Univariate and multivariate logistic regression of combined biomarker score and some baseline predictors of low DAS28 at 3 months.
| Variables | Univariate | Multivariate | ||
|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | |
| RF-positivea | 1.33 | 0.79–2.24 | 1.35 | 0.71–2.57 |
| Anti-CCP–positive | 1.31 | 0.79–2.18 | 1.65 | 0.87–3.14 |
| Sex, male | 2.82 | 1.67–4.76 | 1.76 | 0.88–3.50 |
| Age (per yr increase) | 1.03 | 1.01–1.05 | 1.03 | 1.01–1.05 |
| Current smoking (yes) | 0.92 | 0.50–1.69 | 0.78 | 0.38–1.64 |
| HAQ (per unit increase) | 0.36 | 0.23–0.57 | 0.38 | 0.22–0.65 |
| Combined biomarker score (per unit increase)b | 0.40 | 0.29–0.55 | 0.43 | 0.29–0.65 |
| Results after stepwise logistic regression analysis | ||||
| Sex, male | 1.64 | 0.90–2.98 | ||
| Age (per yr increase) | 1.03 | 1.01–1.05 | ||
| HAQ (per unit increase) | 0.35 | 0.22–0.58 | ||
| Combined biomarker score (per unit increase) | 0.44 | 0.31–0.63 | ||
Values in bold face are statistically significant.
↵a RF was excluded from multivariate model because of its collinearity with anti-CCP.
↵b OR for the combined biomarker score in univariate analysis indicates that for each unit increase, the odds of achievement of LDA decrease by 60%. Similarly, in multivariate model, after considering for other predictors (anti-CCP, sex, age and smoking), the odds of LDA decrease by 56% for each unit increase in the combined biomarker score. RF: rheumatoid factor; anti-CCP: anticyclic citrullinated peptide; HAQ: Health Assessment Questionnaire; DAS28: 28-joint count Disease Activity Score; LDA: low disease activity.