Article Text

Download PDFPDF

Original article
Aortic regurgitation severity after transcatheter aortic valve implantation is underestimated by echocardiography compared with MRI
  1. Stefan Orwat1,
  2. Gerhard-Paul Diller1,
  3. Gerrit Kaleschke1,
  4. Gregor Kerckhoff1,
  5. Aleksander Kempny1,
  6. Robert M Radke1,
  7. Boris Buerke2,
  8. Matthias Burg2,
  9. Christoph Schülke2,
  10. Helmut Baumgartner1
  1. 1Division of Adult Congenital and Valvular Heart Disease, Department of Cardiovascular Medicine, University Hospital Muenster, Muenster, Germany
  2. 2Department of Clinical Radiology, University Hospital of Muenster, Muenster, Germany
  1. Correspondence to Professor Helmut Baumgartner, Division of Adult Congenital and Valvular Heart Disease, Department of Cardiovascular Medicine, University Hospital Muenster, Albert-Schweitzer-Campus 1, Muenster 48149, Germany; helmut.baumgartner{at}


Objective Aortic regurgitation (AR) after transcatheter aortic valve implantation (TAVI) is associated with a poor clinical outcome and its assessment therefore crucial. Quantification of AR by transthoracic echocardiography (TTE), however, remains challenging in this setting. The present study used quantitative flow measurement by cardiac MRI (CMR) with calculation of regurgitant fraction (RF) for the assessment of AR and compared the results with TTE.

Methods and results We included 65 patients with a mean age of 82.2±8.1 years (38 women) who underwent successful TAVI with Edwards SAPIEN valves (52 transfemoral, 13 transapical). The postinterventional degree of AR was assessed by CMR and by TTE. There was agreement between CMR and TTE with regards to the absence of severe AR. However, TTE significantly underestimated the presence of moderate AR classifying it to be mild in 38 and moderate in only 5 patients, whereas CMR found mild AR in 23 and moderate in 16 patients. Overall, there was only fair agreement between CMR and TTE regarding the grading of AR with a weighted κ of 0.33. The rate of detection of TTE for more than mild AR was only 19%.

Conclusions Using CMR for the quantification of AR in a sizeable group of TAVI patients, we demonstrate a strong tendency of TTE to underestimate AR compared with CMR. Since higher AR severity on echocardiography has been associated with worse patient outcome, the potential incremental prognostic value of CMR should be studied prospectively in this setting.

View Full Text

Statistics from


Since its introduction in 2002, transcatheter aortic valve implantation (TAVI) has evolved as a promising alternative to conventional aortic valve replacement in surgical high-risk patients.1 One of the major concerns after aortic valve intervention remains postprocedural aortic regurgitation (AR), which has been linked to adverse outcome after surgical aortic valve replacement as well as TAVI.25 In both settings, even mild paravalvular AR has been associated with increased long-term mortality.6 ,7 After TAVI, several aetiologies for AR have been proposed, including annulus-prosthesis-size mismatch and insufficient sealing in the setting of heavily calcified cusps. While paravalvular AR has become rare with current surgical techniques of valve replacement, it remains common after TAVI with reported trivial or mild AR in up to 70% and moderate AR in more than 10%.8 The negative impact of residual AR on outcome obviously increases markedly with the severity of regurgitation.7 Thus, quantification of AR after TAVI appears crucial. Transthoracic echocardiography (TTE) is currently the standard imaging modality for the detection and quantification of AR after the procedure. While detection of AR is possible with high sensitivity and specificity, its quantification by TTE remains however challenging. The current standard used in TAVI research studies—the Valve Academic Research Consortium (VARC) criteria9—recommend quantitative assessment that has been developed and validated for native AR, such as size of the vena contracta or effective regurgitant orifice area. However, these are difficult to apply in this special setting and may actually not be suitable for evaluation of AR after TAVI.10 Magnetic resonance phase-contrast imaging (CMR), on the other hand, allows for accurate and reproducible blood flow measurement in the ascending aorta and quantitative assessment of the severity of AR by calculation of regurgitant fraction (RF).1115 Applicability and accuracy of this technique should not be affected by the specific setting—native AR versus paravalvular AR. Considering the critical effect of AR severity on outcome after TAVI and the difficulties to quantify AR by TTE in this setting, we sought to assess AR severity in a series of TAVI patients by CMR and compare the results with TTE findings.


We enrolled consecutive patients who had undergone TAVI at our institution and presented for regular follow-up visits. Patients underwent CMR if they had no contraindication such as an implanted pacemaker, the examination was logistically possible and the patient agreed to the additional diagnostic procedure but were otherwise not selected. All patients provided written informed consent, and the institutional review board approved the study. The decision for TAVI was made in accordance with current guidelines and after discussion and consensus within a multidisciplinary Heart Team consisting of cardiologists and cardiac surgeons. The TAVI procedure was performed with the balloon-expandable Edwards-SAPIEN or SAPIEN XT valve prosthesis (Edwards Lifesciences, Irvine, California, USA), which was available at this time in 23 and 26 mm sizes. Preprocedural annulus dimension was evaluated in all patients by ECG-gated CT and transesophageal echocardiography (TEE). Additional balloon sizing (angiography during balloon valvuloplasty) was performed in the case of uncertainty about the appropriate valve size. All procedures were performed under general anaesthesia and echocardiography guidance by TEE. Transfemoral access was used in 52 patients while a transapical access route was chosen in 13 patients because of poor vascular access.

