts have been recalled for isolation of or association with E. coli O157:H7 or Listeria monocytogenes (Palumbo et al. 2014b).

Prevalence and Levels of Foodborne Pathogens

A limited number of retail surveys have been done to screen nuts and edible seeds for the presence of Bacillus cereus, E. coli O157:H7, L. monocytogenes, and Staphylococcus aureus (Palumbo et al. 2014a). Most of the surveys have focused exclusively on Salmonella due to the association of this organism with outbreaks in nuts, seeds, and their products. Most of the published surveys have detected Salmonella in a small proportion of samples; however, many of these surveys have collected samples at retail where product age and handling are unknown and have evaluated a small number of samples of individual nut types and analyzed small (10- or 25-g) units (Brockmann et al. 2004; NSW Food Authority 2012; Willis et al. 2009). Some of the collected samples had been roasted, and thus results are not directly comparable in surveys of raw product (Little et al. 2009, 2010).

A survey of nut products (915 samples) collected from retailers, manufacturers, and growers was performed in Australia by the New South Wales (NSW) Food Authority in 2011 (NSW Food Authority 2012). Almonds, Brazil nuts, cashews, hazelnuts, macadamias, mixed nuts, peanuts, pecans, pistachios, and walnuts were examined. A single sample (macadamias; one out of 76, 25-g samples) was positive for Salmonella. Other retail surveys have been performed in the UK and Brazil (Freire and Offord 2002; Kajs et al. 1976; Little et al. 2009, 2010; Willis et al. 2009); Salmonella was isolated in two out of 469, 25-g samples of Brazil nuts (Little et al. 2010) and in an unreported number of subsamples from a 2-kg sample of Brazil nuts (Freire and Offord 2002).

Several large surveys have determined Salmonella prevalence in raw nuts collected from processors shortly after harvest (Table 4). An 8-year survey in California analyzed 13,972, 100-g samples from individual lots of raw almond kernels that revealed a prevalence of 0.98 % ± 0.32 % for Salmonella (Bansal et al. 2010; Danyluk et al. 2007; Lambertini et al. 2012). Salmonella was detected with similar frequency in in-shell almonds sampled over 2 years (1.5 %; seven positive out of 455, 100-g samples) (Bansal et al. 2010). In contrast, the prevalence of Salmonella in four lots of almonds associated with an outbreak was 65 % (Danyluk et al. 2007).

Table 4 Salmonella prevalence in naturally contaminated nuts sampled after harvest (before retail distribution)a
Full size table

The prevalence of Salmonella in California in-shell walnuts was 0.16 % (three of 1,904 375-g samples), whereas E. coli O157:H7 was not detected in any sample (Frelka 2013). Twenty-two of 944, 375-g samples were positive for Salmonella (2.3 % prevalence) in raw peanut kernels sampled over 3 years (2008–2010) (Calhoun et al. 2013). Sesame seeds, collected from imported shipments entering the USA, were contaminated with Salmonella at 11 % of 750-g samples from 177 shipments (Van Doren et al. 2013a) and 9.9 % of 1,500-g samples from 233 shipments (Van Doren et al. 2013b).

Estimating levels of Salmonella in positive lots is challenging, given the generally low prevalence of this organism. The most probable number (MPN) of Salmonella in 99 initially positive raw almond samples was estimated between 0.0044 and 0.15 per gram (Lambertini et al. 2012). Quantifiable levels of Salmonella (0.09 and 0.23 MPN/g) were reported in two positive samples of Brazil nuts (Little et al. 2010). Levels of Salmonella were estimated to be 6 × 10−4 to 0.04 MPN/g in 22 samples of sesame seeds (Van Doren et al. 2013b). When data are available, levels of Salmonella in outbreak-associated product have also been low. Salmonella at levels of <0.03–2 MPN/g, <3 MPN/g, and 1.5 MPN/g were determined in outbreak-associated in-shell peanuts (Kirk et al. 2004), peanut butter (Scheil et al. 1998), and peanut butter (Scott et al. 2009), respectively. Levels of 0.061 and 0.091 MPN/g were determined in recalled outbreak-associated almonds (Danyluk et al. 2007) but were estimated to be 1.2 MPN/g at the peak of the outbreak (Lambertini et al. 2012). The relatively low levels of Salmonella detected in outbreak-associated samples reflect the difficulty in detecting the pathogen in nuts. Routine testing is not likely to detect all or even most contaminated lots.

