I actually had a call very much like this last week and we were discussing it today.
If the patient is unstable because of the tachycardia, the treatment of choice is synchronized cardioversion. However, if the patient has been in atrial fibrillation for over ~48 hours without adequate anti-coagulation, you risk them throwing a clot which could cause a stroke, heart attack, PE, or embolism somewhere else.
That being said, my patient was on aspirin and coumadin for years, and reported being compliant with it. I contacted medical control, explained to them the situation (including coumadin dosing and compliance) and the physician authorized cardioversion. It fixed the guy and we transported a much better looking and feeling patient to the hospital without further incident. I asked an attending physician about the call when we arrived, and he reported that they regularly cardioverted patients with a-fib, as long as they were adequately coagulated (demonstrated by clinical history or labs).
So, if your patient has new onset a-fib with RVR, or is adequately coagulated, cardioversion isn't as awful as some make it seem. At least according to two physicians last week.
If there's any question about coagulatory (?) status, such as an inadequate history, or onset time; I would be very careful and try your medications first (calcium channel blockers). Even in your unstable patients, Cardizem and similiar can work very quickly and have little risk compared to cardioversion.
I'm interested to hear if there are other thoughts on this because I was always "scared" away from cardioversion for these patients in medic school...