I'm feeling wordy today.
Maybe back in the late 70s and 80s when things were over engineered but not today when you have the accountants hovering over you, you don’t get this luxury anymore i know! It may be 20% but i doubt it due to the UK climate but it was not design for just the UK climate. Marketing men rule the world now 
Although it varys a lot, out environment here in the UK is pretty benign and free of extremes. If anywhere in the world is safe to run a tweaked engine, it's England. Modern engines might run smaller 'safety factors' but they're still going to be designed with life and component failure in mind - if only to keep the bean counters happy about potential warranty claims (see Toyota, "the car in front" and now you know why - the accelerator is jammed wide open
).
I think that the major tuning compromises on modern turbo-charged engines are less to do with component reliability and more to do with fuel consumption and CO2 emissions. But component reliability is what causes your car to break down so you are right to look there, particularly as the gains from remapping are essentially to do with upping boost pressure.
The engineering compromises in the standard engine components may be less in favour of reliability than 30 years ago but there will still be a healthy margin in a car designed to survive 12 years/120,000 miles. Having said that, I agree that the tuning companies will have been unlikely to have analysed the increased stresses on a con rod bearing from raising turbo boost (for example). They won't have had access to VW design data or computer models. In place of calculation and modelling, their knowledge and experience have to be used instead - which can be just as valuable, if not more so. Only time, testing and use can reveal the long term effects of the changes they make but I doubt any of them have driven a dozen remapped Mk6 GTI's 150,000 miles just to be sure.
However, they do have access to (admittedly far less reliable) life data in the form of the experience of customers, owners and other tuners who have collectively driven many hundreds of thousands of miles in modified cars. If, say, turbo failures were starting to happen at 50,000 miles then this information would start to appear on forums like this one and feed back not only to the tuner but into the customer domain. The EA113 is now several years old but I doubt there's much information on 100,000+ mile cars yet.
This is why more than anything when I look for a company to modify my engine I look for the history and pedigree of the company - what have they done before and how long have they been doing it? Do they have experience from motorsport or ties with the car manufacturer, etc. This personally is what gives me confidence that they are going to change my engine in ways which will not significantly compromise its reliability.
It's why I haven't yet made my mind up about Revo and why in the past I've gone to Abt in Germany, who although expensive have close historical ties to VW. Before I decide on Revo, I'll talk to them, get to know them a little, find out what they did before 2004 (when their website was created), understand a bit about the way they go about doing things. Some might say this is overkill but I like to keep my cars for much longer than the 3 years/30,000 miles that seems more typical on here. Thus the long term effect on conrod bearings and the experience of the bloke altering the code in my engine management is of importance.
And of course I'll listen to the feedback of actual owners, who use their cars in anger on and off track, on forums like this one.
