Effect of carcass decontamination at pig slaughterhouses on the number of human

Salmonella cases in Denmark.

Søren Aabo

Section for Food HygieneDepartment of Microbiology and Risk Assessment

National Food InstituteThe Technical University of Denmark

Contributors:

Kristen BarfodTove Christensen Pia ChristiansenTina Bech HansenRikke KragHelle Mølgaard SommerMorten Mørkbak

Sanco workshop february 20092 DTU Food, Technical University of Denmark

Danish project:

“DECONT”: 2005- 2009

“Efficacy cost benefit and consumer perception of post harvest carcass decontamination of slaughter pigs”

• Data generation• Effect of steam, hot water and lactic acid

decontamination• Risk modelling • Cost effect modelling• Consumer perception• Willingness to pay

Sanco workshop february 20093 DTU Food, Technical University of Denmark

Background for focus on carcass decontamination of pork in Denmark:

E.C. regulation 853, article 3, 2004 opens up for physical decontamination in EU.

Political wish to approach the low Swedish Salmonella levels in pork.

Exhaustion of the current control programs.

Increased industrial interest for end-point decontamination. A simple (and relative cheap) way to improve food safety compared to herdintervention.

Experience with Hot Water Decontamination of 0,5-1 % of production.

Possible significant impact on food safety.

Could be outbreak protective.

No information on decontamination of pork available from other EU countries.

Sanco workshop february 20094 DTU Food, Technical University of Denmark

Concerns about decontamination

•Toxicity of compounds?

•May lead to poor slaughterhygiene?

•Discolouration of meat

•Excessive water consumption (Hot Water Decontamination).

•Consumer acceptance

•May not be installed in small slaughterhouses (Practical/economical reasons)

•Small slaughterhouses – significant suppliers of fresh meat in DK

•Unknown effect on human ilnesses

Sanco workshop february 20095 DTU Food, Technical University of Denmark

Development of the prevalence of Salmonella onswine carcasses 2001-2007Bacteriological fresh meat surveillance at slaughterhouses

Source: Danish Veteriunary and Foos AdministrationOpdateret 03-03-2008

0,00

0,50

1,00

1,50

2,00

Jan2001

Jul Jan2002

Jul Jan2003

Jul Jan2004

Jul Jan2005

Jul Jan2006

Jul Jan2007

Jul

Måned

% p

ositi

ve p

røve

r

% positive % positive, moving avg. for 12 month

Sanco workshop february 20096 DTU Food, Technical University of Denmark

Salmonella source attribution of 1775 registered human cases in Denmark in 2006

Ducks (0.2-1.3%)

Imported pork (0.7-2.6%)

Broilers (0.2-1.0%)

Table eggs (4.9-7.5%)

Imported chicken (7.4-11.0%)

Imported turkey (4.1-6.6%)

Imported ducks (0.3-1.2%)

Travel (mean: 25%)

Imported beef (0.7-2.1%)

Beef (0.8-2.0%)Pork(4.6-7.8%)

Outbreak, source unknown (5.9%)

Unknown (33.5-39.4)

Pork associated:107 (82-138).

Sanco workshop february 20097 DTU Food, Technical University of Denmark

Figure 1: CSLM of hair follicle with green fluorescent Yersinia enterocolitica moving into the cavity by capillary force. X10

Source: Rikke Krag, KU/ DTU

Decontamination methods investigated:

Steam ultra sound

Hot Water Treatment

Lactid acid

Bacteria hide themselves – a challenge to decontamination

Sanco workshop february 20098 DTU Food, Technical University of Denmark

Figure 2Pork skin inoculated with gfp-tagged Yersinia enterocolitica and decontaminated with Sono Steam for 2 sec. After decontamination few bacteria are identified and mainly in deeper tissue structures.

Gfp-tagged Yersiniaenterocolitica in hair follicles

Sanco workshop february 20099 DTU Food, Technical University of Denmark

Steam ultrasound - SonoSteam®

Sanco workshop february 200910 DTU Food, Technical University of Denmark

Effect of Sonosteam treatment on Salmonella Typhimurium seeded (10^7cfu/cm2) on skin and meat side of jaw samples

(n=6) ± SD

0,00

1,00

2,00

3,00

4,00

Treatment time (seconds)

log

(10)

redu

ctio

n

SkinMeat

Skin 0,57 2,05 2,33 3,21

Meat 0,44 1,23 1,52 2,03

0.5 1.0 1.5 2.0

Sanco workshop february 200911 DTU Food, Technical University of Denmark

Hot water Decontamination80˚C for 15 seconds

Almost all level III herds and all S. Typhimurium DT104 infected herds

Sanco workshop february 200912 DTU Food, Technical University of Denmark

Prevalence of Salmonella on carcasses after Hot Water Decontamination

Almost all carcasses from Salmonella level III herds and S. TyphimuriumDT104 herds are Hot Water Treated in Denmark = 0,5-1 % of production

