Atrial fibrillation is a heart rhythm disorder that often occurs after heart surgery. During atrial fibrillation blood cloths may form, predominantly in the left atrial appendage, a small sac in the wall of the left side of the heart. Some heart surgeons close this appendage to protect against stroke, particularly in patients with a history of atrial fibrillation, yet there is little evidence to support the efficacy and safety of this practice.

We therefore conducted the Left Atrial Appendage Closure by Surgery (LAACS) study (2010-2016) were patients in whom the appendage was closed (by chance) suffered fewer brain damages that patients where it remained open. Although encouraging, these results were not only based on strokes, but also on scars without symptoms found in brain scans. The following LAACS-2 study will include a sufficient number of patients to determine whether future guidelines should advise to close systematically the left atrium appendage during a heart operation.

Aim:

The overall aim of the project is to provide lifelong protection against brain damage related to atrial fibrillation, to patients who undergo open heart surgery.

Background:

Formation of thrombi in the left atrium appendage (LAA) during atrial Fibrillation (AF) can provoke strokes, (shortened - references 1-14) The investigators hypothesis is that systematic closure of the LAA, in addition to planned open-heart surgery, can reduce cerebral embolisms derived from thrombus formation in the left atrium. The investigators performed the LAACS study (clinicaltrial.gov NCT02378116) to test this hypothesis, which is the first randomized study that shows a protective effect of closure of the LAA during surgery, against ischemic damage of the brain, a question that has eagerly been sought in the scientific community. 1-3 The results of the LAACS trial demonstrate a benefit including clinical significant stroke and silent infarctions, 15 with 5 (5%) primary events in the LAACS-group and 14 (16%) in the control group (Hazard ratio 0.3; 95% CI: 0.1-0.8, p=0.02). Nevertheless, a limited number of patients were randomized in LAACS (N 187) of whom 141 followed the procedure. Expectedly, differences on stroke were not significant. Hence, although encouraging, results should be proven in a larger trial, before a systematic closure of the LAA can be recommended as a routine procedure during open-heart surgery. The main purpose of LAACS2 is to provide evidence to support or not such procedure.

Methods:

The target population is that of patients older than 18 years, that await their first-time open-heart surgery. The investigators design the project as an open, parallel study where patients will be randomized to closure of the LAA, in addition to planned open heart surgery. The LAA will remain open in the control group.

Inclusion and randomization:

Patients are identified among referrals for thoracic surgery in the respective departments and will receive written information and contacted by telephone prior to their admission. Both, patients with and without previous diagnosis of AF will be invited to participate. For inclusion and exclusion criteria (see 'eligibility').

Since use of anti-coagulants provides protection against thrombus formation, the patients will be stratified by ongoing use of anti-coagulant medications, which includes patients with planned mechanic valve replacement, in blocks of 16 patients randomized 1:1 in each strata.

Beyond closure of the LAA or control, patients will be managed according to general practice, without restrictions on the preferred use of medications for rhythm control, or conversion procedures.

Method of LAA closure:

The LAACS procedure is routinely conducted on patients undergoing per-operative ablation for AF during CABG and valve operations. Nevertheless, insufficient closure of the LAA is of major concern, as an incomplete closure may even be harming3. Accordingly, experimental 16,17 and clinical studies 18,19 support the efficacy of closure of the LAA with a clip device rather than a single suture 2,19,20. Pre-clinic studies using the AtriClip device have complete occlusion, at least, for the first 90 days following the procedure 18,21. Therefore, the standard procedure for LAA closure in the LAACS 2 trial is using AtriClip devices (AtriCure Inc., West Chester, Ohio, USA).

Tissue, blood and urine samples:

During operation, a biopsy from the right atrial appendage (approximately 3 g) is taken from all patients, and from the LAA (approximately 1 g) if patients are ungoing LAACS. Furthermore, approximately 15 mL of venous blood and 30 mL of urine is drawn during surgery.

All samples are immediately frozen in liquid nitrogen and transported on dry ice to a -80 °C freezer. Samples will be analyzed when the last sample has been drawn.

Follow-up:

The participants will be followed for at least two years, until the end of the study. Follow-up is based on yearly contact with the patients' hospital records and telephone contact, if medical records are not available.

The first 10 participants in each site will be invited to participate in an imaging study, one year after the operation, to assess the quality of the closure. Imaging will be esophagus echocardiography or a computed tomography scan, according to the imaging facilities available in the site.

All patients with AF onset after the operation, before discharge, where there has not been recorded recurrence of AF by the end of the first year of follow-up will be invited to participate in long-term continuous ambulatory heart-rhythm monitoring (Holter for at least seven days).

Patients may decline to participate in the additional imaging and Holter monitoring and remain in the study for follow-up.

