52 replies to this topic

[QAGB]

Hello,

 

My company is a mostly liquid sugar based process, and we've been trying to grasp APC (Aerobic Plate Count) levels. Our supplier for a specific ingredient has their standard for APC, and we have been following their standard since it seems to be an industry wide level. However, we are unable to consistently meet this. We're washing tanks (mostly hot water rinses at temperatures at or above 165F), and we are holding product at 165F for at least 30 minutes. We're working on heating products a little more (the products that can tolerate it are heated to 180F) for longer than 30 minutes. However, higher heat also causes an increase in thermophilic spore count (not that this directly relates to APC), among other quality related issues. We have seen promise with caustic washes, but those are very difficult to do without a delivery system. Any suggestions would greatly be appreciated.

 

QAGB

[Charles.C]

Hello,

 

My company is a mostly liquid sugar based process, and we've been trying to grasp APC (Aerobic Plate Count) levels. Our supplier for a specific ingredient has their standard for APC, and we have been following their standard since it seems to be an industry wide level. However, we are unable to consistently meet this. We're washing tanks (mostly hot water rinses at temperatures at or above 165F), and we are holding product at 165F for at least 30 minutes. We're working on heating products a little more (the products that can tolerate it are heated to 180F) for longer than 30 minutes. However, higher heat also causes an increase in thermophilic spore count (not that this directly relates to APC), among other quality related issues. We have seen promise with caustic washes, but those are very difficult to do without a delivery system. Any suggestions would greatly be appreciated.

 

QAGB

 

Hi QAGB,

 

Thanks for the Interesting query. Don't recall seeing this process discussed here before. I deduce this is some kind of "Pasteurization"  procedure whose effectiveness  is evaluated via APC?

 

Can you clarify the quantitative aspect of the standard such as specified APC procedure (eg BAM etc) and the required result ? (various factors like sampling / plate incubation time / temperature / calculation can vary and sometimes influence the result).

 

And an idea of yr currently achieved result ?

[QAGB]

Hi QAGB,

 

Thanks for the Interesting query. Don't recall seeing this process discussed here before. I deduce this is some kind of "Pasteurization"  procedure whose effectiveness  is evaluated via APC?

 

Can you clarify the quantitative aspect of the standard such as specified APC procedure (eg BAM etc) and the required result ? (various factors like sampling / plate incubation time / temperature / calculation can vary and sometimes influence the result).

 

And an idea of yr currently achieved result ?

 

Hi Charles,

 

Yes, it is in essence a pasteurization temperature. One of the microbial measures is APC, and we usually do Yeast, Mold, and maybe Coliform testing as well. Yeast, Mold, and Coliform numbers are almost always low and well within specification, but APC can be variably high or low, without a whole lot of rhyme or reason.

 

The quantitative measure for APC is done by AOAC Method 966.23, in which the end product is measured in CFUs per gram. We need to be ≤5,000, or ≤10,000 CFUs per gram, depending on the product. We can test a product and be at APC levels of 500 CFU/g, and then the next time 11,000 CFU/g, and then after that, back down to 3,000 CFU/g.

 

My initial thought has always been biofilm build-up, but if a tank is being cleaned regularly, should there be that much biofilm to cause this sort of randomness?

 

Thanks,

 

QAGB.

[Charles.C]

Hi QAGB,

 

A few obvious questions.

 

Unfortunately don't have aoac manual to hand, often same as BAM in which case should be reliable. 

I assume this is a batch operation.

What determines the process temperature / time / profile applied, ie what is the required lethality criterion, if any ?

do you validate that the process is consistent with respect to the (slowest heating point?) temperature of the contents ?

 

Do you monitor variations in the APC level of input material for different batches (hopefully homogeneous) ? if so what is max-min range ?

 

how many samples for APC do you give per finished batch (hopefully not 1), if >1, how much variation in results ?

 

Sorry if these questions have obvious answers, I have never seen a pasteurization efficiency defined by an APC result before (seeing as APC is not a criterion for safety). 

 

PS - i assume you are aware that APC results are inherently, notoriously, inaccurate. some possible non-process explanations could be due to sampling (heterogeneity) /analysis. i always send some dummy reference samples to check a lab at start of usage. The reference sample preparation can be tricky though.

[QAGB]

Hi QAGB,

 

A few obvious questions.

 

Unfortunately don't have aoac manual to hand, often same as BAM in which case should be reliable. 

I assume this is a batch operation.

What determines the process temperature / time / profile applied, ie what is the required lethality criterion, if any ?

do you validate that the process is consistent with respect to the (slowest heating point?) temperature of the contents ?

 

Do you monitor variations in the APC level of input material for different batches (hopefully homogeneous) ? if so what is max-min range ?

 

how many samples for APC do you give per finished batch (hopefully not 1), if >1, how much variation in results ?

 

Sorry if these questions have obvious answers, I have never seen a pasteurization efficiency defined by an APC result before (seeing as APC is not a criterion for safety). 

 

PS - i assume you are aware that APC results are inherently, notoriously, inaccurate. some possible non-process explanations could be due to sampling (heterogeneity) /analysis. i always send some dummy reference samples to check a lab at start of usage. The reference sample preparation can be tricky though.

