Assist Portal search icon
Assist Portal close icon
Assist screening tools - ASSIST Portal
Assist Portal - Blog




Risk factors for new-onset bipolar disorder in a community cohort: A five-year follow up study.
Published in: Psychiatry Research
By: de Lima Bach et al., 2021


The aim of this study was to assess sociodemographic and clinical risk factors for new-onset Bipolar Disorder (BD) among a community cohort of individuals aged 18-24 year old, living in urban Brazil.


The researchers conducted a longitudinal clinical assessment of over 1200 individuals which involved two assessments across a 5-year span. Initially, researchers identified 1762 individuals living in urban Brazilian community, and approached them to complete in the study. 1560 participants agreed to participate and complete the assessment, which included questions pertaining to sociodemographic (e.g., sex, age, education, skin colour, current work, and relationship status); and clinical predictors of mental health (e.g., substance use, comorbid mental health disorders, familial history of mental health, suicidality). Ineligible participants were those not aged 18-24 at time of assessment, those with severe cognitive disabilities that precluded them from understanding and/or completing the assessment, and those indicating the presence of lifetime manic or hypomanic episodes at baseline. Participants were assessed in their homes by trained interviewers.
The ASSIST was used to screen for substance abuse/dependence, which was identified as meeting either moderate or high-risk use on a particular substance. The MINI was used to assess the other clinical risk-factors of interest. These included depressive, anxiety, obsessive-compulsive and post-traumatic stress disorders and suicide risk. Family history of mental health disorders were also assessed at baseline.
After 5-years, all participants were then contacted to participate at follow-up. Of the 1560 participants, 1244 agreed to participate, representing an impressive retention rate of 79.7%. Fourteen participants were confirmed deceased and 302 participants could not be contacted. At follow-up, participants were assessed over the course of weekly meetings to identify new cases of BD. In the case of uncertain or questionable BD diagnosis at baseline, re-evaluation was conducted by ‘an experienced psychiatrist.


Investigators identified a number of important findings which will be detailed in turn.
First, the cumulative incidence (i.e., the proportion of new cases) of BD at follow-up was 4.6%. Remembering that those with BD or indicators of BD at baseline were excluded.
Second, there were no significant differences between sociodemographic factors and the onset of BD at follow-up. Specifically, among the 50 individuals meeting BD diagnosis at follow-up, there were no differences between them according to sex, age, income, skin colour, current work and relationship status.
Third, there were a number of clinical factors which increased the risk of new-onset BD. Risk of new-onset BD was increased by depressive disorders by 6.6 times, PTSD by 5.6 times, suicide risk by 3.5 times, and anxiety disorders by 3.2 times. Simply stated, those individuals meeting the clinical diagnoses of comorbid mental health disorders were between 3 and 6 times more likely to develop BD compared to those without. A non-significant increase in risk of new-onset BD was found among those with history of familial mental health disorders.
In terms of substance use behaviours, a similar pattern emerged. Risk of new-onset BD was increased by abusive use of tobacco, cannabis, cocaine, and other illicit substances. A non-significant increase in risk of new-onset BD was found among those with risky alcohol use behaviours. Again, simply stated, those individuals using all substances (except arguably alcohol) were found to be at greater risk of developing BD after 5-years.


Our results point to substance abuse as an important risk factor for the onset of BD. As for the clinical implications of these findings, there is a need for accurate clinical assessment of psychoactive substances use and, especially, early intervention to pre-vent the occurrence of mania/hypomania episodes in young adults


One of the primary benefits of longitudinal research over cross-sectional studies is that it allows for the possibility for inferences about causality. In this study, it was found that clinical risk-factors, including the presence of substance use disorders were the strongest predictors of new-onset BD among young people. This study shows empirically that there is a strong relationship between problematic substance use and the onset of mental health disorders. Though in practice, the relationship is often complex and bi-directional. Another promising feature of this research was the assessment of both licit and illicit substance use, which is important in the context of public health broadly, and mental health specifically. Typically, alcohol and tobacco are the focus of such research, but studies such as this demonstrate that illicit substances are just as (if not more) likely to increase the risk of mental health disorders and should be considered as part of assessment by mental health professionals.


Studies such as these demonstrate both the potential and the need for targeted early intervention among community mental health settings. Although this study presents a number of strong points, there are some areas which might warrant further investigation. First, the threshold for meeting presence of abuse/dependence of a given substance was reporting a moderate risk score on the ASSIST. Though moderate-risk substance use can lead to serious harmful consequences in the future, this study might have been enhanced by also comparing moderate to high-risk use to understand the added risk of BD, with respect to high-risk use (i.e., compared to moderate-risk). Second, it is somewhat unclear what qualifications the initial investigators had to conduct the screening and assessment. The report suggests investigators were trained, but does not elaborate further, raising the possibility of questionable diagnoses. Finally, this study identified a number of individuals with moderate- and high-risk use. Future research might seek to investigate the possibility of randomly administering a brief intervention for a proportion of those individuals and subsequently comparing the incidence of new-onset BD to investigate whether brief intervention presents a useful preventive approach.


If you currently work in Community Mental Health settings and want to know more about how screening and brief intervention may help reduce the impact of comorbid mental health issues, please visit our Portal to find out more. There you will find downloadable resources about the risks and harms of substance use in general, as well as handy guidelines on screening and brief intervention. We are also working with an expert advisory group to develop a new screening and brief intervention resource aimed at those working in mental health settings. The manual will be available online and promoted through our website when published.


In the coming weeks, we will be releasing a new training package aimed at community mental health workers. You can sign-up here once available


You can read the full report of the cohort study here