Chapter 11. Prosthetic Valves

Over the past three decades, significant advances have been made with respect to the evolution of prosthetic heart valves and the manner in which they are evaluated in vivo. New materials and configurations of mechanical and biologic valves have improved the quality of life for patients around the world. However, despite the benefits they confer by replacing a diseased valve, all prosthetic valves have inherent limitations and possible complications that may develop in the immediate or late post-implantation period. This chapter provides an overview of (1) the clinical indications for implantation of prosthetic heart valves; (2) the major types of prostheses and (3) their evaluation using transesophageal echocardiography (TEE); (4) some inherent limitations of echocardiography; and (5) common prosthetic pathology.

Similar to the assessment of native cardiac valves (see Chapters 7, 9, and 10), TEE can provide a detailed evaluation of the structure, function, and integrity of prosthetic valves using two-dimensional (2D), color-flow Doppler (CFD), continuous-wave Doppler (CWD), pulsed-wave Doppler (PWD), and three-dimensional (3D) imaging modalities. However, several features unique to the intraoperative environment such as dynamic changes to preload, myocardial contractility, and afterload can create challenges for the TEE assessment of prosthetic valves. In addition, prosthesis malfunction can affect adjacent valves and chambers, and this must be considered in the assessment. Nevertheless, the integration of information from the TEE examination can provide a comprehensive assessment of the patient's overall cardiac status.

A diseased native valve that is symptomatic will ultimately require intervention (repair or replacement) by percutaneous or surgical means. The most commonly replaced valves are those in the aortic and mitral positions commensurate with the high prevalence of diseases that involve these valves. Recently, with significant improvements in mitral valve repair, the numbers of valves implanted for mitral regurgitation have been decreasing and stenosis has become a more common indication for mitral valve replacement. In addition, with a large number of prosthetic valves already implanted, pathologies involving these replaced valves are increasing the frequency of repeat surgery. Often presenting as new or increased valvular regurgitation in the setting of a systemic infection, endocarditis of the native valve or prosthetic grafts is an infrequent yet important indication for placement of a new prosthesis. In addition, inadequate anticoagulation can lead to acute or chronic stenosis of previously placed valves. The myriad of systemic diseases that can cause valvular dysfunction are comprehensively discussed in the ACC/AHA guidelines for evaluation of patients with valve disease.1

When faced with a patient who needs a valve replacement, cardiologists and cardiac surgeons have a wide variety of valves from a number of manufacturers to select from. The decision as to which prosthesis should be implanted depends on a number of factors including the patient's age, gender (women of child-bearing age), life expectancy, comorbidities, location and size of the diseased valve, surgical expertise or preference, and considerations regarding anticoagulation (noncompliance or likely loss to follow-up).2 A detailed description of the process of valve selection is beyond the scope of this text and is discussed elsewhere.1

All valves can be classified as mechanical or biologic depending on the predominant material of composition. While mechanical valves are further classified according to the occlusion device, biologic valves are subclassified according to the presence of synthetic support structures (stented or stentless) and by the origin of the valve tissue (xenografts, homografts, or autografts).2–4

Mechanical Valves

Mechanical valves are available in a variety of sizes and designs. All mechanical valves are comprised of a sewing ring (used to secure the valve in the patient) and an occluding device (designed to provide one-way flow through the prosthesis). The design of the occluding device is used to further subclassify this type of prosthesis into one of three major categories: (1) ball-in-cage, (2) tilting disk, and (3) bileaflet.

Ball‐in-Cage

The first successful prosthetic heart valve implantation in the early 1960s was of the ball-in-cage design. The Starr-Edwards valve was the most widely used, with more than 200,000 prostheses placed since its introduction into clinical practice in 1965.3,5 Although it has undergone several modifications, the basic design remains unchanged, with a silastic ball constrained within a sewing ring comprised of Teflon or polypropylene cloth and two stellite alloy U-shaped arches that form a cage (Figure 11–1). The ball travels forward into the cage during antegrade flow, with the blood passing between the sewing ring and under the trailing edge of the ball between the struts. The occluder then moves back to seat snugly against the sewing ring when pressures equalize between the two chambers. There is no regurgitant flow during elevation of pressures in the receiving chamber.

Their bulky design and flow characteristics (the smaller the valve, the more obstructive in nature the valve becomes) limit their use to the mitral position. They are also not typically implanted in individuals with small left ventricular cavities or with small aortic annular diameters. Although durable and therefore still seen in patients presenting for repeat surgery, these valves are no longer implanted, having been supplanted by valves with a better hemodynamic profile.

Tilting Disk

Representatives of this design include the Medtronic-Hall (Medtronic, Minneapolis, MN), OmniScience (MedicalCV Inc., Inver Grove Heights, MN), and the Bjork-Shiley valves (Figure 11–2). Although each of these valves has its own distinctive design features, a typical tilting disk valve is comprised of a circular sewing ring and an eccentrically hinged or pivoting disk occluder. The pyrolytic carbon disk opens to show two unequal (major and minor) orifices during antegrade blood flow. With reversal of pressures, retrograde flow against the larger leading portion of the disk results in closure by rotation on its hinge or pivot.

The Medtronic-Hall valve, first used clinically in 1977, is the most commonly implanted tilting disk valve and is second only to the St. Jude bileaflet valve as the most frequently implanted mechanical prosthesis. The pyrolytic carbon disk has a small central orifice through which the eccentrically placed hinge mechanism passes. The disk and struts are enclosed within a titanium housing that is attached to the native valve annulus with a Teflon sewing ring. The maximum opening angle of an aortic Medtronic-Hall valve is 75°, whereas a mitral prosthesis generally opens no more than 70°.3,6 The fact that the occluder disk opens less than 90° generates resistance to forward flow and small eddies of stagnant flow proximally that predispose to thrombus development. The small hole in the occluder mechanism results in a characteristic central regurgitant jet that is visible with CFD.

The OmniScience valve also consists of a pyrolytic carbon disk suspended within a titanium housing; however, the sewing ring is comprised of polyester knit rather than of Teflon. Unlike the Medtronic-Hall valve, the OmniScience prosthesis does not have a central hinge; rather, the motion of the disk is restricted by a series of struts. Because there is no central hinge and prerequisite hole, the OmniScience valve does not have a central regurgitant jet when in the closed position. The maximum opening angle of this design is approximately 80°.2

The Bjork-Shiley prosthesis, the first successful low-profile tilting disk design (1969), is no longer available in the United States because of problems with strut fracture. However, more than 360,000 valves of this type were implanted, so one occasionally may encounter a patient with this design in situ. The disk of this design has a convex-concave design, with a maximum opening angle of 70°.2

