Zoonotic and spoilage bacteria in a meat production and a processing line in Ethiopia

Background:
In Ethiopia, little information is available on the status of food safety. It was the aim of this study to detect points of risk in meat and product contamination; to assess the prevalence of Salmonella serotypes and E. coli in meat production and handling chains and to track possible sources and transmission routes of Salmonella; to evaluate the microbiological quality of the final product using the Aerobic Plate Count (APC) and the Enterobacteriaceae count (EBC); and to perform antimicrobial susceptibility/resistance tests on the isolates.

Materials and methods:
The study was conducted from December 2011 to April 2012 in two beef production and supply lines. The first was the Addis Ababa Abattoir Enterprise (AAAE) abbatoir with its recipient city butchers’ in Addis Ababa city - “the abattoir line” - while the second was a beef processing plant located some 47 km out of Addis Ababa at Bishoftu/Debre Zeit town with its product recipient supermarkets in Addis Ababa city - “the processing plant line”. For this purpose, from the abattoir line, different samples were taken along cattle slaughter steps from the abattoir environment (n=101), animal-related materials (n=102) and from raw beef from city butchers’ (n=34). Similarly, samples were collected in the processing plant line along the processing steps from the processing environment (n=194), animal-related materials (n=118) and products from 8 selected supermarkets (n=119) . All samples were examined for
Salmonella and E. coli. In addition, samples from the processing plant line were examined for APC and EBC. Salmonella isolates were serotyped and exposed to Pulsed Field Gel Electrophoresis (PFGE) using XbaI® enzymatic genomic digestion. Possible sources and transmission routes of Salmonella serotypes were tracked based upon genetic similarity and differences of the obtained same serotype relation. All Salmonella and E. coli isolates were tested for susceptibility/resistance to antimicrobial agents of six different chemical classes of drugs and each with their respective concentrations (Oxoid) used for sensitivity testing. The drugs were β-lactams - amoxicillin (AML 25 μg), amphenicols - chloramphenicol (C 50 μg), amino glycosides - gentamycin (CN 10 μg) and neomycin (30 μg), tetracycline - oxytetracycline (OT 30 μg), polymyxins - polymyxin B (PB 300IU), and folate pathway inhibitors - trimethoprim (W 5 μg) and trimethoprim-sulfamethoxazol (SXT 1.25/23.75 μg). Within both lines, sampling occasions, locations and directions in the production, processing and supply of the final product were considered and taken into account for data management and analysis. Numerical data from spoilage bacteria counts were calculated in logarithmic function and compared using single and paired t-tests. Categorical data for prevalence and antimicrobial susceptibility/resistance were calculated using percentages and the 95% midprevalence
exact confidence interval was used to assess associations. PFGE data were analyzed using the BioNumerics® 6.6 version.

