Pulmonary Regurgitation after Tetralogy of Fallot Repair: A Diagnostic and Therapeutic Challenge.

BACKGROUND: Pulmonary regurgitation is the key hemodynamically significant lesion in repaired tetralogy of Fallot contributing to progressive right ventricular (RV) dilatation and biventricular dysfunction. The timing for pulmonary valve replacement remains a controversial topic, and the decision to intervene depends on assessment of RV size and RV function.
OBJECTIVES: This review aims to discuss the echocardiographic techniques that can be used to assess patients with pulmonary regurgitation after the repair of tetralogy of Fallot defect. While cardiac magnetic resonance (CMR) imaging is the clinical reference method, there is an important role of echocardiography in identifying patients with significant pulmonary regurgitation and assessing the RV size and function. The different echocardiographic techniques that can be used in this context are discussed. Newer techniques for assessing RV size and function include three-dimensional (3D) echocardiography, tissue Doppler and strain imaging. 3D RV volumetric reconstruction based on two-dimensional imaging is a promising new technique that could potentially replace CMR for RV volumetric assessment.
CONCLUSIONS: Developments in echocardiographic techniques provide new insights into the impact of pulmonary regurgitation on RV structure and function. Echocardiography and CMR are complementary modalities and further research is required to define the optimal use of both techniques for this indication.

INTRODUCTION

Surgical repair of tetralogy of Fallot (TOF) often results in hemodynamically significant pulmonary regurgitation. This has been associated with right ventricular (RV) dilatation, biventricular dysfunction and arrhythmias.[1] While echocardiography is the first line non-invasive imaging technique, it has limitations in quantifying pulmonary regurgitation and RV size and function. Due to the complex nature of RV geometric shape, quantitative assessment via echocardiography remains challenging and even more so in adult patients with repaired congenital heart disease.[234] While two-dimensional (2D) echocardiography is widely used for screening and follow up of right ventricular dilatation, it cannot accurately define right ventricular volume and global function. Cardiovascular magnetic resonance (CMR) imaging has become the standard in the long term assessment of patients with TOF, and is the clinical reference method of quantifying right ventricular volumes and function.[56] Based on CMR measurements, RV dilatation and dysfunction have been correlated with adverse outcomes including death, sustained ventricular tachycardia and decreased functional status.[7] It has been demonstrated that reverse RV remodeling is less likely once the RV is overly dilated. Threshold values of RV end diastolic volume (RVEDV) of 150-180 ml/m2 have been proposedas an indication for pulmonary valve replacement.[89] The benefits of an appropriately timed pulmonary valve replacement on RV remodeling and recovery of function have also been recognized;[89101112] however, the decision to intervene will depends on many factors including RV end-systolic and end-diastolic volumes, RV ejection fraction (EF), residual RV outflow tract obstruction, tricuspid regurgitation and LV systolic function.[13] We will review the most recent evolutions in echocardiography, which could be helpful in the echocardiographic assessment of patients after TOF repair.

Echocardiographic assessment of the right ventricular outflow tract and pulmonary regurgitation

One of the important aspects in the evaluation of postoperative TOF patients is the echocardiographic assessment of the right ventricular outflow tract, which, apart from the VSD closure, is one of the main aspects of the surgical repair. Residual outflow tract obstruction needs to be excluded, as significant residual narrowing is a risk factor for poor long-term outcome.[14] By combining 2-D imaging with color Doppler, pulsed wave and continuous wave Doppler, it is generally possible to quantify the severity of the obstruction and identify the most important anatomic component (subvalvar, valvar or supravalvar). The intensity and location of the Doppler gradients should take into account the degree of pulmonary regurgitation as well as the anatomical characteristics of the stenosis. Specifically, with severe pulmonary regurgitation, the gradient may be overestimated by the volume load. Also, in long-segment obstruction such as conduit narrowing, the Bernoulli equation cannot be applied and often results in gradient overestimation. Important information can be obtained from the measurement of peak tricuspid regurgitation jet to estimate RV pressure. Assessing the degree of pulmonary regurgitation is a second important component of the echocardiographic evaluation.