Bileaflet

The major representatives of the bileaflet design are the St. Jude (St. Jude Medical, St. Paul, MN), the Carbomedics (Sulzer Carbomedics, Austin, TX) and the On-X (On-X LTI, Austin, TX) prostheses (Figure 11–3). The St. Jude bileaflet prosthesis is the most widely used mechanical valve in the world, with more than 600,000 implantations since its introduction in 1977. Each of these prostheses has two semicircular pyrolytic carbon occluders attached by small midline hinges to a support ring (pyrolytic carbon ring in St. Jude) and a Dacron sewing cuff. The manner in which the two occluder disks are hinged means they require no supporting struts, and therefore have an extremely favorable flow profile with lower transvalvular gradients as compared with the ball-in-cage and single tilting disk designs at similar annular diameters.3 With antegrade flow, both leaflets open to a maximum angle of 85°, resulting in two large, semicircular lateral orifices and a much smaller central rectangular orifice. With sufficient back pressure, the leaflets rotate on their hinges to close with an angle of approximately 25° to the plane of the supporting ring.2

The Carbomedics prosthesis is similar to the St. Jude in design; however, it has an adjustable titanium housing that can be rotated to position the leaflets in such a manner that they avoid contact with subvalvular tissue. The On-X valve is a newer valve made of pure pyrolitic carbon with a longer flow channel and an inlet that is flared outward to improve the flow dynamics through the valve. In addition, two leaflet guards are present to limit pannus encroachment onto the leaflets.7

Biologic Valves

Biologic tissue valves may be stented or stentless, depending on the presence of synthetic support structures. They may also be composed of either porcine or bovine leaflet tissue (xenografts), human cadaveric tissue (homografts), or native human tissue (autografts, as in the Ross procedure). Some biologic valves contain a mix of synthetic and natural biological materials (heterografts). Frequently, xenografts or homografts, especially for the aortic position, will include surrounding aortic tissue for improved natural structural support (composite grafts). The principal advantage of biologic over mechanical valves is related to the need for anticoagulation: mechanical valves require long-term anticoagulation, whereas biologic valves require only a short period (8 to 12 weeks) during endothelialization of the sewing ring.

Stented Biologic Prostheses

Stented biologic valves combine synthetic structural and supporting elements, with leaflets comprised of porcine valve leaflets (Hancock [Medtronic, Minneapolis, MN] and Carpentier-Edwards [Edwards Lifesciences, Irvine, CA] valves) or shaped pericardial tissue (Ionescu-Shiley [discontinued] and Carpentier-Edwards Perimount [Edwards Lifesciences, Irvine, CA] valves; Figure 11–4). The biologic elements of these valves are treated with glutaraldehyde to reduce antigenicity and increase tissue strength. However, this same process renders the tissue less pliable than native human valves and can promote calcification and degeneration in the long term.

The porcine valve leaflets in the Carpentier-Edwards prosthesis are mounted on a flexible Elgiloy frame, with stents manufactured from a single piece of wire attached to a Dacron sewing ring. The individual leaflets are mounted above the sewing ring, which allows for a larger orifice area and an improved hemodynamic profile as compared with valves with a similar ring size. However, the addition of the supporting struts (stents) does reduce the effective orifice area (EOA) as compared with a native human valve.8 The Hancock valve has many similarities to the Carpentier-Edwards valve on gross visual inspection. Nevertheless, each has a distinctive radiographic shape, with the Carpentier-Edwards valve appearing as a crown and the Hancock valve appearing as a circular ring because its stents are constructed from polypropylene, which is not radiopaque. During antegrade flow, the valve leaflets assume a somewhat irregular cone shape secondary to the slight restriction in opening caused by the valve stents. With pressure reversal, the leaflets coapt and commonly show a small central regurgitant jet, even in a normally functioning biologic valve.

Pericardial biologic valves (Ionescu-Shiley and Carpentier-Edwards Perimount) are not restricted by the physiologic size of the donor valve leaflets, thus allowing for the construction of larger valves. The Ionescu-Shiley valve was a low-profile prosthesis with three bovine pericardial cusps mounted to a Dacron-covered titanium frame using retention sutures. This prosthesis had difficulties with leaflet dehiscence and was discontinued after 10 years. The Perimount valve by Carpentier-Edwards has its pericardial leaflets mounted within the stent to maximize leaflet opening and to reduce abrasion between the leaflet and the stent.8

Stentless Biologic Prostheses

The development of stentless bioprostheses progressed with the desire to improve hemodynamics and long-term durability and preserve the benefits afforded by a tissue valve. The removal of the stent and sewing cuff permit the implantation of a relatively larger valve as compared with a stented biologic or mechanical valve of the same circumference. In addition, removal of the stent appears to significantly reduce stress at the base of the leaflets, which reduces calcification and slows valve degeneration. These valves are used only in the aortic position in which the patient's annulus and aortic root provide support and flexibility to the implanted materials. The typical stentless valve is comprised of an intact porcine aortic valve with an outer layer of polyester fabric to lend support and to facilitate implantation (Figure 11–5). Prior to the addition of the fabric, the valves are processed at very low fixation pressures to maintain collagen pliability. In addition, the Medtronic Freestyle valve is treated with amino-oleic acid to decrease calcium deposition. Several stentless valves are currently available, including the Toronto SPV (St. Jude Medical, St. Paul, MN), Medtronic Freestyle (Medtronic, Minneapolis, MN), and the CryoLife-O'Brien (CryoLife, Kennesaw, GA). These valves are technically more challenging to implant, and the exact surgical technique depends on how much of the recipient's aortic root and native sinus tissue are used in the process (full root technique with reimplantation of the coronary arteries, the root inclusion technique with preservation of the native coronary arteries, or a complete subcoronary or modified subcoronary technique).

The operative mortality rate for the stentless valves is slightly higher than that for other biologic valves (3% to 6%) partly due to the increased complexity of surgical technique. However, stentless valves have a very favorable 12-month complication rate (low rates of endocarditis, thrombosis, and hemorrhage) and survival rate of 91 ± 4% at 6 years. Currently, the engineered durability has held true, with no reported structural failures.9

Homografts and autografts

Homografts (human cadaveric tissue) are collected from the aortic and pulmonic positions within 24 hours of donor death and are treated with antibiotics before sterilization and preservation in liquid nitrogen at –196°C. These valves provide excellent hemodynamics, have low rates of thrombogenesis, are relatively resistant to infection, and are therefore ideal for use in the setting of acute native or prosthetic endocarditis.10,11 However, these cryopreserved valves are subject to accelerated degeneration as compared with native and mechanical valves. In addition, they require ongoing cryopreservation, have limited availability (most institutions carry a limited number of sizes, if at all), and require additional surgical time if the valve is not sized and thawed until the annulus is directly measured after arrest of the heart rather than relying on TEE valve sizing.12 The implantation of a homograft requires a highly skilled surgeon. Mitral homografts have been attempted in the past, but are no longer common.

Pulmonary autograft for aortic valve replacement (Ross procedure) uses the patient's native pulmonic valve and proximal main pulmonary artery to replace the diseased aortic valve and ascending aorta, with reimplantation of the native coronary arteries into the neo-ascending aorta. A stentless homograft (or other bioprosthesis) is usually placed in the pulmonic position so that anticoagulation is not required in the long term. Autografts are used most commonly in children, adolescents, adults with a longer than 20-year life expectancy, and those individuals in whom long-term anticoagulation is contraindicated or unwanted because of lifestyle factors.2,3 These valves have excellent hemodynamic profiles and are extremely durable with the “life expectancy” of a native valve. Limitations to its use include a technically more complex operation with the necessary replacement of two heart valves and the degeneration and/or obstruction of the new “pulmonic” valve and proximal pulmonary artery necessitating reoperation.

