© 2015, Wiley Periodicals, Inc. DOI: 10.1111/echo.12973
Echocardiography
Changes in Right Ventricular Shape and Deformation Following Coronary Artery Bypass Surgery—Insights from Echocardiography with Strain Rate and Magnetic Resonance Imaging € sner, M.D., Ph.D.,* Derk Avenarius, M.D.,* Siri Malm, M.D., Ph.D.,*,† Amjid Iqbal, M.D., Ph.D.,* Assami Ro Henrik Schirmer, M.D., Ph.D.,*,† Bart Bijnens, Ph.D.,‡,§ and Truls Myrmel, M.D., Ph.D.*,† *University Hospital North Norway, Tromsø, Harstad, Norway; †Department of Clinical Medicine, University of Tromsø, Tromsø, Norway; ‡Pompeu Fabra, Barcelona University, Spain; and §The Catholic University Leuven, Leuven, Belgium
Background: This study was designed to assess whether altered RV geometry and deformation parameters persisted well into the recovery period after presumably uncomplicated coronary artery bypass grafting (CABG). It was our hypothesis that the altered geometry of and load in the RV following pericardial opening would change both regional and global deformation indices for an extensive period postoperatively. Methods and Results: Fifty-seven patients scheduled for CABG underwent preoperative and 8–10 months postoperative magnetic resonance imaging (MRI) for RV volume measurements, and resting echocardiography with assessment of geometry and RV mechanical function determined by tissue Doppler imaging (TDI) based longitudinal strain. Both MRI and echocardiography revealed postoperative dilatation of the RV apex, shortened longitudinal RV length but unchanged RV ejection fraction. Echocardiography parameters associated with filling of the right atrium showed signs of constraint with a reduced systolic filling fraction and increased right atrial size. Right ventricular segmental strain (20 13% vs. 29 20% preoperatively; mean SD, P < 0.0001) was reduced postoperatively in parallel with TAPSE (1.3 0.3 cm vs. 2.2 0.4 cm; P < 0.0001). Conclusion: Post-CABG longitudinal motion of the RV lateral wall is reduced after uneventful CABG despite preserved RV ejection fraction and stroke volume. The discrepancy in various RV systolic performance indicators results from increased sphericity of the RV following opening the pericardium during surgery. Therefore, longitudinal functional parameters may underestimate RV systolic function for at least 8–10 months post-CABG. Changes in deformation parameters should thus always be interpreted in relation to changes in geometry. (Echocardiography 2015;32:1809–1820) Key words: right ventricle, coronary artery bypass graft surgery, pericardial opening, strain Reduced TAPSE, paradoxical septal motion, and abnormal right ventricular filling are all phenomena observed after open heart surgery.1 Several hypotheses have been proposed to explain these alterations, such as pericardial opening,2 pericardial adhesions,3,4 right atrial (RA) dysfunction after cannulation, and poor RV protection during cardiopulmonary bypass.5,6 However, reduction of TAPSE and septal wall motion has not been associated with decreased exercise capacity or left ventricular (LV) dysfunction.7 Furthermore, 3D measurements have shown Address for correspondence and reprint requests: Truls Myrmel, M.D., Ph.D., Department of Cardiothoracic and Vascular Surgery, The Heart and Lung Clinic, University Hospital of North Norway, 9038 Tromsø, Norway. Fax: +47 77628298; E-mail: [email protected]; [email protected]
preserved RV ejection fraction (EF) and stroke volume (SV) suggesting a preserved global RV performance.8 In this study, we observed the changes occurring in RV geometry, segmental deformation, global wall motion, indicators for right ventricular pressure and RV EF before and 9 months after coronary artery bypass grafting (CABG). Based on existing data, we hypothesized that reduced TAPSE after routine cardiac surgery might not reflect reduced RV myocardial function, but that distorted geometric conditions following pericardial opening can induce altered deformation of the RV. In accordance with this, altered diastolic constrain from pericardial opening can also potentially affect the normal filling pattern of the right heart. Thus, in this prospective study, we 1809
€ sner et al. Ro
analyzed RV dimensions and regional deformation using strain and strain rate. We also observed the effect of surgery on atrial size and venous return indices commonly used in echocardiography. To obtain a reference measure of RV volumes, these were derived from cardiac magnetic resonance imaging (MRI). Methods: Protocol Overview and Patient Characteristics: All patients underwent resting echocardiography and MRI 1 to 7 days before and 8 to 10 months after CABG. We have previously published data from the same patient cohort focusing on dobutamine stress echocardiography for assessment of LV viability.9 As stated, the present part of the study was designed to assess the effect of standard CABG on RV geometry and deformation. Between November 2005 and December 2007, 67 patients with coronary artery disease (CAD) scheduled for CABG were included in the study. Exclusion criteria for eligible patients were valvular heart disease, chronic obstructive pulmonary disease, atrial fibrillation, unstable angina, and myocardial infarction within 3 months prior to CABG, left bundle branch block or any contraindications to MRI (e.g., claustrophobia, pacemaker implants, or significantly reduced renal function). As previously described, five of the 67 enrolled patients were excluded from the study prior to CABG, and an additional five did not complete the postoperative protocol.9 Thus, a total of 