Why doesn't Pharma use GC?

[Victor]

I'd be interested in hearing some opinions on this question. While GC seems to be used in the pharmaceutical industry for determining residual solvents (which would be difficult if not impossible to determine by HPLC) there seems very little use of GC in the pharmaceutical industry to determine APIs or degradation products. When you look through GC manufacturers catalogs, the peaks for many polar drugs look excellent, and the efficiencies are much higher than in HPLC. There is a predominance of use of GC for analysis of drugs - I guess becuase of the ease of coupling with MS, and the definitive nature of the spectra produced. The answers to this question might be:

1) It is difficult to inject water onto GC columns (but this is not at all impossible).
2) the higher resolution of GC is not really necessary in pharma, because a typical formulation contains a limited number of ingredients (compared with something like an environmental analysis).
3) The quantitation aspects of capillary GC are seen as more suspect than HPLC. In split and splitless injection at least the whole sample is not directly introduced onto the column as in HPLC.
4) Many problems with polar compounds on GC are more historical and relate to use of more active packed columns. The pharma world has just not moved to embrace the newer techniques.
5) The flexibility of separations in HPLC is much greater due to the ability to change selectivity with changing the mobile phase- but often, pharma methods are set up and never touched anyway because of validation requirements.
6) Perhaps I got it all wrong and a much larger proportion of drugs is involatile than I had thought.

[Victor]

Sorry -in the first part I meant "There is a predominance in the use of GC for the forensic analysis of drugs of abuse...."

[Noser222]

I think number 6 is the biggest factor.

[DR]

More reasons to use LC -

many APIs are heat labile

LC/NMR, LC/MS are becoming more affordable and more common

Column capacity is higher for LC columns (I'm pretty sure - generally speaking), this comes in handy when trying to do an assay and accurately assay impurities at 10^-3 or lower concentrations vs. the API peak.

Diode array and variable wavelength UV detection allow more flexibility than FID, ECD or FPD, especially when you consider that most of the APIs have chromophores and most excipients do not.

[tom jupille]

I suspect that thermal stability/volatility is the biggest single factor. If you look at GC literature just as HPLC was taking off (circa late '60s - early 70's), there was tremendous interest and activity in derivatization chemistry (silylation, esterification, methylation, etc.) to make polar compounds volatile and thermally stable so that you could do GC. Pre-column derivatization is a PITA (tedious, time consuming, often incomplete or non-specific); one of the reasons for the early acceptance of HPLC in the pharma industry was the convenience that came from avoiding derivatization.

Arguably a second factor is that HPLC is more flexible in terms of selectivity (by the way, DR, I think that the peak capacity of GC can be considerably greater than that of HPLC). In GC, the mobile phase is essentially a passive participant in separation chemistry; it serves merely to carry the analytes through the column. Gases are simply too dilute to do much interacting (changing the carrier gas from N2 to He, for example, has essentially no effect on selectivity or retention). In LC, on the other hand, the mobile phase is concentrated enough to make a major contribution to separation chemistry.

All of that said, I think there is also a fair bit of user inertia; we tend to turn first to the technique(s) with which we are most familiar. If they work even reasonably well, we rarely have the luxury of time to search for something better.

[DR]

I thought I might not be right on that one - (column capacity).

That said, GC capacity is better on a per unit basis. When faced with having to collect fractions of an unknown peak for identification purposes, you can dump a lot more API on a semi prep LC column than you can a GC column (LC scalability being another aspect deemed useful by the pharmers in the room).

Also - I rather dislike having chronically burnt finger tips.

[Bruce Hamilton]

7. Development of drugs traditionally involved normal phase column chromatography and/or TLC.

Normal phase HPLC quickly dominated drug discovery and development with improved resolution, and reverse phase offered even greater access to information about drugs under development, as well as overcoming the degradation issues on silica.

It could be said that the pharmaceutical industry has driven HPLC to match it's needs, and other large chemical industries haven't needed it. GC dominates the petroleum industry, fats and oils industry ( but that's changing ), agrochemicals ( but that's changing ). Low pressure ion exchange chromatography dominated potable water and biochemistry ( nucleic acids, proteins, peptides etc. ), but that's changing.

HPLC's biggest hindrance was the lack of a low cost universal organic detector ( like the FID ), that could be used for gradients. The advent of lower cost MS detectors, ELSD, etc has changed that.

Bruce Hamilton

[ravenwork]

I would add environmental testing to the list of sectors where GC predominates. Indeed, it is so heavily favored, that pre-column derivitization is routinely performed, even when a more reasonable HPLC method is available or could be readily developed. The advent of reasonably priced and reliable LC-MS instruments is just beginning to change this.

Evan L. Cooper, Ph.D.

[tom jupille]

Also - I rather dislike having chronically burnt finger tips.

