Jump to content

  • Quick Navigation
Photo
- - - - -

Heterogeneous micro distribution in powders and dry products?


  • You cannot start a new topic
  • Please log in to reply
11 replies to this topic

EssentialFA

    Grade - Active

  • IFSQN Active
  • 20 posts
  • 0 thanks
0
Neutral

  • United States
    United States

Posted 06 January 2016 - 09:14 PM

Hi everyone,

My company manufactures a vegan seed powder that is high in protein (Hemp seed protein). It is a fairly dry product (8-10% water only, aw < 0.55)

I am trying to better understand lab results we are getting on micro during production and finished product.

 

We take a composite sample of each day of production (milling) by sampling each barrel produced and sending a sample from this mix to our lab, which runs a micro panel (APC, Yeast and mold, b. cereus, salm, S. aureus, coliforms and E. Coli, and C. perfringens)

 

We consistently get low counts for B. cereus (< 100 cfu/g), and about 30% of the time we will have a low count for c. perfringens as well (< 30 cfu/g). Which is odd because the raw material absolutely always tests negative for c perfringens.

What is a little disconcerting to me is that when we ask our lab to retest the sample for perfringens, or if we send another sample from the same batch, it will very often test negative. When dealing with low counts like that, should we expect perfect repeatability of testing?

Is it the lab that is not being consistent, or are we demanding too much?

 

We were told by our lab that powders often have very heterogenous distribution of micros, and therefore testing and getting a "negative" only means that in the 1g they pulled the count was below detection. They also told us that c. perfringens is the same as b. cereus in that it is impossible to get rid of the spores in vegetal derived products. in their mind counts of C. perfringens < 100 cfu/g are not something to worry about in dry products, as the aw would not support growth.

 

What is your take on this? Do we have a sloppy lab giving poor advice?



RMAV

    Grade - SIFSQN

  • IFSQN Senior
  • 406 posts
  • 121 thanks
41
Excellent

  • United States
    United States
  • Gender:Male
  • Location:USA - Midwest
  • Interests:QA, Micro, Sanitation;
    Meats, Juice, Condiments;
    SQF, Audit, and aviation

Posted 07 January 2016 - 05:29 PM

"We were told by our lab that powders often have very heterogenous distribution of micros, and therefore testing and getting a "negative" only means that in the 1g they pulled the count was below detection."

 

Disclaim: I'm not an expert.  I agree with this as it has been my observation.  Also, one can't help but think of the Nestle Cookie Dough recall of 2009...



Thanked by 1 Member:

EssentialFA

    Grade - Active

  • IFSQN Active
  • 20 posts
  • 0 thanks
0
Neutral

  • United States
    United States

Posted 07 January 2016 - 08:31 PM

Thank you for the reply. If I understand you well, we can't have any real certainty that he lab results we are getting will be completely representative of the entire batch... and then statistics and trend analysis would help us know if a criteria is controlled or not, am I right?

Otherwise should we consider that any result we get "negative" for would be equal to "might be containing low levels of contamination"... (not sure how much our customers would like that)

 

I am somewhat of a beginner in the micros area, and just want to be prudent.



RMAV

    Grade - SIFSQN

  • IFSQN Senior
  • 406 posts
  • 121 thanks
41
Excellent

  • United States
    United States
  • Gender:Male
  • Location:USA - Midwest
  • Interests:QA, Micro, Sanitation;
    Meats, Juice, Condiments;
    SQF, Audit, and aviation

Posted 07 January 2016 - 11:15 PM

Thank you for the reply. If I understand you well, we can't have any real certainty that he lab results we are getting will be completely representative of the entire batch... and then statistics and trend analysis would help us know if a criteria is controlled or not, am I right?

Otherwise should we consider that any result we get "negative" for would be equal to "might be containing low levels of contamination"... (not sure how much our customers would like that)

 

I am somewhat of a beginner in the micros area, and just want to be prudent.

