The purpose of this study is to compare lamotrigine with lithium in the long term treatment of bipolar disorder in terms of new episode preventive potentials.

Bipolar disorder is a recurrent lifelong illness. Lithium is still the drug of first choice in the prophylactic treatment. However, lithium's side effects are considerable. Thus the interest in developing alternatives to lithium has increased, and various antiepileptic drugs have been studied. Among these, lamotrigine, which generally is well tolerated, has been evaluated in industry generated trials. However, for various reasons, the generalisability of the results from these studies is limited. Also the follow-up time is limited. This study is a pragmatic effectiveness study, overcoming some of these limitations, and comparing lamotrigine with that of lithium in terms of their potentials for preventing new episodes in patients with bipolar disorder.

Inclusion criteria:

1. Age above 18
2. Inclusion is preceded by an affective episode requiring hospitalisation or outpatient drug treatment (index episode) which can be a manic episode (ICD-10 research criteria), a depressive episode of at least moderate degree (ICD-10 research criteria) or a mixed manic state (manic episode with at least additional three ICD-10 depressive symptoms except for difficulties with concentration or thinking, agitation or dyssomnia). The episode can be with or without psychotic symptoms including Schneiderian first-rank symptoms or bizarre delusions as long as the psychotic symptoms do not occur outside the affective episode.
3. No more than 12 months may pass between the onset of the index episode (or admission if that is the case)and date of randomisation (in order to ensure a current risk of relapse as well as reliable psychopathological information).
4. Besides the index episode, at least one previous episode must have occurred within the last five years which meets the criteria mentioned in point number two above. This episode may not necessarily have led to hospitalisation. Two episodes are separated by at least two months without significant symptoms or change in polarity (depression to mania/mixed mania or vice versa).
5. At least one manic episode (or mixed manic episode) within the last 5 years.

Exclusion criteria:

1. Contraindications to the protocol drugs.
2. Severe somatic disease, e.g. epilepsy, which may interfere with study treatment or effect evaluation.
3. Pregnancy (or risk of pregnancy).
4. Subject has prior to randomization received prophylactic treatment with lithium or lamotrigine conducted adequately in the sense of sufficient time and dose and ensuring compliance, and experienced a definite lack of prophylactic effect
5. An ICD-10 diagnosis of abuse within the last year with a subsequent risk of protocol violation. Therefore, subjects with a secondary abuse can often be included in the study.
6. Anticipated protocol violation for other reasons.
7. No written informed consent from the subject can be obtained.
8. The subject has previously been randomised in the study.