Author + information
- Received January 30, 2006
- Revision received June 13, 2006
- Accepted July 3, 2006
- Published online November 21, 2006.
- Arti J. Choure, MD⁎,
- Mario J. Garcia, MD, FACC⁎,
- Barbara Hesse, MD⁎,
- Matthew Sevensma, DO⁎,
- George Maly, MD†,
- Neil L. Greenberg, PhD⁎,
- Lynn Borzi, RN, MBA⁎,
- Stephen Ellis, MD, FACC⁎,
- E. Murat Tuzcu, MD, FACC⁎ and
- Samir R. Kapadia, MD, FACC⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Samir Kapadia, Department of Cardiovascular Medicine, Associate Professor of Medicine, Cleveland Clinic, 9500 Euclid Avenue/F25, Cleveland, Ohio 44195.
Objectives We sought to determine the in vivo anatomical relationships between mitral annulus (MA) and coronary sinus (CS) as well as CS and left circumflex coronary artery using cardiac computed tomography.
Background Percutaneous treatment of mitral regurgitation (MR) by annuloplasty via CS is under development. Success of such treatment depends on the close anatomical proximity of the MA to the CS. The in vivo data regarding this anatomical relationship in humans are scant. We investigated this relationship using contrast multidetector computed tomography.
Methods We studied 25 normal individuals and 11 patients with severe MR (3 to 4+) due to mitral valve prolapse. Separation between MA and CS was measured in standard planes, in 4-chamber (4C), 2-chamber (2C), and 3-chamber views. Distance from ostium of CS to the intersection with left circumflex (LCX), and anatomical relation of LCX and CS were determined using 3-dimensional mapping (Philips Brilliance, Philips Medical Systems, Amsterdam, the Netherlands).
Results There was significant variance of CS to MA separation at all planes. Separation of CS and MA was increased in lateral location (4C) and decreased in posterior location (2C) in the MR group with increase in MA size. Left circumflex artery crossed between CS and MA in 80% of patients. The LCX crossed CS at a variable distance from the ostium of CS (86.5 ± 21 mm, range 37 to 123 mm)
Conclusions There is significant variability in the relation of CS to MA in humans. Coronary sinus to MA distance increases in patients with severe MR and annular dilation, mainly in the posterolateral location. The left circumflex crosses under the CS the majority of times, but with a significant variability in the location where it crosses the CS. These anatomical features should be taken into consideration while selecting percutaneous treatment strategies for mitral valve repair.
According to the population-based Strong Heart Study, the prevalence of mitral regurgitation (MR) is reported to be as high as 21% (1–3). Current surgical options include either repair or replacement of mitral valve. Studies have suggested that mitral valve repair (MVR) is superior to valve replacement because the former is associated with lower operative mortality, improved late survival, reduced risk of endocarditis, fewer thromboembolic complications, and better preservation of left ventricular function (4–10). The number of MVRs performed each year has increased from 23% in 1990 to 32% in 1999 (11). There is some evidence that MVR with an annuloplasty ring in patients with cardiomyopathy from end-stage heart failure may help restore left ventricular function (12). However, due to concerns regarding significant surgical morbidity and mortality, use of surgical mitral annuloplasty for advanced heart failure is limited (13).
Percutaneous MVR techniques have, therefore, attracted significant attention in recent years. The possibility of minimally invasive correction of symptomatic valvular pathology has obvious attraction for patients. This may offer decreased recovery time and avoid surgical morbidity/mortality for the patients who are correctly selected for these procedures. Currently, several concepts for repair of MR from different etiologies are under various stages of development from bench testing to early clinical trials (14). One of them is to exploit the anatomical relationship of the coronary sinus (CS) with the mitral annulus (MA), to modulate shape and size of the annulus to allow proper coaptation of leaflets. Several animal studies have already been performed, and some human studies are under investigation with different CS devices to correct MR related to ischemia or cardiomyopathy. These animal studies have shown the feasibility of percutaneous CS-based mitral annuloplasty (15–18).
Little in vivo quantitative information is available, however, regarding the anatomical relationship between the CS and MA and whether it changes in various stages of MR. The precise anatomical relationship between the CS and left circumflex coronary artery (LCX) has also not been formally investigated. Some authors have suggested that CS devices may potentially impinge on the LCX during annuloplasty via the CS (19).
Multidetector computed tomography (MDCT) has high spatial resolution that allows precise non-invasive delineation of cardiac anatomy. We sought to study the anatomical relationship between the CS and MA and between the CS and LCX artery using cardiac, contrast-enhanced MDCT and the alteration that occurs in this anatomy from MR due to mitral valve prolapse (MVP).
