Introduction

Mitral valve prolapse (MVP) is the most common valve disease, with an estimated prevalence of 2% to 3% in the general population.¹ Although MVP is generally regarded as a benign condition,^2,3 the outcome is widely heterogeneous, and complications such as mitral regurgitation, atrial fibrillation, congestive heart failure, endocarditis, and stroke are well known. Ventricular arrhythmias and sudden cardiac death (SCD) have been reported.^4–7

Editorial see p 551

Clinical Perspective on p 566

From a pathological anatomy viewpoint, accumulation of proteoglycans (myxomatous mitral valve) is the most common cause of MVP, accounting for leaflet thickening and redundancy, chordal elongation, interchordal hoodings, and annular dilatation.⁸ Although these valve abnormalities well explain mitral regurgitation and mechanical complications caused by enhanced extensibility, the pathogenesis of ventricular arrhythmias/SCD in MVP remains controversial.

The estimated rate of SCD in MVP ranges from 0.2%/y to 0.4%/y in prospective follow-up studies.⁴ Left ventricular (LV) dysfunction resulting from severe mitral regurgitation identifies a patient subgroup at high risk of SCD.⁹ However, life-threatening ventricular arrhythmias also occur in patients with MVP with trivial or absent mitral regurgitation.¹⁰ Previous pathology studies of SCD focused mostly on mitral valve or conduction system abnormalities as the cause of electric instability,^8,11–17 and demonstration of a myocardial source of arrhythmias remained elusive.^18–20

The aim of our report is to demonstrate that MVP is a significant cause of SCD and life-threatening arrhythmias in young adults owing to an underlying myocardial substrate, which is detectable by contrast-enhanced cardiac magnetic resonance (CE-CMR) and may serve for risk stratification and SCD prevention.

Methods

Study Populations

SCD Victims With MVP

In the time interval from 1982 to December 2013, all the hearts of SCD victims ≤40 years of age in the Veneto region in northeast Italy (geographic area, 18 368 km²; overall population, 4 857 210 according to the Italian Census Bureau 2011) were collected, pathologically investigated, and preserved. SCD is defined here as witnessed sudden and unexpected death occurring within 1 hour of the onset of symptoms or death of an individual who had been seen in stable condition <24 hours before being found dead.^21,22 Demographic, clinical, and pathological data were recorded in the electronic database of the Registry of Cardio-Cerebro-Vascular Pathology, which acts as referral center for SCD of the northeast Italy.

Charts were evaluated for age, sex, symptoms, and clinical history. All the hearts were re-examined carefully according to a standardized protocol.²¹ SCD cases were selected in whom MVP caused by myxomatous valve disease was the only cardiac abnormality found at autopsy. Myxomatous valve disease is defined as increased leaflet length and redundancy, with interchordal hoodings and leaflet billowing toward the left atrium and chordae tendineae elongation.⁸ In the absence of extracardiac (cerebral, respiratory) or mechanical cardiovascular explanations, the cause of death was considered cardiac arrhythmic.

Exclusion criterion was clinical or pathological evidence of more than mild mitral regurgitation. Hearts from 15 sex- and age-matched patients (10 female; mean age, 30 years; range, 18–40 years) who died suddenly as a result of extracardiac causes (8 cerebral and 7 respiratory) served as control subjects.

MVP Patients With Complex Ventricular Arrhythmias

The study included consecutive patients referred to the Cardiology Clinic from January 2010 to December 2013 with complex ventricular arrhythmias detected on 12-lead 24-hour Holter monitoring and echocardiographic diagnosis of MVP, defined as >5-mm thickening and >2-mm displacement of 1 or both mitral leaflets into the left atrium as viewed in the LV outflow tract orientation.²³ Twelve-lead ECG 24-hour Holter monitoring was requested because of the presence of either arrhythmic symptoms or 12-lead ECG changes. Complex ventricular arrhythmias consisted of ventricular fibrillation and ventricular tachycardia (VT), either nonsustained or sustained.²⁴ Patients with complex ventricular arrhythmias were further subdivided into 2 groups: those with a run of 3 ventricular premature beats (VPBs) and those with a run of >3 VPBs.