Transcatheter Aortic Valves

Endovascular transcatheter aortic valve implantation (TAVI) has recently become an option for those patients with calcific aortic stenosis who were previously deemed to be too high risk for a conventional valve replacement. The combination of advanced age with multiple comorbidities including renal, pulmonary, and cerebrovascular dysfunction increases the operative risk to over 25%.13 Following the first successful human implant in 2002, several different techniques have been developed (antegrade–transapical, and retrograde–transarterial) to more safely position and deploy the expandable bovine pericardial valves.

Currently there are two types of TAVI valves. The balloon expandable bovine pericardial “Sapien” valve (Figure 11–6) (Edwards Lifesciences, Irvine, California) is mounted onto a stainless steel stent and comes in two sizes (23 and 26 mm). The 23-mm valve is preferred for annular diameters of 18 to 21 mm, and the larger 26-mm valve for larger annular diameters up to 24 mm. The CoreValve (Medtronic, Minneapolis, MN) is also manufactured using bovine pericardium, but it is mounted on a self expanding Nitinol stent and is available in two sizes: a 26-mm valve for an annulus diameter of 20 to 23 mm, and a 29-mm valve for a 23- to 27-mm annulus. The valve chosen must be larger than the annulus diameter to reduce the risk of paravalvular leaks while allowing adequate anchoring of the valve system.

Proper positioning of these valves is critical. Incorrect positioning can lead to embolization into the left ventricular cavity or anywhere along the aorta, paravalvular leaks, coronary ostial obstruction, or interference with the mitral valve (native or prosthetic). Both fluoroscopy and TEE are utilized in many centers, with TEE becoming the preferred imaging modality because of its versatility. TEE allows for a full assessment of ventricular function, a complete evaluation of all valves, sizing of the annulus and other structures, measurement of gradients, and quantification of trans- or paravalvular regurgitant jets. Echocardiography can also be utilized to direct the advancement of guidewires and delivery devices prior to and during the placement and delivery of the valve.14

The evaluation of patients with prosthetic valves begins with an understanding of the patient's symptoms, which often provide clues to the expected pathology. It is also important to record the blood pressure, heart rate, and cardiac output (when available) since Doppler-derived gradients are dependent upon flow and diastolic filling time. Finally, the height, weight, and body surface should be noted in order to assess patient-prosthesis mismatch and cardiac chamber size.15

Evaluation of prosthetic heart valves using TEE can be technically challenging, but it provides a vitally important means to image and interrogate the normal and abnormal structure and function of an implanted valve prosthesis. Rather than being a “standalone” technique, TEE often complements methods such as fluoroscopy, magnetic resonance imaging, transthoracic echocardiography (TTE), and catheterization studies in the evaluation and routine follow-up of patients with prosthetic heart valves.1 TEE is minimally invasive, especially in the intraoperative setting, and provides superior 2D images as compared with traditional TTE because of the use of higher-frequency probes, the proximity of the ultrasound beam to the area of interest, and the lack of intervening structures to absorb or interfere with the beam. The development of real-time three-dimensional TEE (RT3DTEE) has also been shown to be beneficial in the assessment of indwelling prostheses.16,17

Synthetic materials (metals, plastics, fabrics, and pyrolytic carbon) on prosthetic valves present distinct challenges to a thorough echocardiographic examination. Interference with ultrasound transmission and reflection can produce several types of artifacts that make comprehensive assessment challenging. While mechanical valves are more prone to imaging artifacts, biologic valves also include synthetic components that can generate similar difficulties. Ultrasound artifacts are directly related to the reflective and absorptive properties of the prosthetic materials employed. In addition, the speed of sound through the synthetic materials may be faster or slower than that through human tissue, which can lead to alterations in the displayed size and location of the image. One must always keep these artifacts and imaging limitations in mind to avoid the misinterpretation of information that can result in false or missed diagnoses.

Reverberation artifact and acoustic shadowing (see Chapter 3) can conceal anatomic structures distal to the reflector and often render CFD evaluation impossible. The ability to place the transducer directly posterior to the left atrium with TEE allows for an unobstructed view of the mitral surface of the prosthesis and a comprehensive Doppler evaluation. From the midesophageal position, all shadowing and reverberation artifacts reside on the ventricular aspect of the valve. Similarly, the transgastric and deep transgastric views usually allow for the evaluation of the ventricular aspect of the valve, with the artifacts now located within the atria.

TEE imaging of prosthetic valves is not different from regular ultrasound imaging. The best 2D images are generated when the reflector is perpendicular to the ultrasound beam, whereas accurate Doppler studies require an angle of interrogation of less than 20° from parallel to the jet or flow. The measurement of transprosthetic gradients and flows using Doppler ultrasound follows the same limitations and caveats described in Chapters 7, 9, and 10, namely that it may be difficult to obtain the proper position or angle to locate the maximal jet or flow. The multiplane transducer frequently needs to be rotated to adjust the ultrasound beam so that imaging artifacts are minimized and proper identification of normal and pathologic structures can be elucidated. The acquisition of 3D images is described in Chapter 24.

Normal TEE Findings in Prosthetic Valves

While all prosthetic valves are obstructive by the nature of their design, each class and subclass of valve has unique echocardiographic structural characteristics that can be used to identify the specific type, and subsequently the proper function, of the valve. Regardless of the type of valve undergoing evaluation, a complete examination includes imaging the area in question with a variety of views to obtain an accurate assessment. Although usually not immediately available in the operating room setting, comparisons with previous studies can be invaluable in the evaluation of a prosthetic valve. Flow patterns and transvalvular gradients can change over time and new findings should be compared with previous measurements.18,19

Additionally, there are special considerations unique to the intraoperative setting. Loading conditions often fluctuate in the immediate post-CPB phase, making transvalvular gradients difficult to quantify. Moreover, frequent changes in inotrope levels and pacing strategies also result in changing hemodynamics. These must all be considered while interpreting the results of Doppler-based measurements.15

2D Imaging

The sewing ring (if any) should be examined for movement. A rocking motion is suggestive of valve dehiscence, loose or fractured sutures, or a perivalvular abscess. The periannular area is often somewhat obscured by calcification or artifacts from the valve itself. However, it is important to check for small periannular or perivalvular echolucent areas that may indicate fistula or abscess formation. Frequently, the aortic root can appear thickened as a result of hematoma or edema after surgery. This appearance may be prone to misinterpretation and should be distinguished from an abscess.15 Next, valve leaflet or occluding device motion is evaluated. Mechanical occluders should move rapidly and extend to the full limit of the design specifications. Restricted or incomplete opening is suggestive of primary valve dysfunction or of an obstructing material or object (suture material, perivalvular tissue, pannus, thrombus, endocarditis, vegetations, or masses). In addition, residual periannular materials may interfere with the proper closure of the occluders of mechanical prostheses resulting in significant regurgitation through the valve (Figure 11–7). The leaflets on a biologic valve should be thin and have the same appearance as native nondiseased leaflets. A close inspection for tears, perforations, thickening, and calcification should be carried out. Information from the rest of the TEE examination should be incorporated into the assessment of the prosthetic valve. Left atrial “smoke,” or thrombus, is suggestive of a functionally stenotic mitral prosthesis. Similarly, LV dilation or LV hypertrophy with new native mitral regurgitation could be suggestive of prosthetic aortic regurgitation or stenosis, respectively.

