Maxime Verhoeven

50 Chapter 3 The sensitivity analyses using non-imputed data over 24 months were generally in line with the results based on imputed data and showed a statistically significantly lower DAS28 over 24 months in favour of the TCZ arms: TCZ+MTX vs. MTX+Pred: -0.51 [95%CI -0.84 to -0.19]; TCZ vs. MTX+Pred: -0.42 [95%CI -0.75 to -0.09]. No statistically significant differences were found for the other outcomes, see Supplementary Table S3. DISCUSSION Treatment strategies using TCZ, with or without MTX, resulted in better DAS28-based outcomes over 24 months compared to a strategy using MTX with 10mg prednisone, suggesting better disease control, although this could, at least partly, be driven by the specific APR-lowering effect of TCZ since APR (ESR) is an component of DAS28. 18 Therefore, we analysed m-CDAI which does not include an APR. With this outcome, lower disease activity of TCZ-based strategies compared to MTX+Pred over 24 months could not be confirmed. Interestingly, over the first 3 months only, disease activity as measured with m-CDAI was statistically significantly lower in the MTX+Pred strategy. This raises the issue whether the specific APR-lowering effect of TCZ could clinically be relevant, e.g., with respect to cardiovascular diseases. 19,20,21 Despite the absolute differences found in DAS28-based outcomes, differences are likely not clinically relevant. TCZ-based treatment strategies as well as the combination of MTX+Pred in a treat-to-target treatment strategy result in a high proportion of patients being in remission (i.e., 80-60%) depending on remission definition in- or excluding APR. All point estimates were close to 1.00, indicating similar outcomes between treatment strategies. Besides, it is likely that the MTX+Pred treatment strategy is more cost-effective. Eventually, a new relevant treatment target which is becoming more feasible is DFR. The definition of DFR, however, still needs clarification as well as the optimal tapering strategy. In the U-Act-Early trial, DFR was defined as being free from DMARDs or placebo-DMARDs for at least 12 weeks while still in sustained remission (i.e., a DAS28 of less than 2.6 and a swollen joint count of four or fewer joints of the 28 joints assessed, during at least 24 weeks), 12 and a relevant proportion of patients achieved this. Unfortunately it was not possible to compare patients in DFR with the MTX+Pred arm due to different tapering protocols in the trials. For example, in U-Act-Early, MTX and TCZ were both tapered, whereas only MTX was tapered in CAMERA-II. 6,12 It will be relevant to further investigate DFR as it is an important (future) outcome that may positively affect quality of life and work ability, even on longer term. 22 In the current study there was no clear difference regarding effectiveness and safety between the strategy starting TCZ(+MTX) or MTX+Pred, which might be different for established RA patients with a suboptimal response to MTX. Although in U-Act-Early and the present study, 12 effectiveness of TCZ was somewhat less than that of TCZ+MTX, for