Background: Abnormal papillary muscles (PM) are often found in hypertrophic cardiomyopathy (HCM).
Objective: To assess the relationship between morphological alterations of PM in patients with HCM and left ventricular outflow tract (LVOT) obstruction, using magnetic resonance imaging (MRI) and echocardiography.
Methods: Fifty-six patients with HCM (mean age 42 years (interquartile range 27, 51), 70% male) and 30 controls (mean age (42 (30, 53) years, 80% male) underwent MRI on a 1.5 T scanner (Siemens, Erlangen, Germany). Standard cine images were obtained in short-axis (base to apex), along with two-, three- and four-chamber views. The presence of bifid PM (none, one or both) and anteroapical displacement of anterolateral PM was recorded by MRI and correlated with resting LVOT gradients obtained by echocardiography.
Results: Double bifid PM (70% vs 17%) and anteroapical displacement of anterolateral PM (77% vs 17%) were more prevalent in patients with HCM than in controls (p<0.001). Subjects with anteroapically displaced PM and double bifid PM had higher resting LVOT gradients than controls (45 (6, 81) vs 12 (0, 12) mm Hg (p<0.01) and 42 (6, 64) vs 11 (0, 17) mm Hg (p = 0.02), respectively. In patients with HCM, the odds ratio of having significant (⩾30 mm Hg) peak resting gradient was 7.1 (95% CI 1.4 to 36.7) for anteroapically displaced anterolateral PM and 10.4 (95% CI 1.2 to 91.2) for double bifid PM (both p = 0.005), independent of septal thickness, use of β-blockers and/or calcium blockers and resting heart rate.
Conclusions: Patients with HCM with abnormal PM have a higher degree of resting LVOT gradient, which is independent of septal thickness.
Statistics from Altmetric.com
Hypertrophic cardiomyopathy (HCM) is a genetic disorder commonly associated with left ventricular outflow (LVOT) obstruction and systolic anterior motion (SAM) of the anterior mitral valve leaflet.1–3 The presence of a resting LVOT gradient is responsible for many of the symptoms associated with HCM and carries prognostic significance.4 Prior studies have demonstrated an association between SAM, LVOT obstruction, and increased interventricular septal thickness.5 It has been proposed that LVOT narrowing and resultant Venturi effect cause SAM.6–9 Therefore surgical and percutaneous treatments focus on removing or ablating part of the septal myocardium. However, LVOT obstruction and SAM can occur in patients with no septal hypertrophy10–12: in the presence of mitral apparatus abnormalities, including anterior displacement of the papillary muscles (PMs),1 13–15 leaflet elongation5 and anteriorly displaced coaptation of the mitral valve leaflets.16–18 Therefore, subgroups of patients with HCM may have different mechanisms causing LVOT obstruction.
We hypothesised that abnormal PM morphology, often seen in patients with HCM (fig 1), might be associated with increased LVOT obstruction, and sought to assess if this was independent of septal thickness. While Doppler echocardiography provides accurate assessment of the magnitude of LVOT obstruction, cardiac magnetic resonance imaging (MRI) has the ability to acquire high-resolution images in multiple user-defined axes.19–21 The combination of these high-resolution anatomical images with accurate assessment of the physiological aspects of LVOT obstruction using Doppler echocardiography may provide further insights into the pathophysiological mechanisms of HCM. Thus, using echocardiography and MRI, we sought to determine systematically the relationship between PM alterations and the magnitude of LVOT obstruction in patients with HCM.
This observational study was conducted at Cleveland Clinic after appropriate institutional review board approval. We studied 56 consecutive patients with an established diagnosis of HCM who underwent a comprehensive cardiac MR examination between 2002 and 2006. The diagnosis of HCM and dynamic LVOT obstruction was based on standard clinical criteria, using history, physical examination, electrocardiography and echocardiography.22 23 We also studied 30 consecutive subjects, without a diagnosis of HCM, who were undergoing a comprehensive cardiac MR examination for other indications (suspected arrhythmogenic right ventricular dysplasia, Marfan’s syndrome, aortic aneurysms and aortic vasculitis); a similar cine imaging sequence was used in both groups to evaluate left ventricular (LV) morphology and function. All these 30 subjects had a completely normal cardiac MR examination, without any demonstrable cardiac abnormalities, and served as controls. Demographic data, clinical history (including that of septal myectomy) were obtained through chart review. Use of drugs (including β-blockers and calcium channel blockers) was recorded. Hypertension was defined based upon history, raised blood pressure (>120/80 mm Hg) or use of antihypertensive drugs.