Potential Routes of Contamination

Surveys conducted over eight separate harvests revealed that a total of 151 Salmonella isolates representing 49 different serovars were recovered from almonds (Bansal et al. 2010; Danyluk et al. 2007; Bansal and Harris, unpublished). Thirteen different serovars were identified from the Salmonella isolates from peanuts over 3 years (Calhoun et al. 2013). A survey of imported spices, including sesame seeds, revealed 94 different serotypes in 187 Salmonella-positive samples (Van Doren et al. 2013a); 18 different serotypes of Salmonella were present in the 20 positive samples of sesame seeds. The diversity of Salmonella serotypes that have been identified in nut and seed surveys is suggestive of a wide range of environmental contamination sources.

Various steps in the harvesting and processing of nuts provide various opportunities for contamination of the nutmeat. Wet or dry contamination may occur in the field before or during harvest, during postharvest processing, or in the post-processing environment. While there are many similarities, each nut type and production region employs different harvest and processing methods. The harvest and processing steps that are typical for almonds, walnuts, pistachios, and peanuts grown in the USA, where tree nuts and peanuts are almost exclusively mechanically harvested, are outlined in Fig. 1.

Fig. 1
figure 1

Process flow diagram for almonds, walnuts, pistachios, and peanuts from harvest through storage

Full size image

Trees are typically shaken to release the nuts either to the ground (e.g., almonds, walnuts, hazelnuts) or onto catch frames (pistachios), or they may fall naturally to the ground (macadamias). Nuts may dry on the ground after being shaken from the tree, or they may be harvested shortly after shaking by mechanically sweeping into windrows and then into harvest bins or trailers. Nuts may be partially processed (e.g., hulled, shelled) shortly after harvesting and then dried at relatively low (40 °C) to moderate (75–85 °C) temperatures, if necessary, prior to storage. Some nuts (e.g., almonds) are dried before collecting from the orchard and thus may be stockpiled for days to months before hulling and shelling. Stockpiled almonds are covered with tarps and fumigated to control insects; most nuts require fumigation or some other treatment to reduce the potential for damage from orchard and storage insect pests.

Peanuts grow in the soil and are mechanically harvested by lifting the plant out of the ground so that the pods are exposed to the air where the pods dry in windrows for 2–8 days depending upon weather conditions. The peanuts are then threshed from the vines and delivered to buying stations for further curing (forced-air drying to about 10 % moisture), cleaning, and grading before storage in warehouses. At peanut-shelling facilities, harvested product is cleaned to remove dirt, rocks, and other foreign material, followed by shelling and gravity separation that removes all but the peanut kernels, which are sorted, sized, packaged, graded, and stored.

Most nuts are harvested from the ground along with soil and other debris. Activities that have the potential to introduce pathogens onto the orchard floor (e.g., raw manure, contaminated irrigation water, wildlife, and grazing animals) may impact contamination of the product. Almonds harvested onto canvas tarps had lower aerobic plate counts than almonds harvested from the ground (King et al. 1970). Grazing domestic animals on grass or other cover crops in pecan orchards increased the risk of nut contamination by fecal microorganisms (Marcus and Amling 1973). In some regions of the world, there may be a greater reliance on manual labor to collect nuts from the trees or the ground. In these instances, worker hygiene training and the availability and maintenance of adequate sanitary facilities are essential in mitigating contamination by humans.

Commodity-specific guidance documents for Good Agricultural Practices (GAPs) have been published for some nuts (Almond Board of California (ABC) 2009; American Peanut Council (APC) 2011; California Pistachio Research Board (CPRB) 2009; Hazelnut Marketing Board (HMB) 2012), but data to quantify the food safety risk of common production practices for most nuts are generally lacking. Uesugi et al. (2007) determined that a single outbreak-associated Salmonella strain persisted in an almond orchard for at least 6 years. Salmonella multiply readily on wet almond hulls and shells (Uesugi and Harris 2006) and in soil in the presence of almond hull nutrients (Danyluk et al. 2008). Survival in the soil is enhanced when temperatures are cool (Beuchat and Mann 2010b; Danyluk et al. 2008). When wetted, Salmonella can move through intact almond hulls and shells to the kernel (Danyluk et al. 2008). Almonds that drop prematurely may come into contact with irrigation water, rainfall occasionally occurs during harvest when nuts are on the ground, and almonds that are missed during harvest may remain on the ground over the winter months. Growers and processors should be aware of the increased contamination potential associated with wet product so that preharvest (e.g., orchard sanitation to remove nuts that remain after the previous harvest) and postharvest procedures (e.g., product segregation and testing) to control risks can be implemented.

