The study protocol was approved by the Local Ethics Committee for the handling and analysis of retrospective data (ID number: 2018 − 01227, July 20th, 2018). No informed consent was required because the study was untertaken using information consecutively collected in the course of routine care.
Between September 2007 and March 2017, 278 aortic root replacements with the Freestyle® prosthesis were performed by four surgeons at a single centre. Patients were divided in three categories due to the pathology of the aortic root (re-operation vs endocarditis vs dissection). Two groups were identified based on the additional open arch anastomosis resulting in 119 patients (OA) and 159 patients without additional open arch anastomosis (non-OA). Patients with additional valve and coronary artery procedures were also included. Out of the non-OA group, 7 patients were cannulated through the femoral artery and excluded for further analysis. Preoperative baseline characteristics (Table 1), intraoperative and postoperative parameters were extracted from the Dendrite database (Dendrite Clinical Systems Ltd, Reading, UK) and were analysed retrospectively. The follow-up data, including the last echocardiogram, major events, reoperations and death was collected from the hospital software Medfolio® (Nexus AG, Donaueschingen, Germany).
Surgical technique: Freestyle® implantation and open arch anastomosis
Patients were cannulated through the right axillary artery either using an 8mm Dacron® graft, that was sawn beforehand on the artery, or by direct cannulation with the OptiSite® cannula (16-18-20 mm, Edwards Lifesciences Corporation, Irvine, USA) depending on the preference of the surgeon and the quality of the artery. The Freestyle® prosthesis was used as a full root replacement after resection of the native root. The coronary buttons were mobilised and prepared for reimplantation. The left origin of the porcine coronary artery was excised, and the prosthesis orientated towards the native left coronary button to provide a correctly aligned anastomosis. The root prosthesis was implanted using one single 3 − 0 Toplene® suture line (Santec Medicalprodukte GmbH, Grosshostheim, Germany). After reimplantation of the left coronary button with Prolene® 5 − 0 (Ethicon, Sommerville, New Jersey, USA), a new ostium or the original right ostium of the porcine prosthesis was prepared for the anastomosis of the right coronary button.
The indication for replacement of the ascending aorta was made in accordance with the current European guidelines for valvular diseases deciding for a replacement in case of an enlarged aortic diameter of 4.5 cm with a combined valve pathology requireing surgical treatment [11, 12]. The decision to perform hemiarch or open arch anastomosis versus only ascending aortic replacement during the root operation was based on the current available guidelines for the treatment of aortic arch disease focusing on the adjacent aneurysm location in the ascending aorta . If the diameter of the distal ascending aorta measured 4.5 cm affecting the proximal arch, hemiarch replacement was performed.
For the open arch anastomosis, unilateral antegrade cerebral perfusion through right axillary access was initiated after reaching hypothermic conditions of 28°C and by clamping or snaring the brachiocephalic trunc and the left carotid artery leaving the left subclavian artery unblocked. The perfusate was administered with 10–15% of the full flow reaching pressures between 60–120 mmHg and a temperature of 22–28°C. INVOS® Cerebral Oximetry System (Medtronic plc, Dublin, Ireland) was used for transcutanous monitoring of the brain perfusion. CO2 insufflation with 2l/min was used routinely. The aortic clamp was opened in Trendelenburg position and the proximal aortic arch excised as far as indicated. A straight Dacron® prosthesis (sizes between 26-32mm) was anastomosed with 4 − 0 Prolene to the aortic arch. The brachiocephalic trunc and the left carotid artery were declamped and deaired and the aortic clamp was placed on the graft to reinitiate the general perfusion through the axillary access. As the last step, the anastomosis between the Freestyle® prosthesis and the graft was performed during the rewarming of the patient.
After extraction of relevant data from our institutional database, statistical analyses were performed using StatsDirect statistical software, version 3.1 (StatsDirect Ltd, Cambridge, UK). Numerical data were expressed as median and interquartile range (IQR). Nominal and categorical variables were given as absolute numbers and proportions (%). In this study most numerical data were non-normally distributed. Non-parametrical test was used for this reason. The Mann–Whitney U-test was used for the comparison of two groups. When comparing numerical data of more than two groups the Kruskal-Wallis test was performed. When a statistically significant difference was shown, multiple comparisons were done using the Dwass-Steel-Critchlow-Fligner method. Nominal data were compared using the test with the Yates correction. The extended version of the Fisher’s exact test (Fisher–Freeman–Halton) was used for nominal and categorical variables in case of small numbers in some categories. Kaplan-Meier estimate was used to assess mid-term survival.