Hot-air drying processes are used to provide specific quality attributes to products, such as dehydrated apple pieces. To comply with the U.S. Food and Drug Administration Food Safety Modernization Act, there is a need to understand microbial lethality during these processes. The objective of this study was to determine the level of inactivation provided by hot-air drying on a Salmonella cocktail inoculated onto apple cubes and to evaluate the performance of Enterococcus faecium as a surrogate. A cocktail of Salmonella serovars (Agona, Tennessee, Montevideo, Mbandaka, and Reading) and E. faecium were individually inoculated onto cored, peeled Gala apple cubes at 9.2 ± 0.3 and 8.8 ± 0.1 log CFU per sample, respectively. Apple cubes were dried at 104 or 135°C in ∼1.5-kg batches using a hot-air dryer with a vertically directed heat source and without mixing. Three subsamples, consisting of four inoculated cubes, were enumerated at each time point (n ≥ 5) from multiple product bed depths. Water activity decreased throughout the duration of the study, with samples drying faster at 135 than 104°C. Samples at the bottom bed depth, closer to the heat source, dried faster than those at the higher bed depth, regardless of temperature. Significant microbial inactivation was not seen immediately. It took >10 min at the bottom bed depth or >40 min of drying at the top bed depth, regardless of temperature (P < 0.05). By the end of drying, average Salmonella inactivation of greater than 5 log CFU per sample was achieved. At temperature conditions evaluated, E. faecium inactivation was slower than Salmonella, indicating that it would likely serve as a good surrogate for in-plant validation studies. Case hardening did not inhibit microbial inactivation in the conditions tested. Hot-air drying under the conditions evaluated may provide a preventive control in the production of dehydrated products, such as apples.
Hot-air drying at 104 and 135°C provided >5-log Salmonella reduction on apples.
E. faecium serves as a conservative Salmonella surrogate during apple drying.
Case hardening did not affect microbial inactivation during hot-air drying of apples.
Dried or dehydrated apple products are those that undergo an active drying process to remove moisture, extend shelf life, and provide specific quality attributes. Although there have been no recorded outbreaks associated with low-moisture apple products, there have been recalls due to Salmonella on freeze-dried fruit slices (32) and Escherichia coli O157:H7 on sweetened dried cranberries (5), and there have been 20 outbreaks attributed to contaminated apple cider and caramel apples between 2010 and 2020 (8). Outbreak investigations have found that microbial contamination is likely to occur in orchards or packing facilities where soils may serve as a vector for contamination (7, 9, 17, 22, 31). Once products become contaminated, the bacterial pathogens, such as Salmonella, can survive at 5°C (12, 18, 23) and colonize apple surfaces under ambient temperatures from 15 to 22°C, commonly seen during storage and in primary production (12). Apples surfaces are difficult to sanitize and may remain a risk because pathogens are able to survive and/or grow during preproduction storage (11, 14, 22–24).
In the production of low-moisture apple pieces, selected apple varieties are typically peeled, cored, cut to specification (size and geometry), and dehydrated using ambient to hot air temperatures. The efficacy of microbial inactivation of these heat-based processes is greatly unknown because they were originally intended to deliver quality attributes, not pathogen reduction as required by the Food Safety Modernization Act (31). The challenge with low-moisture foods is that bacteria show an increased desiccation resistance and persistence at low water activities (19, 28, 30) as well as an increased thermal resistance (2, 20, 27–29), making them difficult to eradicate once dried. Although Salmonella and E. coli have been identified as the pathogens of concern in fresh produce (22) and are commonly associated with low-moisture foods (2, 24), Salmonella remains the pertinent pathogen to date due to its known increased persistence and resistance at low moisture. Suehr et al. (26) found that Salmonella Anatum, Salmonella Agona, and E. coli O121 had greater survival than E. coli O157:H7 at low moisture and low and neutral pH conditions. Additionally, research has shown that Salmonella Enteritidis PT 30 was more thermally resistant than E. coli O121 in wheat flour (25).
