Conventional indices of right ventricular (RV) function are known to be reduced after cardiac surgery, as a consequence of geometric rather than functional alterations. New techniques, such as three-dimensional (3D) transthoracic and two-dimensional speckle-tracking echocardiography, may be useful in postsurgical RV assessment. The aim of this study was to compare indices of RV function obtained using different echocardiographic modalities, before and after surgery.
Forty-two patients were screened the day before and 6 months after mitral valve repair. Twenty healthy patients were also enrolled as controls. Tricuspid annular plane systolic excursion and peak systolic velocity were calculated from Doppler tissue imaging. Longitudinal and radial strain values were obtained from speckle-tracking echocardiography. RV ejection fraction was calculated from 3D transthoracic echocardiographic RV volumes, and similarly, fractional area change was computed from RV areas.
Tricuspid annular plane systolic excursion (25 ± 4 vs 17 ± 3 mm), peak systolic velocity (17 ± 4 vs 12 ± 2 cm/sec), and fractional area change (43 ± 8% vs 39 ± 7%) significantly decreased after surgery ( P < .01), while 3D RV ejection fraction was preserved (59 ± 7% vs 59 ± 6%). Speckle-tracking echocardiographic results were dependent on the considered direction, with preserved radial but decreased longitudinal strain values. All postoperative two-dimensional longitudinal indices were smaller than in controls. Preoperative parameters were not significantly correlated with RV functional changes.
Although 3D ejection fraction was preserved after surgery, in agreement with the lack of evidence of RV dysfunction, two-dimensional indices showed a functional loss in the longitudinal direction. Fractional area change, as a combination of radial and longitudinal properties, was slightly decreased. Speckle-tracking echocardiography could constitute a useful approach to relate local and space-dependent changes to the global RV function.
The role of right ventricular (RV) function as a prognostic determinant in many cardiovascular diseases and its important prognostic value in the evaluation of the surgical outcome have been demonstrated in several studies.
In routine clinical practice, RV systolic function is commonly assessed by measuring tricuspid annular plane systolic excursion (TAPSE) and peak systolic velocity (PSV) using Doppler and two-dimensional (2D) transthoracic echocardiography (TTE). The clinical and prognostic usefulness of these parameters as well as their relation with RV ejection fraction (EF) are well established. The reduction of these indices after cardiac surgery, although not yet completely explained, is a well-known phenomenon whose correlation with RV systolic dysfunction is still controversial. As a consequence, the reliability of TAPSE and PSV as indices of postsurgical outcome may be limited.
New imaging techniques, such as real-time three-dimensional (3D) TTE and 2D tissue deformation analysis using speckle-tracking echocardiography (STE), have been shown useful in the evaluation of RV function. Accordingly, our aims were (1) to obtain, in a group of patients undergoing mitral valve (MV) repair, indices of RV function by Doppler tissue imaging (DTI), 2D and real-time 3D TTE, and 2D STE, both before and after surgery; (2) to compare the changes induced by surgery in such indices; and (3) to compare the presurgical and postsurgical values with control values obtained from an age-matched group of normal subjects.
Population and Study Protocol
We studied 60 consecutive patients (40 men; mean age, 61 ± 11 years) with severe MV regurgitation secondary to degenerative prolapse or flail and scheduled for MV repair at our medical institution. In accordance with recommendations, MV regurgitation was defined as severe when the effective regurgitant orifice area, estimated using the proximal isovelocity surface area technique, was ≥0.4 cm 2 . Exclusion criteria were the presence of (1) persistent atrial fibrillation, (2) inadequate echocardiographic acoustic apical window, (3) concomitant severe tricuspid regurgitation, and (4) history of cardiac surgery or revascularization.
In addition, a group of 20 normal volunteers (11 men; mean age, 54 ± 9 years) was studied as control group to obtain values for comparison.
All patients underwent complete Doppler, 2D, and 3D transthoracic echocardiographic examinations the day before surgery; similar evaluations were performed 6 months after surgery in all patients not excluded from the protocol. Control subjects underwent a similar but single examination. All participants provided written informed consent.
