Some of these HLA-B27-associated rheumatic or inflammatory conditions have long been known to be accompanied by "cardiac complications," specifically, atrioventricular conduction blocks and lone aortic regurgitation (in contrast with aortic regurgitation combined with a stenotic lesion). During the past decade, we have learned that HLA-B27 is also an important genetic risk factor for these cardiac conditions, regardless of the presence of the typical extracardiac rheumatic syndromes. At one end of the clinical spectrum, we have now identified an HLA-B27-associated cardiac syndrome that consists of severe cardiac conduction system abnormalities and lone aortic regurgitation; the link between this syndrome and HLA-B27 is almost as strong as the link between ankylosing spondylitis and HLA-B27. This paper describes the observations that have led me to propose the replacement of the concept of "cardiac complications" of HLA-B27-related rheumatic disease with the concept of "HLA-B27-associated cardiac disease."

Methods

Top Methods Author & Article Info References

Clinical studies, case reports, reviews, and textbooks were searched for information about the manifestations of cardiac disorders in patients with seronegative spondyloarthropathies who do or do not have HLA-B27. I have collected a personal database since 1977, and the information contained therein was augmented by repeated searches of the English-language literature in the MEDLINE database up to 1995. The reference lists in publications thus identified were used to locate other publications, and the collection was expanded in this way to cover the English-language literature published since 1936. Some papers published in German or French with English abstracts were also included.

Valve Disease

Aortic Valvular Regurgitation

Mallory's 1936 description [9] of aortic valvular regurgitation in two young men with retrospectively clear diagnoses of seronegative spondyloarthropathies has been recognized as the first report of cardiac manifestations in this context, although the connection was far from clear at that time [9]. Aortic regurgitation in the clinical setting of ankylosing spondylitis has subsequently been well characterized and distinguished from aortic valvular dysfunction in other disorders [10, 11]. Three factors can, to a variable extent, contribute to the development of incompetent aortic valves: dilatation of the aortic root, fibrotic thickening and downward retraction of the bases of the cusps, and inward rolling of the edges or margins of the cusps (Figure 1, Table 1). These allegedly pathognomonic features have also been described in patients with other seronegative spondyloarthropathies [12, 13]. The absence of stenotic lesions is notable, although a systolic murmur attributable to increased stroke volume can be prominent, in addition to the high-pitched diastolic murmur from regurgitant flow.

View larger version (0K): [in this window] [in a new window]   Figure 1. Aortic valve from a patient with ankylosing spondylitis. The valve and sinus of Valsalva are cut open. The aortic valve (circumference, 80 mm) and the ascending aorta are both dilated. The base of the aortic cusps is thickened, and the free margins are rolled. All three features contribute to aortic valve insufficiency. The arrow indicates the subaortic ridge caused by fibrosis extending downward from the nonseptal portion of the aorta onto the anterior mitral leaflet. This can impair the mobility of the anterior mitral leaflet and contribute to mitral valve insufficiency. Reproduced from Bergfeldt and colleagues [12] with permission.

An early report from an autopsy series [10] and reports of later echocardiographic studies [14-16] have shown that histopathologic and anatomic abnormalities, respectively, can be detected in a substantial proportion of patients who do not have clinical evidence of cardiac disease and thus have early or mildly expressed aortic disease.

Epidemiology

One of the first large studies on this subject suggested a relation between the duration of ankylosing spondylitis and the appearance of aortic regurgitation, with prevalences of valve dysfunction of 2% after 10 years of disease and 12% after about 30 years of disease [17]. Overall, aortic regurgitation has been diagnosed in 2% to 10% of patients with ankylosing spondylitis [18, 19]. The prevalence of this valvular lesion among patients with other seronegative spondyloarthropathies has not been well defined, but a study of 164 patients with Reiter disease identified 4 cases (2.8%) of aortic regurgitation [20]. Using the opposite approach, Schilder and colleagues [21] found 5 patients with ankylosing spondylitis among 100 patients who were being evaluated for possible surgical correction of aortic regurgitation. Similarly, Qaiyumi and coworkers [22] found 4 patients with ankylosing spondylitis and 3 patients with Reiter disease among 100 patients with lone aortic regurgitation. All of these patients were positive for HLA-B27, and 6 had conduction system abnormalities. Because ethnic background strongly influences the prevalence of HLA-B27, a subgroup analysis is justified: Six of these 7 patients were found among the 61 white patients in this study. In a study of similar size [23], we found that at least 15% of a group of patients with lone aortic regurgitation had HLA-B27-related disorders.