57 patients were included in the final analysis. Table I displays patient characteristics as earlier published,9 with relevant RV data added. Before CABG, six patients had no angina, and of these, two presented with dyspnea. Five patients presented with dyspnea and angina. Twenty patients had a medical history of previous myocardial infarction; however, none of them had suffered a RV myocardial infarction. Intraoperative myocardial infarctions were defined as a postoperative rise in CKMB >50 U/L. Postoperative ECG ST elevations were used as a marker of transmural myocardial infarctions. The study was approved by the Regional Ethical Committee of North Norway (REK NORD 34/2005), and all patients gave written informed consent to participation. Magnetic Resonance Imaging: All MR images were acquired using a 1.5 T scanner (Philips Intera release 2.1, Best, The Netherlands). Multislice four-chamber, single-slice LV outflow tract and multislice short-axis (SAX) scans were obtained using turbo-field-echo (BTFE) true FISP technique, at 8-mm slice thickness, 2-mm interslice gap, 1.25 9 1.25 in-plane pixel
1810
TABLE I Patient Characteristics Count/Values Patients Men Age (years) Height (m) Weight (kg) Cholesterol treated (mmol/l) Diabetes type 2 Hypertension History of smoking Family history of CAD b-Blockers Angina before CABG Positive treadmill test Angina or a positive treadmill test Previous MI Previous PCI 1 Vessel disease alone 2 Vessel disease alone 3 Vessel disease alone Main stem stenosis (MS) alone MS and 1 vessel MS and 2 vessel MS and 3 vessel Right vessel disease Right vessel disease not revascularized Normal EF (>55%) EF (41–54%) EF (22–40%)
57 55 62 1.75 87 4.84
% or SD (Range) 100% 96.5% 9 (39–83) 0.06 (1.58–1.90) 12 (51–117) 0.96 (3.00–6.75)
12 24 25 33 57 51 38 55
21.1 42.1 43.8 57.9 100.0 89.5 66.0 96.5
20 17 12 11 13 3
34.6 29.8 21.1 19.3 22.8 5.3
2 8 8 30 9
3.5 14.0 14.0 52.6% 15.8%
22 29 6
38.5 50.9 10.5
RV = right ventricle; EF = ejection fraction; CAD = coronary artery disease; CABG = coronary artery bypass grafting; MI = myocardial infarction; PCI = percutaneous coronary intervention; MS = main stem stenosis; EDV = end-diastolic volume. Data on the same study population have been previously published.9
size parallel imaging (SENSE) factor 2. Volumes were measured in SAX views using the ViewForum release 5.1 workstation (Philips). The MRI cardiac analysis software was used for semi-automatic calculation of RV systolic and diastolic volumes. SAX volumes were analyzed using 30 cardiac phases. For each short-axis slice, the inner contour of the right heart was manually drawn in both diastole and systole. The position of the most basal axial slice at the tricuspid valve was crosschecked with the four-chamber view. After confirmation of the drawings, the software automatically calculated the end-systolic volume (ESV), end-diastolic volume (EDV), SV, and EF of the RV. The images were transferred to a PACS system (Impax DS3000 release 4.5, Agfa-Gevaert,
Right Ventricular Function after CABG
Mortsel, Belgium) for assessment of systolic and diastolic RV lateral length, defined as the distance between the insertion of the tricuspid plane in the free lateral wall and the RV apex in the fourchamber view. The amount of left ventricular late gadolinium enhancement (LGE) was measured pre- and postoperatively and expressed as percentage of LV mass as described in detail in a previous publication.9 Echocardiography: Data Acquisition: All echocardiographic studies were performed using an iE33 scanner (S5-1 probe, Philips Medical Systems, Andover, MA) and a 1–5 MHz transducer with the patient in the left lateral decubital position. Conventional 2D gray scale images and tissue Doppler imaging (TDI) data were obtained in the apical two-, three-, and four-chamber views. Additional fourchamber views were optimized for the best quality acquisition of the RV and by searching for the highest radial RV diameter. TDI loops of five consecutive cardiac cycles were acquired, wall-by-wall, of the anteroseptal, the inferoseptal, and the free RV wall in two standard apical views, using a sector angle of 25–30° and frame rate of 120–170 s1. Care was taken to align the walls as parallel as possible to the scanline. The cine loops were obtained in proprietary DICOM format and digitally stored for offline analysis at the original number of frame rates. Echo Image Analysis: The readers of both echocardiographic and MR images were blinded for mutual test results. All 2D and TDI measurements were taken by a single, experienced observer using commercially available software (QLAB, Philips Medical Systems). Andover, MA). The TAPSE was measured using M-mode images of the RV lateral free wall. The RV systolic and diastolic areas were measured in an apical fourchamber view visually optimized for imaging of the RV. In the same projection, we measured the tricuspid ring diameter, apical end-diastolic and end-systolic angles, and maximal radial RV diameters. The apical area and diameter were defined as the area distal to and the diameter at the apical third of the central long-axis length (from apex to the center of the tricuspid plane), respectively. Septal motion was assessed in a parasternal M-mode as described in Figure S1. Global LV function was assessed by calculating ESV, EDV, and EF using the biplane Simpson’s method. Stroke volume, cardiac output (CO), and cardiac index (CI) were derived from the calculated volumes and the heart rate. Right atrial areas (RAA) and diameters were measured at end-systole in the four-chamber view.
