Introduction

Animal feeds contaminated with Salmonella pose a risk of infection to livestock and thus to the human food chain (Crump et al. 2002; EFSA 2008). In humans, it is considered one of the most frequent bacterial foodborne diseases posing major public health threats in industrialized countries (D’Aoust 1997). In the USA, it has been estimated that 1.4 million people are infected with nontyphoidal Salmonella, resulting in 15,000 hospitalizations and approximately 400 deaths every year (Voetsch et al. 2004). Approximately 6000 to 12,000 cases of Salmonellosis are reported annually in Canada (http://dsp-psd.pwgsc.gc.ca/collection/HP3-1-32-20.p). There were about 152,000 confirmed cases of human salmonellosis were reported within the EU in 2007 (EFSA 2009), but the decreasing trend in confirmed salmonellosis cases was reported from 2008 to 2012. A total of 92,916 salmonellosis cases were reported by the 27 European Union Member States, with 91,034 confirmed cases in 2012 (EFSA 2014). As a zoonotic agent, animals as well as wildlife can be important contributors to the abundance and distribution of Salmonella in the environment and possible transmission to humans (Smith et al. 2005; Ashbolt and Kirk 2006; Compton et al. 2008; Handeland et al. 2008; Hernandez et al. 2012). It has been estimated that more than 34,500 symptomatic infections with Salmonella spp. occur annually in the Netherlands (about 16.5 million population) (Havelaar et al. 2009). With Enteritidis and Typhimurium being the two most frequently isolated Salmonella enterica subsp. enterica serotypes, accounting for 29 % (Enteritidis) and 47 % (Typhimurium, including its monophasic variant 4) (van den Brandhof et al. 2004). Salmonella contamination in human mostly related to animal product through animal ingestion of Salmonella contaminated feed (Davies and Wales 2010). Studies have shown strong links between Salmonella contamination of feedstuffs or feed mills and infections, with Salmonellas of the same serovars, of groups of chickens (Davies et al. 2001; Jinghui et al. 2014), turkeys (Zecha et al. 1977; Nayak et al. 2003), pigs (Newell et al. 1959, Joost et al. 2013), goats (Duffy et al. 2009), cattle (Glickman et al. 1981; Min et al. 2014), small ruminants, and wild-living birds (Pao et al. 2014).

The importance of animal feed as source for Salmonella dissemination has been highlighted by EFSA (European Food Safety Authority), particularly in countries with a low prevalence of Salmonella in primary production (EFSA 2008). Animal feedstuffs are at risk of Salmonella contamination at several stages in the feed chain starting with the production of ingredients. Different factors are known to influence the risk of introducing Salmonella into animal feed, including contaminated ingredients, contaminated feed mill environments causing re-infection of the feed, wild birds, and rodents (Hoszowski et al. 2008; Andrés-Barranco et al. 2016). According to European feed legislation, Regulation (EC) No. 183/2005, feed business operators should apply Hazard Analysis and Critical Control Points (HACCP) principles and good hygiene practice/good manufacturing procedures (GHP/GMP) at each stage of the feed chain in order to secure safe feed (EFSA 2014). At present, China has taken strict detection and control measurements of Salmonella animal feed. Animal feed could not be contaminated with Salmonella according to Chinese official regulation GB13078-2001 (SAPRC. 2001). It has been reported that the feed yield in China in 2011 was 200 million tons, ranged the first in the world and surpassed firstly United States, feed mill almost distributed the whole country (CFIA 2013). However, the situation of animal feed contaminated with Salmonella in China was relatively serious compared to the USA, Japan, and some European countries. It was found that non-typhoid Salmonella annually caused 9.874 million gastroenteritis cases in China and 91.5 % of them were caused by food transmission (Mao et al. 2011). Several reports have shown animal feed and food contamination with Salmonella existed in different provinces such as Henan, Sichuan, Guangdong, Fujian, Xinjiang, and Gansu (Chunhong et al. 2004; Qiong et al. 2006; Yuanran et al. 2013). In these reports, the degree of contamination with Salmonella in raw materials, especially in animal proteinaceous feed materials, was more serious than that in compound feeds. Jijun and Biao (2009) reported the positive rate of 18.2 % of fish meal and 15.4 % of meat and bone meal contaminated with Salmonella from 299 samples, while 10.1 % positive rate for the finished feeds and 7.8 % positive rate for the concentrated feeds. Additionally, Salmonella contamination could be observed based on the above studies in different sections of food chain.

