Table 2.

Log(NT-proBNP) and global brain volumes in the CARDIA brain magnetic resonance imaging sample.

Dependent variable	βa	95% CI	p-Value
Total brain volume (mL)
Model 1	−2.73	−5.69, 0.23	0.071
Model 2	−2.99	−5.98, −0.0047	0.050
Model 3	−2.31	−5.50, 0.89	0.156
Total gray matter (mL)
Model 1	−3.41	−5.63, −1.19	0.003
Model 2	−3.44	−5.68, −1.20	0.003
Model 3	−2.93	−5.32, −0.53	0.017
Total white matter (mL)
Model 1	0.68	−1.84, 3.19	0.60
Model 2	0.45	−2.13, 3.029	0.73
Model 3	0.62	−2.14, 3.38	0.66
Abnormal white matter loadb
Model 1	0.02	−0.30, 0.34	0.88
Model 2	−0.12	−0.46, 0.22	0.50
Model 3	−0.02	−0.38, 0.34	0.92
White matter integrityc
Model 1	−0.001	−3.48, 0.0009	0.25
Model 2	−0.001	−3.34, 0.001	0.30
Model 3	−0.0007	−0.003, 0.002	0.57

Model 1: adjusted for age, sex, race, and intracranial volume.

Model 2: Model 1 covariates + smoking status, anti-hypertension medication use, systolic blood pressure, diastolic blood pressure, cardiovascular history, renal history, body mass index, total cholesterol, glomerular filtration rate, education, exam center.

Model 3: Model 2 covariates + cardiac output.

^aCoefficient represents the difference in brain tissue volume (milliliter) corresponding to 1 unit increase in logNT-proBNP (picograms/milliliter).

^bLogistic regression models with outcome as dichotomous variable split into high (top 15% of cohort) versus low (bottom 85% of cohort).

^cMeasured by fractional anisotropy.