Fosfomycin analysis with amino column problem

[qfiviera]

Hey guys. I need help. Don´t know what else to do.

I´m tyring to analyse Fosfomycin Tromethamine using USP method. In PF 36(2) says:

BRIEFING
Fosfomycin Tromethamine. Because there is no existing USP monograph for this drug substance, a new monograph is being proposed adopted from the official European Pharmacopoeia 6.0 monograph. The liquid chromatographic procedure in the Assay and in the test for Organic Impurities is based on analyses performed using a Zorbax NH2 brand of L8 column. The typical retention time of fosfomycin is about 9 min.

Conditions: MP Potassium phosphate buffer, monobasic. 10.98 g/L, Flow 1.0 mL/min, detector RID, Temperature (column and optical) 35 °C. My HPLC is an Agilent 1100, and made preventive maintenance 3 months ago and replaced MGCV because it wasn´t working well.

First, I didn´t have the Zorbax column, so I used a like new Luna amino column (less tan 50 inj.). Didn´t work well, broad peaks and unresolved impurities because of tromethemine broad peak at front. Tried to clean it by phenomenex cleaning procedure, and didn´t worked. same chromatogram.
So I purchased Zorbax column. First peak came at 13 min, not at 9 min! But were nicer, thinner and peaks in front were resolved. So I changed Mp, new batch, and same problem. I made a little sequence and peak started to shift, and elutes earlier moving in each injection, always less rt, and starting to see a loss in performance, because peaks started to broad.
Well, I checked the equipment, did a pressure and leak test and didn´t pass. Called the service, and changed the outlet ball valve. The instrument then was OK, pressure and leak passed.
But the problem persists, and now the peak elutes at 7.2 min and decreasing and broadening.

Things already tried and didn´t worked:
Bypass the multi gradient valve
Changed the guard column
Using no guard column
Wash the column like agilent says (twice, first using only ACN, the second try was using DMSO too)

Now I´m getting back to Luna column and see if rt is staying the same or happens the same with Zorbax and starts to move and elute earlier tan before and never stop! Perhaps is the instrument and I´m not seeing it! And next I´m doing the QC test like the one Agilent did with the Zorbax to check performance, because I can´t believe A new column lasted less tan 50 injections.

BUT PLEASE, ANY IDEA!?!?!?!?

[qfiviera]

Forgot to say that excipients are all soluble in water (sugar, lemon essence, sodium bicarbonate, sodium citrate and coloidal silicic acid), and sample is prepared in MP.

[Kreall]

First of all, the monograph in EP are for analysing substances not drugs. They need to be validated to see if they are appropriate for analysis drugs.

The amino columns are a bit tricky. They need a very long time conditioning and they are very sensitive. As you wrote I think the excipients will affect your separation.
There is also a possibility that if the pH of mobile phase is not acidic, than the enolisation of sugars can occur to give aldoses and ketoses which will react with amino group in column to form Schiff base. Less the stationary phase worse the separation. In this case KH2PO4 will give an acidic pH so propably this is not the case.

I propose and experiment:
Take 5-6 injection of drug sample and than inject 5-6 standard samples. If the separation is getting worse with every injection of your drug sample and than when you inject couple of standard solution the separation is getting btter than it is defenetly excipients reacting with/destabilising the column.
If the separation is getting worse with standard sample injection, try to do column loading - inject like 5-6 in every minute and see if it helps.

[qfiviera]

Thank you for your quick reply.
First of all, this is happening to me because I´m trying to validate the product. because there´s no monography, only for API. Second, this is happening with the standard, without excipients. This was my mistake when I posted the second post, so I don´t think excipients are affecting the chromatography. I´m thinking if there is an interaction between the epoxide and the amino group from the column.
I have just finished injecting 6 samples of standard in the Luna column and the rt and performance didn´t changed, so I discarded a problem with the equipment. I think it´s the Zorbax column the problema, but cannot understand why. Why could be reacting or interacting with Zorbax and not with Luna? Why the Zorbax column is useless after 50 injections and Luna doesn´t? Why rt in Zorbax keeps getting lower and in Luna not?

[Kreall]

Don't know about the Zorbax but Luna it is a packing that's hard to kill.
You have got a interesting topic for an article to write. "Stability of retention time on different amino column"

I was wondering what would happen if you will add a little procent of organics in to the mobile phase, like 5% of ACN or MeOH.

[qfiviera]

Thinking the same, gut my boss does not like that much the idea. We both agree that if it is a pharmacopoeial method, we should be able to reproduce it. We purchased the same column! And is not allowed by USP, but we are going to validate, so perhaps we try....
Yesterday I made 6 injections in Luna, and the rt didn´t change. Now I´m going to purchase the standards to do the QC test in the report from Zorbax column to see what happened before going further. We did all that the column care manual says, conditioning, solvents, pH, temperatura, etc. I can´t believe it lasted less than 50 injections. If it´s voided, I´m calling my distributor!

