L K von Segesser

University of Lausanne, Lausanne, Vaud, Switzerland

Are you L K von Segesser?

Claim your profile

Publications (840)2568.96 Total impact

  • Denis Berdajs · Selim Mosbahi · Zalan Forro · Marco Burki · Ludwig K. von Segesser
    [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVES The aim was to determine 3-dimensional (3D) geometrical deformation of the aortic root (AoR) following the David procedure in order to evaluate local haemodynamical conditions of individual AoR elements.
    No preview · Article · Jan 2016 · European Journal of Cardio-Thoracic Surgery
  • Source

    Preview · Article · Dec 2015 · Journal of Cardiothoracic Surgery
  • [Show abstract] [Hide abstract]
    ABSTRACT: Rapid deployment aortic valve replacement (RDAVR) with the use of rapid deployment valve systems represents a smart alternative to the use of standard aortic bioprosthesis for aortic valve replacement. Nevertheless, its use is still debatable in patients with pure aortic valve regurgitation or true bicuspid aortic valve because of the risk of postoperative paravalvular leak. To address this issue, an optimal annulus-valve size match seems to be the ideal surgical strategy. This article describes a new technique developed to stabilize the aortic annulus and prevent paravalvular leak after RDAVR. To confirm the feasibility, this technique was performed in six patients with severe symptomatic aortic stenosis who were scheduled to undergo aortic valve replacement at our center. All patients survived surgery and were discharged from the hospital. There were no new intracardiac conduction system disturbances observed, and a permanent pacemaker implantation was not required in any of the patients. The intraoperative and postoperative echocardiogram confirmed successful positioning of the valve, and no paravalvular leak was observed. In this preliminary experience, RDAVR through a full sternotomy or an upper hemisternotomy approach with the use of aortic annulus stabilization technique was safe, and no leak was observed. Future studies on large series of patients are necessary to confirm the safety and effectiveness of this technique in preventing paravalvular leak in patients with true bicuspid aortic valves or pure aortic regurgitation.
    No preview · Article · Nov 2015 · Innovations Technology and Techniques in Cardiothoracic and Vascular Surgery
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Deep sternal wound infection (DSWI) is a feared complication following cardiac surgery. This study describes clinical, microbiological, and treatment outcomes of DSWI and determines risk factors for complications. Of 55 patients with DSWI, 66% were male and mean age was 68.2years. Initial sternotomy was for coronary artery bypass graft in 49% of patients. Sternal debridement at mean 25.4±18.3days showed monomicrobial (94%), mainly Gram-positive infection. Secondary sternal wound infection (SSWI) occurred in 31% of patients, was mostly polymicrobial (71%), and was predominantly due to Gram-negative bacilli. Risk factors for SSWI were at least 1 revision surgery (odds ratio [OR] 4.8 [95% confidence interval {CI} 1.0-22.4], P=0.047), sternal closure by muscle flap (OR 4.6 [1.3-16.8], P=0.02), delayed sternal closure (mean 27 versus 14days, P=0.03), and use of vacuum-assisted closure device (100% versus 58%, P=0.008). Hospital stay was significantly longer in patients with SSWI (69days versus 48days, P=0.04).
    Full-text · Article · Nov 2015 · Diagnostic microbiology and infectious disease
  • [Show abstract] [Hide abstract]
    ABSTRACT: Objectives: The aim was to define the 3-dimensional (3D) geometrical changes of the aortic root and to determine the local shear stress profile of aortic root elements during the cardiac cycle. Methods: Six sonomicrometric crystals (200 Hz) were implanted into the aortic root of five pigs at the commissures and at the aortic root base (AoB). 3D aortic root deformation including volume, torsion and tilt angle were determined. Geometrical data with measured local flow and pressure conditions was used for computed fluid dynamics modelling of the aortic root. Results: Compared with end-diastole, the sinotubular junction and AoB have maximal expansion at peak ejection: 16.42 ± 6.36 and 7.60 ± 2.52%, and minimal at isovolaemic relaxation: 2.87 ± 1.62 and 1.85 ± 1.79%. Aortic root tilt and rotation angle were maximal at the end of diastole: 17.7 ± 8.8 and 21.2 ± 2.09°, and decreased to 15.24 ± 8.14 and 18.3 ± 0.1.94° at peak ejection. High shear stress >20 Pa was registered at peak ejection at coaptations, and during diastole at the superior two-thirds of the leaflets and intervalvular triangles (IVTs). The leaflet body, inferior one-third of the IVTs and valve nadir were exposed to moderate shear stress (8-16 Pa) during the cardiac cycle. Conclusions: Aortic root geometry demonstrates precise 3D changes of tilt and rotation angle. Reduction of angles during ejection results in a straight cylinder with low shear stress that facilitates the ejection; the increase during diastole results in a tilted frustum with elevated shear stress. Findings can be used for comparative analysis of native and synthetic structures with individual compliance.
    No preview · Article · Sep 2015 · European journal of cardio-thoracic surgery: official journal of the European Association for Cardio-thoracic Surgery
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVES Transapical transcatheter valve procedures are performed through a left minithoracotomy and require apical sutures to seal the apical access site. The use of large-calibre devices compromises any attempt to fully perform the procedure with a thoracoscopic approach or percutaneously. We report our preliminary experience in animals with a new sutureless self-expandable apical occluder, engineered to perform transapical access site closure in a minimally invasive setting with large-size introducer sheaths.
    Preview · Article · Aug 2015 · Interactive Cardiovascular and Thoracic Surgery
  • Source

