In developing countries, rheumatic heart disease remains the chief cause of acquired cardiac disease during the first five decades of life.1 In industrial countries, rheumatic fever and its only sequel, rheumatic heart disease, have become relatively rare because of improved living conditions and effective programs of primary and secondary prevention with penicillin. Significant outbreaks of rheumatic fever, however, occurred in the United States of America during the last two decades of the 20th century,2–13 and sporadic cases continue to be encountered across the nation.
A resurgence of rheumatic fever in Utah, which began in 1985,2 has persisted to the present,14 and has coincided in time with the development of echocardiography. Echocardiography, which is now firmly established as an integral part of the practice of cardiology, is the only new test in the last forty years that aids the clinician in the diagnosis and management of rheumatic fever and rheumatic heart disease. Echocardiography, coupled with Doppler interrogation, is ideally suited for the evaluation of rheumatic heart disease, and has given new insight into the pathogenesis of acute rheumatic carditis.
In endemic regions, cardiologists and cardiac surgeons have gained a new understanding of rheumatic mitral regurgitation.15, 16 This new concept, however, is not discussed in many current paediatric and medical textbooks, or in other authoritative publications.17, 18 Our goal in compiling this review is to present this newer understanding, and to describe how past clinical experience and autopsy findings, as well as recent studies of the function of the mitral valve during life and in the experimental setting, support this new concept. Combining past and present experience into a coherent, unified concept involves a “back to the future” scenario.
Pre-echocardiographic concept
A discussion of rheumatic fever in the most recently published textbooks varies little from what was published in textbooks thirty years ago. Both then and now, chapters begin the discussion of carditis, with a statement that rheumatic carditis is a pancarditis, in other words the disease involves the endocardium, the myocardium and the pericardium. While such a statement is not completely erroneous, it is simplistic, and lacks a specificity we currently demand in the description of any other form of cardiac disease.
The role of the pericardium
In reviewing the involvement of the three layers of the heart, the role of the pericardium merits only a brief discussion, since rheumatic pericarditis never occurs in the absence of significant mitral valvar involvement.19 The full blown clinical picture of chest pain, friction rub, and ST-T wave changes on the electrocardiogram with or without radiographic cardiomegaly, is present in less than one-tenth of patients with rheumatic fever.19 Cardiac tamponade is exceedingly rare. We could find only a single case report.20 Unlike other forms of pericarditis,21–23 rheumatic pericardial inflammation almost never results in constrictive pericarditis.
The role of the myocardium
Myocarditis was thought to be responsible for most of the symptoms, and certainly was held to be the cause of heart failure and death during the acute episode of rheumatic carditis. This concept was endorsed by our most respected authorities, including T. Duckett Jones, Milton Markowitz, Gene Stollerman, Benedict Massell and Jesse Edwards. Although the current buzzword of “evidence-based” was never used in their consideration of acute rheumatic carditis, the evidence from their clinical experience and autopsy findings led to their conclusions. Clinically altered heart tones, elevated sleeping pulse rate, and cardiomegaly seen on a chest film, were considered to be due to myocardial inflammation and resultant chamber enlargement and ventricular dysfunction.24, 25
Convincing evidence of myocardial involvement came from autopsy findings in children and young adults who died during an acute episode. The left atrium and left ventricle were uniformly massively dilated, which could not be explained by minimal deformity of the mitral valve.26, 27 Microscopic examination showed inflammatory cells along with the pathognomic Aschoff bodies in the myocardium.28 Some investigators proposed that the Aschoff bodies were necrotic myocardial cells.29
Early immunological evidence also supported the concept of myocardial involvement. In 1964, cross-reactive antistreptococcal antibodies were found to bind the sarcolemma of the myocyte.30 Additional immunological studies stemming from that early experience continued to support significant myocardial involvement, with a demonstration of cross-reactive antistreptococcal antibodies to myosin31, 32 and molecular mimicry between segments of group A M-protein and myosin33, 34 and tropomyosin.35 Thus, there was strong evidence from a clinical standpoint, autopsy findings, and immunological data that the myocardium was significantly involved.