Aqueous sources of contamination can also occur during postharvest handling. For many nuts (e.g., pistachios and walnuts), the outer hull is removed using a combination of physical abrasion and water sprays. Nuts may be submerged in water to separate the nuts from rocks and other orchard debris (e.g., walnuts and pecans), to separate well-formed heavy nuts that sink from those that may be damaged and float (e.g., pistachios), or to soften the shell before cracking (e.g., cashews and pecans). Whenever water is used, there is the potential to spread contamination from one lot to another and throughout the facility. Sanitizer-free walnut float tank water had aerobic plate counts of 6.5–7 log CFU/ml (Frelka 2013), and Salmonella survived for more than 2 weeks at ambient temperatures in float tank water in the presence of high microbial loads (Blessington 2011).

The use of antimicrobials in either float tanks of water or as sprays has been explored, but high organic loads in float tanks make this approach challenging because the activity of many antimicrobials is mitigated by organic matter (Beuchat and Mann 2011a; Beuchat et al. 2012, 2013b). In a commercial facility, there was no difference in microbial populations on in-shell walnuts that were sprayed with water and various peroxyacetic acid formulations (100 and 200 ppm, four different formulations) (Frelka 2013). Pecans may be soaked for a few minutes in hot water or for several hours in cold or ambient water to hydrate the kernel and reduce damage during cracking. Under laboratory conditions, significant decreases of 0.41–0.98 log CFU Salmonella/g on the shell surface occurred within 1 h after immersion of in-shell pecans in water containing 100 ppm of chlorine; increasing chlorine concentrations to 400 ppm did not impact the results (Beuchat and Mann 2011a; Beuchat et al. 2012).

Almond hulls and shells are removed by sheer rollers in a manner that allows mixing of the kernels with the hulls and shells, increasing the potential for cross-contamination (Du et al. 2007). During this process, large volumes of dust are generated, further increasing the potential for the spread of contaminants among lots, should they be present. Salmonella is capable of multiplying in almond dusts that are combined with small amounts of water or aqueous sanitizer (Du et al. 2010b); hence, efforts should be made to avoid introduction of water into almond huller-sheller facilities.

The shells of most tree nuts provide an important and effective barrier to microbial contamination, and the presence of a hull can further reduce the risk. The kernel inside an intact shell was once thought to be virtually sterile (Chipley and Heaton 1971; Kajs et al. 1976; Meyer and Vaughn 1969); however, there is substantial evidence that the shell provides variable levels of protection from contamination. For example, walnut kernels have low populations of bacteria when aseptically extracted from in-shell walnuts removed directly from the tree (Frelka 2013). Total aerobic plate and E. coli/coliform counts on kernels increase significantly as the walnuts move through the hulling and drying steps (Blessington 2011; Frelka 2013). Shell thickness can vary significantly among different varieties of the same nut, and shell breakage can expose the kernel within and lead to contamination (Beuchat and Mann 2010a; Frelka 2013; King et al. 1970). The almond kernel can be partially or completely exposed at the time of harvest in soft-shell almond varieties. Wetting the shell suture also promotes microbial infiltration of the nut shell and subsequent contamination of the kernel (Beuchat and Heaton 1975; Beuchat and Mann 2010a; Marcus and Amling 1973). Walnut kernels extracted from nuts with broken shells have significantly higher microbial populations than kernels extracted from walnuts with visibly intact shells (Frelka 2013). Drying may influence shell integrity (Frelka 2013; King et al. 1970; Meyer and Vaughn 1969).

Contamination of nuts also may occur during or after further processing. In more than one peanut butter outbreak (Table 3), investigators concluded that contamination of the product with Salmonella most likely occurred within the processing environment. In some cases, ingress of water into the facility (e.g., water leakage through roof and skylights in poor repair) may have contributed to the contamination by providing an opportunity for the organism to multiply in the production environment. Potential for cross-contamination between treated and untreated product, poor ventilation that increased the possibility for contamination of food and food-contact surfaces, lack of documented cleaning and sanitation of equipment, and lack of a validated process that would ensure adequate reduction of Salmonella were also cited as contributing factors (Cavallaro et al. 2011; FDA 2009c, 2013; Sheth et al. 2011).

