Table 2.
Factors influencing serum levels of HMGB1 in univariable and multiple linear regression.
| Predictor variables∗ | Univariable analysis | Multivariable analysis# | ||||
|---|---|---|---|---|---|---|
| b | SE | P value | b | SE | P value | |
| Age | −0.001 | 0.002 | 0.66 | |||
| Gender (female/male) | 0.05 | 0.03 | 0.07 | |||
| Smoking (nonsmoker/smoker) | −0.07 | 0.04 | 0.08 | |||
| Drinking (nondrinker/drinker) | −0.05 | 0.05 | 0.35 | |||
| PP group versus control group | 0.20 | 0.04 | 0.0001 | 0.20 | 0.04 | <0.0001 |
| AE < 10 y group versus control group | 0.23 | 0.04 | <0.0001 | 0.23 | 0.04 | <0.0001 |
| AE ≥ 10 y group versus control group | 0.22 | 0.04 | <0.0001 | 0.22 | 0.04 | <0.0001 |
| Asbestosis group versus control group | 0.44 | 0.06 | <0.0001 | 0.44 | 0.06 | <0.0001 |
| MPM group versus control group | 0.48 | 0.07 | <0.0001 | 0.48 | 0.07 | <0.0001 |
∗The dependent variable was natural logarithm form; #estimated with forward stepwise method; b: regression coefficient; SE: standard error.