|Home | About | Journals | Submit | Contact Us | Français|
To describe the echocardiographic characteristics and investigate the clinical correlates and prognostic significance of left ventricular false tendons (LVFTs).
Although LVFTs are generally considered as anatomic variants, they have been associated with innocent precordial murmurs and electrocardiographic abnormalities in small case series. The correlates of LVFTs in the community are unknown.
We compared 101 Framingham Study participants with LVFTs (mean age 56 years, 45% women) on routine two-dimensional echocardiograms with 151 referents without LVFTs (mean age 57 years, 44% women). We examined the cross-sectional clinical, electrocardiographic (rest and ambulatory), and echocardiographic correlates of LVFTs using logistic regression models, and evaluated the prospective association between LVFTs and all-cause mortality using Cox proportional hazards regression models.
A total of 107 LVFTs (94 simple with two points of attachment, and 13 complex/branching type with three or more points of attachment) were identified in 101 participants. LVFTs were most commonly visualized in the apical four chamber view (81%) and were predominantly localized to apical-third of the LV cavity (78%). LVFTs were associated with the presence of innocent precordial murmurs (multivariable adjusted odds ratio [OR] 5.55, 95% confidence interval [CI] 1.40-21.94), and electrocardiographic LV hypertrophy (OR 4.43, 95% CI 1.08-18.25). Body mass index (BMI) was inversely related to the presence of LVFTs (per kg/m2 increment, OR 0.94, 95% CI 0.88-0.99). LVFTs were not associated with QRS axis deviation, ventricular premature beats or repolarization abnormalities (all p-values >0.20). During a mean (±SD) follow-up of 7.7 (±1.6) years, 15 participants with and 19 without LVFTs died. In multivariable analyses, presence of LVFTs was not associated with the risk of death (p=0.92).
In our community-based sample of middle-aged to elderly white women and men, LVFTs were more likely to be identified in individuals with lower BMI, were cross-sectionally associated with the presence of innocent precordial murmurs and electrocardiographic LV hypertrophy, but were not associated with the risk of mortality.
Left ventricular false tendons (LVFTs) are discrete, fibromuscular structures of varying length and thickness that traverse left ventricular (LV) cavity.1 Although these structures are readily identified by routine two-dimensional echocardiography,2 their morphologic appearance and location in the LV cavity have been reported primarily from autopsy-based studies 2-9 or from echocardiographic studies conducted on patients at referral centers.1,6,10-21 One study has elaborated echocardiographic characteristics of LVFTs in a group of healthy company workers,22 but data on the epidemiology of LVFTs in the community are unavailable.
Although LVFTs have generally been considered as benign anatomic variants,5,6,15,23-26 numerous disease associations have been reported in literature. These include an increased prevalence of precordial murmurs,11,13,24,27-32 repolarization abnormalities on the resting electrocardiogram,33,34 pre-excitation,30 ventricular arrhythmias,11,14,20,22,35-37 mitral regurgitation,38 and a dilated left ventricle.29,39 Most of these clinical associations have been described, however, only in small clinical case series. The clinical correlates of LVFTs in the community are unknown. Furthermore, despite the apparent association of LVFTs with ventricular arrhythmias and repolarization abnormalities, to our knowledge, there are no published reports evaluating the prognosis of participants with LVFTs in the community.
Therefore, the objectives of our investigation were to describe the echocardiographic characteristics of LVFTs, investigate their clinical correlates, and examine their prognostic significance in a community-based sample of middle-aged to elderly white women and men.
The Framingham Heart Study is a prospective longitudinal epidemiologic study of 5209 women and men that began in 1948. The children of the Original cohort and spouses of these children were enrolled into the Offspring study in 1971. The design and enrollment criteria of these studies have been detailed elsewhere.40,41 Participants of the Original cohort are examined every 2 years, whereas those in the offspring study are evaluated approximately every 4 years. At each examination, participants have undergone medical history, physical examination, 12-lead electrocardiogram, and blood tests for the assessment of cardiovascular disease risk factors. All participants are under continuous surveillance for the development of cardiovascular disease and/or death.
Participants of the Original cohort who attended the 20th examination cycle (1988 to 1989, N=1206) and individuals of the Offspring cohort who attended the 4th examination cycle (1987 to 1990, N=3931) were eligible for the present investigation. Routine two-dimensional echocardiography with Doppler color flow imaging was performed at these examinations. The presence or absence of LVFTs on echocardiograms was routinely noted by the sonographers on a data coding sheet. In light of the possibility that LVFTs are better identified when they are specifically searched for,14,22 and for the purpose of enriching our study sample with LVFTs, we selected the following echocardiograms for review by an experienced echocardiographer: all those studies (n=88) coded as demonstrating LVFTs by the sonographers and twice this number (n=176), matched by age and sex and drawn randomly from the remaining echocardiograms that were not coded as having LVFTs. All the chosen echocardiograms were reviewed in a blinded fashion and in a randomized sequence without knowledge of the initial coding (Figure 1).
