Table 3.
Difference between etoricoxib and comparators in change in biomarker concentrations from randomization at month 3
| Relative difference (Etoricoxib minus Comparator) [% (CI)†] | Upper bound of CI meets non-inferiority criterion?‡ | P-Value Superiority testing of etoricoxib vs. comparator§,¶ | |
|---|---|---|---|
| Comparator | |||
| Placebo | |||
| CRP | −7.8 (−30.5, 22.4) | Yes | 0.52 |
| LDL-C | −4.0 (−10.6, 3.2) | Yes | 0.20 |
| Homocysteine | −3.9 (−11.6, 4.6) | Yes | 0.30 |
| Fibrinogen | −3.7 (−9.4, 2.3) | Yes | 0.16 |
| Celecoxib | |||
| CRP | 23.1 (−6.8, 62.7) | Yes | — |
| LDL-C | −1.4 (−7.4, 4.9) | Yes | — |
| Homocysteine | −0.4 (−8.9, 9.0) | Yes | — |
| Fibrinogen | 2.5 (−3.6, 9.0) | Yes | — |
| Ibuprofen | |||
| CRP | 14.0 (−10.1, 44.6) | Yes | — |
| LDL-C | −2.9 (−8.0, 2.6) | Yes | — |
| Homocysteine | −3.0 (−10.5, 5.2) | Yes | — |
| Fibrinogen | 1.0 (−3.9, 6.3) | Yes | — |
CI was 97.5% for etoricoxib vs. placebo and etoricoxib vs. celecoxib, and 95% for etoricoxib vs. ibuprofen.
Test whether upper bound of CI for relative difference is less than non-inferiority margin (i.e., <80% for CRP and <20% for LCL-C, homocysteine, and fibrinogen).
Applicable only for the etoricoxib-vs-placebo comparison and only if etoricoxib was shown to be non-inferior to placebo.
Longitudinal analysis with repeated measures in a per-protocol population. Mixed model included factors for treatment, visit, treatment-by-visit interaction, gender, current tobacco use, diabetes, hypertension, age category (> or ≤median), and Framingham risk score as fixed effects using an unstructured covariance matrix.