2.1 Produce samples

A total of 128 samples of RTE and packed vegetable samples were collected randomly from 3 supermarket chains from June to November 2014 in the city of Loja, Ecuador. The 60 RTE samples (some containing two to three ingredients, including lettuce, cabbage, carrot) consisted of packed, cut vegetables of lettuce (Creole, Roman and American), parsley (plain and curly) and coriander. These samples were obtained unopened in the original containers prior to their expiration date, and the production batches were verified and within the shelflife time frame of up to eight days, as indicated by the labels. Samples were placed in ice chests and transported immediately to the laboratory for analysis.

2.2 Microbiological analysis

The metadata on all the samples were recorded, including the name of the manufacturer, type of vegetable, lot number, date of expiration and type of packaging. Samples that were deteriorated or had visible damaged were discarded and not included in the analysis.

The microbiological quality of the samples was determined with methods of the FDA Bacteriological Analytical Manual (BAM) (^{U.S., 2013}) (http://www.fda.gov/Food/FoodScienceResearch/LaboratoryMethods/ucm2006949) briefly, 50 g of the product (from 3 production batches) were weighed and mixed with 450 mL of buffered peptone water, from which 1:10 serial dilutions were made until the final dilution of 107. The mesophilic aerobic count was performed by standard plate count using plate count agar (Difco BBL). The counts of coliform bacteria and E. coli were determined by the MPN method using Lauryl Sufate Tryptose broth, but modified with the use of ColiComplete disks (Biocontrol Bellevue, WA) (^{AOAC, 2012}) as described in literature (^{Valentin-Bon et al., 2008}). The disks contain X-Gal (5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside) which is cleaved by β-galactosidase to produce a blue color that is indicative of coliform. The discs also contain 4-methylumbelliferone-ß-D-glucuronide, which is cleaved by E. coli ß-glucuronidase activity to produce blue fluorescence visible at 365-nm UV. After 48 h of incubation at 37°C, tubes with blue dye were counted as coliform positive and those with blue fluorescence were positive for E. coli. The level of coliform and E. coli present was estimated by MPN/g based on the combination of positive tubes.

The identification of the genus and the species of the gram-negative bacilli was carried out touching lightly the center of the colony with sterile inoculating needle and inoculate to Triple iron Agar (Difco BBL) slant, Lysine iron agar (Difco BBL) slant, Urea broth (BBL), Citrate medium (Difco BBL), and Sulfide indole motility agar (BBL) by streaking slant and stabbing butt. The culture was incubated 24 ± 2 h at 35°C and for the confirmation the Microgen GN-ID identification system was used.

3.1 The mesophilic aerobic count

In the mesophilic aerobic bacteria count the mean microbial loads for RTE were lettuce 5,66±0,6, spinach 6,45±0,8, carrot 7,23±0,2, cabbage with carrot 7,31±0,3 log₁₀ CFU/g, in the case of packaged vegetables, the mean microbial loads were: parsley smooth 5,58±0,6, parsley curli 6,09±0,4, coriander 6,03±0,5, lettuce creole 5,18±0,8, lettuce romaine 6,00±0,4 and spinach 6,00±0,4 log₁₀ CFU/g, these values are shown in Table 1.

3.2 Total coliforms and E. coli

The number of total coliforms and E. coli in RTE and packaged vegetables are presented in Tables 2 and 3. Total coliforms were not detected in 6,7% of the vegetables (spinach).

All 5 samples (100%) of grated carrot and cabbage with carrot contained coliform levels between 5,04 and >5,04 log₁₀ MPN/g.

E. coli was detected in 13 of 60 (21,7 %) RTE vegetable samples and in 19 of 68 (27,9 %) packaged vegetable samples (Table 3), but the levels were less than 6,2 MPN/g.

3.3 Discussion

The total counts of the lettuce samples we analyzed had a mean value of log 5,66 and ranged from 4,57 to 6,78 log₁₀ CFU/g (Table 1). In Iran, RTE vegetables were examined and reported mean total counts of log 6.7 with ranges from 5,5 to 7,4 log₁₀ CFU/g ^{Jeddi et al., 2014}). In the Campinas area of Brazil, analysis of conventional and organic vegetables reported total count values from 4,00 to 6,89 log₁₀ CFU/g (^{Maffei et al., 2013}) and from 6,07 to 7,48 log₁₀ CFU/g (^{Maistro et al., 2012}), respectively. Our results are consistent with these data and with those reported from the United States in 55 samples of RTE salads (^{Hagenmaier and Baker, 1998}) and in 100 samples of RTE vegetables (^{Valentin-Bon et al., 2008}). In Spain, total counts between 3,1 to 7,81 log₁₀ CFU/g were reported in 140 lettuce samples collected from 16 restaurants (^{Soriano et al., 2006}). It has been noted that in all vegetable samples aerobial mesophylls were developed. In another study from Brazil, 21% of 133 samples of RTE vegetables showed > 6,0 log₁₀ CFU/g for coliforms (^{Fröder et al., 2007}). All these studies showed that RTE vegetables sampled worldwide contains high levels of total bacteria, often >7,0 log₁₀ CFU/g.

Large variability in total counts not only between different samples but also within samples from the same producer and with the same production date have been previously described (^{Valentin-Bon et al., 2008}). We found similar patterns in our study, all samples studied were positive for coliforms, where lettuce and spinach samples had values ranged from 4,57 to 6,78 and from 5,47 to 7,82 log₁₀ CFU/g, respectively.

