Characterization and risk factors of vancomycin-resistant Enterococci (VRE) among animal-affiliated workers in Malaysia

Authors


  • The first and second authors contributed equally for this work.

Correspondence

Latiffah Hassan, Department of Veterinary Pathology and Microbiology, UPM, Serdang Selangor, Malaysia. E-mail: latiffah@vet.upm.edu.my or tiffah_hassan@yahoo.com

Abstract

Aims

This study determined the risk factors and characteristics of vancomycin-resistant Enterococci (VRE) among individuals working with animals in Malaysia.

Methods and Results

Targeted cross-sectional studies accompanied with laboratory analysis for the identification and characterization of resistance and virulence genes and with genotype of VRE were performed. VRE were detected in 9·4% (95% CI: 6·46–13·12) of the sampled populations. Enterococcus faecium, Enterococcus faecalis and Enterococcus gallinarum were isolated, and vanA was detected in 70% of the isolates. Enterococcus faecalis with vanB was obtained from one foreign poultry worker. At least one virulence gene was detected in >50% of Ent. faecium and Ent. faecalis isolates. The esp and gelE genes were common among Ent. faecium (58·3%) and Ent. faecalis (78%), respectively. The VRE species showed diverse RAPD profiles with some clustering of strains based on the individual's background. However, the risk factors found to be significantly associated with the prevalence of VRE were age (OR: 5·39, 95% CI: 1·98–14·61) and previous hospitalization (OR: 4·06, 95% CI: 1·33–12·35).

Conclusion

VRE species isolated from individuals in this study have high level of vancomycin resistance, were genetically diverse and possessed the virulence traits. Age of individuals and history of hospitalization rather than occupational background determined VRE colonization.

Significance and Impact of the Study

This study provides comprehensive findings on the epidemiological and molecular features of VRE among healthy individuals working with animals.

Introduction

Vancomycin-resistant Enterococci (VRE) are among the most important hospital-acquired pathogens worldwide (Leclercq and Leclercq 2009), with the majority of reported outbreaks occurring in the USA (Tacconelli and Cataldo 2008). The emergence of vancomycin resistance among the genus Enterococcus has been linked to an increased use of glycopeptide antibiotics in human medicine and the use of avoparcin as growth promoter in pigs and poultry (Bates et al. 1994; Boerlin et al. 2001). Livestock, especially chickens and pigs, are frequently incriminated as reservoirs of VRE for humans (Bates et al. 1994; Vignaroli et al. 2011); therefore, individuals working closely with animals, such as farmers, farm veterinarians and butchers, may have a higher risk for VRE infection or colonization. The findings on the relationship between the clinical strains to isolates of animal origin have been inconsistent (Stobberingh et al. 1999; Kolar et al. 2005; Seo et al. 2005; Jung et al. 2006; Hammerum et al. 2010; Freitas et al. 2011) with some studies suggesting that VRE are host specific (Willems et al. 2000; Homan et al. 2002). Consequently, the direct public health and clinical impact of VRE in animals to human health is unclear. Furthermore, the occurrence of VRE among the livestock in a country is not concordant with outbreaks of VRE in humans. For example, in the USA, VRE are not prevalent in food animals and animal products, even though clinical VRE infections in humans occur at a higher rate compared to that of European countries where VRE are more commonly found among livestock and healthy humans (Goossens 1998; Tacconelli and Cataldo 2008). At present, the risk factors consistently identified to be associated with the colonization or infection of the organism in humans are prolonged antibiotic usage, surgery and hospitalization because of various illnesses (Tornieporth et al. 1996; Safdar and Maki 2002).

The resistance of Enterococcus to vancomycin is encoded by mobile genetic elements that allow vancomycin resistance to spread clonally or laterally (Werner et al. 2008). Several resistance phenotypes have been identified (Lemcke and Bülte 2000); however, vanA and vanB are the most important acquired resistance genes for Enterococcus. Nosocomial strains are reported to have additional virulence traits such as gelE (gelatinase gene), cylA (haemolytic cytolysin gene) and esp (surface adhesin gene) (Willems and Bonten 2007; Pangallo et al. 2008).

The epidemiology of VRE colonization among individuals working in close proximity with animals has not received much attention from researchers. In Malaysia, VRE, including those possessing the vanA gene, have been detected in meat animals and their products (Son et al. 2002; Getachew et al. 2009, 2010), and although rare, community-acquired and nosocomial VRE infections have been reported (Raja et al. 2005; Pahang 2006). We therefore conducted this study to determine the prevalence and risk factors of VRE in two groups with high levels of contact with pigs and poultry, the two commonly suggested reservoirs for VRE. In addition, we also sampled veterinary students who have a lower level of contact with various animals as the baseline comparison. We then characterized the isolates obtained to establish the level of vancomycin resistance, the frequency of resistance and virulence genes and the molecular characteristics using random amplified polymorphic DNA (RAPD) method.

