- Email: [email protected]
Complications of mitral valve prolapse
Complications of Mitral Valve Prolapse Disproportionate Occurrence in Men and Older Patients
RICHARD B. DEVEREUX, M.D. IRENE HAWKINS, M.D. RANDI KRAMER-FOX, M.S. ELIZABETH M. LUTAS, M.D. ISAAC W. HAMMOND, Ph.D. MARIANE C. SPITZER CLARE HOCHREITER, M.D. RICHARD B. ROBERTS, M.D. ROBERT N. BELKIN, M.D. PAUL KLIGFIELD, M.D. W. TED BROWN, M.D., Ph.D. NATHANIEL NILES, M.D. MICHAEL H. ALDERMAN, M.D. JEFFREY S. BORER, M.D. JOHN H. LARAGH, M.D. New York, New York
From the Departments of Medicine and Public Health, New York Hospital-Cornell Medical Center, New York, New York. This work was supported in part by Grants HL-22006, 18323, and 25604 from the National Heart, Lung, and Blood Institute, Bethesda, Maryland, and presented in abstract form at the Scientific Session of the American College of Cardiology, Anaheim, California, March 13, 1985. Requests for reprints should be addressed to Dr. Richard B. Devereux, Division of Cardiology, Box 222, New York Hospital-Cornell Medical Center, 525 East 68th Street, New York, New York 10021. Manuscript submitted September 26, 1985, and accepted November 11, 1985.
To determine factors influencing the strength of association between mitral valve prolapse and mitral regurgttation, ruptured chordae tendineae, and infective endocarditis, the prevalence of mitral prolapse in patients with disease was compared with both clinical and population control groups. The prevalence of mitral valve prolapse was 4 percent among population and clinical control giroups (eight of 196 and 84 of 2,146, respectively) and was significantly higher (p
November
1986
The American
Journal
of Medicine
Volume
81
751
COMPLICATIONS
OF MITRAL
VALVE
PROLAPSE-DEVEREIJX
ET AL
evaluate the strength of the association between mitral valve prolapse and infective endocarditis, significant mitral regurgitation, and ruptured mitral chordae tendineae.
were designed to reflect a general adult population, the clinical population from which the patient groups were drawn, and an unselected population of adults with mitral prolapse. Our general population group consisted of normotensive subjects chosen randomly within each of 36 predefined strata to reflect the age, sex, and racial composition of a large employed population in New York [ 191. Echocardiography was performed in 1981 and 1982, concurrently with echocardiographic studies in the patient groups. Our clinical control subjects were consecutive patients undergoing echocardiography in 1980 and 198 1, during the period of enrollment of all patient groups. As part of a study of the impact of echocardiography on clinical diagnosis, complete information about medical history and physical findings was provided by each patient’s physician before echocardiography. Because the number of subjects with mitral prolapse identified in our employed population sample was too small to permit reliable comparison of demographic and clinical features with patient groups, we used an additional control group of unselected subjects with mitral valve prolapse. This consisted of 81 consecutive adult first-degree relatives found to have mitral valve prolapse during a prospective family study, previously reported in part [20]. Echocardiographic Methods. Echocardiography was performed with the patients in the partial lefi decubitus position with the head of the bed elevated approximately 30 degrees. We paid particular attention throughout the study to ensure that transducer placement in the left parasternal interspaces was such that the ultrasound beam was oriented perpendicularly to the valve during recordings of the mitral valve. Standard techniques were used to obtain views of the mitral valve, the left ventricle, and the left atrium and aortic root. Echocardiograms were recorded at a paper speed of 50 mm per second using 2.25 mHz focused transducers and either an lrex System 3, a Picker 8OC, a Smith-Kline Ekoline echocardiograph with a Honeywell 1856A stripchart recorder, or an ATL 3001C or a Diasonlcs DRF 400 echocardiograph with Tektronix recorder. The echocardiographic results were interpreted independently of clinical information. Systolic motion of the mitral valve was evaluated on all strips recording the mitral valve on each study, using standard criteria for diagnosis of mitral prolapse [21-231. Mitral valve prolapse was considered to be present if continuous interfaces representing mitral valve leaflets showed both “turning around” with posterior motion (away from the transducer) and movement at least 2 mm behind the line connecting the C and D points of the mitral valve. Late systolic prolapse was diagnosed when the valve “turned around” in midsystole, whereas holosystolic prolapse was diagnosed when posterior motion began with the onset of systole. The depth of prolapse was measured from the C-D line to the first mitral leaflet interface, thus excluding the possibility of measuring to a chordal or left atrial wall echo. Echocardiographic results were not considered to indicate mitral valve prolapse if they demonstrated only minimal motion behind the C-D line, late systolic “flattening” of mitral leaflet motion without actual posterior motion, or discontinuous lines that appeared to prolapse. Transient posterior motion of mitral valve leaflets immediately after mitral valve closure was not considered to be mitral valve
SUBJECTS AND METHODS Study Design. We used the case-control method. All subjects underwent M-mode echocardiography. Mitral prolapse was diagnosed using standard M-mode echocardiographic criteria by investigators blinded to clinical findings. The prevalence of mitral valve prolapse in each patient group was compared with that in control groups to derive overall odds ratios. Because of the known effect of age and sex on the prevalence of mitral valve prolapse in unselected populations, odds ratios were recalculated after each patient was age- and sex-matched to two subjects from each control population. Differences in age, sex, and presence of heart murmurs between patients with complications and another control group of unselected subjects with uncomplicated mitral valve prolapse were also assessed. Subjects. Consecutive patients at the New York HospitalCornell Medica! Center with any of the three following conditions were studied. They were as follows: All patients with conclusive evidence of infective endocarditis on native heart valves between November 1978 and December 1982, who underwent M-mode and twodimensional echocardiogrephy and had complete medical records available, were enrolled. All but one patient had bacteriologic identification of the infecting organism, whereas the lone patient with culture-negative endocarditis presented with fever, arthralgias, petechial rash, anemia, hematurla, erythrocyte sedimentation rate of 128, and a latex fixation test result positive at 15,120. All patients with isolated, pure, moderate to severe mitral regurgitation by clinical criteria enrolled in a prospective natural history study, between December 1980 and July 1983, were enrolled in this study. Each subject underwent complete clinical evaluation and M-mode and two-dimensional echocardiography as part of the study protocol. Evidence of hemodynamically important mitral regurgitation included the presence of a holosystolic or nearly holosystolit murmur of mitral regurgitation, a third heart sound, and enlargement of the left ventricle and left atrium on the basis of physical and echocardiographic findings. Further proof of the severity of mitral regurgltation was provided in most patients by cardiac catheterization or Doppler echocardiography. Twelve-lead electrocardiography, treadmill exercise testing using the Bruce protocol, and radionuclide cineangiograpy at rest and during exercise [ 181, as well as cardiac catheterization when clinically indicated, were performed to exclude ischemic heart disease as the cause of mitral regurgitation. Consecutive patients with ruptured mitral chordae tendineae diagnosed by M-mode and two-dimensional echocardiographic criteria between 1976 and 1982 were enrolled. Full clinical records were available for each patient as well as anatomic confirmation of mitral chordal abnormality in 20 patients at the time of cardiac surgery or autopsy. To minimize potential bias introduced by selection of control subjects, we used three independent groups. They
752
November
1986
The American
Journal
of Medicine
Volume
81
COMPLICATIONS
TABLE I
OF MITRAL
Number Age (years)* Men (percent) Women (percent) f
ET AL
ControlGroups
Groups Mitral Regurgitation
Patient
Mean
PROLAPSE-DEVEREUX
Clinical Characteristics of Patient and Control Groups lnleclive Endocarditis
l
VALVE
standard
67
3af
31
14
49 (73) 18 (27)
50f
14
17 (55) 14 (45)
Ruptured Chordae
Population
Clinical
43 56f 17 30 (73) 13 (27)
196 446 13 106 (54) 90 (46)
2146 54% 19 923 (43) 1,223 (57)
Relativeswith Prolapse
Mitral
81
38 f 14 25 (31) 56 (69)
deviation.