There was, however, strong clinical evidence 25 years ago, long before the widespread use of echocardiography, that childhood rheumatic carditis was essentially a valvar disease, and that the myocardium did not have a dominant role. Most notably, the right-sided chambers were not dilated as should have been encountered with a global myocarditis.36, 37 In addition, the clinical presentation of non-rheumatic myocarditis was distinctly different from rheumatic carditis. Heart failure without significant mitral regurgitation was the usual presentation, and tachyarrhythmias and ventricular ectopy were common.38
Additional evidence against the dominant role of myocarditis comes from the success of valvar replacement. In 1974, Strauss et al.39 reported that four boys suffering from medically intractable heart failure from rheumatic carditis recovered following prosthetic replacement of the mitral valve. This was a remarkable surprise to everyone except, perhaps, the authors. Yet, this landmark event had relatively little impact in changing the thinking that myocarditis played a dominant role in rheumatic carditis. In 1979, Lewis et al.40 using M-mode echocardiography, detected no evidence of myocardial dysfunction in an 18-year-old girl in intractable heart failure with marked aortic and mitral regurgitation. This early M-mode echocardiographic demonstration gave them courage to have the patient undergo double valvar replacement. Subsequent to the recovery of the patient, this group replaced the mitral valves in four additional patients suffering from intractable failure during acute episodes of rheumatic carditis. All recovered.
Currently, myocytic injury from whatever cause, be it ischaemia,41, 42 infection,43 trauma,44 toxic agents,45 or cardiopulmonary bypass,44 is confirmed clinically by echocardiographic demonstration of decreased ventricular function, and by the finding of elevated levels of troponin in the serum. Neither is present during acute rheumatic carditis.
M-mode evaluation, along with cross-sectional and Doppler echocardiographic studies, have also shown that impaired myocardial contractility is not encountered during the initial episode or episodes of rheumatic carditis.15, 46, 47 Conversely, normal or increased left ventricular ejection fraction and percent shortening are typically found. Decreased impedence to left ventricular ejection from mitral regurgitation could affect these determinations, but after the surgical restoration of mitral valvar competence, left ventricular function remains normal. Decreased ventricular function is encountered only in patients with long-standing volume overload due to significant mitral regurgitation, and/or aortic regurgitation.48–50
As already discussed above, in contrast to other myocarditities, and despite reports of antibodies to myosin and tropomysin, rheumatic carditis is not associated with elevated levels of troponin in the serum . The initial report, and the largest clinical experience determining the levels of troponin in acute rheumatic carditis, comes from Kamblock et al. working in Tahiti.51, 52 Additional reports from the United States of America53, 54 and Turkey55 confirm the Tahitian experience. This data provides further evidence that the primary site of rheumatic myocarditis is not the myocyte.
Cunningham56 has demonstrated molecular mimicry between cardiac myosin and streptococcal M-proteins. In addition, she found that Lewis rats immunized with streptococcal M-protein or cardiac myosin develop valvar lesions that mimic rheumatic valvar disease.56 Despite these findings, from a purely clinical standpoint it is difficult to understand how antibodies to cardiac myosin capable of causing cross-reactive valvar disease do not cause sufficient myocytic injury that can be recognized by either altered ventricular function or elevated levels of troponin.