Gala apples are a commonly dried variety that requires higher temperatures and longer residence times. Gurtler et al. (13) found dehydration resistance of Salmonella greatest on two cultivars, Gala and Envy, of the six cultivars studied at a moderate process temperature. Product preparation, such as peeling, shape, and size, can also affect dehydration and microbial inactivation. Studies have shown that the average microbial attachment to apples was 16 to 19% higher on flesh than skin (18). These data indicate the importance of investigating microbial inactivation on apple flesh.
Processing parameters include dryer bed configuration (length, width, height, and conveyor material), air direction, air recirculation, time, temperature, and humidity. Previous studies have investigated the dehydration of apples at ambient conditions. Research has shown that dehydration of apples at moderate temperatures may provide some microbial reduction (4, 10, 13). DiPersio et al. (10) studied survival of a five-strain cocktail on Gala apple slices dehydrated for 6 h at 60°C. They found a 2.7- to 2.8-log CFU/g reduction, with approximately 4.8 log CFU/g remaining. Gurtler et al. (13) found similar results with minimal microbial reduction after drying Gala apple slices at 60°C for 5 h. There was 5.58 log CFU of Salmonella (five-serovar cocktail) per slice remaining postdehydration. Studies on E. coli–inoculated apple rings found similar results, with populations reduced by 2.9 and 3.3 log CFU/g at 57.2 and 62.8°C, respectively (4).
Because moderate temperature alone was unable to serve as a preventive control, understanding the efficacy of higher-temperature dehydration on microbial inactivation is needed. Because dehydration is a dynamic process in which moisture is driven off as product temperature rises, complexities arise in predicting microbial inactivation. Once moisture is driven off of the surface, the rate of dehydration is determined by how fast moisture can diffuse from the interior of the product to the surface (15). If evaporation from the surface occurs faster than internal moisture migration to the surface, case hardening can occur, with areas of lower water activity on the outer surfaces of the product. If there are any bacteria remaining on these outer surfaces subsequent to the case hardening, the remaining bacteria may experience increased thermal resistance (21, 33). This phenomenon has led to knowledge that the traditional “cold” spot does not necessarily correlate to the “least-lethal” spot on the product. Wang (33) evaluated microbial reduction and product temperature at both the surface and core of a Salmonella-inoculated dough throughout a baking process. It was found that greater microbial reduction occurred at the core of the product, with a cooler final temperature, than at the surface, which had a higher measured temperature (P < 0.05). It was hypothesized that case hardening led to a lower water activity on the surface, reducing Salmonella inactivation (33). Ensuring that microbial inactivation occurs prior to case hardening is imperative to implementing hot-air drying as a preventive control in the manufacture of dried apple products. Because there are no data on microbial inactivation of hot-air drying of apples, this study will serve as foundational information to characterize Salmonella inactivation. Additionally, understanding the correlation of the pertinent pathogen with a potential surrogate microorganism will aid in future in-plant validation studies.
The objective of this research was to determine whether the level of inactivation provided by hot-air drying at 104 or 135°C on a Salmonella cocktail inoculated onto apple cubes could be used as a preventive control and to evaluate the performance of Enterococcus faecium as a potential surrogate for Salmonella inoculated onto the surface of apple cubes.
MATERIALS AND METHODS
A cocktail of five Salmonella serovars from outbreaks and/or with known thermal and storage survival were used in this study, including Agona 44767 (associated with puffed rice cereal recall in Minnesota; U.S. Food and Drug Administration [FDA], Office of Regulatory Affairs [ORA], Arkansas Regional Laboratory, Jefferson, AR), Tennessee K4642 (associated with a 2006 peanut butter outbreak; Beuchat, University of Georgia, Athens), Montevideo 488275 (associated with a 2009 to 2010 black pepper outbreak; FDA, ORA, Arkansas Regional Laboratory), Mbandaka 698538 (associated with sesame tahini from Turkey; FDA, ORA, Arkansas Regional Laboratory), and Reading Moff 180418 (with known thermal and desiccation persistence; FDA, Center for Food Safety and Applied Nutrition Laboratory, Bedford Park, IL). In addition, a potential surrogate, E. faecium NRRL B-2354 (U.S. Department of Agriculture, Agricultural Research Service, Peoria, IL), was investigated.