I'm of the same opinion as you, but I do not have direct experience on your side of things - hopefully someone who does and knows will chime in.



Charles.C

    Grade - FIFSQN

  • IFSQN Moderator
  • 18,985 posts
  • 5285 thanks
1,256
Excellent

  • Earth
    Earth
  • Gender:Male
  • Interests:SF
    TV
    Movies

Posted 08 January 2016 - 02:00 PM

Hi EssentialFA,

 

I sincerely hope that yr OP does not mean that the total sample size taken for testing was 1g.

 

You might ask yr lab as to the methodology in use.

 

All micro results have errors due to sampling + analytical variations.

 

Many (most?) numerical micro. results are famous for their (accepted) inaccuracy. And often debated  interpretation. Hence the variety and ranges of  tolerance formats seen in micro. specs.

 

Substantial sample sizes are required to achieve a "reasonable" confidence level in the result where low levels are involved, eg for pathogens like Salmonella. Such limitations are detailed in all micro. textbooks. Most routine in-house micro. labs cannot logistically handle the "optimum" sample sizes.

 

Just imagine if you have 5 red balls randomly mixed with 95 black ones. Guess the probability of finding 1 red one. And then another. :smile:


Kind Regards,

 

Charles.C


EssentialFA

    Grade - Active

  • IFSQN Active
  • 20 posts
  • 0 thanks
0
Neutral

  • United States
    United States

Posted 08 January 2016 - 04:18 PM

Thanks everyone for the feedback.

I wish this were a little more "black and white", but I understand that the uncertainty has to be accepted and that I guess we can't rely on tests results alone to guarantee quality/safety of our product.

 

And for the record, we do not send 1g of sample to the lab :) we send about 300-500g (composite sample of one day of production (about 50g taken from each 300 lbs drum of bulk product))

But from that 300g the lab will only take a very small quantity (from what I understand)

 

There is a line to walk between having a testing/sampling plan that is representative of the production and gives us confidense in the result obtained, and at the same time a costing factor (if we switch so sending 5 or 10 different samples instead of making a composite one, our testing cost would skyrocket so much that it would not be possible to follow up)
 

when you add to this dealing with a somewhat uncommon product for which micros standards have not been established (as far as I can tell), it is sometimes difficult to bring together what is reasonnably feasible, what level of confidence we can have in the results and explaining all that to our customers as well...



Charles.C

    Grade - FIFSQN

  • IFSQN Moderator
  • 18,985 posts
  • 5285 thanks
1,256
Excellent

  • Earth
    Earth
  • Gender:Male
  • Interests:SF
    TV
    Movies

Posted 08 January 2016 - 04:30 PM

Hi EssentialFA,

 

To rephrase, i sincerely hope that yr lab do not take only 1g from yr provided input to them.


Kind Regards,

 

Charles.C


EssentialFA

    Grade - Active

  • IFSQN Active
  • 20 posts
  • 0 thanks
0
Neutral

  • United States
    United States

Posted 08 January 2016 - 06:42 PM

Hi Charles,

 

Sorry I misunderstood your comment the first time.

Yes our lab uses proprietary methods, and they are very vague when it comes to giving details about what they do.

We are thinking of switching to another that would conduct "standart" methods, which would be more transparent for us and for our clients.



Charles.C

    Grade - FIFSQN

  • IFSQN Moderator
  • 18,985 posts
  • 5285 thanks
1,256
Excellent

  • Earth
    Earth
  • Gender:Male
  • Interests:SF
    TV
    Movies

Posted 09 January 2016 - 02:53 AM

Hi EssentialFA,

 

Regarding yr OP, a few more comments since i missed a few lines 1st time around (sorry).

Some info. regarding the actual process may be necessary to attempt to interpret the data which you are querying (see comments below).

 

(1)

We consistently get low counts for B. cereus (< 100 cfu/g), and about 30% of the time we will have a low count for c. perfringens as well (< 30 cfu/g). Which is odd because the raw material absolutely always tests negative for c perfringens.

