Mitral Valve Prolapse: Causes, Clinical Manifestations, and Management Richard B. Devereux, MD; Randi Kramer-Fox, MS; and Paul Kligfield, MD

Transcription

1 REVIEW Mitral Valve Prolapse: Causes, Clinical Manifestations, and Management Richard B. Devereux, MD; Randi Kramer-Fox, MS; and Paul Kligfield, MD Purpose: To assess the causes, methods of diagnosis, clinical spectrum, and management of mitral valve prolapse. Data Identification: Results of prospective study of over 800 subjects at Cornell Medical Center from 1979 to the present were used along with studies published from 1963 to 1989 identified by computerized literature searches of Index Medicus and MEDLINE, and by hand searches. Study Selection: Studies involving controlled design, longitudinal follow-up, or critical assessment of diagnostic methodology, and clinical studies or previous reviews that have contributed most to the understanding of mitral valve prolapse were selected. Data Extraction: Data concerning the causes, clinical manifestations and complications, and prevalence of mitral valve prolapse, as well as the strength of association between mitral valve prolapse and diagnostic signs, were evaluated and used to develop a practical approach to evaluating and managing patients. Results of Data Synthesis: Most instances of mitral valve prolapse are primary and inherited, with possible genetic heterogeneity. Mitral prolapse may be diagnosed by auscultation of midsystolic clicks and late-systolic murmurs that respond typically to maneuvers, or by billowing of mitral leaflets across the mitral anular plane in long-axis, twodimensional echocardiographic views or by a late-systolic, posterior leaflet displacement of at least 2 mm in meticulously targeted M-mode recordings. Mitral valve prolapse is associated with thoracic bony abnormalities, low body weight, low blood pressure, and a modest excess of orthostatic hypotension, syncope, palpitations, and atrial arrhythmias, but not with nonspecific symptoms (atypical chest pain, dyspnea, anxiety or panic attacks). Complications of mitral valve prolapse, including about 4000 mitral valve operations, 1100 cases of endocarditis, and possibly 4000 sudden deaths per year in the United States, are concentrated disproportionately in older men, with about 5% of affected men and 1.5% of affected women ultimately requiring valve surgery. Conclusions: Prophylaxis for endocarditis and closeness of follow-up should be related to the occurrence of the independent risk factors for complications of mitral prolapse (presence of mitral regurgitation, male gender, and age over 45 years), whereas active management and close observation are needed for severe mitral regurgitation and advanced ventricular arrhythmias. Annals of Internal Medicine. 1989;111; From The New York Hospital-Cornell Medical Center, New York, New York. For current author addresses, see end of text. In the 25 years since Barlow (1) first securely established the mitral valvular origin of midsystolic clicks and late-systolic murmurs, mitral valve prolapse has become recognized as one of the most prevalent cardiac abnormalities (2, 3) and as one that frequently underlies such serious conditions as infective endocarditis (4-7) and severe mitral regurgitation (7-10). Early appreciation of mitral valve prolapse was facilitated by cineangiographic demonstration of systolic billowing of mitral leaflets into the left atrium (11, 12), recognition of the effect of posture on auscultatory events and motion of prolapsing mitral leaflets (13-16), and the advent of echocardiography as a noninvasive method of visualizing mitral-leaflet motion (17, 18). Subsequent clinical studies (19-32) linked mitral prolapse to a wide spectrum of symptoms, signs, and complications that have been considered to constitute the "mitral valve prolapse syndrome" (32-34). More recently, systematic research has refined the criteria for diagnosing this disorder (35, 36), established the importance of genetic factors in its origin (37, 38), and elucidated the clinical features and risk of complications associated with prolapse of the mitral valve (4-7, 10, 39-45). Despite these advances, however, uncertainty persists among practitioners concerning several aspects of mitral valve prolapse. Is it a nonspecific phenomenon with many causative factors or a distinct condition with a defined biological basis? Is mitral valve prolapse so prevalent as to preclude any clinical significance or is it only moderately common? Does mitral valve prolapse cause a wide range of cardiac and psychologic symptoms, or is its spectrum of clinical features more limited? Is it a benign "variant of normal" or a dangerous condition with frequent complications? Recent research provides reasonable, although in some instances only provisional, answers to these questions. To address these and related questions, we have used data from a prospective study done at Cornell Medical Center of nearly 300 patients with mitral valve prolapse and of more than 400 members of nearly 100 of these patients' families (including over 100 additional persons with mitral prolapse); we also used data from other control groups (6, 7, 36, 37, 44); from other reports on mitral valve prolapse involving controlled study design, longitudinal follow-up of patients, or critical assessment of diagnostic methodology, and from other relevant studies (a total of more than 1200 from 1963 to 1988). Further information about the prevalence, clinical features, and methods of diagnosis of mitral valve prolapse may be found in recent critical reviews (3, 34, 36) American College of Physicians 305

2 Table 1. Prevalence of Mitral Valve Prolapse in Family Studies Study Families Total First-Degree Variable Diagnostic Prevalence of (Reference) Relatives with Mitral Valve Prolapse Affecting Prevalence Method Mitral Valve Prolapse in Controls n n n(%) % Weiss et al. (66) (47) Sex Echocardiogram Scheeleetal.(69) (31) Age Echocardiogram Devereux et al. (37)* (30) Age, sex Echocardiogram 4 Devereux et al. (36)* (30) Age, sex Echocardiogram 4 Strahan et al. (38) (41) Age Echocardiogram, clinical Chen et al. (67) (27) Echocardiogram Wilcken et al. (68) (20) Echocardiogram 3 Hickeyetal. (70) f (10) Age Echocardiogram 2 * Reference 7 is source of the data on prevalence of mitral valve prolapse in unselected adults. t Relatives with known heart murmur or other evidence of mitral prolapse were excluded. Adapted from Devereux and Kramer-Fox (71). Causes of Mitral Valve Prolapse In common usage, the term "mitral valve prolapse" describes displacement of the mitral leaflets in superior and posterior directions from their normal location during systole (36), in keeping with the definition of prolapse as "the slipping of a body part from its normal position in relation to other body parts" (46). The valve displacement of mitral valve prolapse may be associated with many functional findings, from mild displacement of a valve leaflet without regurgitation to marked "billowing" with loss of systolic apposition and important regurgitation. The abnormal motion patterns of the mitral valve characteristic of mitral valve prolapse result from a disproportion between the valve's connective tissue elements (leaflets, anulus, and chordae tendineae) and their muscular supports (papillary muscles and left ventricular myocardium) (15, 16). The mitral leaflets and anulus are larger in relation to left ventricular size in normal women than in normal men (47-49), which probably contributes to the more frequent expression of mild degrees of mitral prolapse in women than in men (2, 3, 37). Structural enlargement (50-54), abnormal distensibility (49), and distorted local architecture (55-57) of the valve's connective elements have all been documented in patients with mitral valve prolapse. Mitral valve prolapse may be a primary condition or may be secondary to several disorders (3). Mitral valve prolapse is a well recognized phenotypic feature of several heritable disorders of connective tissue, most notably the Marfan syndrome (58, 59) and types I and III of the Ehlers-Danlos syndrome (60). The documented weakness or hyperextensibility of cardiovascular and nonvascular connective tissue in these conditions and their pattern of dominant inheritance suggest that these forms of mitral valve prolapse are expressions of defects in structural connective tissue proteins. Other causes of secondary mitral valve prolapse include both heritable abnormalities of the myocardium (59) and conditions in which left ventricular size is reduced relative to that of the mitral valve. The best documented examples of the latter are ostiumsecundum atrial septal defects (61) and anorexia nervosa (62). Normalization of initially diminished left ventricular size has resulted in disappearance of mitral prolapse in both of these conditions. Conversely, mitral valve prolapse in patients with the Marfan syndrome may be masked by left ventricular enlargement due to progressive aortic regurgitation (63). Although the established causes of secondary mitral valve prolapse are instructive, they are uncommon (3); the best documented of these-the Marfan syndrome-accounts for only about 1 in 500 cases of mitral valve prolapse (64), and the suggested association between segmental left ventricular dysfunction due to coronary artery disease and mitral valve prolapse is not supported by recent data (65). In most instances, therefore, mitral valve prolapse occurs as a primary condition. The precise cause of primary mitral valve prolapse remains undefined, but available evidence indicates that many, if not all, instances of mitral valve prolapse are inherited, with most studies (36-38, 66-71) strongly favoring autosomal dominant transmission (Table 1). The results of family studies (36-38, 66, 69, 70) show that expression of the gene or genes for mitral valve prolapse are affected by both age and sex. Thus, the proportion of affected, first-degree relatives in the first 93 families we studied was approximately 50% among adult, female, first-degree relatives under the age of 50, as would be expected in a fully expressed autosomal dominant disorder, whereas lower prevalences (10% to 30%) were found among adult men, older women, and children of both sexes (36, 37). Thus, expression of genetic mitral valve prolapse is reduced both by male sex and age-related factors that remain unclear. The genetic defect underlying mitral valve prolapse remains unknown. Pathologic studies have shown disruption of collagen bundles in leaflets and chordae tendineae from prolapsed mitral valves (50, 51, 72). Biochemical studies also have shown a spectrum of abnormalities in collagen, including deficiency of collagen type III and AB collagen in one prolapsed valve (73); abnormal ratios of synthesis of collagen types I and III by skin fibroblasts from some patients with August 1989 Annals of Internal Medicine Volume 111 Number 4