prolapse. Posterior motion of the mitral leaflets after the time when the left ventricular internal dimension reached its minimum and gradual posterior motion of the mitral leaflets reaching a midsystolic nadir were also not considered to represent mitral valve prolapse. Echocardiographic findings of ruptured mitral chordae were sought on both M-mode and two-dimensional studies. M-mode manifestations of chordal rupture included the following: abnormal systolic appearance in the left atrium of mitral leaflet or chordal echoes; chaotic diastolic fluttering of the anterior mitral leaflet, more irregular than usually seen with aortic regurgitation; systolic fluttering of the mitral valve; paradoxic anterior motion of the posterior mitral leaflet in early diastole; paradoxic movement of the posterior leaflet both anteriorly in systole and posteriorly in diastole; and irregular early diastolic fluttering of the posterior leaflet. Ruptured chordae and accompanying flail leaflet were identified on two-dimensional echocardiography by the presence of abnormal systolic mitral coaptation with systolic whipping of the involved leaflet into the left atrium [ 24) in either the long axis or the apical four-chamber view. In the presence of ruptured chordae tendineae, mitral prolapse was diagnosed only if two-dimensional echocardiography showed the free edge of the affected mitral leaflet to be directed toward the left ventricle, with the mid-portion of the leaflet billowing farther into the left atrium. In one patient, the protrusion of a chorda tendinea with an attached mass into the left ventricular outflow tract in systole was considered indicative of chordal rupture. Statistical Analysis. To approximate the relative risk of each complication occurring in subjects with mitral valve prolapse as opposed to those without, we calculated the odds ratio [25]. The data used to derive the odds ratio for each comparison were the number of subjects with mitral prolapse among patients with a specific complication (A) and in the relevant control group (B) as well as the number of subjects without mitral valve prolapse in the patient and control groups (C and D, respectively). The odds ratio was then calculated as A X D/B X C. This approach was also used to approximate the relative risk of complications in subgroups of patients with mitral prolapse defined by simple demographic or clinical findings (e.g., male sex or presence of a heart murmur). Statistical significance differences was assessed using the chi-square test, with Yates’ continuity correction when appropriate [ 261. Approximate confidence limits of odds ratios were calculated by Woolf’s method [27]. Because age and sex are known to influence the preva-
November
lence of mitral valve prolapse [20,28,29], we recalculated the odds ratios after matching each patient to two subjects of the same sex and within five years in age from each control group. Assessment of the odds ratio for each complication in subjects with mitral valve prolapse from these case-control triplets was performed by the Mantel-Haenszel test with calculation of approximate confidence intervals [30,31]. Statistical significance was assessed by the chi-square test with one degree of freedom. Differences between groups with regard to continuous variables were assessed by the unpaired Student t test [26].
RESULTS Characteristics
of Patient and Control Groups.
The
number of subjects in each group, their age, and the proportion of men and women are given in Table I. The evident differences in age and sex distribution among groups were taken into account in the analyses of ageand sex-matched case-control triplets, the results of which are presented next. A murmur of mitral regurgitation was known before present illness in six of 11 patients with mitral prolapse and endocarditis (55 percent), and before study enrollment in 22 of 8 1 affected family members (27 percent); late or holosystolic mitral regurgitation murmurs were present during hospitalization in 9 of 11 patients with endocarditis (78 percent), whereas late systolic murmurs were audible in 27 affected relatives (33 percent) and a holosystolic murmur was audible in one (1 percent) additional relative with mitral valve prolapse. The most frequent indications for echocardiography among clinical control subjects were an explicit diagnosis of mitral prolapse or suspicion of this condition (16 percent), established or suspected diagnoses of other forms of valvular heart disease (24 percent), pericardial disease (10 percent), atrial fibrillation or other arrhythmia (7 percent), and congestive heart failure or cardiac enlargement (7 percent).
Odds Ratios for Complications in Subjects with Mitral Valve Prolapse: Overall Analyses (Table II). The prevalence of mitral valve prolapse was virtually identical in our population and clinical control groups (eight of 196 or 4.1 percent and 84 of 2,146 or 3.9 percent, respectively; difference not significant). Mitral prolapse was slightly more common among women (five of 90 or 5.6 percent) than men (three of 106 or 2.8 percent) among the popula-
1986
The American
Journal
of Medicine
Volume
81
753
COMPLICATIONS
OF MITRAL VALVE PROLAPSE-DEVEREUX
TABLE II
Association of Mitral Prolapse with Complications: Overall Analysis
Number Mitral prolapse (percent) No mitral prolapse (percent) Odds ratio versus population control group Odds ratio versus clinical control group Statistical
significance:
* p
ET AL
lnfeclive Endocardilis
Mitral Regurgitation
Ruptured Chordae
Population Control Group
Clinical Control Group
67 11 (16) 56 (84) 4.6’
31 17 (55) 14 (45) 28.5”
43 27 (63) 16 (37) 39.6”
196 8 (4) 188 (96) -
2,146 84 (4) 2,062 (96) 1.0
29.8”
41.4”
1.0
-
4.8” versus
population
(employed)
and clinical
tion control group, whereas prevalences among women and men in the clinical control group were more nearly equal (52 of 1,233 or 4.3 percent versus 32 of 923 or 3.5 percent). In contrast, the prevalence of echocardiographic mitral prolapse was significantly higher (p
754
November
1986
The American
Journal
of Medicine
control
groups.
tis for age and sex with two subjects from our population and clinical control groups. This procedure reduced the prevalence of mitral valve prolapse in each control population to 2 percent and increased the odds ratio for mitral valve prolapse in endocarditis to 6.7. We believe that this change in the odds ratio was due to sex- and not agematching. Approximate 95 percent confidence limits of this odds ratio are 1.9 to 22.9. In contrast, age- and sex-matching of control subjects to patients with mitral regurgitation (Table IV) did not systematically alter the prevalence of mitral valve prolapse among control groups from the 4 percent observed in our overall analysis, or the high odds ratios for this condition in subjects with mitral valve prolapse. Similarly, age- and sex-matching of control subjects to patients with ruptured chordae tendineae (Table V) had no effect on the dramatic odds ratios for mitral valve prolapse observed in patients with ruptured mitral valve chordae tendineae. Relation of Complications to Demographic and Clinical Features of Mitral Valve Prolapse. To assess whether patients with mitral valve prolapse at high risk of complications could be identified on the basis of simple demographic or clinical features, we next compared the patients with mitral valve prolapse in each group with complications to unselected subjects with mitral valve prolapse identified during family studies. Table VI compares the age of the 11 patients with mitral valve prolapse and endocarditis, the 17 with mitral prolapse and mitral regurgitation, and the 27 with mitral prolapse and ruptured mitral chordae to that of the 8 1 first-degree relatives with mitral prolapse in our family studies. Patients with mitral prolapse and endocarditis were similar in age to the unselected subjects with mitral prolapse, whereas patients with mitral regurgitation or ruptured chordae and mitral prolapse were significantly older. Table VII shows the proportion of men among persons with mitral valve prolapse in our patient groups and unselected subjects. Men predominate in each patient group, whereas two-thirds of the unselected subjects with mitral valve prolapse are women, The highest odds ratio for men among patient groups with mitral prolapse occurred with ruptured chordae tendineae (7.4) and infective endocardi-
Volume
61
COMPLICATIONS
TABLE
III
Association
of Mitral Prolapse
with Infective
Endocarditis:
OF MITRAL
Matched-Triplets
Patients With infective Endocarditis Number Mitral prolapse (percent) No mitral prolapse (percent) Odds ratio versus population control group Odds ratio versus clinical control group Statistical
TABLE
significance
IV
by Mantel-Haenszel
Association
test:
of Mitral Prolapse
95 percent
confidence
limits:
Statistical
significance
TABLE V
by Mantel-Haenszel
Association
test:
Matched-Triplets
of Mitral Prolapse
95 percent
with Ruptured
62 1 (2) 61 (96) limits:
Tendineae:
t 9.1 to 126, t 6.3 to 87.8.
Matched-Triplets
Patients with Ruotured Chordae Number Mitral prolapse (percent) No mitral prolapse (percent) Odds ratio versus population control group Odds ratio versus clinical control group Statistical
significance
TABLE VI
by Mantel-Haenszel
Age of Subjects
test:
with Complicated
95 percent
86 4 (5) 82 (95) 0.3
confidence
and Uncomplicated
Analysis
Matched Control Groups Population Clinical
43 27 (63)” 16 (37) 17*+ 53*$
lp
62 3 (5) 59 (95) 3.1 -
0.4
confidence
Chordae
Analysis Matched Control Groups Population Clinical
31 17 (55)’ 14 (45) 34.0*+ 23.7”t “p
134 3 (2) 131 (96) 1.0 -
t 1.96 to 22.9.
