It is probably highly improbable that a nuke would hit a nuke dead on, and even if it did I am not sure that would trigger another explosion ), I think all it would do is add to the radioactive fallout.
Yea, it's my understanding that they have a timing mechanism that actually starts the chain reaction - not an impact trigger. The bombs dropped at Hiroshima and Nagasaki were detonated before they hit the ground. So if two missiles hit (or one missile intercepted the bomb) you likely would forestall the nuclear detonation but the impact would still likely release radiation.
Or I could be talking entirely out of my butt... 
Actually it would depend on a lot of factors
H-Bombs actually work because of plastic- the plastic has a matrix of Hydrogen atoms built into it, and it absorbs the radiation from the fission explosion used as the trigger ( due to how E&M waves like to hop from atom, be absorbed and then re-emit) this superheats the plastic fast enough that it starts the Hydrogen Fusion process before the whole structure is destroyed by the Fission explosion. The Fusion of Hydrogen into helium releases a massive amount of energy as radiation (heat, light, gamma and sub atomic particles)
If another H-Bomb were to get enough radiation from the first explosion, then the Hydrogen matrix could also become superheated and detonate. Since air blast waves take a while to propagate, and radiation does not in comparison, its entirely possible that this COULD happen, though I suspect most war heads are heavily shielded to prevent that - as well as protect those that work around them
The Fission explosions used as triggers for H-bombs are caused by critical mass- where Neutrons "collide" with the fissionable atoms, thus releasing more neutrons to "collide" with others atoms in a chain reaction. It is the energy released by the collision that provides the blast. The critical mass is achieved either by compressing the material (Fatman and its Plutonium core) or by simply adding material (Little Boy and U238) However, a bombardment of Neutrons from another detonation could potentially cause the reaction to start and be self sustaining. Again, perhaps enough to cause the Hydrogen matrix to detonate as well.
They were timed to detonate at altitude because the blast would do more damage that way. I suspect delaying a launch to an already attacked area by 20 minutes had more to do with air temperature for flight and radiation effects on the electronics, than it did with not wanting to waste munitions by accidental detonations..
Sidenote: The Littleboy design was "tested" only by literally dropping a core of U238 through a borehole in a stack (or pile) of U238. Some poor guy would get behind (a very inadequate) lead shield and drop more and more stuff through the pile - measuring radiation levels each time,, till they found "critical mass" - if the core had gotten stuck, it would have resulted in a Nuclear meltdown and explosion. They dropped the final device on Hiroshima without ever actually detonating a similar device. The hard part about building a fission bomb is NOT the construction, but getting enough refined fissionable material. Back then they had miles of gas diffusion chambers to separate the U238 isotope from the yellowcake. It took years to get enough material for one bomb. Now you can duplicate that result with a few years and about $20K worth of laser equipment. The yield of the Hiroshima bomb was less than 1/10 of a percent. The vast majority of the material did not fission. Yes, there are better and more complicated designed for obtaining higher yields, but I believe the bang obtained from Littleboy would be enough for most terrorists to be satisfied.
Now start researching Protactinium and see what conclusions you come to.