Serological level Salmonella prevalence (%)

- HWT +HWT

I 1,76 0,04

II 3,84 0,31

III 5,07 0,11

n=30.000 n=9000

Reduction in prevalence : 40-50 fold

Sanco workshop february 200913 DTU Food, Technical University of Denmark

Risk model

Modelling carcass contamination before and after decontamination

Modelling no. salmonella bacteria per serving

Dose response modelling and adjusting to the Danish source attribution model

Sanco workshop february 200914 DTU Food, Technical University of Denmark

Amount of faeces on carcass

N= 1920

Risk modellingExposure model – salmonella on carcasses

+

Salmonella bacteria in faeces

N=1440

=

Distribution for Salm_carcass/G6

M ean = 36,16174

X <=23,45%

X <=53,4795%

0

0,01

0,02

0,03

0,04

0,05

0,06

0 40 80 120

Sanco workshop february 200915 DTU Food, Technical University of Denmark

Salmonella per serving

• Salmonella per carcass (75 kg)

• Salmonella per serving (200 g)

• Reduction factor for foodpreparation

• Salmonella per prepared serving

Sanco workshop february 200916 DTU Food, Technical University of Denmark

Beta-Poisson Dose-response

00,10,20,30,40,50,60,70,80,9

1

0 1 2 3 4 5 6 7 8 9 10log Dose

P_ill 2.5th 97.5th

Pinf = 1-(1+Dose/β)-α (α,β = 0.132, 51.5, FAO/WHO 2008)

Ducks (0.2-1.3%)

Imported pork (0.7-2.6%)

Broilers (0.2-1.0%)

Table eggs (4.9-7.5%)

Imported chicken (7.4-11.0%)

Imported turkey (4.1-6.6%)

Imported ducks (0.3-1.2%)

Travel (mean: 25%)

Imported beef (0.7-2.1%)

Beef (0.8-2.0%)Pork(4.6-7.8%)

Outbreak, source unknown (5.9%)

Unknown (33.5-39.4)

Sanco workshop february 200917 DTU Food, Technical University of Denmark

Predicting no. of cases from Relative Risk

Method Relative Risk No. of cases/year 2006

No decontamination 1 107

Hot water 80°C 5 secs 0.08 9

Hot water 80°C 15 secs 0.05 5

Hot water 80°C 15 secs 1% lactic acid

0.01 1

Hot water 55°C 2.5% lacticacid

0.12 13

Sanco workshop february 200918 DTU Food, Technical University of Denmark

Factors providing uncertaintyto the estimate

Transport Retail

Temperature abuse/growthDessication /reductionCross contaminationUnder cookingReady to eat sources

ConsumerTransport

Methodology and tissue associated effect of decontamination methods

Sanco workshop february 200919 DTU Food, Technical University of Denmark

Relative Risk for Human Salmonellosis

3,57

2,482,17

1,241,00

0,00

0,50

1,00

1,50

2,00

2,50

3,00

3,50

4,00

E B A D C

Abattoir

RR RR

Distribution for Salm_carcass/G6

M ean = 36,16174

X <=23,45%

X <=53,4795%

0

0,01

0,02

0,03

0,04

0,05

0,06

0 40 80 120

Sanco workshop february 200920 DTU Food, Technical University of Denmark

Perception of different risk reductions strategies

To solve the problem with pathogenic bacteria in meat it is OK to….

0

200400

600800

1.000

Freeze the meat

Treat with steam

Boil the meat in

water

Irrigate with hot water

Marinate meat with soy sauce

Irrigate the meat with chlorous comp...

Definitely234Not at all

Sanco workshop february 200921 DTU Food, Technical University of Denmark

ConclusionDecontamination methods available or under development

Potential significant impact on food safety, estimated to 10-100 fold decrease in human illness if all meat is treated.

Slaughter house hygiene have significant effect on human risk

Retail and consumer hygiene level influence the effect of carcass decontamination

Outbreak not included in model

Decontamination may be used for specific purposes (high risk carcasses, meat for fermented products)

Sanco workshop february 200922 DTU Food, Technical University of Denmark

Thank you for your attension

Sanco workshop february 200923 DTU Food, Technical University of Denmark

Table 2.Cost data for different decontamination technologies# Hot water

slaughtering Steam

ultrasound Lactic acid 2.5%

55 °C in 120 seconds

Carcass per year (1000) 740 740 740 Investment (1000 DKK#) 342 604 12 1000 € per year Capital cost 55 97 2 Energy and water cost 85 27 3 Labour cost 25 0 10 Other variable costs 25 27 5 Total cost 191 150 20 € per carcass Costs per carcass 0.26 0.20 0.03 # Danish Meat Institute (hot water), FORCE Technology (steam ultrasound), SFK Systems, (steam vacuum) and Spraying System (lactic acid spray)