The diagnosis of a LVFT was based on the finding of a distinctive linear echogenic strand, traversing the LV cavity, connecting the LV free wall or papillary muscle and the ventricular septum, not related to the mitral valve apparatus, and identified in at least two echocardiographic planes of view (Figure 2). All equivocal interpretations, such as when a LVFT was noted in only one echocardiographic view, were reviewed by a second experienced echocardiographer, and a LVFT was considered present only if both echocardiographers agreed. Special care was taken to differentiate LVFTs from other entities such as thickened ventricular trabeculations, intraventricular masses/tumors, discrete subaortic membrane, accessory anterior mitral leaflet, and flail mitral chordae tendineae.
Based on the location of the LVFTs in the apical-, mid-, or basal-third segments along the long axis of the left ventricle,42 we classified each LVFT into transverse (localized to one zone), diagonal (extending across two adjacent zones) or longitudinal (extending across all three zones) types. In addition, based on the morphologic appearance, we classified LVFTs that appeared as a single chord traversing the LV with 2 points of insertion as simple, and LVFTs that appeared as branching with 3 or more points of insertion as complex/branching. We calculated end-diastolic thickness of LVFT by averaging 3 measurements at the site of maximum thickness of the tendon and classified LVFTs as thick (≥2 mm) and thin (<2 mm).22 In light of the notion that a systolic musical murmur may result from the turbulence produced by LVFT when pulled taut across the left ventricle,5,17,24 we ascertained the tautness and laxity of the visualized LVFTs at end-systole and at end-diastole.
We examined the cross-sectional association of LVFT with the following: (a) clinical correlates: age, sex, height, body mass index, hypertension, innocent precordial murmur, diabetes mellitus, myocardial infarction, and heart failure; (b) electrocardiographic correlates: QRS axis deviation, ventricular premature beats (VPBs), LV hypertrophy and repolarization abnormalities; and (c) echocardiographic correlates: LV internal diameter (LVID), wall thickness (LVWT), mass (LVM), and ejection fraction (LVEF). Information on all these covariates was coded at the baseline examinations blinded to the current investigators.
Additionally, in a subgroup of participants (N=140) who underwent ambulatory Holter monitoring one examination prior to the index examination (Original Cohort examination-19 and Offspring Study examination-3), we examined the association of LVFT with the presence of VPBs. As LVFTs are primarily congenital structures, we reasoned that use of Holter records from a prior examination would be unlikely to introduce a major bias in assessing cross-sectional associations.
Criteria and methods of measurement of all covariates have been described previously.43 Briefly, an ‘innocent’ precordial murmur was considered as present if a systolic or diastolic murmur was auscultated by an examining Framingham Heart Study clinic physician, in the absence of any significant valve disease (greater than mild stenosis or regurgitation of any valve) on Doppler echocardiography. QRS axis deviation included right (+ 90° to ±180°), left (− 30° to −90°), or extreme (− 90° to ±180°) axis deviation. VPBs were defined as the presence of one or more premature and abnormally prolonged (>0.12 seconds) QRS complexes of abnormal shape without a preceding P wave on a 10-second electrocardiographic rhythm strip, or on ambulatory Holter recordings.44 Diagnosis of electrocardiographic left ventricular hypertrophy (ECG-LVH) was based on one of the following voltage criteria: R wave >1.1 mV in aVL; R wave >2.5 mV in V5 or V6; S wave >2.5 mV in V1 or V2; sum of S in V1 or V2 plus R in V5 or V6 >3.5 mV; or sum of R in I and S in III >2.5 mV with or without ST-T changes indicative of ‘strain’ pattern.45 Repolarization abnormalities were considered present if non-specific ST segment and T wave changes were noted on the resting 12-lead electrocardiogram in the absence of myocardial infarction or LV hypertrophy, and without concurrent use of digitalis and/or diuretics. LV internal dimensions (LVID) and the thicknesses of the interventricular septum (IVST) and the LV posterior wall (PWT) at end diastole (d) were obtained by averaging M-mode measurements in at least 3 cardiac cycles using a leading edge technique as recommended by the American Society of Echocardiography.46 Left ventricular wall thickness (LVWT) was calculated as the sum of end-diastolic IVSTd and PWTd. LVM was calculated as 0.8 [1.04 (IVSTd + LVIDd + PWTd)3 – (LVIDd)3] + 0.6.47 Estimation of LV ejection fraction (LVEF) was based on visual assessment and an LVEF of <50% was considered to indicate LV systolic dysfunction.
