Table 2. Association between brain volumes and change in tCBF.

	Model I			Model II
	tCBF follow-up			tCBF follow-up
	n	β	95% CI	n	β	95% CI
TBV baseline, per s.d. increase
Whole population	3,011	0.296	0.200; 0.393	2,760	0.314	0.212; 0.416
<65 years	2,375	0.300	0.188; 0.412	2,182	0.310	0.191; 0.428
⩾65 years	636	0.264	0.071; 0.457	578	0.305	0.101; 0.510
GM baseline, per s.d. increase
Whole population	3,011	0.177	0.110; 0.244	2,760	0.187	0.116; 0.258
<65 years	2,375	0.178	0.101; 0.254	2,182	0.183	0.102; 0.264
⩾65 years	636	0.152	0.013; 0.291	578	0.181	0.031; 0.330
WM baseline, per s.d. increase
Whole population	3,011	0.161	0.099; 0.222	2,760	0.171	0.106; 0.235
<65 years	2,375	0.163	0.092; 0.234	2,182	0.169	0.094; 0.244
⩾65 years	636	0.147	0.027; 0.268	578	0.169	0.042; 0.296

Abbreviations: CI, confidence interval; GM, gray matter; n, number of persons in the analysis of total study population (N=3,011); TBV, total brain volume; tCBF, total cerebral blood flow; WM, white matter. Values represent adjusted mean difference in Z-scores of tCBF (with 95% CI) per s.d. increase in TBV, GM or WM for the whole population and stratified on age (<65 years vs. ⩾65 years). Values in bold represent significance findings after Sidak correction for four correlated tests (P<0.029).

Model I: adjusted for age, age × age (only in age overall analyses), sex, scan interval, intracranial volume at baseline, tCBF at baseline, GM volume and WM volume were adjusted for each other.

Model II: Model I+additionally adjusted for systolic blood pressure, diastolic blood pressure, antihypertensive medication, body mass index, diabetes mellitus, high-density lipoprotein-cholesterol level in serum, total cholesterol level in serum, lipid-lowering medication, smoking, APOE-ɛ4 status, and carotid stenosis.