Salmonella has been isolated from dried pistachios in both postharvest and retail surveys. The source of Salmonella in pistachios is unknown, but introduction is possible at points during production, harvest, and postharvest activities. To examine the behavior of Salmonella on pistachios during simulated postharvest conditions, early-, mid-, and late-season inhull pistachios were collected from two commercial processors over five different harvests. Pistachios were inoculated with cocktails of nalidixic acid– or rifampin-resistant Salmonella at 0.64 to 1.59 log CFU/g (low) or 2.73 to 3.27 or 4.29 to 4.31 log CFU/g (high) and were incubated for up to 30 h under commercially relevant conditions (23, 35, or 37°C and 50 or 90% relative humidity [RH]). Populations of Salmonella were measured by plating onto tryptic soy agar and CHROMagar Salmonella with added nalidixic acid or rifampin. Individual growth curves at the same temperature and RH differed significantly among different lots of pistachios. Except for a single late-season lot in which no significant growth was observed, Salmonella multiplied under all storage conditions. In the first 3 h after inoculation, insignificant (most cases) to small (0.41 to 0.67 log CFU/g) but significant (P < 0.05) mean increases in Salmonella populations were measured; the mean predicted time to achieve maximum populations (5 to 8 log CFU/g) was 16 ± 4 h. In paired samples, longer lag phases, lower growth rates, and lower maximum increases were observed with inoculated inhull pistachios incubated at 23°C and 50% RH compared with 35 or 37°C and 90% RH. Similar growth curves were observed at the low and high inoculum levels; throughout the 30 h of incubation, Salmonella populations were consistently ∼1 to 2 log CFU/g lower on pistachios inoculated at the low inoculum level. Managing the time between harvesting and hulling will reduce the potential for growth of Salmonella on pistachios during postharvest handling.
Low-moisture foods do not support the growth of foodborne pathogens. However, low-moisture foods, including several different tree nuts, have been associated with outbreaks of salmonellosis and, more rarely, enterohemorrhagic Escherichia coli gastroenteritis (13). Pistachios have been recalled in the United States for the presence of Salmonella (21). Consumption of raw or roasted pistachios was linked to one case of salmonellosis in 2009 (7) and two multistate outbreaks of salmonellosis in 2013 and 2016 (8, 20). Raw inshell pistachios collected shortly after harvest from California storage silos had a 3-year average weighted prevalence of Salmonella of 0.61% (3,966 100-g samples) (12); geometric mean levels of Salmonella in positive samples, as determined by a most-probable-number (MPN) method, ranged from 0.0010 to 0.053 MPN/g. The source of Salmonella in pistachios is unknown; however, similar to other tree nuts, introduction is possible at any step from production, harvest, and postharvest handling to final processing (10, 11). Adequately dried pistachios (<7% moisture [approximate water activity (aw) of <0.70]) do not support the growth of foodborne pathogens. The decline of Salmonella populations during storage of dried inshell pistachios was slow (approximately 0.15 log CFU/g/month) at 23°C, or insignificant over more than a year at 4°C (15, 17).
The United States leads the world in commercial pistachio production (2), and nearly all (99%) of the total U.S. crop is produced in California (3). Production of pistachios in California has grown rapidly over the past two decades, from 81 million kg in 1997 to over 408 and 272 million kg in 2016 and 2017, respectively (1). Pistachio trees are alternate bearing, producing a greater than average crop one year and a lower than average crop the following year. California produces a small number of cultivars (primarily Kerman) that have a relatively short harvest window (18). The overall production increases and somewhat unpredictable annual harvest size have sometimes strained the postharvest handling system during the relatively short harvest period (approximately 6 weeks from late August to early October), increasing the possibility for delays in unloading harvest trailers (2, 18).
In California, with the exception of young trees (≤6 years), pistachio fruits are harvested by mechanically shaking the tree and dropping the pistachios onto a catch frame (18). Maturation may be uneven throughout the tree, and thus an orchard may be shaken more than one time over a period of several weeks. The fragility of the hulls increases with increasing maturity associated with the second or third shake of the trees.
