Table 3.

Multiple linear regression models estimating associations between sleep measures and markers of inflammation.

	CRPa		IL-6a		TNF-αa
	b (SE)	p-value	b (SE)	p-value	b (SE)	p-value
Sleep duration (linear)
Model 1	0.043 (0.018)	0.020	0.029 (0.015)	0.061	0.005 (0.010)	0.598
Model 2	0.046 (0.018)	0.013	0.030 (0.015)	0.050	0.007 (0.010)	0.497
Model 3	0.041 (0.018)	0.026	0.028 (0.015)	0.067	0.007 (0.010)	0.518
Sleep duration (quadratic)b
Model 1	0.024 (0.011)	0.037	0.033 (0.010)	0.001	0.010 (0.006)	0.107
Model 2	0.021 (0.012)	0.073	0.028 (0.010)	0.004	0.007 (0.006)	0.277
Model 3	0.018 (0.012)	0.134	0.026 (0.010)	0.009	0.008 (0.007)	0.241
Insomnia Rating Scale
Model 1	0.000 (0.004)	0.951	−0.002 (0.004)	0.599	−0.002 (0.002)	0.467
Model 2	−0.001 (0.004)	0.826	−0.003 (0.004)	0.359	−0.003 (0.002)	0.270
Model 3	−0.002 (0.005)	0.733	−0.005 (0.004)	0.205	−0.003 (0.003)	0.292

^anatural log transformed

^bsleep duration as a linear term was included as a covariate in all models.

Model 1: covariates include age, gender, education, and body mass index

Model 2: includes additional adjustments for smoking status, alcohol use, presence of cardiovascular disease and diabetes, number of chronic medical conditions, statin use, and anti-inflammatory medication use

Model 3 includes additional adjustments for psychotropic medication use and psychiatric diagnostic status.