The relation between HLA-B27 and lone aortic regurgitation has also been explored in the absence of "typical" clinical rheumatic or ocular manifestations [23-26]. Although no such relation has so far been established, aortic regurgitation has been suggested as a possible forme fruste of ankylosing spondylitis [27].

Juvenile or Fulminant Aortic Regurgitation

Several case reports [28-37] have drawn attention to uncommon features of aortic regurgitation in the setting of seronegative spondyloarthropathy. Aortic regurgitation can present in patients as young as 9 years of age, can precede serious joint manifestations, and can have a fulminant course. The prevalence of cardiac manifestations seems to be similar in children with ankylosing spondylitis and in adults [38].

Treatment

The surgical treatment of aortic regurgitation in patients with ankylosing spondylitis has been reviewed [39]. Substantial dilatation of the ascending aorta above the most proximal centimeters that are usually affected has been a concern, as has the frequent co-occurrence of restrictive pulmonary dysfunction. Issues of potential importance that have not been resolved, or even addressed, are 1) whether the tendency toward myocardial fibrosis in these patients should influence the decision about when to initiate surgical treatment and 2) whether the recovery of left ventricular function after surgery differs in patients whose aortic regurgitation is the result of HLA-B27-related disease rather than another cause.

Mitral Valve Regurgitation

Insufficiency of the mitral valve may result from subaortic fibrosis of the anterior leaflet with impairment of its mobility; a "subaortic hump (bump)" or an actual ridge is the pathoanatomic, angiographic, and echocardiographic feature [40, 41] (Figure 1, Table 1). This manifestation seems to be much less common than aortic regurgitation; to my knowledge, fewer than 10 cases have been reported [29, 33, 40-43]. Mitral insufficiency may also result from left ventricular dilatation secondary to aortic regurgitation.

Arrhythmias

Atrioventricular Block

Epidemiology

Since the 1940s, atrioventricular conduction blocks have been reported in patients with ankylosing spondylitis and Reiter disease [44, 45] and have been regarded as the most common cardiac complication [46]. In the first systematic study on this subject, Bernstein and Broch [44] examined routine electrocardiograms from 190 patients with ankylosing spondylitis. They found that 29 (15%) had first-degree and 3 (1.6%) had third-degree atrioventricular (complete) heart block [44]. Subsequently, cardiac conduction abnormalities, including atrioventricular and intraventricular blocks, have been found in up to one third of patients with ankylosing spondylitis who were followed longitudinally for many years with repeated electrocardiography [47]. This is important in view of a tendency for atrioventricular block to occur intermittently in these patients [18, 47, 48]. The intermittent nature of the block is also an important piece of evidence supporting the involvement of a reversible inflammatory process rather than permanent fibrosis in its causation (Table 1).

Complete heart block has been found in 1% to 9% of patients with ankylosing spondylitis [18, 44, 47, 49, 50]. In Reiter disease and other seronegative spondyloarthropathies, the prevalence of complete heart block is poorly defined. Good [20] found atrioventricular blocks in 10 of 164 patients (6%) with Reiter disease but stated that "the true incidence may be considerably greater." At least 12 patients with Reiter disease and complete heart block have been described [45, 51, 52], and atrioventricular block has been suggested as an early feature of Reiter disease [53].

