Infective endocarditis can be associated with several clinical and anatomic complications. Among these is aortic root abscess. Aortic root abscess is defined by the presence of necrotic tissue in the aortic root, annulus, or as aortoventricular discontinuity. In this chapter, the presentation, diagnosis, and surgical management of endocarditis complicated by aortic root abscess will be discussed.
Clinical presentation and diagnosis
The definitive diagnosis of endocarditis is based upon a combination of clinical, echocardiographic, and microbiology findings. The modified Duke criteria, introduced by Durack et al., are utilized by most clinicians for the diagnosis of endocarditis and they have been reported to have a sensitivity and specificity of over 80% [ ].
Major criteria include, but are not limited to, positive blood cultures and new echocardiographic findings of cardiac valve vegetation or regurgitation, oscillating intracardiac mass, partial dehiscence of a prosthetic valve, and abscess. Aortic root abscess with extension to surrounding structures may also lead to symptoms prompting presentation. These include involvement of the conduction system with arrhythmias or heart block, compression of the coronary arteries with acute coronary syndromes, or pseudoaneurysm formation with resultant chest pain and/or compression of adjacent structures. In extreme situations, the abscess can erode into adjacent cardiac chambers leading to fistulous connections within the heart. Frequently, patients require cardiac surgical attention after presenting with positive blood cultures and one of the aforementioned echocardiographic or radiologic findings.
It is estimated that intracardiac abscess will affect up to 14% of patients suffering from infective endocarditis [ , ]. Staphylococcus aureus is the most frequently isolated pathogen affecting approximately one-third of patients; however, many others have been implicated such as Haemophilus species, Enterococci, Beta-hemolytic streptococci, Bacteroides species, and others [ ].
Transthoracic echocardiogram (TTE) is frequently employed as the first diagnostic test in the setting of suspected endocarditis. However, the ability of TTE to detect new valvular vegetations is only 50%. Transesophageal echocardiogram (TEE), in contrast, has a sensitivity of 90%–100% in the detection of cardiac vegetations ( Fig. 18.1 ). When employed together TTE and TEE demonstrate vegetations in 90% of patients suffering from endocarditis. Further, TEE is the diagnostic test of choice for demonstrating aortic root abscess as well as fistulas and shunts [ , ].
It is our practice to undertake CT scanning of the brain, chest, abdomen, and pelvis to screen patients for possible septic emboli to the brain, lungs, and spleen as well as to evaluate for the presence and extent of a pseudoaneurysm when they present with infective endocarditis. CT scanning also allows for detailed delineation of the aortic root anatomy which is useful in operative planning, such as during dissection of an aortic root pseudoaneurysm ( Figs. 18.2 and 18.3 ).
A left heart catherization may also be useful in operative planning and may be undertaken safely if a large, mobile vegetation is not visualized on the aortic valve by echocardiography.
The first step in successful management of infective endocarditis complicated by aortic root abscess is the assembly of a multidisciplinary team that includes clinicians who specialize in cardiology, cardiac surgery, intensive care, and infectious diseases. The cornerstones of treatment include infection eradication, as evident by negative blood cultures, and prevention of further cardiac damage and systemic complications.
Microbiological data suggest that staphylococci and streptococci are the causative agents in approximately 80% of aortic root abscess cases. Within these geniuses, it should be noted that the incidence of S. aureus –related infection is increasing while the incidence of streptococcus viridans is declining. When fungus is isolated as the causative organism it is most frequently Candida species followed by Aspergillus species [ ].
Prompt antibiotic therapy is required and should be employed as soon as infective endocarditis is clinically suspected. Initial treatment is with broad-spectrum agents, and this is then tailored based upon blood cultures and susceptibilities.
