Microbial Quality of Self Serve Salad Bars

October December 1996

Report prepared by Simon Rockliff and Geoff Millard

Samples collected analysed by the staff at ACTGAL

1. OBJECTIVE

1.1 The survey was designed to determine if the product has improved since the 1995 ACT survey.

1.2 To compare the 1996 ACT survey results with reports prepared by South Australia, Western Australia and "Choice" magazine.

2. Background

2.1 Evidence of microbial cross-contamination was detected in the 1995 Self Serve Salad (SSS) Bar survey which was conducted between January and February and consisted of 21 salad samples. In August 1996, "Choice" magazine published an article dealing with the results of an Australian wide survey (61 samples). Western Australia produced a survey report in May 1992 (315 retail samples) while South Australia compiled its report in August 1996 from the results of 48 samples.

2.2 The Australia New Zealand Food Authority (ANZFA) does not have microbiological standards for Self Serve Salad Bars but other organisations have prepared and used "inhouse" microbiological guidelines including:

2.2.1 South Australia (SA) Health Commission guidelines, used in a survey of food sold from Self-Service Takeaway Salad Bars

 

Test Criteria

Satisfactory

Poor/Unsatisfactory

Total Plate Count (TPC)

<1,000,000 organisms /g

>1,000,000 organisms/g

Coliforms

<100 organisms /g

>100 organisms /g

Salmonella

Not detected

Detected

Campylobacter

Not detected

Detected

Listeria monocytogenes

Not detected

³ 100 organisms /g

Coag+staph#

Not detected

Detected

 

2.2.2 "Choice" survey Toss the Salad "Choice" Magazine guidelines

 

Test Criteria

Acceptable

Borderline

Unacceptable

Borderline Hazard

Potential Hazard

TPC

<1000000

1–10x106

>10x106

   
L. mono.

Absent in 25 g

     

Present in 25 g

B. cereus

<100

   

100–1000/g

>1000/g

E. coli

<10/g

10–100/g

>100/g

   
Type of staphylococcus that causes food poisoning


<100

   


100–1000


>1000/g

 

  • g = gram.

2.2.3 West Australia State Health Laboratory Service guidelines for Microbiological Quality of Salad in the Western Australia survey

 

Test Criteria

Good

Poor

Unacceptable

E. coli count /g

<10

10 to <100

³ 100

Total Plate Count/g

< 1 x 106

1–10 x 106

³ 10 x 106

 

3. Survey

3.1 Between September and November 1996, the ACT Government Analytical Laboratory collected 54 samples from eight different establishments. Samples were purchased as consumer items over the counter by the ACT Government Analytical Laboratory (ACTGAL) staff and analysed by the Microbiology Unit. The samples covered the range of available SSS Bar salads from supermarkets, fastfood outlets etc. The samples were assessed for overall hygiene quality by the Standard Plate Count (TPC), Escherichia coli (E. coli) and coagulase-positive Staphylococcus (Coag+staph) analyses and for specific food pathogens such as Bacillus cereus (B. cereus) Listeria monocytogenes (L. monocytogenes), thermophilic Campylobacter and Salmonella spp.

3.2 The salads consisted of eleven pasta, nine coleslaw, eight potato, four tabouli, four Italian, three rice, three spinach, two Waldorf, and one of each, Caesar salad, crunchy slaw, Greek, lettuce, noodles Alfredo, pesto lasagne, pineapple slaw, salmon, pumpkin tomato and seafood delight.

4. Results

4.1 Table 1. Range of microbiological results obtained by the 1995 ACTGAL survey

 

Test Organism Range
L. monocytogenes All samples <3#
E. coli <2–920#
Salmonella spp Not Detected in any sample*
Coag+Staph <50–300#
B. cereus <50–250#

 

# Units expressed in terms of Colony forming units (cfu) per gram.

* Organism not detected in 25 gms

4.2 Table 2. Range of microbiological results for the 1996 ACTGAL survey

 

Test Organism Range
L. monocytogenes All samples <3#
E. coli <2–350#
Salmonella spp Not Detected in any sample*
Campylobacter spp Not detected in any sample*
Coag+Staph <50–500#
B. cereus <50–3500#

 

# Units colony forming units / gm. TPC 4 samples contained spreading organisms and could not be counted. * Units expressed in terms of detected/ not detected in 25 g of sample.

4.3 Table 3. Range of Total Plate Count results for the 1996 ACTGAL survey

 

Total Plate Count Range #

Number of samples in range

<50–1000

Nil

1001–10,000

12 (22.2%)

10,001–100,000

7 (13%)

100,001–1,000,000

8 (14.8%)

1,000,001–10,000,000

18 (33.3%)

10,000,001–100,000,000

3 (5.6%)

100,000,001–1000,000,000

2 (3.7%)

Spreaders covering plate

4 (7.4%)

 

# Units expressed in terms of Colony forming units (cfu). per gram.

5. Discussion

5.1 Problems arose when attempting to compare the results of the different surveys. The "Choice" Magazine survey did not test for Campylobacter or Salmonella spp. nor did it tabulate the results for each purchased salad. The other surveys also had different criteria and testing regimes eg. Western Australia tested for E. coli and Total Plate Count only while South Australia did not test for E. coli or B. cereus.