Material and methods

Study design and population

We performed a series of cross-sectional studies between 2007 and 2009 in several groups that included veterinary students, poultry and pig workers. Sample sizes were calculated based on the expected prevalence rate of 6% (Eisner et al. 2005) at 95% confidence level using the OpenEpi (Dean et al. 2003) program. The program calculated that a minimum of 80 individuals were required from each subpopulation to enable a precise estimation of VRE prevalence. In addition, opportunistic sampling of chickens and pigs where workers were located was performed whenever permissible.

Sample and data collection

We obtained approval from the Ethics Committee at the Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, for both sampling procedures and questionnaire administration for the study [UPM/FPSK/PADS/T7-MJKETIKApER/F01 (VET_OCT(08)16)]. Participation into the study was voluntary and based on the willingness of farms and abattoir managers to allow the study to be conducted. Consent was also obtained from all study subjects prior to interview and sample collection.

A sterile commercial rectal swab (Meus, Kima, Italy) was provided for each participant with instructions on self-sampling. Each swab was then labelled by specific code to maintain anonymity. From chickens and pigs, cloacal and rectal swabs, respectively, were collected from a minimum of 26 animals from each location where permissible. The swabs were transported on the same day to the Bacteriology Laboratory at the Faculty of Veterinary Medicine, Universiti Putra Malaysia, Serdang, Malaysia, for bacteriological analysis.

Questionnaires were distributed to the participants to complete and were collected at the end of the sampling session. Putative risk factors such as occupation (student, poultry workers, pig workers), nationality (local, foreigner), ethnicity (Malay, Chinese, Indian, others), age (<40 years old, >40 years old), sex (male, female), diet (vegetarian, nonvegetarian), meat preference (chicken, beef, pork) and medical history for the previous 3 months (hospitalization of >72 h, surgery, antibiotic medication) and level of direct contact with livestock in the past 3 months (≤1 day week−1 – low, 2–3 days week−1 – moderate, ≥4 days week−1 – high) and pet ownership (dog, cat, both or none) were collected using a questionnaire. Those with difficulties understanding English and Bahasa Malaysia (the national language) were assisted by translators.

Laboratory analysis

VRE screening and isolation

The VRE isolation method was performed as described by the Malaysia Regional Veterinary Laboratory Test Method document (TMBAct-3, 2004) with slight modification. Briefly, the swabs were immersed individually in the brain heart infusion (BHI) broth (Pronadisa™ Laboratorios, Conda, Spain) and incubated overnight (12–18 h) at 37°C aerobically (Klein et al. 1998). A loopful of the cultured BHI broth was then inoculated onto Slanetz and Bartley agar (Merck Inc., Darmstadt, Germany), supplemented with 8 μg ml−1 vancomycin (Sigma, St Louis, MO, USA). All plates were incubated aerobically for 48 h at 37°C (Domig et al. 2003). From each plate, all red-maroon colonies were selected and transferred to VRE agar (Oxoid, Basingstoke, UK) supplemented with 8 μg ml−1 vancomycin for confirmation. Representative colonies were purified from the agar on BHI agar (Difco, Detroit, MI, USA) for DNA extraction.

DNA extraction

Fresh colonies of presumptive VRE cultures were grown on BHI agar. DNease® Blood and Tissue DNA extraction kit (Qiagen®, Hilden, Germany) was used to extract genomic DNA according to the protocol described for Gram-positive bacteria by the manufacturer. Extracted DNA was stored at −20°C and used for all molecular procedures in this study.