Role of the endocardium
Historically, there was clinical, autopsy, and immunological support for the endocardial component of rheumatic pancarditis. The endocardial component includes “valvitis”, since the endocardium extends over the valvar leaflets. Autopsy findings supported this concept, since the leaflets were oedematous, contained extensive cellular infiltrates, and presented with fibrinous vegetations on the atrial side of the coapting surfaces of the mitral valve, and on the ventricular aspect of the aortic valvar leaflets. In addition, antibodies to group A streptococcus carbohydrate are elevated in patients with rheumatic mitral valvar disease.57 These antibodies disappear when the diseased valve is replaced.58
The term endocarditic valvitis, however, is somewhat misleading, since the leaflets of the mitral and aortic valves are much more than mere extensions of the endocardium. In the past, all components of the mitral valve were incompletely considered. The changes that occur during an acute episode of rheumatic activity are quite specific, and involve much more than the endocardium of the leaflets, since the entire mitral valvar apparatus is involved. The mitral annular dilation seen with rheumatic carditis was thought to occur secondary to left ventricular dilation from mycarditis. The cordal elongation of the aortic leaflet of the mitral valve, when noted, was also thought to be secondary to myocarditis.23
Normal mitral valvar function
Understanding how all components of the mitral valvar apparatus contribute to normal function is necessary to understand how the valve becomes incompetent during the acute rheumatic episode or episodes, as well as the changes that occur after recovery. While the papillary muscles, the left atrium, the left ventricle, and the systemic vascular bed all contribute to the successful function of the mitral valve, we shall limit our discussion to the so-called “intrinsic” components of the mitral valvar apparatus, that is, the leaflets, the annulus, and the tendinous cords.
The two leaflets of the mitral valve are the largest and singularly the most important components of the total mitral valvar apparatus. They are what make the mitral valve a valve. All other components of the apparatus are designed to support the function of the leaflets. While the aortic and mural leaflets have a different gross appearance, individually they vary little in their gross structure.
The aortic leaflet has a shorter length of attachment to the more fibrous anterior component of the annulus when compared to the mural leaflet, but also has greater depth. It has a somewhat delta, or triangular, shape, with a broad base of attachment, and a rounded apex centrally (Fig. 1).59
Figure 1. The aortic leaflet of the mitral valve has a modified delta, or triangular, shape with a broad base attached at the annulus, and a somewhat rounded apex. The tendinous cords inserting in the apical portion of the leaflet are perceptively thinner than the so-called “secondary cords”, which insert more centrally on the ventricular aspect of the leaflet. Photograph courtesy of Professor Robert H. Anderson.
The aortic leaflet, along with the ventricular septum, serves as the outflow tract of the left ventricle.60, 61 With this important function, it is subject to more stress, and appropriately is slightly thicker than the mural leaflet.
The mural leaflet is attached to the fibrous annulus at the parietal junction of the left atrium with the left ventricle. The mural leaflet has a larger length of attachment or circumference, but is narrower and slightly thinner than the aortic leaflet. It has a semicircular shape overall, but exhibits distinct separations along its length which extend from the apposing edge nearly to the annulus.61 This gives the mural leaflet a scalloped appearance, with some going so far as to suggest that the scallops are distinct leaflets.62, 63 The chief function of the mural leaflet is to coapt against the aortic leaflet to create an effective seal, the so-called keystone effect. It is the scallops that permit the mural leaflet, with its longer annular attachment, to fulfill this function, functioning akin to the pleats of a skirt.61
The edges of both leaflets where their surfaces coapt, and at the points of insertion of the primary cords, a feature to be discussed, are thicker than the more central non-coapting surfaces of the leaflets (Fig. 1). This area, known as the “rough zone”, extends around the entire mobile edges of both leaflets, and functions as the competent seal of the valve.