    Full-text · Dataset · Aug 2015
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Objectives: Repair of the right ventricular outflow tract (RVOT) in paediatric cardiac surgery remains challenging due to the high reoperation rate. Intimal hyperplasia and consequent arteriosclerosis is one of the most important limitation factors for graft durability. Since local shear stress and pressure are predictive elements for intimal hyperplasia and wall degeneration, we sought to determine in an oversized 12-mm RVOT model, with computed fluid dynamics simulation, the local haemodynamical factors that may explain intimal hyperplasia. This was done with the aim of identifying the optimal degree of oversizing for a 12-mm native RVOT. Methods: Twenty domestic pigs, with a weight of 24.6 ± 0.89 kg and a native RVOT diameter of 12 ± 1.7 mm, had valve conduits of 12, 16, 18 and 20 mm implanted. Pressure and flow were measured at 75, 100 and 125% of normal flow at RVOT at the pulmonary artery, pulmonary artery bifurcation and at the left and right pulmonary arteries. Three-dimensional computed fluid dynamics (CFD) simulation in all four geometries in all flow modalities was performed. Local shear stress and pressure conditions were investigated. Results: Corresponding to 75, 100 and 125% of steady-state flow, three inlet velocity profiles were obtained, 0.2, 0.29 and 0.36 m/s, respectively. At inflow velocity profiles, low shear stress areas, ranged from 0 to 2 Pa, combined with high-pressure areas ranging from 11.5 to 12.1 mmHg that were found at distal anastomosis, at bifurcation and at the ostia of the left and right pulmonary arteries in all geometries. Conclusions: In all three oversized geometries, the local reparation of shear stress and pressure in the 16-mm model showed a similar local profile as in the native 12 mm RVOT. According to these findings, we suggest oversizing the natural 12-mm RVOT by not more than 4 mm. The elements responsible for wall degeneration and intimal hyperplasia remain very similar to the conditions present in native RVOT.
    Full-text · Article · Apr 2015 · Interactive Cardiovascular and Thoracic Surgery
  • [Show abstract] [Hide abstract]
    ABSTRACT: The mechanism behind early graft failure after right ventricular outflow tract (RVOT) reconstruction is not fully understood. Our aim was to establish a three-dimensional computational fluid dynamics (CFD) model of RVOT to investigate the hemodynamic conditions that may trigger the development of intimal hyperplasia and arteriosclerosis. Pressure, flow, and diameter at the RVOT, pulmonary artery (PA), bifurcation of the PA, and left and right PAs were measured in 10 normal pigs with a mean weight of 24.8 ± 0.78 kg. Data obtained from the experimental scenario were used for CFD simulation of pressure, flow, and shear stress profile from the RVOT to the left and right PAs. Using experimental data, a CFD model was obtained for 2.0 and 2.5-L/min pulsatile inflow profiles. In both velocity profiles, time and space averaged in the low-shear stress profile range from 0-6.0 Pa at the pulmonary trunk, its bifurcation, and at the openings of both PAs. These low-shear stress areas were accompanied to high-pressure regions 14.0-20.0 mm Hg (1866.2-2666 Pa). Flow analysis revealed a turbulent flow at the PA bifurcation and ostia of both PAs. Identified local low-shear stress, high pressure, and turbulent flow correspond to a well-defined trigger pattern for the development of intimal hyperplasia and arteriosclerosis. As such, this real-time three-dimensional CFD model may in the future serve as a tool for the planning of RVOT reconstruction, its analysis, and prediction of outcome. Copyright © 2015 Elsevier Inc. All rights reserved.
    No preview · Article · Mar 2015 · Journal of Surgical Research
  • Enrico Ferrari · Ludwig von Segesser · Denis Berdajs
    [Show abstract] [Hide abstract]
    ABSTRACT: Coronary bypass grafting remains the best option for patients suffering from multivessel coronary artery disease, and the saphenous vein is used as an additional conduit for multiple complete revascularizations. However, the long-term vein graft durability is poor, with almost 75% of occluded grafts after 10 years. To improve the durability, the concept of an external supportive structure was successfully developed during the last years: the eSVS Mesh device (Kips Bay Medical) is an external support for vein graft made of weft-knitted nitinol wire into a tubular form with an approximate length of 24 cm and available in three diameters (3.5, 4.0 and 4.5 mm). The device is placed over the outer wall of the vein and carefully deployed to cover the full length of the graft. The mesh is flexible for full adaptability to the heart anatomy and is intended to prevent kinking and dilatation of the vein in addition to suppressing the intima hyperplasia induced by the systemic blood pressure. The device is designed to reduce the vein diameter of about 15-20% at most to prevent the vein radial expansion induced by the arterial blood pressure, and the intima hyperplasia leading to the graft failure. We describe the surgical technique for preparing the vein graft with the external saphenous vein graft support (eSVS Mesh) and we share our preliminary clinical results. © The Author 2015. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