Survival of Pathogens in Nuts and Nut-Processing Environments

Once Salmonella is introduced into nuts, nut pastes, or the environment in which these products are processed, the organism may persist for extended periods of time (Beuchat and Heaton 1975; Beuchat and Mann 2010a; Blessington et al. 2012, 2013; Burnett et al. 2000; Kimber et al. 2012; Uesugi et al. 2006). In several cases, outbreak strains have been isolated from production facilities months after the affiliated product was processed (CDC 2012; FDA 2013; Isaacs et al. 2005; Sheth et al. 2011).

Lower storage temperatures reduce the rate of oxidation of the fats within nuts, and nuts are often held commercially or by consumers at cooler temperatures (Lee et al. 2011). Levels of E. coli O157:H7, L. monocytogenes, and Salmonella in nuts are usually stable at freezing or refrigeration temperatures with little to no decline observed in months to over a year of storage (Beuchat and Heaton 1975; Beuchat and Mann 2010a; Blessington et al. 2012, 2013; Burnett et al. 2000; Frelka 2013; Kimber et al. 2012; Uesugi et al. 2006). As temperatures increase, rates of reduction of inoculated organisms increase (Beuchat and Heaton 1975; Beuchat and Mann 2010a; Uesugi et al. 2006), but even at ambient temperatures, the declines are slow (Fig. 2). Linear reduction in Salmonella populations on almonds stored under ambient conditions ranged from 0.16 to 0.32 log CFU/g/month (Lambertini et al. 2012). Inoculated E. coli O157:H7 and L. monocytogenes populations generally declined more rapidly than Salmonella (Blessington et al. 2012, 2013; Kimber et al. 2012).

Fig. 2
figure 2

Survival of Salmonella (triangles), E. coli O157:H7 (squares), and Listeria monocytogenes (circles) on inoculated almonds (a) and pistachios (b) stored at 23 °C; asterisk indicates six of six replicates were positive via enrichment of 10-g samples (modified from Kimber et al. 2012)

Full size image

In some cases, the inoculation level of Salmonella does not influence the rate of decline (Uesugi et al. 2006), whereas in other cases, faster rates of decline are observed at lower inoculation levels (Beuchat and Mann 2010a; Blessington et al. 2012, 2013). A relatively rapid die-off after inoculation is sometimes followed by long-term persistence characterized by little to no measureable decline (Fig. 2). The practical significance of this observed “tailing” is currently unknown.

Control of Foodborne Pathogens

Long-term persistence and enhanced thermal resistance on tree nuts make control of foodborne pathogens in nuts and nut products a challenge. The appropriate level of control in these products has been debated (Schaffner et al. 2013). In the absence of extensive data a 5-log reduction is often proposed as long as appropriate procedures to ensure adequate control of the introduction of pathogens into the raw nuts and additional controls to prevent posttreatment recontamination are in place (FDA 2009a, 2011; Schaffner et al. 2013).

The almond industry in California reacted to outbreaks of salmonellosis associated with their product (Table 3) and implemented a food safety action plan that included the promulgation of regulations through the Almond Board of California (a US Department of Food and Agriculture Agricultural Marketing Service Federal Marketing Order). Since September 2007, these regulations require all almonds grown in California and sold in North America (Canada, Mexico, and the USA) to undergo a treatment capable of reducing Salmonella by at least 4 log (Federal Register 2007). The process criterion was based on an initial risk assessment evaluating the risk of salmonellosis from consumption of almonds (Danyluk et al. 2006). A subsequent risk assessment evaluated the implementation of the 2007 rule and concluded that the current production system was capable of preventing illness even if higher than typical levels of Salmonella were occasionally present (Lambertini et al. 2012). The Almond Board of California established a system to validate and verify that processes used by the almond industry were consistently meeting the mandate (ABC 2014b). These documents were used in the development of validation guidance for evaluating thermal treatments for low-aW foods (Anderson and Lucore 2012).

