The purpose of the study was to evaluate the efficacy of lithium prophylaxis in patients with bipolar affective disorder and to determine clinical differences between lithium responders and non-responders. Clinical and demographic data of 103 patients treated with lithium for at least 10 years were analysed. 24 patients had no episodes on lithium (excellent responders--ER), in 48--at least 50% reduction of episode index (number of episodes/year) was observed, and the remaining 31 did not benefit from lithium treatment. Duration of illness before lithium introduction was significantly longer in lithium non-responders. Lithium responders had significantly higher episode index before lithium. Percentage of excellent responders diminished during treatment. The trend toward higher lithium plasma level in excellent responders was observed. Non-responders had significantly more depressions during first year of lithium prophylaxis. Depressive episodes during first year of treatment may be a predictor of non-response.