Clostridium difficile is an anaerobic Gram-positive endospore forming bacterium that is considered to be the most significant cause of nosocomial infection acquired in humans (CDI) [1]. To date, the most significant epidemic strain of C. difficile encountered in healthcare facilities within North America and Europe has been a strain classified by restriction enzyme analysis are BI, pulsed field gel electrophoresis (PFGE) as NAP1 and PCR ribotyping as ribotype 027 (B1/NAP-1/027) [2]. This strain is widely disseminated internationally and as a common cause of both endemic and epidemic CDI. In addition, recent increases in community associated (CA)-CDI in patients with no recent contact with clinical environments has led to speculation of other potential sources of exposure. The potential that CDI could be a foodborne disease has been raised based on variable but often high rates of C. difficile colonization of food animals and identification of C. difficile in retail meat products [3]-[9]. Additionally, ribotype 078, a strain commonly associated with CA-CDI, is a predominant strain in food animals and food heightening concerns of a possible food origin [4],[10].

Data regarding the contamination of food of animal origin are mostly reported from Europe and North America. These studies have focused mostly on the prevalence of C. difficile at the farm or retail level, and only a few surveys have been performed at meat processing plants [6],[7],[11]-[13]. Rodriguez-Palacios et al (2007) reported the presence of C. difficile up to 20% in unseasoned hamburgers in Canada [3]. Further studies indicated that C. difficile spores are relatively unaffected by processing such as freezing, refrigeration and cooking ([14]-[16]. Furthermore, the epidemiology and prevalence of C. difficile in food in Iran are limited, despite the importance of CDI and a previous study that found ribotype 078 to be a leading cause of CDI in people at one hospital in Isfahan, Iran [17]. The rate of CA-CDI reported to be 24% in Iran, with no data available on sources of infection [17].

Therefore, the objective of the present study was to determine the prevalence of C. difficile in seasoned hamburger and examine the sources of C. difficile dispersal in hamburger processing plants in the same location in Iran.

Raw beef meat was the only ingredient found to be contaminated with C. difficile in 3/54 (5.6%; 95% CI [1],[15]).

Detection of C. difficile in raw beef was unsurprising and demonstrates the potential of meat as a source of C. difficile dispersal, something that is logical given its presence in food animals. However, C. difficile has been found in vegetables and other food items like fish, shellfish, edible bivalve molluscs, egg, ready to eat food (RTF) and salads [4],[21]-[26], so while C. difficile is likely always of human or animal origin, it is important not to ignore the various potential sources that might have been contaminated by human or animal sources (e.g. manure contamination of water, indirect contamination by hands). If the main source of C. difficile contamination of hamburger is raw beef and this bacterium is endemic in the cattle population, it is likely that this organism can be introduced continuously.

The relatively low prevalence of C. difficile reported in the present study (5.6%) is in the range of recently published studies from numerous countries reporting the isolation of C. difficile from raw beef spanning 1.65 to 42.4% of samples collected at retail level [4],[5],[9],[27],[28]. Previous Iranian data are limited, with a report of contamination of 1.65% (2/121 samples) raw beef samples (9) and 2.8% (1/35) beef samples in meat packaging plants [29]. In contrast, this organism was not isolated from any of 145 raw chopped beef samples in the Netherlands [30].

C. difficile was not found in any of the 45 non-meat ingredients of hamburger (14 textured soy proteins, 17 seasoning and 14 defrosted onions). Various potential explanations can be hypothesized. One is that the raw ingredients for these materials may be less likely to be contaminated. Another is that processing might have eliminated any C. difficile contamination, such as the heat generated during extrusion of textured soy protein [31].

The seasoning samples collected in this study were a mixture of black pepper, cinnamon, sumac and cumin were free of C. difficile. Seasonings could be contaminated with spores of Clostridium species because of the lack of proper sanitary conditions during collection or as a consequence of open air drying procedures [32]. However, in the processing plants selected in this study, the seasonings were supplied to the meat processing plants after being sterilized by irradiation in 2 KGy. A dose limit of irradiation for decontamination of microbial spores is in the range of 1-4 KGy [33], so the levels used here could have inactivated any C. difficile spores that might have been present.

Failure to isolate C. difficile from onions was not particularly surprising given the small sample size, but this bacterium has been isolated from onions in UK [34]. Similarly C. difficile was not found in a small number of onions in a Canadian study [22].

C. difficile spores are environmentally tolerant and resistant to many disinfectants, so it is not surprising to have found contamination in 2 out of 56 of environmental sites (3.5%; 95% CI [0, 12]). The two strains that were found isolated from one processing plant (D) during two visits approximately 6 months apart, although it is unclear whether this represents long-term persistence or (more likely) repeated contamination. Regardless, these results suggest that the hygiene procedures for cleaning and sanitation were not adequate. The standard operation procedure for sanitation used in processing plant D was washing the grinder with peracetic acid solution (4%) then rinsing with warm potable water. In the other plants where no C. difficile was found, the procedure was washing with sodium hydroxide solution (5%) followed by peracetic acid (4%) washing. The usage of oxidative agents and acids such as hypochlorite, hydrogen peroxide and peracetic acid is recommended to break the chemical bonds of food soils that build up the biofilm [35]. It has also pointed out that any residual organic material involved in biofilm formation could facilitate the attachment of spores and vegetative cells to meat processed in a grinder [11].

