Column COA

[lcguy1]

We just received some new columns. We also have finally received the COA's for these columns. The problem occurs with the specs and the results on the COA's. We have both guard cartridges and 15cm columns from the same batch number and that batch, according to the COA, fails the 90% volumn value for particle size. There is a disclaimer at the bottom which states that "As per technical evaluation the particel size 90% volumn result is acceptable since the material affords column efficiency within our specified range and the column backpressure is normal." The spec is 5.6-6.0 and the result is 6.1. My question is, Why have specs for it of it is OK to fail? Should these be returned and a request made for colums with a different batch number? There is a difference of opinion here.

[DR]

I'd run the assay for which the column was purchased. If all is well, keep 'em. If you see marginal resolution or tailing, return 'em.

If there is no specific assay for which the columns were purchased (you're doing method development with them), I'd return them as you have good reason to suspect that they will not be representative of future batches.

[lcguy1]

I should add that we are a pharm QC lab under GMP's. I guess the problem I have with the whole thing is that it does indeed fail the specs. If any test I run fails, I am not able to just put a statement in there to say , Oh never mind the results, it really passes because it still looks like it should. Limits are limits, that is why we have them.

[tom jupille]

The pharmaceutical industry operates under the tightest possible regulations (with good reason) and does the highest quality work as a result. To paraphrase DR's post, the relevant question is not whether or not they meet the manufacturer's specs, but whether or not they meet your specs.

[Uwe Neue]

That should not happen. A spec is a spec...

[DR]

A spec is a spec until the vendor decides that they've set a spec too tight and for no justifiable reason. Then specs are changed, but this can take a while. The vendor has the luxury of allowing customers to decide if they want the columns or not (hence the disclaimer - I'm certain that if the OP were to call the vendor to get a RMA number and explained that they work in the Pharma industry, they could return the column). If it were the customer having an issue with a spec, there would be no option as a result of adherence to CGMP - they should not ship the product. The vendor is not bound by CGMP regulations but has made a "good faith" attempt to alert customers about the issue so that they can decide for themselves.

Assuming that the vendor is correct and the columns are fine, I would guess that returns on the lot would run about 50% from Pharma customers, half of whom will be concerned about the spec, 25% of whom will read the disclaimer, try and accept the columns and 25% who will fail to notice the disclaimer.

[Uwe Neue]

Hmm... Interesting...
Dr, you think that a column vendor is not bound by CGMP, dispite the fact that his product is used in a regulated environment. Consequence: "it doesn't matter what the spec says, we ship it anyway". Wow! I will propose this at our next production meeting...

Imagine all the cost savings... we can eliminate the entire QC department. Column testing becomes superfluous. The customers are testing the columns anyway. If they don't like it, we take the columns back, and ship them to another customer.

Considering that we got rid of specs, we can also get rid of these R&D people which force us to make things to spec. Heck, who needs them anyway? If we want a new product, we just fool around with the manufacturing procedure of an old product and ship it. Or we mix the stuff that does not sell, and give it a new name.

I think that getting rid of specs creates all kinds of opportunities for creative products and cost savings...

[tom jupille]

Be careful, Uwe! Some people may take you seriously!

[unmgvar]

lcguy1 your post is interesting, mainly for customers under regulation

how serious should we look at specifications?

i think that i would add to your original comment, how serious in my point of view is a vendor that claims a certain standard of quality and then disclaims it?

the reason behind all of those specs is in order to make sure of a certain level of QA that is required especially for the Pharma industry that is under the cGMP regulation but not only,even a non regulated customer such as an academic student should care of the quality of the product is being sold

i would not trust a vendor that finds it correct to give me a product that failed is own procedures. that means that he either does not care or does not understand why we require those specs to begin with.

this is especially important if you were going to do a method validation, where you need to set the specs for your method and you will do it based upon the column's behavior.
you are required to check for batch to batch reproducability, the thing is that altough you do that (and you may get a positive result this time) you know that in the end we rely mainly on the assurance of the vendors for their high quality of production that will give us the overall assurance that the columns from batch to batch will behave pretty much the same, for a method that we might need for several years to come.
today the particle size was 6.1 and it stays like that for sometime then next it will be 5.5 or 5.4 worst sudenly the product is always within their specs but i get results that will no longer be within the specs of my method what to do then?

get a full refund, go for a more serious vendor.

some are going to probably say that i am to severe and that for 99.99% of the cases that spec will not make a difference since like the vendor says the other values are within spec, and they are right most probably. but statistics are all about numbers.
how lucky do you feel today?

and what if the vendor might look and say, we got away with that spec (like DR said), we have a problem with another one, well let's just rewritte the disclaimer to include that too.

also remember that if you have a problem and that an auditor finds that this is the cause, he will not go to your vendor, he will ask you why you decided to use a product from a vendor that did not match your needs for regulation.