3 mitral prolapse (74); and altered ratios of collagen types in some areas of prolapsed mitral leaflets that appeared to be obscured in other regions of the same valves by secondary repair processes (75, 76). However, lack of linkage of mitral valve prolapse to genes for collagen types I, III, and V was shown in two large families (77). Recent evidence that two distinct patterns of abnormal mitral-leaflet motion occur in subjects with mitral valve prolapse, which are characterized by either systolic billowing of mitral leaflets into the left atrium or dynamic systolic expansion of the mitral anulus causing posterior displacement of the leaflets in systole, suggests that two or more subtypes of mitral valve prolapse exist (49). The strongly familial nature of these patterns, with the subtype of mitral prolapse in the index cases also present in nearly 90% of affected relatives in the same family (P < ), suggests that these differences in mitral valve motion are caused by separate genetic entities (49). Mitral valve enlargement and leaflet thickening, which appear to be markers of increased risk for complications (78, 79), occur in a subset of patients with the leaflet-billowing pattern of mitral prolapse (54). Differences in molecular defects may ultimately explain some of the variability in clinical manifestations of mitral valve prolapse. Diagnosis of Mitral Valve Prolapse In view of the importance of midsystolic clicks and late-systolic murmurs in the initial recognition of mitral valve prolapse (1, 13-16, 80), it is fitting that renewed attention should focus on the auscultatory features of this condition. As emphasized recently (81), the most useful auscultatory features of mitral valve prolapse are midsystolic clicks that shift their timing with respect to the first and second heart sounds in response to maneuvers that alter the relationship between left ventricular and mitral valve size; that is, earlier in systole with sitting, standing, or other interventions that reduce ventricular size, or later with those interventions that increase chamber size such as squatting [Figure 1]); and late-systolic murmurs in persons too young to be at risk for mitral anular calcification or papillary muscle dysfunction. One must be attentive to the timing of auscultatory abnormalities; we have found widely split first-heart sounds and mid- Figure 1. Effect of postural changes on auscultatory signs of mitral valve prolapse. Sitting and standing move the midsystolic click (C) closer to the first heart sound (Sj) and prolong the late systolic murmur (SM). On squatting, the click moves toward the second heart sound (S 2 ) and the murmur becomes shorter. From Devereux and colleagues (16); reproduced with the permission of the American Heart Association. Figure 2. Schematic diagram showing the discrepancy in apparent mitral leaflet-anular relationships in echocardiographic recordings. The parasternal long-axis view (upper panel) and apical four-chamber view (lower panel) show the discrepancies produced by the fact that the mitral anulus has a "saddle" rather than a planar shape. Ant = anterior; Ao = aorta; LA = left atrium; LV = left ventricular; post = posterior; RA = right atrium; RV = right ventricle. From Levine and colleagues (35); reproduced with the permission of the American Heart Association. systolic, rather than late-systolic, murmurs to be present in a high proportion of patients with false-positive diagnoses of mitral prolapse (82). It is also noteworthy that auscultatory manifestations are highly variable in subjects with echocardiographic mitral valve prolapse: These include both fluctuation among audible clicks, murmurs, and combinations thereof and shifts back and forth between "silent" mitral prolapse and typical auscultatory findings, which may account for the frequent discordances between these diagnostic modalities in some studies (83, 84). Our experience (36) with 137 subjects who had echocardiographically-documented mitral valve prolapse and were examined twice by a standardized protocol at a mean interval of 4 years is that auscultatory findings change at least slightly in most persons (Figure 2). As a result, several examinations are needed to determine whether a person may intermittently have a murmur of mitral regurgitation, which is an important factor when considering whether to recommend antibiotic prophylaxis (85). Role of Echocardiography Because of its ability to visualize the anatomy and function of the mitral valve, echocardiography has 15 August 1989 Annals of Internal Medicine Volume 111 Number 4 307

4 Figure 3. Variability of auscultatory features between examinations using standard postural maneuvers. Examinations were a mean of 4 years apart in adult patients with echocardiographically documented mitral valve prolapse. Number of patients with each combination of auscultatory findings is indicated at the right-hand end of lines connecting first and second evaluation results. MVR = mitral valve replacement; -I- = death. From Devereux and colleagues (36); reproduced with permission of the American Heart Journal. proved to be the most useful objective method for detection and characterization of mitral valve prolapse. However, the astonishing finding that nearly 20% of some populations have shown echocardiographic mitral prolapse (2, 86) has led to suspicion that this technique may have poor specificity (81, 87). This concern has been addressed by recent echocardiographic studies that have clarified both the dynamic geometry of the mitral valve and the relative merit of different criteria for diagnosis of mitral valve prolapse. Since the early 1970s, the mainstay of echocardiographic diagnosis of mitral valve prolapse has been the demonstration on M-mode recordings of posterior systolic motion of continuous mitral-leaflet interfaces behind the line connecting the valve's closure and opening points (17, 18), by at least 2 mm in late systole or by at least 3 mm for holosystolic prolapse (86). The diagnosis of mitral prolapse by these criteria has been shown to be acceptably reproducible (88-90), provided that tracings are of high technical quality (91); these criteria have also been shown to be more sensitive for detection of mitral prolapse in patients with typical systolic clicks and murmurs than currently accepted two-dimensional echocardiographic criteria (50% to 85% compared with 35% to 57% [36, 90, 92]). The acceptance of a holosystolic posterior displacement of only 2 mm may have contributed to overstatement of the prevalence of mitral prolapse in the Framingham study (2) compared with the 3% to 4% prevalence found in other population samples that have been recently studied (7, 93, 94). An important modification of the original M-mode diagnostic criteria that we and other investigators (70) have adopted is to accept the M-mode diagnosis of holosystolic prolapse only if systolic billowing of one or both leaflets across the plane of the mitral anulus is shown in the parasternal, long-axis, two-dimensional view. Strict application of these criteria is important because a more lenient interpretation can result in dramatic overstatement of the prevalence of mitral valve prolapse (40% compared with 0% in one study [95]). Two-dimensional echocardiography has played an increasingly important but also controversial role in recognition of mitral valve prolapse. Although the original studies detected mitral prolapse by showing systolic billowing of mitral leaflets across the mitral anular plane in the parasternal, long-axis view (96, 97), it was subsequently suggested-in the absence of any direct validation-that the apical four-chamber view was more useful (98). More recent studies have found apparent protrusion of mitral leaflets across the mitral anular plane in the apical four-chamber view in an extremely high proportion of persons who had had normal auscultatory examinations, reaching 34% in one group of adolescents (99). The reason for this surprising result has been clarified by Levine and colleagues (35), who showed that the mitral anulus is not flat but instead has a "saddle" shape. The mitral anulus is farthest from the left ventricular apex in its anterior and posterior portions, where the hinging points of the anterior and posterior mitral leaflets are seen in the parasternal, long-axis view, and is closest to the apex in its medial and lateral portions, approximately where it is transected by the tomographic plane of the apical four-chamber view (Figure 3). Thus, mitral leaflets that lie clearly on the left ventricular side of the mitral anular plane during systole in long-axis views may appear to protrude into the left atrium in the apical, four-chamber view, a phenomenon not associated with mitral regurgitation or other mitral valve abnormalities (79). Accordingly, the diagnosis of mitral valve prolapse by two-dimensional echocardiography should be made only when systolic billowing of mitral leaflets is shown in parasternal or apical, long-axis views (36). This approach is relatively insensitive because it detects billowing in only about 50% of subjects with M-mode echocardiographic and auscultatory evidence of uncomplicated mitral valve prolapse (49, 90, 92); this insensitivity is explained by the fact that posterior displacement of mitral leaflets in systole due to dynamic mitral anular expansion is not readily appreciated unless complex analyses are done (49). In contrast, two-dimensional echocardiograms are positive in nearly all patients who have mitral valve prolapse that is associated with severe mitral regurgitation or other complications (54, 79). Currently recommended echocardiographic criteria for diagnosis of mitral valve prolapse are shown schematically in Figure August 1989 Annals of Internal Medicine Volume 111 Number 4