Genus confirmation, species identification and resistance gene detection

Multiplex PCR (M-PCR) was used for simultaneous detection of Enterococcus genus, species (Enterococcus faecium, Enterococcus faecalis, Enterococcus gallinarum, Enterococcus casseliflavus) and van genes (vanA and vanB). Eight primers (Table 1) as published by several authors (Clark et al. 1993; Dutka-Malen et al. 1995; Cheng et al. 1997; Satake et al. 1997; Ke et al. 1999; Elsayed et al. 2001) were synthesized in AitBiotech, Singapore. Standard controls Ent. faecalis (ATCC 51299), Ent. faecium strain (ATCC 51559), Ent. gallinarum (ATCC 49573) and Ent. casseliflavus (ATCC 25788) were used to standardize the PCR conditions. The M-PCR was carried out in a 25-μl reaction volume that consisted of 1× PCR buffer, 0·2 mmol l−1 each dNTP, 1U HotStar Taq plus DNA polymerase and 5× Q-Solution. All reagents were supplied by Qiagen. The sequences of both forward and reverse primers used in the PCR are given in Table 1. Amplification was carried out using a MyCycler®thermal cycler (Bio-Rad, Hercules, CA, USA) with initial denaturation of 94°C for 5 min, followed by 30 cycles of amplification consisting of three steps: denaturation at 94°C for 1 min, annealing at 54°C for 1 min and extension at 72°C for 1 min. Final extension step was at 72°C for 10 min. PCR products were checked by gel electrophoresis on 2% agarose gel using power supplied at 5 V cm−1 for 90 min. Gels were stained with 3× GelRed™ Nucleic Acid Gel Stain (Biotium, Hayward, CA, USA), and gel photographs were taken under UV by AlphaImager® (Proteinsimple, San Jose, CA, USA). Analysis was made by visual inspection for the presence of expected band size (Table 1) and by comparing with standard controls and 100-bp DNA marker (New England Biolabs, Herts, UK).

Table 1. Primers used for Enterococcus genus and species identification and van gene detection
Primer specificityPrimer pair sequences (forward and reverse)Size (bp)Reference
  1. Desired/working concentration of each primer in 25 μl of reaction volume.

Enterococcus genus (Ent)5′-TACTGACAAACCATTCATGATG-3′112Ke et al. (1999)
5′-AACTTCGTCACCAACGCGAAC-3′
Enterococcus gallinarum (vanC1)5′-GGTATCAAGGAAACCTC-3′822Dutka-Malen et al. (1995)
5′-CTTCCGCCATCATAGCT-3′
Enterococcus casseliflavus/ Enterococcus flavescens (vanC2/C3)5′-CGGGGAAGATGGCAGTAT-3′484Satake et al. (1997)
5′-CGCAGGGACGGTGATTTT-3′
Enterococcus faecalis (ddlE. faecalis)5′-ATCAAGTACAGTTAGTCTTTATTAG-3′941Dutka-Malen et al. (1995)
5′-ACGATTCAAAGCTAACTGAATCAGT-3′
Enterococcus faecium (ddlE. facium)5′-TTGAGGCAGACCAGATTGACG-3′658Cheng et al. (1997)
5′-TATGACAGCGACTCCGATTCC-3′
Resistance gene VanA5′-CATGAATAGAATAAAAGTTGCAATA-3′1030Clark et al. (1993)
5′-CCCCTTTAACGCTAATACGATCAA-3′
Resistance gene VanB5′-AAGCTATGCAAGAAGCCATG-3′536Elsayed et al. (2001)
5′-CCGACAATCAAATCATCCTC-3′
PCR internal control 16S rDNA (rrs)5′-GGATTAGATACCCTGGTAGTCC-3′320Van de Klundert and Vliegenthart (1993)
5′-TCGTTGCGGGACTTAACCCAAC-3′

Determination of vancomycin susceptibility pattern

The susceptibility pattern for VRE isolates was determined by using a standard E-test procedure (AB Biodisk, Solna, Sweden). For the control, Ent. faecalis ATCC 29212 and Ent. faecium ATCC 51559 were used. The breakpoints used for the interpretive criteria for the level of resistance were based on the Clinical and Laboratory Standards Institute (CLSI) performance standard (CLSI 2006), and accordingly, the isolates were categorized as either intermediate (MIC: 8–16 μg ml−1), resistant (MIC: 32–128 μg ml−1) or highly resistant (MIC ≥ 256 μg ml−1) to vancomycin.

Detection of virulence genes

Virulence cylA, esp, gelE genes were detected using the protocols and primers published by Pangallo et al. (2008).

RAPD analysis

RAPD analysis was carried out as previously described (Martin et al. 2005). FPQuest DNA fingerprinting software (Bio-Rad Laboratory Inc) was used to analyse the RAPD-PCR electrophoretic profiles. Dice correlation coefficients were calculated by the unweighted pair group method with arithmetic mean (upmga). Isolates with a value of coefficient >0·7 were considered to be closely related.

Data analysis

The significance of the association between VRE colonization and each hypothesized risk factor was initially evaluated using univariate analysis Pearson's chi-square or the Fisher's exact tests, when appropriate. Multiple logistic regression analysis was performed to evaluate the variables associated with VRE colonization. VRE status (positive/negative) was considered as the dependent variable. The independent variables tested were those that were biologically important and those variables that have P < 0·25 in the univariate analysis. A backward stepwise procedure was performed where the analysis began with a full and saturated model and variables were eliminated from the model in an iterative process. The Wald test was used for statistical significance of inclusion or elimination of each coefficient (β) to ensure that the predictor was useful to the model. The variables were removed using the program default limit until a final model was reached. The fitness of the final model was assessed at each step using the Hosmer–Lemeshow goodness-of-fit test (P > 0·05). Odds ratios and 95% confidence intervals were generated to estimate the probability of VRE status between the variables of interest. All tests were two-sided and conducted at the level of significance α = 0·05. Data obtained were managed and analysed using IBM® Spss® Statistical software ver. 19 (Chicago, IL).