Isolates, maintained as frozen stock cultures at −80°C, were used to prepare a working culture by resuscitating into 10 mL of tryptic soy broth (TSB; BD Difco, Franklin Lake, NJ) followed by incubation at 37°C for 24 h. Cultures were streaked for isolation onto tryptic soy agar with 0.6% yeast extract (TSAYE; BD Difco) and incubated at 37°C for 24 h. These working cultures were stored at 4°C and restreaked onto TSAYE on a monthly basis for a maximum of 6 months. All experimentation was completed within this time frame.
To prepare cultures for experiments, a single isolated colony was aseptically transferred into 10 mL of TSB and incubated at 37°C for 24 h. Overnight cultures, 0.1 mL, were individually spread plated onto TSAYE and incubated at 37°C for 24 h.
Plate-grown cells were harvested by pipetting 1 mL of 0.01% buffered peptone water (Difco) onto each plate. Cells were gently agitated into solution using an L-shaped spreader and were pipetted into a conical tube. Individual serovars were ∼10.5 to 11 log CFU/mL.
To prepare the cocktail, equal volumes of Salmonella serovars were combined together. Cocktails were mixed by vortex, and 0.25 g of Red 40 per mL of inoculum was added. Addition of Red 40 dyed the apples a pale red color, which allowed researchers to distinguish inoculated and uninoculated apples during hot-air drying studies. Prior to hot-air drying studies, preliminary research was conducted to confirm that the addition of Red 40 did not significantly affect microbial survival on inoculated apple cubes 2 h postinoculation. Additionally, inoculated samples were thermally treated at 104°C (n = two time points). All samples were enumerated, as described in “Microbial enumeration.”
Preparation and inoculation of apple cubes
Whole snack-size Gala apples (Tree Top, Inc., Selah, WA) were stored at 4°C to mimic storage of apples as seen in industry. Immediately before each trial, apples were removed from cold storage, visually inspected for damage, and assessed for firmness. Apples that exhibited damage or bruising or that were outside of the acceptable range of 10 to 25 lb/in2 were excluded. Apple pressure was measured on fresh-cut flesh using a 7/16-in. (11.13-mm) tip handheld penetrometer. Apples were peeled, cored, and diced in 0.635-cm cubes using a Dynacube Table Top manual food cutter (Dynamic CL003, Dynamic International, Pewaukee, WI). Cube geometry was selected because it would provide a uniform drying profile in the single-direction, single-pass dryer used in this study. Apple cubes were enumerated on sTSAYE formulated with tryptic soy agar, 0.6% yeast extract, 0.05% ammonium iron citrate, and 0.3% sodium thiosulfate prior to inoculation. The pH of fresh-cut apples was measured with a Mettler Toledo SevenCompact pH Ion S220 meter (Mettler Toledo, Columbus, OH).
Prepared apples, 240 g, were weighed inside a Whirl-Pak bag (Nasco, Fisher Scientific, Pittsburgh, PA). An aliquot of prepared inoculum, 15 mL, was pipetted onto the apples. Apples with added inoculum were hand massaged for 5 min to apply inoculum to the surface of apple cubes. Inoculum homogeneity was assessed via enumeration of 10 four-cube subsamples after inoculation, as described in “Microbial enumeration.” Salmonella and E. faecium populations on apple cubes were 9.2 ± 0.3 and 8.8 ± 0.1 log CFU/g, respectively.
