Table 4.

Associations Between rFSRP and Neuropsychological Test Scores and MRI-Derived Indices: Summary of Regression Results

Functional Severity	β	b (95% CI)	P
CDR Sum of Boxes	0.02	0.005 (0.001, 0.01)	0.028
Neuropsychological Test Raw Score	β	b (95% CI)	P
Mini-Mental State Examination	0.002	0.001 (−0.006, 0.008)	0.87
Trail Making Test Part A	0.05	0.14 (0.08, 0.19)	<0.001
Trail Making Test Part B	0.06	0.51 (0.35, 0.67)	<0.001
WAIS-R Digit Symbol	−0.06	−0.09 (−0.12, −0.06)	<0.001
Digit Span Forward	−0.01	−0.002 (−0.007, 0.002)	0.50
Digit Span Backward	−0.02	−0.005 (−0.009, −0.0003)	0.09
Logical Memory Immediate Recall	−0.01	−0.003 (−0.01, 0.01)	0.60
Logical Memory Delayed Recall	−0.01	−0.01 (−0.02, 0.01)	0.46
Semantic Fluency (Animal and Vegetable Fluency summary)	−0.02	−0.02 (−0.04, 0.004)	0.21
Boston Naming Test	0.01	0.01 (−0.01, 0.02)	0.46
MRI-Derived Indices (as quartiles)	Odds Ratio	95% CI	P
Total Brain Volume, cm3	1.01	0.98, 1.03	0.61
Hippocampal Volume, cm3	1.01	0.99, 1.03	0.19
Log-White Matter Hyperintensities, cm3	1.02	1.004, 1.04	0.015

The rFSRP estimates 10-year probability of incident stroke expressed as a percent. Multivariable linear regression examined the effect of the rFSRP on the CDR Sum of Boxes and each neuropsychological test score, whereas ordinal logistic regression tested the effect of the rFSRP on quartile rankings of the MRI indices. Covariates included in the models were age at baseline, sex (1 = female, 0 = male), racial identity (1 = White, 0 = other), years of education, and APOE e4 allele status (0 = absent, 1 = homozygotes, 2 = heterozygotes). For MRI indices, estimated intracranial volume and age at MRI (as opposed to age at baseline) were included as covariates. Note that no statistically significant effects were found for log-white matter hyperintensities when examined as a continuous variable. P-values for analyses examining neuropsychological test scores are False Discovery Rate (FDR)-adjusted due to the number of analyses performed for this domain (i.e., 10 neuropsychological test scores)

Sample size for the CDR and neuropsychological tests was 19,309 with the exception of the WAIS-R Digit Symbol for which the sample was reduced to 16,447 due to missing data. Sample size for MRI indices was 1,196. Higher scores are worse for CDR Sum of Boxes, Trail Making Test Part A, Trail Making Test Part B, and volume of log-white matter hyperintensities. Lower scores are worse for all other outcomes.

Abbreviations; β = standardized beta; b = unstandardized beta; MRI = magnetic resonance imaging; CDR = CDR Dementia Staging Instrument; WAIS-R = Wechsler Adult Intelligence Scale, Revised version