Indications for surgical intervention
When periannular extension, in the form of aortic root abscess or fistula, is present this alone is an indication for early surgical intervention. Additional indications include congestive heart failure, persistent bacteremia, septic emboli, or electrical rhythm disturbances. This is frequently associated with new-onset aortic and/or mitral valve regurgitation or prosthetic-valve dehiscence and/or obstruction [ ].
Systemic embolization also is an indication for surgical intervention. This also includes cerebrovascular complications including silent complications and transient ischemic attacks. Of note, surgery is relatively contraindicated for 1 month after an intracranial hemorrhage unless a neurosurgical procedure can be undertaken to reduce the risk of subsequent bleeding. However, in patients who are extremely ill from aortic root abscess the relative risks and benefits of waiting versus proceeding with surgery may favor early surgery [ , ].
Finally, persistent sepsis, as defined by persistent fevers and/or positive blood cultures after 5–7 days of appropriate antibiotic therapy, is an indication for surgical intervention [ ].
A small subset of patients may be managed without surgery. They include those with small abscesses measuring less than 1 cm not complicated by heart block, valvular dehiscence, or insufficiency and those that do not progress while the patient is undergoing antibiotic therapy. If a nonoperative approach is undertaken, we recommend serial exams with echocardiography at 2, 4, and 8 weeks after presentation [ ].
Principles of surgical management
Once the decision to pursue operative intervention is made, as will be the situation in the setting of an aortic root abscess in a patient fit for an operation, there are several principles that will facilitate a successful outcome. Even though there is not one “standard” operation for patients suffering from an aortic root abscess, all patients should be prepared in the same fashion. This includes a detailed and thorough TEE in the operating room. Further, given that patients have frequently undergone prior cardiac operations and that the disease process has a predilection for adhesion formation, the surgeon should prepare for difficult dissection. It is prudent to obtain femoral arterial access prior to dividing the sternum, should expeditious cardiopulmonary bypass be required. The bilateral lower extremities should be prepped into the sterile field in case conduit is needed for bypass grafting.
There are many different techniques for safely performing redo sternotomy. The surgeons should choose the technique they are most familiar with, but the importance of femoral access cannot be overstated. Once the median sternotomy has been completed, standard cannulation with an aortic cannula and dual-stage venous cannula is acceptable. However, it is our preference to utilize bicaval cannulation in these complex root operations. On occasion, severe adhesion formation to the right atrium, or fistulous connection to the right-sided cardiac structures requires opening the right atrium. This technique does greatly improve visualization of the aortic root in situations of dense adhesion formation. Given the association of aortic valve insufficiency with this disease process, we also routinely place a left ventricular vent via the right superior pulmonary vein.
Once cardiopulmonary bypass is initiated it is our preference to utilize retrograde cardioplegia followed by ostial cardioplegia once the aorta is opened. Given the long cardiopulmonary bypass and myocardial ischemia times in these complex root or multivalvular reconstructions, we perform systemic cooling, often to 30°C for additional metabolic protection.
After myocardial arrest has been achieved, the infected area should be exposed typically through transverse aortotomy, in order to identify the extent of involvement. This evaluation dictates the extent of root reconstruction required. An added principle to achieving a successful short- and long-term repair is to undertake a thorough and complete debridement of all infected and necrotic material. After the debridement is concluded we frequently place Rifampin-soaked cotton swabs into the aortic annulus and into the left ventricle in an attempt to sterilize the field.
Finally, at the conclusion of the case, it is our practice to utilize an irrigation catheter to infuse antibiotic solution into the mediastinum for 5–7 days after the operation. This is frequently a simple 16 French catheter with tip removed and positioned within the mediastinum. This catheter is tunneled through the chest wall in the second or third intercostal space to right of the sternum. Although the combination of Rifampin-soaked cotton swabs and mediastinal irrigation has not been definitively proven to improve outcomes we have found it useful in our practice, particularly for gram-positive organisms.