A conveyor system moves the fruit from the catch frame to a plastic bin or into a small hopper. The bins or the hoppers are moved to the edge of the orchard. For a few of the smaller growers, the plastic bins are used to transport the pistachios to a hulling facility. For others, either a forklift (bins) or conveyor system (hoppers) is used to transfer product to a bottom dump trailer. These trailers are then transported to hulling facilities, where the pistachios are weighed, sampled, and evaluated for quality. Trailers are unloaded as the processing capacity of the facility allows. After unloading, pistachios are hulled and then dried to 8 to 15% moisture using forced hot air (70 to 105°C) (11). Further drying takes place after the pistachios are transferred to large silos, with application of forced warm, dry ambient air over the first few days of storage. At this point pistachios can be held for up to 14 months before further processing.
The time the harvested pistachios remain in the trailer before unloading is impacted by the speed at which the trailer is filled, the time for the trucker to retrieve the trailer from the orchard, the distance (travel time) from the orchard to the hulling facility, and the hold time between receiving and unloading of the trailer (11). Harvested pistachios are 40 to 50% moisture on a fresh weight basis (18), and their respiration can result in rapid elevation of temperature (19) and relative humidity (RH) in the harvest trailers. Pistachio hulls contain 3 to 4% soluble sugars (primarily glucose and fructose) on a dry weight basis (16), providing a potential food source for microorganisms. The objective of this study was to examine the fate of Salmonella enterica on inhull pistachios during simulated postharvest conditions, from the time of shaking the tree to just before hulling.
MATERIALS AND METHODS
Temperature and humidity monitoring
Daily high and low ambient temperatures were obtained from the nearby Stratford weather station through the California Irrigation Management Information System (5). Ambient conditions at the orchard were determined by a single data logger (TempTale 4, Sensitech Inc., Beverly, MA) placed near the location where pistachios were loaded into bottom dump trailers (capacity, ∼25,000 kg). Temperature and RH in loaded trailers were monitored and recorded using data loggers. To protect the data loggers from physical damage, a stainless steel coupon (5 by 10 cm, 32-mm thick) was placed on each side. The data loggers were then placed inside steel mesh envelopes. On one occasion, a single data logger was suspended on metal chains in the approximate center of each of six different trailers before they were loaded with pistachios; when the trailer was full, the data logger was covered with pistachios to a depth of ∼1.2 m. On a separate day, seven data loggers were suspended along the length of a single trailer at the approximate midpoint of the trailer's width at a depth between 0.3 and 1.0 m from the top of the loaded trailer. Temperature and RH were recorded every 2.5 min for up to 13 h, from before loading, through loading, transportation, and unloading. The trailers were purposely held at the hulling facility to facilitate measurements over a longer time.
Inhull pistachios (Pistacia vera) were collected during five commercial harvests (2010, 2011, 2014, 2016, and 2017) from two different processors in the Central Valley of California. Samples were collected during the “early season” (week 1 or 2), “midseason” (week 3 or 4), or “late season” (week 5) of the harvest. Inhull pistachios were collected as they were unloaded from harvest trailers or bins (receiving pit). The pistachios were placed into polyethylene bags (30.5 by 30.5 cm; Bitran, Com-Pac International, Carbondale, IL), which were sealed and then transported (∼4 h) on ice to the laboratory. The sealed bags were then transferred to lidded plastic tubs and stored at 4°C for up to 12 h before inoculation.
Nalidixic acid–resistant (in 2010 and 2011) or rifampin-resistant variants of the following strains were used: Salmonella enterica Enteritidis phage type 30 (ATCC BAA-1045), isolated from raw almonds associated with a 2000 to 2001 outbreak (14); Salmonella Enteritidis phage type 9c (RM4635), a clinical isolate from a 2004 outbreak associated with raw almonds (6) (provided by Dr. Robert Mandrell, U.S. Department of Agriculture, Agricultural Research Service); Salmonella Anatum LJH1242, isolated from an almond survey (9); Salmonella Tennessee (K4643), a clinical isolate from a 2006 to 2007 outbreak associated with peanut butter (provided by Dr. Larry R. Beuchat, University of Georgia, Griffin); Salmonella Montevideo (GRC1), isolated from pistachios (7) (provided by U.S. Food and Drug Administration); and, in some cases, Salmonella Oranienburg (1839; used in 2010 and 2011 cocktails in addition to the Salmonella strains listed above), isolated from pecans (provided by Dr. Larry R. Beuchat). Isolates were stored at −80°C in tryptic soy broth (TSB; BD, Franklin Lakes, NJ) supplemented with 15% glycerol.
