Table 3.

Associations between Alzheimer’s disease (AD) and late-onset epilepsy (LOE), UCLA patient population^a.

	AD ~ LOE			LOE ~ AD
	N	HR [95% CI]	P-value	N	HR [95% CI]	P-value
UCLA full sample (N = 416,212)
Basic modela
Cox proportional hazard	415,623	6.86 (6.02, 7.81)	< 0.001*	415,364	6.31 (5.49, 7.25)	< 0.001*
Fine and Gray	415,623	2.54 (2.12, 3.06)	< 0.001*	415,364	2.13 (1.36, 3.34)	0.001*
UCLA ATLAS sample (N = 16,500)
Basic modela
Cox proportional hazard	16,480	10.0 (6.43, 15.6)	< 0.001*	16,462	6.71 (4.06, 11.1)	< 0.001*
Fine and Gray	16,480	3.55 (2.27, 5.55)	< 0.001*	16,462	2.26 (1.31, 3.88)	0.003*
Adjusted for AD geneticsb
Cox proportional hazard	16,480	10.2 (6.53, 15.9)	< 0.001*	16,462	6.55 (3.96, 10.9)	< 0.001*
Fine and gray	16,480	3.77 (2.43, 5.83)	< 0.001*	16,462	2.20 (1.28, 3.81)	0.005*

CI confidence interval, EHR electronic health record, HR hazard ratio.

[a]Basic Model adjustments include age at first visit, sex, EHR record length, hypertension, diabetes, stroke, and hyperlipidemia status.

[b]Additionally adjusted for APOE-ε4 count and AD polygenic risk score (without APOE region) in [b].