Hot-air apple drying
Peeled, cored, cubed apples (∼1.5 kg, 8.8- to 9.0-cm depth) were loaded into the removable glass cylinder chamber with mesh screen floor (20.3-cm diameter, 324-cm2 drying area). Not all apples were inoculated due to personnel safety during experimentation. Inoculated samples on the bottom of the chamber were added first; next, uninoculated cored, cubed apples were added; and then inoculated apple cubes were added on the top. Samples for enumeration were within 1.5 cm of either the bottom base or top surface of the chamber.
The loaded chamber was placed onto the heating element of the Wenger Dryer (Wenger Manufacturing, Sabetha, KS). Dryer capabilities include a fan with airflow velocity of 0 to 1.53 m/s (variable frequency drive controlled); electrical heater with process air temperature of up to 260°C; data acquisition of temperature, air velocity, humidity, load cell loss in weight, differential pressure before and after product bed; and a Sussman MBA electric steam generator (Sussman Electric Boilers, Long Island City, NY). Apples were dried for up to 2 h using a hot-air dryer with single-pass vertically directed heat source and a fixed bed depth. Inlet temperature, at either low temperature (104°C) or high temperature (135°C), was varied per trial, and air velocity, 25% of maximum speed, was held constant. High and low process temperatures of 135 and 104°C, respectively, used in this study reflect expected maximum and minimum industrial conditions expected in multiphase hot-air dehydration, based on historical data provided by industry. Previous research (data not shown) found no significant difference in drying rate for products at 25 and 40% of equipment maximum (P > 0.05). The dryer chamber materials and configurations were kept constant throughout each trial. The initial 8.8- to 9.0-cm bed depth allowed researchers to evaluate product located closer (bottom bed depth) and further (top bed depth) from the heat source. Time and temperature are varied to produce specific quality attributes. Lower inactivation is thought to occur with either lower temperatures of drying, which may not be sufficient to provide microbial inactivation, or higher temperatures, which may induce case hardening prior to microbial inactivation. No added moisture via steam generation was introduced into the system.
Effect of hot-air drying on apple product measurements
Representative thermal profiles of apple cubes were conducted at both high and low temperatures for samples located at both the top and bottom of the product bed depth. Thermal probes were inserted into the geometric center of apple cubes. Temperature measurements were collected every 2 s using a remote-transmitting, six-channel datalogger (Datapaq, MultiPaq21, Fluke Process Instruments, Cambridge, UK).
Water activity measurements were conducted at both high and low temperatures for samples located at both the top and bottom of the product bed depth throughout the process. At least three replicate water activity measurements were completed per time interval at each temperature and bed depth (Aqualab 4TEV, Decagon Devices, Pullman, WA).
In a representative test, air velocity and relative humidity (RH) were measured every 15 min for 2 h. Air velocity was measured at the top of the sample chamber (∼25.5 cm above product) using a volume flow anemometer (Omega HHF12, Miltronics Mfg., Norwalk, CT). RH was measured approximately 3.8 cm above the product bed depth within the sample chamber using a handheld measurement indicator with RH probe (M170, Vaisala, Helsinki, Finland).
Effect of hot-air drying on microbial inactivation
Three independent thermal inactivation trials were completed for apples inoculated with Salmonella cocktail and E. faecium at both low (104°C) and high (135°C) temperatures. Specific time intervals were selected for each temperature, microorganism, and bed depth, as appropriate. Preliminary research found that microorganisms inoculated onto apple cubes were inactivated faster at 135 than at 104°C. Therefore, sampling points were different at these two process temperatures, with sampling points ranging from 0 to 127.5 min at 104°C and 0 to 95 min at 135°C. Additionally, it was found that Salmonella is more sensitive to hot-air drying on apple cubes than is E. faecium; therefore, runs inoculated with E. faecium had delayed initial sampling and longer run times than those inoculated with Salmonella. Samples for apple cubes inoculated with Salmonella were taken between 0 and 120 min, and apple cubes inoculated with E. faecium were sampled from 0 to 127.5 min. Each trial consisted of multiple runs to allow investigators to sample product from the appropriate bed depth location. This was typically completed as a “top” run and a “bottom” run, completed independently of each other for each independent trial. Top runs had longer dry times because preliminary research showed slower thermal heating of the apples and slower microbial inactivation. Sampling at the lower bed depth closer to the heat source occurred between 0 and 67.5 min, whereas samples were taken at the higher bed depth further from the heat source from 0 to 127.5 min. Apples were discarded following each experimental run. For each trial, at least three red-hued subsamples consisting of four inoculated cubes were taken from at least five sampling intervals, including time zero.