Preparation of inocula
Unless otherwise specified, all media were obtained from BD. The preparation of inocula and inoculation procedure were based on the method described by Kimber et al. (15). Briefly, before each experiment, frozen stock cultures of each strain were streaked onto tryptic soy agar (TSA) supplemented with 50 μg/mL nalidixic acid (TSAN) or 75 to 100 μg/mL rifampin (TSAR) and incubated at 37 ± 2°C for 24 ± 4 h. From each plate, a single isolated colony was transferred into 10 mL of TSB and incubated at 37 ± 2°C for 24 ± 4 h. A loopful (∼10 μL) of each culture was then transferred into fresh TSB and incubated at 37 ± 2°C overnight. An aliquot (1 mL) of each overnight culture was spread over a large TSAN or TSAR plate (150 by 15 mm; Fisher Scientific, Pittsburgh, PA) and incubated at 37 ± 2°C for 24 ± 1 h. After incubation, the resulting bacterial lawn was collected by adding 9 mL of 0.1% peptone to each TSAN or TSAR plate and then scraping the slurry on the plate surface with a sterile spreader. A multistrain cocktail was prepared by combining equal volumes (5 mL) of the cell suspensions of each Salmonella strain, resulting in levels of 11 log CFU/mL. The cocktail was then diluted with sterile ultrapure water (Milli-Q Advantage A10, MilliporeSigma, Burlington, MA) to the desired final target inoculum level.
Before inoculation, pistachios were removed from refrigerated storage and held at room temperature for ∼1 h. Preliminary data indicated that the temperature of the pistachios approached ambient temperature during this time. Inhull pistachio samples (400 g) were weighed into polyethylene bags (30.5 by 30.5 cm; Com-Pac International), and 25 mL of inoculum was added. Each bag was sealed and then shaken continuously by hand for 2 min to ensure a thorough coating of the nuts with the inoculum. The inoculated nuts were then spread onto two sheets (46 by 57 cm) of filter paper (Qualitative P5, Fisher Scientific) that were folded in half and placed in a biosafety cabinet for 30 min (to a point where they were visibly dry).
Incubation of inhull pistachio samples
In initial experiments (2010 and 2011), different humidity levels were evaluated: after the 30-min drying period the inoculated pistachios were placed into large weigh boats (50% RH or ambient humidity) or in zippered plastic bags (>90% RH) and held for 24 h on a laboratory bench at ambient conditions (23°C) or in an incubator at 35°C. In later experiments (2014 to 2017), inoculated pistachios were transferred to a metal tray inside a plastic bin, and the open bins were placed in an environmental chamber (Percival, Geneva Scientific LLC, Fontana, WI) and incubated for up to 30 h at 37°C and 90% RH. In 2017, inoculated samples were incubated at both 23 and 37°C and 90% RH. Uninoculated control pistachios were held under the same conditions. Temperature and RH were monitored and recorded using data loggers (Sensitech Inc.) during the incubation period.
Enumeration of inoculated cells and background microbiota
At each sampling time, inhull pistachios were mixed, and 10 g (2010 and 2011) or 40 g was combined with 20 or 80 mL of 0.1% peptone, respectively, in a two-chamber filter bag (Nasco, Modesto, CA). Each sample was processed by shaking for 30 s, rubbing for 15 s, and shaking for an additional 30 s. The liquid portion in each bag was serially diluted in 0.1% peptone. Appropriate dilutions were plated in duplicate using spot plating (20 μL of each dilution), spread plating (100 μL per plate), or a spiral plater (Autoplate, Advanced Instruments, Inc., Norwood, MA) onto TSAN or TSAR supplemented with 50 μg/mL cycloheximide to suppress molds, and onto bismuth sulfite agar (2010 and 2011) or CHROMagar Salmonella (CHROMagar, Paris, France) supplemented with 75 μg/mL rifampin. Samples were incubated at 37 ± 2°C for 24 ± 2 h. Colonies were counted manually or with a ProtoCOL 2 colony counter (Synbiosis, Frederick, MD). Because the pistachios were not liquefied during mixing, the calculated CFU per milliliter of diluent multiplied by the ratio of the volume of diluent to weight of the sample was considered equivalent to the CFU per gram of pistachios.
