Theor. Plates as a System Suit and Method Transfer Criterion

[KarenJ]

Hi,

Is it standard practice to include a Theoretical Plate (N) criterion in the System Suitability for an HPLC method that has a single main peak? If so, what limits are typically set? What other parameters are typically used (eg. %RSD, tailing, k', etc.)?

Also, when transferring a method, what type of Theoretical Plate and/or retention time criteria are set in the method transfer protocol?

Thanks,

KarenJ

[Dan]

Hi Karen,

My reply is mostly based on working in a regulated environment that follows USP and FDA.

The USP<621> states that there should be three measures for system suitability: reproducibility (usually determined by peak area RSD), tailing and resolution. The USP does allow for the use of theoretical plates (N, column efficiency) as a substitute for resolution, especially if there is only one peak. However, the USP points out that the column efficiency is a less reliable means of determining resolution.

A dissolution method for a product with only one active may use theoretical plates as a criteria instead of resolution. If resolution was used, then you would need to prepare a suitable resolution solution for the analysis. The company I work for prefers us to use resolution, even if that means making a resolution solution when there is only one analyte in the method (or standard solution).

But, we do use theoretical plates on occassion. In general, we set the value for the limit as 80% of the lowest theoretical plate value that we observed during method validation.

As to method transfer, we use the same system suitability criteria that are in the method. We don't add/change anything for a method transfer.

(Side note: the USP has postponed the addition of a 'detection sensitivity' system suitability requirement. This requirement was to apply to impurity methods. It was postponed as methods in older monographs may not be able to meet the requirement. However, it has been recommended that the detection sensitivity requirement be added to new methods.)

Regards,
Dan

[tom jupille]

My interpretation of the official (USP, FDA, ICH) system suitability "suggestions" is that they are minimal. I see nothing in the documents that precludes adding other criteria as necessary to ensure that the system is suitable for the analysis. It is up to the developer to specify (and justify) those criteria.

Some possibilities:
- retention time (or, for isocratic separations, k') windows for key peaks
- resolution between key pairs of peaks
- tailing/fronting
- detector baseline noise (for trace analysis, this is the major determinant of LOD & LLOQ)
- plate number (if you have specified retention and resolution, then plate number is redundant; I usually throw in a value just to keep the auditors happy)
- system pressure (this can provide early warning of impending column problems)
- repeatability of replicate injections
- linearity of the calibration curve over an appropriate range (you can get into whole discussions about which statistical measures to apply here )

I'm sure people can suggest more.

[Yorkie31]

Probably one of the other things you really want to consider is whether your method is gradient or isocratic.

With a gradient elution, you are essentially changing the efficiency of the analytical column, as such I would really not use this as part of an SST. As it is inpractical to measure.

For a single peak assay in isocratic conditions, I really only use this as additional information as part of the SST. If you have good asymetry, system precision and stardard verification, that to me is justification enough that the system is performing correctly.

Sure, effieciency is a good measure of how well your column/system is performing... But why shoot yourself in the foot?

Anthony

[tom jupille]

Yorkie31 makes a good point: you cannot apply the isocratic formula for plate number to a gradient separation and expect to get a meaningful number. Simple baseline width (or even width-at-half-height) is better in this case.

[Rande]

It depends what type of column you are using for the assay.

Theoretical Plates (or HETP - Height Equivalent Theoretical Plates, normalized for column length) is a method to evaluate how well packed the column is. To summarize the test, you inject a single "slug" of a compund that does not interact with the column and will come out with the void volume (Vo). Then measure how much it "spreads out" as it passes through the column. This is very important after packing a column - especially size exclusion.

If the analyte interacts with the column or you are running a gradient method, this calculation is NOT applicable.

For analytical size exclusion (SE-HPLC) it is one of the important system suitability tests, but not the only one. Also, increasing values is not always a good thing. I have one application where the plate counts increase as column performance degrades.

[HW Mueller]

Rand, is your contribution actually limited to SEC and you are talking about using a substance which is too small to be excluded, but gets (partially) hung up on the column? I have a problem, for instance with RP, to comprehend the characteristic of a compound which does not interact with the column, but spreads out away from to (tm).