Although the onset of bipolar disorder (BD) is usually estimated in early adulthood (Burke et al., 1990; Kessler et al., 1997), it is still not clear which age-specific triggering factors may contribute to the underlying vulnerability.

The present meta-analysis attempts to clarify the number of untreated years of BD, from first symptoms appearance, to first actions carried out by health services (first diagnosis, treatment and hospitalization).

A literature search of three databases PubMed, Web of Science and Psychology and Behavioral Sciences Collection was performed, looking for peer-reviewed publications in English, French, German and Italian language that reported the AOO of BD; the search string adopted was “bipolar and onset”. Standardized mean differences were calculated between (i) mean AOO, operationally defined as onset of first symptoms/episodes, and (ii) mean AOO, operationally defined according to four criteria: first contact with services, date of the diagnosis, first pharmacological treatment, or first hospitalization.

The searches yielded 8710 articles; 2424 of these articles met the inclusion criteria. A final set of 19 studies presenting multiple definitions of AOO has been analyzed, revealing an overall effect size of 6.96 of untreated years (Cohen's d = 0.65)

To our knowledge, this is the first meta-analysis that addresses the definition and the modulation of the AOO in bipolar disorder. Identifying the time-frame of untreated illness is very important for the best planning of timely interventions.

The authors have not supplied their declaration of competing interest.