The quantification of pulmonary regurgitation (PR) requires the interpretation of different echocardiographic parameters combined in a semi-quantitative assessment. Color Doppler flow imaging of the RV outflow tract and the central pulmonary arteries is one of the most commonly used techniques. As illustrated in Figure 1, this is generally obtained from a low parasternal short axis view imaging the RV outflow tract, the main pulmonary artery and the pulmonary artery branches. Depending of the origin of the diastolic backflow, a qualitative assessment of the degree of PR can be made. Diastolic flow reversal in mild PR originates from the main pulmonary artery (PA), moderate PR from the PA bifurcation and severe PR from the distal pulmonary artery branches. However, this can be influenced by distal PA branch stenosis. Renella et al., showed that diastolic flow reversal in the branch pulmonary arteries was a reliable marker of severe PR.[15] Apart from the origin of the flow reversal, the width of the color jet can also be measured at the level of the pulmonary valve. While multiple different views may be used, the parasternal outflow tract view provides the best view of the jet width that also allows measurement of the pulmonary annulus or outflow tract if the annulus cannot be well identified. The width of the PR jet can be compared relative to the RV outflow tract (RVOT). PR can be classified as mild (jet width to RVOT ratio < 0.2), moderate (ratio between 0.2-0.5) or severe (ratio > 0.5).[1617] In patients after repair of TOF the delineation of the pulmonary annulus is often difficult and the outflow tract shape is commonly non circular due to geometric modification related to the patch repair. Additionally there may be multiple jets of regurgitation, which further complicates the use of this technique. Quite often in patients with transannular patch repair there is free pulmonary regurgitation.

Figure 1

Color Doppler of the right ventricular outflow tract and branch pulmonary arteries (PAs) in parasternal short axis. Note the retrograde flow originating from the distal branch PAs

Vena contracta, the narrowest area of blood flow exiting the valve is often utilized in other regurgitant lesions, but it has not yet been validated in pulmonary regurgitation.[18] Vena contracta is proposed as a simple, easy to measure parameter that correlates well with effective regurgitant orifice area and regurgitant volume.[19] However, in repaired TOF, the shape of the vena contracta is complex, further complicating the quantification of PR. Development of three-dimensional (3D) Doppler flow imaging may allow more accurate assessment of complex jet regurgitation.[20] Regurgitant jets can also be brief and challenging to measure in patients with severe PR and or restrictive RV physiology, due to early equalization of pulmonary artery and right ventricular pressures during diastole.

The proximal isovelocity surface area (PISA) is an additional color flow Doppler based measurement that employs the flow convergence principle.[21] PISA is equal to the flow rate at the regurgitant orifice, and can be used to calculate the maximal effective regurgitant orifice area and the regurgitant volume. It has the advantages of being independent of hemodynamic factors including concomitant valvular disease, and accommodating eccentric color jets, albeit with somewhat decreased accuracy. While promising, limitations of PISA such as required angle correction and variation in orifice during cardiac cycle, have been extensively noted.[22] Also, in the setting of severe or free regurgitation, often seen in TOF patients, the PISA method is not applicable.

Continuous wave (CW) Doppler traces of the RVOT flow can also be helpful in the assessment of the degree of PR. More severe PR will result in faster equalization of pressures between the PA and the RV in diastole, resulting in a shorter pressure half-time and progressive shortening of the backflow, with prolongation of the no flow time in diastole as demonstrated in Figure 2. A pressure half time (< 100 milliseconds) and a no-flow time (> 80 milliseconds) have been shown as the useful parameters for identifying severe pulmonary regurgitation.[2324] This is however also limited by its dependence on RV diastolic properties. In a restrictive RV, a short pressure half-time will reflect an increase in RV end-diastolic pressures. Restrictive RV physiology in adult patients has been shown to be associated with less PR as quantified by CMR.[25] This must be considered when interpreting these Doppler timings. As each technique has significant shortcomings, we stress the importance of integrating the use of these different parameters. The semi-quantitative assessment will generally allow identification of patients with severe pulmonary regurgitation who may benefit from further evaluation.

Figure 2

Continuous Doppler through the right ventricular outflow tract showing a mild degree of flow acceleration with a peak gradient of 25mmHg and severe pulmonary regurgitation with a pressure half time of 83 msec and a long end diastolic no-flow time

Other echocardiographic techniques for assessing the significance of PR have been explored. Festa et al., introduced the use of a Pulmonary Regurgitation Index by M-mode echocardiography (PRIME), which reflects the systolic to diastolic variation in the pulmonary artery diameter. They were able to match PR fraction by CMR to PRIME values with high sensitivity and specificity for mild (<15%), moderate (15-25%) and severe (>40%) pulmonary regurgitation.[20] They also observed that in valves with greater degrees of regurgitation, PRIME was better correlated with CMR regurgitant fraction than PR index, vena contracta, PR deceleration time and pressure half time. As this is most certainly dependent on pulmonary artery compliance and stiffness, and further influenced by distal branch stenosis, central PA size and other physiologic factors, which vary after TOF repair, further evaluation is necessary. Color M-Mode can further aid in showing direction and timing of flow events by superimposing color encoded velocity shift on conventional M-mode illustrating time dependency of PR during the cardiac cycle.