The cardiac MR examinations were performed on 1.5 T MR scanners (Siemens Medical Solutions, Erlangen, Germany)—Sonata (for examinations between 2002 and 2005, 40 mT/m maximum gradient strength, 200 mT/m/s maximum slew rate) or Avanto (for 2006 examinations, 45 mT/m maximum gradient strength, 200 T/m/s maximum slew rate). Scout images were acquired to identify the cardiac axes. For assessment of PM morphology, myocardial thickness and global cardiac function, balanced steady-state free precession images were acquired: TE = 1.6 ms, TR = 3.3 ms, flip angle = 70° and slice thickness = 6 mm (long-axis images) or 8–10 mm (contiguous short-axis images from apex to base). For short-axis images, the field of view varied from 228 to 330 in the x-direction and from 260 to 330 in the y-direction and matrix size varied from 140 to 180 in the x-direction and 256 in the y-direction, giving a spatial resolution of 1.5–2.1 mm (x-direction) by 1.1–1.4 mm (y-direction). In long-axis images, the field of view varied from 250 to 320 (x-direction) and 280 to 340 (y-direction). Matrix size varied from 120 to 210 in the x-direction and was 256 in the y-direction (phase encoding direction), giving a spatial resolution of 1.5–2.1 mm (x-direction) by 1.1–1.6 mm (y-direction). For patients able to suspend respiration, breath-hold duration was 10–15 seconds, depending on the heart rate; otherwise, images were acquired using three signal averages.
In all patients, cine images were obtained at the following anatomical locations: three short-axis slices (base, mid-ventricle and apex), along with standard two-, three- and four-chamber long-axis views. Maximal septal thickness and maximal PM diameter were measured manually and recorded at end diastole on short-axis images. The presence of bifid PM (involving none, one or double PM) was recorded if the PM had more than one head seen on multiple cine images (fig 2A). The minimum distance between the septum and anterolateral PM was manually determined in a four-chamber view at end diastole. Anteroapical insertion was qualitatively defined when the origin (base) of anterolateral PM was displaced anteriorly (in relation to the interventricular septum) and distally in two- or four-chamber views and visible on the distal-most apical short-axis image (fig 2B). The presence or absence of significant LVOT obstruction was not known at the time of MRI assessment.
Echocardiographic studies were performed using commercially available instruments equipped with Harmonic Imaging (Siemens Sequoia, Mountview CA, USA; Phillips 5500/7500, Andover, MA, USA; and GE Vivid 7, Milwaukee, WI, USA), as part of standard clinical practice. HCM was defined by two-dimensional echocardiography as a hypertrophied and non-dilated left ventricle in the absence of another cardiac or systemic disease that could produce a similar magnitude of hypertrophy.1–3 The presence of SAM was determined by M-mode or two-dimensional echocardiography, or both. Left ventricular outflow peak velocity was measured by continuous-wave Doppler echocardiography, and LVOT pressure gradient was estimated using the simplified Bernoulli equation.24 LVOT obstruction was considered to be significant when peak instantaneous outflow gradient was estimated to be at least 30 mm Hg with the use of continuous-wave Doppler echocardiography under basal (resting) conditions.4 Care was taken to avoid contamination of the LVOT waveform by the mitral regurgitation jet.4 The degree of resting mitral regurgitation was recorded based upon chart review on a scale of 0–4+ (0, none; 1+, mild; 2+, moderate; 3+, moderately severe; 4+, severe). The average time between echocardiography and MRI was approximately 4 days.