The background microbiota for uninoculated pistachios was determined by plating appropriate dilutions onto TSA for aerobic plate count (APC) and, for 2016 samples, onto CHROMagar ECC for Escherichia coli–coliform counts (ECCs), and incubating at 37 ± 2°C for 24 ± 2 h. All colonies visible on TSA (APC), and all pink (presumptive coliform) and all blue (presumptive E. coli) colonies on CHROMagar ECC were counted. In most cases, uninoculated samples were also plated onto TSAR and CHROMagar Salmonella supplemented with rifampin (75 μg/mL) and incubated at 37 ± 2°C for 24 ± 2 h.
Characterization of pistachio samples
In some cases, inhull pistachios, as obtained from the hulling facility, were characterized on the basis of the weight of different fractions of a ∼1-kg sample. Each sample was weighed, placed in a tray, and then sorted into six fractions: sticks, loose material (leaves, hulls, and unidentified debris), inhull (with hull or unhulled pistachios), inshell (hull-free inshell pistachios), shells (loose), and kernels (loose). Each fraction was weighed, and the weight percentage for each of the components was calculated.
pH of the hull extract
In 2016, hull materials were removed from early-season (two samples), midseason (one sample), and late-season (one sample) pistachios, and each sample was blended for 30 s in a blender (Waring, Torrington, CT). The hull extract was filtered through cheesecloth, and the pH of three subsamples of the resulting filtrate was determined with a digital pH meter (Thermo Fisher Scientific, Waltham, MA).
Moisture content and aw
Moisture content and aw were determined for uninoculated pistachio hulls or kernels at the beginning and end of the incubation period. Inhull pistachios were hulled and shelled. Hulls or kernels (40 g) were processed for 20 s in a commercial food processor (Waring). The aw and percent moisture were determined for triplicate subsamples of the ground material with a water activity meter (Aqualab model 4TE, Decagon Devices, Pullman, WA) and moisture analyzer (3.6- to 4.0-g samples; model HG63, Mettler-Toledo, Columbus, OH), respectively.
Pistachios collected from a single processor on a single day were considered an independent lot. For each incubation condition or inoculum level evaluated for that lot, two separate sublots of 1,000 g of pistachios were inoculated with the same inoculum preparation. At each sampling time, the pistachios were mixed and three samples were randomly taken from each of the two sublots (n = 6). Duplicate, uninoculated control sublots (1,000 g) were also stored under the same conditions as inoculated sublots. One random sample of uninoculated pistachios from each of the two separate sublots was plated onto selective and nonselective agars at each sampling time (n = 2). Three random uninoculated samples were used to determine moisture content and aw at the beginning and end of the incubation period. Data were analyzed using Microsoft Excel 2010 and Prism 6 software (GraphPad Software, Inc., La Jolla, CA). Analysis of variance, Tukey's multiple comparisons test, and t test were performed. Differences between the mean values were considered significant at P < 0.05. Individual growth curves for Salmonella on inoculated inhull pistachios were fit using the DMFit version 3.5 add-in to Microsoft Excel (http://www.combase.cc/) (4). The lag time, growth rate, and maximum population changes (the difference between the mean initial inoculated and mean highest concentrations) of Salmonella were determined.