Because sensory attributes and quality of nuts can be markedly altered with thermal treatment, several nonthermal methods of pathogen reduction also have been explored (Table 5; Pan et al. 2012). The response of Salmonella to desiccation and to inactivation treatments can differ significantly depending on the strain used, growth conditions (including growth on solid or in liquid medium and incubation temperature), method of inoculation, drying protocols and storage conditions between time of inoculation and treatment, product and product composition, treatment conditions, and recovery methods including media selected to recover potentially injured cells (Abd et al. 2012; Beuchat and Mann 2011a; Goodridge et al. 2006; Keller et al. 2012; Ma et al. 2009; Scott et al. 2009; Uesugi et al. 2006). Thus, care should be taken when conducting and interpreting these types of studies and when applying them to commercial settings.

Table 5 Published treatments for potential control of microorganisms in nuts and nut pastes
Full size table

Thermal Treatments

Nuts often are heat treated (e.g., oil or dry roasting or blanching) to change the flavor, texture, or characteristic of the product. These treatments have only recently been examined as a means to control foodborne pathogens. Thermal treatments are not universally applicable, as they may impact the quality of some nuts, particularly those with a high level of unsaturated fat, like walnuts (Rosengarten 1984), by increasing the rate of oxidative rancidity. Thus, thermal treatments capable of reducing foodborne pathogens while preserving the sensory characteristics of the nut have also been investigated, and several technologies are commercially available.

Similar to other low-a w foods, drying foodborne pathogens on nuts confers significant additional resistance of the organisms to thermal destruction (Burnett et al. 2000; He et al. 2011; Horner and Anagnostopoulos 1972; Ma et al. 2009; Podolak et al. 2010). Nuts may be stored for long periods before processing. Storage of inoculated peanut butter for 2 weeks or almonds for 12 weeks did not significantly alter the thermal sensitivity of Salmonella (Abd et al. 2012; Keller et al. 2012). Increases in moisture or water activity of the nut or relative humidity of the heating environment increased the sensitivity of pathogens to heat (Brandl et al. 2008; Beuchat and Mann 2011b; He et al. 2011, 2013; Jeong et al. 2011; Mattick et al. 2000; Villa-Rojas et al. 2013) and have been used as strategies to improve process efficacy. The size of pecan pieces (Beuchat and Mann 2011b) and method of inoculation (Beuchat and Mann 2011a) significantly impacted the efficacy of thermal treatments.

Reported inactivation curves for Salmonella are sometimes linear (Harris et al. 2012; Keller et al. 2012) but are often nonlinear (Abd et al. 2012; Beuchat and Mann 2011b; Du et al. 2010a; He et al. 2011; Keller et al. 2012; Ma et al. 2009; Shachar and Yaron 2006); the Weibull model has been used to fit nonlinear curves. Inactivation curves for two strains of Salmonella in peanut butter were linear when cells were cultured on solid media and upwardly concave when cells were grown in broth (Keller et al. 2012).

Dry heat roasting methods, such as oil immersion and forced air, are common processes used for nuts. Under laboratory conditions, a 5-log reduction of Salmonella on almond kernels was achieved in hot oil in ca. 1.5 min at 121 °C (Abd et al. 2012; Du et al. 2010a); a 4-log reduction of Salmonella on pecan pieces and halves was achieved in ca. 1.5 and 2 min, respectively (Beuchat and Mann 2011b). Heating at 127 °C for 2 min is a recognized process for achieving a 5-log reduction of Salmonella on almonds (ABC 2007c). Peanuts often are dry roasted to facilitate removal of the skin prior to grinding and milling into peanut butter. The grinding process of about 20 min can generate temperatures of 71–77 °C. Under laboratory conditions, these temperatures and times result in reductions of Salmonella in peanut butter that are <1–1.5 log CFU/g (He et al. 2011; Shachar and Yaron 2006) or less than 3 log CFU/g (Ma et al. 2009). Heating for more than 20 min at 90 °C was necessary to achieve a 5-log or greater reduction in Salmonella in peanut butter (Ma et al. 2009).

In dry roasting, heat transfer is much less efficient (Beuchat and Mann 2011b). The efficacy of dry roasting can be improved by adding moisture to either the product or the process (Jeong et al. 2012), and processing equipment that introduce steam prior to or during heating of nuts are commercially available.

Because the dry roasting process is dynamic and there is a wide range of available equipment, validation of dry roasters is usually accomplished onsite with a surrogate bacterium inoculated onto product and introduced into the product stream (ABC 2007b, 2014a; Anderson and Lucore 2012; Jeong et al. 2011). For these and other thermal processes, Enterococcus faecium NRRL B-2354 has been a widely accepted surrogate for Salmonella in almonds (ABC 2014a; Jeong et al. 2011). Use of this surrogate to validate processes in other nuts or low-aW foods should be supported with additional data, using the food and process in question, and laboratory evaluation against the appropriate target pathogen.