Systolic pulmonary artery pressure (PAPsys) was estimated by summation of the tricuspid valve pressure gradient and the atrial pressure derived from the change in the inferior vena cava diameter during respiration.10 All Doppler measurements describing RV function were performed at end-expiration. From pulsed-wave Doppler, placed on the tip of the tricuspid valve, peak E velocity, peak A velocity, and E deceleration time were measured.11 In the hepatic veins, the systolic and diastolic velocity time integral (VTI), the systolic filling fraction (SFF) (VTIsys/ VTIsys + VTIdia), and the regurgitation time between systolic and diastolic flow were derived.12 The corresponding measurements were performed in the superior vena cava and the pulmonary veins. E/e0 was calculated from the pulsed-wave Doppler of the tricuspid blood flow and the pulsed-wave tissue Doppler of the lateral free RV wall.13 Regional septal function was evaluated in the five septal segments from the standard 16 segment model defined by the American Society of Echocardiography.14 Regional RV lateral wall function was assessed in one apical and one basal segment.15 TDI strain analyses were performed using commercially available software (QLAB). The process of TDI analyses and the criteria for discarding measurements have previously been described in detail.9 The aortic valve opening and closure, and thereby the ejection time (ET), were determined by pulsed-wave Doppler of the aortic valve. ET strain was defined as the highest positive or negative peak strain value during ET. In a cine loop of five cardiac cycles, the first cycle was analyzed except in the case of premature beats or insufficient quality of the first beat. Then, the second or one of the following cycles was analyzed. A strain value of 10% was chosen as the cutoff for normal strain values of the left and right ventricle, based on the findings in a large population study.16 Furthermore, the RV free wall was reanalyzed by 2D strain in the same patients (syngo Velocity Vector Imaging, Siemens, Medical Solutions, Mountain View, CA, USA). There we measured mid-myocardial strain in the RV lateral wall only. When comparing TDI and 2D strain, only apical and basal strains are reported. Reproducibility: To determine inter-observer variability for TDI strain, 15 patients were randomly selected and another independent, experienced observer (SM) blinded to all other data, analyzed the preoperative data. To assess intra-observer variability, data from the same 15 patients were reanalyzed by the main observer (AR) in a new 1811
€ sner et al. Ro
random order after at least 6 months. Hence, for RV inter- and intra-observer variability, 30 RV segments were analyzed. Inter-observer and intra-observer variabilities for strain were expressed as the absolute difference between two measurements in percent of their mean. Reproducibility was also tested in comparison with 2D strain, performed on the same dataset on the free RV lateral wall. Statistical Analyses: All statistical analyses were performed using SPSS version 16.0 (SPSS Inc., Chicago, IL, USA). Unless stated otherwise, all data are expressed as mean standard deviation (SD). Paired Student’s ttests were used to test changes in parameters before and after CABG. Dependency of RV strain on clinical data was tested by linear regression for continuous or ANOVA for discrete independent variables. Inter-observer and intra-observer values are presented using Bland–Altman plots. Results: Clinical Outcomes: Following CABG, four patients suffered from angina pectoris and two from dyspnea (Canadian Cardiovascular Society Angina Classification Class [CCACC] 1–2), three developed chronic atrial fibrillation, and one patient had a positive dobutamine stress test without coronary stenoses on coronary angiography. All other patients were free of symptoms (CCACC = 0). None of the patients had more than grade I tricuspid regurgitation or pulmonary hypertension (estimated systolic pulmonary artery pressure >40 mmHg) before or after CABG. Four patients had a perioperative myocardial infarction with CK MB at 130, 101, 87, and 59 U/L, respectively. The total amount of LGE/myocardial mass was unaltered in these four patients as well as in the total study population (from 3.6 to 4.4% (P = 0.557). None of the patients showed ST elevations in the ECG. The degree of ET strain depression after CABG did not correlate with RV EF, RV EDV, LV EF, or clinical data such as age, weight, BSA, hypertension, hypercholesterolemia, diabetes, positive family history for CAD, or smoking. Volumes, Geometry, and Filling Pressures: The geometric and global functional parameters and changes after CABG, measured by MRI and echocardiography, are displayed in Table II. Based on MRI measurements, the RV volumes and EF did not change significantly after surgery. However, there was a significant increase in SV. Postoperatively, MRI revealed a shortened RV lateral longitudinal wall length in diastole 1812