[badla]

Zorbex Amino columns have a serious issue with the retention time, the retention time keeps decreasing. There is a new column Zorbex Carbohydrate I believe, however you will observe the same issue. I would suggest you to use Phenomenex Amino column. Please read the documents for the column. These columns are equilibrated with non polar solvents, so you will need to gradually equilibrate the column from non polar to polar solvents. Once equilibrated to polar solvents, further equilibrate the column with 50% of Acetonitrile and 50% of DI water, followed by your mobile phase.

After each run, make sure to clean the column with DI water (high salt concentration in your mobile phase) followed by through equilibration with 50% Acetonitrile. Also important to clean your seals and needle thoroughly with water followed by 50% acetonitrile. the salt accumulation causes severe trouble with your LC system. And it is highly impossible to regenerate an amino column, don't waste your time.

One other very important thing with amino columns is do not under any condition loose prime or inject an empty vial, an air bubble will lead to sever damage and expect to see broad and ghost peaks. More over air bubble into your RI detector will lead to unknown peaks.

Also verify the filters on your detector, if you are seeing excessive tailing. Minimize the time to 1 sec or less if possible.

One other important thing with this product is the sample temperature, not sure if USP mentions but try to set up your sample temperature at 5°C and prepare your sample just before injection.

not sure if you have already figured this out but thought this would help you, let me know.

[qfiviera]

Thank you for your reply badla.
At the beggining of August, we´ve purchased an amino column and validated the method. We´ve observed the same issues as the Zorbax, the loss in retention time but not that much.
All the samples are injected inmediately after prepared because we do not have a cooler in the injector module.
About washing, we did it, but did not show significant improvement. Tailing is not a big problem, but shoulder at the end of the peak is, and it is a sign that the column is "dying".
About RID detector, we took care of all possible precautions. Air bubbles I´m pretty sure is not a problem.
I really would like to see all the data analysis from the people who validated the method, because I still can´t believe they used a Zorbax Amino! In general we do not have problems with USP monographs, but this one is a headache, it is not robust at all.

[badla]

I believe there is no point struggling with the zorbex and did Agilent help you resolve this issue I tried getting help, but Agilent was surprised to know about the mobilephase and then never came back or I should say I did not want to waste my time following up with them and why not change to a different brand, we used phenomenex and my RRT's were comparable to EP or USP monographs and I was able to use the column for about 500 injections on an average, I have validated both the API and finished product methods, I have used over 5 columns and did not have difficulty in attaining the same RT values however if I loose prime or inject an empty vial, the next minute the column is of no use. Good luck

[qfiviera]

And did you have problems with rt, decreasing? In my previous post I didn´t mentioned that the column we bought was a Luna amino. Happened the same, but not so quickly as Zorbax did (Zorbax lasted less tan 20 injections). So you are telling me is it crucial to wash the column between injections? How much volume? Because a batch of runs seems to last forevever!

[badla]

Well you do not have to wash the column between injections, I believe the run time is about 20 minutes using a Luna Amino but yes between runs make sure first with water followed by Acetonitrile and water 50:50 for an hour.

Once the column is completely cleaned for any nonpolar solvents and equilibrated with the mobile phase I have not observed significant RT changes, Fosfomycin was consistently eluted at about 12min, Tromethamine peaks at about 2min, Imp A at about 10 min, Imp B at about 3.5 and ImpC at about 4 and D about 14.5 -16.(D was hard to detect).

I observed peak shifting from 15 down to 13 the first time and then realized the column actually comes in equilibrated with non polar organic solvents (I believe it was Hexane).

Try checking with a new column, start cleaning with TFA, then with IPA, followed by Acetonitrile and water (50:50). With water acetonitrile I would say wash for about 3 hours followed by your mobile phase for atleast another 3 hours, I used to put my column equilibrating overnight at low flowrate with the mobile phase. And the DI water acetonitrile cleaning really helps.

[qfiviera]

Reading your post makes me know I´m not alone in the world! You described the same trhings that happened to us. Same rts, imp D difficult to see, column overnight stabilisation, decreasing rt from 15 to 13.....
But what about washing? Because we do a similar wash but after a batch run, supose, depending on what part of validation, about 20 injections. How often do you do it? After how many injections?
And what about noise and peak shape? We compared noises and peak shape, Zorbax vs Luna, and Zorbax exhibited better peak shape and less noise (same day same mobile phase), but of course, Zorbax after few injections....