I've vaped 100% VG for over three years so here are some observations.
The term '100% VG' is actually meaningless because it can mean pure VG with no water, or it can mean correctly-diluted VG with distilled water or alcohol. It depends entirely on the vendor, each has a different meaning for it. To be usable for vaping, in my opinion a VG liquid should be almost as thin as a PG-based one, otherwise it won't wick fast enough. However, like everything in vaping, that's a matter of opinion - everyone's is different. It depends what hardware you're using.
To make a VG liquid as thin as it needs to be for my taste, I add up to 20% DW and/or alcohol and then add flavoring, which means the final result might be 30% dilution. As has been mentioned here, almost all flavorings are based on PG, alcohol or water - so if you add 10% flavoring then basically you are adding almost 10% diluent anyway. In other words if you add 20% diluent then 10% flavor then you have added about 30% diluent to the VG, which makes it far more usable (IMO).
It's hard to find flavors that are glycerine based as against PG based. I don't worry about this because IMO a couple of percent PG kills pathogens (both in the base liquid and also in the resulting aerosol, and possibly helps with potential lung infection issues) and is not noticeable in the mix even if you have a PG intolerance issue (for me at any rate). I think some PG in there is a good thing [1].
I mainly DIY so here are a couple of examples of base materials and how they can be used:
RTS Vapes '100% VG' unflavored base
This is an undiluted glycerine [2] base and almost as thick as a gel. It is unusable (for me) until it has been thinned, although I respect the fact that someone somewhere will like it as it is, for their particular usage. Because it needs extensive dilution, you need to buy it many times stronger than your final target strength. To get it as thin as I require, it needs a 20% addition of diluent plus whatever is in the flavoring: for example 15% DW plus 15% flavor. Essentially this means it has had 30% diluent added. The vape quality is not affected in any way by such extensive dilution, if anything it's improved, at least for my taste - although like anything in vaping, any matter of taste is debatable and purely a personal matter. An interesting point is that adding water to an e-liquid does not seem to be a negative, although there might well be a limit. I try to keep the total dilution to 30% but only because I can't be bothered to experiment with higher dilutions, and up to 30% is just fine.
Ecopure Krystal '100% VG' unflavored base
This is a fully-diluted glycerine base and can be used as-is. It has been diluted with DW and EM. Because it is only available at 36mg, those who vape higher strengths are limited in how much flavor they can add, although typically the final mix will be above 30mg. This VG when ready for vaping is almost as thin as a typical PG mix, and is usable in almost all carto tanks etc. as per PG.
[1] After all this is one of the reasons why PG is used in the nebulizers used by lung transplant patients. It seems a useful addition to VG liquids especially when used by vapers with smoking-related lung diseases such as emphysema.
[2] Our use of the term 'VG' is now incorrect, as times have moved on. Although originally a vegetable-source glycerine was the best choice for vaping, this is no longer completely true, as pharma-grade synthetic glycerine is now available that is 100% pure (such as Dow Optim). Dow Chemical now advise Optim as the best excipient for inhalable medicines, in preference to their own 100% pure PG [3], which used to be the material advised for inhalable pharmaceuticals.
[3] Somewhere in this thread, a member has posted that ".....PG contains carcinogens". This is absolutely incorrect. A PG or VG-based pure nicotine source will contain carcinogens as these accompany the nicotine, and are either impossible or too expensive to remove. Once diluted to usable strengths for vaping, the retail liquid (either PG or VG based) will typically contain tobacco-origin carcinogens at approximately 8ng, a level exactly equal to that generally present in NRT products such as nicotine skin patches. Such levels have been described by an epidemiologist and professor of public health who is a noted expert in this field as, "millions of times lower than conceivably harmful to health". Snus generally contains such carcinogens in the range 1,200ng to 2,000ng/gm and has no statistically-detectable cancer effect, according to several meta-analyses of multiple studies such as Lee & Hamlin's various works [4]. In other words, Snus may contain 250 times more carcinogens than e-liquid, but has no detectable cancer-causing effect.
[4] There have been more than 150 clinical studies of Snus in Sweden, over almost 30 years. A very small number of studies (<5) showed a possibility of a pancreatic cancer increase. These studies have either been discredited due to suboptimal methodology, or their results disappear when combined with the other 145 + studies that showed no such effect. The effect is not observable at all in the meta-analyses even though other effects such as a possibility of increased risk for stroke is visible just above the background noise. It should be noted that:
(a) If 2 trials show an effect but 140 other trials don't, in general the larger picture is considered the most likely rather than the outliers.
(b) An equal number of trials showed the opposite, a protective effect against cancer, and these are also discarded as outliers by general agreement.