All echocardiographic examinations were performed using a Philips iE33 ultrasound system (Philips Medical Systems, Andover, MA).
Doppler and 2D TTE
Comprehensive M-mode and 2D TTE was performed according to the clinical laboratory practice using an S5-1 sector array probe. TAPSE, defined as the displacement of the RV base during the systolic phase in the apical four-chamber view, was estimated by positioning the M-mode cursor at the junction of the tricuspid valvular plane with the RV free wall.
Also, DTI analysis was performed in the four-chamber view to assess PSV: the pulsed Doppler sample volume was positioned on the tricuspid annulus at the hinge point of the tricuspid anterior leaflet. Care was taken to obtain an ultrasound beam parallel to the direction of the tricuspid annular motion toward the apex. To allow the detection of low-intensity velocities, settings were modified in accordance with previously reported acquisition protocols. The low Doppler shift frequencies produced by the moving wall were recorded and the PSV was measured.
Systolic pulmonary arterial pressure was noninvasively obtained using the Doppler echocardiographic method from the systolic right atrioventricular gradient, calculated from the systolic transtricuspid regurgitant flow peak velocity using the modified Bernoulli equation. Right atrial pressure was derived by means of the inferior vena cava collapsibility index measured from the subcostal view.
Real-Time 3D Echocardiography
Real-time 3D TTE was performed immediately after the 2D examination, with the same ultrasound unit, using an X3-1 matrix array probe. The 3D data sets were acquired from the apical view, adapted to improve the visualization of the RV or left ventricular (LV) chamber. Acquisitions were obtained in the full-volume mode, in which R wave–triggered wedge-shaped subvolumes were obtained over four to seven consecutive cardiac cycles during a single breath hold at a mean frame rate of 26 ± 6 frames/sec (range, 16–37 frames/sec).
At least two data sets per cavity per patient were obtained, stored, and processed offline. Commercial software packages (4D-LV Analysis and 4D-RV Function; TomTec Imaging Systems, Munich, Germany) were used to obtain both LV and RV end-diastolic and end-systolic volumes. Briefly, after manual tracing of the endocardial borders on end-diastolic and end-systolic frames in three different views, the software automatically generated the RV or LV surfaces throughout the cardiac cycle. If needed, the operator can manually adjust the detected surfaces before quantification.
RV EF was calculated as the percentage change between end-diastolic and end-systolic volumes.
Speckle-Tracking Strain Analysis
The 2D echocardiographic images were analyzed using wall motion tracking software (2D CPA; TomTec Imaging Systems). All measurements were performed by an investigator experienced in the interpretation of echocardiographic images, blinded to the results of the 2D and 3D transthoracic echocardiographic measurements. From each echocardiographic data set, the best apical four-chamber view, including the whole RV cavity and with a frame rate >50 Hz, was selected for analysis.
After manual initialization of the RV end-diastolic endocardial border, the endocardial contour was subdivided into 47 equidistant control points, which were tracked automatically frame by frame throughout the cardiac cycle. The endocardial contour was manually adjusted when necessary, to optimize the boundary position and tracking. As a result, for the whole RV wall and for each control point, curves of radial and longitudinal displacement, strain, and strain rate were displayed ( Figure 1 , top). Next, the quality of tracking was visually judged as adequate or inadequate on the basis of the presence of artifacts on the obtained curves.
Then, the strain parameters were considered separately for the free and the septal walls. For each wall, the regional strain curves were computed by averaging the values obtained in the relevant control points. Finally, the signed minimum and the maximum values of the averaged curve for the longitudinal and the radial strain, respectively, were used as indices of RV systolic function ( Figure 1 , bottom).
Additionally, on the basis of the tracked 2D RV contours, fractional area change (FAC) was defined as the percentage change between the RV end-diastolic and end-systolic areas.
Normal distribution of continuous variables was assessed using the Kolmogorov-Smirnov test. Continuous variables are expressed as mean ± SD or as medians and interquartile ranges (IQRs) when not normally distributed.