 

Hi Charles,

 

We're not necessarily trying to define pasteurization efficiency in this instance. Pasteurization is not a CCP (Critical Control Point) for the process, as we only really refer to it as "heating" and we consider it a CP (Control Point). I absolutely agree with you that APC is not a criterion for safety. The issue is that many of our customers have specific requirements for micro (especially APC). They want low micro products because their finished products need to have low micro (including APC) for one reason or another.

 

As an example, a new customer was buying their sugar from a different supplier. They no longer wish to purchase from them, and they contact us. They will give us specifications for the product they used to buy. They will either provide the old supplier's specifications (which they want us to match - and want to know why we can't), or provide internal specifications (to which they will not budge). We then have specifications for micro, and we can sometimes fall well within the range for APC, and then others, we're somewhat off. Alternatively in some cases, we receive our own supplier specifications, and we try to remain within those specifications because those specifications are typically industry standard. In those cases, customers tend to know the industry wide specifications, and want to know why we can't offer similar micro specs.

 

However, it is indeed a batch operation, and you're absolutely right in that labs can be very inconsistent. We were at one time using a very well-known, well-respected laboratory. At one time we were getting consistent results. Then all of a sudden, we would send them a sample, and get 9000 cfu/g for APC, and send another sample from the same batch, and get 4,000 cfu/g. We switched to another well-known laboratory, and the results have been extremely consistent when we send duplicate samples. Therefore, we know it isn't the aseptic sampling methods in-house.

 

We use the Pasteurization: A Food Industry Practical Guide (the table is listed in the back of the BRC Guide) to determine lethality and time, as well as Dr. Fred Breidt's work to determine the heating time and temperature. In both instances, we really only have to heat to 165 for about 3/4 a minute, but we're holding at 30 minutes to try and reduce counts further. Our products are heated over long hours (usually 4 - 8), since it takes that long to get up to temperature in most cases, and it is held there. Depending on when the product is scheduled to be produced, it can be held even longer at a temperature of 120F or greater, which should remove any doubt about the slowest heating points of the tanks in question.

 

We haven't checked input levels of the materials, only because the raw materials can be moved around so much to make room for more incoming product. There's also a lot of co-mingling with 48,000 - 120,000 lb. tanks (so we fear we won't get much information from this - and could make matters more confusing).

 

I never thought about doing a larger composite sampling of the batch. We'd only been sending out 1 sample, but we have in many instances tested a batch twice (from separate drums or other packaging), and gotten almost exactly the same results (typically with the more consistent lab, and when the other "not so consistent lab" was handling testing better). Perhaps we should do composite testing alongside the typical sample we submit, and see what happens. The product is uniformly distributed though (all homogeneous liquid). Should we really be getting high APC in one part of the batch, but not another (if it is all blended)?

 

I apologize for the long winded reply, but this situation is so convoluted.

 

Thanks,

 

QAGB

[Charles.C]

Hi QAGB,

 

Many thks for the highly detailed intro. Fascinating stuff.

 

IMEX, it is an achievement to consistently maintain APC results within a factor of 10 of each other, eg 1000 - 10,000 cfu/g  on same sample.  Such a range is perfectly acceptable if the spec is 100,000 max but useless for 5000. IMO you have a mighty tough control problem at the levels you mention unless yr raw material is already near the target levels which i presume it normally isn't.

 

I would have thought it almost impossible to devise a heating profile to consistently achieve a range of output  APCs within better  than, say, +/- 50% of a given target unless the raw material has a highly uniform  APC value and an equally uniform/reproducible heating profile is achievable. But only too willing to be disproved. Maybe yr input does have APC levels within a very close band as a result of its previous manufacture which should simplify the possible errors in predicting output results. Would be good to know IMO.

 

Seafood customers IMEX sometimes demand export APC micro. impossibilities which lead to lost contracts, usually after interminable technical arguments. It's sometimes easier if you have in-house micro. labs since you know (only too) well the limitations.

 

Frankly i would never claim 1 sample as representative of any process lot unless i had validated that the internal variations were negligible. And probably not even then. i have great faith in averages and the stories they tell. (+ APC measurements are usually quite cheap).

 

I agree liquids should be reasonably homogeneous but heating profiles may well not be unless there is a mixing unit incorporated (?). 

 

Offhand I suspect the output APC profile will have a considerable variance about the average so that you presumably have to target considerably above below the spec. to prevent failures if later randomly sampled.

(I have an equivalent problem when calculating minimum pack weights of frozen goods so as to comply with legal thawed limits. My own product/industry conclusion is one has to "give away" product based on trial and error and constant monitoring).

 

IMO you initially need to be sure what the statistical reality of yr APC results is like, ie whether yr present result is improvable or basically limited by the system. This relates to yr day-by-day/ lot-by-lot data for batches in relation to yr APC targets. 

 

On the other hand, there may be some people here working in yr product area who are familiar with this scenario and can suggest an immediate fix. Hope so.

[QAGB]

Hi QAGB,

 

Many thks for the highly detailed intro. Fascinating stuff.

 

IMEX, it is an achievement to consistently maintain APC results within a factor of 10 of each other, eg 1000 - 10,000 cfu/g  on same sample.  Such a range is perfectly acceptable if the spec is 100,000 max but useless for 5000. IMO you have a mighty tough control problem at the levels you mention unless yr raw material is already near the target levels which i presume it normally isn't.