We retrospectively studied 27 normal patients and 14 consecutive patients with severe MR (3 to 4+ by echocardiographic criteria) due to MVP. All patients had undergone contrast-enhanced MDCT under an institutional-review-board-approved research protocol using Philips 16- or 40-slice MDCT (Philips Medical Systems, Amsterdam, the Netherlands) between January and April of 2005. All patients provided informed consent. The normal patients were those with no known history of coronary artery disease or valvular heart disease who underwent MDCT for purpose of CT angiography. The patients in the MVP group were scheduled to undergo either surgical valve repair or replacement and had agreed to enroll in CT angiography study. Two patients in the normal group and 3 in the MR group were excluded because the image quality was compromised due to motion artifacts to such an extent that an accurate interpretation could not be made.
We retrospectively reviewed the contrast-enhanced cardiac computed tomography (CT) scans that were already performed for purpose of CT angiography under an institutional review board–approved protocol. These gated contrast-enhanced spiral CT scans were performed using a 16-slice MDCT scanner (Mx8000, Philips Medical Systems) using the following protocol: 120 kVp, 300 mAs, 0.5 s rotation time, collimation of 4 × 1 mm, and a pitch of 0.375. A total of 100 to 120 cc of non-ionic contrast medium was administered at 3 to 3.5 cc/s.
The scans commenced after a delay of 20 s. The acquisition time varied from 33 to 40 s, covering a distance of 100 to 120 mm, respectively. The electrocardiographic (ECG) information was archived simultaneously along with the scans. The data were retrospectively gated to provide reconstructions at desired phases during the cardiac cycle. All images were then transferred to a dedicated workstation (MxView, Philips Medical Systems) for analysis.
All reconstructed images were manually reoriented into standard 2-dimensional planes including: 1) 4-chamber (4C); 2) 2-chamber (2C); and 3) 3-chamber (3C) views for standardization. Mitral annulus diameter was measured for each patient in these planes. The separation between the MA and CS was also measured in each plane, and was defined as the distance from the MA plane to a parallel line drawn through the center of the CS. This gave a reproducible “distance above” the MA (Figs. 1Ato 1C). The largest CS diameter was measured in vertical and horizontal plane. Coronary arteries and CS were manually tracked and reconstructed using 3-dimensional mapping, Philips, Brilliance (Philips Medical Systems) (Fig. 2).Circumferential distance from the ostium of the CS to the point of intersection with the LCX was derived using true, 3-dimensional length (computer-derived). The anatomical relationship of LCX and CS was also recorded (LCX between CS and MA vs. CS between LCX and MA) (Figs. 3Aand 3B and Figs. 4Aand 4B). The distance between the center of the CS and MA at the crossing point between CS and LCX was determined. The circumferential extent of the MA and CS was determined using 3-dimensional mapping. Left atrial volumes were calculated using the biplane area-length modified Simpson formula: V(cc) = 8A1A2/3πl, where A1and A2represent the enclosed area of the atrial chamber from the 2 orthogonal views, and L is the common diameter directed from apex to base (20).
Descriptive statistics are presented as mean values ± SD for continuous variables and as frequencies and percentages for categoric variables. Continuous variables were compared using 2-way Student ttest. Comparison within groups was made using analysis of variance. A p value of <0.05 was considered statistically significant. All analyses were performed using Statistica statistical software, version 6.0 (StatSoft, Inc., Tulsa, Oklahoma).
Twenty-five individuals (16 men and 9 women; mean age 51 ± 10 years) in the normal group and 11 (10 men and 1 woman; mean age 59 ± 10) in the MR group were studied. Patient characteristics are listed in Table 1.
CS to MA relationship in normal group
The separation between the CS and MA was greatest at the inferolateral location (3C) (Fig. 2c) (mean ± SD: 12.2 ± 3.2 mm) and smallest at the lateral location (4C) (Fig. 2a) (7.8 ± 2.8 mm) (p < 0.001). The posterior (2C) (Fig. 2b) separation was 10.4 ± 2.0 mm. The separation between the CS and MA varied at all locations sampled. The CS plane was on the left atrial side of the MA in all patients studied. The CS diameter was 6.4 ± 2.0 × 5.1 ± 1.4 mm in the lateral (4C); 10.3 ± 3.2 × 9.2 ± 2.1 mm in the posterior (2C) and 7.7 ± 2.0 × 6.0 ± 1.6 mm in the inferolateral (3C) location. The circumference of MA was 105.53 ± 12.97 mm (range: 83.8 to 138.8 mm) and that of CS was 100.44 ± 10.93 mm (range: 80.1 to 127.5 mm).
Relationship of LCX to CS in normal group
The LCX artery crossed between the CS and MA in 80% of the patients (Fig. 3). However, there was considerable variation to the circumferential distance from the ostium of the CS to the intersection with the LCX artery (78.2 ± 18.7 mm; range: 37 to 116 mm) (Fig. 5).The distance between the CS and MA at the point where LCX crosses under CS was 8.0 ± 2.0 mm.