In the normal mitral valve, it is the annulus which defines the mitral valvar orifice, having a larger diastolic and a smaller systolic diameter. The annulus serves as a combined anchor and hinge to support the motion of the leaflets. The structure of the anterior and posterior parts of the annulus differ markedly. The anterior annulus has a thick ridge of connective tissue that joins the non-coronary and left coronary leaflets of the aortic valve. This central band of connective tissue extends both to the right and left around the aortic root, and also around the aortic leaflet of the mitral valve. It is called the region of valvar continuity. The extensions from the central portion of this fibrous skeleton of the heart have a triangular appearance, and are appropriately called the right and left fibrous trigones. The fibrous trigones taper in thickness as they go around the aortic leaflet to join the thinner annulus of the mural leaflet.60, 61
The posterior annulus occupies two-thirds of the overall valvar circumference, while the anterior annulus obviously occupies only one-third. Although anatomically the posterior annulus is typically called the fibrous annulus, a fibrous band is not as readily detected as with the anterior annulus.60, 61
The annulus brings the leaflets closer together in systole to allow a large surface of the leaflets to coapt, insuring the keystone effect and valvar competence.59, 60 The movement of the annulus, which makes the orifice of the mitral valve smaller in systole, is not accomplished by a sphincter-like movement. Instead, the annulus changes from a flat planar axis in diastole to an ellipse in systole, giving the well-described saddle-shaped configuration to the mitral valve.64–67 This is accomplished by the annulus in the region of the commissural cords being pulled during systole below the horizontal axis of the annulus along the central attachment of the leaflets. The total length, or circumference, of the annulus does not change, but the diameter decreases, permitting the leaflets to have a larger surface for coaptation. This spatial concept can be more readily understood by considering schematically the midpoint of the attachment of the aortic leaflet to the anterior annulus as 12:00, and the midpoint of the attachment of the posterior annulus to be 6:00 (Fig. 2a). In diastole, the mitral orifice is on a planar axis, and is nearly circular. With systole, the diameter of the axis from 12:00 to 6:00 is decreased, enabling coaptation of the leaflets. The diameter of the axis from 3:00 to 9:00 is essentially unchanged. This critical function occurs when the annulus assumes an elliptical configuration, as the parts of the annulus in the regions of the ends of the solitary zone of apposition between the leaflets, at 10:00 and 2:00, is pulled below the planar axis (Fig. 2b).
Figure 2. A diagram of the orifice of the mitral valve (a), with the mid-point of the attachment of the aortic leaflet attachment to the anterior annulus designated as 12:00, while the midpoint of the attachment of the mural leaflet to the posterior annulus is designated as 6:00. The nearly circular configuration of the annulus permits the same diameter in the 3:00 to 9:00 axis as in the 12:00 to 6:00 axis. The two ends of the solitary zone of apposition between the leaflets, the so-called “commissures”, are roughly at 10:00 and 2:00 position. With systole, the diameter from 12:00 to 6:00 diameter (shaded) is shortened, while the axis from 3:00 to 9:00 is essentially unaltered. The diagram (b) provides a three-dimensional concept of how the annulus changes from a circular, flat position in diastole to an elliptical configuration in systole, thus bringing the mid-portion of the aortic and mural leaflets into closer proximity.
Tendinous cords
The arrangement of the tendinous cords is unique to an individual. The variation includes the number of cords that arise from the two papillary muscles, the site of insertion into the aortic and mural leaflets, the size and thickness of the individual cords, the number that insert into the thickened rough zone of the leaflet, and whether the insertions include just a single or multiple branches of the cord.60 Basically the cords can be assigned to three groups by their site of insertion into the leaflet.60, 61 Those that insert into the thickened edge of the rough zone are called “primary” or “marginal” cords. Those that insert farther into the central portion of the body of the leaflet are termed “secondary” cords. Some secondary cords insert beyond the rough zone into the body of the leaflet without branching, and are called “strut” cords. The third group, the so-called “commissural” cords, usually arise as single cords from each of the papillary muscles, and send branches to both leaflets adjacent to the mitral annulus at the site of their most peripheral attachment.60 These so-called commissural cords, therefore, support the ends of the solitary zone of apposition between the aortic and mural leaflets.
Cordal function varies between these three groups. The primary cords are responsible for bringing the leaflets down into the left ventricle so as to gain a large area of coaptation, and thus to insure the keystone effect.60 Because the primary cords attach to the leaflets where they coapt, they are subject to less stress. Hence they are thinner than secondary cords (Fig. 1). Cutting a primary cord results in immediate mitral regurgitation, but not in loss of ventricular function.68
The secondary cords are necessary to maintain valvar and ventricular systolic interaction. They aid in shortening the ventricular myocardium to decrease the size of the left ventricular cavity. Cutting secondary cords does not result in mitral regurgitation, but does result in decreased ventricular function.68
The commissural cords behave similarly to the action of the primary cords in maintaining coaptation of the leaflets. Primary cords inserting along the edges of the scallops of the mural leaflet are also necessary to maintain valvar competence.61 A qualifying comment concerning the term “commissure” is appropriate. When applied to the valvar leaflets, the term “commissure” defines the area of coaptation. In the mitral valve, therefore, there is but a solitary commissure, albeit that its two ends are usually defined as the “commissures”. The commissural cords, however, are involved with the leaflets at their most peripheral attachments to the annulus. In addition to maintaining coaptation at the base of the leaflets, the commissural cords may also contribute to annular configuration.
