Metrology question (laboratory instrument)

OK, I’ve been fiddling with regression, ANOVA, gauge R & R, and probability distributions, and the only conclusion I’ve come to is that I have no idea what I’m doing.

I have an automated Karl Fischer coulometer scheduled for validation. I’m trying to determine some performance expectations. The device measures moisture content by electrically liberating iodine from a reagent bath, and the iodine is consumed by any moisture in the sample. The endpoint is a return to the initial state, and is detected by a voltmeter; the “titration” measures current x time => charge transferred. Because this is an exact physical relationship, there is no calibration.

The sample is a lyophilized solid, extracted with methanol. The methanol is as nearly anhydrous as we can get. However, every sample handled is subject to absorbing moisture from the atmosphere, as are all standards. All samples are injected from a precision syringe, but are weighed before and after – this weight is used as the sample size in calculations.

Procedure: (1) titrate standard ampoules with certified water content; (2) titrate methanol blanks; (3) titrate control sample (product lot with test history and control chart); (4) titrate samples.

(1) and (3) are performance checks - if the results don’t fall into range, the test is invalid. Tricky part is (2) – this blank value is used to adjust all results from (3) and (4).

Really tricky part: The certified standard contains 100 µg of water per 1 gram sample. For various reasons that don’t matter here, I can’t use less than 0.5 grams of the standard for a validation test - hence, the lower limit of my ability to confirm the instrument’s performance is 50 µg. The blank, the control, and the sample typically contain 35 µg of water per 1 gram sample. This is OK for the sample, because the spec is only that the moisture must be less than the limit – which is 50 µg/g.

What I’m trying to get at is the probability of passing a bad lot, given the variance of the standard measurements I make.

Bonus trickiness: the control chart for the control samples looks horrible – for a lot with mean = 35 µg/g, the standard deviation is about 10% of that. It’s the nature of operating at the limits, but YUK.

I think you want GQ, not elections.

According to this,

Yes, I know that. Nonetheless, there will be a standard deviation, so…?

BrainGlutton, :smack:

Anonymous Moderator Who Moved This, Thanks. I shouldn’t post late.

Have you determined your test/uncertainty ratio (TUR)? If so, it’s just a matter of what your acceptable false-accept/reject ratio is. There is a formula that will give you test limits based on this information. Is that what you are asking?

I’m not really sure I understand the question - is the issue that you are worried that you are testing things in the 35ug range, but can only validate compared to a 50ug standard? In other words, you have no data to show that readings below 50ug are valid?

You could consider modifying your KF procedure - there’s no reason you couldn’t standardize with 25ug of water; just use lab-grade DI water instead of a prepared standard!

Following USP <921>, the pharmaceutical labs that I worked in calibrated our FK machines using the validated deionized water from the lab source (though you can simply buy bottles of DI water from VWR or Aldrich or wherever).

We used small volume Hamilton syringes to dispense 25ug of water whose temperature was known (therefore it’s specific gravity) by volume. The method was validated through a typical repeatability and robustness protocol and the linearity of the method was shown over the ranges the lab was interested in.

I wish I could remember the exact laboratory procedure, but it’s been 6+ years since I did a KF titration!

ETA: the method required a new calibration every 2 hours…if testing took longer than that, we’d have to stop and recalibrate compared to the water standard. The validation protocol showed that this acceptably dealt with humidity issues.

Thanks, mnemosyne – yes, that’s the problem. Unfortunately, neither the sample size nor the method are subject to change. Besides, I’m pretty sure Hamilton doesn’t make a syringe that can accurately dispense 25 µg – 25 µL, yes, all the way down to 0.5 µL, which is 500 µg. Using 5% of the syringe’s volume is not good enough.

nate, thanks for your reply. I don’t think TUR is going to be terribly useful in this case; I’m not calibrating anything, I’m trying to demonstrate the fitness of this instrument for the method. I have nothing like a calibration standard, and the water standard is certified to +/- 4%.

And I forgot the other thing that sucks; for those who don’t know, these reaction cells do not seal perfectly either, so there’s a constant drift of 2-4 µg/minute.

Wow how did I miss that? It most definitely was 25 uL we were dispensing, not 25 ug. Sorry, it’s been a very long time and I don’t remember things all that well!

What kind of titrator do you have? Is it one of the Metrohm ones? I think the one I’m most familiar with was the (googles) KF701 Titrino, though I did use both a volumetic and Coulometer model, so my memories and advice are probably not all that helpful. Have you contacted the manufacturer? Their validation/QC specialists can probably help you. Metrohm itself says their coulometers can measure 10ug of water, so there must be some way for you to validate that!

Is there a reason you can’t scale up the sample size being tested - IIRC our sample sizes were generally about 2.5g of powder.

If your 1g of standard gives you 50 ug of water, then you can toss in 2g of sample and measure 70 ug of water. That way you know that your samples are above the lowest value you can standardize against (and would fall into some linearity if you were to calibrate using higher concentrated standards).

Why can’t you change the method? Has it been fully validated, or is that what you’re doing? Repeatability and robustness criteria should take care of your concerns, no? I realize a new method validation is time-consuming and annoying (and a PITA paperwork wise) but I’d rather take the time to correct a flawed method than continue with questionable results.

I know we never really found a way around drift… you just had to trust the reading and recalibrate often.

Man, I’ve been out of the industry all-together since 2007, and rarely used Karl Fischer in that last year. I used to like it as a test, though…well, as much as I liked any lab work…which is why I quit :wink:

Well, since you ask:

Yes, it’s a Metrohm. I have asked how they justify that 10 µg claim, and have not received a response yet. They’re not very fast about that sort of thing (well, about anything, really).

Samples are extracted in the single-dose vials they were manufactured & packaged in; the lyophilized powder in one vial must be exactly one dose, and is allowed to contain NMT a certain amount of moisture, period. There’s only so much solvent you can add to a vial, and only so much of the solvent you can reliably extract from the vial.

Coulometry is the method of choice for measuring residual moisture in this range – the volumetric method is way too unstable and imprecise – but working with such small amounts means you’ve got to be insanely fussy about exposure to humidity.

Interesting. We regularly combined several tablets/capsules from a single lot to attain a measurable amount of moisture. The FDA had no problems with it, because the point is to test the lot, not the individual dosages (at least, it was for us!).

I wish I could help more!

I don’t suppose using the KF oven attachment is an option? It deals with atmospheric moisture and is much more precise, but can be a real pain in the ass to use. It doesn’t address the calibration range, but does address some of the humidity issues.

Good luck!

No, changing the method and equipment are not an option.

The problem with pooling samples lies in the difference between oral dosage forms that are supposed to be more-or-less stable with respect to moisture, and a freeze-dried parenteral product that is practically designed to absorb moisture instantly. Also, it’s already in the vial, which is never opened throughout the procedure.