 

I would have thought it almost impossible to devise a heating profile to consistently achieve a range of output  APCs within better  than, say, +/- 50% of a given target unless the raw material has a highly uniform  APC value and an equally uniform/reproducible heating profile is achievable. But only too willing to be disproved. Maybe yr input does have APC levels within a very close band as a result of its previous manufacture which should simplify the possible errors in predicting output results. Would be good to know IMO.

 

Seafood customers IMEX sometimes demand export APC micro. impossibilities which lead to lost contracts, usually after interminable technical arguments. It's sometimes easier if you have in-house micro. labs since you know (only too) well the limitations.

 

Frankly i would never claim 1 sample as representative of any process lot unless i had validated that the internal variations were negligible. And probably not even then. i have great faith in averages and the stories they tell. (+ APC measurements are usually quite cheap).

 

I agree liquids should be reasonably homogeneous but heating profiles may well not be unless there is a mixing unit incorporated (?). 

 

Offhand I suspect the output APC profile will have a considerable variance about the average so that you have to target considerably above the average to prevent failures if later randomly sampled.

(I have an equivalent problem when calculating minimum pack weights of frozen goods so as to comply with legal thawed limits. My own product/industry conclusion is one has to "give away" product based on trial and error and constant monitoring).

 

IMO you initially need to be sure what the statistical reality of yr APC results is like, ie whether yr present result is improvable or basically limited by the system. This relates to yr day-by-day/ lot-by-lot data for batches in relation to yr APC targets. 

 

On the other hand, there may be some people here working in yr product area who are familiar with this scenario and can suggest an immediate fix. Hope so.

 

 

Hi Charles,

 

Thank you for your thoughts on this. Our raw materials (which come from the supplier) do have micro limits. For us, they are raw materials, but in this case they have already been processed prior to our receipt. We are processing further at this facility.We use that supplier's micro limits as our own, and have been told we should be able to stay well within the range. I've had numerous discussions with our suppliers of the material, but I haven't gotten anywhere with it. That's why I am here, to see if there are others out there with similar issues. It is unfortunate that this APC range is industry standard, but we don't have too many options, even though APC is not an indicator of food safety.

 

 

We're a small company, and while having our own micro lab in-house would be useful, it's just easier for us to use an already established lab with ISO 17025 certs as our backbone of verification. I would love for us to be able to get our results on a more timely basis, since our customers need to know lead times and we have to have some sort of cushion when the APC results are not in the range needed.

 

There is a mixing unit incorporated towards the end of the process where the blend is actually being heated to the pasteurization temperature. Those are most often the locations for the majority of rinses and caustic washes for us. Therefore, the blend is heated homogeneously (over a lengthy period of time) and I would think that the batch should have similar micro throughout as a result.

 

In regard to the variance you're talking about, I assume that's why the supplier chose the level they did (they figure the CFU/g should be high enough that the results would in most cases fall in the range), because we've been told we shouldn't have trouble meeting that specification. (Side note: I too know well the headaches of calculating MAVs and trying to adjust items due to varying specific gravities and so forth. I imagine frozen and thaw differences must be a much larger headache than my own.)

 

I think you've given me a good explanation of the statistical issues of 1 sample per batch. We may try to do some composite testing on future batches, and see what we can come up with.

 

Again many thanks, this has been a great help.

 

QAGB

[Charles.C]

Hi QAGB,

 

Thks yr summary.

 

I have one more query.

 

I noted the limits you mentioned earlier. Are  these averages for which the number of samples is specified ?

 

Most typical "food" APC micro. specs for contractual / regulatory purposes are usually minimally based on 5 sample units associated with, for example, an interpretation via nmMc method.

 

This theoretically reduces the risk of an outrageously bad sample.  I get the impression that yr industry is more of a sudden death situation, eg 1 for All ?

[QAGB]

Hi QAGB,

 

Thks yr summary.

 

I have one more query.

 

I noted the limits you mentioned earlier. Are  these averages for which the number of samples is specified ?

 

Most typical "food" APC micro. specs for contractual / regulatory purposes are usually minimally based on 5 sample units associated with, for example, an interpretation via nmMc method.

 

This theoretically reduces the risk of an outrageously bad sample.  I get the impression that yr industry is more of a sudden death situation, eg 1 for All ?

 

 

Hi Charles,

 

Your impression is correct. It is definitely a "sudden death" situation. I've actually never seen any sort of data sheet from any supplier in this industry that declares a 5 sample minimum for any micro specs (and we have between 50-100 suppliers). The suppliers probably do a lot of sampling to come up with their specs, but they don't outright say that the numbers provided are based on a particular number of samples. A lot of our specifications come from historical data and other testing along the way (fairly large sample sizes), so I'm sure the sample size (or at least I hope) is >1.

 

Thanks,

 

QAGB

[Charles.C]

Hi QAGB,

 

However, higher heat also causes an increase in thermophilic spore count (not that this directly relates to APC), among other quality related issues. We have seen promise with caustic washes, but those are very difficult to do without a delivery system. 

 

Having become somewhat pessimistic regarding the statistics in yr business, I re-viewed yr original question.

 

To reveal my ignorance, why does increasing heat produce the spore result noted above, basic defence mechanism ? I presume unrelated to APC since you are below the thermophilic temperature germination range ?

 

Is it physically impractical to do any kind of micro. swabbing of the interior surfaces so as to rule out contamination problems. ? Or via final  flushing out measurements ?

