Allostatic Load Effects on Cortical and Cognitive Deficits in Essentially Normotensive, Normoweight Patients with Schizophrenia

Abstract

Reduced cortical gray matter integrity and cognitive abilities are among core deficits in schizophrenia. We hypothesized that higher allostatic load (AL) that accounts for exposure to chronic stress is a contributor to structural and cognitive deficits in schizophrenia. One hundred and sixty-seven schizophrenia patients who were on average with normal weight, normal systolic, and diastolic blood pressure and 72 healthy controls were enrolled in the study. Group differences in subclinical cardiovascular, metabolic, immune, and neuroendocrine biological markers as indexed by AL and contribution of AL components to the structural and cognitive deficits in schizophrenia were explored. Compared with controls, schizophrenia patients who were normotensive, normoweight, and had low total cholesterol levels still had significantly higher AL mainly due to lower high-density lipoprotein cholesterol and higher heart rate, waist-hip ratio, hemoglobinA1c, hypersensitive C-reactive protein, and overnight-urine cortisol levels. Patients also had decreased whole-brain mean cortical thickness, and lower cognition assessed by the MATRICS consensus cognitive battery. AL was inversely correlated with mean cortical thickness and cognition in schizophrenia, while none of these relationships existed in controls. Mediation analyses showed the effect of AL on cognitive deficits in schizophrenia was significantly mediated by cortical thinning, and the most significant mediating cortical area was the left superior frontal gyrus. Cortical thickness may act as a mediator between AL and cognitive deficits in schizophrenia. Early intervention strategies to reduce cortical thinning and cognitive dysfunction in schizophrenia should target specific aspects of their high AL in addition to weight gain, hypertension and high cholesterol levels.

Keywords: Schizophrenia; allostatic load; chronic stress; cognitive function; cortical thickness.

Fig. 1.

Cortical areas significantly associated with allostatic load in schizophrenia patients controlled for age, sex, and education years. (A) Cortical areas labeled with numbers were significantly correlated with AL after Bonferroni correction. (B) Cortical areas labeled with number was significantly correlated with MCCB total score in patients with schizophrenia after Bonferroni correction. The color bars represent the partial correlation coefficient r. Annotation in the figure: 1 and 2: left and right superior frontal gyrus; 3 and 4: left and right lateral orbitofrontal cortex; 5: right precuneus; 6: right rostral middle frontal gyrus.

Fig. 2.

Mediation effects. (A). Mediating effect of whole-brain mean cortical thickness. X = allostatic load; Y = MCCB total score; Mediator = whole-brain mean cortical thickness. Path a represents AL effects on whole-brain mean cortical thickness. Path b represents the effects of whole-brain mean cortical thickness on MCCB total score. Age, sex, education years, and CPZ were covariates. (B). Mediating effect of left superior frontal gyrus. X=allostatic load; Y=MCCB total score; Mediator=left superior frontal gyrus. Path a represents AL effects on the left superior frontal gyrus. Path b represents the effects of superior frontal gyrus on MCCB total score. Age, sex, education years, and CPZ were covariates.

Fig. 3.

Regional cortical thickness mediates the effects of allostatic load on MCCB score in schizophrenia patients, controlled for CPZ, age, sex and education level. The color bar represents the standardized coefficients (β) of the indirect effect. * Indicate P value threshold were the β of indirect path (≤ –0.05) was nominally significant. The 15 brain regions were nominally associate with both AL and MCCB at P < .05 uncorrected: left and right superior frontal gyrus (1, 2), left pars triangularis of inferior frontal gyrus (3), left rostral middle frontal gyrus (4), left precentral gyrus (5), left lateral orbitofrontal cortex (6), left pars opercularis of inferior frontal gyrus (7), right precuneus (8), right insula (9), left and right paracentral lobule (10, 11), left medial orbitofrontal cortex (12), right lateral orbitofrontal cortex (13), right pars opercularis of inferior frontal gyrus (14), and right frontal pole (15). Models using regions 1–5 showed nominally significant mediation effect; region 1 showed the strongest effect.