 

 

 I am guessing that the numbers <100cfu/g and <30cfu/g (if quoted from lab data) actually mean that the lab result was negative for the vegetative species, ie no detection in the lab-utilised sample size. It is possible that these results are not incorrect but are of limited accuracy (see discussion below).

 

However,  i am puzzled why the lab data (seemingly) does not also state <30cfu/g for the 30% of results you mention are "negative" (quoted from lab data?). Perhaps the lab. test methodologies for raw materials / fin.product are different.

 

What are the results for B.cereus in raw material like ? All stated as "< 100cfu/g" ? All stated as  "negative" ? Or ?

How high, ie maximum, do the results for C.perfringens actually go ?

 

One possible explanation for increases in the count from raw material to finished product is further generation caused during "the process" although the literature indicates that neither vegetative growth nor germination will occur at the low aw value you mention (presumably for finished product?).

 

What is the aw value of raw material ?

 

Another explanation may relate to test methodologies if there are any differences such as speculated above.

 

(2)

We were told by our lab that powders often have very heterogenous distribution of micros, and therefore testing and getting a "negative" only means that in the 1g they pulled the count was below detection

 

There may have been some misunderstanding here between lab and customer. The lab comment possibly did not mean that the product  sample size was 1g.

 

The lab method was probably via a plate count technique. This involves blending/diluting the actual sample (often 25g) in water and placing a small amount of the blend  (eg 1ml = ~1g) onto a plate for incubation followed by counting. The net result is that the 1ml often only corresponds to maybe 0.01g of product or even less. Accordingly, for example/0.01g, a nil detection would then mean that result for speciesX = <1cfu per 0.01g, ie <100/g.

(it is possible to get more precise values with other techniques but more effort/time usually needed and maybe the  spec. was speciesX not >100cfu/g)

 

(3)

they also told us that c. perfringens is the same as b. cereus in that it is impossible to get rid of the spores in vegetal derived products. in their mind counts of C. perfringens < 100 cfu/g are not something to worry about in dry products, as the aw would not support growth.

This product not my area but plant based products (and many others) commonly show some levels of vegetative form/spores of bacillus / clostridium species from the environment.  acceptable tolerances are usually defined in the product specification. The ultimate significance may depend on the subsequent use/treatment/consumption of the product.


Kind Regards,

 

Charles.C


Thanked by 1 Member:

EssentialFA

    Grade - Active

  • IFSQN Active
  • 20 posts
  • 0 thanks
0
Neutral

  • United States
    United States

Posted 11 January 2016 - 04:38 PM

Hi Charles.C,

 

Thank you so much for the further comments.

I have not expressed myself right when I was stating that the results we get are <100 cfu/g for cereus and <30 cfu/g for C.perfringens. What I meant is that for b. Cereus, we get anything between "negative" and 100 cfu/g, and perfringens comes back at "negative", 10 cfu/g, 20 cfu/g or 30 cfu/g.

The maximum count we ever saw for c. perfringens was 60 cfu/g.

 

On the raw material (raw, organic hemp seed), the results are always negative for Clostridium, and 90% of the time negative as well for b. cereus (or very low count, 10 cfu/g or 20 cfu/g max. I do not think we ever tested the aw of the seeds themselves.

 

If I read you well, the reason why the count appears higher in the finished product than in the raw material is that there was a very low level of contamination of these 2 pathogens in the raw material, and the process allowed them to reach detection levels, which would make sense somewhat, although  in the process there is some heat applied and the aw diminished.

 

As far as specifications and allowable levels of b. cereus and C. perfringens in our finished products, I have some issues setting these limits since I can't find some standard for it: some of our retail customers will use it as a protein powder (mix with liquid and consumption without further processing), some will use it in baking. For our bulk customers most will use it in a blend with (many) other ingredients as part of a protein drink powder blend, while others will use it as an ingredient for cooked/baked products.