Epicardial Echocardiography

Infrequently an intraoperative TEE exam cannot be performed because either a TEE probe could not be inserted or there was a contraindication to probe insertion. In these circumstances, the evaluation of prosthetic valves can be performed using epicardial echocardiography (EE). The performance of a complete EE exam is discussed further in Chapter 20 and in the guidelines published by the American Society of Echocardiography and the Society of Cardiovascular Anesthesiologists.20 Although not specifically outlined in these guidelines, the majority of normal cardiac structures (including prosthetic materials) can be visualized using a high-frequency epicardial ultrasound probe.

3D Imaging

The sequential evaluation of prosthetic valves using 3D TEE is similar to that of the 2D examination (see Chapter 24). The evaluation of both the prosthesis and the peri-prosthetic structures using 3D echocardiography may facilitate interrogation of structures that are difficult and/or impossible with regular 2D imaging alone. The ability to take a large volume sample, and crop then rotate the image frequently provides improved visualization of structures opposite the transducer, small perivalvular leaks, leaflet/occluder problems, and thrombi/vegetations. Evaluation of both biologic and mechanical prostheses by 3D echocardiography is both feasible and accurate (96% correlation with surgical findings), especially in the presurgical assessment. However, 3D TEE is subject to the same limitations of 2D TEE with regard to ultrasound artifacts.16,17

Color-Flow and Spectral Doppler

A full color-flow and spectral Doppler examination should be carried out on all prosthetic valves. Care should be taken to properly align the transducer and use appropriate machine settings to ensure that the results are truly representative of the actual in vivo flows and gradients. Normally functioning tissue valves typically have no regurgitation. However, trivial to mild central regurgitant jets are visible about 10% of the time and are considered normal.3 Rarely, trace jets may originate from the commissures. New, significant regurgitant jets in tissue valves may signal leaflet degeneration. A moderate to severe regurgitant jet immediately after separation from cardiopulmonary bypass (CPB) may suggest a leaflet tethered by an annular suture (Figure 11–8) or an improperly sized, stentless prosthesis.

All mechanical valves, except for the Starr-Edwards, have normal “washing jets” that originate from between the occluder device and the support ring, which are designed to help reduce the incidence of thrombus formation on the underside of the prosthesis (Figure 11–9). These jets are considered normal if they are shorter than 2.5 cm and are trivial to mild in nature.1,19 An exception to this rule is the Medtronic-Hall single tilting disk valve in which the central regurgitant jet can be as long as 5 cm. The number of these jets is variable, depending on the angle of interrogation; however, one to two jets are most commonly seen, but up to five have been reported. Depending on the type of valve in question, up to a 10% to 15% regurgitant fraction is considered within the design limits. Moderate to severe regurgitant flow can be seen in mechanical valves with malfunctioning or restricted occluding devices. In addition, mechanical valves exhibit what is referred to as closure backflow. This is the small volume of retrograde flow that forces the occluding device(s) closed. It is considered a normal finding and appears as a very brief regurgitant flash immediately within the area of the valve. With mechanical valves, a few microbubbles (microcavitations) may also be seen in the LV. Mechanical and tissue prostheses can exhibit perivalvular leaks, which are suggestive of valve ring or annular dehiscence or endocarditis.

Spectral Doppler assessment of prosthetic valves is mandatory following implantation. Routine evaluation of prostheses should include an estimation of valve gradients using the Bernoulli equation and calculation of the effective orifice area (EOA) using the continuity equation:

where Q represents stroke volume through a reference site proximal or distal to the valve and VTI refers to the velocity-time integral across the valve. The label size of the prosthetic valve is not considered a valid substitute for the cross-sectional area of the annulus.15 Using the simplified Bernoulli equation (peak pressure gradient equals four times the square of the peak velocity: ΔP = 4V2) can yield high trans-prosthesis pressure gradients. It must be remembered that this equation does not account for the velocity proximal to the valve, which can be elevated, especially for prosthetic valves in the aortic position. Higher left ventricular outflow tract (LVOT) velocities should be accounted for by using the modified Bernoulli equation to obtain an accurate gradient [ΔP = 4(V22 – V12), where V1 is the peak LVOT velocity and V2 is the peak transvalvular velocity].15

The standard 2D, color-flow, and spectral Doppler findings for each type of prosthesis are described in detail in the following sections.

Mechanical Valves

Ball‐in-Cage

The silastic ball of the Starr-Edwards valve may demonstrate one or several intense echoes generated from its leading edge, which frequently obliterate or mask the other prosthetic components. It is uncommon to be able to visualize the U-shaped struts, whereas the sewing or support ring can almost always be seen. This valve is identified most easily by the detection of a large, rapidly moving, echo-dense structure that passes in and out of the imaging plane, casting a broad distal acoustic shadow. When color-flow Doppler is applied, highly turbulent blood flow is seen to originate from the lateral edges of the support ring as the blood circumnavigates the high-profile ball (Figure 11–10). Small retrograde eddies at the periphery also may be seen. With pressure equalization and valve closure, a trivial to small amount of central closure backflow (2 to 5 mL) can be appreciated with CFD. Once seated, the snug fit of the ball into the supporting ring limits further retrograde backflow. Spectral Doppler interrogation shows moderately high antegrade pressure gradients along the edges of the prosthesis as compared with tilting disk and bileaflet valves.5

Tilting Disk

All single tilting disk prostheses have an occluder with an asymmetrical opening pattern. Identification of particular subtypes can be extremely difficult when using 2D TEE because all have an identical pattern of motion and overall appearance. The echocardiographer must be methodical in attempts to image the prosthesis by using the long- and short-axis views with fine adjustments of the multiplane transducer to provide a satisfactory evaluation of occluder motion. M-mode evaluation may be helpful because the 2D frame rate is almost always too slow to accurately represent the pattern of occluder motion. In case of difficulties in identifying the actual number of disk occluders, it is helpful to know that tilting disks travel farther into the antegrade side than bileaflet occluders. Prostheses in the mitral position are frequently oriented with the major orifice facing the lateral wall. Tilting disk prostheses in the aortic position are placed in a manner such that the antegrade flow through the major orifice is directed toward the greater curvature of the aortic arch.

