Yeah, as I researched the topic of multiple choice exam design, seems the rule of thumb is to reject outright any distractors that are chosen by less than 5% of test takers.