RESULTS AND DISCUSSION
Temperature and humidity in pistachio harvest trailers
Most pistachios are harvested between dawn and dusk, but some commercial harvesters are equipped to operate around the clock. Pistachio temperature at the time of harvest depends on the ambient temperature and location of the nut within the tree canopy. Temperature and RH in multiple trailers loaded with pistachios were determined on a single day (27 September 2011) beginning at midday, with single data loggers placed in six different loaded trailers (Fig. 1A). Average ambient conditions at the start of loading were 27°C and 54% RH; high and low ambient temperatures reported from the nearby Stratford weather station for that day were 30.7 and 13.2°C, respectively. Temperature and RH in a single loaded trailer with seven data loggers were determined on a different day (11 October 2011) beginning at midday (Fig. 1B). Average ambient conditions at the start of loading were 28°C and 52% RH; high and low weather station temperatures were 28.5 and 14.9°C, respectively, for that day. The initial temperatures recorded by the metal-encased data loggers on the two test days (40 and 34°C, respectively) reflect direct exposure to sunlight before and while the trailer was being filled. During the first 2 h after the data loggers were completely covered with pistachios, the average RH in the loaded trailers increased to over 90 or 88% (Fig. 1A or 1B, respectively); temperatures decreased to 30 or 28°C (Fig. 1A or 1B, respectively). Subsequent increases in temperature up to 37°C were observed in trailers that were held for longer times (e.g., ∼12 h). In a previous study, initial load temperatures for eight different trailers were close to the air temperature at the time of harvest, ranging from 22 to 34°C (19); maximum temperatures at unloading (6.8 to 14.2 h after loading) ranged from 30 to 41°C. On the basis of these data, the present experiments were conducted at 23, 35, or 37°C and 50 or 90% RH to reflect a range of conditions up to a worst-case but commercially relevant scenario. Longer trailer hold times correlate with decreased product quality (e.g., shell staining), especially at higher temperatures (18, 19). At 25°C, significant increases in shell staining were sometimes noted at 40 h of holding (19). At 30 and 40°C, staining damage significantly increased at holding times of between 16 to 24 h.
Colonies were not detected when the lowest dilution of uninoculated inhull pistachios was plated onto TSAN or TSAR and bismuth sulfite agar, or CHROMagar Salmonella. E. coli colonies were not detected on CHROMagar ECC in any sample (<1.3 log CFU/g). Initial background populations for uninoculated early-season and midseason inhull pistachios ranged from 3.00 to 3.84 log CFU/g (APC) and 0.90 to 2.80 log CFU/g (presumptive coliform) (Fig. 2). Initial APC and presumptive coliform counts were significantly higher in one lot (2016) of late-season inhull pistachios (7.42 and 7.38 log CFU/g, respectively). APC and presumptive coliform counts increased within 12 h by 2.19 to 3.04 log CFU/g in early-season and by ∼1 log CFU/g in late-season harvested inhull pistachios, and they either decreased by 0.55 log CFU/g or increased by 1.99 to 4.48 log CFU/g in midseason inhull pistachios (Fig. 2).
Growth of inoculated Salmonella on inhull pistachios
For all experiments in which Salmonella was inoculated onto pistachios, Salmonella counts on TSAN or TSAR and on bismuth sulfite agar or CHROMagar Salmonella for individual pistachio samples were not significantly different at any time (P > 0.05); thus, only data from TSAN or TSAR are presented. However, individual growth curves at the same temperature and RH differed significantly among different pistachio samples, and thus each lot of pistachios collected within or among different years was treated separately.
In initial experiments conducted in 2010 and 2011, inhull pistachios were collected from a single huller during late-season commercial harvest, inoculated with Salmonella at 4.29 to 4.31 log CFU/g, and held at either 23 or 35°C and 50 or 90% RH (Fig. 3). During the first 12 h of incubation at 23°C, maximum mean increases in the population of Salmonella ranged from 0.77 to 0.96 log CFU/g at 50% RH (Fig. 3A) and from 1.90 to 1.92 log CFU/g at 90% RH (Fig. 3B). At 35°C, maximum mean increases in the population of Salmonella in the first 12 h ranged from 0.77 to 2.59 log CFU/g at 50% RH (Fig. 3D) and from 2.13 to 3.53 log CFU/g at 90% RH (Fig. 3E). For all paired samples (same lot of pistachios, same inoculum preparation, different incubation conditions), consistently higher maximum population increases were observed at higher humidity and higher temperature (Fig. 3). Similar observations were made in 2017 when inoculated pistachios were incubated at 23 or 37°C (Fig. 3C and 3F).