Color-flow Doppler allows differentiation between the Medtronic-Hall valve and the Bjork-Shiley and OmniScience valves. The Medtronic-Hall valve has been engineered with a central hinge and perforation in the occluder that results in a narrow central regurgitant jet (Figure 11–11) with the prosthesis in the closed position, whereas the other valves do not. With the valve in the open position, major and minor orifices often can be visualized, with the minor orifice possessing a more turbulent flow pattern. Spectral Doppler demonstrates a lower peak and mean gradient as compared with the ball-in-cage design because the disk occluder is less of an obstruction to antegrade flow.6 The average gradient for each valve type is dependent on the size and location of the prosthesis, and on the compliance and function of the left ventricle (see Appendix F and G for a more detailed listing). Spectral CWD should be directed through the lateral orifices of these valves.

Bileaflet

Two-dimensional evaluation of bileaflet prostheses (St. Jude, Carbomedics, On-X) is similar to that described for tilting disk prostheses. One must ensure that all available imaging planes are used to properly evaluate the prosthesis and keep the valve centered on the screen with the beam bisecting the support ring. When imaging in the long axis, rapid simultaneous oscillating hemi-disks with variably shaped, distal acoustic shadows are seen.12,18 For mitral valve replacement, surgeons generally position these valves in an “anatomic” orientation, with the occluder pivot points located where the anterolateral and posteromedial commissures previously resided in the mitral position. In the aortic position, one pivot point is usually located at the junction of the left and noncoronary sinuses of Valsalva with the other positioned opposite to it in the mid–right sinus. Alternatively, others attempt to position the valve so that flow from the outflow tract is evenly divided between the two major orifices.

When the ultrasound beam is perpendicular to the plane of the sewing ring, a distinctive V pattern is generated by the occluding disks on 2D imaging (Figure 11–12). It is from this orientation that color-flow and spectral analyses are best performed. Color flow demonstrates a reasonably laminar antegrade flow pattern through the two major orifices. Upon disk closure, two regurgitant jets that converge to form a “V” on the low-pressure side of the valve (see Figure 11–9) are commonly seen at the periphery between the occluders and the sewing ring. All regurgitant jets are generally short and are of a low velocity. The CWD beam should be directed along the major orifices toward the lateral edge of the prosthesis. Spectral Doppler consistently demonstrates that the bileaflet valves have the lowest gradient (flow velocities in the 2 m/s range) for any comparably sized mechanical valve. Three-dimensional imaging, especially in the mitral position, clearly identifies the two disks and enables assessment of disk motion with ease (Figure 11–13). The On-X valve has a unique echocardiographic appearance that can lead to misinterpretation of normal echo findings. The leaflet guards in this valve are commonly seen on a short-axis image (Figure 11–14) and should not be confused with the leaflets themselves. Furthermore, the profile of the inlet flare can cause extensive shadowing in the LVOT, and the diastolic washing jets are directed away from the central axis of the valve and towards the walls of the LVOT (see Figure 11–14).7 In contrast, the washing jets in the St. Jude bileaflet valve are typically directed inward, towards the central axis of the valve.

Biologic Valves

Stented Valves

Hancock and Carpentier-Edwards valves appear similar on 2D imaging. At each commissure, a strut is usually visible, appearing as a narrow echogenic structure. All three struts are usually visible within the same view (Figure 11–15). A strong reflective signal is obtained from the support ring when the beam is directed through the annulus. In long-axis views, the leaflets appear similar to native leaflets. The struts now appear as linear reflectors arising from the highly reflective support ring. In general, one can view only two struts at a time due to the finite thickness of the ultrasound beam. Acoustic shadowing and reverberation artifact from the support ring interferes with the acquisition of usable information from the area immediately distal to the ring.

Color-flow Doppler usually demonstrates nonturbulent antegrade flow, with little to no central regurgitation after valve closure. Different interrogation angles may be necessary to remove imaging artifacts out of the area of interest. These valves generally have a low to moderate peak and mean gradients, which are inversely related to prosthesis size. Three-dimensional imaging again provides superior imaging in the mitral position (Figure 11–16).

Stentless Valves

These valves are usually difficult to distinguish from native valves, except for an increase in echo signal from the annulus and around the base of the leaflets. Other 2D findings depend on the surgical technique used to implant the valve. When the root inclusion technique is used, the ascending aorta often can appear to be thicker in the area of the aortic root and proximal ascending aorta because the wall now contains layers from the prosthesis and the native aorta (a tube within a tube; Figure 11–17). In the immediate postoperative period, blood, fluid, or thrombus may occupy the potential space between these two layers, and very rarely has caused problems with valve function. This potential space usually disappears within the first few postoperative months.3,9 Color-flow and spectral Doppler signals from stentless prostheses are similar to those from native valves.6

Homografts and Autografts

These valves can also be very difficult to distinguish from native valves. A small to moderate increase in echogenicity is frequently noted at the annulus from suture material and accelerated calcification. Doppler signals are comparable to stentless prostheses and should be similar to that of native valves.

Percutaneous Aortic Valves

Prior to the placement of an expandable biologic valve in the aortic position, a complete echocardiographic exam is mandatory. Assessment of LV and valvular function should be comprehensive, and LVOT/annular dimensions, aortic valve areas, and gradients should be recorded. During the positioning of the valve, TEE can be used to monitor advancement of guidewires and the delivery device, evaluate the location of the prosthesis, determine complications (tamponade, dissection, device malposition) and observe the LV for dilation/overload following balloon valvuloplasty of the native aortic valve.

Following the deployment of the valve, both long- and short-axis midesophageal aortic valve views should be obtained to assess the final position of the device. The assessment of these valves (Sapien and CoreValve) is similar to that of biologic valves with respect to the leaflets. Significant artifact from the metal stents in which they are housed may require multiple angles and views for a thorough assessment (Figure 11–18). Other important features of a complete evaluation include assessment of leaflet function using 2D or 3D imaging, measurement of transvalvular gradients using CWD, and an evaluation of any transvalvular (within the stent) or paravalvular (around the stent) leaks.

The various techniques for assessment of stenotic or regurgitant native valves described in Chapters 7, 9, and 10 can be used or modified to evaluate prosthetic valves. The physical principles of the continuity equation, modified Bernoulli equation, mean gradients, velocity-time integral (VTI) ratios, pressure half-times (PHTs), and specific flow characteristics are also applicable to implanted valves. Readers are encouraged to refer to the aforementioned chapters for a complete discussion on the qualitative and quantitative assessments of regurgitant or stenotic lesions.

Mitral Position

TEE is particularly suited for the complete imaging and interrogation of prostheses in the mitral position. Midesophageal views allow for unobstructed imaging of the left atrial side for assessment of regurgitation, paravalvular leaks, and mechanical occluder motion.21 Transgastric and deep transgastric views generally afford views of the ventricular aspects of prostheses, thus enabling confirmation of previously gleaned information.

EOA is derived for a prosthetic mitral valve as the stroke volume through the valve divided by the VTI of the transmitral flow:

where, in the absence of significant mitral or aortic regurgitation, stroke volume can be measured at the LVOT. EOA can be underestimated in bileaflet mechanical valves since the peak velocity measured with CWD includes the higher velocity seen in the smaller central orifice.

