Introduction

Despite the fact that rapid cycling and mixed states are common manifestations of bipolar disorder (Kilzieh and Akiskal, 1999; Akiskal et al., 2000; Grunze et al., 2002a) they have very rarely been subjects of interest for clinical and basic research. As far as clinical research is concerned, rapid cycling and mixed state traditionally were exclusion criteria for controlled randomized phase III studies. This changed just recently, where antiepileptics such as lamotrigine (Calabrese et al., 2000) were tested for their prophylactic efficacy in rapid-cycling patients. As far as mixed states are concerned, recent trials on modern antipsychotics (Keck et al., 2003; Sachs et al., 2002; Tohen et al., 2002) also allowed mixed patients in these studies.

A major problem for including mixed states in studies is the lack of a generally excepted definition. The Diagnostic and Statistical Manual of Mental Disorders, 4th edn (DSM-IV) criteria (fulfillment of all criteria of mania and typical depression at the same time: American Psychiatric Association, 1994) are too narrow, as it has been shown in retrospective analysis of the valproate mania study (Bowden et al., 1994) that one depressive syndrome predicts inferior responsiveness to lithium and better response to valproate (Swann et al., 1997). A commonly used definition is the so-called Cincinnati criteria, which require at least three relevant depressive syndromes (McElroy et al., 1992); however, again, a general consensus on how to define mixed states has not yet been achieved as other views may also have clinical advantages (Perugi et al., 1997).