fetching data ... 
Background: Modern trials in RA often have a strict protocolized treat-to-target design, whereas in daily practice shared decision making is the preferred way to manage treatment.
Objectives: To document the protocol violations in an early RA trial that illustrate potential conflicts between these strategies.
Methods: In the COBRA treat-to-target trial, treatment-naive early RA patients were classified into a low- and high-risk group, and treated according to two different protocols. Treatment target was minimal disease activity (DAS44 <1.6) or EULAR good response. Non-responders at 13 weeks were randomized to treatment intensification or continuation, with as primary endpoint the proportion of patients achieving the treatment target at 26 weeks.
Results: In total 190 patients were included. High-risk patients (n=150) were initially treated with COBRA-light (30 mg/day oral prednisolone, tapered in 8 weeks to 7.5 mg/day, and MTX increasing to 25 mg/week): 110 (73%) patients reached the treatment target at 13 weeks. The 31 non-responders were randomized to treatment intensification (n=15) or treatment continuation (n=16) for weeks 14 to 26. Intensification comprised 60 mg/day prednisolone, tapered in 6 weeks to 7.5 mg/day, and addition of sulfasalazine 2000 mg/day and hydroxychloroquine 400 mg/day (COBRA-plus).
The low-risk group (n=40) was treated with MTX monotherapy increasing to 25 mg/week: only 20 (50%) reached the target at week 13. The 15 non-responders were randomized to treatment intensification with a COBRA-light oral pulse of prednisolone (n=8) or continuation (n=7).
Over the whole study period 79 of 190 patients (42%) had a protocol violation. In the first 13 weeks, 27 (18%) high-risk and 8 (20%) low-risk patients had a protocol violation. In the responders at week 13, 15 (14%) high-risk and 3 (18%) low-risk patients had a protocol violation in week 14-26. In the high-risk non-responders, 18 of 31 had a protocol violation (58%) after randomization at 13 weeks, 11 in the intensification group and 7 in the continuation group (
Protocol treatment violations after randomization of non-responders at week 13 by risk group.
| High-risk | Low-risk | |||
|---|---|---|---|---|
| Continuation (16) | Intensification (15) | Continuation(7) | Intensification (8) | |
| Week 14-26 | ||||
| Number of violations | 7 | 11 | 4 | 4 |
| Increased intensity | ||||
| Rheumatologist | 3 | |||
| Other physician | 1 | |||
| Decreased intensity | ||||
| AEs MTX | 1 | 1 | ||
| No MTX due to AEs | 1 | |||
| Rheumatologist | 1 | 3 | 1 | |
| Rheumatologist, patient | 1 | 7 | ||
| Patient | 1 | 2 | ||
| Mixed (increase + decrease) | ||||
| AEs MTX, rheumatologist | 1 | |||
| Rheumatologist, patient | 1 | 1 | ||
Violations are categorized by direction of intensification and by reason/initiator.
AE=adverse event; MTX=methotrexate.
Conclusion: In this study almost half of the patients had a protocol violation. The high frequency is most likely related to the open design of the trial, the risk of adverse effects through intensification with high dose prednisolone and other DMARDs, the randomization phase, and, last but not least, by the shared decision making. Shared decision making by physician and patient may jeopardizes protocol adherence in trials with an open treat-to-target design.
Acknowledgements: We have received an unrestricted grant from Pfizer.
Disclosure of Interests: Linda Hartman: None declared, Linda A. Rasch: None declared, S.A. Turk: None declared, Marieke ter Wee: None declared, Pit J. S. M. Kerstens: None declared, Conny J. van der Laken: None declared, Michael T Nurmohamed: None declared, Dirkjan van Schaardenburg: None declared, Lilian H. D. van Tuyl: None declared, Alexandre Voskuyl: None declared, Maarten Boers Consultant of: Novartis, WIllem Lems Speakers bureau: Pfizer, Galapagos, Lilly, Amgen, UCB, Consultant of: Pfizer, Galapagos, Lilly, Amgen, UCB