Results

Descriptive analysis of sampled populations

Veterinary students

Students (n = 92) from the veterinary college of one public university in the state of Selangor volunteered to be in the study. All participants were local students comprising ethnic Malaysian Malays, Chinese and Indians. Four students (4·3%; 95% CI: 1·4–10·1) tested positive for VRE, of which three were from preclinical years. All students had on average less than 1 day per week of direct contact with livestock in the 3 months prior to the study. Detailed characteristics of this population are shown in Table 2.

Table 2. Univariate analysis of variables in each subpopulation sampled (students, poultry and pig workers) and their vancomycin-resistant Enterococci (VRE) colonization status
VariableVet. students (n = 92)Poultry workers (n = 111)Pig workers (n = 95)Overall
VRE+VRE−VRE+VRE−VRE+VRE−VRE +VRE−P-value
Nationality
Local (n = 202)4 (4·3)88 (95·7)7 (24·1)22 (75·9)9 (11·1)72 (88·9)20 (9·9)182 (90·1)0·83
Foreigner (n = 96)008 (9·8)74 (90·2)0 (6·2)14 (93·8)8 (8·3)88 (91·7)
Ethnicity
Malay (n = 65)5 (5·1)37 (94·9)4 (15·4)22 (84·6)2 (7·1)26 (92·6)6 (9·2)59 (90·8)0·04
Chinese (n = 73)2 (4·4)43 (95·6)007 (15·9)37 (84·1)4 (5·5)69 (94·5)
Indian (n = 51)0 (0)4 (100)3 (100)00 (0)23 (100)10 (9·6)41 (80·4)
Others (n = 109)0 (0)4 (100)8 (9·8)74 (90·2)  8 (7·3)101 (92·7)
Age
19–40 (n = 213)4 (4·3)88 (95·4)6 (8·3)66 (91·7)1 (2·1)48 (98·0)11 (5·3)202 (94·8)0·001
>40 (n = 85)009 (22·0)30 (78·0)8 (16·7)38 (82·6)17 (20·0)68 (80·0)
Sex
Male (n = 204)1 (3·6)27 (96·4)13 (14·0)78 (86·0)9 (10·5)76 (89·5)23 (9·5)181 (88·7)0·27
Female (n = 79)3 (4·3)61 (95·7)2 (40·0)3 (60·0)0 (8·4)10 (100)5 (6·3)74 (93·7)
Vegetarian
Yes (n = 4)0 (0)1 (100)000 (0)3 (100)04 (100)
No (n = 262)4 (4·4)87 (95·6)8 (10·1)71 (89·9)9 (9·8)83 (90·2)21 (8·0)241 (92·0)
Consume pork
Yes (n = 138)2 (4·5)42 (95·5)008 (8·5)86 (91·5)10 (7·2)128 (92·8)0·32
No (n = 160)2 (4·2)46 (95·8)15 (13·5)96 (86·5)1 (100)0 (0)18 (11·2)142 (88·8)
Hospitalization
Yes (n = 39)3 (12)22 (88)2 (33·3)4 (66·7)1 (12·5)7 (87·5)6 (15·4)33 (84·6)0·04
No (n = 225)1 (1·5)66 (98·5)5 (7·0)66 (93·0)8 (9·2)79 (90·8)14 (6·2)211 (93·8)
Antibiotic therapy
Yes (n = 55)0 (0)23 (100)1 (12·5)7 (87·5)2 (8·3)22 (91·7)3 (5·5)52 (94·5)0·46
No (n = 196)4 (6·6)57 (93·4)6 (8·8)62 (91·2)6 (9·0)61 (91·0)16 (8·2)180 (91·8)
Surgery
Yes (n = 13)0 (0)2 (100)003 (27·3)8 (72·7)3 (23·1)10 (76·9)0·06
No (n = 250)4 (4·4)86 (95·6)7 (9·1)70 (90·9)6 (7·2)77 (92·8)17 (6·8)233 (93·2)
Own pet
Yes (n = 120)4 (6·3)59 (93·7)4 (14·8)23 (85·2)4 (13·3)26 (86·7)12 (10·0)108 (90·0)0·24
No (n = 143)029 (100)3 (6·0)47 (94·0)5 (7·8)59 (92·2)8 (5·6)135 (94·4)