Acute rheumatic carditis
What happens to the individual components of the mitral valve during acute rheumatic carditis? The most important component, the leaflets, is heavily involved. The coapting surfaces develop fibrinous vegetations on their atrial side. The body of the leaflets show cellular infiltrates, oedema, and neovascularization.23 These impressive changes probably do not result in any valvar incompetence. Oedema of the leaflets should not cause loss of coaptation. Even the fibrinous vegetations apparently do not cause significant regurgitant flow. Nothing is seen on Doppler studies during the initial bout or bouts of carditis to indicate that the specific involvement of the leaflets results in mitral valvar incompetence. The extensive changes in the leaflets that follow the acute episode do set the table, so to speak, for what follows to establish chronic rheumatic heart disease. This is persistent mitral regurgitation from leaflet and cordal contraction, and mitral stenosis from thickening of the leaflets, and fusion of the cords and the ends of the zone of apposition between the leaflets.
The annulus dilates during acute rheumatic carditis, as described by Carey Coombs in 1924,69 who stated “At this time the cusps are but little deformed but owing to the wide separation of the ring the valvular apparatus is apparently rendered inadequate.” In the 1950s, initial efforts to decrease annular size involved suturing across the ends of the zone of apposition between the leaflets.70, 71 Prosthetic valves well above the expected size of the annulus in younger patients can be well accommodated if replacement is deemed necessary.72 Although the cause of the annular dilation was always assumed to be myocarditis, this seems unlikely in the absence of myocytic injury and ventricular dysfunction. As the volume of mitral regurgitation increases, both the left atrium and left ventricle dilate, producing further annular dilation, and cordal and papillary muscular dysfunction.48 Thus, the adage that mitral regurgitation begets mitral regurgitation.73
Marcus et al.48 described well the pathogenesis of acute mitral regurgitation. They argued that the initial post-streptococcal injury causes annular dilation, leading to a decrease in effective coaptation of the leaflets. This loss of the large surface of coaptation results in increased tension on primary cords. The primary cords, structurally unsuited to excessive tension, elongate, causing the edge of the aortic leaflet to lose effective apposition with the mural leaflet, and resulting in prolapse of the apposing edge of the aortic leaflet into the left atrium. The loss of coaptation of the aortic leaflet results in a regurgitant orifice, with a posterolaterally directed jet of mitral regurgitation striking the left atrium at the site of McCallum's patch.74
Primary cordal stretching, and prolapse of the aortic leaflet of the mitral valve, was described by Carpentier over 20 years ago.75 This remarkable French cardiovascular surgeon, well-known for correcting mitral annular dilation with a plastic ring which bears his name, and which is employed world-wide, added cordal shortening to correct rheumatic mitral regurgitation over two decades ago.76 The combination of reducing annular size with a plastic ring, and plastic repair of cords and leaflets, is now standard for correction of stable rheumatic mitral regurgitation.15, 76–79 Prosthetic replacement is usually reserved for intractable failure during acute carditis, since repair may break down, requiring a second operative procedure. In the 21st Century, where established cardiovascular surgery is available, no patient should die from rheumatic mitral regurgitation or aortic regurgitation, be it acute or chronic.80
While we have stressed involvement of primary cords of the aortic leaflet of the mitral valve, it should be appreciated that infrequently cords of the mural leaflet may also be involved, and may even rupture during an acute episode.81
Recent studies of mitral valvar function
Numerous clinical and experimental studies of mitral valvar function conducted recently, without rheumatic mitral regurgitation in mind, support the new post-echocardiographic concept of the pathogenesis of the initial episode of childhood rheumatic carditis. In 1997, Obadia et al.,68 using a working pig heart, found that primary cords of the aortic leaflet are important for mitral valvar competence, with the secondary cords being more involved with left ventricular function.