The Karst rural area, an undeveloped district located in the southern west of China, has been recently listed as an ecological agricultural developing region by the Chinese government due to its specific geographical and ecological environment. There are more than 200 enterprises for feed production and sales. However, there appears to be very little literature concerning feed production and food chains with Salmonella contamination in this area. To clarify the mechanisms of the contamination, accumulation of basic data is necessary for local government to draw up the plan to control Salmonella contamination in feed production chain. In the present study, samples were collected from the different sections in feed production chain in Karst rural areas of China from the year 2009 to 2012, and the data used for analysis in this research can be as a reference to feed production in the other areas of China. It provides a basic data to evaluate Salmonella contamination in feed production chain and helps to identify the contamination sources and sites. Therefore, the feed quality supervision system may be improved accordingly.

Materials and methods

Sample collection

All samples used for this study were collected from different departments distributed in nine regions of Karst rural areas in China between 2009 and 2012. Sampling sites included feed mills, feedstuff shops, and farms, where represented the feed production chain in whole Karst areas. The total of 1077 samples were collected, 234 of which were taken in 2009, 328 in 2010, 301 in 2011, and 214 in Table 1. Sampling collecting method meet the standard of GB/T14699.1 in China. For each kind of sample, three packs of the same samples (500 g for each) with a total of 1500 g of weight were collected, which could meet the requirements of examination for three laboratories. For the test of Salmonella, 8 analytical samples of 25 g were tested.

Table 1 Number of samples collected from nine regions from 2009 to 2012 in Karst rural areas
Full size table

Reagent and broth preparation

Biochemical reagents including buffered peptone water (BPW), enrichment broth with magnesium chloride, enrichment broth with cystine, and selenite were purchased from China import and export commodity inspection institute; Phenol red agar, deoxycholate hyadrogen sulfiale lactose agar (DHL), magnesium chloride malachite green medium, nutritive agar were from Hangzhou microbial Co. Ltd. China. API ID32 reagent strips were made up of 32 small dry medium tubes, 3 ml ampoule of 0.85 % NaCl solution, JAMES reagent, petroleum wax were produced by French Merieux biological company, purchased from Guangzhou QianJiang Company, China. A–F polyvalent serum was from Lanzhou biological products research institute, the ministry of health of China.

Bacteriological examination

The detection of Salmonella from feed was based on culture methods according to the national standard of the People’s Republic of China (GB - T13091-2002) (19). Briefly, Salmonella was cultured on phenol red agar, the plates were incubated at 37 °C for 24 h, and the suspicious bacteria was identified in DHL plate when colors of cell edges or whole cell appeared black. Then, the suspicious bacterial colonies were picked up for culturing on nutrient agar plate at 37 °C for 18–24 h. The pure cultures were taken out for biochemical test using the API bacteria identification system and results were read automatically on ATB Express/mini API (BIOMÉRIEUX, France); strains suspected of being Salmonella were used to conduct further serological identification.

Biochemical identification of Salmonella bacteria

Twenty-five grams of sample each was emulsified and then inoculated into the 500-ml flask with 225-ml BPW buffering peptone, incubated at 37 °C for 8 to 18 h. Ten milliliters of enrichment broth cultures were inoculated into 100-ml selenite cystine medium and incubated at 37 °C for 24 h. At the same time, 10-ml broth cultures were inoculated into 100 ml of magnesium chloride malachite green medium at 42 °C for 24 h. Two sorts of enrichment broth were separately cultured and inoculated on DHL and phenol red agar plates, then incubated at 37 °C for 20–24 h.

Serological test

The suspected small bacteria colonies growing on phenol red agar whose medium color changing from pink to red were selected to conduct the further biochemical identification using the API automatic verification system with ID32 reagent strips. Salmonella isolates were obtained and identified to serovar level according to the Kauffmann-White scheme (Grimont and Weill 2007).

Results

Contamination of various kinds of feeds with Salmonella

In total, 1077 samples were collected and analyzed from nine regions of Guizhou province in China from 2009 to 2012 during this study (Table 1). In 2009, a total of 234 samples were determined, 42 of which represented 8 sorts of feed materials, 106 of which were commercial finished feed samples, 85 of which were commercial concentrated feed samples, and 1 of which was the sample of self-made by farm. The results showed that among all samples, 11 of which were contaminated with Salmonella, and the Salmonella contamination rate were 4.7 %. The high positive rates were detected in meat meal, meat and bone meal, feather meal, blood meal, and fish meal. No Salmonella contamination was found in commercial finished feeds, commercial concentrated feeds, bone meal, and feeding yeast (Table 2). In 2010, among 328 samples, 11 of which were contaminated with Salmonella (4.3 %). In 2011, only 2 samples were positive among 301 samples (0.66 %), while 6 samples contaminated among 214 samples (2.8 %) in 2012. Compared to the positive rates of contaminated with Salmonella from 2009 to 2012, Salmonella contamination mainly distributed in animal proteinaceous materials such as meat meal, meat and bone meal, feather meal, blood meal, and fish meal. No Salmonella contamination was observed in vegetable protein material like rapeseed or soybean meal, but some Salmonella contamination was found in the concentrated feed and the self-made feed by farm (Fig. 1).