Four cubes, 1.2 ± 0.3 g fresh weight with a surface area of 9.68 cm2, were analyzed per sample. Samples were collected aseptically, serial diluted in buffered peptone water, spread plated in duplicate onto appropriate modified TSAYE, and incubated for 24 h at 37°C. Salmonella-inoculated samples were plated onto sTSAYE formulated with tryptic soy agar, 0.6% yeast extract, 0.05% ammonium iron citrate, and 0.3% sodium thiosulfate. E. faecium–inoculated samples were plated onto eTSAYE formulated with tryptic soy agar, 0.6% yeast extract, 0.05% ammonium iron citrate, and 0.025% esculin hydrate, 97%.
Independent trials were conducted in triplicate, using fresh inoculum on separate days. Water activity measurements were averaged, and standard deviation was calculated. Three independent thermal inactivation trials were completed for apples inoculated with both the Salmonella cocktail and E. faecium, at both low (104°C) and high (135°C) temperatures. When microbial counts fell below the limit of detection (1.7 log CFU per sample), the limit of detection was used for calculation. Log reductions of each microbial population were calculated. One-way analysis of variance and paired t tests between sampling points to determine stability were calculated using Minitab 19 software (State College, PA) at a significance level of 95% (α = 0.05).
RESULTS AND DISCUSSION
Prior to each trial, apple pressure was confirmed to ensure that it fell within industrial specifications of 10 to 25 lb/in2. Bruised apples were discarded prior to experimentation. The average pressure of apples used in experimentation was 17.3 ± 4.1 lb/in2. The pH of fresh-cut apples was 3.80 ± 0.12. No Salmonella was detected on the apples prior to the experiments (limit of detection = 1.7 log CFU/g).
Product characteristics during drying
Dehydration of high-moisture products to low moisture changes quality parameters and extends product shelf life. Figures 1 and 2 show representative thermal profiles of apples located at the top and bottom of the bed depth at 104 and 135°C, respectively. Variations of up to ±10°C in overall temperature were seen. This is likely due to differences in geometric placement of thermocouples within the sample chamber as well as to variations in air flow through product. Apples processed at 135°C reached higher overall temperature than those processed at 104°C (Figs. 1 and 2). Regardless of temperature, samples near the bottom of the chamber heated faster than samples at the top (P < 0.05). This is expected, because samples near the bottom of the chamber are closer to the one-direction forced hot-air flow. Location of apples at the top bed depth resulted in an overall higher temperature by 80 min of drying at 104°C (Fig. 1). This is likely due to a combination of the extended process time and the moisture loss of apples, with apples closer to the heat source dehydrating faster than those farther away from the heat source, although this same phenomenon was not seen at 135°C (Fig. 2). Trials conducted at 104°C had a maximum duration of 120 and 127.5 min for Salmonella and E. faecium, respectively. Trials conducted at 135°C had a maximum duration of 80 and 90 min for Salmonella and E. faecium, respectively.
Average water activity of apple samples was determined throughout the drying process for samples located both at the bottom and top of the bed depth for low and high temperature conditions for the duration of the study (n ≥ 3; Tables 1 and 2). As expected, water activity decreased throughout processing, as moisture was driven off the apple samples. Initial water activity of the fresh-cut apple cubes was 0.990 ± 0.004, and it decreased to a final water activity of 0.247 ± 0.070 after 2 h. Samples at 135°C dried faster than at 104°C at the same bed depth. The water activity of samples at the bottom bed depth was 0.600 and 0.624 for 135 and 104°C, respectively, at 40 min. The same trend was seen at the top bed depth: the water activity of the apple cubes at 104°C was 0.701 and at 135°C was 0.524 at 90 min. Samples at the bottom bed depth, closer to the heat source, dried faster than those at the higher bed depth, regardless of temperature. This can be seen at 104°C, where the water activity of samples located at the bottom bed depth was 0.562 at 50 min and did not reach that same water activity at the top bed depth until 110 min.
