I've never been convinced by the evidence that I've seen for fasting for TFT tests, and whether or not doing so raises or lowers TSH. I'm happy to believe that an effect of fasting/eating on TSH may exist in some way, I'm just not sure what the effect actually is.
However, I do know for certain that there are some tests which should be done fasting e.g. iron testing, fasting glucose testing, cholesterol/lipid testing, and no doubt a few others as well.
If my doctor decides I need some blood testing and arranges for iron testing and TFTs to be done, do I fast (for the iron) or not fast (for the TFT)? I can't do both for the same blood draw, and the doctor is unlikely to allow me to have two lots of blood tests (for cost reasons!) when both can be tested from the same blood draw.
If I don't deliberately fast for TFTs then I can imagine a scenario in which one time I get a blood test I might have a Full English Breakfast beforehand. Another time I might be running late and I just have coffee. The time after that I fancy cereal or porridge. I don't know what effect all these different situations will have on my TFTs. I'm sure it reduces the comparability of results between tests though.
So I side-step the whole question by being as consistent as I can possibly be and by fasting for every single blood test. And for me that means, "Stop eating and drinking after about 10pm the night before (apart from water), until after I have the blood draw".
I have discovered one fly in the ointment though. If the blood test form doesn't specify overnight fasting then some people who take blood interpret this as saying "Thou must not fast!", and they get very twitchy if you have fasted. But if fasting is not required, then I interpret this as saying it doesn't actually matter, and the choice is mine. So I always fast. In my view, it maximises the comparability of my tests.