I know some psychiatrists who suggest 300mg as a minimum. I have used up to 600mg in patients who it appeared to help at lower doses and had breakthrough symptoms. 700mg+ are used in epilepsy. The Maudsley Prescribing Guidelines note that doses up to 1200mg have been used in bipolar disorder. The reason for the 200mg dose is that in the RCTs there was no evidence of any further benefit at 400mg. However, it is possible that higher doses are needed for some.
There is limited evidence supporting lamotrigine for the treatment of acute bipolar depression. 4/5 RCTs failed to show efficacy. When pooled together in meta-analysis there is marginal benefit, but the NNT is 12! So you need to treat 12 people for one to have an effect above placebo. Of course the problem could be the problem with dose titration. I don't think it is a good option for monotherapy, but it has a relatively benign side-effect profile, SJS is rare (had a pt who got a big payout for developing SJS and blew it all on crystal!), and patient's like it. Some people complain of horrendous agitation or activation at higher doses. Also it is associated with hypersexuality which is not mood-related (i.e. not as a result of manic induction and occurs in epileptics as well).