LVFTs have been associated with clinical,11,13,24,27-32 electrocardiographic11,14,20,22,30,33-37 and echocardiographic29,38,39 abnormalities. Therefore, we examined the association of LVFTs with all-cause mortality on follow-up.
We compared the age- and sex-adjusted clinical, electrocardiographic, ambulatory Holter, and echocardiographic characteristics between participants with LVFTs and those without LVFTs using general linear models (least squares means) for continuous variables and logistic regression models48 for categorical variables. We performed additional adjustment for body mass index to examine the association of LVFTs with presence of innocent precordial murmurs, and ECG-LVH. The association of LVFTs with echocardiographic variables such as LVM, LVWT, and LVID were performed after adjustment for age, sex, and height. To identify important multivariable predictors of LVFTs, we constructed stepwise logistic regression models including covariates that were found to have a p-value of <0.10 in age- and sex-adjusted analyses. The variables age and sex were forced into these models. Lastly, we used Cox proportional-hazards regression models49 adjusting for age, sex, body mass index, hypertension, diabetes mellitus, and ECG-LVH to evaluate the association between LVFTs and all-cause mortality during follow-up. We performed all statistical analyses using version 6.2 of SAS software,50 and considered a two sided p-value of <0.05 as statistically significant.
Of the 88 echocardiograms initially suspected of having LVFTs by the sonographers, LVFTs were confirmed to be present in 69 (78%) but absent in 18 (21%); one study was deemed to be inadequate for the evaluation of LVFTs. Of the 176 echocardiograms of age- and sex-matched participants not coded as having LVFTs by the sonographers, LVFTs were confirmed to be absent in 133 (76%) but were identified in 32 (18%); 11 studies had inadequate echocardiographic views to ascertain LVFT status. Thus, our final study sample consisted of 101 participants with LVFTs (cases, Figure 1) and 151 participants without LVFTs (referents).
A total of 107 LVFTs were identified in 101 participants (95 participants with one LVFT, and six participants with two LVFTs). LVFTs were most commonly visualized in the apical four chamber view (81%) followed by apical two chamber view (62%) and the apical long axis view (61%). LVFTs were predominantly of the transverse type, and were located most commonly in the apical zone (Figure 3). Complex/branching appearance of the LVFTs was evident in 12%. The average (±SD) thickness of LVFTs was 1.6 ±0.5 millimeters (range, 0.5 to 3.3 millimeters). Of the 107 LVFTs in our study sample, 85 (79%) were lax at end systole but became taut at end diastole; 15 (14%) were taut throughout the cardiac cycle and 1 (1%) was lax at end systole as well as at end diastole. For 6 (6%) LVFTs the tautness or laxity could not be ascertained.
After adjustment for age and sex, body mass index was inversely related to the presence of LVFTs (mean±SD, 25±5 with LVFTs vs. 27 ±5 without LVFTs, p=0.01). LVFTs were significantly associated with the presence of an innocent precordial murmurs (9% with LVFTs vs. 2% without LVFTs, p=0.02), which remained unchanged after additional adjustment for body mass index (p=0.02). The proportion of participants with hypertension, diabetes mellitus, history of prior myocardial infarction, and heart failure were similar in participants with and without LVFTs (all age- and sex-adjusted p-values >0.20).
In age- and sex- adjusted models, LVFTs were associated with the presence of ECG-LVH (18% with LVFTs vs. 9% without LVFTs, p=0.01), which was attenuated upon additional adjustment for body mass index (p=0.08). There was no significant association, however, between the presence of LVFTs and the prevalence of QRS axis deviation, VPBs, or repolarization abnormalities on the 12-lead electrocardiogram (all age- and sex-adjusted p-values >0.20).
Fifty seven participants with LVFTs and 83 participants without LVFTs underwent evaluation by ambulatory Holter. The mean (±SD) duration of Holter monitoring records available were 7.6 (±9.2) hours for those with LVFTs and 9.0 (±10.1) hours for those without LVFTs. In statistical models adjusting for age and sex, there was no significant association between the presence of LVFTs and VPBs (47% with LVFTs vs. 47% without LVFTs, p=0.99). The frequency (beats per hour) and the form (uniform, multiform, couplets, or ventricular tachycardia) of VPBs were also similar in participants with LVFTs compared with participants without LVFTs (data not shown).
