Help!

I am looking to reduce the run time of a potency assay from 35 minutes to approx 20 minutes with no loss in resolution (some of the peaks are quite close together!)

Here is my method and conditions:

isocratic hplc, column oven at 50 degrees c(would like to get rid of), uv detection at 214nm, column c18 5micron 25cmx4.6mm ID, flow rate 0.8ml/min, mobile phase 49% phosphate buffer pH 3.0/51% ACN, run time 35 minutes

Sample matrix is 50%phosphate buffer pH 7.0/50%ACN

I am open to suggestions!

Anna,

First a few questions: Why two different pH's between the sample prep and the aqueous portion of the mobile phase? What's your current resolution between the critical pair (2 closest peaks) and how many peaks do you have? Do you know the identities and pK's of all of them?

The way I might approach this to get it done as painlessly as possible as follows: Using your current column, improve the resolution between ALL of the peaks such that you have "excess" baseline between all of the peaks. You can do this by varying the % MeCN, then type of organic (ie, switch to Methanol or some combination of MeOH/MeCN and see what effect that might have on selectivity). Switching from an isocratic run to a gradient could also reduce run-time significantly. Changing your pH might also have a radical effect on the elution order and run time.

Why do you want to reduce column temp? Are your analytes labile at 50C? If not, reducing column temp might or might not improve the separation. As a general guide, retention is reduced by 1% for every degree C increase in column temperature, so lowering your temp will likely not move you toward your goal of reducing run time. Higher temps also allow for higher flow rates due to reduced mobile phase viscosity. If your analytes are labile, then by all means, drop the temp and deal with consequences. If not, I'd probably leave well enough alone as higher temps actually aid in acheiving a more rapid separation.

Anyway, once you've improved the separation sufficiently to have excess baseline between all peaks, try a shorter column of the same type (exactly) as the one you're using. If you've switched to a gradient, remember to change your gradient times by the ratio of the new to old column lengths. For example, say you have a 10 minute gradient with a 5 minute hold at the end at 1 mL/min on the 25 cm column and you're switching to a 15 cm column. Your gradient should then be set to run over 6 minutes with a 3 minute hold. If this separation still gives baseline resolution, you've reduced run time by 40%. Now it likely won't be a 40% improvement over the original method, because the tweaks to improve resolution could very well actually lengthen the run time on the longer column. I might then go after increasing the flow rate. If you could increase your flow rate by 50-100% while retaining resolution, you'll be cutting run times further. If you increase flow rate, you'll be changing gradient times once more. If you have a 10 minute gradient at 1.0 ml/min, you'll have a 5 minute gradient at 2.0 mL/min. Reducing the particle size from 5.0 microns to 3-4 microns on a shorter column would also improve resolution (and increase retention), but would also increase backpressure, which may or may not be an issue. With MeCN based mobile phases it usually is not, especially at 50C, but one should account for as many variables as possible.

You may be able to tweak further and reduce run time even more by changing to a different column entirely, but that would involve completely re-developing the method from scratch. Granted, that isn't too far from what I've proposed already, but I've done it more than once with good results.

Good luck and let us know how it works out. If I've missed anything or given erroneous advice, I'm sure our colleagues here will put me straight!

Best,

Chris

Switch to the same column in a 3 micron, 4.6X150mm format. This column should have about the same plate count as the current column but a shorter run time due to being a shorter column. Convince the powers that be that this falls within the guideline for a method adjustment and not a modification, pass system suitability, and your done (maybe).

Of course you could redevelop and revalidate - but you would need to push a lot of samples to justify this if your only goal is to shorten run time.

The difference in pH is probably due to a basic analyte that needed to be run at pH 7 for decent chromatography but solution stability required a lower pH. Good Luck.

Tom,

She's running at pH 3 and dissolving at pH 7, therefore, I might expect that it's an acidic material rather than basic.

Also, if her peaks are quite close together as she stated, I'd predict that the assay will need significant re-engineering and re-validation to acheive her goal of a ~40% reduction in run time.

With that said, I'd likely go for a 50-70% reduction in run-time before being willing to re-validate. FWIW, she's going to be changing method sensitivity (for the better) with any reduction in column size. Would that alone be enough to qualify as a modification rather than adjustment?

Chris

I had wonderful results changing the eluation order in some samples just changing organic phase or increasing the amount of organic phase.

I reduced from 50 minutes chromatogram to 15 minutes only with a high level of gradent.

Hi Chris,

Of course you are right about the pH, I should have read her post more carefully.

The only guidance I have seen for what is an adjustment vs a modification that requires revalidation is the USP draft guidance:

pH of mobile phase: +/- 0.2 units
• Concentration of salts in buffer: +/- 10%
• Ratio of components in mobile phase: for minor
component +/- 10%
• Column length: +/- 60%
• Column inner diameter: +/- 50%
• Particle size: can be reduced by 50%
• Flow rate: +/-50%
• Column temperature: RT-5, RT+10, >RT +/-10%

My idea falls with this draft guidance - switch to a shorter column with a smaller particle size and avoid revalidation. Also, with a smaller particle size, plate count will be less of a function of flow rate so you might be able to increase flow rate to shorten run time without decreasing resolution too much.

This will not work with all column manufacturers because some columns show a change in selectivity with a change in particle size. I guess some column manufacturers use a different synthetic or bonding process with different particle sizes.

Ah, very interesting! Those guidelines leave a fair amount of wiggle-room in which to improve a method without having to completely re-validate. Would one still need to show quantitative equivalence between the two versions of an assay?

This will be most useful as I'm just getting started working on getting my little methods development lab into cGLP compliance and I have alot yet to learn about it.

Many thanks for the clarification!

Chris

At my company we would document equivalency of the two test methods, write a memo for the cGMP file. We've adjusted mobile phase slightly, and also bumped up column temperature a few degrees to about 28 C as in summer our room air-conditioning can't keep the room below that at night when our main building AC is shut off (we have some Agilent 1050s which can heat the column but cannot cool.