Poultry farm

Workers (n = 111) from 11 chicken farms in Selangor and Johor states were sampled between November 2008 and October 2009. The majority of farm workers were foreigners (50% from Selangor and 98% from Johor). Overall, there were 82 (73·8%) foreign workers including those from Indonesia (60%), Bangladesh (27·5%), Myanmar (7·5%), Thailand (2·5%) and Nepal (2·5%). Population composition details of the sample are presented in Table 2. At the time of sampling, the locals had a mean work experience of 9·5 years (range, 1–19; SD, 7·8), and the foreign workers had a mean work experience of 1·6 years (range, 0·6–11; SD, 1·5). All workers had a high level of contact (≥4 days week−1) with chickens in the 3 months prior to the study. Fifteen workers (13·5%; 95% CI: 8·1–20·9) tested positive for VRE. A total of 236 3 to 4 week-old chickens tested negative for VRE.

Pig abattoir

Workers (n = 95) from pig abattoirs from three states (Selangor, Perak and Johor) were recruited for this study. Foreign nationals from Nepal (n = 6) and Vietnam (n = 8) made up 15% of the total sampled population (Table 2). All workers had a high level of contact with pigs in the 3 months prior to the study. Nine workers (9·5%; 95% CI: 4·6–17·0) tested positive for VRE. Opportunistic sampling from 54 finisher pigs at the abattoir revealed that all pigs sampled tested negative for VRE.

Risk factors for VRE colonization

VRE were detected in 28 of 298 (9·39%, 95% CI: 6·46–13·12) of the sampled population. Putative risk factors were initially evaluated for their bivariate association with VRE colonization (positive or negative) (Table 2). Based on the results of the univariate analysis (Table 2), age and previous hospitalization were offered to the logistic regression. The analysis identified age of >40 (OR: 5·39, 95% CI: 1·98–14·61) and previous hospitalization (OR: 4·06, 95% CI: 1·33–12·35) (Table 3) as significant risk factors of VRE colonization among these populations.

Table 3. Binary logistic regression analysis of age and hospitalization history for colonization with vancomycin-resistant Enterococci
FactorβSEWalddfP-valueOROR 95% CI
LowerUpper
  1. SE, standard error; OR, odds ratio.

Hospitalized1·400·576·0710·0144·061·3312·35
Age1·680·5110·9610·0015·391·9814·61

Identification of VRE isolates

M-PCR (Fig. 1) identified VRE isolates such as Ent.faecium (n = 12; 43%), Ent. faecalis (n = 9; 32%) and Ent. gallinarum (n = 7; 25%) in students, poultry and pig workers. A higher rate of Ent. faecium isolation was observed in poultry and pig workers, while the other two VRE species were observed at a similar rate in all groups (Table 4).

Figure 1.

Agarose gel photograph of M-PCR products. Simultaneous detection of genus-specific (Ent), Enterococcus faecium (FM), Enterococcus faecalis (FS), VanA, VanB, vanC1 and vanC2/3 with internal control (rrs). M: 100-bp DNA marker, L1: Ent. faecium (ATCC 51559), L2: Ent. faecalis (ATCC 51299), L4: Enterococcus gallinarum (ATCC 49573), L4: Enterococcus casseliflavus (ATCC 25788), L5: Staphylococcus aureus and L6: water.

Table 4. Vancomycin resistance gene and virulence genes detected from vancomycin resistance Enterococci (VRE) species isolated from veterinary students, poultry and pig workers
 VRE speciesTotal
Enterococcus faecium (n = 12)Enterococcus faecalis (n = 9)Enterococcus gallinarum (n = 7)
  1. gelE, gelatinase gene; cylA, haemolytic cytolysin gene; esp, surface adhesin gene.

VanA
Students1102
Poultry workers72211
Pig workers4307
VanB
Poultry workers0101
VanC1
Students0011
Poultry workers0022
Pig workers0022
Not detected
Students 1-1
Poultry workers 1-1
Virulence genes
cylA 240 
esp 720 
gelA 270 

Vancomycin resistance genes and vancomycin susceptibility pattern

VanA was detected in all Ent. faecium (n = 12), 75% of Ent. faecalis (n = 6) and 28·6% of Ent. gallinarum (n = 2), VanB from 16·7% of Ent. faecalis (n = 1) and vanC1 from 71·4% of Ent. gallinarum (n = 5). VanB had not been reported in Malaysia previously, and in this study, the gene was isolated from a worker from Bangladesh. No resistance gene was detected in two of the Ent. faecalis isolates (Table 4).

