Serum Samples That Have Been Stored Long-Term (>10 Years) Can Be Used as a Suitable Data Source for Developing Cardiovascular Risk Prediction Models in Large Observational Rheumatoid Arthritis Cohorts

Table 2

Effect of storage time (years) on TC levels.

	Estimate	SE	-value	95% CI
				Lower	Upper
Constant	3.408	0.389	<0.0001	2.65	4.17
Storage time	−0.030	0.008	<0.0001	−0.045	−0.015
Time within patients; (follow-up)	0.012	0.003	<0.0001	0.001	0.018
Time^2 (follow-up)	−0.0001	0.00003	<0.0001	−0.0002	−0.0001
Gender	0.023	0.097	0.810	−0.167	−0.213
Age	0.016	0.004	<0.0001	0.087	0.023
Statin use at baseline	0.888	0.188	<0.0001	0.519	1.256
BMI	0.011	0.011	0.343	−0.011	0.033
Glucocorticosteroid use at baseline	0.239	0.115	0.038	0.013	0.465

(i) SE, standard error; BMI, body mass index; and 95% CI: 95% confidence interval.
(ii) Data are adjusted for age, gender, statin use at baseline, BMI, and glucocorticosteroid use at baseline.
(iii) Time² is a quadratic term that was included due to the nonlinear course of cholesterol levels over follow-up time, and this variable also represents time within patients (follow-up time).