5 Other Diagnostic Methods Although of limited clinical utility, pathologic evaluation of mitral valve prolapse continues to yield important insights. It has recently been shown, for example, that lack of chordal support to segments of one or both mitral leaflets may allow regional deformity and prolapse without overall valve enlargement (55, 57). Findings in a large series reported by Roberts and colleagues (53) suggest that this chordal deficiency may contribute more often than actual rupture of chordae tendineae to the development of severe mitral regurgitation in prolapse patients with marked mitral-leaflet and anular enlargement. Thus, both overall valve enlargement and regional "balloon" deformity should be considered when making a gross anatomic diagnosis of mitral prolapse; also, it is now accepted that histologic examination showing disruption of collagen bundles reflects the primary (still unknown) defect and that the commonly observed myxomatous infiltration and endocardial fibrosis appear to be secondary responses to stress and trauma of the distorted valves (52). Although contrast angiography played a historic role in the initial elucidation of mitral valve prolapse, it is now considered less accurate for this purpose than echocardiography because of poor inter-observer reproducibility (100) and the frequent difficulty of differentiating mild mitral valve prolapse from the numerous variants of normal angiographic appearance of the mitral valve (101). Clinical Features of Mitral Valve Prolapse Mitral valve prolapse was first recognized by its auscultatory features and by angiographic and echocardiographic evidence of abnormal mitral valve motion, but studies (19, 24, 29, 32, 33) appeared soon thereafter that reported a high prevalence of nonanginal chest pain, dyspnea, and anxiety-related symptoms in patients with mitral prolapse. The concept of an inclusive "mitral prolapse syndrome" has proved to be clinically useful because it provides an explanation for common, troublesome, and otherwise confusing cardiovascular and psychologic symptoms that is acceptable to patients and physicians alike (102). Although mitral valve prolapse has been associated with many clinical features in studies of highly selected patients (24, 103), controlled studies have not supported this association. These studies have assessed the strength of association between mitral valve prolapse and clinical features of the "mitral prolapse syndrome" by comparing similarly selected or unselected control subjects and patients with echocardiographically documented mitral valve prolapse (39, 43, 44, 93, ). Controlled studies have substantiated the high prevalence of cardiac and psychiatric symptoms in patients with mitral valve prolapse who were evaluated in tertiary-care hospitals (43, 44). However, chest pain, dyspnea, psychological symptoms, and prolongation of the electrocardiographic Q-T interval occurred equally often among patients with mitral valve pro- Figure 4. Schematic diagram showing currently accepted echocardiography criteria for diagnosis of mitral valve prolapse. Top. Twodimensionally targeted M-mode recordings of continuous mitral leaflet interfaces show {top left) late-systolic prolapse, with prolapse beginning in midsystole {P) and characterized by at least a 2-mm posterior displacement of leaflets behind the valve's C-D line; holosystolic prolapse {top right) is characterized by a 3-mm displacement of leaflets behind the C-D line and confirmed by demonstration of leaflet billowing in the two-dimensional, parasternal long-axis view. Bottom. Two-dimensional, parasternal long-axis view showing systolic billowing of mitral leaflets {arrows) into the left atrium {LA), a motion the posterior component of which may be detected by the vertically oriented M-mode beam (M). Hatched areas indicate muscular walls of left ventricular myocardium, and dotted lines show normal position of mitral leaflets and annulus. Ao = aorta; LV = left ventricle. lapse and cardiovascularly normal individuals who were evaluated in the same clinical setting (43, 104, 108). Furthermore, mitral valve prolapse appears to be no commoner in patients with panic and anxiety disorders than in control subjects when similar precautions are taken (105, 107, 109, 110). Similarly, persons with and without echocardiographic mitral valve prolapse in the Framingham general population sample were equally likely to experience chest pain or dyspnea or to have exercise intolerance or repolarization abnormalities in the electrocardiogram (3, 39). Our own studies have taken advantage of the fact that in an inherited condition like mitral valve prolapse one can efficiently obtain information about the clinical features that are truly associated with it, free of the selection biases that influence referral of patients to university hospitals, by comparing affected firstdegree relatives (relatively unselected individuals who have the condition) with the unaffected relatives and spouses of relatives in the families of affected patients (who constitute genetically related and unrelated control groups who share similar environments). In the first 88 families we studied, echocardiographic mitral valve prolapse was present in 81 adult, first-degree rel- 15 August 1989 Annals of Internal Medicine Volume 111 Number 4 309

6 Table 2. Association between Clinical Features of the "Mitral Prolapse Syndrome" and Echocardiographically Documented Mitral Valve Prolapse* Clinical Feature First-Degree Adult First-Degree Adult PValuef Relatives with Relatives and Spouses Mitral Valve Prolapse without (n = 81) Mitral Valve Prolapse (n = 232) /2(%) n(%) Systolic clicks, 55(68) 17 (7) < mitral murmurs, or both Thoracic bony abnormalities 33(41) 34(15) < Body weight < 90% of ideal 26(32) 29(13) < Systolic blood pressure < 120 mm Hg 43(53) 65(28) < Palpitations 32(40) 53(23) < 0.01 Atypical chest pain 14(17) 37(16) NS Dyspnea 5 (6) 21 (9) NS Panic attacks 6 (7) 11 (5) NS Trait anxiety score > 50 5 (6) 14 (6) NS Inferior-lead electrocardiographic 9(11) 23(10) NS repolarization abnormalities * Adapted from Devereux et al. (44). t NS = not significant (P > 0.05). atives and absent in 172 other relatives and 60 spouses (44). When the affected relatives were compared with the control groups (Table 2), the expected strong association between echocardiographic and auscultatory features of mitral valve prolapse was observed, as were less close but still highly significant associations between mitral prolapse and thoracic bony abnormalities, low body weight, low systolic blood pressure, and palpitations, in accord with results from other controlled studies (2, 39, 93, 111). In contrast, there was no difference in the rate of nonanginal chest pain, dyspnea, panic attacks, high levels of anxiety, or electrocardiographic repolarization abnormalities between patients with mitral prolapse and control subjects; however, even a large study such as this cannot exclude the possibility of weak associations. In a separate study of 134 subjects (112), we showed that mitral valve prolapse and panic attacks were associated with contrasting patterns of autonomic dysfunction. When compared with control subjects, a significantly greater proportion of the patients with mitral prolapse showed orthostatic hypotension (17% compared with 3%, P < 0.01) and syncope (20% compared with 0%, P < 0.01), which were possibly related to reduced blood volume (113); however, the group with panic attacks exhibited hyper-reactive heart rate and blood pressure increases in response to orthostatic stress. These data indicate that mitral valve prolapse and panic disorders are biologically distinct in addition to being statistically unassociated (44, 110). The prevalence of autonomic dysfunction in cases of mitral valve prolapse appears low as no significant differences existed between persons with mitral prolapse and normal persons in two smaller studies (114, 115). Thus, carefully controlled studies show the spectrum of clinical features associated with mitral valve prolapse to be narrower than previously thought, but even features truly associated with mitral prolapse, such as thoracic bony abnormalities, low body weight, and palpitations, are not sufficiently specific to be useful in the diagnosis of mitral valve prolapse. Furthermore, we have found that patients who had nonspecific symptoms that led to consideration of mitral valve prolapse are particularly likely to have "false-positive" diagnoses unsupported by objective auscultatory or echocardiographic evidence of mitral prolapse (82). Two factors appear to account for the mistaken impression that prolapse of the mitral valve is associated with a spectrum of nonspecific symptoms. The first may be considered a special case of the more general problem of ascertainment or selection bias (102, 116). Common conditions are more likely to be diagnosed in symptomatic than in asymptomatic persons with the same condition, and among the symptomatic patients, those most severely affected are more likely to be referred for specialized investigation and care. As a result, patients cared for by physicians who carry out clinical research are much more likely to have severe symptoms and signs of their disease, whether it be mitral valve prolapse or not. This phenomenon may exaggerate or even artifactually create associations between diseases and clinical features that influence either the likelihood of the disease being diagnosed or the pattern of subsequent medical care. A second factor shown by our research (44, 117) is that women in the family units were more likely than men, irrespective of the presence or absence of mitral prolapse, to experience atypical chest pain, dyspnea, panic attacks, high-trait anxiety, or to have repolarization abnormalities shown by inferior-lead electrocardiography (Table 3). Complications of Mitral Valve Prolapse Soon after the modern recognition of mitral valve prolapse, clinical observations suggested that patients with mitral valve prolapse were at risk for infective endocarditis (20-23), mitral regurgitation (19, 25), serious arrhythmias (26, 27), sudden death (21), and stroke (22). However, studies of single cases or small August 1989 Annals of Internal Medicine Volume 111 Number 4

