Author + information
- Received January 13, 2000
- Revision received March 15, 2000
- Accepted April 28, 2000
- Published online October 1, 2000.
- Dave T Kim, MD∗,
- Angela C Lai, BS∗,
- Chun Hwang, MD, FACC∗,
- Ling-Tao Fan∗,
- Hyrar S Karagueuzian, PhD, FACC∗,
- Peng-Sheng Chen, MD, FACC∗ and
- Michael C Fishbein, MD, FACC†,*
- ↵*Reprint requests and correspondence: Dr. Michael C. Fishbein, Department of Pathology and Laboratory Medicine, UCLA School of Medicine, 10833 Le Conte Avenue, Los Angeles, California 90095-1732
We sought to study the anatomy of human ligament of Marshall (LOM).
Although the LOM has been implicated in the genesis of focal atrial tachyarrhythmias, its gross anatomic and microscopic features in humans hearts have not been completely defined.
We studied seven postmortem human hearts from five men and two women with a mean age of 52 ± 26 years. Four did not have any heart disease. One woman had dilated cardiomyopathy, and two men had chronic atrial fibrillation. A block of tissue encompassing the LOM from the coronary sinus (CS) cephalad, between the atrial appendage and left pulmonary veins, was dissected. Serial sections from this tissue were then stained with hematoxylin and eosin, trichrome, and/or tyrosine hydroxylase.
The LOM consists of multiple sympathetic nerve fibers, ganglia, blood vessels and multiple myocardial tracts (Marshall Bundles) insulated by fibrofatty tissue. One or more myocardial tracts was inserted directly into the left atrial free wall and CS. The distance between insertion sites was 7.8 ± 2.5 mm. Nerve fibers, some tyrosine hydroxylase positive, were present within the fibrofatty matrix and within the myocardial tracts.
Human LOM 1) is innervated by sympathetic nerve fibers; 2) is more complex than the LOM in canine hearts; and 3) has multiple myocardial tract insertions into the left atrial free wall and CS, forming a substrate of reentry. Radiofrequency catheter ablation from the CS may fail to reach the free wall insertion.
In 1850 John Marshall first described a vestigial fold of the pericardium that contains fibrous bands, small blood vessels and nervous filaments enveloped in fat. This structure came to be known as the ligament of Marshall (LOM) (1). It contains the oblique vein of Marshall (VOM) that drains into the coronary sinus (CS). The LOM, which courses obliquely above the left atrial appendage and lateral to the left superior pulmonary veins, remained in virtual obscurity until Scherlag et al. (2) proposed that the cardiac muscle fibers (Marshall bundles) within this LOM were the terminal end of an inferior interatrial tract in canine hearts. They also demonstrated double potential recordings near the LOM and proposed that this was due to activation from the left atrial tract that is well insulated from the myocardium by fibrofatty tissue. Doshi et al. (3) recently showed that the LOM is a catecholamine sensitive automatic focus in canine left atrium and that it may be a source for adrenergic atrial tachyarrhythmias. It may also play a role in inducing chronic atrial fibrillation (AF) by providing a source of rapid activation. Hwang et al.’s subsequent studies (4,5) strengthened the possible role of the LOM as a source of focal AF. Despite the growing interest in the LOM, a detailed study of its microscopic structure in humans is lacking. Mounting evidence suggests that the LOM may possess a significant role in generating certain atrial tachyarrhythmias. We therefore investigated its anatomic structure in humans to attempt to elucidate possible mechanisms for these tachyarrhythmias.
Seven postmortem human hearts were obtained from five men and two women with a mean age of 52 ± 26 years (Table 1). Four did not have any heart disease; one woman had dilated cardiomyopathy; and two men had chronic AF.
A rectangular block of tissue encompassing the oblique VOM arising from the CS between the atrial appendage and left pulmonary veins was dissected (Fig. 1) and fixed in formalin. Serial tissue samples of the entire region, 2 to 3 mm thick were processed routinely and embedded in paraffin. Samples were cut coronally perpendicular to the epicardium in the five specimens without structural heart disease. In the two with structural heart disease or chronic AF, the sections were cut sagittally.