Patients with Mitral Regurgitation Number Mitral prolapse (percent) No mitral prolapse (percent) Odds ratio versus population control group Odds ratio versus clinical control group
Analysis
134 3 (2) 131 (96) 1.0
with Mitral Regurgitation:
ET AL
Matched Control Groups Population Clinical
67 11 (16)” 56 (64) 6.7*+ 6.7”+
lp
VALVE PROLAPSE-DEVEREUX
limits:
t 7.2 to 43.5,
86
2 (2) 84 (98) 3.0 t 16.7 to 168.0.
Mitral Prolapse
Patients with Mitral Prolapse
Number Age range Mean age” p value versus
relatives
with mitral
prolapse
Infective Endocarditis
Mitral Regurgitation
Ruptured Chordae Tendlneae
Control Group Relatives with Prolapse
11 19-60 37 f 13 NS
17 20-72 48f 14 < 0.005
27 23-79 55f 16
81 16-71 38% 14 -
NS = difference not significant. f standard deviation. l
TABLE VII
Sex of Subjects
with Complicated
and Uncomplicated
Mitral Prolapse
Patients with Mitral Prolapse
Number Male (percent) Female (percent) Odds ratio versus Statistical
significance:
relatives
with mitral
* p
prolapse
Infective Endocarditis
Mitral Reauraitalion
Ruptured Chordae Tendineae
Control Group Relatives with Prolapse
11 8 (73) 3 (27) 6.0*
17 9 (53) 8 (47) 2.5
27 20 (74)T 7 (26) 7.4+
81 25 (31) 56 (69) -
+ p
November
1988
The
American
Journal
of Medicine
Volume
81
755
COMPLICATIONS
TABLE VIII
OF MITRAL
Relation
VALVE
PROLAPSE-DEVEREUX
of Infective
ET AL
Endocarditis
to Previous
or Current
Murmur
Mitral Prolapse with Infective Endocardilis Number Previously known murmur (percent) No known murmur (percent) Odds ratio Present murmur (percent) No present murmur (percent) Odds ratio Statistical
significance:
l
difference
not significant;
11 6 (56) 5 (45) 3.2” 9 (82) 2 (18)
The American
Journal
28 (34) 53 (66)
clinical practice and medical reports [1-4,9-l 61. Two previous studies assessed the relative risk of infective endocarditis occurring in persons with mitral valve prolapse by comparison with clinical control groups. Clemens et al [ 153 found an odds ratio of 4.9 based on comparison of their entire endocarditis and clinical control groups, and an odds ratio of 8.2 based on matched-pairs analysis, This is nearly identical to our results. Hickey et al [ 161 reported a similar odds ratio (5.3) using only clinical control subjects. However, doubt has persisted about these estimates of the relative risk of infective endocarditis in subjects with mitral prolapse because the prevalence of the condition in clinical control groups might have been distorted by selection biases. Our study establishes the accuracy of the estimate of a five- to seven-fold increase in risk of endocarditis in persons with mitral valve prolapse by finding nearly identical odds ratios when our patients with infective endocarditis were compared with control subjects drawn from a general adult population, and with a second clinical control group. Further analysis of our data indicates that men with mitral valve prolapse have a greater risk of infective endocarditis than affected women. This finding is compatible with the results of most [1,3,4,32,33] but not all [2] previous studies. The increase in odds ratios for endocarditis in subjects with mitral prolapse between overall and matched-pair analyses in the study of Clemens et al [ 151, similar to our findings, is also compatible with a male predominance. Our data also support an increased risk of infective endocarditis in patients with mitral prolapse and a murmur of mitral regurgitation, in keeping with the suggestion by Hickey et al [ 161. However, in this retrospective study, it was not possible to define precisely the odds ratio for endocarditis in subjects with mitral valve prolapse who did or did not have murmurs, with estimates ranging from 3.2 (difference not significant) based on the history of an antecedent murmur to 8.5 based on the presence of a mitral regurgitant murmur at the time of evaluation (p
Previous studies, mostly uncontrolled, have supported associations between mitral valve prolapse and several complications, including infective endocarditis, hemodynamically important mitral regurgitation, rupture of mitral valve chordae tendineae, and neurologic ischemic episodes. The present controlled study assessed the relation between mitral valve prolapse and the first three of these complications, finding highly significant associations (p
1986
ai 22 (27) 59 (73)
t p
COMMENTS
November
Relatives with Mitral ProlaDse
8.5t
tis (6.0). An additional observation in support of the apparent greater increase in risk of infective endocarditis among men as opposed to women with mitral prolapse is that one of the 25 men in our family mitral prolapse control group was hospitalized with bacterial endocarditis in February 1985, after completion of these data analyses. We examined a clinical feature previously suggested by others to be useful in identifying patients with mitral valve prolapse at risk for endocarditis-the presence of mitral regurgitant murmur. To avoid confounding effects of damage to mitral valve structure during endocarditis, we first assessed the prevalence in both groups of a history of a heart murmur known before the onset of present illness or enrollment in our family study. Only a modest and statistically insignificant association (odds ratio 3.2) is seen between a previously known murmur and endocarditis (Table VIII). However, when the proportion of patients with mitral prolapse and control subjects who exhibited late or holosystolic murmurs of mitral regurgitation during hospitalization or at the time of enrollment in our family study was compared, the odds ratio for endocarditis in patients with mitral prolapse who had as opposed to those who did not have heart murmurs rose to 8.5 and attained statistical significance at the p
756
in Mitral Prolapse
of Medicine
Volume
81
COMPLICATIONS
infection. Our finding that the prevalence of mitral Valve prolapse demonstrated by well-standardized M-mode echocardiographic criteria was approximately 4 percent in both our control groups-at the lower end of the accepted range for this condition but in keeping with prevalences of 2 to 4 percent recently reported by several investigators using similar methods [34-36]-shows that the proportion of the general population in whom endocarditis prophylaxis might be indicated for mitral prolapse is somewhat lower than previously thought. Taken together with the recent simplification of American Heart Association recommendations for antibiotic prophylaxis [37], these finding have implications for clinical practice. First, the somewhat lower population prevalence of mitral prolapse (4 percent instead of 5 to 7 percent) and the abbreviated course of antibiotics now recommended would reduce the expense and medical risks of routine antibiotic prophylaxis in subjects with mitral prolapse to below those estimated in recent costeffectiveness analyses [38-401. Second, our demonstration of a higher relative risk of endocarditis in men than women with mitral prolapse, and the evidence in favor Of a higher risk in patients with a murmur of mitral regurgitation in our study and that of Hickey et al [ 161, suggests that risk of endocarditis can be usefully compartmented. Thus, odds ratios of infective endocarditis occurring in men and women with mitral valve prolapse were 6.7 and 3.4, respectively, when our population control subjects were used as the reference standard, and 7.6 and 2.7, respectively, when our patients with infective endocarditis were compared with general population data from the Framingham Study [28], in which mitral prolapse was diagnosed by similar criteria, with cross-evaluation of echocardiographic results by one of us (R.B.D.). If, as our family studies suggest, the prevalence of late systolic murmurs is similar in men and women with mitral valve prolapse, it may be possible to approximate the risk of endocarditis on the basis of the combination of sex and the presence or absence of a murmur of mitral regurgitation. Projected odds ratios Compared with those for normal persons would range from 2.0 in women without heart murmurs to as high as 12.5 in men with murmurs of mitral regurgitation if our odds ratio for endocarditis in
OF MITRAL
VALVE PROLAPSE-DEVEREUX
ET AL
mitral prolapse based on a history of previous murmurs (3.4) is used, and from 1.0 to 15.7 in the same groups using the higher odds ratio (8.5) based on actual presence of a murmur at the time of study. Thus, our estimates of risk based on sex and presence or absence of a murmur, supported by the data of Clemens et al [ 151 and Hickey et al [ 161, appear to clarify decision-making for antibiotic prophylaxis: nearly half the patients with mitral prolapse-women without a murmur of mitral regurgitation-would fall into a group with too little risk to warrant this intervention, and an additional 10 percent (men with murmurs) would fall into an unequivocal high-risk group. Our data also confirm that strong associations, with odds ratios from 17 to 53, exist between mitral valve prolapse and clinically significant mitral regurgitation and echocardiographic findings of rupture of mitral chordae tendineae, as suggested in previous studies [5-131. Although the proportion of patients with mitral regurgitation or ruptured chordae who exhibited mitral prolapse was higher in our study than reported by some [5,7], it was lower than the proportion of patients with anatomically redundant mitral valves in two recent studies of the same complication. Thus, Waller et al [6] reported that 48 of 65 (74 percent) patients with isolated, pure, severe mitral regurgitation not due to ischemic heart disease had enlarged mitral leaflets and annuli, whereas Hickey et al [8] found anatomically “floppy” valves in 93 percent of patients in whom ruptured chordae were found at cardiac surgery. The clustering of hemodynamically important mitral regurgitation and ruptured chordae tendineae in older persons with mitral valve prolapse is compatible with the hypothesis [4 1,421 that these complications are related to “wear and tear” of an intrinsically abnormal mitral valve. The male predominance among our patients with these complications of mitral prolapse, similar to that observed by Lucas and Edwards [32] in a necropsy study, might be attributed to the greater hemodynamic load placed on the mitral valve by the known higher systolic blood pressure in men [43]. ACKNOWLEDGMENT We would like to thank Virginia Burns for her assistance in preparation of this manuscript.