7 series are subject to the same type of selection bias described above, which may result in the reporting of interesting coincidences along with true associations. Association of Complications with Mitral Valve Prolapse Mitral valve prolapse has been found in a higher proportion of patients with complications than would be expected from its prevalence of approximately 4% in unselected populations (3, 7, 70, 93, 94). In recent studies from industrialized countries, from 38% to 64% of patients with severe, pure mitral regurgitation have had mitral valve prolapse as the underlying cause (7, 8, ), whereas the proportion ranged from 11% to 29% in patients with infective endocarditis (4-7, 45, ). The data on patients who had neurologic ischemic episodes, suspected to have been caused by fibrin-platelet emboli formed on roughened valve surfaces, have been quite variable; mitral valve prolapse was found by echocardiogram in 2% to 35% of patients ( ), and the evidence suggested that the risk attributable to mitral prolapse is so low that it is discernible only in subjects without other factors predisposing to cerebrovascular disease (131). The situation concerning sudden death is even more controversial; its occurrence in patients with mitral prolapse but without coronary disease has been well documented at autopsy (132), but such patients accounted for only about 1 % of sudden deaths in the forensic series reported by Davies and colleagues (50), which falls below the expected prevalence in the general population. However, this result may still be consistent with a true association between sudden death and mitral prolapse, because mitral valve prolapse appeared to occur in a disproportionate number of the small minority of patients who were free of obstructive coronary artery disease and had sudden death (50, 133). In the families we have studied, four instances of sudden death have occurred in persons who did not have significant coronary artery obstruction: Two young women with uncomplicated mitral prolapse and two men over 50 years old with severe mitral regurgitation or rupture of mitral chordae tendineae had sudden death. Further evidence linking mitral valve prolapse to various complications has been derived from follow-up of patients evaluated in major medical centers (78, Table 3. Findings from the Cornell Family Study on the Predominance in Women of Features of the "Mitral Prolapse Syndrome"* Women Men P Value (n = 216) (n = 185) /2(%) n(%) Nonanginal chest pain 63(29) 24(13) < Dyspnea 50(23) 15 (8) < Panic attacks 29(13) 4 (2) < High trait anxiety 23(11) 6 (3) < 0.01 Inferior-lead ST-T 44(20) 12 (6) < abnormalities * Adapted from data of Devereux et al. (44). Table 4. Annual Occurrence in the United States of the Complications Associated with Mitral Valve Prolapse* Complication Patients Patients Annual per Year with Events Mitral Valve Attribut- Prolapse able to Mitral Valve Prolapse n % n Mitral valve surgery Infective endocarditis Sudden deathf * For sources of data on which these calculations are based, see text. f Figures for sudden death are less stable than for other complications because they are based on a single study (50) ). The prognosis of most patients in all studies has been benign, but serious complications-including infective endocarditis, progressive mitral regurgitation resulting in valve replacement or congestive heart failure, stroke, and sudden death-have occurred in 3% to 33% of adult patients over mean intervals of 6 to 14 years (78, 134, 135, 137). These results have been confirmed by the finding of an appreciable proportion of patients who had evidence of severe mitral regurgitation among series of patients with mitral valve prolapse at autopsy (50, 52), although in one study the average age of death in patients with mitral prolapse actually tended to be greater than that of patients without this condition (50). It is now possible to estimate the aggregate number of persons with mitral valve prolapse who suffer several of its complications annually in the United States. Because mitral prolapse usually has onset during late childhood or adolescence, and because complications are rare in childhood, calculations are restricted to adults. To make sound estimates, one needs reliable estimates of the total number of patients who have each complication annually in the United States and the percentage of patients in whom each complication is apparently caused by mitral valve prolapse. Sufficient information (Table 4) is available to make reasonable estimates for mitral regurgitation (119, 138, 139), infective endocarditis (3-7, 45, 122, 123, ), and, with less certainty, sudden death (50, 141, 142). Relative and Absolute Risk for Complications of Mitral Valve Prolapse To turn knowledge of associations between mitral valve prolapse and certain complications into a basis for the practical management of patients, it is necessary to estimate the actual risk for such complications and to identify potentially preventable or controllable factors that amplify this risk. For this purpose, we must rely on the results of controlled studies, which, to date, have been cross-sectional ones done using case-control methodology. The risk for infective endocarditis in patients with mitral valve prolapse has been the most extensively studied, and many investigators estimate its likelihood to be 3 to 8 times higher in persons with mitral valve 15 August 1989 Annals of Internal Medicine Volume 111 Number 4 311

8 prolapse than in the general population (4-7, 45, 124). Barnett and colleagues (143) reported that the risk for neurologic ischemic episodes was increased approximately sixfold in women under the age of 40 years who had mitral valve prolapse. Even more striking increases in risk have been suggested by our calculations that mitral valve prolapse was associated with approximately 30- to 40-fold increases in the likelihood of developing hemodynamically important mitral regurgitation or ruptured mitral chordae tendineae (7), and that patients with mitral valve prolapse who had severe mitral regurgitation were at 50 to 100 times the average risk for sudden cardiac death (142); such risk is probably related to the concentration of ventricular dysfunction and complex ventricular arrhythmias in this subgroup (144). Estimates of risk are of greatest use when they identify subgroups of patients in particular need of close surveillance or preventive treatment. By comparing the characteristics of 21 patients with mitral valve prolapse who had infective endocarditis with those of 102 adult relatives or spouses found to have mitral valve prolapse in our family studies, we showed that male gender, an age of at least 45 years, and a history of a pre-existing heart murmur were independently associated with infective endocarditis (6). The average incidence of infective endocarditis is 1 case per persons per year in the general population ( ); we estimated that infective endocarditis would occur each year in 1 of 1920 patients with mitral valve prolapse who had a late or holosystolic murmur of mitral regurgitation compared with 1 of who did not have a mitral systolic murmur. Similar calculations would suggest annual incidences of infective endocarditis of 1 in 3640 for affected men and 1 in 2930 for persons 45 years of age or older with mitral valve prolapse. That major morbidity occurred in one third of our prolapse patients with endocarditis during shortterm follow-up (3 deaths, 4 valve replacements) and that endocarditis appeared to be of dental origin in 7 of 21 patients suggests that infective endocarditis as a complication of mitral valve prolapse is more dangerous and more preventable than previously thought (40-42, 145). Fewer data are available concerning other complications of mitral valve prolapse. In a single, controlled study (121) we found the risk for hemodynamically important mitral regurgitation in persons with mitral valve prolapse to be increased by male gender, an age of 45 years or greater, and lack of the low body weight and low systolic blood pressure commonly associated with this condition. These conclusions are supported by other studies (7, 53, , ) that reported a striking male predominance (247 of 356, 69%) among patients with mitral valve prolapse who had severe mitral regurgitation, and by the association between elevated blood pressure and the rupture of chordae of the prolapsed valves reported by Roberts and colleagues (149, 150). Wilken and Hickey (10) have calculated that the lifetime risk for needing mitral valve replacement is approximately 4% among men and 1.5% among women with mitral valve prolapse in Australia, and we have estimated these risks at 5.5% and 1.5% respectively, based on data from the United States (151). For sudden, presumably arrhythmic, death, we have estimated the annual risk at between 1 in 53 and 1 in 106 in patients with mitral valve prolapse who have important mitral regurgitation but at only about 1 in 5400 in patients with little or no mitral regurgitation (142). The risk for sudden death associated with mitral regurgitation, independent of cause, appears to be concentrated among patients with repetitive ventricular arrhythmias who also have depressed ventricular function (152), which is consistent with the primary importance of depressed ventricular function in determining the risk imparted by complex arrhythmias documented in other conditions (153, 154). Evaluation and Management of Mitral Valve Prolapse