There is probably something wrong, either with their conclusion or with other studies of similar kind. There must exist a state of disagreement among scientists regarding the role of AA in arrhythmia. Other studies using real animals (not in-vitro) showed the opposite effect, such as this:
Effects of arachidonic, linoleic, linolenic and oleic acid on experimental arrhythmias in cats, rabbits and guinea-pigs
I am not sure which is correct, but since the wiki article does point out that AA plays multiple role in many processes, thus multiple, even contrary outcomes aren't surprizing.