Background: The VEDERA (Very Early vs. Delayed Etanercept in Rheumatoid Arthritis) randomised controlled trial compared the effect of conventional synthetic disease-modifying anti-rheumatic drug (csDMARD) therapy with biologic DMARD (bDMARD) therapy using the tumour necrosis factor inhibitor etanercept in treatment-naive, early rheumatoid arthritis patients. The CADERA (Coronary Artery Disease Evaluation in Rheumatoid Arthritis) trial was a bolt-on study in which VEDERA patients underwent cardiovascular magnetic resonance imaging to detect preclinical cardiovascular disease at baseline and following treatment. Objectives: To evaluate whether or not patients with treatment-naive early rheumatoid arthritis have evidence of cardiovascular disease compared with matched control subjects; whether or not this is modifiable with DMARD therapy; and whether or not bDMARDs confer advantages over csDMARDs. Design: The VEDERA patients underwent cardiovascular magnetic resonance imaging at baseline and at 1 and 2 years after treatment. Setting: The setting was a tertiary centre rheumatology outpatient clinic and specialist cardiovascular magnetic resonance imaging unit. Participants: Eighty-one patients completed all assessments at baseline, 71 completed all assessments at 1 year and 56 completed all assessments at 2 years. Patients had no history of cardiovascular disease, had had rheumatoid arthritis symptoms for ≤ 1 year, were DMARD treatment-naive and had a minimum Disease Activity Score-28 of 3.2. Thirty control subjects without cardiovascular disease were approximately individually matched by age and sex to the first 30 CADERA patients. Patients with a Disease Activity Score-28 of ≥ 2.6 at 48 weeks were considered non-responders. Interventions: In the VEDERA trial patients were randomised to group 1, immediate etanercept and methotrexate, or group 2, methotrexate ± additional csDMARD therapy in a treat-to-target approach, with a switch to delayed etanercept and methotrexate in the event of failure to achieve clinical remission at 6 months. Main outcome measures: The primary outcome measure was difference in baseline aortic distensibility between control subjects and the early rheumatoid arthritis group and the baseline to year 1 change in aortic distensibility in the early rheumatoid arthritis group. Secondary outcome measures were myocardial perfusion reserve, left ventricular strain and twist, left ventricular ejection fraction and left ventricular mass. Results: Baseline aortic distensibility [geometric mean (95% confidence interval)] was significantly reduced in patients (n = 81) compared with control subjects (n = 30) [3.0 × 10–3/mmHg (2.7 × 10–3/mmHg to 3.3 × 10–3/mmHg) vs. 4.4 × 10–3/mmHg (3.7 × 10–3/mmHg to 5.2 × 10–3/mmHg), respectively; p