Sanco workshop february 200924 DTU Food, Technical University of Denmark

0.0

20.0

40.0

60.0

80.0

100.0

120.0

88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06

Est

imat

ed n

umbe

r of h

uman

cas

es p

er 1

00,0

00

Broilers Pork Table eggsTotal cases

Source: D

anish Zoonoses Centre,

DTU

National Food Institute

Major foodborne human infections in Denmark 1988-2006

Sanco workshop february 200925 DTU Food, Technical University of Denmark

Prevalence of Salmonella positive carcassesfor herds in Level I, II, and III with and withoutHot water Dewcontamination (HWD)

Salmonella prevalence estimate (%)*

Level I Level II Level III

”Routine” slaughter, no HWD 1,76 3,84 5,07(Number of pools) (5130) (560) (6000**)

Sanitary slaughter with HWD 0,04 0,31 0,11(Number of pools) (919) (288) (621)*: adjusted for a sensitivity of 55% due to pooled sampling **: individual samples

Reduction in prevalence : 40-50 fold.

Sanco workshop february 200926 DTU Food, Technical University of Denmark

1,E+00

1,E+01

1,E+02

1,E+03

1,E+04

1,E+05

1,E+06

1,E+07

1,E+08

1,E+09

1,E+00 1,E+01 1,E+02 1,E+03 1,E+04 1,E+05 1,E+06

conc. E.coli in swabs (cfu/ml)

conc

. E.c

oli i

n Fa

eces

(cfu

/gr)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

< 0.07 0.07-0.7 0.7-7.5 7.5-75 75-670 > 670

no. of Salmonella/gr faeces

prob

abili

ty

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

< 0.10 0.10-0.91 0.91-10.1 > 10.1

no. of Salmonella/ml from carcass swabs

prob

abili

ty

(Cswab , Cfaeces)

cfu/ml

cfu/cm2

cfu/carcass / cfu/gr faeces

Distribution of gram faeces

per carcass

Distribution of Salmonella

on carcassX

Salmonella

per gram faeces

=

Data to validate model output

a.)

b.)

c.)

Figure 1. Schematic representation of the simulation model. From a.) paired results of the concentration of E. coli in faeces and in carcass swab samples obtained from the same animal, the faecal contamination of the carcass is estimated (i.e cfu/ml in swab ⇒ cfu on 2800 cm2 swabbed surface ⇒ cfu on total carcass surface). Combining the estimate with b.) the probability distribution of the number of salmonellas per gram faeces from infected swine, results in an estimated number of salmonellas per carcass. The simulation results are validated through carcass measurements of Salmonella contamination described in c.) a probability distribution of the number of salmonellas per ml. from carcass swab samples.

Structure of model: Patogens on carcasses

Sanco workshop february 200927 DTU Food, Technical University of Denmark

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

Slovenia

Sweden

Austria

Poland

Lithuania

Cyprus

Denmark

Latvia

Czech Republic

13 MS-group

The United Kingdom

France

Belgium

Ireland

Prevalence of Salmonella positive carcass swabs in selected EU countries

(EFSA Salmonella baseline in slaughter swine, carcass swabs, 2008)

13 MS group: 8.3% (95%CI: 6.3;11.0)

max: 20.0% (95%CI: 10.8;34.0)

Crude estimate of effect if all carcasses are treated:

0,1% 0,2% 0,4% 0,6% after Hot Water Treatment

5% 10% 15% 20% Before Hot Water Treatment

Sanco workshop february 200928 DTU Food, Technical University of Denmark

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

Slovenia

Sweden

Austria

Poland

Lithuania

Cyprus

Denmark

Latvia

Czech Republic

13 MS-group

The United Kingdom

France

Belgium

Ireland

Prevalence of Salmonella positive carcass swabs in selected EU countries

(EFSA Salmonella baseline in slaughter swine, carcass swabs, 2008)

13 MS group: 8.3% (95%CI: 6.3;11.0)

max: 20.0% (95%CI: 10.8;34.0)

Sanco workshop february 200929 DTU Food, Technical University of Denmark

Gfp-tagged Yersiniaenterocoliticaon the skin surface

Figure 1a,b and c.Pork skin inoculated with gfp-tagged Yersiniaenterocolitica and decontaminated with SonoSteam for 1 sec. After decontamination bacteria are located on skin surfaces and in deeper structures

Gfp-tagged Yersiniaenterocoliticain a hair follicle.