The midline sternotomy incision offers excellent access to all cardiac structures and has been the traditional route for the performance of most cardiac operations. However, the growth and advancement of percutaneous interventional techniques such as coronary artery stenting have been accompanied by a growing patient willingness to undergo procedures with less favorable mid- and long-term outcomes compared with coronary artery bypass grafting (CABG), largely to enjoy lower periprocedural risks and less invasiveness. In response to this, cardiac surgeons have progressively developed less invasive operations to provide the benefits of standard cardiac procedures with less morbidity. Through the use of innovative surgical approaches combined with new technology, cardiac surgeons have introduced a spectrum of novel, minimally invasive operations. The avoidance of cardiopulmonary bypass (CPB) and/or the full median sternotomy incision are the main features common to this new generation of procedures. This chapter describes and discusses the evolution and results of minimally invasive cardiac surgery, with an emphasis on minimally invasive mitral valve operations.

Although the first attempts to revascularize ischemic myocardium were accomplished on a beating heart, the advent of CPB with cardioplegic arrest offered a reliable and highly reproducible means to perform CABG. During the mid-1990s, in an attempt to circumvent the need for sternotomy, surgeons developed a procedure to create a left internal mammary artery (LIMA)-to-left anterior descending coronary artery (LAD) anastomosis through a small left anterior thoracotomy. In this procedure, termed minimally invasive direct coronary artery bypass (MIDCAB), the LIMA is harvested through a left anterior thoracotomy or limited sternal split incision with the assistance of a variety of chest retractors. A stabilizer is subsequently applied on the beating heart, thus facilitating the performance of the coronary anastomosis (Fig. 54-1). Coronary occlusion is obtained proximally by snaring the coronary artery with a Silastic tape. Early experiences reported good graft patency rates.^1,2 In a meta-analysis including 17 studies of over 4000 patients, Kettering reported early and late mortality rates of 1.3 and 3.2 percent, respectively.³ Conversion to sternotomy or CPB was 1.8 percent. At 6 months of follow-up, 3.6 percent of the 445 grafts that were studied angiographically were occluded; 7.2 percent had significant stenosis. Decreased postoperative pulmonary dysfunction and atrial fibrillation, as well as better pain management and quality of life indices have also been reported in patients undergoing MIDCAB.^4–6 MIDCAB procedures have also been applied with success to coronary reoperations.⁷ Despite the purported advantages of MIDCAB, these minimally invasive procedures remain technically more challenging and require longer operative times with a higher rate of anastomotic revision compared with LIMA-to-LAD anastomoses performed through a standard sternotomy.^8,9 Furthermore, wound complications including incisional hernia, wound dehiscence or infection, and chronic pain, have been reported in up to 9.1 percent of patients.¹⁰ Such wound complications, which are often a source of significant morbidity, are mainly linked to the excessive rib spreading required to harvest the LIMA through a limited thoracotomy. To minimize wound complications, surgeons developed a procedure to harvest the LIMA with a standard thoracoscope, thus minimizing rib spreading.¹¹

Multivessel disease is also amenable to minimally invasive approaches without the use of CPB. The LIMA can be harvested and all myocardial territories can be revascularized via a small left anterolateral thoracotomy, with additional stab incisions made for the introduction of epicardial stabilizers and apical positioners. Proximal anastomoses are constructed either to the LIMA or to the ascending aorta using a proximal connector or direct suture after application of a side-biting clamp. A recent two-center consecutive series of 450 patients, including 7 patients with previous CABG, demonstrated acceptable procedural outcomes, with a 1.3 percent mortality rate, conversion to sternotomy in 3.8 percent, use of CPB in 7.6 percent, stroke in 2 patients (0.4 percent), and thoracic wound infection in 1 patient (0.2 percent).¹² However, these were a highly selected group of patients, with three-vessel disease present in only 28.4 percent of patients; 73.8 percent received only one or two grafts. All 34 patients requiring CPB received a graft to the circumflex coronary system. At a mean follow-up of 19 months, 3 percent of patients required percutaneous coronary intervention, mainly related to problems with the proximal LIMA T-graft anastomosis.