Application of hot water or steam to remove almond skins (blanching) is used in almond processing (Harris et al. 2012). Almond kernels are submerged in hot water or steam to break down the pectin connecting the skin to the kernel; a set of rollers is used to facilitate removal of the loosened skins, and the resulting skin-free almonds are then quickly dried using forced hot air (Harris et al. 2012; ABC 2007a). Significant initial reductions of Salmonella were observed upon exposure of inoculated almonds to hot water (Harris et al. 2012). D-values of 1.2, 0.75, and 0.39 min were determined from the linear portion of the curve for temperatures of 70, 80, and 88 °C, respectively, and a z-value of 35 °C. Blanching almonds at 88 °C for 2 min is a recognized process for achieving a 5-log reduction of Salmonella (ABC 2007a). Hot water treatment of in-shell pecans at 75–95 °C during the conditioning step preceding shelling operations reduced Salmonella to various degrees depending on the methodology used for inoculation (Beuchat and Mann 2011a).

Radio-frequency (Gao et al. 2011) and infrared heating have been evaluated as alternatives to roasting to reduce Salmonella on almonds (Brandl et al. 2008; Bingol et al. 2011; Yang et al. 2010). Infrared radiation causes rapid increases in heat on the surface of the almond and, when combined with a short warm temperature hold time, significantly reduced populations of inoculated Salmonella. The process is typically much shorter than traditional thermal methods and can result in a product quality that is similar to untreated product (Brandl et al. 2008).

Nonthermal Treatments

Nonthermal treatments include use of gas fumigation, application of antimicrobial chemicals, and physical treatments. Gaseous ozone reduced by 2–3 logs E. coli and Bacillus cereus on pistachio kernels (Akbas and Ozdemir 2006). Propylene oxide (PPO) gas treatments are permitted in the USA and Canada for tree nuts but not peanuts (Environmental Protection Agency (EPA) 2006). Commercial PPO processes achieve a 5-log or greater reduction of Salmonella on almonds (ABC 2008; Danyluk et al. 2005) and have been evaluated for pecans under laboratory conditions (Beuchat 1973). Although effective, PPO is a batch process limited by the size of the PPO chamber. Bulk nuts (usually in bins, boxes, or totes) must be warmed to at least 30 °C, which can take one or more days; operational parameters include chamber temperature, PPO concentration and time, vacuum after gas injection, and the number of aeration cycles at the end of the PPO exposure time. To achieve the full 5-log reduction and to reduce PPO concentrations to below 300 ppm (a requirement of the US Environmental Protection Agency), nuts must be held for a further 2–5 days before shipping.

High hydrostatic pressure (HHP) has been evaluated for reduction of Salmonella on almonds (Goodridge et al. 2006; Willford et al. 2008), in peanut butter (D’Souza et al. 2012), and on sesame seeds (Wuytack et al. 2003). A 3- to 4-log reduction of Salmonella was achieved when inoculated almonds were suspended in water during HPP at 414 MPa for 6 min and dried at ambient temperature for 25 min after treatment (Willford et al. 2008). Greater reductions were achieved when higher temperatures were used to dry the almonds after HPP. In contrast, less than 2-log reductions of Salmonella were observed in peanut butter treated with high pressure at 600 MPa for 5 min (Grasso et al. 2010) or 400 or 600 MPa for 4–18 min (D’Souza et al. 2012). This is likely due to the low moisture and high oil content of the peanut butter matrix, compared with whole nuts that are suspended in water for HPP processing.

Before shelling, in-shell pecans are commercially cleaned and conditioned in water that is often chlorinated (Beuchat and Mann 2011a). Chorine is effective in maintaining water quality; reductions of inoculated Salmonella are minimal. Free chlorine levels can decline rapidly if nuts are not cleaned prior to conditioning (Beuchat et al. 2012). A 20-min exposure to lactic acid (2 %) or levulinic acid (2 %), in combination with sodium dodecyl sulfate (0.05 %), reduced Salmonella inoculated onto in-shell pecans by 3.4–3.7 log CFU/g. Although this type of treatment would be possible for in-shell nuts with smooth hard shells that are commonly soaked for long periods (e.g., pecans or cashews prior to shelling), it would not be widely applicable to many nuts especially those with porous shells (e.g., almonds or peanuts). Application of 10 % citric or lactic acids to in-shell almonds or almond kernels and holding for 1–5 min reduced Salmonella populations by 0.2–1.4 log CFU/g. The impact of these levels of acids on sensory characteristics of the nuts was not evaluated. Prior to roasting, almonds and pistachios are sometimes sprayed or dipped into saturated salt solutions or flavoring mixtures that may contain citric acid (Kim and Harris 2006). Significantly greater reductions in Salmonella were observed when almond kernels were exposed to water or citric or lactic acids prior to heating.