Results:
Abattoir line
Spoilage bacteria: Due to non-expected heavy contamination of the abattoir spoilage bacteria were only assessed at the processing plant line.
Salmonella: 63 isolates resulting in a 26.6% prevalence were identified in the abattoir line. Salmonella prevalence was higher for abattoir environment (36.6%) than for animal-related materials (14.7%). Prevalence at butchers’ was 32.4%, similarly to abattoir environment but not animal-related materials. Six different serotypes with higher prevalences/proportions of 11.4%/42.8% for S. Saintpaul, followed by 5.9%/22.2% for S. Muenchen were observed. In this line, a prevalence of S. Larochelle (4.6%), S. Dublin (1.7%), S. Kastrup (1.3%), S. London (0.24%) and unidentified ones (1.3%) was found. The PFGE pattern showed 1, 2 and 6 pulsotypes with 1 to 14 ratios of isolates to pulsotypes in serotypes and 1 to 7 pulsotypes with 1 to 2.5 ratios of isolates to pulsotypes in positive locations. Based on genomic similarity, transmission routes were identified for S. Saintpaul (SSaX2), S. Muenchen (SMuX1), S. Larochelle (SKLX1) and S. Dublin (SDuX1). Findings revealed similar pulsotypes over sampling occasions/batches in different sampling locations for the corresponding serotypes. Other isolates of the same serotype but of a different pulsotype were determined as possible sources and/or contaminants of the production line from other sources.
E. coli: With regard to E. coli from this line, an overall prevalence of 45.1% was observed. This prevalence essentially agreed with prevalences of the environment (40.6%), animalrelated materials at the abattoir (52.0%) and with retailed meat at butchers (38.2%). Prevalence at individual sampling locations ranged from 33.3% in tap water to 55.9% in raw beef at the abattoir with no principal differences among and between sampled locations.
Processing plant line
Parameters tested here were spoilage bacteria, Salmonella and E. coli.
Spoilage bacteria: Highest APC of 5.28±0.24 log10 cfu/cm2 in room floor swabs to lowest 3.99±0.75 log10 cfu/g in spices wwere observed. EBC ranged from high 3.19±0.55 log10 cfu/cm2 for room floors to low 2.08±0.19 log10 cfu/g in products at Supermarket-G. The paired t-test on the count difference (log10 APC – log10 EBC) within locations showed higher APC than EBC. This difference ranged from 2.75±0.65 cfu/g in products to 1.71±0.70 cfu/g in spices. Using APC, microbiological quality of products in total was found to be good in 24.4% of cases, 44.5% acceptable and 31.1% unsatisfactory. In contrast, by EBC 64.7% of products were good and 35.5% acceptable.
Salmonella: A total of 23 Salmonella isolates (5.3% prevalence) was observed in this line. Prevalence in theenvironment (5.2%) was similar to that of animal-related materials (10.2%) and final products from supermarkets (0.8%), but in comparison it was higher in animalrelated materials than in final products. Only one isolate was observed in a single supermarket, the others being negative. A total of seven serotypes with a high ratio of prevalence to proportion (1.4%/26.1%) of S. London was observed. Unidentified Salmonella strains were also obtained from this line. Possible transmission of S. London (SLoX1) was observed on the 1st sampling occasion from raw beef samples to working tables which were also positive on the 2nd sampling occasion. A similar pulsotype within each of S. Anatum (SAnX1), S. London (SLoX1) and S. Muenchen (SMuX1) from raw beef on the 1st, 2nd and 6th sampling occasions was observed, respectively showing that the contaminated meat was coming from the same sources.
E. coli: The overall prevalence along the processing plant line (46.4%) was found to be the same as the prevalence in the environment (50.5%) and in animal-related materials at the processing plant (56.8%) but was higher than in the end products at supermarkets (29.4%). Differences were not observed between and among locations, with the only exception beingprocessing plant rooms for whom a higher prevalence was recorded (75%) than in tap water (23.5%), spices (13.3%), Supermarkets-A and H (each with 20%) and Supermarket-F (16.7%). There was no difference in prevalence among and between products from all supermarkets.

Antimicrobial susceptibility/resistance:
Salmonella: Susceptibility to PB, CN, C, W and SXT ranged from 93.2% to 100% and from 30.4% to 50.4% to N for all isolates in general and for those from the abattoir line and processing plant line in particular. Susceptibility to OT was 44.2% overall but it was 31.7% and 78.3% in isolates from abattoir and processing plant lines, respectively. Resistance was high (53.5%) overall. High (65.1%) resistance to oxytetracycline was found in isolates from the abattoir line was low (21.7%) in isolates from the processing plant line. A total of 55 (63.9%) isolates was found to be resistant to at least one of the antimicrobial agents used in this study. High resistance was detected in S. Saintpaul (41.8%) and in S. Muenchen (30.9%). The proportion of Salmonella isolates tested and the corresponding resistant isolates showed no difference (p>0.05) within origin/source. Strains of higher resistance were found at the abattoir line (85.5%) than at the processing plant line (14.5%). Single-drug resistant strains were more frequent (83.6%) than two-drug resistant ones (12.7%). Multiple drug resistance (MDR) in the strains was only 3.6%. Resistance to OT was frequently observed. Distribution of strains resistant to individual drugs in both abattoir and processing plant lines was 1.8% to each of PB, CN, W and SXT, 5.4% to C, 26.6% to N and 83.6% to OT, in increasing order.
E. coli: Susceptibility of E. coli isolates to PB, C, CN and SXT ranged from 92.2% to 100%. Furthermore, their susceptibility to AML was between 70.1% and 79.5% while to OT it spanned between 49.5% and 60.0%. This was true overall for isolates from the abattoir line and the processing plant line. High frequencies of resistant strains to OT (37.5% to 43.9%) and to AML (18.5% to 26.2%) were observed in total as well as in individual isolates from abattoir and processing plant lines. A total of 143 (46.6%) isolates were found resistant to at least one of the antimicrobial agents used in this study. Higher frequencies of resistant strains were observed at the processing plant line (61.5%) than at the abattoir plant line (38.5%). Single-drug resistant isolates (55.9%) were more frequent than two-drug resistant ones (28.7%) and MDR accounted for 15%. Isolates resistant to OT were frequently observed. The total distribution of strains resistant to individual drugs both at the abattoir and at the processing plant line showed differences. The extent of resistance to each drug in increasing order was PB (1.4%), CN (2.1%), C (8.4%), SXT (16.1%), AML (45.5%) and OT (85.3%).