 

I deduce you are practically trying to achieve an “xD” reduction of the initial “total” bacteria in the input using some kind of extrapolation from L.mono based tables (BRC). Maybe the Fred Breidt reference adjusts for the species aspect. The basic reduction equations are simple and well-known but  meaningful system parameters (D/z values) can be harder to select / need validation. Plus there is inevitably some distribution of  bacterial concentration about an  average rather than a single number (the FDA analysis for Salmonella food pasteurization requirements takes the latter factor into account from memory).

 

Regardless, I still suspect yr basic APC headaches may be driven by statistics / accept-reject methodologies rather than heat supply limitations.

Just as for my packed weight problems, there may be 2 requirements - (a) a heat profile sufficient to generate an acceptably low average output APC  and (b) knowledge of the variance of the APC distribution such that the chance of selecting a random sample size of ONE (!)  exceeding yr max. limit can be made adequately low. This is standard textbook stuff but reliable data in yr system is i fear hard to get.

 

I daresay yr current approach tends to be "overkill" in order to eliminate (b) by brute force.

 

It's an interesting dynamics.   

[QAGB]

Hi QAGB,

 

 

Having become somewhat pessimistic regarding the statistics in yr business, I re-viewed yr original question.

 

To reveal my ignorance, why does increasing heat produce the spore result noted above, basic defence mechanism ? I presume unrelated to APC since you are below the thermophilic temperature germination range ?

 

Is it physically impractical to do any kind of micro. swabbing of the interior surfaces so as to rule out contamination problems. ? Or via final  flushing out measurements ?

 

I deduce you are practically trying to achieve an “xD” reduction of the initial “total” bacteria in the input using some kind of extrapolation from L.mono based tables (BRC). Maybe the Fred Breidt reference adjusts for the species aspect. The basic reduction equations are simple and well-known but  meaningful system parameters (D/z values) can be harder to select / need validation. Plus there is inevitably some distribution of  bacterial concentration about an  average rather than a single number (the FDA analysis for Salmonella food pasteurization requirements takes the latter factor into account from memory).

 

Regardless, I still suspect yr basic APC headaches may be driven by statistics / accept-reject methodologies rather than heat supply limitations.

Just as for my packed weight problems, there may be 2 requirements - (a) a heat profile sufficient to generate an acceptably low average output APC  and (b) knowledge of the variance of the APC distribution such that the chance of selecting a random sample size of ONE (!)  exceeding yr max. limit can be made adequately low. This is standard textbook stuff but reliable data in yr system is i fear hard to get.

 

I daresay yr current approach tends to be "overkill" in order to eliminate (b) by brute force.

 

It's an interesting dynamics.   

 

 

Hi Charles,

 

APC basically measures the mesophilic bacteria in a substance (basically a count of active bacteria). Thermophilic spores by nature are vegetative, and tend to live at higher temperatures, so they are not necessarily connected to the point of deducing trends or issues from one to the other (so yes, we are below the germination range). The higher we heat the product, the more we activate thermophilic spores.

 

It is indeed an interesting dynamic. We use the pasteurization temperature references to basically give us guidance as to how long we really need to hold the product. Beyond that, we've basically been working with numbers to try and validate the process, but there are so many variables involved that it's extremely difficult to find a suitable control.

 

We're still going to take your approach on increasing the sample size and performing a composite test along with our usual sample to see where that gets us. It's a great idea, and definitely worth a shot.

 

QAGB

[Charles.C]

Hi QAGB,

 

I had a quick look at some background to liquid sugar (“syrup”?) systems on the IT. It’s certainly a very specialized environment (must admit i didn't realize how specialized). And, as you say, not simple.

 

As i understood the system is a subtle mix of,  I presume, minute quantities of  residual, potentially, mesophilic/thermophilic vegetative species (didn’t see species details) and thermophilic spores (and perhaps minute quantities of mesophilic spores also ?)

 

The AOAC test procedure looks same as BAM and, as you noted, incubation around 35degC will primarily  focus on mesophilic entities (IMEX overlaps are always occurring in the bacterial world).

 

I note you are typically holding around 72degC with some potential boosting to around 80degC (no good with Fahrenheit values). From memory one needs temperatures in the high 60’s to begin producing much lethality for mesophiles (and some psychrophiles). I assume 80degC is still too low for thermophilic spore activity (didn’t check).

 

I have no doubt that this system has been extensively studied already in the economic context of the specs you mentioned, unfortunately I did not see any related, detailed, publications so far on IT. I did note a myriad of  spore specs for solid sugar where there is apparently a wider spectrum.

 

It was not obvious to me if there can be any significant contribution from germinating mesophilic spores (if any present?) at yr holding temperatures or perhaps such is assumed negligible ? The latter would certainly simplify any  maths.

 

I can only make 2 opinions based on my very limited knowledge of this system (and its specs) -

 

(1) Logically it would be nice to possess “effective”  D/z values which could simply be plugged into the same formula as used for species like L.mono/samonella. I deduce reliable data is not yet available for generic liquid sugar systems hence you are following analogies to, eg L.mono, then (a lot?) of  trial and error. The determination of such reference values inevitably involves my 2^nd opinion.

