Author + information
- Received July 18, 2006
- Revision received February 12, 2007
- Accepted February 13, 2007
- Published online June 26, 2007.
- Eirik Pettersen, MD⁎,1,⁎ (, )
- Thomas Helle-Valle, MD†,1,
- Thor Edvardsen, MD, PhD⁎,
- Harald Lindberg, MD, PhD‡,
- Hans-Jørgen Smith, MD, PhD§,
- Bjarne Smevik, MD§,
- Otto A. Smiseth, MD, PhD⁎ and
- Kai Andersen, MD, PhD‡
- ↵⁎Reprint requests and correspondence:
Dr. Eirik Pettersen, Department of Cardiology, Faculty of Medicine, University of Oslo and Rikshospitalet-Radiumhospitalet Medical Center, 0027 Oslo, Norway.
Objectives The aim of the present study was to characterize the contraction pattern of the systemic right ventricle (RV).
Background Reduced longitudinal function of the systemic RV compared with the normal RV has been interpreted as ventricular dysfunction. However, longitudinal shortening represents only one aspect of myocardial deformation, and changes in contraction in other dimensions have not previously been described.
Methods Fourteen Senning-operated patients age 18.4 ± 0.9 years (mean ± SD) with transposition of the great arteries were studied. We compared the contraction pattern of the systemic RV with findings in the RV and left ventricle (LV) of normal subjects (n = 14) using tissue Doppler imaging and magnetic resonance imaging.
Results In the systemic RV free wall, circumferential strain exceeded longitudinal strain (−23.3 ± 3.4% vs. −15.0 ± 3.0%, p < 0.001) as was also the case in the normal LV (−25.7 ± 3.1% vs. −16.5 ± 1.7%, p < 0.001), opposite from the findings in the normal RV (−15.8 ± 1.3% vs. −30.7 ± 3.3%, p < 0.001). Strain in the interventricular septum did not differ from normal. Ventricular torsion was essentially absent in the systemic RV (0.3 ± 1.8°), in contrast to a torsion of 16.7 ± 4.8° in the normal LV (p < 0.001).
Conclusions In the systemic RV as in the normal LV, there was predominant circumferential over longitudinal free wall shortening, opposite from findings in the normal RV. This may represent an adaptive response to the systemic load. Noticeably, however, the systemic RV did not display torsion as found in the normal LV.
In patients operated with atrial switch for transposition of the great arteries (TGA), the morphologic right ventricle (RV) supports the systemic circulation, representing a potential for systemic RV failure (1,2). Findings of reduced longitudinal function in the systemic RV have been proposed to reflect global ventricular dysfunction (3,4). However, reduced longitudinal shortening describes only one aspect of myocardial deformation and is also present in patients without heart failure. This raises the possibility of compensatory increased function in another dimension due to altered loading conditions. If present, such responses should be defined before evidence of ventricular dysfunction can be established.
The aim of the present study was, therefore, to describe systemic RV function in terms of longitudinal and circumferential myocardial shortening and ventricular torsion, employing tissue Doppler (TDI) and magnetic resonance imaging (MRI).
Senning-Operated TGA Patients
Eighteen patients (6 women) with TGA operated as infants with atrial switch as described by Senning were identified from the hospital’s database. One patient died suddenly before the planned examination. Three patients were excluded from analyses: 2 because of hemodynamically significant lesions, the third because of poor quality MRI recordings. Thus 14 patients were included.
Healthy Control Subjects
Fourteen healthy volunteers (4 women) were examined by echocardiography and MRI.
Operated Control Subjects
To exclude that systemic RV findings might merely reflect postoperative effects, we performed echocardiograms in 14 patients (3 women) with systemic left ventricles (LVs) age 18.1 ± 4.8 years successfully operated for other congenital heart disease (3 for aortic valvular disease, 5 for perimembranous ventricular septal defects, 6 for TGA with arterial switch).
All study subjects gave written informed consent (minors by proxy) to participate in the study. The protocol was approved by the regional ethics committee.
Clinical examination and exercise testing
The Senning-operated TGA patients underwent clinical examination and electrocardiogram (ECG). They performed bicycle exercise testing with an initial work load of 25 W or 50 W, individualized to yield an exercise duration of about 10 min, increased by 25 W every 2 min. Ventilatory oxygen uptake was measured with an open circuit-technique (EOS/SPRINT, E. Jaeger, Wurzburg, Germany).