Initially, both air velocity and RH were low. At t = 0 min, the air velocity was 0.00 m/s and RH was 44%. Recorded air velocity slowly increased to 0.51 m/s at 20 min. As drying progressed and moisture was driven off, shrinking product left larger voids through the bed that resulted in increased recorded air velocity post–dryer bed. The fastest recorded air velocity was measured between 30 and 60 min at 3.00 m/s. By the end of drying, dried apple cubes began to settle into the voids, restricting air flow and leading to a final air velocity of 2.70 m/s. RH increased as moisture was liberated from the apple cubes during the constant rate period. The RH reached a maximum of 97% at 10 min of drying. As moisture continued to be driven off of the apples, available water in the apples decreased. Because remaining product moisture was lower over time, less water could be introduced into the air and corresponding RH decreased. RH decreased back to initial levels of 45% by 30 min. The RH continued to decrease as the apple cubes dried; it was 5% at 1 h and was down to 1.1% at 2 h.
Only a portion of the 1.5 kg of diced apples was inoculated per run for safety of personnel and capacity to inoculate enough product for required runs. To identify inoculated product for subsequent sampling, partitioning or visual marking was required. Anderson et al. (1) determined that Red 40 served as a visual marker of inoculated material and did not have an effect on Salmonella survival during extrusion. Preliminary research was conducted to confirm that Red 40 was not lethal to the E. faecium, as well. Populations of inoculum on apples with and without Red 40 were 9.0 ± 0.1 and 9.1. ± 0.1 log CFU/g, respectively (Table 3). To confirm that Red 40 would not affect thermal lethality results, samples were heat treated in a single layer at 104°C for 6 and 20 min. At 6 min, populations were 8.8 ± 0.2 and 8.8 ± 0.1 log CFU/g, with Red 40 and without Red 40, respectively. At 20 min, populations were 8.0 ± 0.1 and 8.0 ± 0.7 log CFU/g, with Red 40 and without Red 40, respectively. There were no significant differences in populations with or without Red 40 at times 0, 6, or 20 min (P > 0.05), justifying the use of Red 40 as a suitable visual marker.
Active hot-air drying is a dynamic process in which moisture is driven off as heat is applied to the system and product temperature rises. These dynamic thermal and water activity conditions make it inappropriate to calculate D-values for microbial inactivation kinetics. Microbial log reductions from enumerated microbial populations at time intervals during hot-air drying at 104 and 135°C are presented in Tables 4 and 5, respectively. Microbial reduction was calculated by comparing populations at processing time points to initial counts. Starting populations for Salmonella and E. faecium were 9.2 ± 0.3 and 8.8 ± 0.1 log CFU per sample, respectively. As subsequent processing removed water, weight and surface area per cube decreased. For standardization of sampling, four cubes were always enumerated per sample per time point. Samples at the bottom bed depth experienced thermal inactivation earlier during the hot-air drying than samples at the top, regardless of microorganism or temperature (P > 0.05; Tables 4 and 5). At 104°C, significant population decreases for both Salmonella and E. faecium occurred by 15 min for samples at the bottom of the bed depth, providing 0.5- and 0.4-log CFU per sample reductions, respectively (P < 0.05; Table 4). At increased temperature (135°C), significant microbial inactivation for samples located at the bottom of the bed depth occurred by 10 min for Salmonella (0.8-log CFU per sample reduction) and by 12 min for E. faecium (0.3-log CFU/g sample reduction) (Table 5). Samples located at the top of the bed depth did not begin to experience significant microbial inactivation until later, 40 to 90 min, depending on microorganism and temperature.
