Summary of baseline covariates and their relation to prescription of tocilizumab (TCZ) as monotherapy
| Prescription of TCZ as monotherapy | ||||
|---|---|---|---|---|
| Covariate | Summary (n=2057) | Proportion of mono (%) | OR (95% CI) for mono (n=1359) | |
| Age (years) | 55 (13.1) 56 (46–64) | 1.38 (1.11 to 1.71) per 20 years more | ||
| Sex (%) (n=2056) | Male | 21 | 24 | Reference |
| Female | 79 | 29 | 1.30 (0.92 to 1.83) | |
| Disease duration (years) (n=1900) | 11.4 (9.5) 9.1 (4.1–16.1) | 1.11 (0.96 to 1.27) per 10 years more | ||
| Seropositivity (%) (n=1891) | No | 18 | 27 | Reference |
| Yes | 82 | 28 | 1.08 (0.75 to 1.54) | |
| Number of prior biologics (%) (n=2056) | 0 | 19 | 29 | Reference |
| 1 | 26 | 26 | 0.82 (0.54 to 1.25) | |
| ≥2 | 55 | 28 | 0.76 (0.51to 1.13) | |
| Corticosteroids (%) (n=2011) | No | 51 | 33 | Reference |
| Yes | 49 | 22 | 0.74 (0.56 to 0.99) | |
| DAS28 (n=1914) | 5.0 (1.4) 5.1 (4.1–6.0) | 1.14 (0.89 to 1.45) per 2 units more | ||
| HAQ (n=1673) | 1.4 (0.7) 1.5 (1.0–2.0) | 0.95 (0.77 to 1.19) per 1 unit more | ||
| Year of TCZ initiation (%) | 2009 | 15 | 19 | Reference |
| 2010 | 23 | 22 | 1.23 (0.78 to 1.96) | |
| 2011 | 20 | 24 | 1.40 (0.87 to 2.25) | |
| 2012 | 22 | 35 | 2.07 (1.30 to 3.28) | |
| 2013 | 20 | 38 | 2.58 (1.62 to 4.10) | |
| Country (%) | Czech Republic | 12.9 | 23 | 0.66 (0.39 to 1.12) |
| Denmark | 35.5 | 31 | 0.92 (0.60 to 1.39) | |
| Finland | 2.3 | 11 | 0.36 (0.10 to 1.27) | |
| The Netherlands | 2.4 | 28 | – | |
| Norway | 3.8 | 52 | – | |
| Portugal | 8.4 | 15 | 0.46 (0.24 to 0.87) | |
| Russia | 4.1 | 11 | 0.30 (0.12 to 0.79) | |
| Slovenia | 9.5 | 21 | 0.28 (0.14to 0.55) | |
| Sweden | 6.4 | 38 | 1.41 (0.82 to 2.44) | |
| Switzerland | 14.7 | 34 | Reference | |
Sample sizes (n) equal the number of eligible patients presented in the header of column ‘Summary’ unless indicated otherwise. The column named ‘Summary’ provides a description of covariates in terms of mean (SD) and median (IQR) for discrete or continuous covariates and percentages for categorical covariates. The column named ‘Proportion of mono’ provides the frequency of monotherapy-initiated TCZ treatment for each category of a categorical covariate. The last column presents estimated ORs and 95% Wald CIs for prescribing TCZ as monotherapy (as compared to any type of combination therapy) based on multiple logistic regression. ORs for discrete or continuous covariates are presented for a difference corresponding approximately to the IQR. For categorical covariates, ORs with respect to the chosen reference category are shown. p Values from likelihood ratio tests for categorical covariates with more than two categories were 0.41 for number of prior biologics, <0.0001 for year of TCZ initiation and <0.0001 for country. The multiple logistic regression is based on all TCs with complete covariate information. The Netherlands (patchy data) and Norway (no HAQ recorded) lack TCs with complete covariate information.
DAS, disease activity score; HAQ, Health Assessment Questionnaire; TC, treatment courses.