Use of the pressure half-time constant of 220 milliseconds in the assessment of a stenotic prosthetic valve overestimates the EOA because this number was derived from studies in native stenotic valves and not validated in prosthetic valves.22 However, a single pressure half-time value greater than 200 milliseconds or a consistent value in excess of 130 milliseconds may indicate pathologic obstruction.

When interpreting velocities across prosthetic valves in the mitral position, one must differentiate an increase in transvalvular velocity due to mitral regurgitation from prosthetic valve stenosis and high output states. The absence of a regurgitant jet or presence of appropriately moving leaflets would obviously suggest a high output state. Another measurement that has been proposed to help in this differentiation is the ratio of VTI across the mitral valve to the VTI across the LVOT (VTIMV/VTILVOT). A normal ratio suggests a high output state since the velocity is increased across the MV and the LVOT, while a high ratio indicates either prosthetic mitral stenosis or regurgitation. A value greater than 2.2 should prompt the search for pathology.15 Significant prosthetic mitral stenosis is suggested by (1) peak transmitral velocity of 2.5 m/s or greater, (2) mean gradient greater than 10 mm Hg, (3) VTIMV/VTILVOT greater than 2.5, (4) EOA less than 1 cm2, and (5) PHT greater than 200 milliseconds.15

In the absence of significant mitral regurgitation, the continuity equation is a valid and better method than using pressure half-time for determining the area of the mitral prosthesis. This may be explained by the fact that most prostheses, with the exception of the annuloplasty rings, are circular in nature. Thus, the assumption of mitral annular circularity in the calculation of the continuity equation is satisfied. Comparisons with previous studies and knowledge of the normal parameters are extremely important in distinguishing a normal from a dysfunctional valve. Indicators of severe mitral prosthetic regurgitation include (1) increased inflow velocities with normal pressure half-time, (2) dense regurgitant CWD signal, (3) pulmonary vein systolic flow reversal, (4) regurgitant jet greater than 40% of the left atrial area or a jet area greater than 8 cm2, (5) regurgitant volume 60 mL or greater, (6) regurgitant fraction 50% or greater, (7) effective regurgitant orifice 0.5 cm2 or greater, and (8) vena contracta greater than 0.6 cm.2,6,15,23 Since regurgitant lesions may involve a dehiscence of the valve suture line, it is imperative to examine the valve from all possible angles with 2D and CFD to determine the origin of the jet.

Aortic Position

A full and accurate evaluation of aortic prostheses usually can be obtained with TEE. Proper interrogation using 2D, 3D, color-flow, and spectral Doppler sometimes can be challenging, even for experienced echocardiographers, because of artifact interference and image acquisition. In the midesophageal planes, the sewing or support ring and/or stents from prosthetic valves are interposed between the transducer and the valve orifice, resulting in unavoidable artifacts in the area of interest. When attempting to evaluate the prosthesis for regurgitation or paravalvular leaks, or for spectral Doppler interrogation, the transgastric long-axis (110° to 130°) and deep transgastric long-axis views are often very helpful. These views remove the echo artifacts out of the region of the LVOT and provide for a more parallel alignment between the CWD or PWD beam and transvalvular flow. However, this places the sample volume in the far field, which predisposes to aliasing and sampling errors. Small manipulations of the probe tip and fine adjustments of the multiplane transducer in all views generally allows for visualization of most, if not all, relevant structures.

Prosthetic aortic valve replacement requires measurement of the aortic annulus. The echocardiographer first obtains a midesophageal long-axis view of the aortic valve and LVOT with the depth setting set so that the structures in question fill the screen. The image is frozen and the frames are advanced sequentially until maximal valve opening is visualized (usually mid-systole on electrocardiography), and the calipers are placed at the junction of the aortic valve leaflet and the LVOT on each side. Several measurements should be made to ensure an accurate assessment; the LVOT is a three-dimensional structure shaped like a cylinder and care should be taken that the measurement is taken at the widest part of the cylinder. When a stentless biologic valve is to be placed, the diameter of the sinotubular junction (STJ) also needs to be carefully measured (Figure 11–19).5 This usually can be accomplished in the same view used for the measurement of the aortic annulus. The STJ diameter is generally about 80% of the annular diameter. If the STJ is dilated, the stentless valve leaflets will have a tendency to separate centrally during diastole secondary to a lack of lateral support from the dilated proximal ascending aorta. If the STJ or ascending aorta is significantly dilated, the surgeon should consider a composite prosthesis that incorporates the ascending aorta. The Ross procedure deserves special mention because the TEE-derived measurements of the aortic annulus, pulmonic annulus, and STJ may determine whether the surgery actually proceeds or whether an alternative prosthetic valve type is selected. If there is a significant mismatch between the pulmonic valve (which is to become the neo-aortic valve) and the aortic annulus or STJ, the valve likely will be incompetent because of distortion of the valve and its leaflets. Post-bypass regional wall motion abnormalities may also be present in these patients, especially in the basal anteroseptal region, due to compromise of the septal perforator branch as it arises from the left ascending descending artery (LAD) in close proximity to the pulmonic valve.

Use of the continuity equation in aortic prostheses has been validated and can be useful to measure the EOA. Applied to the aortic position, the equation is expressed as:

An EOA of less than 0.8 cm2 usually suggests stenosis. However, this value may be considered normal for some mechanical valves. Bileaflet valves usually have higher velocities across their central orifice compared to the larger lateral orifices. Since CWD does not distinguish between the two velocities, and measures the highest values, the calculated peak gradients are often high, leading to an underestimation of EOA.15 This is especially true in high flow situations and with smaller valves. Many centers report mean valve gradients rather than peak gradients because there is less variability over time. The phenomenon of pressure recovery (reconversion of kinetic energy not completely dissipated as turbulence back to pressure energy distal to a stenosis, resulting in decrease in the pressure gradient) should also be considered in patients with a small (<3 cm) aorta or a small (<19 mm) bileaflet tilting disk prosthesis.

The Doppler velocity index (DVI), a ratio of the peak velocity in the outflow tract to that through the prosthesis, is a useful tool in the assessment of prosthetic aortic valves:

where VAVR is velocity through the aortic valve prosthesis.

It may be particularly useful during TEE assessment since it does not need to account for LVOT diameter, which can be difficult to measure with a prosthetic device in situ. In addition, since the implanted valve size is usually proportional to the LVOT size at the time of surgery, DVI is less dependent on valve size. Due to higher velocities across the prosthesis, the DVI is always less than one, but a DVI value less than 0.25 suggests severe prosthetic stenosis.15

The contour of the jet tracing can provide valuable clues to the presence of underlying stenosis. With progressive stenosis, the jet contour changes from a triangular shape with a short acceleration time (AT, time from start of flow to peak velocity) to a smoother contour with a prolonged AT. An AT beyond 100 milliseconds indicates stenotic flow, while an AT-to-ejection time ratio greater than 0.4 also indicates obstruction.15 In summary, significant aortic prosthetic stenosis is suggested by (1) peak transvalvular velocity greater than 4 m/s; (2) mean gradient greater than 35 mm Hg; (3) DVI less than 0.25; (4) EOA less than 0.8 cm2; (5) rounded, symmetrical contour of the jet velocity through the valve; and (6) AT greater than 100 milliseconds.

