Table 2.
Model Coefficients of Statistical Relationship of cSVD Indices With Framingham Risk Score (FRS)
| Framingham risk score | ||||
| PVS-BG | R2 = 0.21 | |||
| Reference: PVS-BG = 0 (mean FRS = 12.80) | ||||
| Mean FRS | OR | CI | p Value | |
| 1 | 14.41 | 1.08 | 1.01–1.17 | 0.026 |
| 2 | 16.46 | 1.27 | 1.16–1.38 | <0.001 |
| 3 | 16.98 | 1.41 | 1.25–1.59 | <0.001 |
| 4 | 18.33 | 1.55 | 1.18–2.01 | 0.001 |
| PVS-CS | R2 = 0.22 | |||
| Reference: PVS-CS = 0 (mean FRS = 12.53) | ||||
| Mean FRS | OR | CI | p Value | |
| 1 | 14.46 | 1.14 | 1.08–1.21 | <0.001 |
| 2 | 14.73 | 1.22 | 1.15–1.29 | <0.001 |
| 3 | 16.99 | 1.38 | 1.28–1.48 | <0.001 |
| 4 | 19.11 | 1.57 | 1.36–1.82 | <0.001 |
| FZKS DWMH | R2 = 0.21 | |||
| Reference: FZKS DWMH = 0 (mean FRS = 13.12) | ||||
| Mean FRS | OR | CI | p Value | |
| 1 | 14.77 | 1.11 | 1.07–1.15 | <0.001 |
| 2 | 16.62 | 1.26 | 1.21–1.33 | <0.001 |
| 3 | 16.95 | 1.43 | 1.27–1.59 | <0.001 |
| FZKS PVH | R2 = 0.22 | |||
| Reference: FZKS PVH = 0 (mean FRS = 13.56) | ||||
| Mean FRS | OR | CI | p Value | |
| 1 | 15.56 | 1.14 | 1.11–1.18 | <0.001 |
| 2 | 17.06 | 1.29 | 1.22–1.35 | <0.001 |
| 3 | 17.58 | 1.46 | 1.23–1.72 | <0.001 |
| CMB lobar | R2 = 0.22 | |||
| Reference: CMB lobar <2 (mean FRS = 14.61) | ||||
| Mean FRS | OR | CI | p Value | |
| ≥2 | 18.12 | 3.151 | 1.57–6.61 | <0.001 |
| CMB deep | R2 = 0.13 | |||
| Reference: CMB deep <1 (mean FRS = 14.64) | ||||
| Mean FRS | OR | CI | p Value | |
| ≥1 | 16.57 | 1.65 | 0.93–2.96 | p = 0.024 |
| Lacunes | R2 = 0.18 | |||
| Reference: lacunes = 0 (mean FRS = 14.67) | ||||
| Mean FRS | OR | CI | p Value | |
| 1 | 16.41 | 1.11 | 1.05–1.17 | <0.001 |
| 2 | 16.61 | 1.19 | 1.05–1.34 | 0.003 |
| >2 | 15.5 | 1.12 | 0.98–1.29 | 0.10 |
Abbreviations: CMB = cerebral microbleed; cSVD = cerebral small vessel disease; FZKS DWMH = Fazekas deep white matter hyperintensity; FZKS PVH = Fazekas periventricular hyperintensity; PVS-BG = perivascular spaces in basal ganglia; PVS-CS = perivascular spaces in centrum semiovale.
FRS was used as a predictor in multinomial or logistic regression models. For these models, odds ratios, confidence intervals (95%), and p values are reported. Odds ratios represent the increase (>1) in the risk of falling in a category (compared with the reference) for a 1-unit increase in the predictor variable. For multinomial models, goodness-of-fit metrics (McFadden R2) are reported. Coefficients refer to comparison with reference groups, and other group contrasts are listed in eTable 7.