According to studies done in cardiology patients [54, 55], as many as 15% to 20% of male patients with permanent pacemakers may have an HLA-B27-related disease process as the cause of their bradycardia. Seronegative rheumatic disease was recognized only about 50% of the time in these patients before a cardiac conduction abnormality was diagnosed. Complete heart block may be the only HLA-B27-related disease manifestation, occurring in the absence of rheumatic and ocular manifestations [56]; this observation was recently confirmed [57]. Similarly, Sahi and Winfield [58] reported early-onset complete heart block in a 35-year-old HLA-B27-positive man with no signs of ankylosing spondylitis whose two brothers had this rheumatic disease. It is important to note that there is no general association between complete heart block and HLA-B27. Instead, these studies suggest that in a subgroup of patients with acquired complete heart block, an inflammatory disease process related to HLA-B27 seems to be the cause of the cardiac conduction system abnormality. The relative risk that an HLA-B27-positive man will need a permanent pacemaker was calculated to be 6.7 compared with a man who has other B alleles [56]. This makes HLA-B27 one of the strongest genetic risk factors for cardiac disease identified to date [59]. In this respect, there is a sex-linked difference in susceptibility to HLA-B27-related heart block; in female patients with high-degree heart block and permanent pacemakers, the frequency of HLA-B27 was not increased [60].

Invasive Electrophysiology

Electrophysiologic studies, summarized in Table 2, show that the preferential level of block is in the atrioventricular node itself ([51, 61-66]; unpublished data). This contrasts with most cases of acquired complete heart block, 80% of which are within or below the bundle of His [67] (Table 3). However, this is at the same level in the conduction system as digitalis-induced block [68], most cases of congenital or very early acquired block [69], and atrioventricular block that occurs in the clinical setting of an acute inferior myocardial infarction [70].

Treatment

Although a transient episode of complete heart block in a setting of HLA-B27 positivity may not recur for up to 26 years [18, 47, 48, 66], no reliable predictor of this favorable outcome has been identified. The obvious treatment in the presence of symptoms is therefore the implantation of a permanent pacemaker. In the absence of symptoms and in the presence of an adequate escape rhythm, a more conservative approach may be justified. In patients with HLA-B27-related disorders receiving long-term pacemaker treatment, prognosis is good and longevity is not adversely affected [71].

Other Arrhythmias

Bradycardia or pauses related to sinus node dysfunction, as well as prolonged sinus node recovery times, have been seen on invasive electro-physiologic testing [65, 66]. A contributing factor is probably obliterative intimal proliferation in the sinus node artery [72] (the same kind that has been described in tissues adjacent to afflicted joints [73-75]), the vasa vasora of the aortic root [11, 12], and the atrioventricular node artery (Davies MJ. Personal communication) (Figure 2, Table 1).

View larger version (0K): [in this window] [in a new window]   Figure 2. Intimal proliferation within the atrioventricular nodal artery of a patient with HLA-B27-associated cardiac disease. (Courtesy of Professor Michael J. Davies, St. George's Hospital, London.).

Atrial fibrillation has been seen in several patients with HLA-B27-associated rheumatic manifestations who had neither other cardiovascular abnormalities nor other noncardiovascular disorders that could explain the presence of this arrhythmia [76]. However, a systematic study of the relation between HLA-B27 and "lone" atrial fibrillation has not been reported, and any relation is therefore unproven, particularly in view of the substantial frequency of this arrhythmia in the general population. No systematic studies on ventricular arrhythmias in the setting of HLA-B27-related cardiac disease have been done.

