11 replies to this topic

[svg87]

hi everyone I have some questions, we process raw unpasteurized and pasteurized almonds. recently we got a visit from a customer, we use a salmonella detection test (Romer Labs Rapid Check) the food sample size for this test is 25g, but the customer said that the minimum amount should be 100g, so on the reports the specs should be "salmonella Negative/100g" but the test its been created just for 25g, so my question is if we have to get 4 samples of 25g from the same lot?

[Scampi]

I think you should get 4 random 25 g samples from the lot..........if you take a grab sample of 100 from 1 spot and get a negative result, it just means that 1 spot was negative and isn't indicative of the whole

 

Also, if you're lots are 1000's of kgs, a 100 gram sample is not near enough, it should be large enough to represent the whole

 

wouldn't 25 grams be 5 almonds?  Not very statistically useful I wouldn't think

[FSQA]

hi everyone I have some questions, we process raw unpasteurized and pasteurized almonds. recently we got a visit from a customer, we use a salmonella detection test (Romer Labs Rapid Check) the food sample size for this test is 25g, but the customer said that the minimum amount should be 100g, so on the reports the specs should be "salmonella Negative/100g" but the test its been created just for 25g, so my question is if we have to get 4 samples of 25g from the same lot?

 

Work with your customer how they would prefer it to be reported on your COA or if they require any different kind of product testing (which is adaptable by your lab),

 

IMO, my preferred way would be to test 4 sample and report as following on the COA:

 

Salmonella (1) : Negative  AOAC Method # ______ or 25 grams protocol

Salmonella (2) : Negative  AOAC Method # ______ or 25 grams protocol

Salmonella (3) : Negative  AOAC Method # ______ or 25 grams protocol

Salmonella (4) : Negative  AOAC Method # ______ or 25 grams protocol

 

Besides your customer request , also, look into your sampling method and the test method you are following. If this is a widely acceptable sampling method and testing in your industry. 

Some companies prefer AOAC method, however, FDA prefers the BAM method as stated:

 

https://www.fda.gov/...s/ucm070149.htm

[SQFconsultant]

I would suggest contacting the customer and asking if they want a grab of one sample at 100 or more or 4 separate at 25 or more.

[K.Oostra]

I own an ISO 17025 laboratory (Exact Scientific Services) and we test almond and peanut butter product daily.  When I see a result that says per 100 grams, that means there was 100 grams of product tested for that sample.  You would run your normal dilution (1:10) except you are now using 100 grams of sample to 900 mL of enrichment broth.  We run all almond and peanut butter on a 375 gram sample to 3375 mL of enrichment broth.  This is based on the FDA BAM Salmonella sampling procedure.  We run a composite on 15 jars per lot.  I would say at a minimum you should have 4 jars per lot composited into 1 100 gram sample for testing.  We use an AOAC-OMA method (PCR) for our analysis, with confirmation if the PCR is positive.  Almond butter can have interfering properties that will cause false positives and negatives.  If you haven't done a verification positive spike, I would do that to make sure you Romer analysis is working.  A low level spiked sample (<25 cfu/gram) would be used to verify recovery.

[Rener De Jesus]

Follow the manufacturer’s instruction, which is 25 g. But you need to do composite sampling, take many samples to make up 100g from same lot/batch. Then, test for salmonella - 25g. By the way, there’s a difference between sample size and analytical unit. In your lab, 25g for salmonella is define as the sample size or analytical unit?

[Charles.C]

hi everyone I have some questions, we process raw unpasteurized and pasteurized almonds. recently we got a visit from a customer, we use a salmonella detection test (Romer Labs Rapid Check) the food sample size for this test is 25g, but the customer said that the minimum amount should be 100g, so on the reports the specs should be "salmonella Negative/100g" but the test its been created just for 25g, so my question is if we have to get 4 samples of 25g from the same lot?

 

Hi svg

 

I "second" Post 4.

 

Nominally yr customer's wish is satisfied by analysing either a one-grab sample, (ie1x100g) or four separate samples of 25 gram each, (ie 4x25gram) or a composite of the aforementioned 4 sub-samples, ie 1x100g).

(It has been previously established that the analytical accuracy/sensitivity achieved when making separate analyses on 4, 25 gram samples as compared to analysing a composite of all 4, ie 100g is not significantly different).

 

But -

If all the lot (assuming there is a lot) were perfectly homogeneous the difference in sampling accuracy between the one-grab/4-grab options would be zero.

However in practice lots are not  homogeneous so taking several random sub-samples gives a "higher" probability of obtaining a representative sample.

[svg87]

Follow the manufacturer’s instruction, which is 25 g. But you need to do composite sampling, take many samples to make up 100g from same lot/batch. Then, test for salmonella - 25g. By the way, there’s a difference between sample size and analytical unit. In your lab, 25g for salmonella is define as the sample size or analytical unit?

we collect 600 grams for the entire lot sample and from those 600g we use 25g to test for salmonella

[svg87]

we collect 600 grams for the entire lot sample and from those 600g we use 25g to test for salmonella

 

the sample size is 600g and the analytical unit for the test is 25g

[svg87]

I forgot to mention, the representative sample size from a 44,000 lb lot is 600g but the analytical test requires 25g to perform the test. does anyone knows if theres any specification about the size from the lot and the size of the sample?

[Charles.C]

Hi svg,

 

It appears you have some confusion as to the objective of taking a larger sample.

 

i suggest you study the BAM chapters on Salmonella/Sampling as already referenced/described here to get an appreciation of the standard Procedure.

 

You might also find this older thread of interest -

 

http://www.ifsqn.com...almonella-test/

[moskito]

Hi,

 

I agree with Charles. Please look into BAM. Very important is the statistics behind that and the sampling plan to be applied.

To overcome such discussion it is better to define the requirements with n, c, m M. The refered sample size is in any case 25 g, but n varies. 

E.g. with are taking samples from chocolate continuously during loading or unloading, i.e. we are collecting 30 x oder 60 x samples of 25 g each means 750 or 1500 g per 25 to. Because for salmonella the requirement is negative (c=0), we perform one test of the whole sample. The CoA differs from supplier to supplier - neg/25 g  or neg/750g (based on sampling plan agreed on). In NO case 1 sample of 750 g is meant.

E.g. for nuts we collect 3 x 10 kg (over the whole lot) and homogenize 10 kg each and taking samples from that for analytics (micro,mycotox etc.)

 

Rgds

moskito