REFERENCES 1.
2.
3.
LeBauer EJ, Perloff JK, Keliher TF: The isolated systolic click with bacterial endocarditis. Am Heart J 1967; 73: 534-537. Lachman AS, Bramwell-Jones DM, Lakier JB, Pocock WA, Barlow JB: infective endocarditis in the billowing mitral leaflet syndrome. Br Heart J 1975; 37: 326-330. Corrigall D, Bolen J, Hancock EW, Popp RL: Mitral valve prolapse and infective endocarditis. Am J Med 1977; 64: 215-222.
November
4.
5.
6.
1986
Nolan CM, Kane JJ, Grunow WA: infective endocarditis and mitral prolapse. Comparison with other types of endocarditis. Arch Intern Med 1981; 141: 447-450. Guy FC, MacDonald RPR, Fraser DB, Smith ER: Mitral valve prolapse as a cause of hemodynamically important mitral regurgitation. Can J Surg 1980; 23: 166-170. Waller BF, Maron BJ, DelNegro AA, Gottdiener JS, Roberts WC: Frequency and significance of M-mode echocardiographic evidence of mitral valve prolapse in clinically
The
American
Journal
of Medicine
Volume
61
757
COMPLICATIONS
7.
6.
9.
10.
Il.
12.
13.
14.
15.
16.
17. 18.
19.
26.
21.
22.
23.
24.
758
OF MITRAL
VALVE
PROLAPSE-DEVEREUX
ET AL
isolated pure mitral regurgitation: analysis of 65 patients having mitral valve replacement. Am J Cardiol 1964; 53: 139-147. Oltveira DBG, Ddwkins KD, Kay PH, Paneth M: Chordal rupture. I: Aetiology and natural history. Br Heart J 1963; 50: 312-317. Hickey AJ, Wilcken DEL, Wright JS, Warren BA: Primary (spontaneous) chordal rupture: relation to myxomatous valve disease and mitral valve prolapse. J Am Coll Cardiol 1965; 5: 1341-1346. 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: 525-532. Mills P, Rose J, Hollingsworth J, Amara I, Craige E: Longterm prognosis of mitral-valve prolapse. N Engl J Med 1977; 297: 13-18. Appelblatt N, Willts P, Lenhart J: Ten to 45 year follow-up df 69 patients with systolic click with or without apical late systolic murmur (abstr). Am J Cardiol 1975; 35:‘119. Belardi J, Lardani H, Manubeus S, Sheldon WC, Moreyon A: fdiopathic prolapse of the mitral vaive: a follow-up study In 137 patients studied by angiography (abstr). Am J Cardiol 1976; 37: 120. Koch FH, Hancock EW: Ten year follow-up of forty patients with midsystolic click/late systolic murmur syndrome (abstr). Am J Cardiol 1978; 37: 149. Nishimura RA, McGoon MD, Shub C, Miller FA Jr, llstrup DM, Tajik AJ: Echocardiographically documented mitral valve prolapse. Long-term follow-up of 237 patients. N Engl J Med 1985; 313: 13051309. Clemens JD, Horwitz RI, Jaffe CC, Feinsteih AR, Stanton BF: A controlled evaluation of the risk of bacterial endocarditis in persons with mitral valve prolapse. N Engl J Med 1982; 307: 776-781. Hickey AJ, MacMahon SW, Wilcken DEL: Mitral valve prolapse and bacterial endocarditis: when is antibiotic prophylaxis necessary? Am Heart J 1985; 169: 431-435. Leatham A, Brigdeh W: Mild mitral regurgitation and the mitral prolapse fiasco. Am Heart J 1980; 99: 659-664. Borer JS, Bacharach SL, Green MV, et al: Real-time radionuelide cineangiography in the non-invasive evaluation of global and regional left ventricular function at rest and during exercise in patients with coronary artery disease. N Engl J Med 1977; 296: 839-844. Alderman MH, Schoenbaum EE: Detection and treatment of hypertension at the work site. N Engl J Med 1975; 293: 65-68. Devereux RB, Brown WT, Kramer-Fox R, Sachs I: Inheritance of mitral valve prolapse: effect bf age and sex on gene expression. Ann Intern Med 1982; 97: 826-832. Popp RL, Brown OR, Silverman JF, Harrison DC: Echocardiographic abnormalities in the mitral valve prolapse syndrome. Circulation 1974; 49: 428-433. 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: 1091-1096. Devereux RB, Brown WT, Sachs I, Kramer-Fox R: Evolution of echocardiographic findings in patients with mitral valve prolapse. J Cardiovasc Ultrasonogr 1983; 2: 29234. Mintz GS, Kotler MN, Segal BL, Parry WR: Two-dimensional
November
1986
The American
Journal
of Medicine
25.
26. 27.
28.
29.
30.
31.
32. 33.
34.
35.
36.
37.
38.
39.
40.
41. 42.
43.
Volume
81
echocardiographic recognition of ruptured chordae tendineae. Circulation 1978; 57: 244-250. Schlesselmann JJ: Case-control studies: design, conduct, analysis. New York: Oxford University Press, 1982; 3255. Armkage P: Statistical methods in medical research. New York: Wiley, 1971. Schlesselmann JJ: Case-control studies: design, conduct, analysis. New York: Oxford University Press, 1982; 176178. Savage DD, Garrison RJ, Devereux RB, et al: Mitral valve prolapse in the general population. I. Epidemiologic features. The Framingham Study. Am Heart J 1983; 106: 571-576. Strahan NV, Murphy EA, Fortuin NJ, Come PC, Humphries JO’N: Inheritance of the mitral valve prolapse syndrome: discussion of a three-dimens/onal penetrance model. Am J Med 1983: 74: 967-972. Mantel N, Haenszel W: Statistical aspects of the analysis of data from retrospective studies of disease. JNCI 1959; 22: 719-748. Schlesselmann JJ: Case-controls studies: design, conduct, analysis. New York: Dxford University Press, 1982; 2i2216. Lucas RV JR, Edwards JE: The floppy mitral valve. Curr Probl Cardiol 1982; 7: l-48. Durack DT, Kaplan EL, Bisno AL: Apparenf failures of endocarditis prophylaxis. Analysis of 52 cases submitted to a national registry. JAMA 1983; 250: 2318-2322. Hickey AJ, Wolfers J, Wilcken DEL: Mitral valve prdlapse. Prevalence in an Australian population. Med J Aust 1981; 1: 31-33. Wann LS, Grove JR, Hess TR, et al: Prevalence of mitral valve prolapse by two-dimensional echocardiography in healthy young women. Br Heart J 1983; 49: 334-340. Bryhn M, Perrson S: The prevalence of mitral valve prolapse in healthy men and women in Sweden. Acta Med S&d i984; 215: 157-160. 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-1127A. Bor DH, Himmelstein DU: Endocarditis prophylaxis for patients with mitral valve prolapse. A quantitative analysis. Am J Med 1984; 76: 711-717. Retchin SM, Fletcher RH, Waugh RA: Endocarditis and mitral valve prolapse: what is the risk? Int J Card&l 1984; 5: 653-659. Clemens JD, RanSohoff DF: A quantitative assessment of pre-denfal antibiotic prophylaxis for patients with mitral valve prolapse. J Chronic Dis 1984; 37: 531-544. Devereux RB, Perloff JK, Reichek N, Josephson ME: Mitral valve prolapse. Circulation 1976; 54: 3-14. Hickey AJ, Wilcken DEL: Age and the clinical profile of idiopathic mitral valve prolapse. Br Heart J 1986; 55: 582-586. Rowland ML, Fulwood R: Corbnary heart disease risk factor trends in blacks between the first and second National Health and Nutritional Examination Surveys, United States, 1971-1980. Am Heart J 1984; 108: 771-779.