Following a period of stable population, populations began to die as product temperature increased. Rapid inactivation of both Salmonella and E. faecium occurred as apple temperatures reached ∼100°C. Greater than a 5-log CFU per sample reduction was observed for both Salmonella and E. faecium inoculated onto apple pieces that were hot-air dried at 104 or 135°C, regardless of bed depth, by the end of drying (Tables 4 and 5). Results show that the time required to achieve >5-log Salmonella reduction at 104°C was between 45 and 50 min for the bottom and between 110 and 120 min for the top bed depth (Table 4). When temperature was increased to 135°C, the time required to achieve an equivalent reduction decreased to 20 to 25 min for the bottom and 70 to 75 min for the top bed depth (Table 5). Similar reductions of E. faecium were seen as well. At 104°C, it took ∼10 min longer to achieve a 5-log reduction of E. faecium at either bed depth: between 45 and 60 min at the bottom and between 120 and 127.5 min at the top bed depth. At 135°C, reduction of E. faecium required between 24 and 36 min on apples at the bottom and between 85 and 95 min on apples at the top bed depth. Over 5-log CFU per sample reduction was achieved at all conditions examined. In some cases, as much as 7.4-log CFU/g Salmonella reduction was found (Table 5). Maximum log reduction was dependent on initial inoculation level, limit of detection, and time point enumerated. Greater reductions may have been found, but sampling intervals were set to provide a 5-log reduction.
Research has shown that microbial reduction is greatly affected by both process temperature and moisture (water activity and/or RH). Previous studies at 60°C found a 2.0-log Salmonella reduction after 5 h (13), and 2.7- to 2.8-log Salmonella reduction after 6 h (10) on Gala apples. Similar results were seen with E. coli inoculated onto Gala apples (4). After 6 h, E. coli on Gala apples was reduced by 2.9 and 3.3 log CFU/g at 57.2 and 62.8°C, respectively (4). This study found greater microbial reductions at higher temperatures compared with those at the moderate temperatures used in previous published apple dehydration research.
Additionally, microbial reduction at the elevated process temperatures in this study provided greater reduction during apple drying than was seen in other low-moisture products processed at high temperatures, due to the higher initial water activity. Whereas this study was able to provide a >5-log reduction, processing of pecans and in-shell pistachios under similar conditions was able to provide only limited microbial reduction. These tree nuts and seeds have a lower native water activity, which affects reduction. Hot-air drying at 120°C for 20 min reduced a five-serovar cocktail of Salmonella on pecans by 1.18 to 2.04 log CFU/g (3). Increasing product moisture via a presoak, with all other parameters held constant, led to a significantly faster inactivation of Salmonella Enteritidis PT 30 on in-shell pistachios at 104.4°C (6). These results suggest the importance of ensuring microbial reduction during the earlier stages of apple drying, before significant water activity changes lead to increased microbial resistance.
During hot-air processing, there can be case hardening, induced by surface drying, which may inhibit microbial inactivation. Case hardening occurs when the outer surface hardens because it has dried more completely than the inner portion, which remains softer, with higher moisture. It was unknown whether higher temperatures and longer residence times would negatively affect microbial inactivation as seen during dynamic baking conditions experienced during cooking of meat (21) and dough (33). Marks (21) investigated inactivation of Salmonella on (surface samples) and in (internal samples) beef patties at low, medium, and high humidity at 218 and 232°C. That study found that humidity level significantly impacted microbial inactivation. In beef patties cooked in low humidity conditions, performance standards (internal temperature of 158.8°C and ≥6.5-log reduction) were met using internal samples, but were not met on the surface because the majority of samples had <6.5-log reduction. Similar results were seen during baking of a high-protein dough formulated with wheat flour, soy protein, and soy oil processed at 100°C. Results indicated that the multicomponent dough experienced case hardening prior to microbial inactivation, which limited microbial reduction on the product surface. After 90 min of thermal treatment, the survival population of Salmonella decreased from 9.57 to 4.90 log CFU/g in the sample surface and 4.06 log CFU/g in the sample center (33). Similar case hardening was observed by the end of processing for the apples at 104 and 135°C. Unlike for meat and dough, this was not found to affect microbial inactivation because case hardening did not occur until populations decreased below detectable limits (1.7 log CFU per sample).