A highly detailed study of five excised porcine mitral valves mounted in a left heart simulator was reported by a collaborating group from Aarhus University, Denmark and the Georgia Institute of Technology, Atlanta.82 In this study, the position of the papillary muscles was changed to alter cordal tension which, in turn, caused functional mitral regurgitation.
A recent report by Nazari et al. from Pavia University83 summarized the effects of varying systolic stress distribution on the cords supporting the aortic leaflet of the mitral valve under normal and pathologic situations. This insightful theoretical analysis, similarly done without specific consideration of rheumatic carditis, supports the new post-echocardiographic concept of childhood rheumatic mitral regurgitation. Figure 3 is reproduced (with permission) from their article.
Figure 3. This schematic illustrates how annular dilation results in loss of coaptation between the surfaces of the leaflets, which in turn causes an increased tension on the primary tendinous cords. Reproduced with permission from Nazari et al. J Cardiovasc Surg 2000; 41: 193–202.
Lomholt et al. at the Aarhus University in Denmark84 reported an elegant experimental study in a porcine model using miniature C-arm pressure transducers to measure cordal tension in primary and secondary mitral cords. The miniature transducers were inserted into cut cords during cardiopulmonary bypass. When the animals became stable after removal from bypass, they were subjected to manoeuvres to raise and lower left ventricular pressure, thereby altering cordal tension. With partial aortic occlusion causing an increase in left ventricular pressure from 109 to 152 millimetres of mercury, cordal tension in the strut cords rose by 45 percent. Conversely, partial caval obstruction effecting a drop of left ventricular pressure from 100 to 90 millimetres of mercury, reduced cordal tension by approximately half. A strikingly similar study, albeit infrequently cited, was reported in 1963 by Salisbury et al.85 Their study was done with 15 mongrel dogs, using essentially the same measurements of cordal tension subsequent to bypass with a small pressure transducer. Their study also showed that cordal tension was proportional to left ventricular pressure. In addition, they found that increased left ventricular volume also caused increased cordal tension.
While it is clearly evident that annular dilation is an early event leading to mitral regurgitation, we do not have a precise explanation for why it occurs. A possibility, not previously considered, is that injured commissural cords might stretch similarly to what occurs with primary cords. This could result in the annulus failing to assume an elliptical configuration. If the diameter of the annulus from 12:00 to 6:00 were not reduced, a smaller area of leaflet coaptation would result (Fig. 2a, b). With minimal injury, resolution of the acute episode could permit the cords and the annulus to return to normal or near normal, which could explain the frequent disappearance of the murmur of mitral regurgitation. We emphasize that such an initial event, while logical, remains theoretical, and will require studies for confirmation.
The new concept of the pathogenesis of childhood rheumatic mitral regurgitation admittedly is not the total answer, but does explain many aspects of acute rheumatic carditis which are infrequently or incompletely addressed. Molecular mimicry between components of group A streptococcus and components of human cardiac valvar tissue is an accepted concept explaining the “valvitis” of acute rheumatic carditis. Molecular mimicry, however, should involve both atrioventricular valves equally, which clearly is not the case. Significant tricuspid regurgitation virtually never occurs with the initial episode or episodes of rheumatic carditis, whereas mitral regurgitation is the benchmark of rheumatic carditis. The principal difference between the two valves is the tension to which they are subject from their respective ventricles. Left ventricular pressure is roughly four times that of the right ventricle. Significant tricuspid regurgitation occurs in the setting of mitral valvar disease and secondary pulmonary hypertension.86 Rheumatic tricuspid regurgitation, as with rheumatic mitral regurgitation, follows annular dilation86 and surgical repair is technically the same, that is, annuloplasty.