The control group consisted of patients with MVP with minor ventricular arrhythmias, that is, isolated VPB, couplets, and bigeminal VPB.

Exclusion criteria were significant mitral regurgitation, tricuspid dysplasia or regurgitation, cardiomyopathies or congenital heart abnormalities, hemodynamic unstable conditions, and contraindication to CMR. The study was approved by the institutional review board, and all patients gave informed consent.

Protocols of Investigation

Pathological Anatomy Study

Formalin-fixed hearts were restudied according to a protocol previously reported.²¹ Leaflet involvement (whether anterior, posterior, or bileaflet) and the presence of endocardial fibrous plaque (friction lesion) on the LV inferobasal wall were assessed. Multiple samples of the LV and right ventricular free walls and septum, including the papillary muscles (PMs), were obtained for histology. Additional samples were taken in the LV inferobasal free wall, under the posterior mitral leaflet. Sections 5 μm thick were stained with hematoxylin-eosin, Weigert–van Gieson, Heidenhain trichrome, and Alcian–periodic acid–Schiff. Morphometric analysis was performed with an Image-Pro Plus program (version 4.0. Media Cybernetics) to quantify the fibrous tissue percent area of LV myocardium on Heidenhain trichrome–stained sections at ×25 magnification. Mean cardiomyocytes diameter was calculated on hematoxylin-eosin–stained sections at ×400 magnification. Quantitative analysis was performed by 2 blinded expert pathologists (C.B. and S.R.) with an interobserver variability <5%.

Clinical Study

All patients underwent cardiovascular evaluation that included history, physical examination, 12-lead ECG, 2-dimensional transthoracic echocardiography, 12-lead 24-hour Holter monitoring, and CE-CMR. Coronary angiography was performed in selected cases. The 12-lead ECG at rest and the 24-hour Holter monitoring were independently assessed as previously reported²⁵ by 2 experienced observers (M.D.L. and D.C.) who were blinded to patient data.

Nonsustained VT was defined as ≥3 consecutive VPBs with a rate >100 bpm that lasted <30 seconds during 24-hour Holter monitoring. Sustained VT was defined as tachycardia originating in the ventricle with a rate >100 bpm and lasting >30 seconds or requiring an intervention for termination.

CMR was performed on a 1.5-T scanner (Magnetom Avanto, Siemens Medical Solutions, Erlangen, Germany). All patients underwent a detailed CE-CMR study protocol as previously described.²⁵ The presence and location of late gadolinium enhancement (LGE) were independently assessed by 2 experienced observers (M.P.M. and B.G.) who were blinded to clinical data. To exclude artifact, LGE was deemed present only if visible in 2 orthogonal views (long axis and short axis). LGE was identified by the use of a signal intensity threshold of >5 standard deviation (SD) above a remote reference region and quantified according to a previously reported method.

Statistical Analysis

Data are expressed as mean±SD or median with 25th to 75th percentiles for normally distributed and skewed variables, respectively. Normal distribution was assessed with the Shapiro-Wilk test. Categorical differences between groups were evaluated by the χ² test or the Fisher exact test as appropriate. Paired and unpaired t tests were used to compare normally distributed continuous variables obtained from the same patient and different patients, respectively; Wilcoxon signed-rank test (same patient) and Wilcoxon rank-sum test (independent samples) were used for skewed continuous variables.

A value of P<0.05 was considered significant. The minimal detectable effect at a significance level of 5% and power at 80% is equal to 1 (with a nonparametric test) for quantitative variables (the Cohen effect); for binary data, an odds ratio of at least 10 can be detected if the proportion of the characteristic test is equal to 10% in the no complex ventricular arrhythmias group (Fisher exact test). Statistics were analyzed with SPSS version 19 (SPSS Inc, Chicago, IL).