RICHARD B. DEVEREUX, M.D. IRENE HAWKINS, M.D. RANDI KRAMER-FOX, M.S. ELIZABETH M. LUTAS, M.D. ISAAC W. HAMMOND, Ph.D. MARIANE C. SPITZER CLARE HOCHREITER, M.D. RICHARD B. ROBERTS, M.D. ROBERT N. BELKIN, M.D. PAUL KLIGFIELD, M.D. W. TED BROWN, M.D., Ph.D. NATHANIEL NILES, M.D. MICHAEL H. ALDERMAN, M.D. JEFFREY S. BORER, M.D. JOHN H. LARAGH, M.D. New York, New York
From the Departments of Medicine and Public Health, New York Hospital-Cornell Medical Center, New York, New York. This work was supported in part by Grants HL-22006, 18323, and 25604 from the National Heart, Lung, and Blood Institute, Bethesda, Maryland, and presented in abstract form at the Scientific Session of the American College of Cardiology, Anaheim, California, March 13, 1985. Requests for reprints should be addressed to Dr. Richard B. Devereux, Division of Cardiology, Box 222, New York Hospital-Cornell Medical Center, 525 East 68th Street, New York, New York 10021. Manuscript submitted September 26, 1985, and accepted November 11, 1985.
To determine factors influencing the strength of association between mitral valve prolapse and mitral regurgttation, ruptured chordae tendineae, and infective endocarditis, the prevalence of mitral prolapse in patients with disease was compared with both clinical and population control groups. The prevalence of mitral valve prolapse was 4 percent among population and clinical control giroups (eight of 196 and 84 of 2,146, respectively) and was significantly higher (p
November
1986
The American
Journal
of Medicine
Volume
81
751
COMPLICATIONS
OF MITRAL
VALVE
PROLAPSE-DEVEREIJX
ET AL
evaluate the strength of the association between mitral valve prolapse and infective endocarditis, significant mitral regurgitation, and ruptured mitral chordae tendineae.
were designed to reflect a general adult population, the clinical population from which the patient groups were drawn, and an unselected population of adults with mitral prolapse. Our general population group consisted of normotensive subjects chosen randomly within each of 36 predefined strata to reflect the age, sex, and racial composition of a large employed population in New York [ 191. Echocardiography was performed in 1981 and 1982, concurrently with echocardiographic studies in the patient groups. Our clinical control subjects were consecutive patients undergoing echocardiography in 1980 and 198 1, during the period of enrollment of all patient groups. As part of a study of the impact of echocardiography on clinical diagnosis, complete information about medical history and physical findings was provided by each patient’s physician before echocardiography. Because the number of subjects with mitral prolapse identified in our employed population sample was too small to permit reliable comparison of demographic and clinical features with patient groups, we used an additional control group of unselected subjects with mitral valve prolapse. This consisted of 81 consecutive adult first-degree relatives found to have mitral valve prolapse during a prospective family study, previously reported in part [20]. Echocardiographic Methods. Echocardiography was performed with the patients in the partial lefi decubitus position with the head of the bed elevated approximately 30 degrees. We paid particular attention throughout the study to ensure that transducer placement in the left parasternal interspaces was such that the ultrasound beam was oriented perpendicularly to the valve during recordings of the mitral valve. Standard techniques were used to obtain views of the mitral valve, the left ventricle, and the left atrium and aortic root. Echocardiograms were recorded at a paper speed of 50 mm per second using 2.25 mHz focused transducers and either an lrex System 3, a Picker 8OC, a Smith-Kline Ekoline echocardiograph with a Honeywell 1856A stripchart recorder, or an ATL 3001C or a Diasonlcs DRF 400 echocardiograph with Tektronix recorder. The echocardiographic results were interpreted independently of clinical information. Systolic motion of the mitral valve was evaluated on all strips recording the mitral valve on each study, using standard criteria for diagnosis of mitral prolapse [21-231. Mitral valve prolapse was considered to be present if continuous interfaces representing mitral valve leaflets showed both “turning around” with posterior motion (away from the transducer) and movement at least 2 mm behind the line connecting the C and D points of the mitral valve. Late systolic prolapse was diagnosed when the valve “turned around” in midsystole, whereas holosystolic prolapse was diagnosed when posterior motion began with the onset of systole. The depth of prolapse was measured from the C-D line to the first mitral leaflet interface, thus excluding the possibility of measuring to a chordal or left atrial wall echo. Echocardiographic results were not considered to indicate mitral valve prolapse if they demonstrated only minimal motion behind the C-D line, late systolic “flattening” of mitral leaflet motion without actual posterior motion, or discontinuous lines that appeared to prolapse. Transient posterior motion of mitral valve leaflets immediately after mitral valve closure was not considered to be mitral valve
SUBJECTS AND METHODS Study Design. We used the case-control method. All subjects underwent M-mode echocardiography. Mitral prolapse was diagnosed using standard M-mode echocardiographic criteria by investigators blinded to clinical findings. The prevalence of mitral valve prolapse in each patient group was compared with that in control groups to derive overall odds ratios. Because of the known effect of age and sex on the prevalence of mitral valve prolapse in unselected populations, odds ratios were recalculated after each patient was age- and sex-matched to two subjects from each control population. Differences in age, sex, and presence of heart murmurs between patients with complications and another control group of unselected subjects with uncomplicated mitral valve prolapse were also assessed. Subjects. Consecutive patients at the New York HospitalCornell Medica! Center with any of the three following conditions were studied. They were as follows: All patients with conclusive evidence of infective endocarditis on native heart valves between November 1978 and December 1982, who underwent M-mode and twodimensional echocardiogrephy and had complete medical records available, were enrolled. All but one patient had bacteriologic identification of the infecting organism, whereas the lone patient with culture-negative endocarditis presented with fever, arthralgias, petechial rash, anemia, hematurla, erythrocyte sedimentation rate of 128, and a latex fixation test result positive at 15,120. All patients with isolated, pure, moderate to severe mitral regurgitation by clinical criteria enrolled in a prospective natural history study, between December 1980 and July 1983, were enrolled in this study. Each subject underwent complete clinical evaluation and M-mode and two-dimensional echocardiography as part of the study protocol. Evidence of hemodynamically important mitral regurgitation included the presence of a holosystolic or nearly holosystolit murmur of mitral regurgitation, a third heart sound, and enlargement of the left ventricle and left atrium on the basis of physical and echocardiographic findings. Further proof of the severity of mitral regurgltation was provided in most patients by cardiac catheterization or Doppler echocardiography. Twelve-lead electrocardiography, treadmill exercise testing using the Bruce protocol, and radionuclide cineangiograpy at rest and during exercise [ 181, as well as cardiac catheterization when clinically indicated, were performed to exclude ischemic heart disease as the cause of mitral regurgitation. Consecutive patients with ruptured mitral chordae tendineae diagnosed by M-mode and two-dimensional echocardiographic criteria between 1976 and 1982 were enrolled. Full clinical records were available for each patient as well as anatomic confirmation of mitral chordal abnormality in 20 patients at the time of cardiac surgery or autopsy. To minimize potential bias introduced by selection of control subjects, we used three independent groups. They
752
November
1986
The American
Journal
of Medicine
Volume
81
COMPLICATIONS
TABLE I
OF MITRAL
Number Age (years)* Men (percent) Women (percent) f
ET AL
ControlGroups
Groups Mitral Regurgitation
Patient
Mean
PROLAPSE-DEVEREUX
Clinical Characteristics of Patient and Control Groups lnleclive Endocarditis
l
VALVE
standard
67
3af
31
14
49 (73) 18 (27)
50f
14
17 (55) 14 (45)
Ruptured Chordae
Population
Clinical
43 56f 17 30 (73) 13 (27)
196 446 13 106 (54) 90 (46)
2146 54% 19 923 (43) 1,223 (57)
Relativeswith Prolapse
Mitral
81
38 f 14 25 (31) 56 (69)
deviation.
