Abstract TH884: Pulse Pressure as a Mediator of the Association Between Cognition and Bone Health in Older Women: Findings from the Study of Muscle, Mobility and Aging- Bone Ancillary Study (SOMMA Bone)

Background: Stiffening arteries, a risk factor of cognitive impairment, leads to increased pulse pressure (PP, difference between systolic and diastolic blood pressure). PP directly relates to bone metabolism by microvascular perfusion and mechanotransduction. We hypothesized that PP may mediate the known association between performance on episodic memory and bone health. Methods: Participants in the Study of Muscle, Mobility and Aging (SOMMA) study were > 70 years, able to complete a 400-meter walk at baseline, and free of advanced chronic disease. At the baseline SOMMA visit, participants completed a California Verbal Learning Test II short form (CVLT-II), a test of episodic memory. The SOMMA-Bone Study at the Pittsburgh site measured bone structure/microarchitecture using High-Resolution peripheral Quantitative Computed Tomography (HR-pQCT) at the distal radius and tibia after the 12-month SOMMA visit (N= 327; aged 76.2± 4.6 years; 60.5% women; 87.2% White). Total BMD (Tt.BMD, mg HA/cm 3 ), total area (Tt.Ar, mm 2 ), and failure load (F.ult, N) from micro-finite element analysis (µFEA) measurements were estimated and standardized by sex. A causal mediation analysis method was used to examine if PP mediated the association of CVLT-II performance and HR-pQCT measurements in women using the mediation package in R, adjusting for relevant confounders. To estimate the indirect (Average Causal Mediation Effect, ACME) and direct (Average Direct Effect, ADE) effects, we used nonparametric bootstrapping with 5000 simulations, assuming no unmeasured confounding. Results: There was no association between episodic memory and bone parameters in men. Of the women in SOMMA-Bone (n=198, 76.2 ± 4.7 years, 83.8% White), mean PP was 59.1±13.8 mmHg. Better performance on the immediate (but not short or long) recall was associated with higher F.ult in the tibia. Better performance on short and long recall tests (but not immediate) were associated with higher Tt. BMD (Table 1). At the radius, better performance on the long recall test was associated with lower Tt.Ar only. In all models, PP was not a significant mediator. Conclusions: In associations between better episodic memory and better skeletal parameters, there was no evidence of mediation through PP. This suggests that cognitive and skeletal health may share underlying biological mechanisms independent of PP in older women. Serum proteomic analysis is underway to identify the biology of this relationship.