Table 2 Number of samples contaminated with Salmonella from different types of feeds from 2009 to 2012
Full size table
Fig. 1
figure 1

Positive rate of Salmonella contamination from feedstuff and feed ingredients in Karst rural areas of China during 2009–2012

Full size image

Contamination of production sections with Salmonella in feed chain

Salmonella contamination status from different sections of feed production chain was also investigated from 2009 to 2012. Table 3 shows a total 1077 samples inspected, 33 sample appeared to be positive (3.1 %), 4 of them were from feed ingredients, 5 of which were from feedstuff in feed mill, 7 were from sale sections, 13 of them were from processing of feed ingredients, 4 from feedstuff in farms. These results imply that there was Salmonella contamination existed in each section of feed chain. In 33 positive samples, the animal proteinaceous materials with Salmonella contamination reached 63.6 %, the concentrated feeds were 30.3 %, self-made feed by farm was only 3.0 %. No positive samples were detected in vegetable proteinaceous materials. As to sections of feed production chain, section of feed materials processing had the highest proportion Salmonella contamination (39.4 %), followed by section of compound feed in sales (21.2 %). Section of compound feed in feed mill, section of feed materials in feed mill, and section of compound feed in farm were 15.2, 12.1, and 12.1 %, respectively.

Table 3 Salmonella contamination in different sections of feed production chain
Full size table

Effects of processing methods on feed contaminated with Salmonella

Effects of two kinds of processing methods on feed contaminated with Salmonella are listed in Table 4; the results indicated that no Salmonella contamination was detected in 775 samples collected from the feed processed with bulking method, while 21 samples were found to be positive from 35 feed samples processed with pelletizing method. The temperature for bulking ranged from 93.3 to 178.6 °C for 45–60 s, moisture content of puffing feed was 25–35 %, while the temperature for pelletizing was 71.1 to 82.2 °C for 2–16 s, moisture content of pelleting feed was 11–19 %, the temperature for bulking was significantly higher than that for pelletizing, the time spent for puffing also more than that for pelletizing. This may explain why no Salmonella contamination rate was detected for bulking method.

Table 4 Comparison of Salmonella contamination of two processing methods
Full size table

Discussion

There was an overall reduction tendency in positive rate of Salmonella contamination. It reflects that the feed producers gradually improved the consciousness of avoiding Salmonella contamination in feed production. Although there were differences in sampling numbers and sampling types for examination from 2009 to 2012, the sources of Salmonella contamination mainly came from animal feed raw material especially from meat and bone meal and blood meal. Vegetable protein materials such as rapeseed meal and soybean meal had not been found positive. This result is consistent with the previous reports (Hacking et al. 1978; Nabbut 1978; Harris et al. 1997). The reason is twofold: (1) the vegetable fat has been shown to inhibit the growth of Gram negative bacteria (Khan and Atamy 1969; Fay and Farias 1975) and (2) high temperatures (up to 105 °C) may eliminate Salmonella when rapeseed meal and soybean meal being processed (Wierup and Kristoffersen 2014). However, it has been reported that Salmonella was frequently isolated from consignments of vegetable proteins as feed materials (Martin and Per 2010; Wierup and Häggblom 2010; Wierup and Kristoffersen 2014). In a Dutch report, 3.2 and 6.7 % of Brazilian extracted soybeans were found positive for Salmonella during 2002 and 2003, respectively (Anon 2003). In a recent comprehensive study based on an annual examination of up to 80,000 lots of feed, Kwiatek et al. (2008) reported that in Poland up to 15.0 and 15.4 % of imported lots of soy and rapeseed were respectively Salmonella contaminated in 2005–2007. From 1994 to 2012 in a Norwegian, 34 % of samples collected during unloading of ships delivering soybeans yielded Salmonella; the proportion of samples from ships that yielded Salmonella varied from 12 to 62 % each year. Corresponding data for products produced in Poland were 6.3 and 7.7 %. In the present study, the rapeseed meal was from the southwestern provinces, while soybean meal came from northeastern China. Heat treatment is the mandatory procedure in producing of materials. Steam (water vapor) is used during process of pelleted feedstuff, and it can be used for sterilizing of feed material as well. Although the situation, to some extent, seems to be safety, the potential risks still exist for feed producers when imported vegetable proteinaceous materials from other provinces or the other countries because oil seed feed ingredients are often contaminated by Salmonella. Besides, it is difficult to compare the level of contamination between different studies because the results depend on the sampling and culture techniques applied (Martin and Per 2010). Therefore, it is essential for feed producers in Karst rural areas to introduce HACCP-control to reduce the risk of Salmonella contamination. Heat treatment can also prevent Salmonella contamination of feed ingredients to be compounded feed.

