Towards a multifactorial approach for prediction of bipolar disorder in at risk populations

Abstract

Background: The high prevalence, recurrence rate, chronicity, and illness burden in bipolar disorder (BD) are well documented. Moreover, insufficient response with conventional pharmacological and manual-based psychosocial interventions, as well as evidence of illness progression and acceleration, invite the need for early detection and primary prevention.

Methods: Herein we comprehensively review extant studies reporting on a bipolar prodrome. The overarching aim is to propose a predictive algorithm (i.e. prediction of BD in at-risk populations) integrating genetic (i.e. family history), environmental (e.g. childhood maltreatment) and biological markers (i.e. BDNF, inflammatory and oxidative stress markers). Computerized databases i.e. Pubmed, PsychInfo, Cochrane Library and Scielo were searched using the followed terms: bipolar disorder cross-referenced with prodromal, preclinical, at risk mental states, clinical high risk, ultra high risk, biomarkers, brain-derived neurotrophic factor, inflammation, cytokines, oxidative stress, prediction and predictive model.

Results: Available evidence indicates that a prodrome to bipolar disorder exists. Commonly encountered features preceding the onset of a manic episode are affective lability, irritability, anger, depression, anxiety, substance use disorders, sleep disorders, as well as disturbances in attention and cognition. Non-specificity and insufficient sensitivity have hampered the development of an adequate prediction algorithm.

Limitations: Limitations include biases associated with retrospective studies, poor characterization of clinical high risk, inadequacy of prospective studies regarding sample selection and absence of specificity of risk states.

Conclusion: We propose a hypothetical prediction algorithm that is combinatorial in approach that attempts to integrate family history, early adversity, and selected biomarkers.