Pathologic regurgitation in aortic prostheses may be more difficult to quantify than native valves. Acoustic shadows from the prosthetic valve may render quantification attempts more difficult. However, criteria used for native valves may be reasonably applied to prostheses, especially spectral Doppler parameters such as the pressure half-time of the regurgitant jet (see Chapter 9).15 Additionally, the qualitative evaluation of the nature of regurgitant jets with CFD and 2D examination is more valuable. Identification of the origin of the regurgitant jet with CFD may help determine pathology such as pannus, endocarditis, or prosthetic valve leaflet dysfunction.

Tricuspid and Pulmonic

Placement of prosthetic valves on the right side of the heart is much less common than on the left. Patients are often able to tolerate significant regurgitation without major signs and symptoms because pressures are lower in the right heart. Replacement of the pulmonic valve is performed most commonly secondary to congenital malformations of the valve or outflow tract (usually done in infancy) or as part of the Ross procedure for aortic valve pathology. Frequently, tricuspid pathology is amenable to valve repair with sutures alone (De Vega annuloplasty) or with a ring. When valve replacement is necessary, biologic valves are almost always placed because of the higher risk of thrombosis with mechanical valves and the fact that low right-side pressures cannot reliably open and close the occluders of mechanical valves.2,3 After tricuspid replacement, it is not uncommon to see a small paravalvular jet emanating from the septal portion of the ring because surgeons frequently leave a small gap in the fixation sutures so as not to interfere with the bundle of His and thereby reduce the rate of iatrogenic third-degree heart block. Pathologic regurgitation or stenosis of the tricuspid prosthesis is difficult to quantify but conventional techniques as applied to native valves or prosthetic mitral valves may be used with reasonable confidence. It is important to use an averaged measurement of at least five beats for the tricuspid prosthesis. A pressure half time of 230 milliseconds or greater, a peak E velocity of greater than 1.7 m/s, and a mean gradient 6 mm Hg or greater suggest prosthetic tricuspid stenosis.15 Other than CFD patterns indicating severe tricuspid regurgitation (TR) (jet area >10 cm2, vena contracta [VC] >0.7 cm), secondary signs suggestive of prosthetic regurgitation include a dilated right atrium and holosystolic flow reversal in the hepatic veins. Calculation of the EOA should not be performed using the pressure half-time algorithm applied to the mitral valve since this technique has not been validated in the tricuspid position. However, estimation of EOA using a combination of Doppler techniques such as trans-tricuspid VTI and LVOT stroke volume is reasonable as long as significant aortic and tricuspid regurgitation do not coexist.

The pulmonic valve position can be difficult to examine even in the absence of a prosthetic valve because it is the valve farthest from the TEE probe. Nevertheless, one usually is able to identify the type of prosthetic valve and any associated pathology. The best views for assessing the pulmonic valve include the upper esophageal aortic arch short-axis view (approximately 70°) with the depth adjusted to 10 to 12 cm, which shows the valve and the main pulmonary artery trunk, and allows for measurement of transvalvular gradients and color-flow assessment; and the right ventricular inflow–outflow view at approximately 60°, which provides satisfactory windows for valve identification and qualitative assessment of regurgitant jets. Color-flow Doppler will usually indicate turbulent flows in prosthesis regurgitation or stenosis, while 2D examination will reveal leaflet thickening or obstructive lesions. Peak velocities greater than 3 m/s for prosthetic valves or 2 m/s for homografts are considered suspicious for prosthetic stenosis, while a prosthetic pulmonary valve regurgitant jet that exceeds 50% of the pulmonary annulus diameter is suggestive of severe regurgitation.15 There are limited data validating these techniques with TEE, and careful inspection of the valve using all available modalities is mandatory for a complete assessment.

It is not uncommon for surgeons to have to replace more than one valve, at the same time or sequentially. Diseases such as rheumatic fever and myxomatous degeneration frequently affect more than one valve in the same individual. Surgical implantation of multiple prosthetic valves is an added technical challenge. Combined aortic and mitral valve replacement carries a 70% higher risk and poorer survival rate than replacement of either valve alone. The Society of Thoracic Surgeons national database committee (STS National Database Fall 2008—Executive Summary) gives an operative mortality of 9.6% for all multiple valve surgeries, compared with 3.2% and 5.7% for aortic and mitral valves alone, respectively. Long-term survival depends on preoperative functional status.

TEE evaluation of prosthetic valve structure and function with multiple valves in situ also provides a challenge to the echocardiographer. Imaging artifacts from one valve frequently obscure the examination of the other, sometimes making a complete examination impossible.

See Chapter 8 for a full discussion of mitral valve repair. One method of repairing a leaking mitral valve uses prosthetic materials to reestablish anatomic coaptation of the anterior and posterior mitral leaflets. Rings also are used frequently to rectify mitral regurgitation secondary to annular dilation (cardiomyopathy or end-stage coronary heart failure) or leaflet prolapse. Different prosthetic rings are available with different amounts of flexibility and overall shapes. Many are complete D-shaped rings, whereas others are shaped like a C or even attempt to reproduce the saddle-shape of the mitral annulus. Surgical preference usually dictates the type of prosthesis placed. The ring itself is visible around the periphery of the native mitral valve, as are the sutures used to secure it in place (Figure 11–20). Prosthetic mitral rings do not require long-term anticoagulation.

Patient-Prosthesis Mismatch

Patient-prosthesis mismatch (PPM) is best characterized using EOA indexed to body surface area. PPM in the aortic position is considered to be hemodynamically insignificant if EOA is greater than 0.85 cm2/m2, moderately significant at 0.65 to 0.85 cm2/m2, and severe at less than 0.65 cm2/m2. In the mitral position, EOA should be greater than 1.2 cm2/m2 to avoid high postoperative gradients and persistent pulmonary hypertension. PPM has also been associated with reduced short- and long-term survival.15

Endocarditis

Bacterial infections of implanted prosthetic materials are problematic regardless of their location, and infection of prosthetic heart valves can lead to significant mortality, with reported rates as high as 70% for acute (<6 months of implantation) and 45% in late (>6 months) endocarditis.10,11 Prosthetic valves are at an increased risk for endocarditis for two reasons: abnormal flow patterns and foreign material. The annual rate for late endocarditis is approximately 0.5%, with no significant difference between mechanical and stented biologic valves.10,11 Compared with native valves, prosthetic valves are more likely to have ring abscesses, conduction abnormalities, and fistulae, and have a poorer prognosis. Endocarditis has two common appearances on mechanical valves: vegetations on leaflets, occluders, or support material; and echolucent ring abscesses with or without fistulae. Tissue valves most frequently develop new stenotic or regurgitant lesions but also can develop vegetations (Figure 11–21) and/or abscesses. New or worsening periprosthetic leaks are highly suggestive of endocarditis and are a particularly ominous sign.

