. Author manuscript; available in PMC: 2022 Oct 5.

Published in final edited form as: Sci Transl Med. 2020 Nov 11;12(569):eaay1913. doi: 10.1126/scitranslmed.aay1913

Table 3:

Association of NFL concentrations with ICH volume in patients with ICH from the validation cohorts

		Association between NFL and ICH volume
	N	Unadjusted analysis		Multivariable analysis¹
Group		β (95% CI)	P-value	β (95% CI)	P-value
Yale ICH study series	96	0.62 (0.30, 0.94)	<0.001	0.60 (0.27, 0.93)	<0.001
UPenn ICH study series
0–12 hours from stroke to blood draw	36	0.13 (−0.42, 0.68)	0.65	0.12 (−0.53, 0.77)	0.72
12–36 hours from stroke to blood draw	40	0.63 (0.18, 1.07)	0.009	0.74 (0.21, 1.27)	0.010
60–84 hours from stroke to blood draw	29	0.71 (0.12, 1.31)	0.026	0.51 (−0.17, 1.19)	0.16

β=regression coefficient; CI=confidence interval; β values, 95% CIs, and p-values result from linear regression models. β values are interpreted as the change in mean ICH volume (on the square root scale) for each doubling in NFL concentration.

Multivariable models were adjusted for age, sex, hypertension, diabetes, and BMI for the Yale ICH study series, and for age, sex, hypertension, and diabetes for the longitudinal UPenn ICH series. P-values < 0.0125 are considered statistically significant after correcting for multiple testing.