Performance of E. faecium as a surrogate for Salmonella
Overall, E. faecium inactivation was significantly slower or not significantly different from Salmonella during apple drying (P < 0.05; Tables 4 and 5). In direct comparisons, there was significantly less E. faecium reduction than Salmonella reduction on apples processed at 104°C at 15 min, with 0.4- and 0.5-log reductions, respectively (P < 0.05). This was also seen at 30 min, with 1.9- and 0.7-log reductions for Salmonella and E. faecium, respectively (P < 0.05). At 45 min, there was no significant difference in microbial reduction, with Salmonella reduction of 4.6 ± 1.0 log CFU per sample and E. faecium reduction of 3.0 ± 1.7 log CFU per sample (P > 0.05). This was also seen in samples enumerated from the top bed depth at 90, 100, and 110 min, because there was no significant difference between microbial reductions (P > 0.05). At 110 min there were reductions of 2.3 ± 1.2 log and 2.5 ± 2.63 log for Salmonella and E. faecium, respectively. At 2 h, 10 min later, there was a greater than 6.5-log reduction of Salmonella but a less than 4-log CFU/g reduction in E. faecium (P < 0.05). Similar results were seen at 135°C (Table 5). There was significantly greater Salmonella reduction at 10, 25, 75, and 85 min (P < 0.05). Whereas there was 7.4-log reduction of Salmonella at 25 min, E. faecium reduction was only 1.3 ± 0.7 log CFU per sample at 24 min (Table 5). At 75 min, Salmonella reduction was 6.5 ± 1.5 log CFU per sample, but E. faecium reduction was only 0.8 ± 0.3 log CFU per sample.
The time required for a 5-log reduction per sample was greater for E. faecium than for Salmonella regardless of temperature and bed depth. On average, 5-log CFU per sample reductions required at least 7 and 11 min longer for E. faecium than for Salmonella at 104 and 135°C, respectively. This indicates that E. faecium serves as a conservative surrogate under the product and process conditions evaluated. In cases where it appears E. faecium inactivation was greater than Salmonella inactivation (104°C at 100 and 110 min), microbial reductions were not significantly different (P > 0.05). Jeong et al. (16) quantified the performance of Pediococcus sp. (reclassified as E. faecium, which was used in this study) as a surrogate for Salmonella Enteritidis PT 30 in moist air convection heating of almonds. Overall, the mean log reductions for Pediococcus were 0.6 and 1.4 log lower than those for Salmonella Enteritidis PT 30 (P < 0.05) at predicted reductions of 3 and 5 log, respectively, which supported current findings.
In summary, sample bed depth affected the thermal profile of the product but did not affect the overall inactivation, given appropriate threshold temperatures were achieved. Once product temperature achieved a threshold temperature (∼100°C), rapid microbial inactivation occurred. Microbial inactivation of greater than 5 log CFU per sample of Salmonella or E. faecium inoculated onto the surface of Gala apple cubes was achieved during the hot-air drying process at ranges from 104 to 135°C. Case hardening did not inhibit microbial inactivation during the hot-air drying process. These results suggest that hot-air drying may serve as a process preventive control for Salmonella during apple dehydration and that E. faecium, because its inactivation was similar to or slower than that of Salmonella, would likely serve as a good surrogate for in-plant validation studies.
This work was supported by the National Institute of Food and Agriculture, U.S. Department of Agriculture, under award no. 2015-68003-23415, and by Tree Top, Inc. Guidance on product and process conditions was provided by Tree Top, Inc. Authors have no known conflicts of interest.