Table 2.

Influence of potential confounders on the rate of VRS detected on both in vivo and post-mortem MRI.

	Deep gray matter model			White matter model
Predictors	Est.(β1)	CI (95%)	p	Est.(β1)	CI (95%)	p
Intercept	98.45	–81.27–278.17	0.242	67.30	29.66–104.94	0.001
Single hemispheric block	53.73	23.34–84.12	0.004	1.40	–12.40–15.20	0.840
n VRS (p.m.−i.v.)/(p.m.+i.v.)	–14.27	–30.85–2.31	0.082	14.07	–1.72–29.86	0.080
Time: MRI to death [d]	0.04	–0.48–0.56	0.854	–0.08	–0.20–0.04	0.165
Time: death to MRI [d]	–0.10	–0.38–0.19	0.440	–0.01	–0.08–0.06	0.728
Total number of blocks/patient	–2.99	–18.97–12.98	0.663	1.46	–2.23–5.14	0.389

Note: Linear mixed effect models, with tissue block nested in subjects as a random factor, were computed for the following confounders: total number of VRS in vivo, total number of VRS post-mortem, the time interval from in vivo MRI to death, and the time interval from death to the post-mortem MRI. None of the confounders had a significant effect size.

MRI: magnetic resonance imaging; VRS: Virchow-Robin spaces.