and systole, while radial and anterior–posterior mid-ventricular diameters were not significantly changed. This was in accordance with the echocardiographic unchanged RV FAC as well as changed dimensions shown in Table II, where post-CABG longitudinal diameters were markedly reduced with higher apical and mid-ventricular diastolic and systolic RV diameters. The difference between systolic and diastolic apical diameters increased as well as the systolic apical angle reduction. Figures 1 and 2 show examples of echocardiographic images and MRI with geometric measurements of the RV before and after CABG. Right ventricular geometric measurements reveal a significant postoperative increase of the mid- and apical diastolic diameters, the apical angle, and area, as well as a larger global diastolic area. RA diameters and areas were also increased after CABG. A shortened deceleration time of the tricuspid E wave and a decreased systolic filling fraction in hepatic veins and superior vena cava were demonstrated, as shown in Figure 3. Both increased RA dimensions and decreased hepatic vein SFF indicate increased RV end-diastolic and RA pressures. However, post-CABG vena cava dimensions and estimated RA pressures by respiratory changes remained unchanged. Global and Segmental RV Wall Deformation: Figure 4 displays examples of pre- and postCABG TAPSE and tissue velocities. Figure 5 depicts an example of typical pre- and postCABG strain curves. Figure 6 displays apical and basal strain in the presence or absence of right coronary artery (RCA) disease. TAPSE and ET strain reduction after CABG was independent of the presence of RCA disease and complete revascularization of a significant RCA stenosis. Table III indicates that post-CABG TAPSE was markedly reduced as well as other parameters expressing global systolic wall deformation, that is, ET strain of the free lateral RV wall and septum and regional deformation in the RV free wall, that is, basal and apical RV ET strain, despite normal or slightly reduced preoperative strain. RV tissue velocity (TV) (E) was reduced indicating reduced diastolic wall motion. Figure 7 and Table III show a higher variance (i.e. SD) of pre-CABG compared to post-CABG ET strain. Only segments with high values reduced their strains, whereas most segments with moderate strains stayed unchanged and low strains tended to increase. This effect was similarly demonstrated both measured by TDI or 2D strain. In general, only 10 of 89 preoperative segments were dysfunctional (ET strain > 10%) despite a predominant presence of RCA disease
Right Ventricular Function after CABG
TABLE II Clinical, Geometrical, and Doppler Changes between Pre- and Post-CABG in the Right and Left Ventricle Pre-CABG
BP sys (mmHg) BP dia (mmHg) MRI: RV volumedia (ml) MRI: RV volumesys (ml) MRI: RV EF (%) MRI: RV SV (ml) MRI: RV diameter longdia (cm) MRI: RV diameter longsys (cm) MRI: RV diameter radialdia (cm) MRI: RV diameter radialsys (cm) MRI: RV diameter ant-postdia (cm) MRI: RV diameter ant-postsys (cm) US: HR (1/min) US: RV areadia (cm2) US: RV areasys (cm2) US: RV FAC (%) US: Apical angledia (°) US: Apical angle reduction (°) US: Apical RV diameterdia (cm) US: Apical diameter reduction (cm) US: Apical areadia (cm2) US: Apical FAC (%) US: Mid-RV diameterdia (cm) US: Mid-RV diametersys (cm) US: RA area (cm2) US: Septal motion LV (mm) US: Septal motion RV (mm) US: LV volumedia (ml) US: LV volumesys (ml) US: LV EF (%) US: CO (ml) US: SV (ml) US: CI (ml/m2) US: TV: PAPsys (mmHg) US: TV: E deceleration time (ms) US: HV: Systolic filling fraction (%) US: HV: Regurgitation time (ms) US: SCV: Systolic filling fraction (%) US: SCV: Regurgitation time (ms) US: TR: E/E0 RV Wall US: TV (E`) (cm/s)
Post-CABG
Mean
SD
Mean
SD
n
P
135 80 130 80 40 50 7.0 5.3 3.7 2.9 6.7 6.0 70 20.2 12.5 38 48 7 2.5 0.4 4.5 13.0 3.7 2.8 16.2 8.6 5.2 138 69 52 4823 69 2367 28 235 58.9 32 62 47 6.3 8.7
20 12 32 29 9 12 1.2 1.1 0.8 0.8 1.0 1.0 13 4.6 3.5 15 13 11 0.6 0.5 1.4 29.8 0.5 0.6 3.8 2.8 2.9 40 31 10 1220 17 658 8 86 8.8 50 10 59 2.4 3.1
137 80 134 78 42 56 6.1 5.0 3.5 2.9 6.8 6.0 68 22.0 13.9 37 53 13 2.7 0.6 5.2 22.8 4.0 3.2 18.4 2.8 0.1 130 66 50 4504 66 2172 27 199 48.4 48 52 76 9.4 6.1
20 11 32 24 8 15 0.8 0.8 0.6 0.5 0.7 0.7 12 4.6 3.9 12 10 9 0.5 0.4 1.2 21.1 0.7 0.6 3.9 2.9 2.2 32 25 9 1477 20 636 8 44 8.5 65 11 63 3.4 1.7
57 57 54 54 54 54 54 54 54 54 54 54 57 57 57 57 55 55 55 55 55 55 55 55 57 54 53 49 49 50 50 50 50 32 50 41 42 49 47 50 57
n.s. n.s. n.s. n.s. n.s. 0.008 0.002 n.s. n.s. n.s. n.s. n.s. n.s. 0.010 0.009 0.05 0.018 0.005 0.026 0.024 0.008 n.s. 0.002
Echocardiography
Changes in Right Ventricular Shape and Deformation Following Coronary Artery Bypass Surgery—Insights from Echocardiography with Strain Rate and Magnetic Resonance Imaging € sner, M.D., Ph.D.,* Derk Avenarius, M.D.,* Siri Malm, M.D., Ph.D.,*,† Amjid Iqbal, M.D., Ph.D.,* Assami Ro Henrik Schirmer, M.D., Ph.D.,*,† Bart Bijnens, Ph.D.,‡,§ and Truls Myrmel, M.D., Ph.D.*,† *University Hospital North Norway, Tromsø, Harstad, Norway; †Department of Clinical Medicine, University of Tromsø, Tromsø, Norway; ‡Pompeu Fabra, Barcelona University, Spain; and §The Catholic University Leuven, Leuven, Belgium