    No preview · Article · Jan 2015 · Multimedia Manual of Cardiothoracic Surgery
  • Source
    Ludwig K. von Segesser · Enrico Ferrari · Piergiorgio Tozzi · Saad Abdel-Sayed · Denis Berdajs
    [Show abstract] [Hide abstract]
    ABSTRACT: The advent of retrievable caval filters was a game changer in the sense, that the previously irreversible act of implanting a medical device into the main venous blood stream of the body requiring careful evaluation of the pros and cons prior to execution suddenly became a "reversible" procedure where potential hazards in the late future of the patient lost most of their weight at the time of decision making. This review was designed to assess the rate of success with late retrieval of so called retrievable caval filters in order to get some indication about reasonable implant duration with respect to relatively "easy" implant removal with conventional means, i.e., catheters, hooks and lassos. A PubMed search (www.pubmed.gov) was performed with the search term "cava filter retrieval after 30 days clinical", and 20 reports between 1994 and 2013 dealing with late retrieval of caval filters were identified, covering approximately 7,000 devices with 600 removed filters. The maximal duration of implant reported is 2,599 days and the maximal implant duration of removed filters is also 2,599 days. The maximal duration reported with standard retrieval techniques, i.e., catheter, hook and/or lasso, is 475 days, whereas for the retrievals after this period more sophisticated techniques including lasers, etc. were required. The maximal implant duration for series with 100% retrieval accounts for 84 days, which is equivalent to 12 weeks or almost 3 months. We conclude that retrievable caval filters often become permanent despite the initial decision of temporary use. However, such "forgotten" retrievable devices can still be removed with a great chance of success up to three months after implantation. Conventional percutaneous removal techniques may be sufficient up to sixteen months after implantation whereas more sophisticated catheter techniques have been shown to be successful up to 83 months or more than seven years of implant duration. Tilting, migrating, or misplaced devices should be removed early on, and replaced if indicated with a device which is both, efficient and retrievable.
    Preview · Article · Nov 2014 · Swiss medical weekly: official journal of the Swiss Society of Infectious Diseases, the Swiss Society of Internal Medicine, the Swiss Society of Pneumology
  • Source
    Liang Li · Saad Abdel-Sayed · Denis Berdajs · Enrico Ferrari · Ludwig K von Segesser
    [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVES A new caval tree system was designed for realistic in vitro simulation. The objective of our study was to assess cannula performance for virtually wall-less versus standard percutaneous thin-walled venous cannulas in a setting of venous collapse in case of negative pressure.
    Preview · Article · Nov 2014 · Interactive Cardiovascular and Thoracic Surgery
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Concomitant aortic and mitral valve replacement or concomitant aortic valve replacement and mitral repair can be a challenge for the cardiac surgeon: in particular, because of their structure and design, two bioprosthetic heart valves or an aortic valve prosthesis and a rigid mitral ring can interfere at the level of the mitroaortic junction. Therefore, when a mitral bioprosthesis or a rigid mitral ring is already in place and a surgical aortic valve replacement becomes necessary, or when older high-risk patients require concomitant mitral and aortic procedures, the new 'fast-implantable' aortic valve system (Intuity™ valve, Edwards Lifesciences, Irvine, CA, USA) can represent a smart alternative to standard aortic bioprosthesis. Unfortunately, this is still controversial (risk of interference). However, transcatheter aortic valve replacements have been performed in patients with previously implanted mitral valves or mitral rings. Interestingly, we learned that there is no interference (or not significant interference) among the standard valve and the stent valve. Consequently, we can assume that a fast-implantable valve can also be safely placed next to a biological mitral valve or next to a rigid mitral ring without risks of distortion, malpositioning, high gradient or paravalvular leak. This paper describes two cases: a concomitant Intuity™ aortic valve and bioprosthetic mitral valve implantation and a concomitant Intuity™ aortic valve and mitral ring implantation.
    Preview · Article · Jul 2014 · Interactive Cardiovascular and Thoracic Surgery
  • Léa Hochstrasser · Patrick Ruchat · Nicole Sekarski · Michel Hurni · Ludwig K von Segesser