Results

SCD Victims With MVP

Among 650 consecutive young patients with SCDs recorded in the Veneto region registry, 43 patients (26 female; median age, 32 years; range, 19–40 years) with isolated MVP caused by myxomatous valve disease were identified. They represent 7% of all SCD patients and 13% of women who died suddenly, with MVP the first structural cause in the latter group. Main clinical and pathological data are reported in Table 1. SCD occurred mostly at rest or during sleep (n=35, 81%). Twenty patients (47%) had an in vivo diagnosis of MVP, with auscultatory click in 18 (90%) and palpitations in 14 (70%). Nine (21%) were taking β-blocker therapy for nonsustained ventricular arrhythmias. ECG was available in 12 (28%), showing negative/isodiphasic T waves on inferior lead in 10 of the 12 (83%; Figure 1A and 1B); all (100%) had right bundle-branch block (RBBB) morphology ventricular arrhythmias.

• Download figure
• Download PowerPoint

In SCD patients with MVP, valve leaflets were redundant, thick, and elongated, with either isolated posterior (n=13, 30%) or bileaflet (n=30, 70%) involvement (Figure 1C). The involvement of the posterior leaflet was diffuse in 23 (53%) and confined to the medial scallop in 20 (47%). Endocardial fibrous plaques in the posterolateral wall were found in 25 (58%).

Microscopic examination of the LV myocardium showed an increased endoperimysial and patchy replacement-type fibrosis at the level of PMs and adjacent free wall in all (Figures 1D, 1E, and 2). Similar findings with a subendocardial-midmural layer distribution were detected in the inferobasal wall under the posterior mitral valve leaflet in 38 patients (88%). The mean fibrous tissue percent area in MVP SCD victims was 30.5% at the level of PMs and 33.1% in the inferobasal wall myocardium (versus 6.3% and 6.4% in control subjects; P<0.001). In the same areas, the cardiomyocytes showed increased diameter (19.2±6.0 versus 12.8±0.4 μm; P<0.001) and dysmorphic and dysmetric nuclei.

• Download figure
• Download PowerPoint

MVP Patients With Complex Ventricular Arrhythmias

The baseline clinical and CMR findings are summarized in Table 2. Fourteen patients with MVP with or without minor ventricular arrhythmias (ie, isolated VPBs, couplets, and bigeminal VPBs) served as control subjects.

Thirty living patients with MVP (22 female; median age, 41 years) with complex ventricular arrhythmias, that is, ≥1 ventricular fibrillation (n=2, who had also nonsustained VT) and VT (n=28), either nonsustained (n=27) or sustained (n=1), were enroled. VT of an LV origin (RBBB morphology) was present in all, with either an inferior (43%) or a superior (87%) axis. Among the 27 patients with nonsustained VT, the mean length was 4 beats (range, 3–11 beats). Complex ventricular arrhythmias occurred at rest in 26 of 30 patients (87%). All patients had normal QTc (mean ms, 423; range, 409–440). Exercise stress test, performed in 20 patients, was negative for effort-induced ventricular arrhythmias.

Bileaflet MVP was present in 21 patients (70%) with complex ventricular arrhythmias compared with 5 control subjects (36%; P=0.031).

On postcontrast sequences, LV LGE was identified in 28 patients (93%) versus 2 (14%; P<0.001). By dividing the MVP population with complex ventricular arrhythmias into 2 subgroups, 20 patients had 3-VPB run and 10 patients >3-VPB run (P>0.05). No difference was found in terms of LV LGE between those with a 3-VPB run and those with a >3-VPB run (P>0.05).

The LGE was localized on the PMs in 25 patients (83%), with a midapical distribution in 16 and basal adjacent free wall in 24 cases, and on the LV inferobasal segment under the posterior leaflet in 22 (73%; Figure 3A–3D). A focal endocardial LGE in the same region, featuring a fibrous plaque, was found in 12 patients (40%).

