Allostatic Load and Severe Clinical Phenotypes in Major Depressive Disorder: A Cross-Sectional Statistical Mediation Analysis of Perceived Stress, Cognitive Impairment, and Suicidal Ideation.
Perceived stress is associated with adverse clinical features of major depressive disorder (MDD), but the role of multisystem physiological dysregulation remains uncertain. We examined whether allostatic load (AL) was statistically associated with the relationship between perceived stress and cognitive impairment or severe suicidal ideation.
This retrospective, cross-sectional electronic medical record study included 385 adults with DSM-5 MDD assessed within a single three-day index window. Perceived stress was measured with the 14-item Perceived Stress Scale (PSS-14), and chronic exposure context was characterized from documented negative life events. AL was calculated from 15 cardiovascular, metabolic, immune-inflammatory, and neuroendocrine indicators. Cognitive impairment was defined as Montreal Cognitive Assessment (MoCA) <26 and severe suicidal ideation as Columbia-Suicide Severity Rating Scale (C-SSRS) ideation severity ≥4. Sequential logistic regression and bootstrap product-of-coefficients statistical mediation analyses were performed.
Compared with Low AL, High AL was associated with cognitive impairment (adjusted odds ratio [aOR] 6.17, 95% CI 2.89-13.15) and severe suicidal ideation (aOR 6.43, 95% CI 2.98-13.88; both P<0.001). Continuous AL showed consistent associations with both outcomes. The indirect effect through AL was supported for cognitive impairment (0.39, 95% bootstrap CI 0.16-0.69), but not for severe suicidal ideation (0.14, 95% bootstrap CI -0.12 to 0.41).
Higher AL was independently associated with both adverse clinical phenotypes. Statistical mediation was supported for the perceived stress-cognitive impairment association but not for suicidal ideation. Because exposure, mediator, and outcomes were assessed within the same index window, temporal and causal mediation cannot be inferred.
This retrospective, cross-sectional electronic medical record study included 385 adults with DSM-5 MDD assessed within a single three-day index window. Perceived stress was measured with the 14-item Perceived Stress Scale (PSS-14), and chronic exposure context was characterized from documented negative life events. AL was calculated from 15 cardiovascular, metabolic, immune-inflammatory, and neuroendocrine indicators. Cognitive impairment was defined as Montreal Cognitive Assessment (MoCA) <26 and severe suicidal ideation as Columbia-Suicide Severity Rating Scale (C-SSRS) ideation severity ≥4. Sequential logistic regression and bootstrap product-of-coefficients statistical mediation analyses were performed.
Compared with Low AL, High AL was associated with cognitive impairment (adjusted odds ratio [aOR] 6.17, 95% CI 2.89-13.15) and severe suicidal ideation (aOR 6.43, 95% CI 2.98-13.88; both P<0.001). Continuous AL showed consistent associations with both outcomes. The indirect effect through AL was supported for cognitive impairment (0.39, 95% bootstrap CI 0.16-0.69), but not for severe suicidal ideation (0.14, 95% bootstrap CI -0.12 to 0.41).
Higher AL was independently associated with both adverse clinical phenotypes. Statistical mediation was supported for the perceived stress-cognitive impairment association but not for suicidal ideation. Because exposure, mediator, and outcomes were assessed within the same index window, temporal and causal mediation cannot be inferred.