afib is not an inherently dangerous arrhythmia. And I"m not sure what mcgill was trying to say about it's frequency being too high to get past the AV node? The only way it doesn't get past the AV node is if you have an AV block on top of it. In that case, the danger is not from the afib but rather from the Mobitz type II or the tertiary block (primary block and Mobitz type I are not particularly dangerous rhythms) There are two major dangers of afib:
1. as the atria have effectively lost their kick, the person has blood stasis in this chamber, allowing for thrombus to form. This thrombus may cause emboli, resulting in stroke. Thus, patient's in afib require anticoagulation to reduce their risk of emboli and resultant stroke.
2. afib with rapid AV response is dangerous because you may have a tachycardia that is too rapid for effective pumping and too rapid for coronary artery filling (which occurs during diastole). For this reason, people in afib may need to take a medicine that slows AV conduction such as a CCB or a BB.
Third, there is no evidence that rhythm control is necessary for afib. Sometimes they will cardiovert a younger person with "holiday heart" after ensuring no thrombus has formed, because otherwise he will not be able to be as active as he wants to be (hard to be a runner if you're losing atrial kick). Rather, emphasis is on rate control and anticoagulation.