Help! How can I improve sensitivity?

[PMDOC]

Hi!

I'm a physiscian and PhD student, and have about 1,5 years experience with HPLC-MS/MS and method development. Lately, I have been struggeling with developing a method for quantitation of several steroid hormones (quite non-polar). My problem is to achieve high enough sensitivity so that i can measures subnonamolar concentration of these substances in biological matrices.

We have an MDS Sciex API 4000 connected to an Agilent 1100 HPLC stack. Currently I'm using a Zorbax Eclipse CDB-C18 (3.0x250mm, 5-Micron) ananlytical column.

I've been trying different sorts of approaches. One of them is injecting large amounts of sample (100-500 uL) doing "On-column extraction". I.e I inject 300uL sample onto the analytical column conditioned in H2O. Using a flowrate of 300uL/min I change the MeOH gradient to start elution. Somehow MS repsonse is not improving as much as I estimated. It seems to increase by just 10-30 % by doubling sample volume. Am I doing something wrong here, or have I just reached the maximal sample volume for this column?

Furthermore I have also tried doing an online-cleanup by first applying the sample on an Oasis HBL extraction column, wash with water and eluting to the analytical column. This gives me two problems: First, applying samples with less than 30 % ACN or MeOH seems to retain well in the extraction column, but samples consisiting of more that do not. Is this normal? And if so, is there any other way around this?

Secondly, when eluting fra the extraction column onto the analytical column I experience increased (3-5 fold) noise (compared to not using the extraction column). The peaks are broader than just using the analytical column(flow and gradient parameters unchanged). And, loading larger amount of sample (200-300uL) does just give marginaly higher respons., altough the peak RT and with remains the same.

Where I work there is knowone that really know HPLC. Is there nothing to gain by increasing sample volume? Doesn't on-line extraction help sensitivity (not for me, so far)? Anyone have a clue or suggestion of other thing I can do?

[Mattias]

Just a quick comment:

Do you have any form of acid in the mobile phase? That can improve the ionisation a lot.

[PMDOC]

Just a quick comment:

Do you have any form of acid in the mobile phase? That can improve the ionisation a lot.

Thx for swift reply!

Yes, I have! I've tried both 0.1% formic acid, 0.1% acetic acid and 0.1% ammonium acid. All produce about the same respons.

Is there any point in increasing/decreasing the amount?

[Mark Tracy]

What are the samples dissolved in? If the samples are in a solvent that is strong enough to elute the compounds, large-volume injections will result in the peak expanding in width instead of height. (In extreme cases, the peak is smeared beyond recognition.) For on-column extraction to work, the sample must be in a solvent too weak to elute the compounds.

Your observation about the online SPE is correct. Again if the sample solvent is too strong, you get no retention.

The elution strength is a nonlinear function of organic content that can work in your favor. Try diluting 100µL of sample to 300µL with water and inject all 300µL. You may get a sharper, taller peak than the undiluted sample.

[Tomasz]

What kind of interface do you use? For "quite non-polar" compounds APCI can be considered better than ESI. Your tries with acids in mobile phase suggest that you use ESI.

Best Wishes,
Tomasz

[PMDOC]

Hi again...and thx again for replying.

Tomasz: I am using TurboSpray. Do your really think APCI will give better sensitivity? Worth trying?

Mark Tracy: I have tried to dissolve samples in different solutions of ACN (67 - 10%) and MeOH (10-50%). The lower the concentrations of ACN or MeOH, the better peaks I get, allthough the dilution in the end makes it impossible to inject enough quantities of the analytes. Is there any way around this? Of course, i pre-condition my analytical column for several minutes with pure water.

Is there any limitation on how much sample volume I can put on the analytical column, given that it is diluted enough to be retained? And weak should the sample solvent be to be retained on the column during an on-column extraction?

[Mark Tracy]

In one case, with the sample in water, I injected 5 mL onto a 4.6x150mm column. Not typical. As for the limit of how much solvent you can tolerate, only you can determine that.