5.2 The following comparisons could however, be made.

5.2.1 Total Plate Count (TPC) comparison

TPCs were not performed in the 1995 ACTGAL survey.

The 1996 ACTGAL survey found eighteen (33.3%) samples were in the 110 x 106 cfu/g range, three (5.6%) samples in the 10100 x106 cfu/g range and two (3.7%) samples in the 110 x108 cfu/g range for TPC. Meanwhile "Choice" found the TPC of 25 (40.9%) of its salads in the >10 x 106 cfu/g range and Western Australia found 63 (20%) samples were in the 110 x 106 and 100 (32.0%) samples in the >10 x106 cfu/g ranges. South Australia found seven (14.5%) and five (10.4%) samples were in the 110 x 106 and 10100 x106 cfu/g ranges for TPC respectively.

5.2.2 E. coli comparison ("Choice" (1) guidelines were used to assess the E.coli results).

E. coli was detected in three (14%) of the ACTGAL 1995 samples at the 10100 cfu/g (borderline) level and in one sample (4.7%) at the >100 cfu/g (unacceptable) level while the 1996 ACTGAL survey detected E. coli in four (7.4%) samples at the borderline level and one sample (1.8%) at the unacceptable level. All of the unacceptable E. coli results were from either Spinach or Tabouli salads.

Meanwhile "Choice" found 31% of its salads had 10100 cfu/g E. coli i.e. in the borderline level and 6.5% as unacceptable with Western Australia finding 16% of its salads in the borderline category of 10100 cfu/g and 7.3% of sample >100 cfu/g or unacceptable for E. coli. The South Australian survey did not test for E. coli.

5.2.3 Coagulase-positive Staphylococcus comparison ("Choice" (1) guidelines were used to assess the Coagulase-positive Staphylococcus results).

Coagulase-positive Staphylococcus was detected in one (4.7%) of the ACTGAL 1995 samples at the 100.1000 cfu/g (borderline hazard level). The 1996 survey detected Coag + Staph in one (1.8%) of the samples at the borderline hazard level.

"Choice" found one (1.6%) of its salads contained >100 cfu/g Coag + Staph while Western Australia survey did not test for Coag + Staph. South Australia did not detect Coag + Staph in any of its 48 samples.

5.2.4 B.cereus comparison ("Choice" (1) guidelines were used to assess the B.cereus results).

B. cereus was detected in two (9.5%) of the ACTGAL 1995 samples at the 10100 cfu/g (acceptable level) and two (9.5%) at the >100 cfu/g (borderline hazard). The 1996 survey detected B. cereus in five (9.2%) of the samples at the acceptable level and in five samples (9.2%) at borderline hazard.

"Choice" found 2 (3.3%) of its salads were borderline hazard (>1001000 cfu/g), whilst, Western Australia found 1 sample (4.8%) had 480 cfu/g B.cereus. The Western Australia survey tested rice containing salads only. Once again South Australia did not test for B. cereus.

5.2.5 Salmonella spp was not detected in either of the ACTGAL or South Australian surveys.

5.2.6 Campylobacter spp was not detected in either of the ACTGAL or South Australian surveys.

5.2.7 Listeria monocytogenes "Choice" and South Australian surveys detected this pathogen in 18% and 25% of their samples respectively whereas ACTGAL did not detect any Listeria monocytogenes in any of the tested salads.

6. Conclusion

6.1 Overall the results indicate that SSS in the ACT compare favourably or are better then those surveyed by other State laboratory or "Choice" magazine. This is demonstrated by the absence of L. monocytogenes in ACT salads compared with an 18% and 25% isolation rate for "Choice" magazine and South Australian surveys respectively. A comparison of the ACT 1995 and 1996 surveys showed an improvement in the ACT salads over the nineteen months.

6.2 Compared to the "Choice" survey the 1996 ACTGAL survey demonstrated ACT salads had lower levels of TPC and E. coli, while being comparable for Coagulase-positive Staphylococcus and slightly higher for B. cereus.

6.3 When compared with Western Australia, the 1996 ACT survey showed better results for E. coli but not quite as good for the TPC.

6.4 Compared to the South Australian Survey the ACT salads were slightly poorer for TPC, Coagulase-positive Staphylococcus and B. cereus criterion but better for E.coli and much better for levels of Listeria monocytogenes.

6.5 As some of the TPCs were high more care needs to be taken to ensure that organisms do not proliferate.

6.6 The presence of E. coli in the tabouli/spinach salads could indicate a need for better washing of the ingredients.

7. References

Toss the Salad, Choice magazine, August 1996 p611.

South Australia (SA) Health Commission Survey of Food sold from Self-Service Takeaway Salad Bars, August 1996 Project Coordinated and Report Complied by: Fay Jenkins, Elena Anear and Zabeta Manatakis.

West Australia State Health Laboratory Service, Microbiological Quality of Salad in the Western Australia, May 1992.