 

(2) As already discussed, it seems crucial to me to validate the “accuracy” of yr existing APC data. I’m rather surprised if no-one has not done this already since it’s a logical parallel to challenge tests.

 

No sugar people on this forum ??

[QAGB]

Hi QAGB,

 

I had a quick look at some background to liquid sugar (“syrup”?) systems on the IT. It’s certainly a very specialized environment (must admit i didn't realize how specialized). And, as you say, not simple.

 

As i understood the system is a subtle mix of,  I presume, minute quantities of  residual, potentially, mesophilic/thermophilic vegetative species (didn’t see species details) and thermophilic spores (and perhaps minute quantities of mesophilic spores also ?)

 

The AOAC test procedure looks same as BAM and, as you noted, incubation around 35degC will primarily  focus on mesophilic entities (IMEX overlaps are always occurring in the bacterial world).

 

I note you are typically holding around 72degC with some potential boosting to around 80degC (no good with Fahrenheit values). From memory one needs temperatures in the high 60’s to begin producing much lethality for mesophiles (and some psychrophiles). I assume 80degC is still too low for thermophilic spore activity (didn’t check).

 

I have no doubt that this system has been extensively studied already in the economic context of the specs you mentioned, unfortunately I did not see any related, detailed, publications so far on IT. I did note a myriad of  spore specs for solid sugar where there is apparently a wider spectrum.

 

It was not obvious to me if there can be any significant contribution from germinating mesophilic spores (if any present?) at yr holding temperatures or perhaps such is assumed negligible ? The latter would certainly simplify any  maths.

 

I can only make 2 opinions based on my very limited knowledge of this system (and its specs) -

 

(1) Logically it would be nice to possess “effective”  D/z values which could simply be plugged into the same formula as used for species like L.mono/samonella. I deduce reliable data is not yet available for generic liquid sugar systems hence you are following analogies to, eg L.mono, then (a lot?) of  trial and error. The determination of such reference values inevitably involves my 2^nd opinion.

 

(2) As already discussed, it seems crucial to me to validate the “accuracy” of yr existing APC data. I’m rather surprised if no-one has not done this already since it’s a logical parallel to challenge tests.

 

No sugar people on this forum ??

 

 

Hi Charles,

 

Yes, we pretty much work with sugar in all of its different forms here. Sugar is very specialized, but in many ways is very simple. There are parts of the sugar industry where micro isn't a concern (like corn syrups). There are other parts of the sugar industry where micro presents many challenges.

 

In this case, the product will have the entire spectrum of organisms. Industry speaking, we are mostly concerned with mesophiles due to storage. Mesophiles are the concern at around room temperature, so that is where you really want to reduce numbers. Mesophiles are going to be the happiest between 20-50C (70-120F). Thermophiles are happiest between 45-125C (115-250F). So at the temperatures we're using for heating, we're killing mesophiles (they won't be germinating too much at that temperature), but in return, we're giving thermophiles somewhat of a pleasant growth environment. We just don't want to keep trying to heat higher and higher and have more issues with thermophilic spores.

 

We don't have an actual specification for spores, as it isn't something widely tested in the industry for liquid sugar. We just don't want to try to solve one problem and create a new one in the process. It's just been a circular process for us, trying to figure out what the solution would be.

 

QAGB

[Charles.C]

Hi QAGB,

 

Thks yr comments. Taken on board.

 

Have been attempting to slightly self-educate myself on this topic. I located the attached paper which is clearly not a recent publication (as are many other refs in this field) and may/may not relate to yr particular process style. You may be familiar with it already.

 

 Liquid Sugar - counting methods.pdf   701.72KB   37 downloads

 

The product discussed is clearly of a much lower APC than you are producing so the “customized” analytical procedures are probably not relevant to yr situation however the comments regarding non-uniform distribution of microbial characteristics / sampling aspects could be of  general interest. The significant question is whether the comments may relate to higher microbial levels as in yr product.

 

Just in passing, I also noted that some processors are routinely supplying product at very low APC levels such as <200cfu/g in this attachment –

 

 Liquid_Sucrose.pdf   501.11KB   26 downloads

 

I noticed this textbook comment -

 

A substantial portion of the refinery's output is sold as liquid sugar. This can be in the form of (1) dissolved granulated sugar, (2) fine liquor or (3) intermediate liquors from the various crystallising stages.

 

I'm only speculating via the limits you previously gave but I guess 1 > 3 may represent a (substantially) increasing progression in APC value (and similarly in max. limit values). > 10:1 perhaps ? (not too sure what "fine" liquor actually is)

 

The product currently under discussion is perhaps from "3" ?

 

I don't know how you are determining / applying the implemented heat treatment but as I understand, yr main problem is not that you cannot get to the desired low APC level, it's more that you cannot get there consistently due to the large swings in results for the ONE sample unit submitted per batch.

 

Frankly, assuming you apply some kind of xD logic to meet the APC limit, I find it difficult to see how you can plan any heat treatment without some assumption regarding the max. starting APC level. But perhaps you've found a way, eg Big Bazooka (BB). I do wonder how you convince yourselves that the APC results for successive batches should be the same ?.

