A consumer risk assessment model of Salmonella in Irish fresh pork - PowerPoint PPT Presentation

A consumer risk assessment model of Salmonella in Irish fresh pork sausages: Transport and refrigeration modules Dr. Ursula Gonzales Barron Ilias Soumpasis Dr. Grainne Redmond Prof. Francis Butler Biosystems Engineering, UCD School of Agriculture, Food Science and Veterinary Medicine University College Dublin, Ireland

Background  Between 10-30% of all cases of foodborne salmonellosis had pork and pork products incriminated as the actual source.  In Ireland, a pork product that merits attention is the fresh pork sausage for being a raw comminuted product that is widely consumed.

Background  From an Irish consumption database, a person eats on average 90 g of sausages per week  12 800 tons of sausages would be consumed each year.  A risk assessment model estimated that on average 4.0% (95% CI: 0.3 – 12.0%) of the pork cuts produced in Ireland would be contaminated with Salmonella spp.

Objective  To develop a consumer risk assessment model for estimating the exposure and risk of salmonellosis associated with Irish fresh pork sausages.  Second-order model is underpinned by: Predictive microbiology data of Salmonella in sausage  Irish data of Salmonella prevalence and numbers in fresh pork  sausages at retail Temperature profiles of the refrigerated product  Consumer surveys on transport and refrigeration conditions  An Irish consumption database. 

Methodology

Structure of the consumer model Prevalence Salmonella data ( Prev ) concentration at t 0 ( λ 0 ) in a pack Retail Temperature Growth parameters profile ( T(t) ) ( T min , µ , N max , h o ) Transport Dynamic Baranyi’s & model Refrigeration Salmonella Time ( t ) concentration ( λ R ) D-value , Ea Dynamic exponential death model Cooking Salmonella concentration ( λ C )

Initial concentration of Salmonella in fresh pork sausage at retail ( λ 0 ) MPN data of Salmonella from Mattick et al. (2002)  Pack Sausage replicate Mean (MPN/g) 1 2 3 4 5 6 from contaminated packs 1 1-1-1 (110) 2-0-1 (140) 1-1-1 (110) 1-1-1 (110) 1-1-1 (110) 2-0-1 (140) 120 2 2-0-0 (90) 2-1-0 (150) 2-0-0 (90) 2-0-0 (90) 2-0-0 (90) 2-2-0 (210) 120 Variability in initial λ 0 3 1-1-1 (110) 2-0-0 (90) 2-0-1 (140) 1-2-0 (110) 2-0-0 (90) 2-0-0 (90) 115 4 2-0-0 (90) 1-1-1 (110) 1-0-1 (70) 1-0-0 (40) 1-0-1 (70) 1-0-1 (70) 75 5 1-0-1 (70) 1-0-1 (70) 1-0-1 (70) 2-0-0 (90) 1-0-0 (40) 2-0-0 (90) 72 6 1-0-0 (40) 1-0-1 (70) 1-0-1 (70) 1-0-0 (40) 1-0-1 (70) 1-0-0 (40) 55 7 1-0-1 (70) 1-0-0 (40) 1-0-1 (70) 1-0-1 (70) 1-0-0 (40) 1-0-0 (40) 55 8 0-1-0 (30) 1-0-0 (40) 1-0-0 (40) 1-0-0 (40) 1-0-0 (40) 0-1-0 (30) 37 9 0-0-0 (<30) 0-0-0 (<30) 0-0-0 (<30) 0-0-0 (<30) 0-0-0 (<30) 0-0-0 (<30) <30 10 0-0-0 (<30) 0-0-0 (<30) 0-0-0 (<30) 0-0-0 (<30) 0-0-0 (<30) 0-0-0 (<30) <30 Posterior distribution for the uncertainty around MPN

Modelling initial concentration ( λ 0 ) For every triplet (every sausage tested), a likelihood  function l(x| λ ) or conditional probability of observing the positive tube counts X={xi} given true Salmonella concentration λ was calculated. Out of n i serial dilution analysis tubes, the numbers of  positive tubes x i are independent random variables distributed as Binomial ( n i , p i ) with, p 1 exp v / df i i

Modelling initial concentration ( λ 0 )  Likelihood function l(x| λ ) m n i x n x l x | 1 exp v / df exp v / df i i i i i x i i 1  Each triplet (sausage tested) produced a likelihood function

For instance, pack #4: l(x| λ ) 0.5 90 MPN/g Sausage 1: X={2, 0, 0} 0.4 Uncertainty 0.3 0.2 0.1 0.0 0 50 100 150 200 250 300 350 400 450 Concentration of Salmonella in fresh raw sausages (CFU/g)

For instance, pack #4: l(x| λ ) 0.5 Sausage 1: X={2, 0, 0} 0.4 Sausage 2: X={1, 1, 1} Uncertainty 0.3 0.2 0.1 0.0 0 50 100 150 200 250 300 350 400 450 Concentration of Salmonella in fresh raw sausages (CFU/g)

For instance, pack #4: l(x| λ ) 0.5 Sausage 1: X={2, 0, 0} 0.4 Sausage 2: X={1, 1, 1} Uncertainty Sausages 3,5,6: X={1, 0, 1} 0.3 0.2 0.1 0.0 0 50 100 150 200 250 300 350 400 450 Concentration of Salmonella in fresh raw sausages (CFU/g)