Myocardial Disease

"Carditis" and "cardiomyopathy" were reported in patients with ankylosing spondylitis more than 25 years ago, although the characteristic features of myocardial involvement remained unclear for many years [17, 77]. The first detailed reports on this condition were from studies that used several investigational techniques, such as echocardiographic and histopathologic examination, nuclear angiography, and the assessment of systolic time interval [78-80]. The first study included patients with ankylosing spondylitis and Reiter disease [78]; the others included patients with ankylosing spondylitis and healthy controls [79, 80]. The patients with rheumatic disorders did not have signs of aortic regurgitation or serious conduction system abnormalities, nor did they have other diseases, such as hypertension or diabetes, that could have caused cardiac dysfunction. Although the main finding in the first study was a dilated and poorly contracting left ventricle [78], the other studies emphasized early diastolic filling abnormalities [79, 80]. Histologic examination showed a mild, diffuse increase of interstitial connective tissue but no inflammatory changes or amyloid deposits [79]. Systolic time interval evidence of subclinical systolic dysfunction was found in 17% of patients with ankylosing spondylitis who did not have valvular abnormalities [81]. Two additional echocardiographic reports [82, 83] confirmed diastolic filling abnormalities in patients with ankylosing spondylitis; these abnormalities were not present in controls. No relation to age, disease duration, or peripheral joint problems was seen among those with cardiac abnormalities [82]. Follow-up data are not available, and the clinical significance of these observations is uncertain with regard to illness, both in patients with lone left ventricular dysfunction and in patients with concomitant aortic valve regurgitation.

Pericarditis

Both clinically and on histopathologic examination, pericarditis is a rare feature that has been recognized in 1% or less of studied patients [81, 84]. The proportion of afflicted patients and the relative clinical importance of different cardiac abnormalities in patients with seronegative spondyloarthropathies are shown in Table 4.

Pathophysiology

Etiologic Agents

An intensive search for an etiologic agent in ankylosing spondylitis has been in progress for many years. No single agent has been definitely incriminated as the cause of this disease. A possible role for Klebsiella species in the development of ankylosing spondylitis was first suggested in 1976 [85]. The occurrence of arthritis as part of Reiter disease or as a more isolated feature has been demonstrated after infections with other gram-negative bacteria, such as Shigella [86-88], Salmonella [89, 90], and Yersinia species [91, 92].

Pathogenetic Role of HLA-B27

The role of HLA-B27 in the development of disease is the subject of a large body of literature but is beyond the scope of this paper; the interested reader is directed to several reviews [93-96] and the proceedings of a recent symposium [97]. However, a few points may be of general interest. It seems that not only susceptibility to disease but also the disease course can be modified by the patient's genetic background, that is, by the presence or absence of HLA-B27 [98]. In addition, an animal model of human disorders associated with HLA-B27 has been described in transgenic rats that express HLA-B27 and human ß₂ µglobulin [99].

Serologic Activity

Unfortunately, specific serologic markers are lacking, both for diagnostic purposes and for assessment of disease activity [100, 101]. This is reflected by the name "seronegative spondyloarthropathies," which alludes to the absence of substantial titers of rheumatoid factor. This does not mean that serologic markers are absent in all patients; rather, it means that interindividual variation in serologic reactivity is considerable and has not, to date, allowed the definition of specific patterns.

Similar variability in serologic status was seen in patients with cardiac manifestations [102]. In this study, however, a significant increase was found in in vitro platelet-aggregating activity in patients with HLA-B27-associated cardiac disease compared with controls. This observation needs to be confirmed both in vitro and in vivo, and the cause or causes of this activity must be elucidated.

Development of Disease

The present consensus is that the HLA-B27 gene itself is the major disease susceptibility gene (Table 5). Environmental triggers (antigens) are probably present. Although most patients do not have a history of infectious disease before the onset of rheumatic disease (which is often insidious), some persons develop clinically recognizable rheumatic disease after a clinical infection with gram-negative bacteria, as noted above. The importance of ongoing subclinical infection in rheumatologic and other inflammatory disease pathogenesis is still being debated. It is intriguing that tissues enduring repeated or sustained tensile forces seem to be preferentially affected. This suggests that iterated physiologic trauma plays a role or that the composition of the involved connective tissue is crucial.

Characteristically, the disease course varies both between and within individual persons, over time as well as with respect to the distribution of organ manifestations and their severity (measured either as residual deformities or as a need for therapy), and serologic evidence of disease activity. Relatively well-defined syndromes are recognizable, but the temporal dispersion of various manifestations often makes it difficult to establish a correct diagnosis and identify a relation among different organ manifestations. Frustration on the part of both patient and physician often ensues, not only because of trouble-some symptoms but also because an integrating framework for interpreting multiple and seemingly unrelated disease manifestations is lacking. It is important to note that 75% to 80% of HLA-B27-positive persons do not develop any disease that today can be related to this genetic trait. The reason or reasons for a penetrance of HLA-B27-associated disease of 20% to 25% is unknown but is of obvious importance.