 

I can think of a variety of possible causes to explain the variations you are getting, eg  -

 

(1) different batches have significantly different start APCs, causes similar variations in output results - ie may be genuine reported variations if lab is reliable

(2) diferrent batches have similar start APCs but the applied heat treatment varies - ie may be genuine reported variations if lab is reliable

(3) there is a significant variation in APC within any given lot - need data

(4) there is a significant variation in APC between different lots - need data

(5) the calculated/applied heat treatment is not consistently correlated to the starting APC so cannot achieve a similar result - need method

(6) the sampling procedure for submitted sample is not aseptic - need method.

(7) the lab handling/analytical procedure is not consistent/correct - need to test with random duplicates

 

The classic statistical method to solve such problems, ie identify the weak point, is via methods like ANOVA but they entail a lot of work and analysis. Probably easiest route is to examine the most accessible elements first.

 

Personally i would validate the lab first. If Ok then the sampling. if OK then the process. Difficult to offer much thouight on the process due limited knowledge (2-ways).

 

One popular, albeit crude, solution is a BB approach, eg target low enough so that the swings do not give results out of compliance. This is the approach i am forced to use for packed weight control of some products where major texture variations within/between lots simply rule out close control. Problem is that it may cause a financial hiccup. Maybe you are doing this now ?

 

I also noticed a couple of sites with micro. sampling plans for liquid sugar based on lot sizes although i daresay you are already well informed on this subject. Can post if you are interested (and i can re-find).

[QAGB]

Hi QAGB,

 

Thks yr comments. Taken on board.

 

Have been attempting to slightly self-educate myself on this topic. I located the attached paper which is clearly not a recent publication (as are many other refs in this field) and may/may not relate to yr particular process style. You may be familiar with it already.

 

Liquid Sugar - counting methods.pdf

 

The product discussed is clearly of a much lower APC than you are producing so the “customized” analytical procedures are probably not relevant to yr situation however the comments regarding non-uniform distribution of microbial characteristics / sampling aspects could be of  general interest. The significant question is whether the comments may relate to higher microbial levels as in yr product.

 

Just in passing, I also noted that some processors are routinely supplying product at very low APC levels such as <200cfu/g in this attachment –

 

Liquid_Sucrose.pdf

 

I noticed this textbook comment -

 

 

I'm only speculating via the limits you previously gave but I guess 1 > 3 may represent a (substantially) increasing progression in APC value (and similarly in max. limit values). > 10:1 perhaps ? (not too sure what "fine" liquor actually is)

 

The product currently under discussion is perhaps from "3" ?

 

I don't know how you are determining / applying the implemented heat treatment but as I understand, yr main problem is not that you cannot get to the desired low APC level, it's more that you cannot get there consistently due to the large swings in results for the ONE sample unit submitted per batch.

 

Frankly, assuming you apply some kind of xD logic to meet the APC limit, I find it difficult to see how you can plan any heat treatment without some assumption regarding the max. starting APC level. But perhaps you've found a way, eg Big Bazooka (BB). I do wonder how you convince yourselves that the APC results for successive batches should be the same ?.

 

I can think of a variety of possible causes to explain the variations you are getting, eg  -

 

(1) different batches have significantly different start APCs, causes similar variations in output results - ie may be genuine reported variations if lab is reliable

(2) diferrent batches have similar start APCs but the applied heat treatment varies - ie may be genuine reported variations if lab is reliable

(3) there is a significant variation in APC within any given lot - need data

(4) there is a significant variation in APC between different lots - need data

(5) the calculated/applied heat treatment is not consistently correlated to the starting APC so cannot achieve a similar result - need method

(6) the sampling procedure for submitted sample is not aseptic - need method.

(7) the lab handling/analytical procedure is not consistent/correct - need to test with random duplicates

 

The classic statistical method to solve such problems, ie identify the weak point, is via methods like ANOVA but they entail a lot of work and analysis. Probably easiest route is to examine the most accessible elements first.

 

Personally i would validate the lab first. If Ok then the sampling. if OK then the process. Difficult to offer much thouight on the process due limited knowledge (2-ways).

 

One popular, albeit crude, solution is a BB approach, eg target low enough so that the swings do not give results out of compliance. This is the approach i am forced to use for packed weight control of some products where major texture variations within/between lots simply rule out close control. Problem is that it may cause a financial hiccup. Maybe you are doing this now ?

 

I also noticed a couple of sites with micro. sampling plans for liquid sugar based on lot sizes although i daresay you are already well informed on this subject. Can post if you are interested (and i can re-find).

 

 

Hi Charles,

 

 

Thank you for the information on Liquid Sugar - Counting Methods. That definitely helps us, and strengthens your point about sample size. I had no idea that yeasts tend to migrate to the surface of the product, and that the organisms aren't necessarily evenly distributed in the liquid. Yes, we also work with liquid sucrose. APC in that product tends to be low (even <10/cfu/g APC) as long as the product is stable. The example you provided looks to be 10g dry basis.

 

However, the products under discussion range from 2 - 3 (intermediate and fine liquors). The reason why we "assume" we can stay below the required APC level is that the supplier has indicated as such. The supplier has a range, and upon discussions with them, they feel we should be able to reasonably be in those levels. I can mostly agree with that, but I know there will be variation. It depends on the APC of the product leaving the supplier, the APC levels of the tanker, APC levels in the tank upon arrival, how many products pass through between tank rinses, receiving dirtier loads compared to cleaner loads, checking pipelines, etc. We have a lot of variables, which also lead to our confusion and issues sorting this out. I would definitely not think consecutive batches would have the same APC, because of the variables. However, I do think that with the heating and rinsing being done, we should be able to reduce mesophile levels significantly.