Additionally, sites are requested to perform equivalent imaging of the heart on patients that have undergone closure of the left atrial appendage and experienced strokes, to evaluate the completeness of the LAA closure.

Definition of AF and recurrence of AF:

AF onset in the days following the operation, until discharge, detected in clinical settings are considered as "per-operative AF". Recurrence of AF detected during the first three months following the operation are considered "early recurrence" and recurrence beyond three months after the operation detected in clinical settings are considered "late recurrence". In patients undergoing Holter monitoring in the study, the occurrence of AF for 30 seconds or longer will be considered diagnostic for AF recurrence.

Power calculation for primary endpoint stroke and minor stroke (transitory cerebral ischemia):

Stroke has an incidence of 1-5% in the years following heart surgery 7,22-24. Chang et al. pooled data from studies comparing outcomes after PCI and CABG (from the SYNTAX, BEST and PRECOMBAT trials), which provides an estimate of 3,7% incidence of stroke in the first three years following coronary by-pass operations 25. In the LAACS study, incidence of stroke was 3.2% (3 of n 64) in patients who underwent surgical closure of the LAA, while in control patients in whom LAA remained open, stroke incidence was 11.3% (8 in n 77) after 3,6 median follow-up. A non-randomized study reported a 33% risk reduction of thromboembolic events.

To achieve a corresponding reduction in the LAACS trial, with a significance level of 0.05 and 90% power, over three years, it would be necessary to include 600 patients in each group (calculated manually and using web calculators www.biomath.info/power/chsq.htm, www.dssresearch.com, and www.optimizely.com ).

We deemed a relative risk reduction of 60% with LAA closure plausible. In a balanced design (1:1), with a Chi-square approximation (comparing two independent binomial proportions) with a two-sided significance level of 0.05 and power of 90%, and two years of observation time, the sample size needed is 1302. We include a 'drop-out' rate of 10%, as unexpected intraoperative findings may prompt surgeons to deviate from the allocated treatment, and the primary outcome is subject to death as a competing risk. In total, 1,500 patients should be randomized.

After two years of follow-up, 39 (651*6.0%) and 16 (651*2.4%) primary events are expected in the control and LAA closure groups, respectively. The trial may be extended if 55 primary events have not occurred after two years of follow-up.

The steering committee will decide whether the trial should continue enrolment if more than 55 events have occurred before 1500 patients have been included. Likewise, the committee will decide whether to increase the number of patients included if less than 55 primary events have occurred after the last randomized patient has been followed for two years.

Power calculation for harm of the LAACS procedure:

Incomplete LAA closure may render a more thrombogenic surface than the innate LAA surface.2,19,20 The relative risk for stroke related to incomplete LAA closure has been estimated between 10 to 25% compared to those with complete LAA closure.13,35 Accordingly, the investigators can evaluate whether an equivalent problem occurs in the LAACS2, including between 359 to 1455 patients in the LAACS2 study, with a 90% power and a 0.05 significance level. Hence, the 2000 patients the investigators plan to include, as calculated for the primary endpoint should suffice to evaluate the possible harm of the procedure.

Prespecified substudies The wavECG substudy utilizes a preoperative Myovista electrocardiogram (HeartScienc-es, Southlake, Texas, USA) to diagnose diastolic dysfunction in consecutive patients and will evaluate if these indices can improve current prediction models of (1) postoperative AF and (2) postoperative stroke including TIA. An anticipated 300 patients will be en-rolled.

The biomarker substudy will evaluate blood, urine, and atrial tissue samples in a sub-group of patients to explore systemic signs of atrial cardiomyopathy associated with the development of (1) postoperative AF and (2) postoperative stroke, including TIA. An anticipated 75 patients will be enrolled.

The imaging substudies will evaluate the quality of LAA closures in a subgroup of ten consecutive patients from each site utilizing echocardiography, CT, or MRI according to the availability of imaging facilities. Moreover, imaging control of LAA closures is recommended in patients with LAA closure developing ischemic stroke, including TIA. An antic-ipated 50 patients will be enrolled.

The Holter substudy will be conducted in patients with new-onset AF from surgery until discharge and where another episode of AF has not been recorded approaching the end of follow-up. Patients from selected sites will be invited to participate in long-term continuous ambulatory heart-rhythm monitoring to assess the burden of subclinical recurrent AF. An anticipated 100 patients will be enrolled.

Inclusion Criteria:

• Patients referred for planned first-time heart surgery:
• Coronary artery by-pass surgery (CABG)
• Valve surgery
• Combined CABG and valve surgery

Exclusion Criteria:

• Endocarditis
• No possible follow-up
• Planned closure of the left atrium appendage as part of surgery

Study data will be shared upon reasonable request after study end.

From study end until data storage permissions from the Danish Data Protection Agency expire.

Reasonable request