Recommended approaches to detection and management of mitral valve prolapse have tended to oscillate between the opposite poles of a nihilism based on the premise that mitral prolapse is ubiquitous and benign and an intense concern derived from recognition of frequent complications in highly selected groups of patients (155, 156). The findings reviewed above suggest that an intermediate course is both appropriate and practical. Auscultation Auscultation remains the commonest method by which mitral valve prolapse is first recognized. When both a midsystolic click and late-systolic murmur are present and vary appropriately in timing and intensity with maneuvers (Figure 1) or a loud midsystolic click shows appropriate mobility, the diagnosis may be considered definitive. If other less specific auscultatory features such as a soft or immobile midsystolic click or a late-systolic murmur in a middle-aged or older person prompt consideration of mitral valve prolapse, echocardiographic confirmation is desirable. Echocardiographic diagnosis of mitral valve prolapse should be based on either demonstration of at least a 2-mm, late-systolic, posterior displacement of continuous mitral leaflet interfaces by two-dimensionally targeted M-mode recordings or of unequivocal systolic billowing of one or both mitral leaflets across the mitral anulus in the parasternal or apical, long-axis view (35, 36). Echocardiography The role of echocardiography in screening for mitral valve prolapse in unselected populations or symptomatic patients without typical auscultatory features can be placed in perspective by simple calculations. Based on the finding of typical midsystolic clicks or late or holosystolic murmurs on a single careful examination in roughly two thirds of relatives with mitral valve prolapse in our family studies (44) and an approximate population prevalence of 3% to 4% for this condition (7, 36, 93, 94), only 1% to 1.5% of unse August 1989 Annals of Internal Medicine Volume 111 Number 4

9 lected adults or patients with nonspecific cardiovascular or psychiatric symptoms would have auscultatorily "silent" mitral valve prolapse; this proportion would rise to only 2% to 3% in persons with symptomatic palpitations, a group in which the prevalence of mitral prolapse is modestly increased. Because the complications of mitral valve prolapse are most prevalent in persons with mitral regurgitation, echocardiographic screening to detect occasional persons with low-risk forms of the condition is not cost-effective unless objective testing is required to expunge a dubious diagnosis and free the patient from unfounded concerns about heart disease and unwarranted treatment. Conversely, echocardiographic screening of first-degree relatives of patients with unequivocal mitral valve prolapse is likely to be cost-effective as 30% of such persons were affected in our family studies (36, 37, 44). Risk Level and Management Management of the patient with mitral valve prolapse should be matched to the risk level for both infective endocarditis and progressive mitral regurgitation in the particular patient. Because the risk for these complications is related to the presence of at least mild mitral regurgitation (6, 9, 45, 124), no specific treatment may be needed for patients with mitral valve prolapse, particularly women under the age of 45 who do not have a mitral systolic murmur on any of several examinations using auscultation in multiple positions and isometric handgrip exercise, or evidence of mitral regurgitation by Doppler echocardiography. We reassure such persons that the outlook is benign and may even be enhanced if they have the commonly associated low body weight and low blood pressure (39, 44, 93, 106, 157); antibiotic prophylaxis is not routinely recommended unless the person wishes maximum protection against even the remotest risk; and re-evaluation by auscultation and echocardiogram is recommended at moderate intervals (perhaps every 5 years) to be certain the patient has not passed into a higherrisk group. Based on present evidence, patients with echocardiographic mitral valve prolapse who have soft, latesystolic murmurs of mitral regurgitation even intermittently or during standard maneuvers (Figure 1) constitute a group at discernibly, albeit modestly, increased risk for endocarditis or progressive mitral regurgitation. We recommend antibiotic prophylaxis for such patients: Before dental procedures, penicillin should be given to women with consistently audible mitral murmurs and men who have even intermittent evidence of mitral regurgitation; and erythromycin (possibly less effective than penicillin but without risk for inducing anaphylaxis [40]) should be given to women with intermittent, mitral systolic murmurs and to older men with isolated, midsystolic clicks on examinations as well as to all patients with penicillin allergy. In view of evidence suggesting that elevated blood pressure may predispose to chordal rupture and progressive mitral regurgitation in patients with mitral valve prolapse (149, 150) and the possibility that known differences in blood pressure between genders (158, 159) may underlie the male predominance among patients with mitral valve prolapse and severe mitral regurgitation, we recommend antihypertensive treatment for all patients in this group who have even very mild, established systemic hypertension. Doppler echocardiography is an important adjunct to imaging techniques because it defines precisely the extent of mitral regurgitation, and this evaluation as well as auscultatory examination is warranted at more frequent intervals (perhaps every 3 years) to assess possible progression of mitral regurgitation. High-Risk Cases The greatest risk for endocarditis, sudden death, and need for mitral valve surgery is concentrated among patients with mitral valve prolapse who have hemodynamically important mitral regurgitation; such patients constitute approximately 2% to 4% of affected adults (3, 6, 142, 157). On physical examination, severe mitral regurgitation is suggested by a holosystolic or nearly holosystolic mitral regurgitant murmur, commonly accompanied by a left ventricular third heart sound and leftward displacement of a dynamic left ventricular impulse. A diagnosis can be confirmed by the demonstration of significant mitral regurgitation by pulsed and color flow Doppler echocardiography (160, 161) in conjunction with imaging echocardiographic evidence of mitral valve prolapse and left heart-chamber enlargement. Prophylaxis for infective endocarditis is mandatory; penicillin should be used in the absence of a specific allergy, and it is theoretically attractive, although not of proven value, to treat even borderline systemic hypertension with antihypertensive drugs in such patients. Regular follow-up is required; annual examinations using both imaging and Doppler echocardiography and selected other methods such as nuclear angiography and treadmill exercise tests are recommended. Corrective valvular surgery, either valve replacement or repair (which is being used in an increasing proportion of cases), is recommended when patients either develop dyspnea of a class II or greater severity (New York Heart Association classification) or when left ventricular systolic function falls into the lower part of the normal range in the absence of symptoms. As reported by Zile and colleagues (162) and confirmed in our laboratory (163), a simple partition value for recognition of the latter is an M-mode echocardiographic, left ventricular fractional shortening of less than 31%, which can be used to predict a suboptimal outcome after mitral valve replacement for severe mitral regurgitation. Clearly subnormal ventricular function should not preclude corrective valvular surgery, which may improve the poor survival associated with medical management of patients with severe mitral regurgitation and ventricular dysfunction (164); the benefits may be similar to those reported for cardiac surgery in other high-risk groups (165). Arrhythmias Patients with mitral valve prolapse who experience arrhythmias may require treatment to relieve symptoms 15 August 1989 Annals of Internal Medicine Volume 111 Number 4 313

10 or to reduce risk of sudden death. Palpitations and salvos of atrial premature complexes and brief bursts of atrial tachycardia are slightly more common in patients with mitral prolapse than in normal persons (166, 167). Suggested mechanisms of arrhythmogenesis include stimulation of atrial pacemakers by the impact of prolapsing leaflets or mitral regurgitant jets, and origin of impulses from electrically active cells, recently shown to have beta adrenoceptors, in the mitral leaflets (168, 169). However, the coincidence of symptoms and atrial arrhythmia is unusual. Awareness of palpitation in other patients with mitral valve prolapse may coincide with simple ventricular premature complexes, but the prevalence of ventricular arrhythmias in controlled studies of patients with mitral valve prolapse is not strikingly higher than in normal subjects (142, 166, 167). In many cases, arrhythmias will respond to treatment with beta-blocking drugs, although untreated persons may have periods of remission during which symptoms are not experienced. Some patients with atrial arrhythmias may respond favorably to digitalization. Whether and when to use antiarrhythmic drugs to prevent sudden death in patients with mitral valve prolapse remains controversial (142). Sudden death appears to occur most often in the 2% to 4% of patients with hemodynamically severe mitral regurgitation (50, 137, 142, 152), but even in this high-risk group no evidence exists that treatment averts sudden death. The rare occurrence of arrhythmic death among the larger population of persons with otherwise uncomplicated mitral prolapse ( or more adults in the United States) poses an even more difficult problem in the absence of risk factors accurately predictive of arrhythmic death. Although sudden death is generally an arrhythmic event, complex ventricular arrhythmias or abnormal ventricular repolarization have not been shown to