The most appropriate prosthetic-valve choice for patients suffering from endocarditis complicated by aortic root abscess requiring valve replacement and left ventricular outflow tract (LVOT) reconstruction has not been definitively shown in the literature [ , ]. Depending upon the patient’s age, comorbid conditions, and the extent of annular destruction a mechanical or biological prosthetic with patching may be the optimal choice. Biological or mechanical composite aortic root conduits, stentless xenograft root grafts, cadaveric homograft, or pulmonary autografts may all be appropriate choices. Most studies delineating the outcomes associated with these choices, as well as comparing one to another, are small in sample size and confined to single-institution series. As such, interpretation of these data is difficult. Nevertheless, no compelling data have shown one graft choice to be more associated with reinfection over another. Jassar et al. noted in a study comparing the use of composite grafts (both biological and mechanical) to homografts that the type of conduit or prosthetic chosen did not have a significant influence upon in-hospital or long-term mortality [ ]. Avierinos et al., in a similar study, noted similar findings [ ].
Each prosthesis option has relative advantages and disadvantages which will be discussed further.
Many surgeons consider a cadaveric homograft with a preserved anterior mitral valve leaflet as the standard prosthesis choice in patients with aortic root abscess. Homografts have several appealing features in this situation. They have significant tissue flexibility and can be conformed to fit an aortic annulus and root that has been compromised by infection and subsequent debridement. Further, the preserved mitral valve leaflet can be utilized for intracardiac repairs such as ventricular septal defect closure and/or fistula exclusion. Finally, the biologic nature of the graft may aid in the prevention of early reinfection and simplify management of reinfection at a later time point.
Homografts do have several drawbacks. The grafts themselves have limited availability and require a technically complex operation for implantation. They are therefore typically performed at select centers with the ability for cryopreservation as well. Homografts also frequently fail by structural deterioration with significant associated calcification and torn leaflet(s), which makes reoperation difficult and fraught with risk. While transcatheter aortic valve replacement is an option in a degenerated homograft aortic valve, there are often anatomic barriers due to extreme calcification and narrowing of the homograft conduit, making high-risk open reoperation necessary.
Stentless xenograft aortic roots have also been compared to homografts. No definitive evidence suggests that stentless aortic roots are superior to the other available options. However, stentless xenografts have the advantage of being readily available in several sizes and have demonstrated excellent hemodynamic performance and long-term survival similar to homografts. Stentless xenograft failure, similar to homografts, is also frequently associated with significant calcification and torn leaflet(s) with associated severe aortic insufficiency [ ].
As no definitive evidence exists demonstrating the superiority of one valve and/or conduit choice over another, the decision rests mostly with surgeon preference. Aortic root replacement with or without annular reconstruction in the setting of infective endocarditis is a technically challenging operation and it is prudent that the surgeon utilizes the prosthesis that he or she is most comfortable with noting that perioperative survival should take precedent over long-term results or reoperation risks.
The need for patch material is frequently encountered during the course of surgery for aortic root abscess and biologic or synthetic patch material should be readily available. A large variety of patch material is available to the surgeon, including autologous pericardium, glutaraldehyde-fixed equine pericardium, aortic root homograft wall, Dacron, autologous rectus abdominis fascia, and bovine pericardium.
It is our preference to utilize glutaraldehyde-fixed bovine pericardium for reconstruction in the setting of an infected field. This option limits the amount of prosthetic material placed within the field and is easily handled and modified, allowing it to be used in many different settings and types of reconstruction. David et al. have noted good mid- and long-term results with this approach [ ]. However, some investigators have raised concerns about the long-term durability of this material with some citing late degeneration and associated paravalvular leaks. Others have demonstrated Dacron to be a suitable material with good long-term outcomes [ ].
As with prosthetic-valve choices, there is no definitive evidence that one material is superior to another. As such, surgeons should choose a material that they are familiar with and will allow for a successful repair in their hands.