Paraffin blocks were serially sectioned (269 ± 29 slides per case). Every tenth section of atrial tissue was stained with hematoxylin and eosin and examined by light microscopy. After regions of interest that included the LOM were identified, unstained adjacent sections were stained with Masson’s trichrome stain and for tyrosine hydroxylase using a standard immunohistochemical technique (avidin-biotin-peroxidase). Measurements of the muscle tract dimensions were made using a 7× objective monocular lens with a built-in ruler accurate to within 0.1 mm. The inscribed ruler was used to measure the diameter and length of the myocardial tracts.
In all seven specimens, the LOM was identified along the VOM that originates at the proximal CS. Hematoxylin and eosin and trichrome staining revealed myocardial tracts, well insulated in fibrofatty tissue, inserting into the left atrial free wall and the musculature surrounding the proximal CS (Fig. 1). One or more myocardial tracts were present coursing either superficial or deep to the VOM.
Multiple insertion of the Marshall bundles
Table 2is a summary of the dimensions of the myocardial tracts present within the LOM. The diameter was measured as the greatest width along the myocardial tract and the length as the straightest distance between the two insertion points, that is, from the insertion point into the atrial free wall and extending to the intima of the CS. The distance from the CS intima was used so that a reference point could be used via CS venograms in future procedures when attempting to locate the approximate distance of the insertion point into the atrial free wall. Four specimens exhibited both superficial and deep tracts in relation to the VOM, two with a superficial tract and one with a deep tract. In specimens 2 and 7, the point of insertion into the atrial free wall could not be established. The visible tracts were measured to be 8.0 mm in length in specimen 2 and 10 mm in specimen 7. These measurements were used in calculating the average length; however, the actual lengths would be longer. In specimen 6 the superficial tract ended at 3 mm without reinserting into the atrial free wall. The diameters of these tracts were also measured at their greatest width on the microscopic field. The mean superficial tract measured 7.7 ± 2.2 mm in length and 0.7 ± 0.2 mm in diameter. The mean deep tract measured 8.0 ± 2.8 mm in length and 0.7 ± 0.3 mm in diameter. The mean of all the tracts combined was 7.8 ± 3.9 mm in length and 0.7 ± 0.2 mm in diameter.
A major finding of the study is that the LOM in humans consists of Marshall bundles that have multiple insertions into the CS musculature and the free wall of the left atrium. The presence of multiple insertions provides a possible anatomical substrate for reentrant excitation. These findings also indicate that the human LOM is anatomically more complex than the LOM of dogs (2,3). In addition, similar to the canine LOM, the human LOM contains multiple sympathetic nerve fibers. The anatomical proximity of the LOM to the sympathetic nerves may provide a mechanism for adrenergic atrial tachyarrhythmias (6) in humans.
The anatomical structure of the LOM
With the development of the left innominate vein, the majority of venous blood from the cranial portion of the human embryo returns to the heart through the right common cardinal vein, destined to become the superior vena cava. The LOM is the developmental vestige of the obsolete left common cardinal vein and associated structures. The ligament described by Marshall (1) consists of fibrous bands, small blood vessels and nervous filaments. Scherlag et al. (2) corroborated these findings in canine hearts and demonstrated muscle tracts within the LOM. They also recorded double potentials in this region. The first potential is from the left atrial myocardium. The second potential comes from an electrically active muscle bundle within the LOM. Similar double potentials can also be recorded within the persistent left superior vena cava in humans (7). In Scherlag’s studies, the vast majority of the muscle tracts did not reinsert into the adjacent musculature. Our anatomic findings from human hearts differ in that the Marshall bundle forms multiple complex connections with adjacent atrial myocardium and the CS musculature. These muscle tracts are well insulated by fatty tissue and are innervated by sympathetic nerve fibers.
Focal source of atrial tachyarrhythmias
Haissaguerre and colleagues (8,9) reported successful termination of AF in humans by radiofrequency catheter ablation at focal sites within the atrium. These sites were identified by double potentials around the pulmonary veins, especially near the left superior pulmonary vein (LSPV) (4). The left superior pulmonary vein is adjacent to the LOM. Rapid focal discharges from those sites trigger AF. These studies support a proposal made by Prinzmetal et al. (10) in 1950, who showed that a focal source of activation could explain the mechanisms of atrial arrhythmias, including AF.