predict sudden death in patients with uncomplicated mitral valve prolapse. Because at least 5% to 10% of these persons, even when asymptomatic, will be found to have ventricular couplets, salvos, or ventricular tachycardia by ambulatory electrocardiography (166, 167), evaluation and treatment of these arrhythmias is a major clinical problem. Based on the estimated incidence of sudden death in persons with uncomplicated mitral prolapse, under ideal circumstances over 300 such persons with repetitive ventricular arrhythmias would require treatment to prevent one sudden death each year (142). This estimate assumes, questionably, that complex ventricular arrhythmias accurately predict mortality in these subjects, that no offsetting proarrhythmic effect of drug therapy would occur, and that their suppression by drugs prevents sudden death-a benefit that has yet to be shown in patients with mitral prolapse. These estimates argue against the use of empiric antiarrhythmic therapy to prevent sudden death in patients with uncomplicated mitral valve prolapse. However, we are acutely aware of the distinction between population estimates and clinical management of persons in practice. Arrhythmia suppression with betablocking drugs may be tried in patients with complex arrhythmias, when tolerated, but empiric treatment with drugs that have more frequent proarrhythmic effects or other untoward consequences should probably be avoided. When the need for treatment is further suggested by an ominous symptom such as syncope or a finding such as sustained ventricular tachycardia, this should be instituted under careful monitoring, or electrophysiologic guidance should be sought. Other Problems Management of the patient with mitral valve prolapse and symptoms other than palpitations or dyspnea that are related to mitral regurgitation may require various approaches. When chest pain, palpitations, and dyspnea are concurrent with repeated episodes of severe anxiety termed "panic attacks," treatment directed toward either pharmacologic or behavioral therapy of panic disorder under the guidance of an experienced psychiatrist is often effective, whereas use of standard cardiac medications generally is not (110). The occasional patient with mitral valve prolapse who has recurrent dizziness or even syncope due to orthostatic hypotension (31, 112, 170) usually has a reduced blood volume (113). Both the deficit in blood volume and the resultant symptoms may be alleviated in some instances by treatment with clonidine (171) or by the combination of sodium supplementation and fiuorinef. Careful evaluation of the patient with atypical chest pain may reveal a treatable noncardiac cause such as abnormal esophageal motility (172). Acknowledgments: The authors thank Virginia Burns for assistance in manuscript preparation. Requests for Reprints: Richard B. Devereux, MD, Division of Cardiology, Box 222, The New York Hospital-Cornell Medical Center, 525 East 68th Street, New York, NY Current Author Addresses: Drs. Devereux and Kligfield, and Ms. Kramer-Fox: New York Hospital-Cornell Medical Center, 525 East 68th Street, New York, NY References 1. Barlow JB, Pocock WA, Marchand P, Denny M. The significance of late systolic murmurs. Am Heart J. 1963;66: Savage DD, Garrison RJ, Devereux RB, et al. Mitral valve prolapse in the general population. 1. Epidemiologic features: the Framingham Study. Am Heart J. 1983;106: Levy D, Savage D. Prevalence and clinical features of mitral valve prolapse. Am Heart J. 1987;113: Clemens JD, Horwitz RI, Jaffe CC, Feinstein AR, Stanton BF. A controlled evaluation of the risk of bacterial endocarditis in persons with mitral valve prolapse. N EnglJ Med. 1982;307: Hickey A J, MacMahon SW, Wilcken DE. Mitral valve prolapse and bacterial endocarditis: when is antibiotic prophylaxis necessary? Am Heart J. 1985;109: MacMahon SW, Roberts JK, Kramer-Fox R, Zucker DM, Roberts RB, Devereux RB. Mitral valve prolapse and infective endocarditis. Am Heart J. 1987;113: Devereux RB, Hawkins I, Kramer-Fox R, et al. Complications of mitral valve prolapse. Disproportionate occurrence in men and older patients. Am J Med. 1986;81: Guy FC, MacDonald RP, Fraser DB, Smith ER. Mitral valve prolapse as a cause of hemodynamically important mitral regurgitation. Can J Surg. 1980;23: Kolibash AJ Jr, Kilman JW, Bush CA, Ryan JM, Fontana ME, Wooley CF. Evidence for progression from mild to severe mitral regurgitation in mitral valve prolapse. Am J Cardiol. 1986;58: Wilcken DE, Hickey A J. Lifetime risk for patients with mitral August 1989 Annals of Internal Medicine Volume 111 Number 4

11 prolapse of developing severe valve regurgitation requiring surgery. Circulation. 1988;78: Criley JM, Lewis KB, Humphries JO, Ross RS. Prolapse of the mitral valve: clinical and cine-angiographic findings. Br Heart J. 1966;28: Ranganathan N, Silver MD, Robinson TI, et al. Angiographicmorphologic correlation in patients with severe mitral regurgitation due to prolapse of the posterior mitral valve leaflet. Circulation. 1973;48: Fontana ME, Pence HL, Leighton RF, Wooley CF. The varying clinical spectrum of the systolic click - late systolic murmur syndrome. Circulation. 1970;41: Epstein EJ, Coulshed N. Phonocardiogram and apex cardiogram in systolic click-late systolic murmur syndrome. Br Heart J. 1973;35: Fontana ME, Wooley CF, Leighton RF, Lewis RP. Postural changes in left ventricular and mitral valve dynamics in the systolic click - late systolic murmur syndrome. Circulation. 1975;51: Devereux RB, Perloff JK, Reichek N, Josephson ME. Mitral valve prolapse. Circulation. 1976;54: Kerber RE, Isaeff DM, Hancock EW. Echocardiographic patterns in patients with the syndrome of systolic click and late systolic murmur. N Engl J Med. 1971;284: Dillon JC, Haine CL, Chang S, Feigenbaum H. Use of echocardiography in patients with prolapsed mitral valve. Circulation. 1971;43: Hancock EW, Cohn K. The syndrome associated with midsystolic click and late systolic murmurs. Am J Med. 1966;41: LeBauer EJ, Perloff JK, Keliher TF. The isolated systolic click with bacterial endocarditis. Am Heart J. 1967;73: Shappell SD, Marshall CE, Brown RE, Bruce TA. Sudden death and the familial occurrence of mid-systolic click, late systolic murmur syndrome. Circulation. 1973;48: Barnett HJ, Jones MW, Broughner D. Cerebral ischemic events associated with prolapsing mitral valve. Trans Am Neurol Assoc. 1975;100: Lachman AS, Bramwell-Jones DM, Lakier JB, Pocock WA, Barlow JB. Infective endocarditis in the billowing leaflet syndrome. Br Heart J. 1975;37: Malcolm AD, Bougher DR, Kostuk WJ, Ahuja SP. Clinical features and investigative findings in presence of mitral leaflet prolapse. Study of 85 consecutive patients. Br Heart J. 1976;38: Higgins CB, Reinke RT, Gosink BB, Leopold GR. The significance of mitral valve prolapse in middle-aged and elderly men. Am Heart J. 1976;91: Winkle RA, Lopes MG, Popp RL, Hancock EW. Life-threatening arrhythmias in the mitral valve prolapse syndrome. Am J Med. 1976;60: DeMaria AN, Amsterdam EA, Vismara LA, Neumann A, Mason DT. Arrhythmias in the mitral valve prolapse syndrome. Prevalence, nature and frequency. Ann Intern Med. 1976;84: Campbell RW, Godman MG, Fiddler GI, Marquis RM, Julian DG. Ventricular arrhythmias in syndrome of balloon deformity of mitral valve. Definition of possible high risk group. Br Heart J. 1976;38: Pariser SF, Pinta ER, Jones BA. Mitral valve prolapse syndrome and anxiety neurosis/panic disorder. Am J Psychiatry. 1978; 135: Coghlan HC, Phares P, Cowley M, Copley D, James TN. Dysautonomia in mitral valve prolapse. Am J Med. 1979;67: Santos AD, Mathew PK, Hilal A, Wallace WA. Orthostatic hypotension: a commonly unrecognized cause of symptoms in mitral valve prolapse. Am J Med. 1981;71: Wooley CF. Where are the diseases of yesteryear? DaCosta's syndrome, soldiers heart, the effort syndrome, neurocirculatory asthenia-and the mitral valve prolapse syndrome. Circulation. 1976;53: Wooley CF. The mitral valve prolapse syndrome. Hosp Pract [Off]. 1983;18:163-9, Boudoulas H, Wooley CF. Mitral valve prolapse and the mitral valve prolapse syndrome. Progr Cardiol. 1986;14: Levine RA, Triulzi MO, Harrigan P, Weyman AE. The relationship of mitral annular shape to the diagnosis of mitral valve prolapse. Circulation. 1987;75: Devereux RB, Kramer-Fox R, Shear MK, Kligfield P, Pini R, Savage DD. Diagnosis and classification of severity of mitral valve prolapse: methodologic, biologic, and prognostic considerations. Am Heart J. 1987;113: Devereux RB, Brown WT, Kramer-Fox R, Sachs I. Inheritance of mitral valve prolapse: effect of age and sex on gene expression. Ann Intern Med. 1982;97: Strahan NV, Murphy EA, Fortuin NJ, Come PC, Humphries JO. Inheritance of the mitral valve prolapse syndrome. Discussion of a three-dimensional penetrance model. Am J Med. 1983;74: Savage DD, Devereux RB, Garrison RJ, et al. Mitral valve prolapse in the general population. 2. Clinical features: the Framingham study. Am Heart J. 1983;106: Bor DH, Himmelstein DU. Endocarditis prophylaxis for patients with mitral valve prolapse. A quantitative analysis. Am J Med. 1984;76: Retchin SM, Fletcher RH, Waugh RA. Endocarditis and mitral valve prolapse: what is the "risk"? Int J Cardiol. 1984;5; Clemens JD, Ransohoff DF. A quantitative assessment of pre-dental antibiotic prophylaxis for patients with mitral valve prolapse. JChronDis. 1984;37: Retchin SM, Fletcher RH, Earp J, Lamson N, Waugh RA. Mitral valve prolapse. Disease or illness? Arch Intern Med. 1986; 146: Devereux RB, Kramer-Fox R, Brown WT, et al. Relation between clinical features of the mitral prolapse syndrome and echocardiographically documented mitral valve prolapse. / Am Coll Cardiol. 1986;8: MacMahon SW, Hickey AJ, Wilcken DE, Wittes JT, Feneley MP, Hickie JB. Risk of infective endocarditis in mitral valve prolapse with and without precordial systolic murmurs. Am J Cardiol. 1987;59: Webster's Third New International Dictionary. Springfield, Massachusetts: G. & C. Merriam Co.; 1967: Davies MJ. Pathology of Cardiac Valves. London: Butterworth; 1980: Ranganathan N, Lam JH, Wigle ED, Silver MD. Morphology of the human mitral valve. II. The valve leaflets. Circulation. 