Recordings were obtained with a GE Vingmed Vivid 7 scanner (GE Vingmed Ultrasound, Horten, Norway). Gray scale and color TDI images (frame rate 90 to 240 frames/s) of both ventricles were analyzed off-line using dedicated software (EchoPac, GE Vingmed Ultrasound) by an observer blinded to the MRI data.
Ventricular geometrywas described by the position of the interventricular septum (IVS) as the ratio between end-systolic RV and LV septum-to-free-wall diameters and by free wall radius of curvature measured using a dedicated Matlab application (MathWorks Inc., Natick, Massachusetts) (Fig. 1).
Regional myocardial function was assessed by peak systolic strain, expressing the percent change in segment length from end-diastole, and strain rate, expressing the rate of deformation, derived from TDI recordings (5). End-diastole was defined at the peak R of the electrocardiographic QRS complex, end-systole at the first negative crossover of the velocity curve. Negative strain represents segmental shortening, while positive represents lengthening. Longitudinal regional function was measured in the apical, mid-, and basal segment of the free wall of both ventricles (Fig. 1) as well as in the IVS. Circumferential regional function was measured in the RV and LV free wall at the midventricular level. All measurements were averaged over 3 cardiac cycles.
Magnetic resonance imaging scans were performed in patients and healthy control subjects using a 1.5-T scanner (Magnetom Vision Plus, Siemens, Erlangen, Germany). Right ventricular mass, volumes, and ejection fraction (EF) were calculated from area measurements of multiple short-axis breath-hold images covering the entire RV.
Myocardial Strain by MRI Tagging
Tagged MRI images were obtained as described previously (6). Recordings were analyzed by Harmonic Phase Imaging (HARP, version 1.0, Diagnosoft Inc., Palo Alto, California) (7). Circumferential peak systolic strain at the midventricular level of the RV lateral wall was measured as a reference method for Doppler-derived circumferential strain.
Global basal and apical systolic rotations were measured from tagged MRI images using HARP (6). Ventricular torsion was calculated as the difference between basal and apical rotation. Global rotation could not be calculated for the normal RV, as the rotation algorithm presupposes a relatively circular short-axis shape. However, rotation of the entire heart could be measured using the common center of gravity of the 2 ventricles as a reference point, allowing extraction of regional data for rotation of the normal RV free wall.
All data are presented as mean ± SD. Statistical analysis was performed using SPSS 12.0.1 (SPSS Inc., Chicago, Illinois). Student ttest was used for comparisons between 2 groups and 1-way analysis of variance with post-hoc Bonferroni correction for comparisons between more than 2 groups. Pearson’s correlation coefficient was used where appropriate. A value of p < 0.05 was considered significant. Tissue Doppler imaging and MRI strain measurements were compared by a Bland-Altman plot.
Clinical, echocardiographic, and MRI data of the Senning-operated TGA patients and healthy control subjects are presented in Table 1.All patients were in New York Heart Association functional class 1, except for 1 in class 2 due to sinus node dysfunction. All were in sinus rhythm at rest, and none had undergone additional cardiac surgery since the initial operation. None had ECG changes at rest suggesting myocardial scar or during exercise suggesting ischemia. There was no correlation between RV mass and measures of diastolic function.
In the Senning-operated TGA patients, there was a markedly higher ratio between the RV and LV diameter than that in normal subjects (1.28 ± 0.32 vs. 0.54 ± 0.10, p < 0.001). Furthermore, the radius of curvature of the systemic RV free wall was less than that of the normal RV (2.11 ± 0.39 cm vs. 3.09 ± 0.49 cm, p < 0.001) but similar to the normal LV free wall radius of curvature (2.11 ± 0.39 cm vs. 2.37 ± 0.29 cm, p = NS).
Longitudinal and circumferential shortening
Senning-Operated TGA Patients vs. Healthy Control Subjects
Strain and strain rate values are presented in Tables 2 and 3.⇓⇓All values in the text are from the midsegment, unless otherwise stated. In the systemic RV free wall, circumferential strain was greater than longitudinal strain (−23.3 ± 3.4% vs. −15.0 ± 3.0%, p < 0.001), opposite from findings in the normal RV free wall (−15.8 ± 1.3% vs. −30.7 ± 3.3%, p < 0.001), but similar to the contraction pattern of the normal LV (−25.7 ± 3.1% vs. −16.5 ± 1.7%, p < 0.001) (Fig. 2).Strain rate showed a similar pattern with greater circumferential than longitudinal rate of shortening in the systemic RV free wall (−1.5 ± 0.5 s–1vs. −1.1 ± 0.2 s−1, p = 0.001). However, strain rate values in the systemic RV were significantly less than in the normal LV. In the IVS, there was no difference in strain between the patients with systemic RVs and the healthy control subjects. However, as in the free wall, strain rate in the IVS was significantly reduced in the Senning-operated TGA patients compared with the control subjects.