TEE is vastly superior to TTE for the evaluation and diagnosis of prosthetic endocarditis, with significantly greater sensitivity and specificity (TTE sensitivity is 60% to 80% and specificity is 98%; TEE sensitivities are 100% for native valves and 86% to 94% for prosthetic valves, with a specificity of 88% to 100%).11 TEE better characterizes vegetations, abscesses (Figure 11–22), and fistulae, thereby providing important diagnostic information related to medical versus surgical intervention. The recent introduction of real-time three-dimensional echocardiography has shown significant promise in imaging infected prosthetic valves because of the ability to crop and rotate a volumetric sample of echocardiographic information.24 This can allow the operator to generate views that may not be feasible using conventional 2D imaging. Endocarditis should be suspected in all patients with prosthetic valves who develop septicemia. Perivalvular abscess formation should be considered in those who do not respond to aggressive antibiotic therapy. Other clinical indicators prompting a thorough TEE “hunt” for an abscess include new conduction abnormalities and worsening heart failure (possible fistula formation). Findings commonly associated with abscess formation on a 2D study include dehiscence manifested as a rocking of the valve (Figure 11–23), perivalvular lucency, and periaortic root thickening.4,10,11 Abnormal color-flow patterns can be suggestive of fistulas and blood flow into abscess cavities. Abscess ruptures into the LVOT, right atrium, left atrium, ascending aorta, pericardial space, and main pulmonary artery have been described in the literature.

Thrombosis and Hemorrhage

Thrombosis and hemorrhage account for more than 50% of all reported complications in left-side biologic valves and nearly 75% of all mechanical prostheses.2 Maintenance of therapeutic anticoagulation levels continues to be one of the most daunting challenges related to valve replacement. Even “perfect” anticoagulation levels do not guarantee freedom from these serious and potentially deadly complications. The early detection of periprosthetic thrombosis has been improved significantly with TEE, which can help guide management, be it medical (thrombolytic therapy) or surgical. However, previously described imaging artifacts and limitations can interfere with the proper diagnosis.25 Furthermore, differentiating a thrombus from a pannus can be difficult but is often critical to directing therapy. In general, pannus formation is more common in the aortic position and is characterized by a small, dense mass, whereas thrombi are larger, have a soft ultrasound density similar to myocardium, and are more likely to be associated with abnormal prosthetic valve motion.15

All mechanical valves carry approximately the same incidence of thrombosis and require long-term anticoagulation and/or aspirin administration to help prevent clot formation (the greatest incidence is during the first postoperative year). Even with therapeutic levels of anticoagulation, the rate of thrombosis is 0.6% to 1.8% per patient-year for bileaflet valves.5 The risk of spontaneous thromboembolism is three to six times higher if anticoagulation is not administered or is sub-therapeutic. When placed in the tricuspid position, thrombosis of a mechanical valve is seen in more than 20% of patients because of the lower pressures in the right heart. Thrombus adherent to prosthetic materials usually can be visualized, but sometimes the only indication of thrombosis is inappropriate occluder motion with failure to completely open or close (Figure 11–24). Faulty occluder motion can be visualized directly with 2D echo or 3D echo (particularly in the mitral position), or be surmised by new obstructive or regurgitant lesions on color-flow and/or spectral Doppler.

Thromboembolism

Embolic events in patients with mechanical valves have been reduced to 1% to 4% per patient-year with strict anticoagulation protocols. However, the risk of a serious bleed while on anticoagulants is 1% to 5% per patient-year.26

Fibrin Strands

Occasionally, thin filamentous strands can appear to be attached to the atrial side of mitral, and the ventricular side of aortic, mechanical prostheses (and occasionally tissue valves). These strands are comprised of fibrin and are usually a few millimeters in length. Their motion is unrelated to that of the valve itself, and they can be distinguished from vegetations and thrombi by their movement in and out of the imaging plane. Several studies have suggested higher rates of systemic embolization when these strands are present.2

Primary Prosthesis Failure

Fortunately, a primary failure of mechanical prosthetic valves is a rare occurrence. Case reports of occluder device and strut embolization have been published for most available valves. An uncommon complication with the Starr-Edwards valve was “ball variance,” in which small cracks would develop in the silastic occluder ball, leading to thrombosis within the stellite cage and subsequent valve obstruction.

Prosthesis failure in tissue valves usually is related to calcific degeneration of the leaflets themselves, leading to restricted cusp motion (stenosis) and, more commonly, regurgitation secondary to tears and malocclusion. All non-autograft tissue valves will eventually degenerate to the point of requiring replacement in the long term. However, the newer stentless biologic valves appear to have a much slower degeneration rate.

Paravalvular Regurgitation

Paravalvular regurgitation appears as one or more regurgitant jets that originate from outside the prosthetic valve annulus (Figure 11–25). Mechanical failures, suture fracture, inadequate approximation and fixation of the prosthetic valve annulus, valve dehiscence, a heavily calcified annulus, and endocarditis may result in paravalvular leaks. In addition to the CFD findings, valves with large paravalvular leaks often appear to have a “rocking” motion of the valve sewing ring on the 2D examination. Identification of one finding should lead to a search for the other.

LVOT Obstruction

Although quite rare, obstruction of the LVOT has been described after implantation of prosthetic mitral valves or mitral annular rings.21 Obstruction can be caused by residual submitral valvular tissue (leaflets or chordae), strut or tissue leaflets in the case of stented bioprostheses, or the native leaflets themselves in the case of mitral valve ring annuloplasty. The presence of LVOT obstruction should be considered in all cases, especially in situations of poor hemodynamics after cessation of CPB in a previously normal ventricle (see Chapters 8 and 14). This is most easily assessed from the transgastric long-axis or deep transgastric long-axis views. Residual subvalvular tissues also may interfere with the valve function, resulting in mitral regurgitation rather than outflow tract obstruction.27

Obstruction Secondary to Tissue Adhesives or Glues

The last several years have seen a dramatic increase in the use of a variety of sealants, glues, and other materials that are either directly applied to or sprayed on the surgical field for the purpose of controlling bleeding at suture lines or repairs of cardiac tissues. Unfortunately, from time to time some of these materials can make their way into a cardiac chamber or vessel and pose the risk of embolization or interference with the proper function of native or prosthetic valves.28 A thorough examination following the separation from CPB should demonstrate the presence of foreign materials.

The evaluation of prosthetic valves is a necessary tool in the armamentarium of a transesophageal echocardiographer. Whether evaluating a new valve immediately after CPB or assessing a prosthesis that has been in situ longer than 10 years, echocardiography provides critical information that allows clinical decisions to be made solely on the basis of echo findings. The evaluation of prosthetic valves does not stop at successfully identifying the type and location of the valve. A comprehensive interrogation by using all modalities available should be carried out so that significant pathology is not overlooked.