The area near the LSPV or the LOM may serve as a source of AF, but the mechanisms by which rapid activations occur in that region remain unclear. Based on the anatomical data of this study, there are at least two possibilities. Because abundant sympathetic nerve fibers co-localize with muscle bundles within the LOM, it is possible that sympathetic activity may bring these muscle bundles to threshold by increased automaticity. The sympathetic stimulation could also result in triggered activity in that region. Both Hwang (11) and Doshi (3) showed that isoproterenol could induce automatic atrial tachycardia from the LOM in humans and in dogs. Our anatomical study also points to a second possible explanation for the rapid activation in that region. The complex patterns of connections among insulated muscle bundles at the lower end of the LOM provide substrates for reentrant excitation. This local reentrant activity could serve as a source of AF. Because of the small size of the reentrant circuit, this mechanism may escape detection by routine electrophysiological recordings using large electrode catheters.
Radiofrequency ablation of myocardial tract insertion sites
An important clinical implication of this study relates to radiofrequency ablation of AF (4,8,9,12,13). For arrhythmia arising from the LOM, the ablation could be performed within the CS (14) or within the LSPV (4). Complications such as perforation (15) or stenosis (16) have been reported with ablation within vascular structures.
Our present studies show that there are two common insertion sites of Marshall bundles: 1) the Marshall bundle could directly insert into the musculature surrounding the proximal CS near the origin of the VOM, or 2) the muscle bundle could insert distally into the posterior atrial free wall, approximately 7.8 ± 2.5 mm superior to the CS. Based on our findings, we postulate that ablation could be applied at the second site, that is, from the endocardium of the inferior posterolateral left atrium. Radiofrequency ablation at that site could sever the connection between Marshall bundles and the CS and the left atrial free wall, resulting in the cure of atrial arrhythmia. This approach may be preferable to the application of energy directly within the CS or the pulmonary veins, because it minimizes the possibility of vascular perforation or stenosis. Recently, Hwang et al. (5) reported successful ablation of AF arising from the LOM by radiofrequency energy application from the left atrial endocardium with the guidance of a catheter within the VOM. The success of that procedure is predicted by our anatomical studies.
This study was done with a relatively small number of specimens. Distal (cephalad) insertions of the myocardial tracts in two specimens were cut off and could not be visualized. In addition, two of the seven specimens did not stain positively for TH, possibly due to the use of autolyzed tissue obtained at autopsy.
There were two major findings in this study: 1) human LOM consists of multiple myocardial tract insertions that may form substrates for reentrant excitation, and 2) radiofrequency catheter ablation aimed at the insertion sites of Marshall bundle in the left atrium free wall may cure focal AF originated from the LOM.
We thank Nina Wang for her technical assistance and Elaine Lebowitz and Judy Wiltz for expert secretarial assistance.
☆ This study was supported by grants from the Pauline and Harold Price Endowment to Dr. Chen, a Cedars-Sinai ECHO Foundation Award to Dr. Karagueuzian, a Piansky Family Endowment to Dr. Fishbein, an NIH SCOR grant in Sudden Death (P50-HL52319), an AHA National Center Grant-in-Aid (9750623N, 9950464N), a UC-TRDRP 6RT-0020 and a grant from the Ralph M. Parsons Foundation, Los Angeles, California.
- atrial fibrillation
- coronary sinus
- ligament of Marshall
- left superior pulmonary vein
- tyrosine hydroxylase
- vein of Marshall
- Received January 13, 2000.
- Revision received March 15, 2000.
- Accepted April 28, 2000.
- American College of Cardiology
- Marshall J
- Scherlag B.J,
- Yeh B.K,
- Robinson M.J
- Doshi R.N,
- Wu T.-J,
- Yashima M,
- et al.
- Hwang C,
- Wu T.-J,
- Doshi R.N,
- Peter C.T,
- Chen P.-S
- Coumel P
- Jais P,
- Haissaguerre M,
- Shah D.C,
- et al.
- Prinzmetal M,
- Corday E,
- Brill I.C,
- et al.
- Brignole M,
- Menozzi C,
- Gianfranchi L,
- et al.
- Tracy C.M,
- Swartz J.F,
- Fletcher R.D,
- et al.
- Kay G.N,
- Chong F,
- Epstein A.E,
- Dailey S.M,
- Plumb V.J
- Nakagawa H,
- Matsudaira K,
- Scherlag B.J,
- et al.
- Robbins I.M,
- Colvin E.V,
- Doyle T.P,
- et al.