[If a patient has never experienced depression, should we tell him he has bipolar disorder? An updated systematic review on recurrent mania].

CONTEXT: Bipolar disorder (BD) is a severe and recurrent mood disorder. It is characterized by episodic changes in mood and energy/activity levels that are increased during mania/hypomania or decreased during depression. Recurrent mania (RM) is a mood disorder, which would be defined by at least two manic/hypomanic without depressive episodes. Despite a rich body of clinical descriptions, RM is still not integrated into the latest editions of disease classifications and continues to be subsumed under BD in clinical practice.

OBJECTIVES: We conducted a systematic review of the literature to pool data about RM prevalence within BD groups, identify differences between RM and BD and develop reliable knowledge about specificities of RM. Furthermore, we sought to identify the methodological bias inherent to RM studies.

METHOD: Relevant publications were identified by a systematic search of PubMed, Embase, ScienceDirect and PsychInfo databases according to PRISMA criteria, with no limitation of date. The following MESH terms were used: (mania OR manic) AND (unipolar) NOT (depress*) OR ("unipolar mania" OR "unipolar manic" NOT "depress*").

RESULTS: Twenty-three (23) of 186identified studies met eligibility criteria for our systematic review. The total sample included 1118RM subjects among 4796BD subjects. The weighted mean of RM prevalence was 23.2%. Compared to BD, RM was characterized by a predominance of men, an earlier age at illness onset, less rapid cycles and seasonal variations, longer manic episodes, less specific clinical features (suicide attempts, anxious disorders, catatonic symptoms, irritability, hyperactivity, racing thoughts), less family history of depression, more addictive comorbidities and worse response to lithium prophylaxis (P<0.05). However, many studies failed to replicate these significant differences.

LIMITS: RM studies were mainly retrospective. The major bias of RM studies were the lack of consensus on the defining criteria for RM and the risk of unreported depressive episodes, both in charts that were reviewed in retrospective studies and in prospective studies with insufficient follow-up duration.

CONCLUSION: Although the literature on RM remains sparse, many authors agree that RM should be distinguished from BD. RM would concern almost 1 in 4 BD patients. Furthermore, several clinical variables could differentiate this mood disease from BD and may orient the specific therapeutic choice. However, clinical criteria are still not reliable enough to make a diagnosis of RM. Further studies are required to replicate the results of existing studies and to adjust for the effect of methodological biases.