Data source
From July 2017 to July 2019, data from a single institution were collected in an open, double arm, retroprospective manner. The final database was constituted by pre-operative demographic variable, intra-operative interventional data and post-operative measurements.
Only the patients who underwent cardiac surgery and discharged with oral anticoagulant were rolled in the study; patients with preexisting bleeding disorders were excluded. Pre-operative use of any oral anticoagulant did not constitute an exclusion criteria. All major, elective, cardiac surgical procedures were considered; patients who underwent urgent or emergency operation were excluded from the study.
The VKA included exclusively warfarin. The NOACs used were dabigatran, rivaroxaban and apixaban.
Reasons for antithrombotic treatment at discharge were atrial fibrillation, deep venous thrombosis, atrial myxoma and left ventricular thrombus.
A bleeding event was defined as evidence of pericardial effusion, with or without need of surgical intervention. The incidence of pleural effusion, cerebral and abdominal hemorrhagic events were recorded but not reported as not primary outcomes of the present study.
All the ischemic or thromboembolic events were collected and reported. As ischemic event it was considered the incidence of cerebrovascular events, bowel ischemia, lower limb ischemia or myocardial infarction. A thromboembolic event was defined as incidence of deep venous thrombosis, pulmonary embolism, cerebral infarct or systemic venous thrombosis.
The primary end point was the incidence of immediate postoperative bleeding, defined as pericardial effusion requiring intra-hospital re-exploration. This term was used to define any pericardial drainage performed before discharge; this was then differentiated as early, if within the first 72 hours from surgery, or late, if after 72 hours.
The key secondary composite end point was the late re-admission for pericardial effusion, with or without need of surgical drainage.