    [Show abstract] [Hide abstract]
    ABSTRACT: Objectives: To evaluate long-term outcome of initial aortic valve intervention in a paediatric population with congenital aortic stenosis, and to determine risk factors associated with reintervention. Patients and methods: From 1985 to 2009, 77 patients with congenital aortic stenosis and a mean age of 5.8±5.6 years at diagnosis were followed up in our institution for 14.8±9.1 years. Results: First intervention was successful with 86% of patients having a residual peak aortic gradient <50 mmHg, and the proportion of patients with grade >1 regurgitation increased by 7%. Long-term survival after the first procedure was excellent, with 91% survival at 25 years. At a mean interval of 7.6±5.3 years, 30 patients required a reintervention (39%), mainly because of a recurrent aortic stenosis. Freedom from reintervention was 97, 89, 75, 53, and 42% at 1, 10, 15, 20, and 25 years, respectively. Predictors of reintervention were residual peak aortic gradient (p=0.0001), aortic regurgitation post-intervention >1 (p=0.02), prior balloon aortic valvuloplasty (p=0.04), and increased left ventricular posterior wall thickness (p=0.1). Conclusions: Aortic valve intervention is a safe and effective procedure for congenital aortic stenosis with excellent survival results. However, rate of reintervention is high and influenced by increased left ventricular posterior wall thickness pre-intervention, prior balloon valvuloplasty, higher residual peak systolic valve gradient, and more than mild regurgitation post-intervention. The study highlights that long-term follow-up is recommended for these patients.
    No preview · Article · Jul 2014 · Cardiology in the Young
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVES The reconstruction of the right ventricular outflow tract (RVOT) with valved conduits remains a challenge. The reoperation rate at 5 years can be as high as 25% and depends on age, type of conduit, conduit diameter and principal heart malformation. The aim of this study is to provide a bench model with computer fluid dynamics to analyse the haemodynamics of the RVOT, pulmonary artery, its bifurcation, and left and right pulmonary arteries that in the future may serve as a tool for analysis and prediction of outcome following RVOT reconstruction.
    Full-text · Article · Jun 2014 · Interactive Cardiovascular and Thoracic Surgery
  • [Show abstract] [Hide abstract]
    ABSTRACT: There is controversy over the use of the Ross procedure with regard to the sub-coronary and root replacement technique and its long-term durability. A systematic review of the literature may provide insight into the outcomes of these two surgical subvariants. A systematic review of reports between 1967 and February 2013 on sub-coronary and root replacement Ross procedures was undertaken. Twenty-four articles were included and divided into (i) sub-coronary technique and (ii) root replacement technique. The 10-year survival rate for a mixed-patient population in the sub-coronary procedure was 87.3% with a 95% confidence interval (CI) of 79.7-93.4 and 89.1% (95% CI, 85.3-92.1) in the root replacement technique category. For adults, it was 94 vs 95.3% (CI, 88.9-98.1) and in the paediatric series it was 90 vs 92.7% (CI, 86.9-96.0), respectively. Freedom from reoperation at 10 years was, in the mixed population, 83.3% (95% CI, 69.9-93.4) and 93.3% (95% CI, 89.4-95.9) for sub-coronary versus root replacement technique, respectively. In adults, it was 98 vs 91.2% (95% CI, 82.4-295.8), and in the paediatric series 93.3 vs 92.0% (95% CI, 86.1-96.5) for sub-coronary versus root replacement technique, respectively. The Ross procedure arguably has satisfactory results over 5 and 10 years for both adults and children. The results do not support the advantages of the sub-coronary technique over the root replacement technique. Root replacement was of benefit to patients undergoing reoperations on neoaorta and for long-term survival in mixed series.
    No preview · Article · Apr 2014 · European journal of cardio-thoracic surgery: official journal of the European Association for Cardio-thoracic Surgery
  • [Show abstract] [Hide abstract]
    ABSTRACT: In extreme situations, like hyperacute rejection of heart transplant or major heart trauma, heart preservation may not be possible. Our experimental team works on a project of peripheral ECMO support in acardia as a bridge to heart transplantation or artificial heart implantation. An extracorporeal membrane oxygenation (ECMO) support was established in 5 calves (58.6 ± 6.9kg) by the transjugular insertion to the caval axis of a self-expanded cannula, with carotid artery return. After baseline measurements, ventricular fibrillation was induced, great arteries were clamped, heart was excised and right and left atria remnants, containing pulmonary veins, were sutured together leaving an atrial septal defect over the caval axis cannula. Measurements of pump flow and arterial pressure were taken with the pulmonary artery (PA) clamped and anastomosed with the caval axis for a total of 6 hours. PA anastomosis to the caval axis provided an acceptable 6 hours hemodynamic stability, permitting a peripheral access ECMO support in extreme scenarios indicating a heart explantation.