Change in relationship of CS to MA in MR due to MVP patients
The separation of CS and MA was increased in the lateral location (4C) and decreased in posterior location (2C) in the MR group (Table 2).This could suggest a flattening of the MA “saddle” shape; our current programming tools, however, do not allow us to measure non-planarity to prove this statistically. The MA diameter in all planes was increased in the MR group (Table 2). Left atrial volume was larger in the MR group compared with the normal group (92.5 ± 33.5 mm3vs. 45.6 ± 11.2 mm3, p < 0.013). The increase in CS to MA separation in the lateral location (4C) weakly correlated with the increase in left atrial volume (r2= 0.167, p = 0.01) and annular diameter in the anteroposterior plane (2C) (r2= 0.167, p = 0.01). The circumferential extent of the MA in patients with MVP was 138.06 ± 16.21 mm (range: 116.3 to 161.2 mm) and that of the CS was 124.88 ± 16.96 mm (range: 97 to 154.8 mm).
This study has 3 major findings. First, we found significant variability in the distance between the CS and MA in humans (Table 2). Second, the LCX artery crossed under the CS in 80% of patients, but with a significant variability in the location where it crosses the CS. Third, the distance from the CS to the MA increased with annular dilation, mainly in the posterolateral location, suggesting a possible flattening of “saddle of mitral annulus” (20,21). These anatomical features should be taken into consideration while selecting percutaneous treatment strategies for MVR.
The mechanism of valvular regurgitation can be classified based on leaflet dysfunction to provide a practical framework for the evaluation of different approaches (22). Patients with type I dysfunction have normal leaflet motion, and MR is secondary to annular dilatation or leaflet perforation. Those with type II dysfunction have increased leaflet motion with the free edge of the leaflet overriding the plane of the annulus during systole (leaflet prolapse). The most common pathology in this group is the chordal elongation or rupture and papillary muscle elongation or rupture. Patients with type IIIa dysfunction have restricted leaflet motion during both systole and diastole, commonly resulting from leaflet thickening or retraction, chordal thickening, shortening, or fusion and commissural fusion resulting from rheumatic etiology. Patients with type IIIb dysfunction have restricted leaflet motion during systole caused by left ventricular enlargement with apical papillary muscle displacement. Surgical repair has been studied for all these mechanisms of MR, and most of the surgical techniques include mitral annuloplasty as an integral part of repair techniques (23). Current approaches for percutaneous repair of mitral valve include edge-to-edge repair or annuloplasty via the CS, transventricular, or transpericardial approach. There are 3 devices that are being studied for annuloplasty via the CS approach—the Viacor (Viacor Inc., Wilmington, Massachusetts), the Cardiac-Dimension (Cardiac Dimensions Inc., Kirkland, Washington), and the Edwards Viking (Edwards Life Sciences, Irvine, California).
The success of CS annuloplasty approach depends on close proximity of CS to MA. Therefore, unusual anatomic separation between CS and MA may result in a device being inefficient in altering the MA when implanted in the CS. Further, the separations of CS and MA planes and their orientations in the 3-dimensional space may also determine the success of these devices. The exerted force may be distributed differently along the annulus depending on the orientation and separation of the CS and MA planes. It is interesting to note that the left atrial enlargement in patients with MR altered the relation of the MA to the CS differently in specific locations. The separation between them was increased in the lateral location and decreased in the posterior location, whereas it did not change in the inferolateral location. There are no similar studies looking at the in vivo anatomical relationship, although electron-beam cardiac tomography has been utilized in the past to analyze the coronary venous configuration (24). Variation in the CS-MA relationship has been described in autopsied formalin-fixed hearts. Anatomic description is mainly based on postmortem specimens (25–27).
Shinbane et al. (26) reported an anatomic relation between CS and MA in 10 structurally normal cadaver hearts for left accessory pathway localization. They measured the shortest distance between the discrete fibrous point of attachment of the mitral valve leaflet to the MA and the closest intimal surface of the CS starting at 20-mm increments from the CS ostium. Mean distance was reported as 14.1 ± 3.1 mm, 10.2 ± 4.9 mm, and 10.7 ± 3.5 mm at 20, 40, and 60 mm from the CS ostium, respectively. In a separate study reported by Yamanouchi et al. (25), the distance from the ventricular endocardium under MA to the nearest wall of CS ranged between 8.2 ± 2.9 mm to 10.9 ± 3.3 mm depending on the location of the measurements. In yet another study, El-Maasarany et al. (27) reported the shortest distance between the wall of the CS and the endocardium adjacent to the MA at the level of the anterolateral commissure to be 5.2 ± 1.6 mm. Although these data highlight significant variability in the distance between MA and CS, the exact comparison to our in vivo study data is not possible because these were necropsy studies, and the points of measurements were not similar to our study.