In patients where the severity of pulmonary regurgitation remains in doubt, quantification by CMR is considered the clinical reference technique.[13] The regurgitation fraction can be quantified by phase contrast imaging and is calculated as retrograde flow volume divided by antegrade flow volume in the proximal main pulmonary artery. This is depicted graphically in Figure 3. Assuming competence of other valves and lack of residual shunt lesions, the difference between right and left ventricular stroke volume can also be attributed to pulmonary regurgitation in this patient population. Calculation of pulmonary regurgitation fraction should be internally consistent between the two alternative methods.[2627] In addition to the pulmonary regurgitation fraction, Wald et al., recently suggested that pulmonary regurgitation volume more accurately reflects the RV preload and thus the physiologic impact on the RV.[28] We suggest measuring both parameters. Further delineation of regurgitation in the branch pulmonary arteries was performed by Kang et al.[29] Their phase contrast cine MR comparison of regurgitation showed greater regurgitation fractions in the left pulmonary artery compared to the right, consistent with Figure 3.

Figure 3

Pulmonary regurgitant flow assessed by MRI in left and right pulmonary arteries. Notice the more important retrograde flow originating from the left pulmonary artery (left side)

Assessment of right ventricular dilatation and right ventricular volumes

Recent guidelines have been published on cross-sectional RV measurements in both children and adults[3031] Measurements of RV size from the apical 4 chamber view and the right ventricular outflow tract in the parasternal short and long-axis axis views can routinely be performed in the assessment of patients with right heart disease.[31] Lack of normal reference values and z-scores for all of these parameters is a current limitation. Normal data are available for the short-axis and long-axis RV dimensions as well as the tricuspid and pulmonary annulus.[303132] Further well-standardized data collection in large patient groups will be required to establish comprehensive normal values. The correlation of these measurements with CMR determined dimensions remains variable. Lai et al., compared 2D echocardiographic measurements with corresponding CMR dimensions and observed that the echocardiographic measurements generally underestimated the CMR measurements. The 2D measurements correlated, but the agreement between the measurements weakened with progressive RV dilatation.[333435]

Beyond simple 2D methods, different formulae have been proposed based on distinctive RV geometrical models to calculate RV volumes based on two-dimensional measurements. Helbing et al., showed that these methods generally do not agree well with the volumes measured by CMR, likely related to inaccuracies in both the measurements and models.[34] Our group recently proposed a relatively simple method based on measuring the RV -end-diastolic area indexed for body surface area (RVEDAi), to identify patients with a dilated RV who could benefit from further imaging. We showed that an RVEDAi <20 cm2/m2 was highly predictive of a RV end-diastolic volume index (EDVi) <170 ml/m2.[36] Greutmann et al., used a similar approach in adult patients after TOF repair. They first demonstrated that when comparing different RV two-dimensional measurements, RVEDAi correlated best with CMR-based RV volumes. Based on a strong linear correlation, they proposed a simple regression equation that could be used to estimate RV volumes based on RV end-diastolic area measurements (RVEDVi = 11.5 + ( 7 Χ RVEDAi)).[37] We showed however that this regression analysis did not work well in a pediatric population.[38]

Three-dimensional Echocardiography (3DE)

In recent years Three-dimensional Echocardiography (3DE) has developed as a useful technique for quantification of LV volumes with nearly fully automated algorithms that have been developed for measuring LV volumes. RV volumetric assessment is more challenging due to the more complex 3-D shape and the poor endocardial definition related to the extensive RV trabeculations. Currently semi-automated methods have been developed and validated and are available for clinical use as illustrated in Figure 4. One of the problems is the feasibility of the method especially in post-operative patients with poor echocardiographic windows. 3DE quality depends largely on 2D image quality. Good spatial and temporal resolution is absolutely necessary as RV endocardial border identification is generally more difficult due to the extensive trabeculations. RV anterior and lateral walls can be especially difficult to image when the RV is significantly dilated. Some studies reported a 3DE feasibility of only 55% of postoperative TOF patients[39] while other studies seemed to have fewer problems with analysis of the obtained 3DE datasets.[4041] Feasibility limitations aside, the accuracy of the 3D volumetric calculation method is also problematic. Most studies report a significant bias when comparing echocardiographic 3D volumes with CMR-derived volumes with an underestimation of the CMR values by the echocardiographic measurements, ranging between 10-25%. This underestimation is more important when the RV is more dilated, which limits the clinical applicability of this technique in patients with progressive RV dilatation.[35394142]