In-shell walnuts are sometimes exposed to very high concentrations of sodium hypochlorite to lighten the shell (common for North American markets where light shells are traditional). Walnuts with cracked or damaged shells are removed prior to treating, as the treatment would adversely affect the exposed kernels. Treating in-shell walnuts for 2 min in 3 % (30,000 ppm) sodium hypochlorite reduced populations of inoculated Salmonella by 2.5 log; further reductions were observed when the treated nuts were subsequently dried and stored (Blessington et al. 2013). This treatment is probably limited to in-shell walnuts and is becoming less common as the North American in-shell walnut market continues to decrease in favor of shelled product.

X-ray irradiation (Jeong et al. 2012), ionizing radiation (Prakash et al. 2010; Wilson-Kakashita et al. 1995), electron beam radiation (Duong and Foley 2006; Hvizdzak et al. 2010), and nonthermal plasma (Deng et al. 2007; Niemira 2012) have also been evaluated. A 4-log reduction in Salmonella on whole almonds using ionizing radiation required a dose of 5 kGy (Prakash et al. 2010); however, at this level of treatment, significant negative sensory changes were detected by a trained panel, and the nut samples were rejected by a consumer panel. In peanut butter, a 4-log reduction in Salmonella was achieved with 3 kGy (Hvizdzak et al. 2010), but sensory effects were not evaluated.

Sanitation

Regardless of the method used to control pathogens in nuts and nut pastes, it is important to prevent the recontamination of treated product via robust sanitation and environmental monitoring programs (GMA 2010). These include taking precautions when designing equipment that can be easily cleaned and keeping the infrastructure in good repair to avoid ingress of water and development of niches that collect debris. Establishing a Primary Salmonella Control Area (generally the processing area after the inactivation step) is usually recommended (GMA 2010).

Cleaning and sanitizing within nut-processing environments are complicated by the significant amounts of dust generated during the postharvest handling of nuts, which consists of residual orchard soil as well as nut hull and shell particulates. Regular dry cleaning to reduce dust buildup and using appropriate air filters to prevent movement of microbial contaminants into the Primary Salmonella Control Area are important. Oil residues from some nuts (e.g., walnuts) also can coat equipment surfaces during operation.

Introduction of moisture can promote growth and survival of microorganisms in nut particulates (Beuchat and Mann 2010b; Du et al. 2010 b; Uesugi and Harris 2006). Almond dust inactivated water-based quaternary ammonium sanitizers (Du et al. 2010 b); therefore, using water or water-based sanitizers is not typically recommended unless equipment can be disassembled, dust is completely removed, and reassembly takes place after thorough drying. For wet-cleaning activities, separate cleaning rooms are sometimes used to clean and sanitize equipment, thereby reducing the opportunity for introduction of water into the facility. When disassembly or removal of equipment is not possible, a common practice is to control dusts by physical removal along with application of an isopropyl alcohol-based sanitizer on product-contact surfaces (Du et al. 2007, 2010 b). Isopropyl alcohol-based sanitizers also reduce microbial loads on shoes when used in footbaths (Burnett et al. 2013). No significant reductions of microbial levels on the shoes over untreated controls were observed when water-based or dry quaternary ammonium sanitizers were used.

Conclusions

Nuts and nut pastes have long been a significant part of the human diet. Recent increases in production and consumption have been coupled with an increased awareness of microbial food safety issues associated with these products. Substantial research has been devoted to understanding pathogen behavior and control on nuts since 2001; data from these studies have directly enhanced food safety guidance and have led to improvements to sanitation programs for these products (e.g., GMA 2010; Scott et al. 2009). Various thermal and nonthermal control methods and the means to appropriately validate these systems have been and continue to be developed and implemented (ABC 2014b; Anderson and Lucore 2012).