prolapse. Posterior motion of the mitral leaflets after the time when the left ventricular internal dimension reached its minimum and gradual posterior motion of the mitral leaflets reaching a midsystolic nadir were also not considered to represent mitral valve prolapse. Echocardiographic findings of ruptured mitral chordae were sought on both M-mode and two-dimensional studies. M-mode manifestations of chordal rupture included the following: abnormal systolic appearance in the left atrium of mitral leaflet or chordal echoes; chaotic diastolic fluttering of the anterior mitral leaflet, more irregular than usually seen with aortic regurgitation; systolic fluttering of the mitral valve; paradoxic anterior motion of the posterior mitral leaflet in early diastole; paradoxic movement of the posterior leaflet both anteriorly in systole and posteriorly in diastole; and irregular early diastolic fluttering of the posterior leaflet. Ruptured chordae and accompanying flail leaflet were identified on two-dimensional echocardiography by the presence of abnormal systolic mitral coaptation with systolic whipping of the involved leaflet into the left atrium [ 24) in either the long axis or the apical four-chamber view. In the presence of ruptured chordae tendineae, mitral prolapse was diagnosed only if two-dimensional echocardiography showed the free edge of the affected mitral leaflet to be directed toward the left ventricle, with the mid-portion of the leaflet billowing farther into the left atrium. In one patient, the protrusion of a chorda tendinea with an attached mass into the left ventricular outflow tract in systole was considered indicative of chordal rupture. Statistical Analysis. To approximate the relative risk of each complication occurring in subjects with mitral valve prolapse as opposed to those without, we calculated the odds ratio [25]. The data used to derive the odds ratio for each comparison were the number of subjects with mitral prolapse among patients with a specific complication (A) and in the relevant control group (B) as well as the number of subjects without mitral valve prolapse in the patient and control groups (C and D, respectively). The odds ratio was then calculated as A X D/B X C. This approach was also used to approximate the relative risk of complications in subgroups of patients with mitral prolapse defined by simple demographic or clinical findings (e.g., male sex or presence of a heart murmur). Statistical significance differences was assessed using the chi-square test, with Yates’ continuity correction when appropriate [ 261. Approximate confidence limits of odds ratios were calculated by Woolf’s method [27]. Because age and sex are known to influence the preva-
November
lence of mitral valve prolapse [20,28,29], we recalculated the odds ratios after matching each patient to two subjects of the same sex and within five years in age from each control group. Assessment of the odds ratio for each complication in subjects with mitral valve prolapse from these case-control triplets was performed by the Mantel-Haenszel test with calculation of approximate confidence intervals [30,31]. Statistical significance was assessed by the chi-square test with one degree of freedom. Differences between groups with regard to continuous variables were assessed by the unpaired Student t test [26].
RESULTS Characteristics
of Patient and Control Groups.
The
number of subjects in each group, their age, and the proportion of men and women are given in Table I. The evident differences in age and sex distribution among groups were taken into account in the analyses of ageand sex-matched case-control triplets, the results of which are presented next. A murmur of mitral regurgitation was known before present illness in six of 11 patients with mitral prolapse and endocarditis (55 percent), and before study enrollment in 22 of 8 1 affected family members (27 percent); late or holosystolic mitral regurgitation murmurs were present during hospitalization in 9 of 11 patients with endocarditis (78 percent), whereas late systolic murmurs were audible in 27 affected relatives (33 percent) and a holosystolic murmur was audible in one (1 percent) additional relative with mitral valve prolapse. The most frequent indications for echocardiography among clinical control subjects were an explicit diagnosis of mitral prolapse or suspicion of this condition (16 percent), established or suspected diagnoses of other forms of valvular heart disease (24 percent), pericardial disease (10 percent), atrial fibrillation or other arrhythmia (7 percent), and congestive heart failure or cardiac enlargement (7 percent).
Odds Ratios for Complications in Subjects with Mitral Valve Prolapse: Overall Analyses (Table II). The prevalence of mitral valve prolapse was virtually identical in our population and clinical control groups (eight of 196 or 4.1 percent and 84 of 2,146 or 3.9 percent, respectively; difference not significant). Mitral prolapse was slightly more common among women (five of 90 or 5.6 percent) than men (three of 106 or 2.8 percent) among the popula-
1986
The American
Journal
of Medicine
Volume
81
753
COMPLICATIONS
OF MITRAL VALVE PROLAPSE-DEVEREUX
TABLE II
Association of Mitral Prolapse with Complications: Overall Analysis
Number Mitral prolapse (percent) No mitral prolapse (percent) Odds ratio versus population control group Odds ratio versus clinical control group Statistical
significance:
* p
ET AL
lnfeclive Endocardilis
Mitral Regurgitation
Ruptured Chordae
Population Control Group
Clinical Control Group
67 11 (16) 56 (84) 4.6’
31 17 (55) 14 (45) 28.5”
43 27 (63) 16 (37) 39.6”
196 8 (4) 188 (96) -
2,146 84 (4) 2,062 (96) 1.0
29.8”
41.4”
1.0
-
4.8” versus
population
(employed)
and clinical
tion control group, whereas prevalences among women and men in the clinical control group were more nearly equal (52 of 1,233 or 4.3 percent versus 32 of 923 or 3.5 percent). In contrast, the prevalence of echocardiographic mitral prolapse was significantly higher (p
754
November
1986
The American
Journal
of Medicine
control
groups.
tis for age and sex with two subjects from our population and clinical control groups. This procedure reduced the prevalence of mitral valve prolapse in each control population to 2 percent and increased the odds ratio for mitral valve prolapse in endocarditis to 6.7. We believe that this change in the odds ratio was due to sex- and not agematching. Approximate 95 percent confidence limits of this odds ratio are 1.9 to 22.9. In contrast, age- and sex-matching of control subjects to patients with mitral regurgitation (Table IV) did not systematically alter the prevalence of mitral valve prolapse among control groups from the 4 percent observed in our overall analysis, or the high odds ratios for this condition in subjects with mitral valve prolapse. Similarly, age- and sex-matching of control subjects to patients with ruptured chordae tendineae (Table V) had no effect on the dramatic odds ratios for mitral valve prolapse observed in patients with ruptured mitral valve chordae tendineae. Relation of Complications to Demographic and Clinical Features of Mitral Valve Prolapse. To assess whether patients with mitral valve prolapse at high risk of complications could be identified on the basis of simple demographic or clinical features, we next compared the patients with mitral valve prolapse in each group with complications to unselected subjects with mitral valve prolapse identified during family studies. Table VI compares the age of the 11 patients with mitral valve prolapse and endocarditis, the 17 with mitral prolapse and mitral regurgitation, and the 27 with mitral prolapse and ruptured mitral chordae to that of the 8 1 first-degree relatives with mitral prolapse in our family studies. Patients with mitral prolapse and endocarditis were similar in age to the unselected subjects with mitral prolapse, whereas patients with mitral regurgitation or ruptured chordae and mitral prolapse were significantly older. Table VII shows the proportion of men among persons with mitral valve prolapse in our patient groups and unselected subjects. Men predominate in each patient group, whereas two-thirds of the unselected subjects with mitral valve prolapse are women, The highest odds ratio for men among patient groups with mitral prolapse occurred with ruptured chordae tendineae (7.4) and infective endocardi-
Volume
61
COMPLICATIONS
TABLE
III
Association
of Mitral Prolapse
with Infective
Endocarditis:
OF MITRAL
Matched-Triplets
Patients With infective Endocarditis Number Mitral prolapse (percent) No mitral prolapse (percent) Odds ratio versus population control group Odds ratio versus clinical control group Statistical
TABLE
significance
IV
by Mantel-Haenszel
Association
test:
of Mitral Prolapse
95 percent
confidence
limits:
Statistical
significance
TABLE V
by Mantel-Haenszel
Association
test:
Matched-Triplets
of Mitral Prolapse
95 percent
with Ruptured
62 1 (2) 61 (96) limits:
Tendineae:
t 9.1 to 126, t 6.3 to 87.8.
Matched-Triplets
Patients with Ruotured Chordae Number Mitral prolapse (percent) No mitral prolapse (percent) Odds ratio versus population control group Odds ratio versus clinical control group Statistical
significance
TABLE VI
by Mantel-Haenszel
Age of Subjects
test:
with Complicated
95 percent
86 4 (5) 82 (95) 0.3
confidence
and Uncomplicated
Analysis
Matched Control Groups Population Clinical
43 27 (63)” 16 (37) 17*+ 53*$
lp
62 3 (5) 59 (95) 3.1 -
0.4
confidence
Chordae
Analysis Matched Control Groups Population Clinical
31 17 (55)’ 14 (45) 34.0*+ 23.7”t “p
134 3 (2) 131 (96) 1.0 -
t 1.96 to 22.9.