 

As far as the causes you've listed:

(1) different batches have significantly different start APCs - Absolutely agree with this point. The supplier has a micro specification though for the raw product incoming. We do feel that there are most certainly significantly different start APCs, but heating, rinses, and/or caustic washes when biofilm buildup may occur should alleviate the problem (I assume).

(2) diferrent batches have similar start APCs but the applied heat treatment varies - I doubt this would be the case. It is unlikely to have similar start APCs for this liquid sugar product. The applied heat treatment is actually not variable, unless we introduce higher heat as a test.

(3) there is a significant variation in APC within any given lot - we will be working to see if this is the case, considering our low sample size.

(4) there is a significant variation in APC between different lots - yes, there would be significant variation in APC, but still we should be able to stay within the range needed.

(5) the calculated/applied heat treatment is not consistently correlated to the starting APC so cannot achieve a similar result - I can't confirm that either, but the heat treatment should be reducing mesophile counts.

(6) the sampling procedure for submitted sample is not aseptic - sampling method is indeed aseptic. We use sealed specimen containers, heating, sterilization, etc. Duplicate testing has shown this to be the case.

(7) the lab handling/analytical procedure is not consistent/correct - we have done random duplicates before, and our current lab is consistent. The previous lab handling was consistent for a while, and perhaps some internal changes occurred causing inconsistencies. That's why we switched to the lab we work with now.

 

 

Unfortunately we cannot target low, because of customer requirements. I would love to do that though, as it would save a lot of headaches.

 

 

Thanks,

 

QAGB

[Charles.C]

Hi QAGB,

 

Thks for the very detailed comments. I think I get the basic elements of yr process. (only the very basics !). As you say, there are a lot of potential contributions to the average APC values and their variances . On the other hand, it is statistically not so unusual to find one or two factors dominate.

 

Yr final comment slightly puzzled me. I deduce yr customer also has a minimum requirement, presumably because other relevant product characteristics excessively degrade  if “over-heated”. That really places demands on the process IMO. My own products input/output APC profiles can display (internal)   ranges (ie max/min) of APC of 10:1 and be fine.

It seems to me that yr operational set-up relies considerably on plant experience / trust in yr material source. Not so unusual IMEX and works if the results are satisfactory. But if problems arise one needs data.

 

Couple of comments –

 

To validate a lab, duplicates are useful but it is also relevant to know the “true” result for the sample given ( Precision / Accuracy). Unfortunately “reference” micro. samples are not always easy to generate. If financing allows, one useful procedure is to do a ring test, produce a reliably mixed representative sample and divide/send portions to several labs, preferably in duplicate and randomly within a group. Then compare the results. You may get a shock.

 

IMO it would be highly informative to estimate the APC profile of yr initial material. And similarly for the output. Just for a few lots to get an idea of the internal ranges involved and to compare batches. Easy to say………

 

I’m curious how your starting material is produced ? A blend from various arbitrary sources ? This is a common problem in my business. And unavoidable in many cases.

[QAGB]

Hi QAGB,

 

Thks for the very detailed comments. I think I get the basic elements of yr process. (only the very basics !). As you say, there are a lot of potential contributions to the average APC values and their variances . On the other hand, it is statistically not so unusual to find one or two factors dominate.

 

Yr final comment slightly puzzled me. I deduce yr customer also has a minimum requirement, presumably because other relevant product characteristics excessively degrade  if “over-heated”. That really places demands on the process IMO. My own products input/output APC profiles can display (internal)   ranges (ie max/min) of APC of 10:1 and be fine.

It seems to me that yr operational set-up relies considerably on plant experience / trust in yr material source. Not so unusual IMEX and works if the results are satisfactory. But if problems arise one needs data.

 

Couple of comments –

 

To validate a lab, duplicates are useful but it is also relevant to know the “true” result for the sample given ( Precision / Accuracy). Unfortunately “reference” micro. samples are not always easy to generate. If financing allows, one useful procedure is to do a ring test, produce a reliably mixed representative sample and divide/send portions to several labs, preferably in duplicate and randomly within a group. Then compare the results. You may get a shock.

 

IMO it would be highly informative to estimate the APC profile of yr initial material. And similarly for the output. Just for a few lots to get an idea of the internal ranges involved and to compare batches. Easy to say………

 

I’m curious how your starting material is produced ? A blend from various arbitrary sources ? This is a common problem in my business. And unavoidable in many cases.

Hi Charles,

 

It's actually more than one customer for us. It's a number of them, and almost all of them want that same APC specification. The process does rely heavily on the materials sourced; which is why this has caused so much confusion for us.

 

Most of the end products in the category of concern are blends of different materials, varying sources depending on the formula. You're right; we should check the APC profile of the initial material, but the initial material gets co-mingled upon arrival. We could check the tankers, but that doesn't necessarily imply that the issue comes from the initial material; it could mean the tanker itself had an issue. However, testing that material won't necessarily tell us what's happening at the end of the process anyway, due to co-mingling and blending of the materials we receive. It would be different if the material was not a blend, and each batch went into its own tank.

