Table 3.

Association between retinal vessel diameters and incident obesity, by current smoking

	Current smoker, absent		Current smoker, present
	No. at risk (cases)	Multivariable-adjusted relative risk (95% confidence interval)*	No. at risk (cases)	Multivariable-adjusted relative risk (95% confidence interval)*
Retinal arteriolar diameter, μm
Tertile 3 (156.0-199.7)	456 (107)	1 (referent)	97 (17)	1 (referent)
Tertile 2 (144.6-155.9)	547 (128)	1.17 (0.87-1.58)	150 (44)	1.65 (1.00-2.72)
Tertile 1 (89.8-144.5)	595 (141)	1.39 (0.99-1.96)	244 (49)	0.92 (0.45-1.88)
p-trend		0.06	491 (110)	0.76
Retinal venular diameter, μm
Tertile 1 (165.9- 220.0)	627 (122)	1 (referent)	72 (9)	1 (referent)
Tertile 2 (220.1-237.3)	548 (136)	1.38 (1.04-1.84)	153 (38)	2.46 (1.10-5.52)
Tertile 3 (237.3-339.7)	423 (118)	1.70 (1.21-2.38)	266 (63)	2.31 (1.01-5.32)
p-trend		0.002	491 (110)	0.14

^*Adjusted for age (years), sex (men, women), education categories (<high school, high school, >high school), alcohol intake (never, former, current), exercise frequency (number of times per week), mean arterial blood pressure (mm Hg), serum total cholesterol (mg/dL), C-reactive protein (mg/dL), fellow vessel diameter (μm, central retinal arteriolar in models of central retinal venular diameter and vice-versa)