• Download figure
• Download PowerPoint

The median LV LGE was 1.2% in MVP with versus 0% in MVP without complex ventricular arrhythmias (P<0.01).

Two patients with MVP experienced aborted SCD caused by ventricular fibrillation despite β-blocker therapy owing to previous sustained VT. Detailed invasive and noninvasive evaluation ruled out cardiac causes other than MVP. Both had RBBB-pattern ventricular arrhythmias with a superior axis and T-wave abnormalities on inferior leads. CE-CMR, performed 6 and 10 months before aborted SCD, revealed LV LGE of PMs and the inferobasal wall (Figure 3E and 3F). Both patients received an implantable cardioverter-defibrillator.

One patient had presyncopal episodes despite bisoprolol therapy. She underwent electrophysiological study with induction of sustained VT with the same RBBB morphology of VPBs (Figure 4A–4C). The CE-CMR showed a nonischemic LGE pattern in the LV inferobasal wall (Figure 4D). She also underwent cardioverter-defibrillator implantation.

• Download figure
• Download PowerPoint

Of the 3 patients with implantable cardioverter-defibrillator (mean follow-up, 10 months), 2 had nonsustained VT: 1 patient with spontaneous interruption and 1 patient who required antitachycardia pacing.

Discussion

MVP is an underrecognized cause of SCD in young adults, accounting for 7% of total fatal events and 13% of female victims in our large cardiac registry experience. The patient with MVP and ventricular arrhythmias at risk of SCD is usually a young adult woman with a midsystolic click at auscultation, bileaflet involvement of the mitral valve, T-wave abnormalities on inferior leads, and RBBB-type or polymorphic ventricular arrhythmias on ECG. Clear-cut evidence of a substrate of electric instability in MVP is provided here for the first time and consists of myocardial scarring targeting the PMs and the inferobasal LV free wall under the posterior leaflet, well in keeping with the site of origin of RBBB-type ventricular arrhythmias. The LV myocardial fibrosis observed at histology in SCD victims was then confirmed in the clinical arm of the study, with evidence of LGE at CE-CMR in arrhythmic patients with MVP, thus pointing to a promising role of this noninvasive technique for risk stratification beyond traditional prognostic markers.

MVP: An Underappreciated Cause of SCD

The absence of uniform diagnostic criteria of MVP in the general and forensic pathology practice and the frequent consideration of this valve disease as an uncertain cause of SCD21 are major obstacles to provide data on the real burden of MVP according to a meta-analysis of published studies. With these shortcomings, the prevalence of MVP in pathology series of SCD in the young ranges from 0% to 24%.^20,22,26,27 With the adoption of strict criteria for the definition of myxomatous mitral valve, in the Veneto region SCD Registry, MVP accounted for 7% of all cases in young adults (<40 years of age) and 13% among women, representing the first structural cause in the latter group. The diagnosis can be established easily on macroscopic examination and then confirmed by routine histology, but it might be overlooked by superficial inspection, resulting in the heart being identified as normal. Our data are likely to change the current thinking about MVP as a benign condition and point to the need to draw the attention of forensic pathologists to an entity that has been largely underestimated so far.

Ventricular Arrhythmias in MVP

In MVP series with prolonged ECG recording, a variable prevalence of ventricular arrhythmias has been reported, reflecting the different MVP definitions, the populations studied, and the complexity of ventricular arrhythmias considered.^{5,9,10,28–35} In particular, the clinical evidence of hemodynamically important regurgitation greatly affects the occurrence of ventricular arrhythmias. However, the detection of MVP in survivors of life-threatening arrhythmias suggests that a true association between hemodynamically uncomplicated MVP and arrhythmic SCD may exist.⁹ Thus, we decided to focus on “pure” MVP, excluding MVP associated with valve incompetence and LV remodeling, not to defile the message by overreporting ventricular arrhythmias. Early electrophysiological studies demonstrated that the most common site of origin of VPB is the inferobasal portion of the LV.³⁶ In the recent study of malignant MVP by Sriram et al,³⁷ frequent VPBs originated from the outflow tract and PMs. Moreover, electrophysiology mapped the site of origin to the PMs, the LV outflow tract, and the mitral annulus, suggesting that VPBs arising close to the prolapsing leaflet and adjacent structures are the arrhythmic triggers.