Background: This study was designed to assess whether altered RV geometry and deformation parameters persisted well into the recovery period after presumably uncomplicated coronary artery bypass grafting (CABG). It was our hypothesis that the altered geometry of and load in the RV following pericardial opening would change both regional and global deformation indices for an extensive period postoperatively. Methods and Results: Fifty-seven patients scheduled for CABG underwent preoperative and 8–10 months postoperative magnetic resonance imaging (MRI) for RV volume measurements, and resting echocardiography with assessment of geometry and RV mechanical function determined by tissue Doppler imaging (TDI) based longitudinal strain. Both MRI and echocardiography revealed postoperative dilatation of the RV apex, shortened longitudinal RV length but unchanged RV ejection fraction. Echocardiography parameters associated with filling of the right atrium showed signs of constraint with a reduced systolic filling fraction and increased right atrial size. Right ventricular segmental strain (20 13% vs. 29 20% preoperatively; mean SD, P < 0.0001) was reduced postoperatively in parallel with TAPSE (1.3 0.3 cm vs. 2.2 0.4 cm; P < 0.0001). Conclusion: Post-CABG longitudinal motion of the RV lateral wall is reduced after uneventful CABG despite preserved RV ejection fraction and stroke volume. The discrepancy in various RV systolic performance indicators results from increased sphericity of the RV following opening the pericardium during surgery. Therefore, longitudinal functional parameters may underestimate RV systolic function for at least 8–10 months post-CABG. Changes in deformation parameters should thus always be interpreted in relation to changes in geometry. (Echocardiography 2015;32:1809–1820) Key words: right ventricle, coronary artery bypass graft surgery, pericardial opening, strain Reduced TAPSE, paradoxical septal motion, and abnormal right ventricular filling are all phenomena observed after open heart surgery.1 Several hypotheses have been proposed to explain these alterations, such as pericardial opening,2 pericardial adhesions,3,4 right atrial (RA) dysfunction after cannulation, and poor RV protection during cardiopulmonary bypass.5,6 However, reduction of TAPSE and septal wall motion has not been associated with decreased exercise capacity or left ventricular (LV) dysfunction.7 Furthermore, 3D measurements have shown Address for correspondence and reprint requests: Truls Myrmel, M.D., Ph.D., Department of Cardiothoracic and Vascular Surgery, The Heart and Lung Clinic, University Hospital of North Norway, 9038 Tromsø, Norway. Fax: +47 77628298; E-mail: [email protected]; [email protected]
preserved RV ejection fraction (EF) and stroke volume (SV) suggesting a preserved global RV performance.8 In this study, we observed the changes occurring in RV geometry, segmental deformation, global wall motion, indicators for right ventricular pressure and RV EF before and 9 months after coronary artery bypass grafting (CABG). Based on existing data, we hypothesized that reduced TAPSE after routine cardiac surgery might not reflect reduced RV myocardial function, but that distorted geometric conditions following pericardial opening can induce altered deformation of the RV. In accordance with this, altered diastolic constrain from pericardial opening can also potentially affect the normal filling pattern of the right heart. Thus, in this prospective study, we 1809
€ sner et al. Ro
analyzed RV dimensions and regional deformation using strain and strain rate. We also observed the effect of surgery on atrial size and venous return indices commonly used in echocardiography. To obtain a reference measure of RV volumes, these were derived from cardiac magnetic resonance imaging (MRI). Methods: Protocol Overview and Patient Characteristics: All patients underwent resting echocardiography and MRI 1 to 7 days before and 8 to 10 months after CABG. We have previously published data from the same patient cohort focusing on dobutamine stress echocardiography for assessment of LV viability.9 As stated, the present part of the study was designed to assess the effect of standard CABG on RV geometry and deformation. Between November 2005 and December 2007, 67 patients with coronary artery disease (CAD) scheduled for CABG were included in the study. Exclusion criteria for eligible patients were valvular heart disease, chronic obstructive pulmonary disease, atrial fibrillation, unstable angina, and myocardial infarction within 3 months prior to CABG, left bundle branch block or any contraindications to MRI (e.g., claustrophobia, pacemaker implants, or significantly reduced renal function). As previously described, five of the 67 enrolled patients were excluded from the study prior to CABG, and an additional five did not complete the postoperative protocol.9 Thus, a total of 57 patients were included in the final analysis. Table I displays patient characteristics as earlier published,9 with relevant RV data added. Before CABG, six patients had no angina, and of these, two presented with dyspnea. Five patients presented with dyspnea and angina. Twenty patients had a medical history of previous myocardial infarction; however, none of them had suffered a RV myocardial infarction. Intraoperative myocardial infarctions were defined as a postoperative rise in CKMB >50 U/L. Postoperative ECG ST elevations were used as a marker of transmural myocardial infarctions. The study was approved by the Regional Ethical Committee of North Norway (REK NORD 34/2005), and all patients gave written informed consent to participation. Magnetic Resonance Imaging: All MR images were acquired using a 1.5 T scanner (Philips Intera release 2.1, Best, The Netherlands). Multislice four-chamber, single-slice LV outflow tract and multislice short-axis (SAX) scans were obtained using turbo-field-echo (BTFE) true FISP technique, at 8-mm slice thickness, 2-mm interslice gap, 1.25 9 1.25 in-plane pixel