Discussion:
Observation of Salmonella and E. coli in all sampling locations on one or more sampling occasions in this investigation demonstrated the presence of zoonotic and spoilage bacteria, particularly with higher loads along the abattoir line. Higher Salmonella prevalence (26.6%) at the abattoir line than at the processing plant line (5.3%) identified the abattoir to be the main source of Salmonella transmission along beef production, processing and handling lines. In this study, occurrence of S. Kastrup and S. London was reported for Ethiopia for the first time. With XbaI® PFGE profiles, S. Kastrup and S. Larochelle showed indistinguishable profiles. Antimicrobial susceptibility/resistance of Salmonella and E. coli from both studied lines showed high susceptibility to PB, CN, C, SXT and W. Resistance was high to oxytetracycline.

Conclusions:
This investigation involved two beef lines to study spoilage and zoonotic bacteria and their antimicrobial susceptibility/resistance. Findings revealed prevalence of the agents and the need for improvement, regular monitoring and application of HACCP during production, processing, handling and supply of products. Furthermore, resistance of Salmonella and E. coli to oxytetracycline was high and accounted for 83.6% and 85.3%, respectively. This drug is marketed, it is popular and widely used in the veterinary sector in Ethiopia. The further use of this drug thus should be questionned in light of the high prevalence of resistant strains in the country. Future emphasis should be put tohygiene toimprove beef production line, to antimicrobial susceptibility testing, to further screening of the present isolates of E. coli strains and to the extension of this type of study to meat production lines of other food animals of the country. - Zoonose- und Verderbnisserreger in zwei Fleisch produzierenden und verarbeitenden Linien in Äthiopien (Rind) -

Hintergrund:
Für Äthiopien existieren nur wenige Informationen zur Lebensmittelsicherheit. Ziele der vorliegenden Untersuchung waren:
- Abschätzung der Salmonella- und E. coli-Prävalenz in Fleischproduktions- und –bearbeitungsketten mit Zurückverfolgung der Transferrouten von Salmonella,
- Beurteilung der mikrobiologischen Qualität des Endprodukts mit Hilfe von Gesamtkeimzahl (GKZ) und Enterobacteriaceae-Zahl (EBZ) sowie
- Resistenztests an den gewonnenen Isolaten

Material und Methoden:
Die Untersuchungen wurden von Dezember 2011 bis April 2012 in zwei Fleischproduktionsketten in Äthiopien durchgeführt. Die „Abattoir Line“ wurde aus dem städtischen Schlachthof in Addis Ababa (Addis Ababa Abattoir Entreprise (AAAE)) sowie Verkaufsstätten gebildet, die Fleisch direkt von diesem Schlachtbetrieb für den Weiterverkauf erhielten. Die „Processing Plant Line“ bestand aus dem Zerlege- und Verarbeitungsbetrieb in Bishoftu / Debre Zeit mit 3 unterschiedlichen Zulieferschlachtbetrieben sowie Supermärkten in Addis Ababa, die fertige Produkte aus dem Bearbeitungsbetrieb erhielten.
In der „Abattoir Line“ wurden entlang der Schlachtlinie an unterschiedlichen Positionen Proben genommen (n = 101) sowie Proben von Tieren bzw. den Schlachttierkörpern (n = 102). In den Verkaufsstätten wurden Fleischproben genommen (n = 34). In der „Processing Plant Line“ wurden von den einkommenden Tierkörpern (n = 118) und Umgebungsproben an unterschiedlichen Positionen im Verarbeitungsbetrieb (n = 194) genommen. In insgesamt 8 Supermärkten wurden Produktproben (Mortadella) (n = 119) entnommen.
Alle Proben wurden auf Salmonella und E. coli untersucht, die Proben der „Processing Plant Line“ wurden zusätzlich auf GKZ und EBZ geprüft. Die Salmonella-Isolate wurden serotypisiert und mit Hilfe einer Pulsfeldgelelektrophorese (XbaI) mögliche Übertragungsrouten mit Hilfe bestimmter Serotypen verfolgt. Die gewonnenen Salmonella und E. coli-Isolate wurden auf antimikrobielle Resistenzen / Empfindlichkeiten gegenüber sechs Chemotherapeutika überprüft: β-Laktame: Amoxicillin (AML 25 μg); Amphenicole: Chloramphenicol (C 50 μg); Aminoglykoside: Gentamycin (CN 10 μg), Neomycin (30 μg); Tetracycline: Oxytetracyclin (OT 30 μg); Polymyxine: Polymycin B (PB 300 IU) und Folsäureinhibitoren: Trimethoprim (W 5 μg), Trimetoprim-Sulfamethoxazol (SXT 1,25/23,75 μg).