Baseline demographics, risk factors and clinical variables are descriptively summarised for each group. Continuous variables are expressed as mean (interquartile range, 25%, 75%). Categorical data are presented as percentage frequency. A χ2 or Fisher exact test was used to compare proportions. Continuous variables are compared between groups using t-test (paired or unpaired) or analysis of variance. Linear regression was used to assess correlations between continuous variables. In 10 randomly selected patients, inter- and intraobserver agreement for PM morphology (double bifid and anteroapical displacement of anterolateral PM) was calculated using κ statistic. To assess the independent impact of LV septal thickness and PM abnormalities, we first scored the anteroapical displacement of the anterolateral PMs as 0 if absent and 1 if present, and of double bifid PMs as 0 if either one was normal and 1 if both were bifid. Finally, the PM abnormality score (0, 1 or 2) was obtained by summing these two scores. Subsequently, we applied generalised linear model analysis of covariance with septal thickness, use of β-blocker and/or calcium blocker, resting heart rate and PM score as covariates and resting LVOT gradient as a dependent variable. To calculate the odds ratio of having a significant peak resting gradient, we first dichotomised peak resting gradient using 30 mmHg as a cut-off point. Then we performed standard univariate and multivariate (using above mentioned covariates) forward stepwise logistic regression analysis (SPSS 10.0; SPSS, Chicago, IL, USA). A p value <0.05 was considered significant.
Table 1 shows the demographic data and the MRI findings for patients with HCM and controls.
As expected, the mean end-diastolic septal and PM thickness, measured by MRI was significantly higher in patients with HCM than in controls. The prevalence of anteroapical displacement of the anterolateral PM and double bifid PMs was also significantly higher in the HCM group than in controls. Figures 3A and B show the association between maximal septal and anterolateral PM thickness (both measured by MRI at end diastole) and maximum septal thickness and minimum distance between anterolateral PM and septum, both measured by MRI.
Reproducibility of papillary muscle morphology assessment by MRI
The intra- and interobserver reproducibility for anteroapical displacement of anterolateral PM was high (κ statistic 0.80 and 0.75, respectively). The intra and interobserver reproducibility for double bifid PM morphology was very high (1.0 and 1.0, respectively).
Analysis of MRI and echocardiographic findings within the HCM group
Within the HCM group, 20 (36%) had significant resting LVOT gradients, 12 (21%) had documented history of syncope, 2 (4%) had history of sudden cardiac death and 43 (77%) subjects showed the presence of SAM on echocardiography: 28 (50%) had resting SAM, 3 (5%) had only chordal SAM and 12 (21%) had provocable SAM after exercise. Mean heart rate and resting LVOT gradient were 67 bpm (60, 72) and 35 mm Hg (0, 55), respectively. Subjects with HCM with SAM had a significantly higher resting LVOT gradient than those without SAM (41 (7, 57) vs 5 (0, 10) mm Hg, p<0.001). Both anteroapical displacement of anterolateral PM and double bifid PMs were more prevalent in subjects with an elevated resting LVOT gradient (⩾30 mm Hg, n = 21) than in those without (n = 35) (90% vs 57% and 95% vs 66%, respectively, both p<0.02). Patients with HCM with anteroapical displacement of anterolateral PMs had significantly higher peak resting LVOT gradient than those without (45 (6, 81) vs 12 (0, 12) mm Hg, p<0.01). Similarly, subjects with double bifid PMs had significantly higher peak resting LVOT gradient than those without (42 (6, 64) vs 11 (0, 17) mm Hg, p = 0.02). Figures 4A and B show the correlation between septal and anterolateral PM thickness (measured by MRI) and resting LVOT gradient (measured by Doppler echocardiography). Within the HCM group, the degree of resting mitral regurgitation was similar between patients with and without anteroapical displacement of anterolateral PM (0.79 (0, 1.5) vs 0.65 (0, 0.5)) and double bifid PMs (0.81 (0, 1.5) vs 0.53 (0, 0.5), both p = NS)
Subsequently, we divided the HCM group into four subgroups based upon maximal septal thickness (<2 or ⩾2 cm) and presence or absence of anteroapical displacement of anterolateral PMs: (a) group A: septal thickness <2 cm and non-displaced PMs (n = 9); (b) group B: septal thickness ⩾2 cm and non-displaced PMs (n = 7); (c) group C: septal thickness <2 cm and displaced PMs (n = 11) and (d) group D: septal thickness ⩾2 cm and displaced PMs (n = 29). Figure 5 shows that there was a significant difference in the mean resting LVOT gradient (mm Hg) in the subgroups.