1970;41: Pini R, Greppi B, Kramer-Fox R, Roman MJ, Devereux RB. Mitral valve dimensions and motion and familial transmission of mitral valve prolapse with and without mitral leaflet billowing. J Am Coll Cardiol. 1988;12: Davies MJ, Moore BP, Braimbridge MV. The floppy mitral valve. Study of incidence, pathology, and complications in surgical, necropsy, and forensic material. Br Heart J. 1978;40: King BD, Clark MA, Baba N, Kilman JW, Wooley CF. "Myxomatous" mitral valves: collagen dissolution as the primary defect. Circulation. 1982;66: Lucas RV Jr, Edwards JE. The floppy mitral valve. Curr Probl Cardiol. 1982;7: Roberts WC, Mcintosh CL, Wallace RB. Mechanisms of severe mitral regurgitation in mitral valve prolapse determined from analysis of operatively excised valves. Am Heart J. 1987;113: Pini R, Devereux RB, Greppi B, et al. Comparison of mitral valve dimensions and motion in mitral valve prolapse with severe mitral regurgitation to uncomplicated mitral valve prolapse and to mitral regurgitation without mitral valve prolapse. Am J Cardiol. 1988;62: van der Bel-Kahn J, Duren DR, Becker AE. Isolated mitral valve prolapse: chordal architecture as an anatomic basis in older patients. J Am Coll Cardiol. 1985;5: Hutchins GM, Moore GW, Skoog DK. The association of floppy mitral valve with disjunction of the mitral annulus fibrosus. N Engl J Med. 1986;314: Virmani R, Atkinson JB, Byrd BF 3d, Robinowitz M, Forman MB. Abnormal chordal insertion: a cause of mitral valve prolapse. Am Heart J. 1987;113: Pyeritz RE, Wappel MA. Mitral valve dysfunction in the Marfan syndrome. Clinical and echocardiographic study of prevalence and natural history. Am J Med. 1983;74: Devereux RB, Brown WT. Structural heart disease. In: King RA, Motulsky AG, Rotter JI, eds. The Genetic Basis of Common Disease. New York: Oxford University Press; (In press). 60. Leier CV, Call TD, Fulkerson PK, Wooley CF. The spectrum of cardiac defects in the Ehlers-Danlos syndrome, types I and III. Ann Intern Med. 1980;92: Schreiber TL, Feigenbaum H, Weyman AE. Effect of atrial septal defect repair on left ventricular geometry and degree of mitral valve prolapse. Circulation. 1980;61: Meyers DG, Starke H, Pearson PH, Wilken MK. Mitral valve prolapse in anorexia nervosa. Ann Intern Med. 1986;105: Lima SD, Lima JA, Pyeritz RE, Weiss JL. Relation of mitral valve prolapse to left ventricular size in Marfan's syndrome. Am J Cardiol. 1985;55: Roman MJ, Devereux RB, Kramer-Fox R, Spitzer MC. Comparison of cardiovascular and skeletal features of primary mitral valve prolapse and Marfan syndrome. Am J Cardiol. 1989;63: Peller OG, Devereux RB, Schreiber TL, McNulty A. Lack of association between acute myocardial infarction and mitral valve prolapse. Am J Cardiol. 1988;62: Weiss AN, Mimbs JW, Ludbrook PA, Sobel BE. Echocardiographic detection of mitral valve prolapse. Exclusion of false positive diagnosis and determination of inheritance. Circulation. 1975;52: Chen WW, Chan FL, Wong PH, Chow JS. Familial occurrence of 15 August 1989 Annals of Internal Medicine Volume 111 Number 4 315

12 mitral valve prolapse: is this related to the straight back syndrome. Br Heart J. 1983;50: Wilcken DE, Hickey A J, Cole WG, Chan D. The pathogenesis of the floppy mitral valve [Abstract]. Circulation. 1984;70(Suppl 2): Scheele W, Allen HN, Kraus R, Rubin PJ. Familial prevalence and genetic transmission of mitral valve prolapse [Abstract]. Circulation. 1977;56(Suppl3):lll. 70. Hickey AJ, Wilcken DE. Age and the clinical profile of idiopathic mitral valve prolapse. Br Heart J. 1986;55: Devereux RB, Kramer-Fox R. Inheritance and phenotypic features of mitral valve prolapse. In: Wooley CF, Boudoulas H, eds. Mitral Valve Prolapse and the Mitral Valve Prolapse Syndrome. Mt. Kisco, New York: Futura; 1988: Baker PB, Bansal G, Boudoulas H, Kolibash AJ, Kilman J, Wooley CF. Floppy mitral valve chordae tendineae: histopathologic alterations. Hum Pathol. 1988;19: Hammer D, Leier CV, Baba N, Vasko JS, Wooley CF, Pinnell SR. Altered collagen composition in a prolapsing mitral valve with ruptured chordae tendineae. Am J Med. 1979;67: Handler CE, Child AH. Increased prevalence of mitral valve prolapse associated with an elevated type 111:111 and I collagen ratio [Abstract]. Br Heart J. 1985;53: Lis Y, Burleigh MC, Parker DJ, Child AH, Hogg J, Davies MJ. Biochemical characterization of individual normal, floppy and rheumatic mitral valves. Biochem J. 1987;244: Cole WG, Chan D, Hickey AJ, Wilcken DE. Collagen composition of normal and myxomatous human mitral valves. Biochem J. 1984;219: Henney AM, Tsipouras P, Schwartz RC, Child AH, Devereux RB, Leech GJ. Genetic evidence that mutations in the COL1A2, COL3A1, or COL5A2 collagen genes are not responsible for mitral valve prolapse. Br Heart J. 1989;61: Nishimura RA, McGoon MD, Shub C, Miller FA Jr, Ilstrup DM, Tajik AJ. Echocardiographically documented mitral-valve prolapse. Longterm follow-up of 237 patients. N Engl J Med. 1985;313: Levine RA, Stathogiannis E, Newell JB, Harrigan P, Weyman AE. Reconsideration of echocardiographic standards for mitral valve prolapse: lack of association between leaflet displacement isolated to the apical four chamber view and independent echocardiographic evidence of abnormality. J Am Coll Cardiol. 1988;11: Reid JV. Mid-systolic clicks. S Afr Med J. 1961;35: Perloff JK, Child JS. Clinical and epidemiolgic issues in mitral valve prolapse. Overview and perspective. Am Heart J. 1987;113: Devereux RB, Kramer-Fox R, Shear MK, Lutas EM, Spitzer MC, Kligfield P. Relation of panic attacks and midsystolic murmurs to over-diagnosis of mitral valve prolapse. Clinical Cardiol. 1989; (In press). 83. Barron JT, Manrose DL, Liebson PR. Comparison of auscultation with two-dimensional and Doppler echocardiography in patients with suspected mitral valve prolapse. Clin Cardiol. 1988;11: Abascal VM, Hagege AA, Brady CL, Levine RA. Mitral valve prolapse: lack of sensitivity and specificity of clicks and murmurs for leaflet displacement by 2-D echocardiography [Abstract]. Circulation. 1987;(SupplIV): Shulman ST, Amren DP, Bisno AL, et al. Prevention of Bacterial Endocarditis. A statement for health professionals by the Committee on Rheumatic Fever and Infective Endocarditis of the Council on Cardiovascular Disease in the Young. Circulation. 1984; 70:1123A-7A. 86. Markiewicz W, Stoner J, London E, Hunt SA, Popp RL. Mitral valve prolapse in one hundred presumably healthy young females. Circulation. 1976;53: Levine RA, Weyman RA. Mitral valve prolapse: a disease in search of, or created by, its definition. Echocardiography. 1984;1: Devereux RB, Sachs I, Kramer-Fox R, Brown WT. Evolution of echocardiographic findings in patients with mitral prolapse. / Cardiovasc Ultrasonography. 1983;2: Wann LS, Gross CM, Wakefield RJ, Kalbfleisch JH. Diagnostic precision of echocardiography in mitral valve prolapse. Am Heart J. 1985;109: Alpert MA, Carney RJ, Munuswamy K, et al. Observer variation in the echocardiographic diagnosis of mitral valve prolapse. Am Heart J. 1986;111: Meltzer RS, Devereux RB, Goldman ME, et al. Observer variability in M-mode echocardiographic diagnosis of mitral valve prolapse: effect of tracing quality and depth of prolapse. / Cardiovasc Ultrasonography. 1987;6: Abbasi AS, DeCristofaro D, Anabtawi J, Irwin L. Mitral valve prolapse: comparative value of M-mode, two-dimensional and Doppler echocardiography. J Am Coll Cardiol. 1983;2: Hickey A J, Narunsky L, Wilcken DE. Bodily habitus and mitral valve prolapse. Aust N Z J Med. 1985;15: Bryhn M, Persson S. The prevalence of mitral valve prolapse in healthy men and women in Sweden. An echocardiographic study. Acta Med Scand. 1984;215: Gorman JM, Shear MK, Devereaux RB, King DL, Klein DF. Prevalence of mitral valve prolapse in panic disorder: effect of echocardiographic criteria. Psychosom Med. 1986;48: Gilbert BW, Schatz RA, VonRamm OT, Behar VS, Kisslo JA. Mitral valve prolapse. Two-dimensional echocardiographic and angiographic correlation. Circulation. 1976;54: Sahn DJ, Wood J, Allen HD, Peoples W, Goldberg SJ. Echocardiographic spectrum of mitral valve motion in children with and without mitral valve prolapse: the nature of false positive diagnosis. Am J Cardiol. 1977;39: Morganroth J, Jones RH, Chen CC, Naito M. Two dimensional echocardiography in mitral, aortic and tricuspid valve prolapse. The clinical problem, cardiac nuclear imaging considerations, and a proposed standard for diagnosis. Am J Cardiol. 1980;46: Warth DC, King ME, Cohen J, Tesoriero VL, Marcus E, Weyman AE. Prevalence of mitral valve prolapse in normal children. / Am Coll Cardiol. 1985;5: Kennett JD, Rust PF, Martin RH, Parker BM, Watson LE. Observer variation in the angiocardiographic diagnosis of mitral valve prolapse. Chest. 1981;79: Spindola-Franco H, Bjork L, Adams DF, Abrams HL. Classification of the radiological morphology of the mitral valve. Differentiation between true and pseudoprolapse. Br Heart J. 1980;44: Gottlieb SH. Mitral valve prolapse: from syndrome to disease. Am J Cardiol. 1987;60:53J-8J Beton DC, Brear SG, Edwards JD, Leonard JC. Mitral valve prolapse: an assessment of clinical features, associated conditions and prognosis. QJ Med. 1983;52: Uretsky BF. Does mitral valve prolapse cause nonspecific symptoms? Int J Cardiol. 1982;1: Hartman N, Kramer R, Brown WT, Devereux RB. Panic disorder in patients with mitral valve prolapse. Am J Psychiatry. 1982;139: Devereux RB, Brown WT, Lutas EM, Kramer-Fox R, Laragh JH. Association of mitral valve prolapse with low body-weight and low blood pressure. Lancet. 1982;2: Hickey A J, Andrews G, Wilcken DE. Independence of mitral valve prolapse and neurosis. Br Heart J. 1983;50: Cowan MD, Fye WB. Prevalence of QTc prolongation in women with mitral valve prolapse. Am J Cardiol. 1989;63: Mavissakalian M, Salerni R, Thompson ME, Michelson L. Mitral valve prolapse and agoraphobia. Am J Psychiatry. 