Ergebnisse:
Abattoir Line
Salmonella: Insgesamt wurden 63 Salmonella-Isolate in der „Abattoir Line“ gewonnen, dies entsprach einer Prävalenz von 26,6 % (Umgebungsproben: 36,6 %, Tierproben: 14,7 %; Proben aus den Verkaufsstätten: 32,4 %). Sechs unterschiedliche Serotypen wurden gefunden: S. Saintpaul, S. Muenchen, S. Larochelle, S. Dublin, S. Kastrup und S. London sowie unidentifizierbare Serotypen. Mit Hilfe der PFGE konnten bei S. Saintpaul, S. Muenchen, S. Larochelle und S. Dublin jeweils ähnliche Pulstypen, die auf eine genetische Verwandtschaft der jeweiligen Serotypen hinweisen, identifiziert werden. Für diese Serotypen wurden mögliche Transferrouten gezeigt.
E. coli: Für E. coli lag die Prävalenz bei 45,1 % (Umgebungsproben: 40,6 %, Tierproben: 52 %; Metzgereien: 38,2 %).
Processing Plant Line
Hygieneparameter GKZ: Die GKZ lag zwischen 3,99 ± 0,75 log10 KbE/g (Gewürze) und 5,28 ± 0,24 log10 KbE/cm2 (Bodentupferproben). Die EBZ lag zwischen 2,08 ± 0,19 log10 KbE/g (Produktproben aus Supermarkt G) und 3,19 ± 0,55 log10 KbE/cm2 (Bodentupferproben).
Salmonella: Insgesamt wurden 23 Salmonella-Isolate in der „Processing Plant Line“ gefunden (Prävalenz: 5,3 %). Für S. London, S. Anatum und S. Muenchen konnten mögliche Transferrouten identifiziert werden.
E. coli: Für E. coli lag die Prävalenz bei 46,4 % (Umgebungsproben: 50,5 %; Tierproben: 56,8 %; Produktproben: 29,4 %).

Empfindlichkeitsprüfungen gegenüber Chemotherapeutika:
Salmonella:Die Empfindlichkeit von den aus beiden Linien gewonnenen Salmonella-Isolaten gegenüber PB, CN C, W und SXT schwankte zwischen 93,2 und 100 %; gegenüber N zwischen 30,4 und 50,4 %.
Insgesamt 55 Isolate waren gegen mindestens ein Chemotherapeutikum resistent (63,9 %).
E. coli: Von den gefundenen E. coli-Isolaten waren 92,2 bis 100 % empfindlich gegenüber PB, C, CN und SXT. Insgesamt waren 143 Isolate gegen mindestens ein Chemotherapeutikum resistent (46,6 %).
83,6 % der Salmonella-Isolate bzw. 85,3 % der E. coli-Isolate waren gegenüber Oxyterazyklin resistent.

Diskussion:
Salmonella und E. coli wurden an unterschiedlichen Probenahmetagen innerhalb beider Ketten gefunden, die Nachweisrate in der „Abattoir Line“ war höher (26,6 %) als in der Processing Plant line (5,3 %). Erstmals in Äthiopien wurden S. Kastrup und S. London nachgewiesen. Die XbaI® PFGE Profile beider Serotypen waren nicht unterscheidbar. Bezüglich der Antimikrobiellen Empfindlichkeit / Resistenz war bei Salmonella und E.coli eine hohe Empfindlichkeit gegenüber PB, CN, C, SXT und W erkennbar sowie eine hohe Resistenz gegenüber Oxytetracyclin.

Schlussfolgerungen:
Diese Untersuchung bezog sich auf zwei Prozeßlinien (Schlacht- und Verarbeitungslinie) und auf Zoonoseerreger, Verderbsmikroflora und das Resistenzgeschehen. Die Ergebnisse lassen die Notwendigkeit regelmäßiger Kontrollen, so etwa eines HACCP- Systems erkennen. Es ergab sich eine hohe Resistenz gegemnüber Oxytetracyclin in beiden Linien (83,6% und 85,3%). Oxytetracyclin ist in Äthiopien häufig in Benutzung, was angesichts der hohen Resistenzen hinterfragt werden sollte. In Bezug auf die Hygiene sollten die Abläufe verbessert werden, es sollten Resistenztests und Untersuchungen auf die nähere Natur der auftretenden E.coli durchgeführt werden. Untersuchungen wie die hier vorgelegte Studien sollten auch in anderen Nutztier- Linien durchgeführt werden.