For the SMD type, you don't want to go smaller than a 1210 package. I've had trouble with the smaller ones before. They're not as tough and resistance can skew after a few trips.
Otherwise the lower the resistance the better so you'll do better in that regard with a higher trip current. The limit there is if you go too high on the trip current you can get response times too slow when a fault occurs. It depends on the power supply, but if using a a high drain 18650 there's generally not a big concern with trip times. They'll quickly trip fuses well above the trip currents you've listed. Fault currents are generally 50 to 100 Amps which will get your trip times under a second.
You want your hold current ratings in the area of the maximal draw for the DNA which is 12A, doesn't have to be right on, you can go under an amount putting maximal demand 50% over the hold rating. That would be 8A hold in this case.
There's a wide margin between hold and trip currents, typically hold current is around half the trip current. With an e-cig you can run fairly close to the trip current without getting a false trip since an atomizer is not powered continuously. You do want some margin there to eliminate the possibility of a false trip.
You can run a couple fuses in parallel to double the hold and trip current and also halve the resistance. I would recommend two of the 1210L450SLWR. Run two in parallel for a single cell, one on each cell for dual parallel cells.
I would not recommend running more than two fuses in parallel since there are derating issues that come into play with parallel fuses. With two it's not an issue, but it can become significant with more than that.