Figure 4

Right ventricular volume by three dimensional echocardiography. (a) Right ventricular end diastolic volume. (b) Projection of the volume in apical four chamber view. (c and d) projections of the reconstructed volume at two different levels in short axis

To overcome this problem our group has recently further explored an alternative method to quantify RV volumes using 2D-based 3D reconstruction illustrated in Figure 5.[43] This novel 3D reconstruction method uses standard 2D echocardiographic images that are localized in 3D space using a magnetic tracking device. The 2D images together with the 3D spatial information is analyzed using a software analysis tool that applies a database of RV shapes and volumes to the data.[3843] Ten to fifteen 2D echocardiographic views are obtained in short clips at end expiration. On these views certain RV landmarks are identified including the tricuspid and pulmonary valve annulus, the subvalvar pulmonary and tricuspid regions, the RV side of the interventricular septum, the RV free wall and the RV apex. Based on the spatial localization of the anatomic landmarks in 3D space, a full reconstruction of the RV volumes is performed based on a database of RV shapes obtained by CMR in patients after TOF repair. Our group demonstrated a high feasibility, excellent reproducibility and good accuracy of this method.[43] We noted a very small underestimation in RV volumes versus CMR volumes (2.5% EDV and 3.6% ESV) but these small differences are probably clinically irrelevant also taking the variability of CMR measurements into account. Disadvantages of this include the additional hardware and software and scan time required. Acquisition times are approximately five minutes with an additional fifteen minutes for data analysis. The accuracy and reproducibility of knowledge based reconstruction was analyzed by Sheehan et al. with favorable consistency with CMR findings.[44] This technique has the potential to be useful in a wide range of heart disease where image quality can be a problem.

Figure 5

Right ventricular end diastolic volume using the 2D based 3D reconstruction method. The different point represent landmarks of ventricular anatomy, i.e. pink represents pulmonary valve, purple — the tricuspid valve, blue — the interventricular septum, red — the free wall endocardium and yellow the right ventricular apex. The yellow lines are the planes of the 2D images projected in the reconstructed volume

Exciting advances in 3D reconstruction provides us the tools to increase our understanding of the RV remodeling process.[45] Data has shown that the outflow tract aside, remodeling occurs in the basal part of the RV where a ‘basal bulge’ develops and in the apical part which becomes more dilated and rounded compared to a normal RV apex. The progression of remodeling over time has not well been studied and may provide remarkable information on pathologic forms of remodeling occurring in a subgroup of patients.

Echocardiographic methods to assess right ventricular function in post-operative TOF patients

Apart from RV size and volume, the effect of pulmonary regurgitation on RV function needs to be studied. Different methods are utilized and these were recently reviewed by our group.[46]

2-D based methods

As RVEDA seems to correlate best with RV end-diastolic volume, it seems logical to measure the RV fractional area change (FAC) from the apical four-chamber view as a method to assess RV systolic function.[31] This method has been shown to correlate well with RV EF although in TOF patients the correlation seems less consistent. This is probably related to the fact that this method does not include the right ventricular outflow tract, which can be significantly dilated and dysfunctional and therefore responsible for reduction in RV EF in some post-operative TOF patients.[47] Therefore, any method that does not include the RVOT can potentially overestimate global RV pump function. Another problem with FAC measurements is that it can be exceedingly difficult to trace the RV especially in the more apical parts contributing to a relatively higher variability. In particular, as the trabeculations contract in systole, the RV base can become very difficult to trace.

A second relatively easy method is the use of the tricuspid annular motion as represented by the tricuspid annular systolic planar excursion (TAPSE), which can be obtained from an RV-centric view from the apical position.[31] As shown by Figure 6, we use color M-mode in our laboratory to define the timing of events, which increases the reproducibility of the method. Recently normal pediatric values for TAPSE have been published. TAPSE has also been studied in patients followingTOF repair.[4849] In this patient group, TAPSE was shown to correlate only moderately with RV EF. This is not surprising since thisparameter exclusively studies the longitudinal function of the basal part of the RV lateral wall. This does not reflect global RV function, as regional dysfunction is often present in the outflow tract and possibly other RV segments.