HLA-B27-Associated Cardiac Disease

Since the late 1940s, cardiac involvement has been recognized in patients with seronegative spondyloarthropathies and has usually been regarded as a "complication" of the extracardiac illness [46]. However, clinical observations over the past several decades indicate that the same underlying disease process can independently involve different organs to varying degrees in different patients. Thus, patients with these disorders sometimes present to specialists other than rheumatologists. The possibility that the heart is a target organ for an HLA-B27-associated, as yet undefined disease process was suggested early in the 1980s [103, 104], and the concept of "HLA-B27-associated cardiac disease" subsequently developed [12, 22, 23, 55-58]. This concept is now supported by immunogenetic, histopathologic, clinical, and electrophysiologic evidence.

Immunogenetic Evidence

The frequency of HLA-B27 is significantly increased in male patients with pacemakers compared with the general population. This is especially true in those with heart block, even in the absence of typical articular and extraarticular manifestations [56, 57]. Additional immunogenetic evidence has been presented, as noted above [58]. In addition, the syndrome of severe conduction system abnormalities plus aortic regurgitation is strongly associated with HLA-B27, which is found in 67% to 88% of patients with these two cardiac conditions [22, 23]. The concomitant occurrence of conduction system abnormalities and aortic regurgitation was noted before HLA typing was widely used [49, 105-107]. Furthermore, in the animal model of HLA-B27-associated disorders, both the heart and the aortic root have been targets for spontaneously developing inflammatory disease [99].

Histopathologic Evidence

Cardiac and aortic tissues show histopathologic abnormalities in the small arteries; these abnormalities are similar to those found in tissues adjacent to afflicted joints [11, 12, 72-75] (Table 1).

Clinical Evidence

Both conduction system abnormalities and aortic regurgitation have been seen in patients with various HLA-B27-related rheumatic and ocular manifestations, without regard to the severity of these manifestations. Approximately 50% of HLA-B27-positive patients with these two cardiac lesions had not previously received a diagnosis of an HLA-B27-related rheumatic disorder [23, 54, 55]. Thus, the initial suggestion [17] of a relation between the duration of rheumatic disease and the occurrence of cardiac abnormalities does not seem to be generally valid. In addition, in one study of 74 patients with postinfectious Yersinia arthritis, "definite carditis" was diagnosed only among HLA-B27-positive patients (it was found in 5 of 49 HLA-B27-positive participants) [98].

Electrophysiologic Evidence

In HLA-B27-associated disease, high-degree atrioventricular blocks are typically located in the atrioventricular node, even in patients with concomitant fascicular or bundle-branch blocks [51, 61-66] (Table 2 and Table 3). The explanation for this relative selectivity is probably an abnormality of the atrioventricular node artery rather than septal fibrosis (which can be widespread in these patients) (Figure 2) because the blocks caused by septal fibrosis should be lower down in the conduction system.

Implications

Defining the genetic background of the many manifestations of HLA-B27-associated inflammatory disease has not yet led to the definition of the basic mechanisms underlying disease development, a specific treatment for the disease, or measures that will prevent the disease short of preventing infection with certain bacterial "triggers." However, it has been possible to define a clinical framework that allows recognition of a genetically linked systemic process. This process can produce widely varying disease profiles, including clinically important cardiac disease, which indicates that the heart is one major target organ. This framework makes it possible to inform patients properly about the nature of this disorder and about its variability with regard to manifestations, time course, intensity, and severity. On the positive side, spontaneous remission is possible, and established, reliable techniques for the treatment of the cardiac manifestations are available. Future interdisciplinary research projects may also be facilitated by this integrating concept of a complex, multifaceted disease process.