1810
TABLE I Patient Characteristics Count/Values Patients Men Age (years) Height (m) Weight (kg) Cholesterol treated (mmol/l) Diabetes type 2 Hypertension History of smoking Family history of CAD b-Blockers Angina before CABG Positive treadmill test Angina or a positive treadmill test Previous MI Previous PCI 1 Vessel disease alone 2 Vessel disease alone 3 Vessel disease alone Main stem stenosis (MS) alone MS and 1 vessel MS and 2 vessel MS and 3 vessel Right vessel disease Right vessel disease not revascularized Normal EF (>55%) EF (41–54%) EF (22–40%)
57 55 62 1.75 87 4.84
% or SD (Range) 100% 96.5% 9 (39–83) 0.06 (1.58–1.90) 12 (51–117) 0.96 (3.00–6.75)
12 24 25 33 57 51 38 55
21.1 42.1 43.8 57.9 100.0 89.5 66.0 96.5
20 17 12 11 13 3
34.6 29.8 21.1 19.3 22.8 5.3
2 8 8 30 9
3.5 14.0 14.0 52.6% 15.8%
22 29 6
38.5 50.9 10.5
RV = right ventricle; EF = ejection fraction; CAD = coronary artery disease; CABG = coronary artery bypass grafting; MI = myocardial infarction; PCI = percutaneous coronary intervention; MS = main stem stenosis; EDV = end-diastolic volume. Data on the same study population have been previously published.9
size parallel imaging (SENSE) factor 2. Volumes were measured in SAX views using the ViewForum release 5.1 workstation (Philips). The MRI cardiac analysis software was used for semi-automatic calculation of RV systolic and diastolic volumes. SAX volumes were analyzed using 30 cardiac phases. For each short-axis slice, the inner contour of the right heart was manually drawn in both diastole and systole. The position of the most basal axial slice at the tricuspid valve was crosschecked with the four-chamber view. After confirmation of the drawings, the software automatically calculated the end-systolic volume (ESV), end-diastolic volume (EDV), SV, and EF of the RV. The images were transferred to a PACS system (Impax DS3000 release 4.5, Agfa-Gevaert,
Right Ventricular Function after CABG
Mortsel, Belgium) for assessment of systolic and diastolic RV lateral length, defined as the distance between the insertion of the tricuspid plane in the free lateral wall and the RV apex in the fourchamber view. The amount of left ventricular late gadolinium enhancement (LGE) was measured pre- and postoperatively and expressed as percentage of LV mass as described in detail in a previous publication.9 Echocardiography: Data Acquisition: All echocardiographic studies were performed using an iE33 scanner (S5-1 probe, Philips Medical Systems, Andover, MA) and a 1–5 MHz transducer with the patient in the left lateral decubital position. Conventional 2D gray scale images and tissue Doppler imaging (TDI) data were obtained in the apical two-, three-, and four-chamber views. Additional fourchamber views were optimized for the best quality acquisition of the RV and by searching for the highest radial RV diameter. TDI loops of five consecutive cardiac cycles were acquired, wall-by-wall, of the anteroseptal, the inferoseptal, and the free RV wall in two standard apical views, using a sector angle of 25–30° and frame rate of 120–170 s1. Care was taken to align the walls as parallel as possible to the scanline. The cine loops were obtained in proprietary DICOM format and digitally stored for offline analysis at the original number of frame rates. Echo Image Analysis: The readers of both echocardiographic and MR images were blinded for mutual test results. All 2D and TDI measurements were taken by a single, experienced observer using commercially available software (QLAB, Philips Medical Systems). Andover, MA). The TAPSE was measured using M-mode images of the RV lateral free wall. The RV systolic and diastolic areas were measured in an apical fourchamber view visually optimized for imaging of the RV. In the same projection, we measured the tricuspid ring diameter, apical end-diastolic and end-systolic angles, and maximal radial RV diameters. The apical area and diameter were defined as the area distal to and the diameter at the apical third of the central long-axis length (from apex to the center of the tricuspid plane), respectively. Septal motion was assessed in a parasternal M-mode as described in Figure S1. Global LV function was assessed by calculating ESV, EDV, and EF using the biplane Simpson’s method. Stroke volume, cardiac output (CO), and cardiac index (CI) were derived from the calculated volumes and the heart rate. Right atrial areas (RAA) and diameters were measured at end-systole in the four-chamber view.