The potential problems with CS annuloplasty include LCX artery impingement, dissection or perforation of the CS, and CS thrombosis. The LCX artery impingement could occur due to a close anatomic relationship between the artery and CS. Appreciation of the relationship between the CS and LCX has been defined as a critical factor to the safety of these devices (15). The CS runs in the posterior atrioventricular groove from its origin in the right atrium to the origin of the anterior interventricular vein in the anterior interventricular groove. The LCX artery runs in the atrioventricular groove above the CS and crosses the CS at a variable distance and in a variable path. In 80% of our patients, the LCX artery crossed under the CS to reach the left ventricle. The mechanical action of percutaneous mitral valvuloplasty devices is a “cinching” or pulling in of the MA along the length of the device. The direction of this force could potentially cause impingement on the LCX artery in patients if it crosses under the CS. There is, however, significant variability in the length to the point where the CS intersects with the LCX artery. Contrast-enhanced MDCT can help identify patients in whom the LCX artery will not be a problem when annuloplasty via the CS is attempted. The precise understanding of this anatomy may allow us to select the most appropriate method for an individual patient. Also, knowledge of the individual anatomy including the CS caliber, MA and CS circumference may aid us in designing devices appropriately.
One of the major limitations of our study is the small sample size. This is an initial attempt to define anatomic variation, and our results highlight the need for anatomic assessment of patients for proper selection of treatment approaches. Two patients in the normal group and 3 in MR group were excluded because the image quality was compromised due to motion artifacts to such an extent that an accurate interpretation could not be made. This number can be further minimized with faster scanners, proper sedation, and detailed instructions at the time of scanning. In this study, we investigated MR due to MVP only, and, therefore, the results may not be generalized to patients with MR due to other etiologies such as ischemic or functional MR. Multidetector CT evaluates the heart in ECG-gated static frame. Therefore, the dynamic nature of the relationship between the CS and MA in different phases of the cardiac cycle cannot be determined in this study. Cardiac MRI and 3-dimensional echocardiography are other potential modalities that can be potentially utilized to study these anatomical relationships. However, currently MDCT is the only modality that allows enough resolution to study CS, LCX, and MA relationships. The potential advantage of cardiac MRI and 3-dimensional echo could be the possibility of investigating relation of these structures during different phases of the cardiac cycle. Further, we could not comment on the branches of the CS because our CT scans were not optimized to visualize venous anatomy.
This study underscores the anatomic variation in the relation of CS and MA and suggests dynamic change in the relationship between CS and MA with MR in humans. Contrast-enhanced MDCT is a non-invasive, low-risk test that can readily determine the anatomical relationships to be manipulated with percutaneous mitral valvuloplasty via the CS approach, and may potentially become an intricate part of pre-procedure planning, device development, and patient selection.
The authors thank Amy Moore from Department of Scientific publications for her contribution toward language editing.
- Abbreviations and Acronyms
- coronary sinus
- computed tomography
- left circumflex coronary artery
- mitral annulus
- multidetector computed tomography
- mitral regurgitation
- mitral valve prolapse
- mitral valve repair
- 2 chamber
- 3 chamber
- 4 chamber
- Received January 30, 2006.
- Revision received June 13, 2006.
- Accepted July 3, 2006.
- American College of Cardiology Foundation
- Enriquez-Sarano M.,
- Schaff H.V.,
- Frye R.L.
- Thourani V.H.,
- Weintraub W.S.,
- Guyton R.A.,
- et al.
- Goldman M.E.,
- Mora F.,
- Guarino T.,
- Fuster V.,
- Mindich B.P.
- Maniu C.V.,
- Patel J.B.,
- Reuter D.G.,
- et al.
- Daimon M.,
- Shiota T.,
- Gillinov A.M.,
- et al.
- Kaye D.M.,
- Byrne M.,
- Alferness C.,
- Power J.
- Block P.C.
- Salgo I.S.,
- Gorman J.H. 3rd.,
- Gorman R.C.,
- et al.
- Vassiliades T.A. Jr..,
- Block P.C.,
- Cohn L.H.,
- et al.,
- Society of Thoracic Surgeons,
- American Associations for Thoracic Surgery,
- Society for Cardiovascular Angiography and Interventions,
- American College of Cardiology Foundation,
- American Heart Association
- Alfieri O.,
- Maisano F.,
- Colombo A.
- El-Maasarany S.,
- Ferrett C.G.,
- Firth A.,
- Sheppard M.,
- Henein M.Y.