Patients with Mitral Regurgitation Number Mitral prolapse (percent) No mitral prolapse (percent) Odds ratio versus population control group Odds ratio versus clinical control group
Analysis
134 3 (2) 131 (96) 1.0
with Mitral Regurgitation:
ET AL
Matched Control Groups Population Clinical
67 11 (16)” 56 (64) 6.7*+ 6.7”+
lp
VALVE PROLAPSE-DEVEREUX
limits:
t 7.2 to 43.5,
86
2 (2) 84 (98) 3.0 t 16.7 to 168.0.
Mitral Prolapse
Patients with Mitral Prolapse
Number Age range Mean age” p value versus
relatives
with mitral
prolapse
Infective Endocarditis
Mitral Regurgitation
Ruptured Chordae Tendlneae
Control Group Relatives with Prolapse
11 19-60 37 f 13 NS
17 20-72 48f 14 < 0.005
27 23-79 55f 16
81 16-71 38% 14 -
NS = difference not significant. f standard deviation. l
TABLE VII
Sex of Subjects
with Complicated
and Uncomplicated
Mitral Prolapse
Patients with Mitral Prolapse
Number Male (percent) Female (percent) Odds ratio versus Statistical
significance:
relatives
with mitral
* p
prolapse
Infective Endocarditis
Mitral Reauraitalion
Ruptured Chordae Tendineae
Control Group Relatives with Prolapse
11 8 (73) 3 (27) 6.0*
17 9 (53) 8 (47) 2.5
27 20 (74)T 7 (26) 7.4+
81 25 (31) 56 (69) -
+ p
November
1988
The
American
Journal
of Medicine
Volume
81
755
COMPLICATIONS
TABLE VIII
OF MITRAL
Relation
VALVE
PROLAPSE-DEVEREUX
of Infective
ET AL
Endocarditis
to Previous
or Current
Murmur
Mitral Prolapse with Infective Endocardilis Number Previously known murmur (percent) No known murmur (percent) Odds ratio Present murmur (percent) No present murmur (percent) Odds ratio Statistical
significance:
l
difference
not significant;
11 6 (56) 5 (45) 3.2” 9 (82) 2 (18)
The American
Journal
28 (34) 53 (66)
clinical practice and medical reports [1-4,9-l 61. Two previous studies assessed the relative risk of infective endocarditis occurring in persons with mitral valve prolapse by comparison with clinical control groups. Clemens et al [ 153 found an odds ratio of 4.9 based on comparison of their entire endocarditis and clinical control groups, and an odds ratio of 8.2 based on matched-pairs analysis, This is nearly identical to our results. Hickey et al [ 161 reported a similar odds ratio (5.3) using only clinical control subjects. However, doubt has persisted about these estimates of the relative risk of infective endocarditis in subjects with mitral prolapse because the prevalence of the condition in clinical control groups might have been distorted by selection biases. Our study establishes the accuracy of the estimate of a five- to seven-fold increase in risk of endocarditis in persons with mitral valve prolapse by finding nearly identical odds ratios when our patients with infective endocarditis were compared with control subjects drawn from a general adult population, and with a second clinical control group. Further analysis of our data indicates that men with mitral valve prolapse have a greater risk of infective endocarditis than affected women. This finding is compatible with the results of most [1,3,4,32,33] but not all [2] previous studies. The increase in odds ratios for endocarditis in subjects with mitral prolapse between overall and matched-pair analyses in the study of Clemens et al [ 151, similar to our findings, is also compatible with a male predominance. Our data also support an increased risk of infective endocarditis in patients with mitral prolapse and a murmur of mitral regurgitation, in keeping with the suggestion by Hickey et al [ 161. However, in this retrospective study, it was not possible to define precisely the odds ratio for endocarditis in subjects with mitral valve prolapse who did or did not have murmurs, with estimates ranging from 3.2 (difference not significant) based on the history of an antecedent murmur to 8.5 based on the presence of a mitral regurgitant murmur at the time of evaluation (p
Previous studies, mostly uncontrolled, have supported associations between mitral valve prolapse and several complications, including infective endocarditis, hemodynamically important mitral regurgitation, rupture of mitral valve chordae tendineae, and neurologic ischemic episodes. The present controlled study assessed the relation between mitral valve prolapse and the first three of these complications, finding highly significant associations (p
1986
ai 22 (27) 59 (73)
t p
COMMENTS
November
Relatives with Mitral ProlaDse
8.5t
tis (6.0). An additional observation in support of the apparent greater increase in risk of infective endocarditis among men as opposed to women with mitral prolapse is that one of the 25 men in our family mitral prolapse control group was hospitalized with bacterial endocarditis in February 1985, after completion of these data analyses. We examined a clinical feature previously suggested by others to be useful in identifying patients with mitral valve prolapse at risk for endocarditis-the presence of mitral regurgitant murmur. To avoid confounding effects of damage to mitral valve structure during endocarditis, we first assessed the prevalence in both groups of a history of a heart murmur known before the onset of present illness or enrollment in our family study. Only a modest and statistically insignificant association (odds ratio 3.2) is seen between a previously known murmur and endocarditis (Table VIII). However, when the proportion of patients with mitral prolapse and control subjects who exhibited late or holosystolic murmurs of mitral regurgitation during hospitalization or at the time of enrollment in our family study was compared, the odds ratio for endocarditis in patients with mitral prolapse who had as opposed to those who did not have heart murmurs rose to 8.5 and attained statistical significance at the p
756
in Mitral Prolapse
of Medicine
Volume
81
COMPLICATIONS
infection. Our finding that the prevalence of mitral Valve prolapse demonstrated by well-standardized M-mode echocardiographic criteria was approximately 4 percent in both our control groups-at the lower end of the accepted range for this condition but in keeping with prevalences of 2 to 4 percent recently reported by several investigators using similar methods [34-36]-shows that the proportion of the general population in whom endocarditis prophylaxis might be indicated for mitral prolapse is somewhat lower than previously thought. Taken together with the recent simplification of American Heart Association recommendations for antibiotic prophylaxis [37], these finding have implications for clinical practice. First, the somewhat lower population prevalence of mitral prolapse (4 percent instead of 5 to 7 percent) and the abbreviated course of antibiotics now recommended would reduce the expense and medical risks of routine antibiotic prophylaxis in subjects with mitral prolapse to below those estimated in recent costeffectiveness analyses [38-401. Second, our demonstration of a higher relative risk of endocarditis in men than women with mitral prolapse, and the evidence in favor Of a higher risk in patients with a murmur of mitral regurgitation in our study and that of Hickey et al [ 161, suggests that risk of endocarditis can be usefully compartmented. Thus, odds ratios of infective endocarditis occurring in men and women with mitral valve prolapse were 6.7 and 3.4, respectively, when our population control subjects were used as the reference standard, and 7.6 and 2.7, respectively, when our patients with infective endocarditis were compared with general population data from the Framingham Study [28], in which mitral prolapse was diagnosed by similar criteria, with cross-evaluation of echocardiographic results by one of us (R.B.D.). If, as our family studies suggest, the prevalence of late systolic murmurs is similar in men and women with mitral valve prolapse, it may be possible to approximate the risk of endocarditis on the basis of the combination of sex and the presence or absence of a murmur of mitral regurgitation. Projected odds ratios Compared with those for normal persons would range from 2.0 in women without heart murmurs to as high as 12.5 in men with murmurs of mitral regurgitation if our odds ratio for endocarditis in
OF MITRAL
VALVE PROLAPSE-DEVEREUX
ET AL
mitral prolapse based on a history of previous murmurs (3.4) is used, and from 1.0 to 15.7 in the same groups using the higher odds ratio (8.5) based on actual presence of a murmur at the time of study. Thus, our estimates of risk based on sex and presence or absence of a murmur, supported by the data of Clemens et al [ 151 and Hickey et al [ 161, appear to clarify decision-making for antibiotic prophylaxis: nearly half the patients with mitral prolapse-women without a murmur of mitral regurgitation-would fall into a group with too little risk to warrant this intervention, and an additional 10 percent (men with murmurs) would fall into an unequivocal high-risk group. Our data also confirm that strong associations, with odds ratios from 17 to 53, exist between mitral valve prolapse and clinically significant mitral regurgitation and echocardiographic findings of rupture of mitral chordae tendineae, as suggested in previous studies [5-131. Although the proportion of patients with mitral regurgitation or ruptured chordae who exhibited mitral prolapse was higher in our study than reported by some [5,7], it was lower than the proportion of patients with anatomically redundant mitral valves in two recent studies of the same complication. Thus, Waller et al [6] reported that 48 of 65 (74 percent) patients with isolated, pure, severe mitral regurgitation not due to ischemic heart disease had enlarged mitral leaflets and annuli, whereas Hickey et al [8] found anatomically “floppy” valves in 93 percent of patients in whom ruptured chordae were found at cardiac surgery. The clustering of hemodynamically important mitral regurgitation and ruptured chordae tendineae in older persons with mitral valve prolapse is compatible with the hypothesis [4 1,421 that these complications are related to “wear and tear” of an intrinsically abnormal mitral valve. The male predominance among our patients with these complications of mitral prolapse, similar to that observed by Lucas and Edwards [32] in a necropsy study, might be attributed to the greater hemodynamic load placed on the mitral valve by the known higher systolic blood pressure in men [43]. ACKNOWLEDGMENT We would like to thank Virginia Burns for her assistance in preparation of this manuscript.