All isolates possessing vanA were highly resistant to vancomycin (MIC > 64). A higher resistance to vancomycin MIC > 256 μg ml−1 was observed in all Ent. faecium isolates. Ent. faecalis isolates had MIC between 12 and 256 μg ml−1, and Ent. gallinarum isolates had MIC ranging between 8 and 64 μg ml−1. According to CLSI guidelines, one Ent. faecalis and three Ent. gallinarum were categorized as intermediately resistant to vancomycin.

Detection of virulence genes

All Ent. faecium (n = 12), Ent. faecalis (n = 9) and Ent. gallinarum (n = 7) isolates were screened for virulence genes cylA, esp, gelE, and identification of the genes was confirmed by the presense of the expected band size. Overall, 58·3% of Ent. faecium and 77·7% of Ent. faecalis isolates (7/12 and 7/9, respectively) carried at least one virulence gene. Among Ent. faecium, esp gene was the most abundant (7/12, 58·3%), while gelE and cylA genes were detected in 16·7% (2/12) of the isolates each. Two of the Ent. faecium isolates carried all three virulence genes. Among Ent. faecalis, gelE gene was detected in 7 (77·7%) isolates followed by cylA (4/9, 44·4%) and esp (2/9, 22·2%). No virulence gene was detected in the Ent. gallinarum isolates (Table 4).

RAPD analysis

RAPD profiles of Enterococcus faecium, Enterococcus faecalis and Enterococcus gallinarum isolates

The electrophoretic patterns suggested that most of the Ent. faecium isolates were similar, while a few were identical. Four RAPD types (Fig. 2a) were identified at 70% coefficient of similarity. Strains HP45, HP34 and ST415 in RAPD type I and AB52 and AB46 in RAPD type III were identical. Enterococcus faecalis had a higher level of genotypic dissimilarities and formed five RAPD types (Fig. 2b). All Ent. faecalis from pig workers were RAPD type III, and isolates from two poultry worker were type II. Interestingly, two isolates with no resistance genes (ST206 and JB9) diverged from the others. Enterococcus gallinarum isolates were genetically more diverse whereby similarities (80%) were observed only between vanA Ent. gallinarum isolated from poultry workers (JB30 and JB39) (Fig. 2c). Overall, except for isolates from the veterinary students, VRE strains (especially for Ent. faecium and Ent. faecalis) tend to loosely cluster within the individual's epidemiological backgrounds.

Figure 2.

Dendrogram from UPGMA cluster analysis of random amplified polymorphic DNA profiles of Enterococcus faecium (a) (n = 11), Enterococcus faecalis (b) (n = 9) and Enterococcus gallinarum (c) (n = 7) isolates. AB/JB, pig abattoir worker isolates; HP, poultry worker isolates; ST, student isolates.

Discussion

The documented prevalence or colonization rate of VRE among humans (hospital-based populations or healthy communities) varies widely between countries and localities. Most of the studies among humans were performed on hospital-based populations and therefore may not necessarily represent healthy communities within the same localities. In work where hospitalized patients and healthy communities were sampled, the colonization rate of VRE in general was higher among the former group (Bates 1997; Wendt et al. 1999; Gambarotto et al. 2000), although the rates differ considerably between countries. The highly variable rates observed among countries indicate the possibility of a confounding effect of the environmental and other factors not accounted for in these studies. However, based on published data on healthy populations and communities, our finding of a 9·4% (95% CI: 6·46–13·12) prevalence rate among the healthy livestock-affiliated population is consistent with published reports on healthy populations in many countries such as Germany (Klare et al. 1995), Austria (Eisner et al. 2005) and France, whereby between 6–13% of populations were VRE-colonized. However, the prevalence rate in our finding is higher when compared to those reported in some parts of the Netherlands (Endtz et al. 1997), USA (Coque et al. 1996; D'Agata et al. 2001), Korea (Song et al. 2009), United Kingdom (Jordens et al. 1994), Czech Republic (Kolar et al. 2005) and Saudi Arabia (Qadri and Postle 1996). The highest documented VRE colonization rates occur in Palestine (Hijazi et al. 2009) and Belgium (Van Der Auwera et al. 1996) whereby 39% and 28% of healthy individuals, respectively, were colonized by VRE – a significantly higher proportion than others documented. In most of these studies, the status of the individuals sampled in relation to their level of contact with animals or livestock was unknown or unmeasured. Therefore, the potential effect of the contact with the proposed reservoirs if it exists cannot be asserted. Among the few studies that attempted to establish the prevalence and epidemiology of VRE among communities in close contact with livestock such as the studies conducted in Norway, the Netherlands and France, the VRE prevalence rates among poultry and turkey farmers were 37% and 28% (Van den Bogaard et al. 1997; Sørum et al. 2006), respectively, and the VRE prevalence rate among communities living in cattle-rearing areas was 11·8% (Gambarotto et al. 2000). In our study, VRE were detected in 13·5% (95% CI: 8·1–20·9) of the poultry workers, which appeared to be lower than those reported in the two former studies. The prevalence of VRE among individuals working with pigs in the current study is similar to that reported in a study in France where the VRE prevalence rate among pig farmers was 8·5%. The same study also found no statistical difference between the prevalence of VRE in pig farmers and that in nonpig farmers (12·5%) (Aubry-Damon et al. 2004). In our study, although the lowest rate of VRE was observed among veterinary students, who in general have less contact with livestock and their environments, the rate was not significantly different. Interestingly, all pigs and chickens sampled at the site of study tested negative for VRE and thus did not correlate with the findings among the workers sampled. However, as sampling was only carried out once, it is possible that the workers had been exposed to the organism from previous batches of animals as our previous studies indicated that VRE are present in pigs and chickens in Malaysia (Getachew et al. 2009, 2010). Alternatively, it is also possible that the workers may have been exposed to other environmental or community sources not investigated in this study. In general, the rates among the individuals affiliated with the proposed reservoir animals in our study were similar to rates reported among those in healthy communities elsewhere. Similar levels of VRE among communities with and without close contact with livestock have also been observed elsewhere (Aubry-Damon et al. 2004). Given the moderate level of VRE colonization among these populations in Malaysia, one would expect more occurrence of clinical VRE infection. However, community- and hospital-based VRE infections in Malaysia have remained uncommon (Raja et al. 2005; Pahang 2006), comparable to many countries where VRE can be detected at a higher level in the community but VRE infection occurs infrequently (Goossens 1998).