Changes between presurgical and postsurgical measurements were tested using paired Student’s t tests or Wilcoxon tests, while differences between the parameters and the reference values from the control group were investigated using unpaired Student’s t tests or Mann-Whitney tests as appropriate. The sample size was calculated to achieve 90% power with a significance of type 1 error of α = 0.05. Correlation between clinical parameters and the changes in RV function (the difference between preoperative and postoperative values of each parameter) was investigated using Pearson’s or Spearman’s coefficient of correlation.
To quantify the repeatability of 2D speckle-tracking echocardiographic and 3D measurements, the analysis was also performed by a second independent observer in a randomly chosen subgroup of 15 data sets, and the first observer repeated the analysis in the same data subgroup ≥2 weeks later, blinded to previous results. Interobserver and intraobserver variability are reported in terms of intraclass correlation coefficients (ICCs). Test-retest reproducibility was evaluated in a randomly chosen subgroup of 15 patients. A second data set of the same patient, acquired during the same examination and under the same conditions, was retrieved from the picture archiving and communication system of our institution and analyzed by the same operator ≥2 weeks after the first analysis, to obtain the ICC between the two measurements.
SPSS software (SPSS, Inc., Chicago, IL) was used for statistical analysis.
Of the 60 enrolled patients, eight were excluded from the protocol and did not undergo the follow-up examination because of concomitant tricuspid annuloplasty (five patients) or minimally invasive mitral surgery (three patients). Of the remaining 52 patients reevaluated during follow-up, 10 were excluded from the study: in three cases, poor acoustic windows prevented any further analysis, while speckle-tracking echocardiographic analysis and 3D RV analysis were not feasible in three and four distinct patients, respectively. As a result, 42 patients were included in the study.
Global feasibility of the study was 52 of 60 (87%), 46 of 52 for STE (88%), and 45 of 52 (87%) for the evaluation of RV volume and function on the basis of 3D transthoracic echocardiographic data sets. The average times required for 3D RV and speckle-tracking echocardiographic analyses were 4 and 2 min, respectively, including manual adjustment.
Preoperative clinical characteristics are shown in Table 1 . All patients were in New York Heart Association class I or II and were referred to MV repair at an early stage, even when no symptoms or signs of LV dysfunction were evident.
|BSA (m 2 )||1.78 ± 0.18|
|Tricuspid diameter/BSA (mm/m 2 )||16.8 ± 2.1|
|Right atrial area (cm 2 )||16.1 ± 3.6|
|ACE inhibitors/angiotensin II receptor antagonists||16 (38%)|
|Ca 2+ antagonist||4 (10%)|
None of the patients had intraoperative complications, and valvular repair was successful in all cases, with trivial or mild residual regurgitation either immediately after the procedure or during follow-up. The mean cardiopulmonary bypass time was 121 ± 37 min, while cross-clamp time averaged 99 ± 34 min. No significant modifications in tricuspid regurgitation severity were observed after surgery. Postoperative right atrial pressure was in the range 6 to 9 mm Hg.
Table 2 lists the results of the presurgical and postsurgical measurements, as well as the reference values obtained from controls.
|Parameter||Control||Before MV repair||After MV repair|
|TAPSE (mm)||25.8 ± 4.4||25.2 ± 4.1||16.6 ± 3.0 ∗ †|
|PSV (cm/sec)||16.2 ± 2.3||17.2 ± 3.8||12.0 ± 2.4 ∗ †|
|PAPs (mm Hg)||26.4 ± 4.4||35.7 ± 4.7 †||26.1 ± 4.6 ∗|
|RV EDV/BSA (mL/m 2 )||44.8 ± 10.0||51.3 ± 10.8 †||49.3 ± 8.9|
|RV ESV/BSA (mL/m 2 )||15.4 ± 3.2||21.3 ± 6.5 †||20.0 ± 5.0 †|
|RV EF (%)||65.6 ± 5.7||59.4 ± 6.8 †||58.9 ± 5.9 †|
|RV FAC (%)||45.4 ± 10.2||42.7 ± 8.1||39.1 ± 6.9 †|
|LV EDV/BSA (mL/m 2 )||52.2 ± 8.4||77.4 ± 27.3 †||54.3 ± 11.8 ∗|
|LV ESV/BSA (mL/m 2 )||21.2 ± 4.1||30.6 ± 11.6 †||22.7 ± 7.5 ∗|
|LV EF (%)||60.3 ± 5.6||60.5 ± 6.3||59.1 ± 6.4|
M-Mode, 2D Transthoracic Echocardiographic, and Tissue Doppler Measurements
As expected, MV repair led to significant decreases in LV volumes, which were increased before surgery as a consequence of mitral regurgitation, while preserving LV EF, resulting in a normalization of LV volumes and function after 6 months.