REFERENCES 1.
2.
3.
LeBauer EJ, Perloff JK, Keliher TF: The isolated systolic click with bacterial endocarditis. Am Heart J 1967; 73: 534-537. Lachman AS, Bramwell-Jones DM, Lakier JB, Pocock WA, Barlow JB: infective endocarditis in the billowing mitral leaflet syndrome. Br Heart J 1975; 37: 326-330. Corrigall D, Bolen J, Hancock EW, Popp RL: Mitral valve prolapse and infective endocarditis. Am J Med 1977; 64: 215-222.
November
4.
5.
6.
1986
Nolan CM, Kane JJ, Grunow WA: infective endocarditis and mitral prolapse. Comparison with other types of endocarditis. Arch Intern Med 1981; 141: 447-450. Guy FC, MacDonald RPR, Fraser DB, Smith ER: Mitral valve prolapse as a cause of hemodynamically important mitral regurgitation. Can J Surg 1980; 23: 166-170. Waller BF, Maron BJ, DelNegro AA, Gottdiener JS, Roberts WC: Frequency and significance of M-mode echocardiographic evidence of mitral valve prolapse in clinically
The
American
Journal
of Medicine
Volume
61
757
COMPLICATIONS
7.
6.
9.
10.
Il.
12.
13.
14.
15.
16.
17. 18.
19.
26.
21.
22.
23.
24.
758
OF MITRAL
VALVE
PROLAPSE-DEVEREUX
ET AL
isolated pure mitral regurgitation: analysis of 65 patients having mitral valve replacement. Am J Cardiol 1964; 53: 139-147. Oltveira DBG, Ddwkins KD, Kay PH, Paneth M: Chordal rupture. I: Aetiology and natural history. Br Heart J 1963; 50: 312-317. Hickey AJ, Wilcken DEL, Wright JS, Warren BA: Primary (spontaneous) chordal rupture: relation to myxomatous valve disease and mitral valve prolapse. J Am Coll Cardiol 1965; 5: 1341-1346. 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: 525-532. Mills P, Rose J, Hollingsworth J, Amara I, Craige E: Longterm prognosis of mitral-valve prolapse. N Engl J Med 1977; 297: 13-18. Appelblatt N, Willts P, Lenhart J: Ten to 45 year follow-up df 69 patients with systolic click with or without apical late systolic murmur (abstr). Am J Cardiol 1975; 35:‘119. Belardi J, Lardani H, Manubeus S, Sheldon WC, Moreyon A: fdiopathic prolapse of the mitral vaive: a follow-up study In 137 patients studied by angiography (abstr). Am J Cardiol 1976; 37: 120. Koch FH, Hancock EW: Ten year follow-up of forty patients with midsystolic click/late systolic murmur syndrome (abstr). Am J Cardiol 1978; 37: 149. Nishimura RA, McGoon MD, Shub C, Miller FA Jr, llstrup DM, Tajik AJ: Echocardiographically documented mitral valve prolapse. Long-term follow-up of 237 patients. N Engl J Med 1985; 313: 13051309. Clemens JD, Horwitz RI, Jaffe CC, Feinsteih AR, Stanton BF: A controlled evaluation of the risk of bacterial endocarditis in persons with mitral valve prolapse. N Engl J Med 1982; 307: 776-781. Hickey AJ, MacMahon SW, Wilcken DEL: Mitral valve prolapse and bacterial endocarditis: when is antibiotic prophylaxis necessary? Am Heart J 1985; 169: 431-435. Leatham A, Brigdeh W: Mild mitral regurgitation and the mitral prolapse fiasco. Am Heart J 1980; 99: 659-664. Borer JS, Bacharach SL, Green MV, et al: Real-time radionuelide cineangiography in the non-invasive evaluation of global and regional left ventricular function at rest and during exercise in patients with coronary artery disease. N Engl J Med 1977; 296: 839-844. Alderman MH, Schoenbaum EE: Detection and treatment of hypertension at the work site. N Engl J Med 1975; 293: 65-68. Devereux RB, Brown WT, Kramer-Fox R, Sachs I: Inheritance of mitral valve prolapse: effect bf age and sex on gene expression. Ann Intern Med 1982; 97: 826-832. Popp RL, Brown OR, Silverman JF, Harrison DC: Echocardiographic abnormalities in the mitral valve prolapse syndrome. Circulation 1974; 49: 428-433. 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: 1091-1096. Devereux RB, Brown WT, Sachs I, Kramer-Fox R: Evolution of echocardiographic findings in patients with mitral valve prolapse. J Cardiovasc Ultrasonogr 1983; 2: 29234. Mintz GS, Kotler MN, Segal BL, Parry WR: Two-dimensional
November
1986
The American
Journal
of Medicine
25.
26. 27.
28.
29.
30.
31.
32. 33.
34.
35.
36.
37.
38.
39.
40.
41. 42.
43.
Volume
81
echocardiographic recognition of ruptured chordae tendineae. Circulation 1978; 57: 244-250. Schlesselmann JJ: Case-control studies: design, conduct, analysis. New York: Oxford University Press, 1982; 3255. Armkage P: Statistical methods in medical research. New York: Wiley, 1971. Schlesselmann JJ: Case-control studies: design, conduct, analysis. New York: Oxford University Press, 1982; 176178. Savage DD, Garrison RJ, Devereux RB, et al: Mitral valve prolapse in the general population. I. Epidemiologic features. The Framingham Study. Am Heart J 1983; 106: 571-576. Strahan NV, Murphy EA, Fortuin NJ, Come PC, Humphries JO’N: Inheritance of the mitral valve prolapse syndrome: discussion of a three-dimens/onal penetrance model. Am J Med 1983: 74: 967-972. Mantel N, Haenszel W: Statistical aspects of the analysis of data from retrospective studies of disease. JNCI 1959; 22: 719-748. Schlesselmann JJ: Case-controls studies: design, conduct, analysis. New York: Dxford University Press, 1982; 2i2216. Lucas RV JR, Edwards JE: The floppy mitral valve. Curr Probl Cardiol 1982; 7: l-48. Durack DT, Kaplan EL, Bisno AL: Apparenf failures of endocarditis prophylaxis. Analysis of 52 cases submitted to a national registry. JAMA 1983; 250: 2318-2322. Hickey AJ, Wolfers J, Wilcken DEL: Mitral valve prdlapse. Prevalence in an Australian population. Med J Aust 1981; 1: 31-33. Wann LS, Grove JR, Hess TR, et al: Prevalence of mitral valve prolapse by two-dimensional echocardiography in healthy young women. Br Heart J 1983; 49: 334-340. Bryhn M, Perrson S: The prevalence of mitral valve prolapse in healthy men and women in Sweden. Acta Med S&d i984; 215: 157-160. 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-1127A. Bor DH, Himmelstein DU: Endocarditis prophylaxis for patients with mitral valve prolapse. A quantitative analysis. Am J Med 1984; 76: 711-717. Retchin SM, Fletcher RH, Waugh RA: Endocarditis and mitral valve prolapse: what is the risk? Int J Card&l 1984; 5: 653-659. Clemens JD, RanSohoff DF: A quantitative assessment of pre-denfal antibiotic prophylaxis for patients with mitral valve prolapse. J Chronic Dis 1984; 37: 531-544. Devereux RB, Perloff JK, Reichek N, Josephson ME: Mitral valve prolapse. Circulation 1976; 54: 3-14. Hickey AJ, Wilcken DEL: Age and the clinical profile of idiopathic mitral valve prolapse. Br Heart J 1986; 55: 582-586. Rowland ML, Fulwood R: Corbnary heart disease risk factor trends in blacks between the first and second National Health and Nutritional Examination Surveys, United States, 1971-1980. Am Heart J 1984; 108: 771-779.