The role of type 2 diabetes in the association between habitual glucosamine use and dementia: a prospective cohort study

Table 3 Interaction effect of glucosamine use and APOE ε4 genotype on the risk of dementia onset

	Risk of dementia by exposure, HR (95% CI)				Additive interaction (95% CI)			Multiplicative interaction (95% CI) ^a
	Glucosamine users and APOE ε4 non-carrier	Glucosamine users and APOE ε4 carrier	Glucosamine non-users and APOE ε4 non-carrier	Glucosamine non-users and APOE ε4 carrier	AP	RERI	SI	Multiplicative interaction (95% CI) ^a
Model 1^c	1 (Reference)	2.74 (2.48–3.03)	1.18 (1.10–1.28)	2.95 (2.72–3.19)	−0.142 (−0.641, 0.136)	−0.388 (−0.641, −0.136)	0.818 (0.718, 0.931)	1.25 (1.11–1.41) P<0.001
Model 2^d	1 (Reference)	2.93 (2.63–3.27)	1.20 (1.10–1.31)	3.04 (2.78–3.32)	−0.105 (−0.206, −0.004)	−0.308 (−0.592, −0.023)	0.863 (0.754, 0.988)	1.22 (1.08–1.38) P=0.002
Model 3^e	1 (Reference)	2.75 (2.49–3.04)	1.14 (1.05–1.24)	2.80 (2.58–3.04)	−0.072 (−0.167, 0.024)	−0.197 (−0.451, 0.058)	0.899 (0.784, 1.031)	1.26 (1.12–1.42) P<0.001
Dementia subtype^b
AD	1 (Reference)	4.07 (3.38–4.89)	0.95 (0.80–1.13)	3.83 (3.26–4.50)	0.069 (−0.063, 0.201)	0.280 (−0.271, 0.832)	1.101 (0.909, 1.333)	1.28 (1.01–1.61) P=0.039
VD	1 (Reference)	2.72 (2.03–3.63)	1.24 (0.99–1.56)	3.00 (2.38–3.78)	−0.193 (−1.073, 0.688)	−0.523 (−2.459, −1.412)	0.766 (0.234, 2.515)	1.30 (0.92–1.83) P=0.137

Abbreviations: HR Hazard ratio, CI Confidence interval, AP Attributable proportion due to interaction, RERI Relative excess risk due to interaction, SI Synergy index, AD Alzheimer’s disease, VD Vascular dementia
^aNull hypothesis for each interaction is AP = 0, RERI = 0, SI = 1, and multiplicative interaction = 1
^bSubgroup analysis for dementia subtype was based on Model 3
^cModel 1: Cox proportional hazards regression adjusted for age (timescale) and sex
^dModel 2: Cox proportional hazards regression adjusted for age (timescale), sex, ethnicity, BMI, education attainment, Townsend deprivation index, smoking status, alcohol consumption, family history of dementia, hypertension, arthritis, total cholesterol, and healthy diet
^eModel 3: Cox proportional hazards regression adjusted for age (timescale), sex, ethnicity, BMI, education attainment, Townsend deprivation index, smoking status, alcohol consumption, family history of dementia, hypertension, arthritis, total cholesterol, healthy diet, mineral and vitamin supplements use, and aspirin use

ISSN: 1758-9193