A trick that I used with online SPE was to inject 1000µL onto a 4.3mm dia extraction column and transfer that onto a 2.1mm analytical column. This yields quite a boost in sensitivity. But the sample matrix was tap water, so that may not work for you.

How much organic is in the mobile phase when your compounds elute? If it is more than 75% then APCI is worth a try.

If you want to go all the way with APCI, you could try HILIC-mode chromatography. Many vendors (including Dionex) have new products in this line.

[Uwe Neue]

I have several separate suggestions. These are completely different approaches that might improve your situation. However, I need to understand what the reason is for using organic solvent in your sample. Is it solubility, or does it have something to do with a sample preparation protocol before the LC

The simplest idea follows what Mark has suggested. You need to dilute the sample to a solvent composition that is maybe 10% less than the solvent composition at which the analyte is eluting. You can play with this using the largest injection volume that your system allows, and see at what organic concentration you get a nice sharp peak for your analyte. If this solvent composition does not allow you to inject the amount of analyte that you want, then this path is either limited, or you need a larger injection loop.

You can enrich your analytes on an Oasis precolumn. In this case, you do not need to do a protein precipitation or similar things before the injection. However, I need to understand you sample matrix to understnad what you need to do. However, there is an excellent trick that will help you to get focussed peaks from the extraction column onto the analytical column: you enrich the sample on the extraction column in a solvnet composition that retains your analytes completely, then wash with initial mobile phase, and then turn the extraction column around and elute the analytes from the extraction column onto the analytical column. From what you are describing, it seems that the extraction column does only a limited extraction (your solvent composition is not weak enough to capture all the analytes on the extraction column). I suspect that this is the case since you are only improving the response to a limited degree. The rest is probably breaking through the extraction column.

There is one more trick that will improve this situation completely and without question, but it requires that you change your setup, and use an additional pump. If you are inclined to play with hardware, this is the way to go. This is called at-column dilution. If you are a person that loves wrenches, let me know and I explain in more detail.

Another approach is to do all this work of sample enrichment and cleanup up front and off line, maybe with parallel processing. Again, I need to understand your matrix to give you further advice.

We have done things like this with plasma samples, urine samples, environmental waters, food etc. Don't be shy to ask questions.

[Uwe Neue]

One final thought: ultimately, the only thing that determines your detection limit is the amount of analyte available, together with details of the method. What is the expected concentration of your analyte in the (raw) sample, what is the (raw) sample volume? If you could get all of this amount into the LC/MS, will you get the results that you need? If the answer is yes, the rest is detail...

[PMDOC]

Good morning!
I am very happy and amazed by the interest you have in my problems.

What I am doing is making an method for quantitation of many different steroid hormones from human samples. The matrices is saliva, homogenized tissues, plasma/serum, urine and may be even hair. Of course I have to do some kind of sample preparation. As for blood and saliva samples I have precipitated the proteins by adding ACN and centrifuge. Besides getting rid of most proteins this also make the samplematrix less viscous, which I think probably is a good thing (although water dilution may have the same effect). After the preciptitation I have tried:
- directly inject the sample
- dilute it with water and inject
- evaporate and reconstitute in 10-50% MeOH + H2O.

I have also tried LLE (using eter or MBTB).

This simple cleanup step is why I am using an organic solvent in my sample.

I think I will try Marks idea and dilute the sample to 10-20 % less than the solvent composition at which the analyte is eluting. I guess I have to do an experiment where I inject the sample and gradually increases the amount av the organic mobilphase until the sample elutes into my MS.

I can inject up to 1500uL, which should be more than enough! The limitation is rather the amount of "raw" sample. As I am going to use this method in clinical studies I need to use as little sample as possible. For serum/saliva samples I hope to limit this to 300-400uL per sample.