Operated Control Subjects Versus Healthy Control Subjects and Senning-Operated TGA Patients
There was no significant difference in strain or strain rate between the operated control subjects and the healthy control subjects, neither for the RV nor the LV. Accordingly, there was a significant difference in contraction pattern between the systemic RV and the RV of the operated control subjects, similar to the difference between the systemic RV and the RV of the healthy subjects.
Strain by TDI and MRI
There was good agreement between strain obtained by TDI and MRI (Fig. 3).The correlation coefficient for strain by the 2 techniques was r = 0.82 (p < 0.001). There was no significant difference in circumferential strain by the 2 techniques. Importantly, MRI confirmed the finding of higher circumferential strain in the systemic RV than in the normal RV (−22.4 ± 2.4% vs. −16.4 ± 2.6%, p < 0.001).
Ventricular rotation and torsion
In the normal LV, MRI demonstrated global clockwise rotation of the base of the ventricle and counterclockwise rotation of the apex (Fig. 4),5.1 ± 1.5° and −11.6 ± 4.6°, respectively (p < 0.001), with resulting torsion of 16.7 ± 4.8°. Similarly, the RV free wall in the normal heart rotated with the LV, with basal and apical rotation of 3.1 ± 1.2° and −8.2 ± 2.7°, respectively (p < 0.001), and resulting torsion of 11.4 ± 2.6°. In contrast, global rotation of the systemic RV was essentially absent both at the basal and apical levels, 0.3 ± 1.3° and 0.0 ± 2.1°, respectively (p = NS). Consequently, there was no significant global torsion of the systemic RV (0.3 ± 1.8°). In the subpulmonary LV, rotation was essentially absent at the basal level, but present, although to a lesser degree than in the normal LV, at the apical level, −0.4 ± 2.4° and −5.8 ± 2.8°, respectively (p < 0.001). The resulting torsion was 5.4 ± 2.0°.
The present study demonstrates a shift in the systemic RV free wall from longitudinal to circumferential shortening when compared with the normal RV but without any difference in IVS shortening. Consequently, the systemic RV contraction pattern resembles that of the normal LV, although without the ventricular torsion of its normal systemic counterpart. Strain rate, however, was significantly decreased in the systemic RV compared with the normal LV. The predominant circumferential over longitudinal free wall contraction might represent an adaptive response to the systemic load, while the lack of torsion and reduced strain rate may suggest incipient myocardial dysfunction.
Geometrical changes and patterns of myocardial shortening
The leftward shift of the IVS with decreased radius of curvature gives the systemic RV a more circular short-axis shape comparable to that of the normal LV. This may facilitate circumferential shortening through reduced regional wall stress. Furthermore, in right ventricular pressure overload, it is mainly the middle circumferential layer that hypertrophies (8,9). Also in this regard, the systemic RV is more similar to the normal LV, which has a well-developed middle circumferential layer (10). A relative increase in circumferential fiber mass may, thus, also contribute to the predominant circumferential RV free wall shortening.
Another issue to be discussed is whether open chest surgery per se affects the RV strain pattern. The normal strain and strain rate in the operated control subjects rule out the possibility that the observed contraction pattern in the systemic RV could be merely a postoperative effect.
In the normal LV, torsion contributes to energy-efficient ejection (11). Our findings on ventricular torsion in the healthy control subjects are consistent with recent findings by MRI tagging and subsequent 3-dimensional reconstruction (12). A previous study reported regional differences in rotation in the systemic RV (13). However, when values are averaged over segments, the resulting global rotation is close to 0 both at the basal and the apical level with no definite global ventricular torsion.
Torsion results from contraction of myocardial fibers arranged obliquely to the ventricular axis (14). However, the fraction of obliquely oriented muscle fibers in the chronically pressure loaded RV is approximately half of that in the normal RV, while the fraction of circumferentially oriented fibers is higher (9). Furthermore, in the normal heart, the systemic ventricle may be partly responsible for the rotation of the subpulmonary ventricle. Reduced torsion of 1 ventricle may, therefore, affect the other through ventricular coupling effects. These factors might contribute to the observed changes in ventricular torsion with consequently less energy-efficient ejection, possibly representing a potential for myocardial dysfunction.