 

I think our first point of action will be to have a larger sample size; that will give us some sort of indication as to whether we're just sampling incorrectly and the average micro results are better than we think. The next step will be to complete the ring testing as you mention. Our last course of action would be checking incoming material due to the nature of the process. In between all of that, I still need more data on our caustic washes, to get an idea on how that reduces counts.

 

Thanks,

 

QAGB

[Charles.C]

Hi QAGB,

 

Yes, co-mingling is a classic headache for investigative purposes in many industries.

Sadly, micro.measurements are generally also useless for on-the-spot evaluations of course.

Knowledge as to what is actually coming from tankers may not be completely uninteresting though even if logistically challenging, one bad apple/source...

(it's so much easier/defendable with in-house micro. facilities)

 

Not directly related (and I appreciate you are not processing via crystals) but in the context of yr caustic treatment, i thought this extract might be of interest (I daresay common knowledge in the business) -

 

 liquid sugar2.png   274.73KB   1 downloads

[QAGB]

Hi QAGB,

 

Yes, co-mingling is a classic headache for investigative purposes in many industries.

Sadly, micro.measurements are generally also useless for on-the-spot evaluations of course.

Knowledge as to what is actually coming from tankers may not be completely uninteresting though even if logistically challenging, one bad apple/source...

(it's so much easier/defendable with in-house micro. facilities)

 

Not directly related but in the context of yr caustic treatment, i thought this extract might be of interest (I daresay common knowledge in the business) -

 

liquid sugar2.png

 

 

Hi Charles,

 

I agree, every bit of information can be helpful in some way, but I don't want to make the situation more cloudy yet. For us, it's better to work backwards to find the root of the problem due to the co-mingling.

 

I just read through that document and yes, a lot of spoilage organisms live with high water activity. Most of the products we work with have low water activity, give or take a couple. We also do use UV lights on the tops of our tanks as well. We've tried to take most of the classic precautions of the industry so we can at least rule out that aspect of it.

 

Thanks,

 

QAGB

[MsFS]

I would also suggest performing micro testing of the equipment rinse water following the cleaning process.  Using a rinse or caustic wash may not be enough to keep the counts down - you may have to follow the caustic wash with a sanitizer.

[QAGB]

I would also suggest performing micro testing of the equipment rinse water following the cleaning process.  Using a rinse or caustic wash may not be enough to keep the counts down - you may have to follow the caustic wash with a sanitizer.

 

Hi MsFS,

 

That is a great idea. I hadn't even thought about the residual water, that would be extremely helpful to us as well.

 

Thank you for your input!

 

QAGB

[Charles.C]

Hi QAGB,

 

Ahem. Post 10.

 

Probably lost in the verbosity.

[QAGB]

Hi QAGB,

 

Ahem. Post 10.

 

Probably lost in the verbosity.

 

Hi Charles,

 

Sorry about that, it wasn't lost in the verbosity, I completely didn't see it.    I saw the part about swabbing the interior only and must have skipped the sentence during processing. We can't swab the insides due to the sizes of the tanks (and confined space training), but we can do flushing measurements as you said.

 

Thanks to you and MsFS both!

 

Update: Today is our first day of increasing sample size. We're going to compare it with our typical one sample analysis and see how it turns out. Also, a caustic wash was done prior to preparing the batch, as a further notation.

 

 

Thanks,

 

QAGB

[Charles.C]

hi QAGB,

 

No problem.

 

I did have one more thought.

 

If expense permits and the division involved can be done adequately aseptically, you might consider selecting one or preferably  2 (well-mixed) units from the total of hopefully >2 samples  and splitting the selected samples in half.

 

Could then send the two halves, randomly labelled, to separate labs, ideally increased with one more as a dummy (but no dummy if a risk of confusion / over-budget).

 

Was just a thought. i can imagine this sampling operation is not simple already (depends how readily different areas are accessible, bottle on a long string ?)

 

I also assume that incoming tanker lots have nothing like COAs attached, esp. from a micro POV ? Do you do any  sampling/analysis at all ?

[QAGB]

hi QAGB,

 

No problem.

 

I did have one more thought.

 

If expense permits and the division involved can be done adequately aseptically, you might consider selecting one or preferably  2 (well-mixed) units from the total of hopefully >2 samples  and splitting the selected samples in half.

 

Could then send the two halves, randomly labelled, to separate labs, ideally increased with one more as a dummy (but no dummy if a risk of confusion / over-budget).

 

Was just a thought. i can imagine this sampling operation is not simple already (depends how readily different areas are accessible, bottle on a long string ?)

 

I also assume that incoming tanker lots have nothing like COAs attached, esp. from a micro POV ? Do you do any  sampling/analysis at all ?

 

 

Hi Charles,

 

We don't want to cover too many variables at one time. We may do some ring testing on laboratory analysis soon but not just yet. I really wouldn't be shocked to find two different labs supposedly following the same methods come up with totally different results. We have had that happen in the past on simple things such as pH or moisture tests. I was more saddened than amazed by the findings considering most of the labs we use are ISO 17025 certified, and the methods for these things are simple and well-known.

 

We actually have a sample point, which happens to be the filling stations. What you get at the filling stations is representative of what is in going into packaging. Those areas get cleaned and sanitized and heated as well, and the initial product is also run through before taking the first sample.

 

The incoming lots don't come with micro values on COAs but we do get COAs for the material. We do some testing on it, but nothing from a micro standpoint.

 

Thanks,

 

QAGB