TABLE 5.

Relative risk (95% CI) of sudden cardiac death by quartile (Q) of plasma magnesium¹

	Plasma magnesium (mg/dL)
	Q1	Q2	Q3	Q4	P for trend2
Range of magnesium (mg/dL)	<1.9	1.9–2.0	2.1–2.1	>2.1
Cases/controls (n)	30/54	31/89	14/56	24/95
Median magnesium in cases (mg/dL)	1.8	1.9	2.1	2.2
Median magnesium in controls (mg/dL)	1.7	1.9	2.1	2.3
Multivariate model 1	1.0 (ref)	0.61 (0.33, 1.12)	0.42 (0.20, 0.90)	0.39 (0.20, 0.78)	0.006
Multivariate model 2	1.0 (ref)	0.47 (0.23, 0.95)	0.31 (0.13, 0.74)	0.23 (0.10, 0.53)	0.001
Multivariate model 3	1.0 (ref)	0.50 (0.23, 1.09)	0.33 (0.13, 0.86)	0.19 (0.08, 0.50)	0.001
Multivariate model 4	1.0 (ref)	0.56 (0.25, 1.25)	0.41 (0.15, 1.10)	0.23 (0.09, 0.60)	0.003

¹ref, reference. Multivariate model 1 was a Cox proportional hazards model adjusted for age and fasting. Multivariate model 2 was adjusted as for model 1 plus BMI (in kg/m²; <25, 25–29.9, or ≥30), parental history of myocardial infarction before age 60 y (yes or no), alcohol intake (<0.1, 0.1–14.9, 15–29.9, or ≥30 g/d), physical activity (quintiles of metabolic equivalent task hours/wk), postmenopausal hormone use, use of thiazide diuretics (yes or no), aspirin use >22 d/mo (yes or no), and intake of magnesium (mg/d), long-chain omega-3 (n−3) fatty acids (% of energy), calcium (mg/d), potassium (mg/d), and vitamin D (IU/d) (all in quartiles). Multivariate model 3 was adjusted as for model 2 plus total:HDL cholesterol, glomerular filtration rate, C-reactive protein, and N-terminal pro-B type natriuretic peptide. Multivariate model 4 was adjusted as for model 3 plus history of diabetes and hypertension.

²Estimated by assigning the median value of plasma magnesium in each quartile and modeling this as a continuous variable in regression models.