Figure 6

Color Doppler M-mode at the tricuspid valve level to assess the tricuspid annular plane systolic excursion

Tissue Doppler imaging (TDI) can measure tricuspid annulus velocities which also reflect the longitudinal function of the inlet part of the RV. Toyono et al., showed that decreased myocardial acceleration during isovolumic contraction (IVA) correlated well with decreased RV function and worsening PR.[50] This proposed index allows assessment of RV function independent of ventricular shape or loading conditions. Kutty et al,[51] more recently showed that systolic tissue Doppler velocities at the base correlate well with RVEF measured by CMR if the RV outflow tract function, defined by regional outflow tract EF, is preserved. In patients with reduced outflow tract EF, the correlation between systolic tissue Doppler velocities and global EF was weak. Van Der Hulst et al., used tissue Doppler velocities to study the RV temporal activation sequence post-operative TOF patients. They showed that mechanical activation of the RV outlet was significantly decreased compared to normal controls. No abnormal time delay was observed in the inlet part of the RV. The outlet time delay correlated significantly with global longitudinal RV strain measurements.[52] The clinical implications of these findings require further study.

Myocardial deformation and strain imaging has also been used to study regional and global RV longitudinal function after TOF repair. For a recent overview of this topic we refer to Friedberg et al.[53] Initial studies using tissue Doppler derived strain measurements showed that regional right and left longitudinal function in this patient population is decreased[54] and that this decrease was more pronounced in patients with more significant pulmonary regurgitation.[55] The interpretation of strain results is complicated due to the physiologic effect of different parameters influencing strain measurements. Strain does not directly reflect intrinsic myocardial contractile force development and is influenced by geometry and loading. The increased RV output caused by pulmonary regurgitation should result in increased myocardial deformation while RV dilatation has the opposite effect on strain measurements. This complicates the interpretation of absolute strain values and possibly explains some of the effects of pulmonary valve replacement on myocardial strain measurements in postoperative TOF patients.[56] Immediately after surgical pulmonary valve replacement, a reduction in longitudinal RV strain was observed with a recovery in function to pre-operative values during further follow-up. Similar findings were obtained in a small group of patients following pulmonary valve replacement with decreased strain early after replacement without improvement in midterm follow-up.[57] Reduced stroke volume and reverse remodeling probably could partially explain these findings. In patients after TOF repair, longitudinal strain measurements as measured from the apical four-chamber view, were shown to correlate only weakly with RVEF.[58] This is probably related to measuring only the inlet and trabecular parts of the RV as well as the fact that radial and circumferential RV deformation is not taken into account. Scherptong et al., showed that strain values decreased more significantly compared to changes in RV volumes and RV EF in a relatively small cohort of adult TOF patients during longitudinal follow-up.[59] The clinical importance of this finding requires further study but may indicate that following longitudinal strain could detect early changes in RV function.

Strain imaging can also be used to study ventricular interactions in post-operative TOF patients. The dilated RV influences left ventricular (LV) function and a close relationship between LV and RV dysfunction has been demonstrated. The development of LV dysfunction is a poor prognostic sign in post-operative TOF patients and is associated with unfavorable clinical outcomes.[7] A recent study by Diller et al., showed that reduced LV longitudinal function as measured by LV strain, is associated with an increased risk for life-threatening arrhythmia and death.[60] This is the first study demonstrating the prognostic value of strain measurements in an adult congenital population.

CONCLUSION

Different echocardiographic techniques can be used in the assessment of post-operative TOF patients:

Color Doppler and CW Doppler can be combined to identify patients with severe pulmonary regurgitation.

RV z-scores and RV EDA can be used to identify patients with significant RV dilatation.

2D-based 3D reconstruction currently is the most reliable accurate echocardiographic method for quantifying RV volumes, and is emerging as a potential alternative to CMR.

Fractional area change can be used for assessment of RV function. Tissue Doppler and TAPSE correlate only weakly with RV EF as they only asses the RV inlet. Strain imaging is emerging as a potentially useful clinical technique, but will require further optimization and prospective evaluation.

Selected patient identified using these echocardiographic techniques could further be evaluated by CMR to quantify the degree of PR, and the hemodynamic impact on RV volumes and global function.