Adaptive response or ventricular dysfunction?
It must be addressed whether our findings reflect an adaptive response or systemic RV dysfunction. None of our patients had clinical evidence of heart failure, and the mean systemic RV EF was close to that reported as normal in this setting (15). Moreover, the contraction pattern was uniform in all patients with systemic RV, regardless of their RV EF. Furthermore, there was no significant difference in strain between the systemic RV and normal LV.
The TGA patients had reduced exercise capacity, consistent with previous findings (16). However, these patients also have pulmonary abnormalities (17) and reduced atrioventricular filling rates during exercise (18). The latter study demonstrated an increase in contractility during inotropic stimulation as seen in the normal LV, indicating preserved systolic systemic RV function despite reduced exercise capacity.
These aspects considered, the reduced longitudinal strain in the systemic RV should not necessarily be interpreted as ventricular failure. Since strain values were similar to those in the normal LV, the shortening pattern may represent an adaptive response to the systemic load. However, strain rate was significantly reduced and rotation absent compared with the normal LV. These findings may imply incipient myocardial dysfunction.
Our study group is selected, and only includes patients without heart failure. Further changes in myocardial deformation may be expected with the development of myocardial dysfunction, and this warrants further study.
The control subjects were slightly older than the TGA patients. As RV longitudinal function is reduced with age (19), it is unlikely that the lower longitudinal strain in the systemic RV of the younger TGA patients could be the result of age-dependent changes.
Strain might be underestimated, Doppler strain by angle dependency and MRI strain by low temporal resolution. Biological variability may also influence the results. As seen from the Bland-Altman plot, variability from this does not seem to represent a problem in the present study.
Circumferential strain and strain rate were assessed only in 1 region. There might be regional variations not discovered by our approach. Furthermore, measurements were performed only in 2 dimensions. A 3-dimensional analysis could provide further insights into the deformation of the systemic RV.
As described, the reference point differs somewhat for calculation of normal RV free wall rotation and global rotation of the systemic ventricles. Thus, although these measures are not directly comparable, they can be compared with regard to presence, direction, and approximate magnitude of rotation.
The present study demonstrates a shift in the systemic RV free wall from longitudinal to circumferential shortening when compared with the normal RV but no change in septal shortening. This contraction pattern resembles that of the normal LV and may be explained by alterations in ventricular geometry and/or myocardial hypertrophy. In contrast to the normal LV, ventricular torsion is essentially absent and strain rate is reduced. While the predominant circumferential contraction may represent an adaptive response to the systemic load, the lack of torsion and reduced strain rate might imply incipient myocardial dysfunction. Since circumferential shortening predominates in the systemic RV, it may prove more valuable in demonstrating myocardial dysfunction. However, this warrants further study.
The authors thank Stein Inge Rabben, PhD, for providing the Matlab application for analysis of radius of curvature.
↵1 Drs. Pettersen and Helle-Valle are recipients of fellowships from the Norwegian Research Council and the Norwegian Council on Cardiovascular Diseases, respectively.
- Abbreviations and Acronyms
- ejection fraction
- interventricular septum
- left ventricle
- magnetic resonance imaging
- right ventricle
- tissue Doppler imaging
- transposition of the great arteries
- Received July 18, 2006.
- Revision received February 12, 2007.
- Accepted February 13, 2007.
- American College of Cardiology Foundation
- Derrick G.P.,
- Josen M.,
- Vogel M.,
- Henein M.Y.,
- Shinebourne E.A.,
- Redington A.N.
- Urheim S.,
- Edvardsen T.,
- Torp H.,
- Angelsen B.,
- Smiseth O.A.
- Helle-Valle T.,
- Crosby J.,
- Edvardsen T.,
- et al.
- Garot J.,
- Bluemke D.A.,
- Osman N.F.,
- et al.
- Sanchez-Quintana D.,
- Anderson R.H.,
- Ho S.Y.
- Greenbaum R.A.,
- Ho S.Y.,
- Gibson D.G.,
- Becker A.E.,
- Anderson R.H.
- Beyar R.,
- Sideman S.
- Haber I.,
- Metaxas D.N.,
- Geva T.,
- Axel L.
- Fogel M.A.,
- Weinberg P.M.,
- Fellows K.E.,
- Hoffman E.A.
- Streeter D.D. Jr..,
- Spotnitz H.M.,
- Patel D.P.,
- Ross J. Jr..,
- Sonnenblick E.H.
- Gilljam T.,
- Sixt R.