Profile of copper-associated DNA methylation and its association with incident acute coronary syndrome

Clinical Epigenetics

Table 4 Adjusted odds ratios for the incident ACS according to quartiles of DNA methylation level

	Quartiles of methylation level^a				p-Trend^b	Linear model^c	p-Linearity^d
	Q1	Q2	Q3	Q4	p-Trend^b	Linear model^c	p-Linearity^d
cg20995564	< 0.3834	0.3834–0.4141	0.4141–0.4473	≥ 0.4473
n (case/control)	85/85	88/82	82/89	86/85
Model 1^e	1.00 (ref)	1.07 (0.70, 1.64)	0.92 (0.60, 1.41)	1.01 (0.66, 1.54)	0.89	1.02 (0.88, 1.19)	0.78
Model 2^f	1.00 (ref)	0.96 (0.59, 1.58)	0.92 (0.56, 1.51)	0.77 (0.47, 1.27)	0.30	0.95 (0.80, 1.13)	0.55
cg18608055	< 0.4760	0.4760–0.5140	0.5140–0.5560	≥ 0.5560
n (case/control)	75/95	92/79	80/90	94/77
Model 1^e	1.00 (ref)	1.48 (0.96, 2.27)	1.13 (0.73, 1.73)	1.55 (1.01, 2.38)	0.12	1.09 (0.94, 1.27)	0.26
Model 2^f	1.00 (ref)	1.53 (0.93, 2.52)	1.06 (0.64, 1.75)	1.23 (0.75, 2.03)	0.74	1.00 (0.84, 1.19)	0.98
cg26470501	< 0.3905	0.3905–0.4170	0.4170–0.4395	≥ 0.4395
n (case/control)	92/79	82/88	80/91	87/83
Model 1^e	1.00 (ref)	0.80 (0.52, 1.22)	0.75 (0.49, 1.15)	0.89 (0.57, 1.37)	0.51	0.90 (0.77, 1.05)	0.19
Model 2^f	1.00 (ref)	0.74 (0.45, 1.23)	0.86 (0.52, 1.43)	1.06 (0.63, 1.79)	0.75	0.96 (0.79, 1.15)	0.63
cg05825244	< 0.4458	0.4458–0.5636	0.5636–0.6648	≥ 0.6648
n (case/control)	74/97	89/81	86/85	92/78
Model 1^e	1.00 (ref)	1.45 (0.95, 2.23)	1.34 (0.87, 2.06)	1.58 (1.02, 2.44)	0.06	1.17 (1.00, 1.37)	0.05
Model 2^f	1.00 (ref)	1.45 (0.88, 2.39)	1.40 (0.85, 2.31)	1.78 (1.07, 2.98)	0.04	1.23 (1.02, 1.48)	0.03

^aDNA methylation level of CpGs was presented as relative level after normalization
^bP for trend across quartiles of CpGs were obtained by including the median of each quartile as a continuous variable in logistic regression models
^cOdds ratios (95% CI) for incident ACS correspond to one-unit increase in inverse normal transformation of methylation level
^dP for linearity were obtained by including inverse normal transformation methylation values as continuous variable in linear regression models
^eModel 1 was adjusted for age and sex in the unconditional logistic regression model
^fModel 2 was additionally adjusted for body mass index, smoking status, drinking status, hypertension, hyperlipidemia, and diabetes in the unconditional logistic regression model

ISSN: 1868-7083