Several dedicated centers have actively pursued the development of robot-assisted totally endoscopic coronary artery bypass grafting (TECAB) on the beating and arrested heart. The first successful TECAB on a human arrested heart was performed by Loulmet and colleagues in 1998.¹³ With this approach, the LIMA is harvested, femoro-femoral CPB is initiated, and the heart is arrested by occluding the ascending aorta with an endovascular balloon and infusing cardioplegic solution into the aortic root. The LIMA-to-LAD anastomosis is performed in a running fashion using robotic instrumentation passed into the chest via modified thoracoscopic ports.

Despite its initial success, TECAB is associated with significant technical limitations. First, aortic cross-clamp and CPB times are usually much greater than with standard techniques.¹⁴ Second, the duration of LIMA harvest is significantly prolonged and LIMA injury can occur in up to 6 percent of patients.¹⁵ Third, the procedure is generally limited to single-vessel bypass of the LAD, although successful multivessel robot-assisted operations have also been reported.¹⁶ In experienced hands, conversion rates to sternotomy or larger thoracic incisions are less than 10 percent. In one study, graft patency was 100 percent at discharge among the 22 out of 45 patients with predischarge angiograms.¹⁷ Furthermore, recently published 5-year clinical outcomes in 100 patients demonstrate greater than 90 percent survival and freedom from angina.¹⁸ However, multicenter randomized controlled data are needed to better define the role of TECAB in the current management of patients with coronary artery disease, especially given the higher prevalence of multivessel and diffuse disease, diabetes, previous percutaneous coronary intervention, and higher urgency in contemporary surgical cohorts. To ensure a safer and more reliable operation, further technological refinements are needed. Improvements in imaging technologies, telemanipulation systems, and endoscopic stabilizers will optimize surgical exposure and anastomotic construction. Direct suturing of the coronary anastomosis remains difficult and time-consuming. Development of distal anastomotic devices may facilitate the performance of TECAB anastomoses on the beating heart. However, variations in patient anatomy, such as intramyocardial LAD or excessive epicardial fat, significantly limit the applicability of these techniques. In such circumstances, the use of endoscopic ultrasound or other imaging modalities may prove useful for the accurate identification of coronary vessels. Although the current literature supports the feasibility of TECAB on the beating or arrested heart using computer-assisted surgical robotic systems, cost issues, limited clinical indications, and a steep learning curve currently limit the widespread use of this technology. Robot-assisted minimally invasive coronary artery bypass surgery remains in its evolutionary stage. Its success will be predicated on a methodical approach, combining improvements in both technology and in surgical techniques.

The era of mitral valve surgery began with the treatment of severe mitral stenosis by closed commissurotomy.¹⁹ However, it was not until the development of the heart–lung machine in 1953 that direct access to the mitral valve became possible. Commissurotomy was then performed under controlled direct vision through a left or right atriotomy. Development of a replacement prosthesis for valves not amenable to repair progressed rapidly after the first mitral valve replacement performed in 1961 by Starr and Edwards.²⁰ The complications encountered with valve replacement with a prosthesis led surgeons to investigate alternative techniques to repair the mitral valve. The establishment of a physiologic classification system of mitral valve disease by Carpentier in the 1970s allowed for a systematic approach to repair the mitral valve.²¹ Further developments in prosthetic rings ensured reproducible and durable mitral valve repairs.²² With refinements in both operative technique and perioperative patient management, mitral repair surgery through a sternal approach can now be performed with mortality rates ranging from 1 to 4 percent with minimal morbidity.^23,24 Furthermore, freedom from reoperation after mitral repair is excellent, especially when the disease is localized to the posterior leaflet and the postoperative echocardiogram shows minimal residual mitral regurgitation.

Although right lateral thoracotomy with femoral cannulation was utilized to expose the mitral valve during the early days of open cardiac surgery, standard sternotomy became the preferred method to access the mitral valve. During the mid-1990s, efforts to avoid a midline sternotomy led to the development of alternate ways to expose the mitral valve. The Cleveland Clinic group initially advocated a parasternal approach but then shifted to an upper midline partial sternotomy. Using this technique, Cosgrove and colleagues reported results similar to a standard sternotomy.²⁵ Other partial sternotomy incisions have also been described, such as the subxiphoid approach consisting of a transverse skin incision overlying the xiphoid process with an inverted J-type mini-sternotomy.²⁶ Advantages of these approaches include central cannulation for CPB as well as good valve exposure. However, the need for sternal division is not obviated and incisions are less esthetically pleasing to patients when compared with the right mini-thoracotomy incision.

