Identifying cognitive subgroups in bipolar disorder: A cluster analysis

Background: Evidence has shown heterogeneity of cognitive function among patients with bipolar disorder (BD). Our study aims to replicate recent findings of cognitive subgroups, as well as we assessed subjective cognitive difficulties and functioning in each cluster. Methods: Hierarchical cluster analysis was conducted to examine whether there were distinct neurocognitive subgroups based on neurocognitive battery. Cognitive Complaints in Bipolar Disorder Rating Assessment (COBRA) and Functioning Assessment Short Test (FAST) were used to assess subjective cognitive difficulties and functional impairment. Results: We found three distinct subgroups: a first cluster with intact cognition (n = 30, 43.5%), a second cluster with selective cognitive impairment (n = 23, 33.3%), and a third cluster with globally cognitive impairment (n = 16, 23.3%). The intact group had more years of education (p < .001) and higher estimated IQ (p < .001) than globally and selectively impaired subgroups. Additionally, they were younger (p = .011), had an earlier age at bipolar diagnosis (p < .037) and earlier age of first hospitalization (p < .035) compared to individuals with globally cognitive impairment. Limitations: This is a cross-sectional design with a small sample including only patients from a tertiary hospital. Conclusion: Our results give support to the existence of a continuum of severity from patients without impairment to those with poor cognitive functioning. Patients in the intact group seem to have higher cognitive reserve than other two groups. However, they also experienced cognitive complaints and some degree of functional impairment. These findings suggest the importance of using a combo of instruments (e.g., objective and subjective cognitive measures plus functioning instruments) for a complete assessment of patients with BD.