In some samples, condensations in the bags, probably from humidity, were observed at the bottom of the bags, and this may have provided moisture which can stimulate microbial growth. These results suggest that the microbiological qualities of the packaged vegetables reported from the various countries appear to be similar. The mesophyll aerobic count is used by some countries as an indicator of safety, but has no direct relation with the presence of pathogens or their toxins. In Ecuador, there are no regulations for mesophyll aerobic count for this type of products, but others have suggested 5,0 log₁₀ CFU/g as acceptable for aerobic counts (^{Solberg et al., 1990}). The National Standard of Iran uses aerobic counts of less than 6,0 log₁₀ CFU/g as being considered safe for consumption. In this context, our study showed that 50% of the produce samples in Loja, Ecuador would have exceeded this norm. Moisture was observed in the bottom of many of the bags of lettuce and spinach we tested, which may have favored microbial growth and contributed to the high counts. The RTE plant products may be contaminated at each stage of their production process (both pre- and post-harvest and processing). ^{Berthold-Pluta et al., (2017)} indicated that potential sources of contamination include soil, water, feces (human and animal origin), animals (for example insects and birds), handling of the products, harvesting and processing equipment and transport.

The greatest applications of coliform, Enterobacteriaceae, and E.coli testing is in assessment of the overall quality of a food and the hygienic conditions present during food processing. The Enterobacteriaceae, coliforms, and Escherichia coli are used with as Quality and Safety indicators.

The total coliform counts we obtained were from 0,47 to >5,04 log10 MPN/g (Table 2), which are higher than the 0,47 to 3,38 log10 MPN/g reported in Spain (^{Soriano et al., 2006}) and the <0,47 to >4,0 log10 MPN/g reported in the United States (^{Valentin-Bon et al., 2008}). In our study, it was not possible to obtain a mean coliform value because 13 of the 60 samples have exceed the counting limit of >5,05 log10 MPN/g. There are no permissible limits of indicator bacteria for RTE in Ecuador, but again, using the National Standard of Iran, 98% of our produce samples have exceeded that standards.

Generic E. coli bacteria were detected in 9 samples of lettuce, 2 samples of spinach, 1 sample of cut carrot and 1 sample of cabbage with chopped carrot (Table 3). In Brazil, the standard for minimally processed products for consumption has a limit of 100 CFU/g for fecal coliforms and in one analysis of 133 salad samples, 73% had exceeded that limit (^{Fröder et al., 2007}). In the United Kingdom, the Public Health laboratory has set E. coli limits for RTE foods, including bagged produce, where E. coli of 20 CFU/g is satisfactory, from 20 to <100 CFU/g as acceptable and >= 100 CFU/g as unsatisfactory (^{Sagoo et al., 2003}). In a few studies, 3200 organic salads and 3852 conventional RTE salads were sampled and analyzed in the United Kingdom and the results showed that only 0,5% had exceeded the permissible maximum E. coli limit of 100 CFU/g and were qualified as unsatisfactory (^{Sagoo et al., 2001}, ²⁰⁰³).

In our study, 25% of the samples had E. coli but did not exceeded the value of 6,2 MPN/g. This is similar to the results obtained in the United States (^{Jeddi et al., 2014}; ^{Valentin-Bon et al., 2008}) where E. coli was detected in 21 samples (18,1%) but at low levels, suggesting that E. coli are not usually present at high levels in bagged produce.

The common foodborne genera of the Family Enterobacteriaceae include Citrobacter, Enterobacter, Erwinia, Escherichia, Hafnia, Klebsiella, Proteus, Providencia, Salmonella, Serratia, Shigella, and Yersinia, we also found the presence of other Gram negative bacteria in produce but in different percentages, Pseudomonas spp (13%), Enterobacter aerogenes (45%), Enterobacter cloacae (8,3%), Klebsiella Oxytoca (1,6%), Citrobacter freundi (3,3%), and Enterobacter spp (1,6%). These results are consistent with those reported in literature (^{Seow et al., 2012}; ^{Soriano et al., 2006}). The presence of these microorganisms in foods, may be indicators of lack of food quality, indicator microorganisms are considered to be all groups or species of microorganisms involved in diseases of alimentary origin that are usually enteric pathogens, which means that they can survive in the gastrointestinal tract of humans, as well as animals and birds, a directly or indirectly contaminated with fecal material from any of these sources, is indicative of the possible presence of enteric pathogens (^{Ray and Bhunia, 2010}).

The results from all these studies showed that the flora and bacterial content in RTE products are highly variable and complex. The microbiological quality of fresh RTE products from production to packaging has been studied (^{Johnston et al., 2006}) and showed that indicator levels in mustard and spinach leaves remained relatively constant, whereas in coriander, parsley and, above all, melons had varied counts demonstrating that the microbial load not only vary according to processing, but also depends on the type of product. The processing procedures for each one of the produce products is different moreover, variations in the microbial load can also occur during storage.

Prevention of foodborne contamination in fresh-cut processing starts in the field by identifying and eliminating possible sources of contamination. Principles of Good Agricultural Practices (GAP) and Good Handling Practices (GHP) remain the pre-requisite cornerstones of food safety management strategies to address challenges posed at the pre-harvest level (^{Castro-Ibañez et al 2017}).

In conclusion, there is a wide range and variations in the microbial counts in the fresh produce samples in Ecuador and even between samples with the same use-by-dates. These RTE produce products can contain very high levels of microbial content and these products are difficult to disinfect, hence can pose health risks if contaminated with pathogens. To improve the safety of fresh vegetables in the industry some effective measures should be applied such as chemical and physical treatments, refrigeration and packaging in modified atmosphere can inhibit or retard bacterial growth.

For the prevention of contamination in fresh processed vegetables, possible sources of contamination should be identified, Principles of Good Agricultural Practices (GAP) should be applied, Good Handling Practices (GHP), food safety management based on the previous analysis of the harvest.