For instance, pack #4: l(x| λ ) 0.5 Sausage 1: X={2, 0, 0} 0.4 Sausage 2: X={1, 1, 1} Sausages 3,5,6: X={1, 0, 1} Uncertainty 0.3 Sausage 4: X={1, 0, 0} 0.2 0.1 0.0 0 50 100 150 200 250 300 350 400 450 Concentration of Salmonella in fresh raw sausages (CFU/g)

Each sausage from a pack was stuffed from same mix, 6 and assuming no clustering  f | X l x | i s 1 0.5 Sausage 1: X={2, 0, 0} Sausage 2: X={1, 1, 1} 0.4 Sausages 3,5,6: X={1, 0, 1} Sausage 4: X={1, 0, 0} Uncertainty 0.3 Uncertainty distribution f(λ|X) 0.2 0.1 0.0 0 50 100 150 200 250 300 350 400 450 Concentration of Salmonella in fresh raw sausages (CFU/g)

Modelling variability in initial concentration ( λ 0 )  Uncertainty around the 10 distributions of within-pack Salmonella concentration f j (λ|X ) (j=1…10) was propagated to a log -normal distribution using Bootstrap.

Estimating growth parameters of Salmonella in raw pork sausage Data from  Ingham et al. 1.0 (2009), where Raw bratwurst 0.9 Broth pH=5.9, Aw=0.97 Salmonella (log CFU/h) 0.8 Typhimurium, Heidelberg, 0.7 Infantis, Hadar 0.6 and Enteritidis were inoculated 0.5 R 2 =0.77 in fresh bratwurst. 0.4 Square-root or  0.3 Belehradek-type 10 15 20 25 30 35 40 45 50 equation Temperature (°C) T b T T max min

Estimating Salmonella growth Y(t) in raw sausage  Dynamic microbial growth (Y(t)) was estimated using Baranyi and Robert’s differential equation d 1 Y T t 1 exp Y Y max max dt 1 exp Q t d Q T t max dt Y 0 Y , Q 0 ln q 0 0 1 h T LP T ln 1 0 max q 0

Temperature profile (T(t)) of sausages during transport Transport time (t T ) from retail purchase to cold storage  modelled from Irish consumers’ survey t T ~ InvGaussian(36.037, 38.761)  Internal temperature of the sausage pack (T(t)) was  modelled with one-directional transient heat transfer equations with T 0 = product’s initial temperature (T at retail) = Normal(5,0.8)  h T = convective heat transfer coefficient (11.0 W/m 2 C)  α ( k/(ρcp) ) = thermal diffusivity of pork sausage (1.41x10 -7 m 2 /s)  k = thermal conductivity of pork sausage (0.48 J/m-s- C)  T ∞ = T amb + 3 C 

Temperature profile (T(t)) of sausages during refrigeration Total refrigeration time, t R ~Gamma(1.1, 15)  Experiments were conducted to capture the oscillations in  the internal temperature of a sausage pack stored in domestic refrigerators  sets of (t, T) Modelled T(t) during refrigeration mimicked exp. data  A T(t) for a sausage pack was modelled in two stages:  Temperature adjustment period  brief and governed  by heat transfer equations until approaching T avg Temperature oscillation stage  consisted of a  temperature history section (t, T) S randomly sampled from the above experiment until the completion of the total refrigeration time (t R ).

T(t) of a refrigerated sausage pack  In every iteration 12 (sausage pack) 10  T avg is sampled Temperature (°C) 8 6 T avg =4.2°C 4 2 0 0 5 10 15 20 25 Refrigeration time (h)

T(t) of a refrigerated sausage pack  In every iteration 12 I II (sausage pack) 10  T avg is sampled Temperature (°C) 8  First stage 6 T avg =4.2°C 4 2 (t I-II , T I-II ) 0 0 5 10 15 20 25 Refrigeration time (h)

T(t) of a refrigerated sausage pack  In every iteration 12 I II (sausage pack) 10  T avg is sampled Temperature (°C) 8  First stage 6  Second stage T avg =4.2°C 4  Common (t I-II , 2 T I-II ) in (t I-II , T I-II ) algorithm 0 0 5 10 15 20 25 Refrigeration time (h)

T(t) of a refrigerated sausage pack  In every iteration 12 I II (sausage pack) 10  T avg is sampled Temperature (°C) 8  First stage 6  Second stage T avg =4.2°C 4  Common (t I-II , 2 T I-II ) in (t I-II , T I-II ) algorithm 0 0 5 10 15 20 25 Refrigeration time (h)

T(t) of a refrigerated sausage pack  In every iteration 12 I II (sausage pack) 10  T avg is sampled Temperature (°C) 8  First stage 6  Second stage T avg =4.2°C 4  Common (t I-II , 2 T I-II ) in (t I-II , T I-II ) algorithm 0 0 5 10 15 20 25 Refrigeration time (h)

T(t) of a refrigerated sausage pack  In every iteration 12 I II (sausage pack) 10  T avg is sampled Temperature (°C) 8  First stage 6  Second stage T avg =4.2°C 4  Common (t I-II , 2 T I-II ) in (t I-II , T I-II ) algorithm 0 0 5 10 15 20 25 Refrigeration time (h)

Growth of Salmonella (Y(t)) under changing temperature T(t) 8 14  Broad 95% CI 7 given high 12 6 variability in b, 10 Temperature (°C) Y(t) (log CFU/g) 5 T min and h 0 8 4 6 3 4 2 2 1 0 0 0 10 20 30 40 50 60 70 80 Time after purchase (h)