Table 2

Factors associated with drug levels and anticertolizumab antibodies longitudinally over 12 months using GEE

VariableCertolizumab drug level
β coefficient (95% CI)
p ValueADAb level
β coefficient (95% CI)
p Value
Univariate analysis
Age0.14 (−0.017 to 0.29)0.080.021 (−2.16 to 2.21)0.99
Female gender*4.76 (0.21 to 9.29)0.040−67.36 (−146.14 to 11.41)0.094
BMI−0.46 (−0.89 to −0.041)*0.032−0.92 (−6.03 to 4.19)0.72
CRP−0.099 (−0.17 to −0.029)*0.0050.57 (−0.32 to 1.47)0.21
ESR−0.12 (−0.22 to −0.020)*0.0190.40 (−0.95 to 1.74)0.58
Baseline methotrexate use−0.11 (−4.68 to 4.47)0.96−3.65 (−53.38 to 46.08)0.89
Baseline methotrexate dose0.23 (0.33 to 0.78)0.434.94 (3.74 to −13.6)0.27
Baseline oral steroid use3.15 (−1.89 to 8.20)0.22−29.77 (−69.54 to 10.0)0.14
Any nbDMARD use at baseline†3.15 (−1.89 to 8.20)0.22−54.66 (−125.37 to 16.05)0.14
Antidrug antibody level−0.037 (−0.055 to −0.018)*<0.0001
Certolizumab drug level−2.56 (−4.09 to −1.03)*0.001
Adherence10.43 (4.76 to 16.11)*<0.0001−45.05 (−108.35 to 18.25)0.16
Multivariate model
Antidrug antibody level−0.044 (−0.059 to −0.028)*<0.0001
Adherence7.08 (0.71 to 13.45)*0.029
Female gender1.77 (−4.21 to 7.76)0.56
BMI−0.13 (−0.68 to 0.43)0.65
CRP−0.065 (−0.14 to 0.013)0.102
  • *p<0.05.

  • †nbDMARD use included methotrexate, sulfasalazine, leflunomide or hydroxychloroquine at baseline.

  • ‡Adjusted for variables significant in the univariate analysis. CRP was used instead of ESR in the multivariate model, as was measured in all patients. In the second column, certolizumab drug level is used as the dependant variable in the GEE model; in the fourth column, anticertolizumab antibodies are used as the dependant variable in the GEE model.

  • ADAb, antidrug antibody; BMI, body mass index; ESR, erythrocyte sedimentation rate; GEE, generalised estimating equation; nbDMARD, non-biologic disease-modifying antirheumatic drug.