In this study, foreign nationals from South Asian countries encompassed about 47% of the labour force in the poultry farms and pig slaughterhouses. Among the workers, VRE were detected in only eight (8·3%) individuals. We found that foreign workers and the locals had an equal opportunity of being colonized by VRE. Likewise, among the locals, ethnicity/race was not significantly associated with VRE when other factors were controlled for. Sex also did not appear to play a role in the VRE colonization in contrast to the significant association reported by Yang et al. (2007) among hospitalized patients. We found that hospitalization and age significantly contributed to the overall prevalence of VRE when other variables were controlled for. Participants aged >40 were five times more likely to be colonized by VRE as compared to younger participants, consistent with the horizontal nature of VRE transmission (Stobberingh et al. 1999; Bonten et al. 2001), which over time increases the likelihood of an individual to be exposed to this organism. The significance of age as a risk factor for the colonization of VRE has been consistently reported in many hospital-based population studies (Tornieporth et al. 1996; Wendt et al. 1999; Webb et al. 2001; Safdar and Maki 2002). The history of hospitalization found to increase the risk of VRE among participants in our study is comparable to many others (Tornieporth et al. 1996; Weinstein et al. 1996; Wendt et al. 1999) and supports the notion that the colonization of VRE may be linked to hospital contamination by which individuals may then become healthy carriers in the community (Bonten et al. 2001).

Enterococcus faecium is the most common enterococci species causing nosocomial outbreaks and sporadic cases of VRE infection; however, this is probably because it is also the most dominant VRE species found in humans (Franz et al. 1999; Wendt et al. 1999; Kühn et al. 2005). Our findings on the dominance of Ent. faecium are comparable to a few others studying populations in close contact with livestock (Van den Bogaard et al. 1997; Sørum et al. 2006; Freitas et al. 2011). Vancomycin-resistant vanA gene was detected in all Ent. faecium, six (66·7%) Ent. faecalis and two (28·6%) Ent. gallinarum isolates, and accordingly, all vanA-positive isolates were resistant to vancomycin. In particular, 10 of the 12 Ent. faecium were highly resistant (MIC > 256 μg ml−1), while Ent. faecalis isolates had a wide range of MIC, of which two isolates without resistance gene were tested as ‘intermediate’ and ‘resistant’ to vancomycin. Enterococcus faecium has consistently been found to be highly resistant to vancomycin, and most nosocomial strains were categorized as highly resistant (Van Der Auwera et al. 1996; Goossens 1998; Gambarotto et al. 2000). Elsewhere, vancomycin-resistant Ent. faecium and Ent. faecalis without vanA or vanB resistance genes had also been observed in humans (Cuellar-Rodríguez et al. 2007; Malhotra-Kumar et al. 2008), and we believe these isolates may possess other less common resistance genes (vanD, vanE, vanG and vanL) not included in our PCR protocol.

