Table 3.

Results of linear mixed regression models predicting the frailty index (FI) from difference-based methylation age acceleration

Covariables	Dataset 1	Dataset 2	Overall	p a
None	−0.023 (−0.208, 0.162)	0.087 (−0.107, 0.281)	0.039 (−0.092, 0.170)	0.56
Age	0.167 (−0.019, 0.353)	0.214 ( 0.021, 0.407)	0.183 ( 0.053, 0.313)	0.0059
Age, sex	0.183 (−0.005, 0.371)	0.242 ( 0.046, 0.439)	0.201 ( 0.069, 0.333)	0.0028
Age, sex, leukocyte distribution (LD)	0.250 ( 0.047, 0.453)	0.274 ( 0.068, 0.481)	0.255 ( 0.115, 0.396)	0.0004
Age, sex, LD, smoking	0.243 (0.038, 0.448)	0.282 (0.074, 0.491)	0.256 (0.113, 0.398)	0.0004
Age, sex, LD, alcohol	0.289 (0.080, 0.497)	0.280 (0.070, 0.490)	0.277 (0.133, 0.421)	0.0002
Age, sex, LD, history of cancer	0.234 (0.030, 0.437)	0.275 (0.067, 0.484)	0.250 (0.108, 0.391)	0.0006
Age, sex, LD, interaction (age accel. with sex)
Estimate in females	0.304 ( 0.036, 0.572)	0.246 (−0.026, 0.519)	0.269 (0.082, 0.455)	0.0048
Estimate in males	0.190 (−0.089, 0.469)	0.307 ( 0.013, 0.601)	0.241 (0.046, 0.436)	0.016

Shown is the estimated change (95 % confidence interval) in FI (expressed in %) per year of age acceleration. All models are adjusted for the methylation array batch using a random effect

^a p values refer to t distribution tests of the estimates obtained by multiple imputation in the overall model