We further tested whether the association between abnormal PMs and resting LVOT gradient was independent of maximal septal thickness. The presence of anteroapical displacement of anterolateral PM was significantly associated with resting LVOT gradient (p = 0.025), with septal thickness not affecting this relationship (p = 0.316). The presence of double bifid PM was less strongly, but still significantly associated with resting LVOT gradient (p = 0.046), with septal thickness not affecting this relationship (p = 0.17). Also, the PM score was significantly associated with resting LVOT gradient (p = 0.007), with higher score having a higher resting LVOT gradient, and with septal thickness not affecting this relationship (p = 0.49). The odds ratio of having a significant (⩾30 mm Hg) peak resting gradient was 7.1 (95% CI 1.4 to 36.7) for anteroapically displaced anterolateral PM and 10.4 (95% CI 1.2 to 91.2) for double bifid PM (both p = 0.005). The PM score correctly predicted a significant resting gradient in 71.4% of cases (odds ratio (OR) = 14.25 (95% CI 2.9 to 71.2); p<0.001). In a multivariate forward stepwise logistic regression model that evaluated anteroapical displacement of anterolateral PMs, septal thickness, use of β-blockers and/or calcium blockers and resting heart rate as possible predictors, only anteroapical displacement (OR = 7.1 (95% CI 1.4 to 36.7)) was a significant predictor. Similarly, double bifid PM, when included in the model instead of anteroapical displacement, was a significant predictor (OR = 10.4 (95% CI 1.2 to 91.2)). Finally, in a model that evaluated PM score, septal thickness, and use of either β-blockers or calcium blockers as possible predictors, only PM score was a significant predictor (OR = 14.25 (95% CI 2.9 to 71.2)).
The presence of resting LVOT gradient is a common and important pathophysiological manifestation responsible for many of the symptoms associated with HCM. It has been shown that an elevated resting LVOT gradient carries prognostic significance in patients with HCM.25 Patients with HCM with a resting gradient of ⩾30 mm Hg had a fourfold increased risk of death or progression to severe congestive symptoms, compared with those without obstruction.4 Furthermore, relief of LVOT gradient has been shown to improve survival.4 26 In patients with HCM, SAM and resting LVOT gradient have been attributed to increased turbulence caused by LVOT narrowing, which occurs owing to a combination of septal hypertrophy5 and the Venturi effect (increased flow velocity and decreased pressure above the aortic valve).6–9 These mechanisms, however, do not explain the following observations: patients with SAM have more leaflet slack than those without,5 onset of SAM at or before aortic valve opening occurs when outflow velocity is low27 and SAM occurs in patients with no significant septal hypertrophy.10–12 Case reports have described abnormalities affecting the mitral valve apparatus in patients with obstructive HCM, including that involving PM displacement.13–15
The patients with HCM in the current study group demonstrate a high frequency of PM abnormalities (anteroapical displacement or double bifid PMs, or both; figs 1 and 2). The PM and septal thickness were higher in patients with HCM, resulting in a significantly reduced distance between the anterolateral PM and the septum, and effectively reducing the LV cavity volume. Within the HCM group, abnormal PM morphology (anteroapical displacement of the anterolateral PM and double bifid PM) was associated with an increased prevalence of SAM and higher resting LVOT gradients; this association was independent of septal thickness. Also, the odds of having an abnormal resting LVOT gradient were significantly higher in those patients with HCM with anteroapically displaced anterolateral PM or double bifid PMs, independent of septal thickness. Furthermore, of the patients with HCM who underwent septal myectomy (without repositioning of PMs), 23% had a residual provocable LVOT gradient.
