Diffrence in retentions time

[mk12]

We contracted some stability studies to one CRO and gave them method . Retention time of analyte in method is 15_+1 min but they had different rentention time 19 min. They used the same chromatographic conditions( including same column).

According to them it could be due to dwell volume because they are using diffrent HPLC then us (agilent).

few data has already been done, Which can be re run..what should be the solution to this problem.

I was thinking if I can give them my forced degradation sample and compare the method that it resolved all the impurties and RRT are same as in house run.

Than data can be valid or not?

What could be the best solution to this problem

I would appreciate your help.

[Uwe Neue]

You did not say that the methods are gradient methods. If thye are, the gradient delay volume or dwell volume may be the problem.

You can measure the dwell volume on both systems to verify. There have been posts around giving you instructions on how to do that.

[mk12]

Thanks Uve My method is gradient ...but my question is... data which are already run and can not be reproduce again How can I make those data valid because system suitability for Retention time does not meet. If I show that difference in retention time does not effect method performance means..it still resolves all the impurities from the main peak and method is still linear ( which we do have), will that make my data valid?..if yes what is best option to compare method performance.

[Kimico]

I agree with Uwe on the gradient delay. You can address the problem as follows:
1)Evaluate Vd on your system and their system. If there is a significant difference, learn if their system has unnecessary tubing before the column. If their system allows, optimize their Vd.
2) Another alternative would be to adjust the injection time on their method so that the gradient hits the column the same way or at a similar time it hits on your system.

As for your question on the validity of the results already obtained, it seems to me you are working in the regulated industry. If you have a method that is written out and approved by the regulatory authorities and you said you were going to jump on one foot, then you must jump on one foot. However, this is not to say that your results are not accurate as you are trying to demonstrate in a scientific fashion. If your stability samples did not meet the system suitability spelled out in your written method then your results are invalid. You could do a risk assessment as far as accepting the results.
Additional questions though; were they running accelerated or LT stability? How deep into the study are they now? t=1,2 or n months?

[mk12]

Thanks Kimico...

Its a long term stability study upto 12 month...and only one month data is done till now...we need to fix this issue before time runs out.

[tom jupille]

The dwell volume (gradient delay volume) is chromatographically equivalent to an isocratic hold at the beginning of the gradient. So:
- if you have more than one peak on the chromatogram, are they all offset by the same time? (peaks near the beginning of the gradient can be ignored for this purpose!). If not, then the matter is not simply one of dwell volume differences, and you have a real problem on your hands!.
- if all the peaks are offset by the same amount, then this is good support for the dwell volume difference hypothesis. As suggested earlier, measure the dwell volume on the two systems and see if the difference in dwell times matches the observed time offset. If it does, that pretty much confirms the hypothesis. At this point, it's a question of what you have to do to placate the auditors/regulators. Your choices are:
1. Write up an SOP that justifies correcting retention times for dwell volume differences. This is dangerous because early peaks can be affected differently (i.e., it can only be applied to peaks that elute well into the gradient)P.
2. As suggested by Kimico, add an isocratic hold (or delay to injection) to the method on one of the systems to normalize the time delay between injection and arrival of the gradient at the column.

[Bruce Hamilton]

Before spending too much time comparing systems, you should point the whole issue at your quality group.

They may want to (re)audit the CRO to discover why the work programme progressed, and may require a restart to the programme. You definitely should obtain Quality sign-off for changing the conditions to achieve suitable RT.

As Kimico notes above, it's going to be hard to convince regulators that "after the fact" approval for the retention time difference is acceptable, even if it could be shown as a dwell volume issue.

Good Luck,

Bruce Hamilton

[mohan_2008]

Typically, if you have a good stability indicating method - resolving all impurities and appears to be robust,

I do not see an issue. Some times, besides the instrumental conditions, organic modifier % in the mobile phase can be really sensitive to the compound and affect its retention time.

And, also assuming the compound pKa is pretty close to the mobile phase pH ( A possibility!).

We need to identify the situation causing this drift. Assign a possible cause for it, and prove that there is no quality impact on this assay due to the RT drift. This can be attached as an amendment to your method protocol.

However, all this needs to be duly notified to the chief chemist at the other CRO's and obtain his approval.

[mk12]

I talked to my CRO already they are not going to fix their dwell volume even if thats the issue but my plan is If I forced degrade one sample and analze the same sample in our lab and their lab and compare impurity profile,RRT and peak purity. If their method resolve all the impurity like our method we can show that retention time drift does effect method performance. Is that the right plan?

Please guide

[Bruce Hamilton]

That would show both systems detect the same peaks, and presumably the validation package is OK for other critical performance criteria.

If your quality group agree that the extra comparison is an adequate procedure and explanation, then it's all go...

Good luck,

Bruce Hamilton