For subsequent experiments, inhull pistachios were collected from one of two hullers on different days during the early-, mid-, or late-season commercial harvests in 2014 and 2016. Pistachios were inoculated with Salmonella at low (0.64 to 1.59 log CFU/g) or high (2.73 to 3.27 log CFU/g) levels. On early-season and midseason inhull pistachios, in the first 3 h after inoculation, small but significant (P < 0.05) mean increases in Salmonella populations (0.41 to 0.67 log CFU/g) were observed at the low inoculum level, whereas no significant (P > 0.05) increases in Salmonella populations (0.10 to 0.40 log CFU/g) were observed at the high inoculum level (Fig. 2). Salmonella populations increased by over 4 log CFU/g to >5 log CFU/g (low inoculum level) or to >7 log CFU/g (high inoculum level) after 12 h. Thereafter, populations continued to increase, plateaued, or decreased. Similar growth curves were observed at both the low and high inoculum levels; throughout the 30 h of incubation, Salmonella populations were consistently ∼1 to 2 log CFU/g lower on pistachios inoculated at the low inoculum level. In contrast, for the single lot of late-season inhull pistachios that were evaluated in 2016, no significant change in populations of Salmonella was observed during 30 h of incubation at either the low or high inoculum levels (Fig. 2). It is possible that high background populations observed in the 2016 late harvest pistachios (>7 log CFU/g) depleted readily available nutrients, thus prohibiting growth of inoculated Salmonella (Fig. 2).
Salmonella growth parameters
Sigmoid functions (DMFit or asymmetric sigmoidal five-parameter logistic) were used to fit growth curves where typically a no-growth linear phase (lag) is followed by a steep mid-phase (log) and then a stationary phase. In most cases, similar curve fits were derived from the two fit functions. However, owing to an insufficient number of early time points in 2010 and high population variability at each time point in 2011, DMFit predicted linear curve fits for five growth curves (2010, 23°C and 50% RH; 2010, 23°C and 90% RH; 2011, 23°C and 50% RH; 2011, 35°C and 50% RH; 2011, 35°C and 90% RH). In these cases, only the asymmetric sigmoidal five-parameter logistic function was used to construct fits more like a classic growth curve (Table 1 and Supplemental Figs. S1 and S2).
Longer lag phases, lower growth rates (all but one case), smaller highest concentrations, and smaller maximum increases were observed for inoculated inhull pistachios incubated at suboptimal growth conditions, i.e., at a lower incubation temperature and/or RH (23°C and 50% RH), compared with the corresponding pistachios (same lot) incubated at higher temperature and RH combinations in the 2010, 2011, and 2017 harvest years (Table 1 and Figs. S1 and S2A and S2B). The time to highest predicted concentration ranged from 9.6 to 23 h.
In harvest years 2014 and 2016, inoculated inhull pistachios were incubated at 37°C and 90% RH, which was at the upper end of measured temperature and RH for harvest trailers (Fig. 1) (19). Lag times of 2.0 to 4.4 h, growth rates of 0.52 to 0.70 log CFU/g/h, and maximum increases of 3.5 to 5.3 log CFU/g were predicted for early-season and midseason harvested inhull pistachios at the high initial inoculation level.
In laboratory-based studies, artificially high levels of inoculum are often used to facilitate enumeration. Lag time, growth rate, maximum populations, or time to maximum population derived from these inoculum levels are assumed to reflect theoretical outcomes for much lower and more realistic contamination levels. In the current study, the “high” inoculum level (3 log CFU/g) might reflect a point source contaminant, such as a localized small amount of fecal material from a bird or rodent.
In 2016, low and high initial inoculum levels of Salmonella were compared on the same lot of early-season and midseason inhull pistachios. High inoculum levels were associated with longer predicted lag phases, higher growth rates, larger highest concentrations, and lower maximum increases (Table 1). Highest predicted populations of Salmonella were 0.69 to 1.49 log CFU/g higher at the high initial inoculum level. The time to achieve the highest predicted concentration was longer at the low inoculum level.
