Exploring the predictive role of the first mood episode on the predominant polarity in bipolar disorder: insights from a path analysis.

Abstract

Background: The predominant polarity in bipolar disorder (BD) is defined by the skewness of mood episodes towards either the manic or depressive pole. However, since the predominant polarity can only be established over the long term, it is crucial to identify predictors of illness trajectory. Among these factors, the polarity at onset has been suggested to hold important implications, even though research in this field is not entirely consistent so far. In this retrospective study, we thus explored whether the polarity of the first episode can predict the predominant polarity in BD.

Methods: We included subjects with BD consecutively referred to two acute inpatient units in the Milan metropolitan area from May 2020 to January 2024. Following Barcelona criteria, a manic (mPP) and a depressive (dPP) predominant polarity were defined as having a ratio ≥ 2:1 of past manic/hypomanic or depressive episodes, respectively. The relationship between first episode polarity and either mPP or dPP was examined using multivariable logistic regression models. A path analysis was then performed to jointly test the associations between putative variables and the predominant polarity.

Results: This study included 128 participants. Regression models estimated an association between a manic onset and a mPP (β = 3.23, p < 0.001) as well as between a depressive onset and a dPP (β = 3.65, p < 0.001). Participants with a mPP showed a lower age at onset (β = - 0.13, p = 0.004), while subjects diagnosed with BD type I were less likely to show a dPP (β = - 2.09, p = 0.024). The path analysis highlighted an association between earlier onset and the likelihood of a first episode of manic polarity (coeff. = - 1.39, p = 0.021). A manic onset was associated with a higher likelihood of mPP (coeff. = 3.46, p < 0.001) and a lower likelihood of dPP (coeff. = - 3.71, p < 0.001). Consistently, participants with a manic onset were more likely to experience a lower number of depressive episodes (coeff. = - 1.36, p < 0.001). Finally, cannabis use disorder was associated with a lower number of depressive episodes (coeff. = - 0.57, p = 0.011).

Conclusions: These results provide important insights into the likely predictive value of first episode polarity in relation to the predominant polarity in BD. Though future studies validating these findings are needed, the polarity at onset may serve as an early marker for illness trajectory.