Concentrated feed is also important source of Salmonella contamination. It may be caused by the difference feed processing methods: no heat treatment process in production of concentrated feed, only a mixture of proteinaceous materials and the multiple kinds of vitamin and minerals. In China, concentrated feed has a large market space because of its convenience, particularly in Karst rural areas, small farms can make full use of agricultural byproducts to reduce feed costs through mixturing concentrated feed. It is not necessary for small farms to invest money to establish feed mill. However, attentions must be paid in Salmonella contamination when using concentrated feed because of its high levels of proteins materials. In addition, the humid air especially in Karst rural areas easily leads to moldy feed. For large farms, pelleting and puffing feeds are often used because they have their own feed mills. Both pelleting and puffing need a heat treatment in feed processing. Pelleting usually involves temperatures between 70 and 90 °C (though pelleting at 90 °C may result in “slushy” feed) and may destroy 99 % of total bacteria (Furuta et al. 1980; Mccapes et al. 1989). In the EU, the control and elimination of Salmonella contamination in the crushing plants are important ways to prevent Salmonella introduction into EU farms. Control of Salmonella in crushing plants is feasible if recontamination is prevented since the production process reaches temperatures (up to 105 °C) that would eliminate Salmonella (Wierup and Kristoffersen 2014). This suggests that physical or chemical approaches are of importance for the raw materials and concentrated feed to eliminate Salmonella contamination prior to mixture of ingredients. A large number of reports demonstrated that Salmonella contamination could be reduced by chemical amendments, heat treatment, and irradiation (Leeson and Marcotte 1993; Jaquette et al. 1996). Heat treatments may also be more effective when used in conjunction with chemical treatments such as propionic acid, even though the effect may not be synergistic (Mattlho et al. 1997). Cold plasma is a relatively new process being developed for applications to fresh and fresh-cut fruits, vegetables, and nuts (Niemira and Gutsol 2010). Brendan (2012) reported that short treatments with cold plasma significantly reduced the viability of E. coli O157:H7 and Salmonella on the surfaces of almonds.

In our study, samples taken from all sections of feed production chain had been inspected to be positive for Salmonella, particularly in sections of feed materials processing, and the positive rate of Salmonella contamination was relatively high (39.4 %). The consequence is in line with other investigations (Chunhong et al. 2004; Qiong et al. 2006; Yuanran et al. 2013). It is worth mentioning that feed inspection is not only an effective way to check Salmonella contamination. It has been verified that HACCP associated with Salmonella contamination control is the best way (Martin and Per 2010). In China, there are entirely differences between manufacturing of finished feed and processing of feed materials in different provinces, and no consistent criterion can be provided for preventing Salmonella contamination. Until now, the feed quality tracing system for Salmonella contamination has not been established. This can attributed to the following aspects: firstly, although most of feed mills possess the basic conditions for inspection of Salmonella contamination, the consciousness of feed quality assurance is not strong. Most of mini-manufacturers of feed materials in Karst rural areas have poor conditions in professional people and facilities for Salmonella inspection. Secondly, the level to control feed quality does not meet the requirements of national standards, causing some unqualified animal raw materials were used. Thirdly, potential safety risk exists in sections of transportation and sales in feed chain and lack of related control measures. Fourthly, feed mills were not able to invest enough money to improve inspection technology and facilities. Finally, most of animal farms have their own rooms for feed processing with a simple grinder and mixer; they used to use blended mash without heat treatment.

Considerable researches reveal that heat treatment involving cooking and pelleting may destroy 99 % of total bacteria (Halls and Tallentire 1978). Bulking feed is an important type for piglets and chicks due to its high digestive ability and hygiene. In this study, two kinds of processing methods with different temperatures and heating time were compared and result showed a higher effectiveness in bulking procedure. There was a higher temperature (93.3–176.8 °C), a longer heating time (45–60 s), and a higher moisture content (25–35 %) in bulking procedure than that in pelleting procedure. This might be a reason why no positive rate detected in bulking procedure while 60 % positive rate detected in pelleting procedure. This result is in consistent with the report by Himathonkham et al. (1996). Additionally, the other method such as irradiation has been investigated as the alternative feed treatment. When vegetative bacteria are exposed to gamma rays, the structure of bacterial DNA is changed, resulting in a reduction in bacterial populations (Clavero et al. 1994; Aude et al. 2013). This method, here, has not been designed to conduct a trial in this study but it may be considered as a selective measure to reduce feed contaminated with Salmonella.

Conclusion

Feed contaminated with Salmonella in Karst rural areas, the southwest of China, mainly comes from animal feed ingredients, particularly from animal proteinaceous materials such as meat meal, meat and bone meal, feather meal, and blood meal. Most of the sections of feed chain have been monitored to be positive for Salmonella in these areas.