Table 5

Linear regression analysis of sRAGE levels among 138 patients with established rheumatoid arthritis (RA)

Variable

β

SE

t

P value

95% Confidence interval

Univariate analysis:

RF

-0.23

0.09

-2.56

0.01

-0.41 to -0.05

Current smoker

-0.24

0.11

-2.14

0.03

-0.47 to -0.02

Serum creatinine^a

0.23

0.13

1.77

0.08

-0.03 to 0.49

History of vasculitis

-0.29

0.14

-2.07

0.04

-0.57 to -0.01

Family history CVD

-0.20

0.10

-2.03

0.04

-0.39 to -0.005

HDL^a

-0.30

0.17

-1.77

0.08

-0.64 to 0.04

CRP^a

-0.13

0.05

-2.59

0.01

-0.23 to -0.03

Diastolic blood pressure

-0.01

0.004

-2.73

0.007

-0.02 to -0.003

Carriage of RAGE 82S

-0.32

0.10

-3.06

0.003

-0.53 to -0.11

S100A8^a

-0.06

0.03

-2.51

0.01

-0.11 to -0.01

S100A9^a

-0.08

0.04

-1.94

0.06

-0.15 to 0.002

S100A12^a

-0.08

0.05

-1.70

0.09

-0.18 to 0.01

Multivariate analysis:

Carriage of RAGE 82S

-0.33

0.12

-2.88

0.005

-0.56 to -0.10

HDL^a

-0.35

0.16

-2.13

0.04

-0.68 to -0.02

CRP^a

-0.14

0.05

-2.63

0.01

-0.24 to -0.03

History of vasculitis

-0.41

0.15

-2.68

0.009

-0.71 to -0.10

Variables that were not listed in the simple linear regression model did not achieve P < 0.1. The selected covariates from simple linear regression (P < 0.1) of current smoker, RF, family history of CVD, diastolic blood pressure, creatinine, S100A8 and S100A12 levels were removed from the multiple linear regression model because these covariates did not independently correlate with the dependent variable. R² of the multiple linear regression model = 0.24. Adjusted R² = 0.1

^aLogarithmic base e transformation; sRAGE was logarithmically (base e) transformed.

CRP, C-reactive protein; CVD, cardiovascular disease; HDL, high-density lipoprotein; RAGE, receptor for advanced glycation end products; RF, rheumatoid factor.

Chen et al. Arthritis Research & Therapy 2009 11:R39   doi:10.1186/ar2645