Table 4.

Results from ‘full’ logistic regression models used to construct the risk scoring system. Weighted analyses are based on weights derived from the saturated selection model. P-values marked with a * indicate variables retained by the backwards stepwise procedure in ‘final’ prediction models.

	NP measures only						NP measures and demographics
	Unadjusted			Weighted (Saturated)			Unadjusted			Weighted (Saturated)
	OR	95% CI	p-value	OR	95% CI	p-value	OR	95% CI	p-value	OR	95% CI	p-value
BRAAK stage
0/I/II/III/IV	1.00	REF	< 0.001	1.00	REF	< 0.001*	1.00	REF	< 0.001	1.00	REF	< 0.001*
V/VI	25.26	8.70, 73.35		27.38	9.42, 79.63		24.77	6.79, 90.34		25.48	6.28, 103.37
Cerebral microinfarcts
0,1,2	1.00	REF	< 0.001	1.00	REF	< 0.001*	1.00	REF	0.001	1.00	REF	< 0.001*
≥ 3	7.17	2.60, 19.74		10.00	3.33, 29.96		7.15	2.31, 22.14		9.30	2.99, 28.90
Neocordical Lewy bodies
0	1.00	REF	0.050	1.00	REF	0.001*	1.00	REF	0.208	1.00	REF	0.019*
≥ 1	5.76	1.00, 33.04		13.45	2.87, 63.00		4.04	0.46, 35.55		9.79	1.46, 65.53
Cystic infarcts
0	1.00	REF	0.111	1.00	REF	0.017*	1.00	REF	0.027	1.00	REF	0.003*
≥ 1	1.90	0.86, 4.17		2.70	1.19, 6.11		3.25	1.15, 9.24		5.08	1.75, 14.81
Amyloid angiopathy
None	1.00	REF	0.189	1.00	REF	0.089	1.00	REF	0.190	1.00	REF	0.080
Mild/Moderate/Severe	1.88	0.73, 4.84		2.34	0.88, 6.25		2.16	0.68, 6.83		2.83	0.88, 9.08
Age (at death) a							1.35	1.15, 1.59	< 0.001	1.38	1.17, 1.63	< 0.001*
Birth cohorta Gender							1.27	1.09, 1.47	0.002	1.27	1.09, 1.48	0.002*
Female							1.00	REF	0.293	1.00	REF	0.355
Male							1.53	0.69, 3.37		1.48	0.65, 3.38
Education
High school or less							1.00	REF	0.204	1.00	REF	0.122
At least some college							0.59	0.26, 1.34		0.51	0.21, 1.20
APOE ε4 allele
None							1.00	REF	0.683	1.00	REF	0.812
At least one							1.21	0.48, 3.05		1.13	0.42, 3.03

^aModeled as a continuous linear term