Isolated aortic valve replacement
Several centers have reported their outcomes with isolated aortic valve replacement in the setting of an aortic root abscess. These highly selected patients include those with an abscess cavity limited to one cusp of the aortic valve, without extension into adjacent structures or cardiac cambers as well as an annulus amenable to simple reconstruction. Frequently the defect created by the abscess is closed via plication or patched using pericardium. Given that these reports frequently demonstrate inferior outcomes with aortic valve replacement alone we tend to favor full root replacement in patients presenting with an abscess [ ].
The double patch technique
The double patch technique, first described by David et al., is a surgical approach utilized to treat endocarditis that involves the aortic and mitral valves as well as the intervalvular fibrous body (IFB) [ ]. It has been reported and modified by several centers and is known by several names, including the “Commando Procedure” and the “UFO Operation.”
Typically, this approach utilizes a standard median sternotomy, bicaval cannulation, antegrade, retrograde, and ostial cardioplegia as well as systemic cooling to 32°C. Once an arrest is achieved, a transverse aortotomy is performed and the aortic valve is inspected. The aortic valve is then excised and confirmation of IFB involvement is undertaken. Once confirmed and the surgeon has committed to a full aortic root replacement the aortotomy is extended into the noncoronary sinus and across the annulus to the level of the dome of the left atrium. The tissue of the aortic root is fully excised with the exception of the coronary buttons. A 2–4 mm rim of aortic tissue around each coronary ostium is typically preserved for the coronary reimplantation. Coronary mobilization is then undertaken. It is critical to fully mobilize the coronary arteries in order to facilitate the anastomosis and eliminate any associated tension. The mitral valve then is exposed and excised along with the IFB and any grossly involved tissue. The extent of resection of the IFB can vary and is determined by the extent of the disease. It may be limited to the central portion associated with the anterior leaflet of the mitral valve or may involve the complete resection of the mitral valve, the IFB, and the dome of the left atrium. It is critical to remove all infected and prosthetic material.
Next, the mitral prosthesis is inserted. Sizing of the valve is critical as too large a valve can cause LVOT obstruction. In the setting of a prior mitral valve prosthesis, it is prudent to simply place a new valve of the same size. If implanting a new prosthesis it is easier to size the valve prior to excision of the IFB, when possible. Ideally, the valve should be seated such that two-thirds of the sewing ring rest within the posterior mitral annulus. Further, care should be taken to orient the valve such that a strut post does not protrude into the LVOT.
The IFB is then reconstructed using a “two-tongued” bovine pericardial patch which is attached to the anterior aspect of the new mitral valve. Extreme care should be taken with needle placement of this suture line to avoid potentially serious bleeding after separation from bypass which can be difficult to visualize once the heart has been reanimated. One wing of the patch is then sewn to the open aspects of the left atrium and the other is secured to the free margins of the LVOT and aortic wall thus creating a new IFB. If the patient only requires aortic valve replacement, this portion of the patch may then be incorporated into the aortotomy closure. If an aortic root replacement is being undertaken the patch is divided a few centimeters above the aortic annulus and can serve as an anchoring point for the aortic root sutures.
Aortic homograft implantation
If an aortic root replacement is being undertaken, several centers have argued that an aortic homograft is the ideal choice, as it utilizes no prosthetic material and limited resections and repairs of the IFB and mitral valve may be accomplished with the associated anterior mitral valve leaflet that accompanies the homograft. The initial stages of the operation, that is, exposure, cannulation, and protection, are the same as described above. Once all infected tissues have been removed, the homograft is then sized. Sizing is a critical step and undersizing should be avoided as it is associated with several complications, including early and late graft failures, infection, pseudoaneurysm formation, and leaflet prolapse. Once all infected tissues have been removed the aortic annulus is sized. It is our preference to utilize Hegar dilators to approximate the internal diameter of the aortic annulus. We then select a homograft roughly equal in diameter to the measured annulus size. As opposed to homograft implantation for noninfective primary aortic disease, often the annulus is not markedly dilated in endocarditis. Therefore annular reduction is not required, and a homograft diameter roughly equal to the measured annulus can be selected.