Between GC and vacuum rack glassblowing in grad school, I think I have no nerve endings left in my fingertips.

[Victor]

Thank you very much for these comments. Clearly, there is an issue with volatility and thermal stability issues with a proportion of APIs and degradants etc. The ease of prep separations pointed out by DR is also important.
Tom's comments about user inertia are interesting. However, if this was the case, why was the change made so enthusiastically by Pharma from GC to HPLC in the first place?

My suspicion is that HPLC was far superior to packed column GC (even in terms of efficiency-only a max of about 10,000 plates for standard packed columns if you are lucky). When capillary GC really took hold ( I guess around 1980 when fused silica columns were brought into use) HPLC had already become deeply entrenched and no-one wanted to go back to GC, even though 100,000 plates is routine.

DR's point about FID and PDA comparisons is interesting; however, I would compare a GC-MS with an HPLC-PDA, because both are approximately the same price these days. This is a much more difficult decision.

I think Evan also makes a good point. There are areas, e.g. like environmental testing, where the same analytes are determined by GC that pharma uses HPLC for, but I see reasons for this as in my original post.

My conclusion is that there is probably a lot more that Pharma could do with GC. But why bother, if the HPLC method performs adequately?

[james little]

We do a lot of things by GC-MS and GC-FID that many people would do by LC-MS and LC-UV.

The GC gives better resolution of components in most cases by just employing a few different columns. So, very little method development.

Down side is that we have to derivatize a lot of components. Dervitizations are chemical reactions so they have by-products. Often the concentration of these by-products (artifacts) vary dramatically with sample history. Thus, develop something in the lab with standards works great, start running samples and have artifacts.

This is why I have collected a large list of artifacts over the years, see

http://users.chartertn.net/slittle

see section on diazomethane and silylation.

[guenzia]

I strongly vote for reason #6 in the original post.
I would also add a couple of additional reasons, mainly connected to the field of bioanalytics (analysis in biological fluids):
(8) when dealing with very complex matrices, sample preparation is normally much simpler for LC than for GC (compare protein precipitation vs liquid-liquid extraction or SPE)
(9) LC is more merciful than GC and people are trained faster in LC than GC. GC still needs true specialists to optimize all the parameters, while a reasonable LC (especially -MS/MS) method is achievable in a short time by relatively experienced people

[ravenwork]

Having worked with GC and HPLC extensively, I cannot agree with the statement that HPLC is simpler than GC. GC involves the well-behaved gas-phase, where the only interaction that matters is that between the analyte and the stationary phase. Usually, method development means nothing more than proper column selection. The separation efficiencies possibly on your average capillary column are rarely approached in HPLC. With HPLC, we must contend with a condensed phase where solvent interactions and mass transport issues become important. System variables become important. In my experience, GC columns frequently work just as advertised by the manufacturer. HPLC columns, on the other hand, frequently require adjustments from what the manufacturer states.

My earlier post on this thread about the environmental testing biz applies here. Enviro testing is very low margin, and consequently often employs marginally educated technicians. These folks do just fine with GC and GC/MS, but they avoid HPLC like the plague.

Evan Cooper

[sfe-co2]

Hi all,

I read the above discussion with interest.

I think one must also remember that whilst GC analysis is, in the general sense, destructive to the analyte, LC isn't. This is arguably a very important feature that allows the pharma industry to isolate/purify actives routinely on a preparative scale.

[jzt]

I work for a big pharma (that's why I can't afford to spend much time here). From my perspective, Tom was right on the mark when he said "I suspect that thermal stability/volatility is the biggest single factor."

we do use GC beyond just residual solvents test. we have GC methods for starting materials and some low molecular weight intermediates. However, when it comes to the drug substance, most of them just can't survive the high temperature in the injector. We have to quantify low level of impurities and degradant in the drug substance too. These compounds are not always in the same class as the drug substance, which makes consistent derivatization almost impossible. GC-MS is also troublesome for us. In EI mode you hardly see the molecular ion, and it takes too much effort to run in CI mode (at least with our instrument). Soft ionization method such as ESI in LC-MS gives us more useful information.

One incident we encountered recently put some doubt on using GC for starting material analysis. It’s a small compound (5 carbons in an aromatic ring) with a primary amine group on the side chain (MW 110). All of our suppliers used GC to analyze this compound, and the purity factor was normally > 99.5%. however, not every batch of this starting material worked the same in the reaction. We couldn’t find anything by the GC method, so we turned to HPLC and CE. Sure enough, there were some late eluting peaks, the amount corresponded with how well that batch worked in the reaction. We later identified those peaks as dimer and trimers (yes, more than one) of that small compound. These impurities will never show up in the GC method, but have great impact on reaction rate and product quality. The final method for quality control is an ion-pair HPLC method. The precursor to this starting material, the synthetic impurities, its positional isomers, the dimer, and trimers are all separated.