During the same time period, other surgeons revisited the right lateral thoracotomy approach to expose the mitral valve. Initial work demonstrated that the mitral valve could be accessed and operations adequately performed through a small right thoracotomy incision.²⁷ Experimental work by the Stanford group led to the development of the Port-Access Endo-CPB system.²⁸ This platform utilizes peripheral cannulation through the femoral vessels for CPB (Fig. 54-2). A branched femoral arterial cannula allows insertion of an endovascular balloon catheter which is positioned in the ascending aorta. Aortic occlusion is performed by inflating the balloon and cardioplegia may be administered through the distal lumen of the catheter. A percutaneous retrograde cardioplegia catheter was also developed and is positioned in the coronary sinus through a right internal jugular venous puncture (Fig. 54-3). Similarly, a venting catheter may be percutaneously positioned in the main pulmonary artery to ensure a bloodless operative field. The mitral valve is exposed using specialized left atrial retractors. Clinical studies performed with the Port-Access system (Edwards Lifesciences, Irvine, CA, USA) demonstrated the feasibility of the procedure.²⁹ However, device-specific complications such as endoclamp-induced retrograde aortic dissection led to further modification of the technique. In 1997, Chitwood introduced a percutaneous transthoracic aortic clamp (Fig. 54-4A), thus eliminating use of the endoclamp and its attendant potential risks such as balloon migration, rupture, and retrograde aortic dissection.³⁰ To enhance mitral valve visualization during the procedure, other groups have proposed the use of two-dimensional video endoscopes.³¹ Robotic approaches involving the da Vinci system (Intuitive Surgical, Inc., Sunnyvale, CA, USA) have also emerged with good short- and midterm outcomes (Fig. 54-5).³²

Throughout the development of the right mini-thoracotomy technique, the establishment of dedicated teams was recognized as a prerequisite to a successful outcome. Anesthesiology expertise in transesophageal echocardiography is of paramount importance for percutaneous catheter positioning. Effective communication between team members during the various steps of the operation is essential. Furthermore, perfusionists need to be facile in the management of novel CPB techniques to ensure optimal functioning of the Port-Access platform and a smooth conduct of the operation.

Currently, most centers favor the right lateral mini-thoracotomy approach for minimally invasive mitral valve operations. CPB is instituted through cannulation of the femoral vessels. Long-shafted specialized instruments are used for mitral valve manipulation (Fig. 54-4B). Although the majority of centers tend to avoid using the endovascular aortic clamp, its use has been proven to be safe in high-volume centers well-experienced in this technique.³³ Conversely, the Chitwood clamp adopted by numerous groups, has been shown to be safe and effective and requires less aortic manipulation. Although cardioplegia may be administered by direct needle puncture of the ascending aorta, the use of a percutaneously inserted coronary sinus catheter for retrograde cardioplegia as well as percutaneous pulmonary artery catheterization for left ventricular venting facilitate exposure by allowing for a bloodless operative field without the presence of cumbersome catheters in the limited work space (Fig. 54-6).

Indications for minimally invasive mitral valve surgery are similar to those that are in effect during conventional mitral surgery through a median sternotomy. However, patient-related factors may influence the decision to proceed with a minimally invasive approach. For example, when considering retrograde aortic perfusion through the femoral artery during CPB, the presence of elevated atherosclerotic plaques >2 mm in height in the descending thoracic aorta or arch may increase the risk of retrograde cerebral and other systemic embolization and constitutes a contraindication to femoral artery-perfused minimally invasive mitral valve surgery. The surgeon must then evaluate alternative arterial cannulation sites or, if not available, resort to a conventional median sternotomy. Other relative contraindications are the presence of significant obesity, a previous right thoracotomy with dense pleural adhesions, a severe chest deformity such as a pectus excavatum or severe scoliosis, and previous breast implant or reconstruction. In our center, over 80 percent of patients referred for an elective isolated mitral valve operation are candidates for a right mini-thoracotomy. The principles of mitral valve replacement and repair during minimally invasive exposure of the mitral valve are no different than those that are now well-established in mitral surgery by standard sternotomy. A decision-making flowchart is shown in (Fig. 54-7).