Recent years solution for higher throughput

[unmgvar]

As laboratories strive to increase throughput, the industry has come up with different ways to answer the needs of the industry.

I wanted to start a general conversation in order to better see the pros and cons of each solution, based both on manufacturers and end user experience.

the different ways that I know of (feel free to add to the list in the future) are as follows:

1. simply use the same hardware technology, but improve existing methods. Here, softwares like DryLab and Chromsword are a very efficient tool for better achieving faster and improved methods.

2. Use Monolith technology like the Chromolith columns from Merck, of increased flow rate for shorten run times.

3.Use of smaller particle size columns, which brought about the introduction of the UPLC system Acquity from Waters.

4. "Moving back" more accurate I think would be "rediscover" TLC technology with the use of CAMAG hardware and HPTLC plates, or OPLC thecnology.

5. Slightly modifying existing hardware and using the dual pump system from Dionex, the SummitX2 system.

of all those 5 points 4 require method validation. I might talk about them in future replies but I actually want to start with the 5th point which to my point of view requires GLP adaptation for correct use.

The concept of the Dual system from Dionex is quite simple by concept:
take existing hardware and improve its throughput capacity.
the "problem" is that this throughput is achieved by a tandem work of two columns.
I can see the pros of that concept for R&D labs, but I did not found an answer to GLP issues regarding the use of two columns in tadem in a QC lab. And as we know in the end the bulk of throughput effeciency has to be achieved in the QC.

the column is the heart and soul of the system. would an anayses performed in tandem be GLP compliant at all? in order to make it compliant what would need to be changes in lab SOP's. would those changes cause such a loss in throughput in such a way that the gain would be insgnificant compare to the investment?

Does any one out there has any ideas?

[Alex Buske]

Hi,

Just a few thoughts: the Summit X2 will not improve performance with isocratic method.
Second: For many applications system suitability tests are required. In that case the system is defined as "the hardware" and "the solvents" and the column.
Having two columns leads to two SSTs...

Alex

[gemm]

there is another method to increase througput. purchase more systems. but use identical ones like you already have. this minimizes time consumption for validating or developing new methods and SOPs.
considering overall cost, this might be an option.

[unmgvar]

Alex,
you are right this solution will not help for isocratic systems, yet there are enough gradient methods used (especially for impurities) by all companies to make use of the idea.
in GLP environment all work requires SST's. but can you with this system perform a new kind of SST.
for example:
you would have a method that requires tailing, resolution, plate counts, RSD. you would also check for acceptance of std with a second one. your SOP's would require you to perform ID injections for all available known compounds. let say that we need to inject 3 of such ID's. let say that the run time is 30 min with 10 min of equilibration.
one way to look at it is to say what ever we used to do with one column we must do with 2. in that case there is no true saving in time. worst unless the sequence isn't huge then it will take more time.

but if we do look at it differently then maybe we can cut on some of the injections.

the first requirements would have to be for both column evidently tailing, resolution and plate count are intrisic to the column.
but the replicate injection RSD, and the std acceptance would be done with one column.
one column would aslo do the ID injections. for time saving matters it could be the second one, and extrapolation would be done on the secont for RT of peaks.
one column would closely check RSD of std through the entire work and on the secont column i would have more sparcely injected std.

over all this would give a good saving in time.

for example if my work is overall 25 injections (all included) for one pump, then with the dual it would be 27 (too long to explain why right now but it is based on the assumption aboves and in correlation with my SOP's requirements).
25x30=750 minutes
27x20=540 minutes.
i have saved 210 min. (3.5 hours) 28% increase in throughput.

and as Gemm "pointed out" i would save on a second instrument (the summitX2 costs only about 24% more to a new normal HPLC), further more i would save on lab space (about 12-15 meter square, which include all, table, chair, pc, printer, working space...) and i would not need to hire and train more personnel as well.

on the grand picture it looks swell, really, but is it GLP viable.

[Alex Buske]

Great throughput increase!
Assuming the technician needs more or less 8 h to preapare solutions, write methods, examine data and do all the documentaion and assuming further that she/he starts the sequence at 4 p.m. then I cant see any performance throughput: the last run ends either at 4:30 am. (with 25x30 min) or 1:00 pm (with 27x20 min).
Basically I like the idea of having a second pump and I certainly see applications: I screening situations one might really utilize the throughput increase, especially with more expensive detectors.

Alex

[unmgvar]

Alex,
you are right there is always the possibiliyt that the system is less then efficient and will bring no relevant output and especially will save time.
on the other hand this is true with every solution.
it also happens to us right now without the use of novel technics.
still the reason that vendors are developing new ways to increase productivity is because customers have a need for it and create a market for it .
i have not gone all the way to make all the numbers stick in my example because i am not a payed Dionex employee and it is not my task to promote their product on this site. i only gave a general example.

anyway, what do you think on the GLP issue?

[Alex Buske]

Unmgvar,

I would think, your GxP concept is right, you might need a special SOP. From a practical point we often use calibration runs for the system suitability test (repeatability of peak areas and retention times) and with two different columns there could be issues on the retention time deviation.
Next point: If you run a a really complicated related substance method and retention differs on both columns just a little bit. How would you defend that the peak with the retention time 2.05 min is labeled as a degradation product in one run and as a excipient (and thuse ignored) in the next run?

Having two pumps in one is a good idea (certainly better e.g. than parallel columns an these multiple-inlet-MS-sources).
There are other products that could speed up the data aquisition: multi-lambda fluorescence dectors, highly sensitive PDAs, ultra-fast autosamplers (with build-in sample prep) easy- and save-to-use CDS. They just dont work together and we end up again with last centuries configuration. From my point of view the greatest potential is the use of robust methods and simple sample prep. The bottleneck in the laboratory throuput are manual operations (sample prep, eluent preparation, data evaluation etc.), instrument time is less restricted.

Alex

[unmgvar]

"and with two different columns there could be issues on the retention time deviation.
Next point: If you run a a really complicated related substance method and retention differs on both columns just a little bit. How would you defend that the peak with the retention time 2.05 min is labeled as a degradation product in one run and as a excipient (and thuse ignored) in the next run? "

I agree with you, i am also pondering on the same matter.
if not specifically said before, this tandem work is not optimal for all existing possibilities, like isocratic work or small worload.

the simple solution to your point is that duplicate injections must be made on the same column. also you have to make sure that all sample are injected on both columns after you have passed a repeatability test with one of the columns. all this evidently requires a smarter organistion of the sequence run.
still i figure that you can this way utilise about 95% of the "wasted reequilibration time" of a gradient injection.

the system is also interesting because as you remarked, a lot of time goes into sample preparation. Dionex also offer the use of a roobotic autsampler, which will further cut down on lab work.

as for a solution to isocratic work, my heart is actually set on the HPTLC approach, right now, especially for dissolution, uniformity and some types of impurity testing (mostly go/no go type)
The camag product line is good. also their scanner has screening capability, and thus you introduce a PDA in the QC. this in it's self is a plus.