Characterization of inhull pistachios
Uninoculated inhull pistachio lots collected in 2016 were sorted into different fractions, including the inhull nuts, loose material (mostly free hull and leaves), the hulled inshell nuts, free kernels, free shells, and sticks or twigs (Fig. 4). The proportion of material in each fraction was similar for early-season and midseason pistachios (Fig. 4); more than 98% of the material by weight was made up of pistachios with an intact hull. In contrast, for the late season lot, 62.8% of pistachios had intact hulls; large amounts of loose material (12.1%) and hulled inshell pistachios (22.9%) were also present. At full maturation, the inshell pistachio will easily separate from the hull when pressure is applied along the axis of the nut (18). As maturity progresses, the hulls become increasingly susceptible to mechanical injury. The large percentage of inshell pistachios in the 2016 late-season lot is indicative of over-mature pistachios.
Moisture content and aw
The initial moisture levels of hulls and kernels from uninoculated inhull pistachios were significantly different among samples, ranging from 75.5% (midseason, 2014) to 80.7% (midseason, 2016) for hulls and from 34.3% (midseason, 2014) to 46.7% (midseason, 2016) for kernels (Table 2). The initial aw was 0.98 or 0.99 for both hulls and kernels. Decreases in moisture content were significant, ranging from 2.0 to 13.3% (hulls) and 2.0 to 8.3% (kernels), between 1 and 30 h of incubation; aw did not change or decreased by a maximum of 0.02. The pH of hulls from samples collected over 4 weeks in 2016 was 5.14, 4.95, 5.03, and 5.07 (average, 5.05 ± 0.10).
The current study demonstrates that Salmonella can multiply on inhull pistachios under commercially relevant times and temperatures. The lag times, growth rates, and maximum and final populations observed varied significantly among independent replicate experiments. However, with paired samples from the same lot of pistachios, growth was positively influenced at a higher temperature and humidity. Lower initial inoculum levels resulted in lower maximum population outcomes. In most cases, insignificant or small but statistically significant increases in populations were observed in the first 3 h of incubation; predicted lag times ranged from 3 to 7 h at 23°C and from 0 to 4 h at 35 and 37°C. The predicted mean time to achieve maximum population was 16 ± 4 h. No growth was observed in a single lot of late-season inhull pistachios. There were significantly greater amounts of inshell and loose material in this lot (Fig. 4), and the initial APC was far higher than in other inhull pistachios evaluated (>7 log CFU/g versus 3 to 4 log CFU/g; Fig. 2), which potentially resulted in a depletion of readily available nutrients.
These findings stress the importance of managing postharvest handling of pistachios to minimize the time between harvest and hulling. In all but two cases, significant growth of Salmonella was observed at 23, 35, and 37°C and at both 50 and 90% RH when incubation times exceeded 3 to 6 h. Harvest during cooler times of day (e.g., late evening and early morning) or under conditions that prevent or minimize temperature and humidity increases in the trailers would reduce the potential for increases in Salmonella populations in inhull pistachios prior to hulling. This study focused on the upper measured temperature and humidity in harvest trailers. Additional information on individual trailer hold times, as well as receipt and unloading temperatures, would further help to inform the risk from delays in unloading trailers. Significant growth of Salmonella could occur before quality impacts, such as shell staining, are evident. Further work is needed to assess the impact of hulling and drying on microbial populations in inshell pistachios.
This research was supported, in part, by the California Pistachio Research Board, the Center for Produce Safety, and the Specialty Crop Block Grant Program at the U.S. Department of Agriculture (USDA) through grant 14-SCBGP-CA-0006. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the USDA. The contributions made by Dr. Anne-laure Moyne, Lillian Khan, Ethan Morgan, and Sylvia Yada in technical support and during the writing process are greatly appreciated.
Supplemental material associated with this article can be found online at: https://doi.org/10.4315/0362-028X.JFP-18-351.s1.