From a pathophysiological perspective, the mechanism of ventricular arrhythmias in patients with MVP with trivial or absent mitral regurgitation remains speculative.^9,38 MVP-related factors have been advocated such as the excessive traction on the PMs by the prolapsing leaflets, the mechanical stimulation of the endocardium by the elongated chordae with afterdepolarization-induced triggered activity, the diastolic depolarization of muscle fibers in redundant leaflets with triggered repetitive automaticity, and the endocardial friction lesions with extension into the myocardium.^39–41 Moreover, the coexistence of extravalvular diseases has been suggested, including autonomic nervous system dysfunction,⁴² conduction system abnormalities,¹³ fibromuscular dysplasia of small coronary arteries,¹⁹ and occult cardiomyopathies.^10,43

The Myocardial Substrate of Electric Instability in MVP

Previous pathology studies in patients with MVP who died suddenly focused mostly on mitral valve structural alterations, suggesting a role for annular circumference, leaflet length and thickness, and the presence and extent of endocardial plaques.^8,14–17 Surprisingly, no investigation systematically addressed the LV myocardium to search for the substrate of electric instability except for few anecdotic cases.^{11,13,14,44,45} For the first time, we extended the histopathological investigation beyond the valve in all SCD cases and provided convincing evidence of fibrosis in the LV myocardium, which is closely linked to the mitral valve, that is, the PMs with adjacent free wall and the inferobasal wall. The LV myocardial scarring is qualitatively different from that observed in ischemic heart disease, where it is usually compact and confluent, instead being patchy and interspersed within surviving, hypertrophic cardiomyocytes. It is noteworthy that previous pathology studies addressed the so-called idiopathic myocardial fibrosis in SCD victims.⁴⁶ By definition, this entity is not associated with other structural heart diseases and remains without explanation. However, the LV fibrosis described in our patients with MVP differs in terms of type (ie, scarring) and location (ie, LV PMs and basal posterolateral wall).

Furthermore, we demonstrate here that CE-CMR can detect LV LGE in patients with MVP with complex ventricular arrhythmias, closely overlapping the histopathological features observed in SCD victims. At the level of PMs, 2 LGE sites have been found: the midapical portion and the base/adjacent LV wall. Herein PM LGE has been reported by Han et al²⁴ in patients with MVP with a history of arrhythmias, most of these patients had moderate to severe mitral regurgitation. Herein we confirm these data in purely arrhythmic patients with MVP without hemodynamic impairment, and we first provide convincing evidence of LGE in the inferobasal LV wall. The arrhythmogenic role of the LV myocardial scarring is supported by the morphology of arrhythmias and by electrophysiological studies in MVP indicating that the most common site of VPB origin is the inferobasal LV wall.^36,37

Most CE-CMR studies of arrhythmic risk stratification come from either ischemic heart disease or cardiomyopathies, with the notion that a larger LGE burden indicates a worse prognosis. Our quantitative data suggest that the volume of LV scarred tissue in MVP is relatively small but still associated with SCD. We should recognize that MVP differs from other nonvalvular diseases in terms of LGE distribution (stretched areas) and amount. Furthermore, the mechanical stretch by the prolapsing leaflet and elongated chordae could act as a trigger of electric instability. Further studies with a larger number of patients with MVP are needed to confirm these preliminary data of LV LGE.