[tom jupille]

also remember that if you have a problem and that an auditor finds that this is the cause, he will not go to your vendor, he will ask you why you decided to use a product from a vendor that did not match your needs for regulation. (Italics added)

Not all specification are meaningful to all people. Whether the column meets all the vendor specs is (from my point of view) irrelevant. What matters is whether it meets my specs. If it doesn't, back it goes. And if I were so naive/stupid/desperate as to use it anyway, then an auditor would be quite justified in "busting" me.

If it does meet my specs, then there's no problem.

In this specific case, my question would be whether or not the exact particle size range was written into lcguy1's specification.
- If it was not, and if the column lets him meet "system suitability" for his method, then I would say he was safe from a regulatory perspective.
- If it was not, and the column fails "system suitability", then the column is unsuitable (blinding flash of the obvious! )
- If it was part of his spec, and the column fails system suitability, then the column is unsuitable (even more blinding flash of the obvious! ) .
- If it was part of his spec, but the column lets him meet system suitability, then he's in a gray area. At that point, what to do depends a bit on the regulatory agency:

USP basically says that you can use any "equivalent" column so long as you pass system suitability. Just to put the importance they attach to particle size, the "L1" category of equivalent columns is defined as "Octadecyl silane chemically bonded to porous silica or ceramic micro-particles, 3 to 10 μm in diameter." (USP 27, sect 621).

The US FDA (for internal use) says: "Particle Size (HPLC): May be reduced by as much as 50%." (ORA-LAB.5.4.5, Attachment A).

I would assume a column vendor does not have to operate under cGMP, but is most likely under ISO 9xxx. They may be in full compliance with their own procedures and SOPs by releasing the column with a disclaimer on the COA and letting the customers decide what to do.

[As a side note, I'm convinced that a manufacturer of life jackets could be ISO-9xxx compliant with a line of concrete-filled life vests, so long as they used a consistent grade of concrete and had documented paperwork that the next-of-kin could use in filing claims ]

[khw]

[ As a side note, I'm convinced that a manufacturer of life jackets could be ISO-9xxx compliant with a line of concrete-filled life vests, so long as they used a consistent grade of concrete and had documented paperwork that the next-of-kin could use in filing claims ]

This was true in the 1992 version of ISO 9xxx. ISO 9xxx:2000 requires customer input to close the loop. If you can't prove satisfaction of customers they won't certify you. (Well, at least in theory)

Your life vest example used to be my favourite argument in discussions about usefulness of ISO until voice of customer came into the game.

[lcguy1]

I talked to the customer service at Waters about this, and the product manager was supposed to get back to me about it. I still have not heard anything. Thanks for all of your input here. I think I wll just get a RA# and send them back and get colums thqt do indeed meet their specs. That way I will forgo any regulatory problems. The particle size specs are not written into our SOP's, but I would expect if a batch failed specs, it would not be used to make product. That is the way it goes here anyway.

[Bruce Hamilton]

I agree with sending them back. My perception is as follows.

The approved supplier requirements of cGMP expect column purchasers to ensure that the column manufacturer ( or any other supplier of critical materials ) has quality systems in place to cope with out-of-specification results, and documented investigations of possible consequences of specification changes.

The manufacturer produced and sold the columns to a specification. If they change the specification, they should have to justify the change internally through their quality system.

In this case, the manufacturer should have performed those internal quality processes, changed the specification, included a specific reference to the specific OOS investigation and results, or rejected the products.

As a column purchaser, the regulators will require you to show why you had confidence in the note on the COA - if your quality people haven't audited the supplier or investigated the note, you will have problems.

Return the columns, and it becomes the supplier's problem - and you can also demonstrate to regulators/auditors that your purchasing quality systems worked - a big plus, just don't show them any of the subsequent internal debate .

Please keep having fun,

Bruce Hamilton

[Uwe Neue]

Let me know if there is a problem with such an exchange. As I said, a spec is a spec...