    No preview · Article · Mar 2014 · ASAIO journal (American Society for Artificial Internal Organs: 1992)
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: During open heart surgery, so-called atrial chatter, a phenomenon due to right atria and/or caval collapse, is frequently observed. Collapse of the cava axis during cardiopulmonary bypass (CPB) limits venous drainage and may result downstream in reduced pump flow on (lack of volume) and upstream in increased after-load (stagnation), which in turn may both result in reduced or even inadequate end-organ perfusion. The goal of this study was to reproduce venous collapse in the flow bench. In accordance with literature for venous anatomy, a caval tree system is designed (polyethylene, thickness 0.061 mm), which receives venous inflow from nine afferent veins. With water as medium and a preload of 4.4 mmHg, the system has an outflow of 4500 ml/min (Scenario A). After the insertion of a percutaneous venous cannula (23-Fr), the venous model is continuously served by the afferent branches in a venous test bench and venous drainage is augmented with a centrifugal pump (Scenario B). With gravity drainage (siphon: A), spontaneously reversible atrial chatter can be generated in reproducible fashion. Slight reduction in the outflow diameter allows for generation of continuous flow. With augmentation (B), irreversible collapse of the artificial vena cava occurs in reproducible fashion at a given pump speed of 2300 ± 50 RPM and a pump inlet pressure of -112 mmHg. Furthermore, bubbles form at the cannula tip despite the fact that the entire system is immersed in water and air from the environment cannot enter the system. This phenomenon is also known as cavitation and should be avoided because of local damage of both formed blood elements and endothelium, as well embolization. This caval model provides a realistic picture for the limitations of flow due to spontaneously reversible atrial chatter vs irreversible venous collapse for a given negative pressure during CPB. Temporary interruption of negative pressure in the venous line can allow for recovery of venous drainage. This know-how can be used not only for testing different cannula designs, but also for further optimizing perfusion strategies.
    Preview · Article · Jan 2014 · European journal of cardio-thoracic surgery: official journal of the European Association for Cardio-thoracic Surgery
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The aim of this study was to evaluate the risk factors associated with Contegra graft (Medtronic Minneapolis, MN, USA) infection after reconstruction of the right ventricular outflow tract. One hundred and six Contegra grafts were implanted between April 1999 and April 2010 for the Ross procedure (n = 46), isolated pulmonary valve replacement (n = 32), tetralogy of Fallot (n = 24), double-outlet right ventricle (n = 7), troncus arteriosus (n = 4), switch operation (n = 1) and redo of pulmonary valve replacement (n = 2). The median age of the patients was 13 years (range 0-54 years). A follow-up was completed in all cases with a median duration of 7.6 years (range 1.7-12.7 years). There were 3 cases of in-hospital mortality. The survival rate during 7 years was 95.7%. Despite the lifelong endocarditis prophylaxis, Contegra graft infection was diagnosed in 12 (11.3%) patients at a median time of 4.4 years (ranging from 0.4 to 8.7 years). Univariate analysis of preoperative, perioperative and postoperative variables was performed and the following risk factors for time to infection were identified: female gender with a hazard ratio (HR) of 0.19 (P = 0.042), systemic-to-pulmonary shunt (HR 6.46, P < 0.01), hypothermia (HR 0.79, P = 0.014), postoperative renal insufficiency (HR 11.97, P = 0.015) and implantation of permanent pacemaker during hospitalization (HR 5.29, P = 0.075). In 2 cases, conservative therapy was successful and, in 10 patients, replacement of the infected valve was performed. The Contegra graft was replaced by a homograft in 2 cases and by a new Contegra graft in 8 cases. Cox's proportional hazard model indicated that time to graft infection was significantly associated with tetralogy of Fallot (HR 0.06, P = 0.01), systemic-to-pulmonary shunt (HR 64.71, P < 0.01) and hypothermia (HR 0.77, P < 0.01). Contegra graft infection affected 11.3% of cases in our cohort, and thus may be considered as a frequent entity that can be predicted by both intraoperative and early postoperative factors. After the diagnosis of infection associated with the Contegra graft was confirmed, surgical treatment was the therapy of choice.
    Preview · Article · Jan 2014 · European journal of cardio-thoracic surgery: official journal of the European Association for Cardio-thoracic Surgery
  • Source
    Ludwig K von Segesser · Friedhelm Beyersdorf

