Routine anticoagulation with rivaroxaban after TAVI harmful: GALILEO trial results

George Dangas

George Dangas (Mount Sinai Hospital, New York, USA) told delegates in a late-breaking trial session at the American Heart Association Scientific Sessions (AHA 2019; 16–18 November, Philadelphia, USA) that “in patients without an established indication for oral anticoagulation after successful transcatheter aortic valve implantation (TAVI), a treatment strategy including rivaroxaban at a dose of 10mg daily was associated with a higher risk of death or thromboembolic complications and a higher risk of bleeding than an antiplatelet-based strategy”.

The GALILEO trial had been terminated early in August 2018 by the data and safety monitoring board because of safety concerns. Dangas presented the main results of the trial on behalf of his fellow investigators, and the study was published simultaneously in the New England Journal of Medicine.

He outlined the reasons for carrying out the study: “Patients undergoing TAVI are typically elderly and frail, and at increased risk for both ischaemic and bleeding complications. Subclinical leaflet thrombosis may occur with bioprosthetic valves, and may be associated with an increased risk of cerebrovascular events that may be prevented or reversed by anticoagulation. While guidelines recommend the use of dual antiplatelet therapy early after TAVI, there is a dearth of evidence for routine use of anticoagulation after TAVI.”

Researchers undertook the open-label, international, multicentre, event-driven, randomised, controlled trial to compare a rivaroxaban-based antithrombotic strategy with an antiplatelet-based strategy post-successful TAVI. The primary efficacy endpoint was death, stroke, myocardial infarction (MI), systemic thromboembolism, symptomatic valve thrombosis, or deep venous thrombosis or pulmonary embolism. The primary safety endpoint was Valve Academic Research Consortium (VARC) 2 criteria of major, disabling or life-threatening bleeding.

In the trial, 1,644 adults with a mean age >80 years who had undergone successful TAVI for aortic valve stenosis were randomised to either the rivaroxaban group (n=826) and were given rivaroxaban 10mg and aspirin 75–100mg daily for 90 days, followed by rivaroxaban 10mg daily or to the antiplatelet group (n=818) and were given clopidogrel 75mg and aspirin 75–100mg daily for 90 days, followed by aspirin 75–100mg daily.Tthe 10mg daily dose that was used is lower than the stroke prevention dosage in atrial fibrillation.

Dangas said: “The primary hypothesis of the trial was that the rivaroxaban-based strategy would be superior to the antiplatelet-based strategy with respect to incidence of death or thromboembolic events. We estimated that 440 composite primary outcome events would provide the trial with 80% power to detect a 20% lower relative risk in the rivaroxaban group than in the antiplatelet group. However, because the trial was terminated early, only 183 patients had reached the primary efficacy outcome [42% of the planned 440].”

In the intention-to-treat analysis, the primary efficacy outcome of death or first thromboembolic event was higher in the rivaroxaban arm (n=105) than in the antiplatelet arm (n=78) (9.8 vs. 7.1 events per 100 person–years), with a hazard ratio (HR) of 1.35 (95% confidence interval [CI]1.01–1.81, p=0.04). All-cause mortality rates in the intention-to-treat analysis were also significantly increased, with 64 deaths in the rivaroxaban arm vs. 38 in the antiplatelet arm (5.8 vs. 3.4 per 100 person–years), with a HR of 1.69 (95% CI 1.13–2.53, p=0.009).

The primary safety endpoint in the intention-to-treat analysis of life-threatening, disabling or major bleeding occurred in 46 patients in the rivaroxaban arm vs. 31 in the antiplatelet arm (4.3 vs. 2.8 per 100 person–years, HR 1.5, 95% CI 0.95–2.37, p=0.08).

In the on-treatment analysis, there were 26 deaths in the rivaroxaban arm and 24 in the antiplatelet arm (HR 1.23, 95% CI 0.71–2.15).

Dangas outlined the limitations of the trial, which included that it used open-label treatment which is potentially subject to reporting and ascertainment bias, it was prematurely terminated so treatment effects and confidence intervals need to be interpreted with caution, and on-treatment analyses are subject to bias due to the high rates of treatment discontinuation. In addition, patients undergoing TAVI with an established indication for anticoagulation were not included in the trial.

A substudy of GALILEO that used 4D computed tomography three months after randomisation found subclinical hypoattenuated leaflet thickening and reduced leaflet motion of bioprosthetic aortic valves were less frequent in patients treated with rivaroxaban than in those in the antiplatelet group. However, there were too few clinical events to permit any meaningful interpretation of these results. The findings were also presented at AHA 2019.

“The mechanism underlying the higher mortality in the rivaroxaban arm observed in the intention-to-treat analysis in this trial is unclear,” noted Dangas. He concluded: “The mortality rate differences were attenuated in the on-treatment analysis and many occurred late after discontinuation of the study drug. These results of the GALILEO main trial are irrespective of the potential effects on valve imaging findings from the GALILEO 4D-CT Ancillary Study.”

Dangas emphasised to Cardiovascular News that GALILEO included patients without any established indication for oral anticoagulation. “Therefore,” he clarified, “it’s results affect patients after TAVI who do not require an anticoagulant for any clinical reason. Other ongoing clinical trials investigate different oral anticoagulants in patients in atrial fibrillation and TAVI.”


Please enter your comment!
Please enter your name here