While I use a Zip Loc bag to spread the paste, a fingerprint shouldn't matter. Thermal paste itself is made of ceramic or metal particles suspended in silicone grease.
http://en.wikipedia.org/wiki/Thermal_grease
TIM has been in computers for a while, but they are not the same as before. TIM used to have a settling period (around 2 weeks) to start working at peak performance, and users can clearly see the difference before and after this period as it can be 5-10 degree apart. Nowadays TIM works almost instantly, and I have no idea how those chemical reaction to make this possible, but I do know there are reactions.
The question is, what happened during the settling period? Is it chemical reaction? or does TIM actually moves in between heat cycles? Whatever it is, it must be down to atomic level(okay okay, not atomic level, but in a very very small scale), and fingerprint is actually really big within that scale. Fingerprint doesn't dissolves into the TIM, but whether repelling against the TIM, leaving itself imprinted on the surface. That means, the surface with the fingerprint actually don't have TIM on.
Don't get me wrong, the fingerprint doesn't damage the CPU or the HS, but it does effect radiativity. Temperature is not the right unit to reflect its effect as heat will eventually find thermal bridge by itself. Its effect however is visible when the radiativity reaches terminal temp.
If you will use alcohol to clean the surface and believe the residue of the previous thermal paste will have an effect, you may as well not to stamp your CPU with your finger. Don't you agree?