Figures 4a, 4b and 4c demonstrate the wide spectrum of echocardiographic findings in childhood acute rheumatic mitral regurgitation, from gross prolapse of the apposing edge of the aortic leaflet (Fig. 4a), to the posterolateral jet of mitral regurgitation associated with a grade 3 murmur of mitral regurgitation (Fig. 4b), and to a holosystolic posterolateral jet that is not audible in a patient with Sydenham's chorea (Fig. 4c).
Figure 4. (a) shows prolapse of the apposing edge (arrow) of a thickened aortic leaflet as seen in a 12-year-old boy with a grade 3 murmur of mitral regurgitation. (b) shows a turbulent mosaic jet of mitral regurgitation directed posterolaterally as seen on color flow Doppler in a 10-year-old boy with a grade 3 murmur of mitral regurgitation, and a grade 2 mid-diastolic murmur, the so-called Carey Coombs murmur. (c) shows a posterolateral jet of mitral regurgitation detected in a 10-year-old girl with Syndenham's chorea, in whom no murmur could be heard.
The posterolateral jet is typical of rheumatic mitral regurgitation in childhood. When large, this jet causes a lesion of thickened endocardium known as MacCallum's patch.74 This lesion is seen only on postmortem examination or at open heart surgery. The lesion differs from other endocardial jet lesions in that cellular infiltrate is present in the subendocardial tissue. It is not known if smaller jets of mitral regurgitation also cause subendocardial cellular infiltrate. With chronic rheumatic mitral regurgitation, fusion of the leaflets and cords, and contracture, can result in more central jets and smaller patches of endocardial thickening.87
In a significant number of patients with “pure” chorea and isolated rheumatic polyarthritis, a holosystolic posterolaterally directed jet of mitral regurgitation that cannot be heard by auscultation can be detected with Doppler echocardiography (Fig. 4c). Some of these patients later develop murmurs of mitral regurgitation. In others who initially have a murmur of mitral regurgitation which disappears on follow-up, a posterolateral jet of mitral regurgitation may still be detected with Doppler echocardiography. This phenomenon is called “subclinical carditis”. Its importance has not been established, and a worthwhile discussion of this somewhat controversial topic is beyond the scope of this presentation. The American Heart Association, with a justifiable concern that over-interpretation of echocardiographic findings could lead to iatrogenic disease, has recommended that mitral regurgitation must be heard to validate the presence of carditis.17, 88 Experience to date suggests their concern is not justified, and that the use of echocardiography is more likely to lead to a correct diagnosis than to an over-diagnosis of rheumatic heart disease.89, 90
Institutions that cared for large numbers of cases of rheumatic fever during the first half of the last century had regimes of prolonged and strict bed rest. This practice was based largely on the empirical observation that patients improved with bed rest, sometimes dramatically. There were no randomized studies, except for a “control” study which showed patients cared for at home had more residual rheumatic heart disease than those who received institutional convalescent care.91 At that time, myocarditis was considered to be the dominant component of rheumatic pancarditis, and the most effective way to decrease cardiac output was to put the patient on bed rest. As discussed, both recent and past experimental studies of mitral valvar function have shown that cordal tension is decreased with lower left ventricular pressure. Thus, the beneficial effect of bed rest in patients with acute carditis is more likely due to reduced tension of the primary cords of the aortic leaflet. Removing children from active play, and placing them on bed rest, is an effective way of decreasing cordal tension. While such studies have yet to be done, this newer understanding would suggest that afterload reducing agents may also be beneficial in this setting.
Children with polyarthritis tend to develop less severe rheumatic carditis. There are probably two reasons why this is true. Children with polyarthritis are inactive because of pain and are also more likely to seek medical care and be placed on bed rest. This observation also offers additional support of the new concept.
We stress that this new concept does not explain specifically how the annulus and cords are initially injured, nor does it explain how mitral regurgitation progresses to mitral stenosis. It does, however, give us reason to consider valvar structure and function in seeking a more complete explanation of both acute and chronic rheumatic heart disease.
Acknowledgements
Dr. Edward B. Clark and Dr. Harry R. Hill reviewed the manuscript and made editorial suggestions. Members of the University of Utah Post-Streptococcal Syndrome Study Team offered critical suggestions.
