Cross-sectional associations of plasma vitamin D with cerebral β-amyloid in older adults at risk of dementia

Table 3 Sensitivity analyses exploring the association of 25(OH)D with cerebral Aβ

	Unadjusted model			Adjusted model
	B coefficient or odds ratio*	SE or 95% CI*	p	B coefficient or odds ratio*	SE or 95% CI*	p
Sensitivity analysis A: 25(OH)D in classes^a
25(OH)D < 20 ng/ml	0.005	0.033	0.869	0.010	0.034	0.775
25(OH)D ≥ 20 ng/ml ≤ 30 ng/ml	0.061	0.036	0.089	0.039	0.035	0.273
Sensitivity analysis B: logistic regression^b
25(OH)D < 20 ng/ml	1.021*	0.477,2.188*	0.957	1.038*	0.421,2.557*	0.936
25(OH)D ≥ 20 ng/ml ≤ 30 ng/ml	1.059*	0.468,2.395*	0.891	0.769*	0.299,1.978*	0.585
Sensitivity analysis C: de-seasonalized 25(OH)D (continuous)^c	−0.001	0.001	0.600	−0.001	0.001	0.421

^aMultiple linear regression was run to explore the associations between cortical Aβ and 25(OH)D classified according to the cut-off values < 20 ng/ml, ≥ 20 but ≤ 30 ng/ml and > 30 ng/ml adjusting for all confounders
^bLogistic regression was performed with Aβ dichotomized with a positivity threshold of mean cortical standard uptake value ratio ≥ 1.17 and 25(OH)D in classes with adjustment for all confounders
^cDe-seasonalized 25(OH)D values were used to assess the relationship between cortical Aβ and 25(OH)D both as continuous variables using a multiple linear regression model adjusted for all confounders. The adjusted models contained fewer subjects (n = 176) due to missing data on confounders
Aβ beta-amyloid, CI confidence interval, SE standard error, 25(OH)D 25-hydroxyvitamin D

ISSN: 1758-9193