Systolic pulmonary artery pressure (PAPsys) was estimated by summation of the tricuspid valve pressure gradient and the atrial pressure derived from the change in the inferior vena cava diameter during respiration.10 All Doppler measurements describing RV function were performed at end-expiration. From pulsed-wave Doppler, placed on the tip of the tricuspid valve, peak E velocity, peak A velocity, and E deceleration time were measured.11 In the hepatic veins, the systolic and diastolic velocity time integral (VTI), the systolic filling fraction (SFF) (VTIsys/ VTIsys + VTIdia), and the regurgitation time between systolic and diastolic flow were derived.12 The corresponding measurements were performed in the superior vena cava and the pulmonary veins. E/e0 was calculated from the pulsed-wave Doppler of the tricuspid blood flow and the pulsed-wave tissue Doppler of the lateral free RV wall.13 Regional septal function was evaluated in the five septal segments from the standard 16 segment model defined by the American Society of Echocardiography.14 Regional RV lateral wall function was assessed in one apical and one basal segment.15 TDI strain analyses were performed using commercially available software (QLAB). The process of TDI analyses and the criteria for discarding measurements have previously been described in detail.9 The aortic valve opening and closure, and thereby the ejection time (ET), were determined by pulsed-wave Doppler of the aortic valve. ET strain was defined as the highest positive or negative peak strain value during ET. In a cine loop of five cardiac cycles, the first cycle was analyzed except in the case of premature beats or insufficient quality of the first beat. Then, the second or one of the following cycles was analyzed. A strain value of 10% was chosen as the cutoff for normal strain values of the left and right ventricle, based on the findings in a large population study.16 Furthermore, the RV free wall was reanalyzed by 2D strain in the same patients (syngo Velocity Vector Imaging, Siemens, Medical Solutions, Mountain View, CA, USA). There we measured mid-myocardial strain in the RV lateral wall only. When comparing TDI and 2D strain, only apical and basal strains are reported. Reproducibility: To determine inter-observer variability for TDI strain, 15 patients were randomly selected and another independent, experienced observer (SM) blinded to all other data, analyzed the preoperative data. To assess intra-observer variability, data from the same 15 patients were reanalyzed by the main observer (AR) in a new 1811
€ sner et al. Ro
random order after at least 6 months. Hence, for RV inter- and intra-observer variability, 30 RV segments were analyzed. Inter-observer and intra-observer variabilities for strain were expressed as the absolute difference between two measurements in percent of their mean. Reproducibility was also tested in comparison with 2D strain, performed on the same dataset on the free RV lateral wall. Statistical Analyses: All statistical analyses were performed using SPSS version 16.0 (SPSS Inc., Chicago, IL, USA). Unless stated otherwise, all data are expressed as mean standard deviation (SD). Paired Student’s ttests were used to test changes in parameters before and after CABG. Dependency of RV strain on clinical data was tested by linear regression for continuous or ANOVA for discrete independent variables. Inter-observer and intra-observer values are presented using Bland–Altman plots. Results: Clinical Outcomes: Following CABG, four patients suffered from angina pectoris and two from dyspnea (Canadian Cardiovascular Society Angina Classification Class [CCACC] 1–2), three developed chronic atrial fibrillation, and one patient had a positive dobutamine stress test without coronary stenoses on coronary angiography. All other patients were free of symptoms (CCACC = 0). None of the patients had more than grade I tricuspid regurgitation or pulmonary hypertension (estimated systolic pulmonary artery pressure >40 mmHg) before or after CABG. Four patients had a perioperative myocardial infarction with CK MB at 130, 101, 87, and 59 U/L, respectively. The total amount of LGE/myocardial mass was unaltered in these four patients as well as in the total study population (from 3.6 to 4.4% (P = 0.557). None of the patients showed ST elevations in the ECG. The degree of ET strain depression after CABG did not correlate with RV EF, RV EDV, LV EF, or clinical data such as age, weight, BSA, hypertension, hypercholesterolemia, diabetes, positive family history for CAD, or smoking. Volumes, Geometry, and Filling Pressures: The geometric and global functional parameters and changes after CABG, measured by MRI and echocardiography, are displayed in Table II. Based on MRI measurements, the RV volumes and EF did not change significantly after surgery. However, there was a significant increase in SV. Postoperatively, MRI revealed a shortened RV lateral longitudinal wall length in diastole 1812