Along with small studies and case reports describing PM abnormalities in patients with HCM, there have been attempts to reproduce the impact of altered PM morphology in an animal model. One study reproduced SAM in dogs with structurally normal hearts and mitral valves by anteriorly displacing the PMs through a suture mechanism.28 SAM in these dogs was a result of the decrease in effective posterior restraint (increased leaflet slack) caused by anterior redirection of PM tension; increased length of the residual leaflet, and coaptation of the mitral valve leaflets in LVOT, subjecting the leaflets to drag forces29 which propel the leaflet anteriorly. Sherrid et al have also demonstrated that the anteriorly displaced mitral valve leaflet coaptation point is important in the evolution of SAM.27 They argue that anteriorly displaced leaflets are pushed into the LVOT by drag forces, as opposed to the Venturi effect, as the onset of SAM occurs when the velocity in the LVOT is normal or low. A study by Klues et al described persistent symptoms and LVOT gradient (60–70 mm Hg) in a few patients who underwent septal myectomy owing to mid-cavitary apposition of PM to the ventricular septum.15 The current study, while not directly demonstrating an increased leaflet length/slack or drag forces, certainly corroborates the findings of prior experimental observations. Finally, to further understand the complex geometry of PMs (which might not be fully appreciated by echocardiography), a recent study has demonstrated the potential clinical utility of three-dimensional tomographic imaging to better delineate PM morphology and insertion.30
Limitations and future outlook
Because this is an observational study conducted at a tertiary referral centre, there is significant selection bias. The patients in our study had significant symptoms and were therefore referred to our centre for a cardiac evaluation. A considerable proportion of patients with HCM are asymptomatic and do not have significant resting LVOT gradients. These asymptomatic patients with HCM may or may not have PM abnormalities. Also, the resting LVOT gradient is dynamic and is affected by multitude of factors, including heart rate, arrhythmia, body position and volume status. Thus, a snapshot assessment of the gradient at the time of echocardiography (performed in lying position) might not be a true reflection of the dynamic LVOT gradient. Assessment of PM morphology/dimensions by MRI is not standardised, and might potentially be subjective based upon the views obtained. However, the assessment in this study was consistent for all the subjects (including patients with HCM and controls).
Another limitation is that of selection of controls. While the controls did not have HCM (and were undergoing cardiac MRI for other cardiovascular indications), it is theoretically possible that they had morphological changes within the heart that might potentially have biased the comparison. However, in our institution, it would be impossible to obtain MRI data in “normal healthy volunteers” in an observational manner, without a prospective study design and appropriate institutional approval.
The observations from this study raise the following questions which would have to be answered using a prospective longitudinal study design: (a) Does abnormal PM morphology progress over time and correlate with progression of clinical symptoms? (b) Does it result from a particular genetic mutation? (c) Is it associated with adverse outcomes? Furthermore, could surgical correction of abnormal PMs provide incremental value to septal myectomy/mitral valve repair alone? Also, a prospective study to assess whether different morphological subtypes of patients with HCM—(a) normal/mildly increased septal thickness with abnormal PMs; (b) significant septal thickness with normal PM morphology and (c) significant septal thickness with abnormal PM morphology—have a different impact on dynamic LVOT gradient is warranted.
Patients with HCM with altered PM morphology—that is, anteroapical displacement of anterolateral PM or double bifid PMs, have a significantly higher prevalence of SAM or an elevated resting LVOT gradient, compared with those without. The association between altered PMs and resting LVOT gradient appears to be independent of septal thickness, resting heart rate or use of drugs (β-blockers or calcium blockers, or both). A significant proportion of patients with HCM with abnormal PM morphology continued to have provocable LVOT gradient despite septal myectomy. Further prospective longitudinal studies are required to determine if abnormal PM morphology is progressive and associated with adverse outcomes. Also, it needs to be determined if surgical correction of abnormal PM morphology provides incremental value to septal myectomy/mitral valve repair alone.
We thank Ms Joan Weaver RT (MRI) and Ms Angel Lawrence RT (MRI) for their help in acquisition of MRI images.
Funding: Modest research support from Siemens Medical Solutions.
Competing interests: None.
Ethics approval: Approved by the Cleveland Clinic institutional review board.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.