Of 28 hamburger samples taken at each step after molding and freezing, 4 (7.1%) were positive for C. difficile. As all positive isolates of C. difficile found in the same meat processing plant (D) belonged to an identical clone (RT 22), which was found on two separate occasions. This could suggest a common source of contamination, although inadequate data are available about strains found in food animals in Iran to properly interpret this finding. If this is a predominant food animal strain, the strain distribution noted here could simply reflect the background contamination of incoming meat products, although that would not explain the discrepancies in prevalence between facilities. Therefore, there must be concern that detection of the identical genotype (RT 22) from raw meat and final product at the same processing plant may indicate the persistence of this genotype during processing. In addition, colonization of the identical genotype in this processing plant may indicate the ineffectiveness of cleaning and sanitation. Complementary typing method such as multilocus variable-number tandem-repeat analysis (MLVA) or PFGE are needed to further investigate the persistence of C. difficile in the environment [36].

Detection of C. difficile from hamburger patties in 2/28 (7.1%; 95% CI [0, 23]) after 24 h freezing at - 20°C indicates the survival of spores at freezing temperature, something that is well established in human fecal samples [14]. However there are no data on effect of freezing on survival of spores in meat.

The results of molecular characterizing of the 9 C. difficile strains are presented in Table 2. All isolates contained tcd B while 7 isolates from 2 strains also possessed tcd A. Interestingly, all isolates lacked cdtB. This was unexpected given the high prevalence of binary toxin positive strains reported in other studies of raw meat, the predominance of ribotype 078 (a cdt possessing strain) in food animals in various countries and the commonness of cdtB in human isolates from a recent study [6],[9],[13]. These results indicate that further study of the strain distribution of C. difficile in animals and humans in the region is required.

PCR ribotyping of 9 isolates categorized into four ribotypes (IR 21, IR22, IR23 and IR24) (Table 2). Ribotype IR 22 was the most frequently 6/9 (66.6%) encountered in our study. This genotype isolated from raw meat, environmental samples and hamburger both after forming and freezing in one processing plant (D), suggests that raw meat may be the major source of contamination. However the specific source of C. difficile in meat needs to be established. The gastrointestinal tract is the most important source of C. difficile contamination [9].Transmission through animals hides, the slaughterhouse environment, the processing facility environment, processing equipment and the hands of personnel handling meat, must also be considered [9].

Inactivation of C. difficile spores by most of the cleaning and sanitation practices is difficult; therefore its accumulation in the environment increases the possibilities for contamination of meat. Ribotype IR 22 was isolated from various stages in processing plant D in both visits in the summer and autumn suggesting the excellent survival of spores in the processing plant. Ribotypes IR23 and IR24 were also found in processing plant D suggesting a relatively higher genetic diversity among C. difficile in this processing plant. Ribotype IR21 was the only distinct genotype found in beef meat imported from Brazil indicating the possibility of a geographical relationship of the genotypes.

Recently, hyper-virulent PCR ribotype 027 has emerged in North America and Europe in links to the hospital outbreaks [2]. This strain has not been reported yet in Iran in either food or clinical samples [9],[17], potentially due to the lack of sufficient research in this area. Ribotype 027 also was not found in the present study, nor was ribotype 078, based on inference of the lack of cdtB in any isolate. We have previously reported ribotype 078 as a common strain in both humans and meat in Iran [9],[17], so the absence of this strain in the current study was surprising.

This study is subjected to some limitations. First, sampling were conduced based on expiry date of ingredients where non-meat products kept for a significantly longer times. Therefore these products sampled less frequently than meat, resulting unbalanced sampling. Second limitation of this study is that only ribotyping used as strains characterization. Complementary typing methods such as MLVA or ideally whole genome sequencing would be needed to further investigate the epidemiology of C. difficile in the hamburger production facilities. Third, the possibility of lab cross-contamination of samples cannot be dismissed, although it may be unlikely for reasons cited above.

In conclusion, the study demonstrated the existence of toxigenic C. difficile in hamburger processing plants through different potential reservoirs such as raw meat and facilities. Occurrence of C. difficile in hamburger as a commonly consumed food in Iran is of public health significance although hyper-virulent strains of ribotype 027 and 078 were not found in this study. For reduction or prevention of C. difficile prevalence in food products, Good Manufacturing practices (GMP) and Hazard Analysis Critical Control Point (HACCP) system should be applied in food industries and followed by post-production control procedures by consumers such as proper cooking based on the adequate time for destruction of C. difficile spores [37]. Susceptible individuals with increased risk for development of CA-CDI should be also educated to minimize the exposure to this pathogen in the food supply [37].

The project was performed by ZE, JSW, HE, MJ and MCh. DNA extraction, PCR techniques were performed by ZE and MJ. Sample collection, culture, statistical analysis and manuscript writing were performed by ZE, JSW and MJ. All authors read and approved the final manuscript.