    Preview · Article · Dec 2013 · Interactive Cardiovascular and Thoracic Surgery

Publication Stats

15k Citations
2,568.96 Total Impact Points

Institutions

  • 2001-2015
    • University of Lausanne
      • • Division of Cardiovascular Surgery
      • • Faculty of Biology and Medicine (FBM)
      Lausanne, Vaud, Switzerland
  • 1996-2015
    • University Hospital of Lausanne
      • Service de chirurgie cardio-vasculaire (CCV)
      Lausanne, Vaud, Switzerland
    • Triemli City Hospital
      Zürich, Zurich, Switzerland
  • 1989-2007
    • University of Zurich
      • • Clinic for Small Animal Surgery
      • • Department of Cardiovascular Surgery
      • • Internal Medicine Unit
      Zürich, ZH, Switzerland
  • 2006
    • École Polytechnique Fédérale de Lausanne
      • Laboratory of Hemodynamics and Cardiovascular Technology
      Lausanne, Vaud, Switzerland
  • 2003
    • Centre hospitalier universitaire de Sherbrooke
      Sherbrooke, Quebec, Canada
  • 1989-1997
    • Schulthess Klinik, Zürich
      Zürich, Zurich, Switzerland
    • University Hospital Zürich
      Zürich, Zurich, Switzerland
  • 1994-1996
    • Inselspital, Universitätsspital Bern
      • Department of Cardiology
      Berna, Bern, Switzerland
  • 1991
    • Zürcher Höhenklinik Wald
      Zürich, Zurich, Switzerland
  • 1988-1991
    • Universitätsspital Basel
      • Clinical Trial Unit
      Bâle, Basel-City, Switzerland