LVFTs were more likely to be found in participants with LV systolic dysfunction (LVEF <50% in 10% with LVFTs vs. 4% without LVFTs, borderline statistical significance, p=0.049, in models adjusting for age and sex). After adjustment for age, sex and height, there was no significant difference between LVID, LVWT, or LVM between cases and referents (all p-values >0.10).
In multivariable analyses, LVFTs were inversely associated with body mass index, and positively associated with precordial murmur and ECG-LVH. LV systolic dysfunction was not associated with the presence of LVFTs (p-value >0.10). In statistical models evaluating LV mass instead of ECG-LVH as a covariate, the multivariable odds ratio for the presence of LVFTs for every one gram increment in LV mass was 1.01 (95% confidence interval 1.00 to 1.01, p-value 0.08).
During a mean ± SD follow-up of 7.7 ± 1.6 years (maximum 10.8 years), 15 of 101 (15%) individuals with LVFTs and 19 of 151 (13%) participants without LVFTs died. In multivariable models, presence of LVFTs was not associated with the risk of mortality (hazards ratio 1.04, 95% confidence interval 0.48 to 2.23, p-value 0.92).
In our community-based sample of middle-aged to elderly white women and men, LVFTs were predominantly located in the apical zone, and were more common in participants with lower body mass index, perhaps due to better visualization. LVFTs were cross-sectionally associated with the presence of ‘innocent’ precordial murmurs and with ECG-LVH. Presence of LVFTs was not related to other ECG features such as QRS axis deviation, ventricular premature beats or repolarization abnormalities. There was no association of the presence of LVFTs and the risk of mortality during follow-up.
In several small case series from referral centers, it has been reported that among participants with innocent precordial murmur, the prevalence of LVFTs was considerably higher,11,17,29,31 although two case control studies did not note such an association.26,51 Our community-based investigation confirms the association of LVFT with innocent precordial murmurs in a white middle-aged to elderly study sample. It has been postulated that a transient24 or a persistent30 precordial systolic murmur may occur when a LVFT is located perpendicular to the stream of blood flow and when a LVFT is pulled taut by a dilated ventricle.30 Phonocardiographic correspondence between oscillations of LVFT and the frequency of the precordial murmur has also been demonstrated.13 Our data suggest that LVFTs may be considered in the differential diagnosis of innocent precordial murmurs in adults.
Using a broad definition of VPBs, i.e., presence of at least one VPB on 24-hour Holter monitoring, a cross-sectional association between the presence of LVFTs and occurrence of VPBs has been reported in small referral series14,20 and in a study on 179 healthy volunteers.22 Based on the demonstration of Purkinje fibers in LVFTs of both animals and humans,4,10,21,35,52,53 it has been hypothesized that VPBs might be triggered due to increased automaticity of these specialized conducting cells during the mechanical stretch of the LV wall at the point of insertion of the LVFT. In our study, we found no significant association between the presence of LVFTs and the existence of VPBs.
Our finding of a modest association between LVFTs and ECG-LVH, but without a concomitant statistically significant association between LVFTs and LV mass, is noteworthy. An individual's body habitus may simultaneously influence both electrocardiographic detection of LVH and echocardiographic visualization of LVFTs. Increased transmission of cardiac electoral forces through the chest wall in thin individuals may result in sufficiently large QRS amplitude to meet the criteria for ECG-LVH. Thin people also have better acoustic windows that may result in improved visualization of LVFTs on echocardiography, a notion consistent with the inverse association between body mass index and the presence of LVFTs noted in our study.
Our study was conducted on a community-based sample, and is inherently less biased compared with hospital-based or referral series. However, because of the extent of unidentified LVFTs during initial screening for the presence or absence of LVFTs on routinely performed echocardiograms, we are unable to provide estimates of the prevalence of LVFTs in the community. Transducer angulations were not routinely employed to improve visualization of LVFTs. Phonocardiographic confirmation of precordial murmurs was not available. Because of the modest sample size we may have lacked sufficient power to detect modest associations. The possibility of multiple comparisons yielding spurious associations is another limitation. Lastly, the Framingham Study is largely Caucasian and the generalizability of our findings to other races/ethnicities is unknown.
In our community-based sample of middle-aged to elderly white women and men, LVFTs were more commonly visualized in individuals with lower body mass indices, and were associated with innocent precordial murmurs. Additional studies with a larger number of individuals are needed to confirm the association between LVFTs and ECG-LVH. Nonetheless, LVFTs do not appear to be associated with the risk of all-cause mortality.
Disclosures: This work was supported in part by a contract (N01-HC-25195) with the National Heart, Lung, and Blood Institute and by a research career award (2K24 HL04334) and R01HL080124 from the National Heart, Lung and Blood Institute, Bethesda, Maryland (to Dr. Vasan).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.