Homografts may be implanted in a variety of ways; however, a complete aortic root replacement operation is the most commonly utilized technique in the setting of an aortic root abscess. Implantation is frequently complicated by the loss of normal anatomic landmarks to infection. In order to address this issue, the homograft anterior mitral valve leaflet can be oriented with the patient’s fibrous trigones and anterior mitral valve leaflet.
Prior to insertion of the homograft, it is our preference to trim the associated LVOT muscle and a strip of bovine pericardium around this structure. It is secured using a running, horizontal, 3-0 prolene suture. The homograft is then inserted and secured, ensuring that undue tension or torsion of the LVOT has not occurred. Homograft can be anastomosed to the patient’s native LVOT with running 3-0 prolene sutures, with interrupted horizontal mattress sutures or with simple interrupted sutures. Coronary buttons are then reimplanted in standard fashion and bovine pericardium is then utilized to close any associated defect within the left atrium.
Stentless aortic valve implantation
Although stentless aortic valve implantation is technically more challenging than traditional aortic valve replacement, stentless valves, such as the Freedom prosthesis (LivaNova, Milan, Italy), offer several attractive features for surgery in the setting of an aortic root abscess.
The Freedom prosthesis may be inserted with the valve inverted within the LVOT, allowing for a better sealing suture line and deeper implantation, if needed. Further, the valve has no synthetic fabric reinforcement, likely reducing the incidence of ongoing or reinfection. However, patients suffering from endocarditis complicated by an aortic root abscess frequently require full root replacement and thus limit the role of the Freedom prosthesis in this setting. In these cases a porcine root, such as the Freestyle prosthesis (Medtronic, Minneapolis, MN), can be utilized.
We prefer to use Hegar dilators to size the Freestyle prosthesis. The dilators are used to measure the annulus diameter, and consideration is also made regarding the amount of scar tissue and space in the aortic root. An advantage of the Freestyle prosthesis is that it is stentless and conformable to narrow spaces, with excellent hemodynamics. Therefore, it is not necessary to oversize the prosthesis. Typically, we select a Freestyle prosthesis 2–3 mm larger than the largest size Hegar dilator that can be accommodated in the aortic root. The accompanying Freestyle sizers can be used for measurement purposes as well.
It is our preference to insert the Freestyle prosthesis utilizing simple, interrupted 2-0 Ethibond sutures ( Fig. 18.4 ). The porcine root is then oriented such that one of the coronary ostia is directly aligned with the patient’s left main coronary artery. The valve is then parachuted into place and small strip of felt is placed between the valve sutures, creating a felt reinforcement of the aortic root suture line as a gasket ( Fig. 18.5 ). The valve sutures are then tied and the left coronary button is reimplanted with running 4-0 or 5-0 prolene. Frequently, the second coronary ostium on the prosthesis does not align well with the patient’s right coronary artery. Here, we prefer to oversew the second unused porcine root ostium with a 3-0 prolene suture and create a new, circular hole within the prosthesis with a 15 blade scalpel and then reimplant the right coronary button in the standard fashion. It is imperative to mobilize both coronary ostia extensively to avoid undue tension and also to reimplant them in appropriate anatomic location to avoid kinking which may compromise diastolic coronary perfusion when the heart is reanimated. Many surgeons recommend restoring aortic continuity prior to right coronary reimplantation in order to pressurize the aorta and fill the heart before selecting the site for right coronary reimplantation. A potential drawback to this approach is difficulty in viewing the porcine valve leaflets when making the opening, and thus extreme care should be taken during this maneuver. It is prudent to mark the location of the right-non commissure on the external surface of the xenograft prior to performing the aortic anastomosis. This serves as a guide for where to avoid when making the opening for the right coronary reimplantation.