Sensitive content warning: This paper contains sensitive content related to suicide.

Suicide is one of the leading causes of global mortality, making risk detection a critical public health priority. Although psychological studies have identified a link between personality and suicide, this relationship has received little attention and has not yet been systematically examined or applied in the context of large language models (LLMs). In this study, we investigate whether LLMs can capture and leverage this relationship for suicide severity detection. We propose incorporating persona profiles based on the Big Five personality traits to simulate user characteristics in a zero-shot setting, enabling us to assess the impact of personality on detection performance. Experimental results show that Neuroticism yields a modest but consistent improvement across most models, with statistically significant gains in several cases. These findings suggest that LLMs can capture inter-domain relationships, and that further refinement of personality-aware modelling may lead to performance improvements.