UC San Diego

The study was to explored potential mechanisms underlying the association between sleep disordered breathing (SDB) and clinical outcomes in a convenience sample heart failure (HF) patients (n=66) and controls (n=53). Subjects participated in a sleep study, and physical assessments and blood draws were performed. Longitudinal outcomes were extracted from the HF patient medical records, and deaths were confirmed using the Social Security Death Index. Depressive symptoms were assessed using the Beck Depression Inventory (BDI). Mediation models were performed, and a test of product coefficients (PRODCLIN) was used to assess statistical significance of compound paths. Regression analyses showed disease status, but not SDB severity, predicted BDI scores and inflammation levels (p<.01). There was a significant second-order effect for two measures of hypoxia (average % oxygen saturation and lowest % oxygen saturation) by disease status for somatic BDI scores (p<.05), and a significant second-order effect for indices of obstructive and mixed apnea-hypopnea by disease status for C-Reactive Protein (CRP; p<.05). Simple slopes analyses revealed that these interaction effects were only significant for controls. Controls evidenced an expected association where worse SDB related to higher somatic scores and higher resting mean levels of CRP, respectively. In HF patients somatic symptom scores remained relatively constant, and CRP levels decreased in the presence of higher SDB severity. During the 3 year follow-up, HF patients experienced an average of 3 (SD=3.9) hospital admissions and 18% (n=12) died. There was a moderated-mediated effect of cognitive-affective depressive symptomatology and resting levels of circulating CRP on the association between total minutes oxygen saturation fell below 90% and mortality (p<.05). Among patients with lower cognitive-affective BDI scores, higher resting levels of CRP corresponded with more minutes that oxygen saturation was below 90%. Among those with higher cognitive-affective scores, lower resting levels of CRP corresponded with more minutes oxygen saturation was lower than 90% (p<.05). In those with lower cognitive-affective scores, higher resting CRP levels corresponded with greater percentage of mortality. Among those with higher cognitive-affective scores, higher resting mean levels of CRP corresponded to fewer deaths. Simple slopes analyses were not significant. The patterns seen in HF patients may be evidence of dysregulation of inflammatory responses resulting from HF syndrome and/or cognitive-affective aspects of depression. Inflammatory mechanisms may mediate associations between SDB and morbidity and mortality, although this association appears to vary by levels of cognitive-affective depressive symptoms