I know that our product being used as a protein drink would allow for rehydration and could allow growth of both perfringens and cereus. However, do we (as the manufacturer) have to account for the possibility of someone mixing the drink and then letting it sit on a shelf/top of a radiator for 7 hours? That could potentially allow perfringens to reach infectious levels, but then even a product that would have tested "negative" could still have low contamination (below detection levels), but I don't think any company would advise for this to be done...

 

I have looked at other manufacturer's of the same product's spec sheets, none of them seem to worry about perfringens, and it isn't even on the spec sheet. I don't know if they don't test for it. Some others have set limits for b. cereus that seems crazy high to me (1000 cfu/g).



Charles.C

    Grade - FIFSQN

  • IFSQN Moderator
  • 18,985 posts
  • 5285 thanks
1,256
Excellent

  • Earth
    Earth
  • Gender:Male
  • Interests:SF
    TV
    Movies

Posted 11 January 2016 - 10:25 PM

Hi EssentialFA,

Thks for the data clarification.

It may depend on the method used by lab but if the initial procedure given in BAM (FDA standard) was used,  "Negative" (ie no growth on the plate) could correspond to  <10cfu/g.
Repeated values around 10-50cfu/g will be of limited accuracy but probably adequate to support compliance with a spec.like max. 100cfu/g.
 

If I read you well, the reason why the count appears higher in the finished product than in the raw material is that there was a very low level of contamination of these 2 pathogens in the raw material, and the process allowed them to reach detection levels, which would make sense somewhat, although  in the process there is some heat applied and the aw diminished.

Yes, precisely. Just to speculate, heating/slow cooling might eliminate the vegetative bacterial content but allow some spore germination if not blocked by aw. The only way to find out more may be to examine the process in detail + do some micro.testing along the process. Latter not so difficult if you have an in-house micro. lab with relevant experience but otherwise a lot of work / money.

I had a look around regarding micro.specs for your specific product / analogous products. As you say, micro. info. on protein powders seems scarce. An older forum thread with a vaguely related product is here -

http://www.ifsqn.com...nished-product/

Referring to post 16, the document mentioned does not contain yr specific product but "covers" many typical food categories. This is a standard  reference source in UK for RTE food micro. specs and is attached below. The typical limits for C.perfringens (cp) and B.cereus (bc) are shown on pg 18-19/34 however the "dried' foods category (11) on pg 26/34  includes Bacillus species but not C.perfringens as sig hazards.(the limits relate to the levels at which pathogenic activities may initiate + a safety factor. There is obviously some subjectivity in this decision.). The reason may relate to differing responses (ie increases) between cp/bc in the case of product mishandling (see below).

 

Attached File  Guidelines_for_assessing_the_microbiological_safety_of_ready-to-eat_foods_on_the_market.pdf   998.98KB   12 downloads

 

 

An example of such a difference is given in the  attachment below which investigates the effect of mishandling a dried soup product due to (cp/bc) in herbal/spice ingredients  As a result C.perfringens is discounted as a potential hazard. The micro limits in tables 1-3 (see M) for  cp/bc match those in upper attachment.

 

Attached File  micro. specs for dried soups.pdf   246.75KB   12 downloads

 

Perhaps manufacturers have (randomly) either (a) validated that onward usage presents no likelihood of dangerous elevation of either of cp/bc and therefore omitted it from the spec altogether, or (b) have simply "intuited" such a conclusion from the available monitoring data and acted similarly, or (c) have entered a limit such as given in upper attachment. (I note that the B.cereus value could be > than the one you mention.) :smile:


 


Kind Regards,

 

Charles.C


Thanked by 1 Member:

EssentialFA

    Grade - Active

  • IFSQN Active
  • 20 posts
  • 0 thanks
0
Neutral

  • United States
    United States

Posted 11 January 2016 - 10:43 PM

Thank you Charles.C, I really appreciate the explanations and documents!






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users