Uwe, I think you are right when you suspect that the sample is breaking through my online extraction column (Oasis HBL). Therefore, I have already tried injecting sample with lower amounts of organic solvents, but still the respons is very pore. I also have a theory why.... When the enriched sample elutes from the Oasis extraction column onto the analytical column, this elution takes too long time. What I am seeing coming from the analytical column is a very broad peak. When doing "on column" extraction the peaks are very sharp and symetric. Am I right in my suspicions? What could I do to get focused peaks??? One funny thing is that when the sample I put on the extraction column is without organic solvent (like urine), I get pretty focused peaks. But if it contains only 20-30% ACN the peaks are low and broad.

I have to pumps and have tried different set-ups. I am more than willing to play with my hardware (it's fun!). Actually, my current "on-line extraction" involves turning the molbil phase flow on the Oasis, and thereafter eluting. What is this "at-column" extraction? Do I need 3 pumps?

Anyway....my best effort so far is by "on column" extraction(100uL sample)..even at very low concentration. I get very small, but nice shaped peaks. Still the S:N ratio needs to be improved by a factor of about 5-10. So I am close, but not yet there.

Tnx again for your effort so far...

[Mattias]

Regarding the acid in the mobile phase you seem to have tried it all. It should work fine.

When you inject the sample in 20-30% ACN you will get elution in the extraction column during the injection which will make the "band" of analyte unfocused.

If you need a high% ACN in your sample, you could try a column that binds the analytes harder than an Oasis column. I have used Hypercarb precolumns for that purpose. But then you need to switch to a Hypercarb analytical column also (otherwise the analytes will just rush through).

[Bintang]

Author(s): Appelblad P, Irgum K
Source: JOURNAL OF CHROMATOGRAPHY A 955 (2): 151-182 MAY 10 2002
Document Type: Review

[Bintang]

OOPS seems the subject headline does not show when posting replies.

The reveiw is about various analytical aspects of separation and detection of neuroactive steroids in biological matrices.

[PMDOC]

Hi Bintang

Thanks for the article link. I was not aware of it, but I've read it know. Very interesting. I especially found this bit interesting:

Using electrospray ionization, the best sensitivity and lowest detection limit for steroids of groups I and II were obtained with an eluent based on methanol and water only...In addition, it was also found that an addition of ammonium acetate and/or acetonitrile to the eluent decreased the sensitivity

Has anyone else experienced this..small amount of ACN or AmAc decreasing the sensitivity?

Another day has soon passed without any progress Applied a brand new analytical column and guard column without that helping much. Oh...actually, I ran one sample that came out extremely well. Nice and high peaks, little noise, well seperated. But running it again, under the exact same conditions, it was all the same again. Wonder why? This was using the on-column extraction method.

[Uwe Neue]

A saliva sample can be injected directly, maybe after 1:1 dilution with internal standard. There should not be a problem with this. The only thing that I would consider is to use a guard column which you are using anyway. Since this is a highly aqeuous sample, your analytes will focus nicely on the column. If there are still any issues left, I would recommend to substitute the column that you are using with one that is compatible with 100% water (such as the Atlantis T3 column from Waters).

Also consider that in the case of saliva, you do not need to create standards dissolved in organic. Whatever is not soluble in water is also not soluble in saliva, right?

The same story for urine, although urine does contain some amount of larger MW material. Here, I would suggest the direct injection (after addition of internal standard) onto the Oasis precolumn. Wash out the precolumn with water and 5% organic. Then switch the precolumn inline to your analytical column in BACKFLUSH mode. In this way the analytes are sitting in a nice focussed band near the OUTLET of your precolumn, and the gradient elutes then in a sharp band onto the main column, even if the retention proterties of the precolumn are higher than that of the analytical column.

For plasma/serum, a similar approach is possible, but your precolumn may need to be replaced more often due to junk building up on it. Dilute you sample 1:1 (here I really mean 1:1) with the internal standard solution (preferentially an acidic solution, acidified with phosphoric acid or something similar), centrifuge, and inject everything onto the precolumn. Wahs after the injection with 5% organic, and then proceed as above, backflushing the sample onto the analytical column.

I don't know yet what to do with hair. Need to think...