nonmonotonic dissolution results using Agilent HPLCs

[kayjbee]

Hi

Has anyone experienced nonmonotonic dissolution results when using Agilent HPLCs. Our experience is that they occur randomly and cannot be replicated. Colleagues have not observed this on Waters alliance or acquity systems. Suggesting it is only Agilent!

A nonmonotonic result is when an anomalous result is observed eg. the 30 minute timepoint below is nonmonotonic

Timepoint ---- % Dissolution
10 -------------- 45
20 -------------- 75
30 ------------ 116
45 ------------- 100


On reinjection the anomalous result has changed and follows the dissolution profile
Timepoint ------ % Dissolution
10 ------------- 45
20 ------------- 75
30 ------------ 99
45 -------------- 100

[HPLCaddict]

I don't think that this is a general problem of Agilent HPLCs. Are there any differences beteen the procedure you and you're colleagues are using?
Are you using 6 vessels for dissolution, resulting in 6 single values for each time point? If yes, what are the variances between these single determinations?
Do you use SST to verify your HPLC is suitable? Single or double injections of samples/references? Regular OQ/PQ?

[kayjbee]

To answer your questions, the procedure followed is the same between colleagues.

6 vessels have been used for the dissolution analysis so yes we have 6 individual values for each dissolution time point. Typically 5 of these values will show good agreement and be in trend with the dissolution profile and one vessel will have a non-monotonic data point

We do run a SST prior to the analysis and also throughout the Empower sample set when analysing the dissolution samples and these always pass criteria. We only perform single injections of samples and standard solutions. The LC's are calibrated on an 12 month interval

This non-monotonic behaviour has now been seen on 4 different Long Term Stability Study projects with no common theme being identified as a root cause. The non-monotonic results occur fairly infrequently (reviewing the latest LTSS data, 30 instances of non-montonic behavior observed for dissolution analysis of approximately 500 sample sets - each dissolution sample set has 30 sample vials injected, 5 timepoints, 6 vessels) and are completely random as to when they do occur - they cannot be replicated as re-injection of samples gives expected dissolution profiles.

Analyses performed on the Waters LC systems have not produced non-monotonic data which leads us to think it is Agilent LC related. This is a big problem for us as 90% of our LC's are Agilent 1100s

[unmgvar]

Do you inject the vials in order of dissolution, going from the lower conc. to the highest?
in those applications that you get the problem, how long are the runs?
are you injecting 100% buffer solutions, or are you diluting with organics as well?
what are the injections volumes of those methods?

[Peter Apps]

You have 30 aberrant results out of 15000 analyses ?

This a 0.2 % rate, and depending on your repeatability / reproducibility these might just be in the tail of the normal distribution of results, rather than an indication of system failure.

If most of your instruments are Agilent presumably nearly all the analyses are run on Agilent, which is why you have not (yet) seen an aberrant result on a Waters.

Peter

[kayjbee]

Vials are injected in order of dissolution from the lower conc to the highest
Each sample injection is typically 5-6 minutes, with each sample set consisting of standard solution injections interspersed throughout the run separating each set of dissolution time point samples.
The injections are not diluted with organic and are simply samples in dissolution media - the injection volume for the current project is either 40 or 15ul (it is a dual active tablet).

This current project is a joint venture between two companies and company B do not have any Agilent LCs only Waters and they have never observed this behaviour on this or any of their previous projects.
As I have stated earlier we have now observed this problem on 4 different LTSS projects all running on Agilent LCs and this problem has now been seen at a additional site also running Agilent LCs

[unmgvar]

have you seen the behavior after with a standard injection as well,
or right after a standard?
are you using the vials solution option of the agilent in the sampler? if yes is it closed with a septa or not in the cap?
in some cases, especially in the development i know of companies using some very "interesting" solutions for longer dissolution times, that can very easily clog the column and as well server to create carry over in the systems.
there are several types of polymer used that especially are nasty.

[carls]

are the non-monotonic results always high or are they random?

Are these Agilent 1100, 1200, 1260 or a combination?

How many of these instruments do you have?

[kayjbee]

This behaviour has not been seen in a standard solution injection. It does occur randomly throughout the sample set so doesn't always immediately follow a standard injection. The results are always high

The LC vials are closed with a septa in the cap.

The Agilent LCs are the 1100 model. We have approx 25 of these systems

[Peter Apps]

Anomalies always being high nearly rules out random variation.

The only thing that I can think of that can cause random, rare and always high results that cannot be repeated is the presence of particles of the analyte in the injection solution. Given that these are dissolution trials this would be due to a failure in filtration.

Given that particles tend to settle, the difference between Agilent and Waters might just be in the way that the sample is removed from the vial by the autoinjector, or even how the vial is moved around just before the sample is withdrawn.

Peter

[Peter Skelton]

To add to Peter Apps' point, Agilent systems use a bottom-draw needle whereas most Waters systems have side draw needles, so there is a potential reason there.

[DR]

Or something is building up within Agilent systems that does not tend to build up in the Waters systems.

How often are the Agilents passivated?

[StephenO]

The injection mechanism of the Agilent are different from the Waters. Agilents use an injection technique that pulls 0.6ul of air in to the tip of the needle from inside of the vial near the septum, referred to sometimes as the "leading bubble technique" its used to decrease band broadening and dispersion during sample injection.

Due to shaking of the vial for example and the high surface tension of aqueous buffers and extra layer of solution forms in the neck of the vial. If this was happening you could be injecting a larger volume of sample than you think thus giving the nonmonotonic results.

This could be a reason for not seeing these results in the Waters systems.