and systole, while radial and anterior–posterior mid-ventricular diameters were not significantly changed. This was in accordance with the echocardiographic unchanged RV FAC as well as changed dimensions shown in Table II, where post-CABG longitudinal diameters were markedly reduced with higher apical and mid-ventricular diastolic and systolic RV diameters. The difference between systolic and diastolic apical diameters increased as well as the systolic apical angle reduction. Figures 1 and 2 show examples of echocardiographic images and MRI with geometric measurements of the RV before and after CABG. Right ventricular geometric measurements reveal a significant postoperative increase of the mid- and apical diastolic diameters, the apical angle, and area, as well as a larger global diastolic area. RA diameters and areas were also increased after CABG. A shortened deceleration time of the tricuspid E wave and a decreased systolic filling fraction in hepatic veins and superior vena cava were demonstrated, as shown in Figure 3. Both increased RA dimensions and decreased hepatic vein SFF indicate increased RV end-diastolic and RA pressures. However, post-CABG vena cava dimensions and estimated RA pressures by respiratory changes remained unchanged. Global and Segmental RV Wall Deformation: Figure 4 displays examples of pre- and postCABG TAPSE and tissue velocities. Figure 5 depicts an example of typical pre- and postCABG strain curves. Figure 6 displays apical and basal strain in the presence or absence of right coronary artery (RCA) disease. TAPSE and ET strain reduction after CABG was independent of the presence of RCA disease and complete revascularization of a significant RCA stenosis. Table III indicates that post-CABG TAPSE was markedly reduced as well as other parameters expressing global systolic wall deformation, that is, ET strain of the free lateral RV wall and septum and regional deformation in the RV free wall, that is, basal and apical RV ET strain, despite normal or slightly reduced preoperative strain. RV tissue velocity (TV) (E) was reduced indicating reduced diastolic wall motion. Figure 7 and Table III show a higher variance (i.e. SD) of pre-CABG compared to post-CABG ET strain. Only segments with high values reduced their strains, whereas most segments with moderate strains stayed unchanged and low strains tended to increase. This effect was similarly demonstrated both measured by TDI or 2D strain. In general, only 10 of 89 preoperative segments were dysfunctional (ET strain > 10%) despite a predominant presence of RCA disease
Right Ventricular Function after CABG
TABLE II Clinical, Geometrical, and Doppler Changes between Pre- and Post-CABG in the Right and Left Ventricle Pre-CABG
BP sys (mmHg) BP dia (mmHg) MRI: RV volumedia (ml) MRI: RV volumesys (ml) MRI: RV EF (%) MRI: RV SV (ml) MRI: RV diameter longdia (cm) MRI: RV diameter longsys (cm) MRI: RV diameter radialdia (cm) MRI: RV diameter radialsys (cm) MRI: RV diameter ant-postdia (cm) MRI: RV diameter ant-postsys (cm) US: HR (1/min) US: RV areadia (cm2) US: RV areasys (cm2) US: RV FAC (%) US: Apical angledia (°) US: Apical angle reduction (°) US: Apical RV diameterdia (cm) US: Apical diameter reduction (cm) US: Apical areadia (cm2) US: Apical FAC (%) US: Mid-RV diameterdia (cm) US: Mid-RV diametersys (cm) US: RA area (cm2) US: Septal motion LV (mm) US: Septal motion RV (mm) US: LV volumedia (ml) US: LV volumesys (ml) US: LV EF (%) US: CO (ml) US: SV (ml) US: CI (ml/m2) US: TV: PAPsys (mmHg) US: TV: E deceleration time (ms) US: HV: Systolic filling fraction (%) US: HV: Regurgitation time (ms) US: SCV: Systolic filling fraction (%) US: SCV: Regurgitation time (ms) US: TR: E/E0 RV Wall US: TV (E`) (cm/s)
Post-CABG
Mean
SD
Mean
SD
n
P
135 80 130 80 40 50 7.0 5.3 3.7 2.9 6.7 6.0 70 20.2 12.5 38 48 7 2.5 0.4 4.5 13.0 3.7 2.8 16.2 8.6 5.2 138 69 52 4823 69 2367 28 235 58.9 32 62 47 6.3 8.7
20 12 32 29 9 12 1.2 1.1 0.8 0.8 1.0 1.0 13 4.6 3.5 15 13 11 0.6 0.5 1.4 29.8 0.5 0.6 3.8 2.8 2.9 40 31 10 1220 17 658 8 86 8.8 50 10 59 2.4 3.1
137 80 134 78 42 56 6.1 5.0 3.5 2.9 6.8 6.0 68 22.0 13.9 37 53 13 2.7 0.6 5.2 22.8 4.0 3.2 18.4 2.8 0.1 130 66 50 4504 66 2172 27 199 48.4 48 52 76 9.4 6.1
20 11 32 24 8 15 0.8 0.8 0.6 0.5 0.7 0.7 12 4.6 3.9 12 10 9 0.5 0.4 1.2 21.1 0.7 0.6 3.9 2.9 2.2 32 25 9 1477 20 636 8 44 8.5 65 11 63 3.4 1.7
57 57 54 54 54 54 54 54 54 54 54 54 57 57 57 57 55 55 55 55 55 55 55 55 57 54 53 49 49 50 50 50 50 32 50 41 42 49 47 50 57
n.s. n.s. n.s. n.s. n.s. 0.008 0.002 n.s. n.s. n.s. n.s. n.s. n.s. 0.010 0.009 0.05 0.018 0.005 0.026 0.024 0.008 n.s. 0.002