1983; 140: Margraf J, Ehlers A, Roth WT. Mitral valve prolapse and panic disorder: a review of their relationship. Psychosom Med. 1988;50: Schutte JE, Gaffney FA, Blend L, Blomqvist CG. Distinctive anthropometric characteristics of women with mitral valve prolapse. Am J Med. 1981;71: Weissman NJ, Shear MK, Kramer-Fox R, Devereux RB. Contrasting patterns of autonomic dysfunction in patients with mitral valve prolapse and panic attacks. Am J Med. 1987;82: Gaffney FA, Bastian BC, Lane LB, et al. Abnormal cardiovascular regulation in the mitral valve prolapse syndrome. Am J Cardiol. 1983;52: Chesler E, Weir EK, Braatz GA, Francis GS. Normal catecholamine and hemodynamic responses to orthostatic tilt in subjects with mitral valve prolapse. Correlation with psychologic testing. Am J Med. 1985;78: Lenders JW, Fast JH, Blankers J, de Boo T, Lemmens WA, Thien T. Normal sympathetic neural activity in patients with mitral valve prolapse. Clin Cardiol. 1986;9: Motulsky AG. Biased ascertainment and the natural history of diseases. N Engl J Med. 1978;298: Devereux RB, Kramer-Fox R. Gender differences in mitral valve prolapse. In: Douglas PS, ed. Cardiovascular Health and Disease in Women. Philadelphia: F.A. Davis; 1988: Waller BF, Morrow AG, Maron BJ, et al. Etiology of clinically isolated, severe, chronic, pure mitral regurgitation: analysis of 97 patients over 30 years of age having mitral valve replacement. Am Heart J. 1982;104: Olson LJ, Subramanian R, Ackermann DM, Orszulak TA, Edwards WD. Surgical pathology of the mitral valve: a study of 712 cases spanning 21 years. Mayo Clin Proc. 1987;62: Danielsen R, Nordrehaug JE, Vik-Mo H: High occurrence of mitral valve prolapse in cardiac catheterization patients with pure isolated mitral regurgitation. Acta Med Scand. 1987;221: Capucci R, Devereux RB, Hochreiter C, et al. Risk factors for severe mitral regurgitation in patients with mitral valve prolapse [Abstract]. Clin Res. 1987;35:785A Corrigall D, Bolen J, Hancock EW, Popp RL. Mitral valve prolapse and infective endocarditis. Am J Med. 1977;63: August 1989 Annals of Internal Medicine Volume 111 Number 4

13 123. McKinsey DS, Ratts TE, Bisno AL. Underlying cardiac lessons in adults with infective endocarditis. The changing spectrum. Am J Med. 1987;82: Danchin A, Voiriot P, Briancon S, et al. Mitral valve prolapse with systolic murmur, a true risk factor for infective endocarditis. A case-controlled study [Abstract]. Circulation. 1988;78 (Suppl II): de Bono DP, Warlow CP. Potential sources of emboli in patients with presumed transient cerebral or retinal ischaemia. Lancet. 1981;1: Hachinski V, Norris JW. The Acute Stroke. Philadelphia: F.A. Davis; 1985: Plum F, Baringer JR, Gilman S, eds. Contemporary Neurology Series; vol Egeblad H, Sorensen PS. Prevalence of mitral valve prolapse in younger patients with cerebral ischaemic attacks: A blinded controlled study. Acta Med Scand. 1984; 216: Kouvaras G, Bacoulas G. Association of mitral leaflet prolapse with cerebral ischaemic events in the young and early middle-aged patient. Q J Med. 1985;56: Adams HP Jr, Butler MJ, Biller J, Toffol GJ. Nonhemorrhagic cerebral infarction in young adults. Arch Neurol. 1986;43: Bogousslavsky J, Regli F. Ischemic stroke in adults younger than 30 years of age. Arch Neurol. 1987;44: Kelley RE, Pina I, Lee SC. Cerebral ischemia and mitral valve prolapse: case-control study of associated factors. Stroke. 1988;19: Chesler E, King RA, Edwards JE. The myxomatous mitral valve and sudden death. Circulation. 1983;67: Topaz O, Edwards JE. Pathologic features of sudden death in children, adolescents, and young adults. Chest. 1985;87: Allen H, Harris A, Leatham A. Significance and prognosis of an isolated late systolic murmur: a 9- to 22-year follow-up. Br Heart J. 1974;36: Mills P, Rose J, Hollingsworth J, Amara I, Craige E. Long-term prognosis of mitral valve prolapse. N Engl J Med. 1977;297: Bissett GS 3d, Schwartz DC, Meyer RA, James FW, Kaplan S. Clinical spectrum and long-term follow-up of isolated mitral valve prolapse in 119 children. Circulation. 1980;62: Duren DR, Becker AE, Dunning A J. Long-term follow-up of idiopathic mitral valve prolapse in 300 patients: a prospective study. J Am Coll Cardiol. 1988;11: Kozuk LJ, Moien M. Detailed diagnoses and procedures for patients discharged from short-stay hospitals, United States, Hyattsville, Maryland: National Center for Health Statistics (U.S.); DHHS Publication no. (PHS) Lawrence L. Detailed diagnoses and procedures for patients discharged from short-stay hospitals, United States, Hyattsville, Maryland: National Center for Health Statistics: DHHS publication no. (PHS) Griffin MR, Wilson WR, Edwards WD, O'Fallon WM, Kurland LT. Infective endocarditis. Olmsted County, Minnesota, 1950 through JAMA. 1985;254: Ruskin JN, DiMarco JP, Garan H. Out-of-hospital cardiac arrest: electrophysiologic observations and selection of long-term antiarrhythmic therapy. TV Engl J Med. 1980;303: Kligfield P, Levy D, Devereux RB, Savage DD. Arrhythmias and sudden death in mitral valve prolapse. Am Heart J. 1987;113: Barnett HJ, Boughner DR, Taylor DW, Cooper PE, Kostuk WJ, Nichol PM. Further evidence relating mitral valve prolapse to cerebral ischemic events. N Engl J Med. 1980;302: Kligfield P, Hochreiter C, Kramer H, et al. Complex arrhythmias in mitral regurgitation with and without mitral valve prolapse: contrast to arrhythmias in mitral valve prolapse without mitral regurgitation. Am J Cardiol. 1985;55: Nolan CM, Kane J J, Grunow WA. Infective endocarditis and mitral prolapse: a comparison with other types of endocarditis. Arch Intern Med. 1981;141: Tresch DD, Doyle TP, Bonchek LI, et al. Mitral valve prolapse requiring surgery. Clinical and pathologic study. Am J Med. 1985;78: Caidahl K, Larsson S, Sudow G, Wallentin I, Angelhed JE, Olsson BS. Conservative surgery for mitral valve prolapse with regurgitation: clinical follow-up and noninvasive assessment. Eur Heart J. 1987;8: Jeresaty RM, Edwards JE, Chawla SK. Mitral valve prolapse and ruptured chordae tendineae. Am J Cardiol. 1985;55: Waller BF, Maron BJ, Del Negro A A, Gottdiener JS, Roberts WC. Frequency and significance of M-mode echocardiographic evidence of mitral valve prolapse in clinically isolated pure mitral regurgitation: analysis of 65 patients having mitral valve replacement. Am J Cardiol. 1984;53: Roberts WC. Mitral valve prolapse and systemic hypertension [Editorial]. Am J Cardiol. 1985;56: Devereux RB. Mitral valve prolapse and severe mitral regurgitation [Editorial]. Circulation. 1988;78: Kligfield P, Hochreiter C, Niles N, Devereux RB, Borer JS. Relation of sudden death in pure mitral regurgitation, with and without mitral valve prolapse, to repetitive ventricular arrhythmias and right and left ventricular ejection fractions. Am J Cardiol. 1987;60: Schulze RA Jr, Strauss HW, Pitt B. Sudden death in the year following myocardial infarction. Relation to ventricular premature contractions in the late hospitals phase and left ventricular ejection fraction. Am J Med. 1976;62: Holmes J, Kubo SH, Cody RJ, Kligfield P. Arrhythmias in ischemic and nonischemic dilated cardiomyopathy: prediction of mortality by ambulatory electrocardiography. Am J Cardiol. 1985;55: Oakley CM. Mitral valve prolapse: harbinger of death or variant of normal? Br Med J [Clin Res]. 1984;288: Kessler KM. Prolapse paranoia [Editorial]. J Am Coll Cardiol. 1988;11: Devereux RB, Cappucci R, Kramer-Fox R. Clinical features of mitral valve prolapse associated with low body weight. Am J Cardiol. 1988;61: MacMahon SW, Blacket RB, Macdonald GJ, Hall W. Obesity, alcohol consumption and blood pressure in Australian men and women. The National Heart Foundation of Australia Risk Factor Prevalence Study. J Hypertens. 1984;2: Rowland ML, Fulwood R. Coronary heart disease risk factor trends in blacks between the first and second National Health and Nutritional Examination Surveys, United States, Am Heart J. 1984;108: Helmcke F, Nanda NC, Hsiung MC, et al. Color Doppler assessment of mitral regurgitation with orthogonal planes. Circulation. 1987;75: Peller OG, Wallerson DC, Devereux RB. Role of Doppler and imaging echocardiography in selection of patients for cardiac valvular surgery. Am Heart J. 1987;114: Zile MR, Gaasch WH, Carroll JD, Levine HJ. Chronic mitral regurgitation: predictive value of preoperative echocardiographic indexes of left ventricular function and wall stress. J Am Coll Cardiol. 1984;3: Furer J, Hochreiter C, Niles NW, et al. Prediction of symptom status following mitral valve replacement for mitral regurgitation by preoperative echocardiographic measurements of the end-systolic stress to end-systolic volume ratio. Am J Noninvasive Cardiology. 1988;1: Hochreiter C, Niles N, Devereux RB, Kligfield P, Borer JS. Mitral regurgitation: Relationship of noninvasive descriptors of right and left ventricular performance to clinical and hemodynamic findings and to prognosis in medically and surgically treated patients. Circulation. 1986;73: Edmunds LH Jr, Stephenson LW, Edie RN, Ratcliffe MB. Openheart surgery in octogenarians. N Engl J Med. 1988;319: Savage DD, Levy D, Garrison RJ, et al. Mitral valve prolapse in the general population. 3. Dysrhythmias: the Framingham study. Am Heart J. 1983;106: Kramer HM, Kligfield P, Devereux RB, Savage DD, Kramer-Fox R. Arrhythmias in mitral valve prolapse. Effect of selection bias. Arch Intern Med. 1984;144: Wit AL, Cranefield PF. Triggered activity in cardiac muscle fibers of the simian mitral valve. Circ Res. 1976;38: Pinto JE, Nazarali A J, Saavedra JM. Quantitative autoradiographic characterization of B-adrenoceptors in mitral valve leaflets of the rat heart [Abstract]. Am J Hypertension. 1988;1:9A Gaffney FA, Bastian BC, Lane LB, et al. Autonomic dysfunction in women with mitral valve prolapse syndrome. Circulation. 1979;59: Gaffney FA, Lane LB, Pettinger W, Blomqvist CG. Effects of longterm clonidine administration on the hemodynamic and neuroendocrine postural responses of patients with dysautonomia. Chest. 1983;83: Koch KL, Davidson WR Jr, Day FP, Spears PF, Voss SR. Esophageal dysfunction and chest pain in patients with mitral valve prolapse: a prospective study using provocative testing during esophageal manometry. Am J Med. 1989;86: August 1989 Annals of Internal Medicine Volume 111 Number 4 317