Results of the Cox regression analyses to assess the association between BMI category and drug survival per TNFi
| Adjusted HR (95% CI) | P value | Crude HR (95% CI) | P value | Time on treatment (months) median (IQR) | |
|---|---|---|---|---|---|
| All TNFi | 34.8 (12.6; 72.1) | ||||
| BMI category | |||||
| Underweight | 1.3 (1.07; 1.58) | 0.011 | 1.2 (0.97; 1.44) | 0.091 | 32.6 (9.4; 59.4) |
| Normal weight | Ref | Ref | 34.9 (12.4; 72.3) | ||
| Pre-obesity | 0.97 (0.89; 1.05) | 0.451 | 1.0 (0.95; 1.12) | 0.471 | 34.9 (12.9; 71.7) |
| Obesity class I | 1.09 (0.97; 1.21) | 0.145 | 1.2 (1.06; 1.31) | 0.003 | 34.7 (12.3; 73.6) |
| Obesity class II | 1.28 (1.06; 1.54) | 0.012 | 1.4 (1.16; 1.68) | <0.001 | 29.7 (10.3; 50.8) |
| Obesity class III | 1.67 (1.29; 2.18) | <0.001 | 1.9 (1.44; 2.42) | <0.001 | 21.1 (9.7; 43.0) |
| Adalimumab | 26.8 (10.3; 60.6) | ||||
| BMI category | |||||
| Underweight | 1.2 (0.84; 1.58) | 0.366 | 1.01 (0.79; 1.46) | 0.640 | 20.5 (5.9; 58.2) |
| Normal weight | Ref | Ref | 24.0 (9.7; 56.0) | ||
| Pre-obesity | 0.86 (0.75; 0.99) | 0.036 | 0.92 (0.81; 1.05) | 0.233 | 32.5 (12.2; 67.9) |
| Obesity class I | 0.91 (0.75; 1.10) | 0.309 | 1.02 (0.85; 1.21) | 0.865 | 27.1 (9.0; 64.4) |
| Obesity class II | 1.21 (0.88; 1.67) | 0.234 | 1.33 (0.97; 1.82) | 0.076 | 26.2 (11.3; 47.0) |
| Obesity class III | 1.31 (0.87; 1.97) | 0.192 | 1.52 (1.03; 2.23) | 0.034 | 16.5 (6.2; 62.0) |
| Certolizumab | Number of patients per group too small | 10.5 (3.7; 18.3) | |||
| Etanercept | 29.5 (11.2; 60.0) | ||||
| BMI category | |||||
| Underweight | 1.05 (0.80; 1.38) | 0.734 | 0.99 (0.76; 1.29) | 0.919 | 38.3 (12.3; 59.5) |
| Normal weight | Ref | Ref | 31.9 (12.0; 64.4) | ||
| Pre-obesity | 1.01 (0.90; 1.14) | 0.820 | 1.07 (0.96; 1.19) | 0.254 | 30.5 (11.3; 60.5) |
| Obesity class I | 1.13 (0.97; 1.32) | 0.108 | 1.25 (1.09; 1.45) | 0.002 | 30.2 (11.5; 62.7) |
| Obesity class II | 1.27 (0.98; 1.65) | 0.069 | 1.33 (1.04; 1.70) | 0.024 | 22.4 (5.5; 47.3) |
| Obesity class III | 1.79 (1.25; 2.55) | 0.001 | 1.93 (1.36; 2.74) | <0.001 | 20.2 (4.0; 49.1) |
| Golimumab | Number of patients per group too small | 12.0 (4.7; 28.0) | |||
| Infliximab | 34.9 (12.5; 77.8) | ||||
| BMI category | |||||
| Underweight | 1.82 (1.20; 2.76) | 0.005 | 1.66 (1.11; 2.48) | 0.014 | 25.8 (6.2; 44.5) |
| Normal weight | Ref | Ref | 37.0 (11.1; 78.4) | ||
| Pre-obesity | 0.96 (0.80; 1.16) | 0.685 | 1.02 (0.86; 1.21) | 0.840 | 32.3 (13.7; 84.1) |
| Obesity class I | 1.21 (0.96; 1.53) | 0.115 | 1.29 (1.03; 1.61) | 0.027 | 27.8 (10.5; 69.3) |
| Obesity class II | 1.49 (0.98; 2.26) | 0.064 | 1.71 (1.14; 2.55) | 0.009 | 19.3 (11.4; 37.8) |
| Obesity class III | 1.46 (0.79; 2.71) | 0.225 | 1.81 (1.01; 3.23) | 0.045 | 27.7 (13.1; 34.0) |
Cox regression analyses results based on multiply imputed data. Adjusted analyses were corrected for gender, smoking, age, baseline DAS28 and GC use. Before imputation, the proportion of missing data was 19% for BMI, 0.2% for gender, 0.3% for age, 49% for smoking and 18% for baseline DAS28. Increased HR indicates a higher hazard to stop medication. Time on treatment is the crude time on treatment, with follow-up time censored at 5000 days.
BMI, body mass index; DAS, disease activity score; GC, glucocorticoid; TNFi, tumour necrosis factor inhibitor.