Since the early descriptions, abnormal LV contraction pattern and ECG abnormalities suggested that MVP has a significant myocardial involvement.^47–51 The hypothesis that the so-called MVP syndrome is a cardiomyopathy in which regional hypercontractility acts as the primum movens of mitral valve geometry disruption with abnormal tension on the chordae and leaflets and a secondary increase in myxomatous tissue and leaflet thickening has been even advanced.^50,51 Our pathology and CE-CMR data support the theory that LV abnormalities are instead the consequence of MVP owing to a systolic mechanical stretch of the myocardium closely linked to the valve, that is, PMs and inferobasal wall, by the prolapsing leaflets and elongated chordae, accounting for a localized hypercontractility, with myocyte hypertrophy and injury eventually leading to fibrous tissue repair. The increased cardiomyocyte diameter, together with replacement-type fibrosis in the same areas, are in keeping with this theory.

Considering that performing CE-CMR in all patients with MVP would be an expensive proposition, some clinical markers that could target a high-risk subgroup destined for screening by CE-CMR are needed. ECG depolarization abnormalities on inferolateral leads, complex ventricular arrhythmias (≥3-VPB run) with RBBB morphology on 12-lead ECG Holter monitoring, and a history of presyncope, syncope, and aborted SCD seem to represent indications for CE-CMR.

Finally, we recognize that our data support an association between anatomic substrate and risk in an entity that is underappreciated as a cause of SCD. The SCD event in MVP has a low enough incidence that any marker of increased risk might be of significant value to the clinician.

β-Blockers are commonly used to treat arrhythmias in patients with MVP. The fact that 21% of young adult SCD victims and 2 living patients had aborted SCD despite β-blocker therapy is disappointing but not surprising.¹⁰ Prospective multicenter studies are warranted to support the role of CE-CMR and electrovoltage mapping for risk stratification and to assess the efficacy of antiarrhythmic therapy and targeted catheter ablation in selected cases.

Study Limitations

While acknowledging the small number of patients with MVP without complex ventricular arrhythmias, we should recognize that it is difficult to collect patients with pure MVP without either valve incompetence or ventricular arrhythmias both clinically and at postmortem. Prospective multicenter studies enrolling a higher number of MVP with and without complex ventricular arrhythmias are warranted to evaluate the exact prevalence of LGE in the overall MVP population.

Genetic data were not available in our SCD population. It is noteworthy that in our series of SCD victims, there was macroscopic evidence of myxomatous mitral valve, and nearly half (47%) had a previous in vivo diagnosis of MVP with a cardiologic checkup ruling out channelopathies. Moreover, the ECG, which was available for revision in 28%, did not show any evidence of long- or short-QT or Brugada syndrome. Of the remaining 31 patients with MVP, 25 (80%) had first-degree family members referred for cardiologic screening without any evidence of channelopathies but MVP in 4 patients (16%).

Although we are strong supporters of the relevance of molecular autopsy in the study of SCD,^21,22 we follow the indication by the Heart Rhythm Society/European Heart Rhythm Association/Asia Pacific Heart Rhythm Society expert consensus statement.⁵² According to these guidelines, arrhythmia syndrome–focused postmortem genetic testing can be useful for all sudden unexplained death syndrome victims as Class IIa indication; furthermore, evaluation of first-degree blood relatives with resting ECG with high right ventricular leads, exercise stress testing, and echocardiography is a Class I recommendation.

Conclusions

This study suggests that MVP is a significant cause of SCD in young adults and is the leading structural cause in women. Arrhythmic patients with MVP are mostly females with ventricular arrhythmias of LV origin and frequent ECG repolarization abnormalities on inferior leads. The hallmark of arrhythmic MVP is fibrosis of PMs and inferobasal LV free wall, which correlates well with arrhythmia morphology, pointing to a myocardial stretch by the prolapsing leaflets and elongated chordae. CE-CMR allows the identification of this arrhythmic substrate and is a promising noninvasive tool for risk stratification and SCD prevention.

Acknowledgments

We would like to acknowledge the skilful technical assistance of Anna Saracino and Daniele Iannazzone for histology and Marco Pizzigolotto for illustrations.

Footnotes