Conversely, both TAPSE and PSV showed values similar to controls in the presurgical examination and significant reductions after surgery, leading to values lower than in controls. The mean reductions between preoperative and postoperative measurements for TAPSE and PSV were 8.6 mm (range, 4–18 mm) and 5.2 cm/sec (range, 1–12 cm/sec), respectively.
RV FAC showed a similar but less pronounced change: whereas before surgery, FAC was comparable with reference values, after MV repair, RV FAC was reduced by 4% on average, significantly lower than in controls.
Three-Dimensional Transthoracic Echocardiographic Measurements
Preoperative RV volumes were enlarged compared with controls but still in the normal range. After surgery, no significant changes in RV volumes were observed, with values still slightly greater than in controls. Similarly, RV EF, which was approximately 5% lower in patients than in controls before MV repair, did not show any significant difference in the preoperative versus postoperative comparison.
Two-Dimensional Speckle-Tracking Echocardiographic Strain Measurements
Results of speckle-tracking echocardiographic analysis ( Figure 2 ) showed that before surgery, patients with MV prolapse had similar strain values compared with the control group in both the free and septal walls, independently of the considered direction.
After surgery, STE showed different results when considering radial or longitudinal strain. Both in the free wall (before vs after surgery: median, 20.1% [IQR, 8.8% to 30.4%] vs 16.4% [IQR, 6.1% to 28.0%]; P = .25) and in the septum (median, 18.3% [IQR, 13.3% to 29.3%] vs 19.7% [IQR, 12.5% to 27.6%]; P = .38), radial strain was preserved after surgery and similar to controls (free wall vs septum: median, 25.2% [IQR, 17.8% to 30.0%] vs 19.6% [IQR, 11.2% to 27.4%]). Conversely, longitudinal strain was reduced both in the free wall (before vs after surgery: median, −26.5% [IQR, −21.1% to −32.0%] vs −18.0% [IQR, −14.7% to −20.9%]) and the septal wall (median, −20.8% [IQR, −17.5% to −22.3%] vs −14.1% [IQR, −11.4% to −16.6%]), compared with presurgical values. As a consequence, postsurgical longitudinal strain was smaller than the reference values both for the free wall (median, −26.5% [IQR, −23.3% to −30.1%]) and the septal wall (median, −20.4% [IQR, −17.4% to −22.3%]).
Correlates of RV Function
Table 3 shows the correlation between several clinical, preoperative and perioperative parameters, and the change in RV function measured using different echocardiographic methods, defined as the difference (Δ) between preoperative and postoperative values. Except for RV end-diastolic volume (positively correlated with ΔTAPSE and Δ radial strain), RV FAC (positively correlated with ΔTAPSE), and TAPSE (positively correlated with Δ radial strain), preoperative parameters were not significantly correlated with the change in RV function. Also, despite the statistically significant results, the correlations noted were rather weak.
|Parameter||ΔTAPSE||ΔPSV||ΔFAC||ΔEF||Δ radial strain||Δ longitudinal strain|
|Free wall||Septal wall||Free wall||Septal wall|
|RV FAC||.41 ∗||.13||.16||.05||−.03||−.22||−.03||−.16|
|RV EDV||.35 ∗||.45||.19||−.16||.42 ∗||−.20||−.24||−.10|