Prior research suggests that certain psychiatric symptoms could be associated with increased risk of death. However, it remains unclear whether this association could rely on all or specific symptoms. In this report, we used data from a multicenter 5-year prospective study (N = 641) of older adults with an ICD-10 diagnosis of schizophrenia, bipolar disorder or major depressive disorder, recruited from French community psychiatric departments. We used a latent variable approach to disentangle the effects shared by all psychiatric symptoms (i.e., general psychopathology factor) and those specific to individual psychiatric symptoms, while adjusting for sociodemographic and clinical factors. Psychiatric symptoms were assessed face-to-face by psychiatrists trained to semi-structured interviews using the Brief Psychiatric Rating Scale (BPRS). Among older adults with major psychiatric disorders, we found that all psychiatric symptoms were associated with increased mortality, and that their effect on the 5-year mortality were exerted mostly through a general psychopathology dimension (β = 0.13, SE = 0.05, p < 0.05). No BPRS item or lower order factor had a significant effect on mortality beyond and above the effect of the general psychopathology factor. Greater number of medical conditions, older age, male sex, and being hospitalized or institutionalized at baseline were significantly associated with this risk beyond the effect of the general psychopathology factor. Since psychiatric symptoms may affect mortality mainly through a general psychopathology dimension, biological and psychological mechanisms underlying this dimension should be considered as promising targets for interventions to decrease excess mortality of older individuals with psychiatric disorders.