Resistance gene vanB, which is common in many parts of the world (Coque et al. 1996; Bell et al. 1998; Yang et al. 2007) but has not been reported in Malaysia, was detected in one Ent. faecalis isolate (3·5%) from a foreign poultry farm worker. This finding emphasizes the possibility of introducing a new bacterial resistant strain into a population (Kumarasamy et al. 2010). VanA- and vanB-containing enterococci, especially the primarily multiresistant Ent. faecium strain, have a much higher clinical impact than vanC VRE, such as Ent. casseliflavus or Ent. gallinarum. Enterococcus faecalis and Ent. faecium are the predominant nosocomial Enterococcus species (Tacconelli and Cataldo 2008; Werner et al. 2008), and although rare, Ent. gallinarum can cause clinical infections (Camargo et al. 2004). In the present study, three Ent. gallinarum/vanC1 showed intermediate resistance to vancomycin, but two of the Ent. gallinarum isolates carried vanA gene and therefore were resistant. The emergence of multiple antimicrobial resistances, including high-level resistance to glycopeptides among Ent. gallinarum, highlighted the need to increase awareness of the potential of this organism as a reservoir of transmissible resistance genes to other organisms in the environment.

We screened our isolates for three virulence determinants (esp, cylA and gelE genes) and found more than half of the Ent. faecium and Ent. faecalis isolates possessed at least one of the virulence traits. Among the Ent. faecium isolates, esp genes were the most common (58·3%) and two of the Ent. faecium isolates had all three virulence genes. The findings of esp genes among VRE isolates have been inconsistent, as a few authors have found that esp gene is highly prevalent among nosocomial VRE isolates (Willems et al. 2001; Woodford et al. 2001; Camargo et al. 2006; Billström et al. 2008) and not present in VRE isolates among apparently healthy populations. Others, however, have found that esp gene is common in healthy and hospitalized populations and can easily be found in community settings (Borgmann et al. 2007; Whitman et al. 2007; Worth et al. 2008). Our findings are comparable to the latter studies and support the finding that strain possessing esp gene can be detected in the general human population and could serve as a potential source of nosocomial VRE. Among Ent. faecalis isolates, gelE was most prevalent (78%), followed by cylA (44%) and esp (22%) gene. Gelatinase gene is affiliated with clinical Ent. faecalis (Bittencourt De Marques and Suzart 2004) but has also been detected in isolates from healthy individuals (Coque et al. 1995, 1996; Biavasco et al. 2007) such as was shown from our study. The findings of VRE virulence genes among healthy populations further strengthen the link between clinical and community reservoirs, leading us to believe that the presence of these virulence genes may serve to enhance the capabilities of VRE as an opportunistic organism that can cause disease in humans only when other necessary factors are present in a setting. The putative role of virulence factors for Ent. gallinarum has not previously been reported. In this study, Ent. gallinarum isolates were negative for all virulence genes, consistent with the findings of Pangallo et al. (2008) who used the same methodology. However, gelE had been detected in Ent. gallinarum in other studies, and it was suggested that the gene may have been acquired via horizontal transfer from Ent. faecalis or Ent. faecium (Biavasco et al. 2007).

Molecular characterization by RAPD method highlighted the presence of genetic diversity in VRE species. Enterococcus faecium isolates appeared to be more genotypically similar than the other two species. For both Ent. faecium and Ent. faecalis of poultry and pig workers, the molecular similarities of these isolates appeared to generally cluster within individuals of similar epidemiological backgrounds, consistent with the findings of Willems et al. (2000); however, the isolates from the students appeared to be randomly affiliated with any of these apparent clusters. We speculate that the genotypic similarities could arise from cross-contamination between individuals within the same environment or from exposure to similar sources over time.

In conclusion, VRE were detected in the healthy students and communities working with chickens and pigs at a rate that was not significantly different. The prevalence rate of VRE was comparable to the rates reported from other countries among healthy individuals and communities living in close proximity with livestock. Previous hospitalization and age were the only two risk factors significant for the colonization of VRE in these populations. VanA was found in most of the VRE isolated, and vanB was discovered in one isolate. At least one virulence gene was recovered from Ent. faecium and Ent. faecalis isolates. Our findings support the fact that those populations in close contact with livestock are not at higher risk for the colonization of VRE.

Acknowledgements

We thank the veterinarians and staff of the farms and abattoirs sampled in this study. This project was jointly funded by grant number 04/01/07/0078RU of UPM and the DVS Malaysia internal fund. We also thank Dr Steven Krauss of UPM for reading this manuscript.

Conflict of interest

There is no conflict of interest on the publication of this manuscript.

Ancillary

Advertisement