At the end of the TAVI procedure, the degree of postprocedural AR was evaluated angiographically and by TEE after final device deployment and removal of the catheter and the guidewire.

After hospital discharge, clinical and TTE follow-up was obtained in all patients at 1–3 months, 6 months, 1 year and then annually. At one of these follow-up visits, patients had CMR and TTE generally within 24 h of each other.

Echocardiographic assessment

All echocardiograms were performed by one of two experienced echocardiographers (SO and AK) using a commercially available echocardiographic machine (Vivid E9 System, General Electric, Milwaukee, Wisconsin, USA) according to a standardised local protocol. Loops were recorded in accordance with published recommendations with subjects in the left lateral position.16 ,17 All recordings were stored digitally for offline analysis. The AR severity was graded according to the recommendations of the European Association of Echocardiography18 and the VARC criteria9 ,19 with a special focus on semiquantitative parameters like diastolic flow reversal in the descending aorta measured by pulse-wave Doppler and circumferential extent of prosthetic valve paravalvular regurgitation. Unfortunately due to limited acoustic windows, adequate echocardiographic quantification of RF was only possible in a minority of patients and has therefore not been included as part of the analysis.

All Doppler measurements were evaluated as the average of at least three cycles in patients with sinus rhythm or more than five cycles in those with atrial fibrillation. For the detection of regurgitant jets, the parasternal long-axis and short-axis views, the apical long-axis view and the five-chamber view were used (figure 1). Similar to the approach for native aortic valvular regurgitation, an integrative approach was applied to grade the AR. The degree of AR was classified into one of four grades: absent, mild, moderate or severe. In the presence of multiple AR jets, the overall degree of both paraprosthetic and central components was estimated in accordance with the VARC recommendations.9 In addition, a prominent holodiastolic flow reversal in the descending aorta classified a patient to have a more than mild AR.20 This evaluation of AR was performed by an echocardiographer who did not attend to the TAVI procedure and who was blinded to the results of the CMR evaluation of AR.

Figure 1

Example of transthoracic echocardiography in a patient with an Edwards SAPIEN prosthesis. Aortic regurgitation was graded by colour Doppler-echocardiography and CW-Doppler as moderate in this example.

Magnetic resonance imaging

All CMR examinations were performed in our radiology department on a 1.5-T MRI system (Achieva, Philips Healthcare, Best, the Netherlands) equipped with a standard five-element cardiac phased array coil for signal reception and a vector electrocardiograph for cardiac synchronisation. All scans were accomplished without sedation. The image acquisition and subsequent analysis was carried out according to current guidelines.21 For cine imaging, a single-slice two-dimensional (2D) balanced steady-state free precession (SSFP) sequence in breath-hold technique and with retrospective ECG triggering was used. Imaging parameters were chosen as follows: echo time (TE) and repetition time (TR) were set to shortest resulting in an average TR of around 4 ms and a TE of 2 ms slightly varying with slice orientation, typically 25 phases per cardiac cycle and with a reconstructed in-plane resolution of 1 mm. The slice thickness usually was in the range of 6–8 mm. The typical temporal resolution of the cine b-SSFP sequences was 30–40 ms depending on the heart rate. The slice for the through-plane phase-contrast flow imaging was placed perpendicular to the direction of flow approximately 10 mm above the aortic prosthesis. This adequate distance to the prosthesis was kept, as phase-contrast acquisitions may be prone to magnetic field inhomogeneities. Sequences for orthogonal images in at least two views were used to ensure the image plane is truly perpendicular to the flow direction. To avoid aliasing, velocity encoding was individually adapted, starting at 200 cm/s, and if aliasing occurred, the maximum velocity was increased by 50 cm/s steps until aliasing did not occur. Image acquisition is gated to the ECG signal and acquired over several cardiac cycles during a 10–20 s breath-hold. Outlining the region of interest within the aortic lumen for each cardiac phase, the instantaneous flow volume (cm3/s) can be calculated and graphically displayed over the entire cardiac cycle (figure 2).

Figure 2

(A and B) Example of cardiac MRI (CMR) images obtained with steady-state free precession acquisition (blue lines=planning lines for imaging planes of velocity-encoded CMR imaging). (C and D) Magnitude image and phase velocity map. (E) Resultant time–velocity curve of aortic flow rate (red part represents regurgitant flow).

The software calculates forward and reversed flow volumes, and from them the regurgitation fraction as follows: aortic regurgitation fraction (RF, %)=diastolic reversed flow volume×100/systolic forward flow volume. As previous comparisons in native valves with both modalities suggest, AR severity was classified with RF below 10% as absent/minimal, 10–20% as mild, 20–40% as moderate and greater than 40% as being severe (table 1).13 ,14 ,2224

Table 1

Classification and grading of aortic regurgitation by cardiac MRI regurgitation fraction (%) in recent studies

Statistical analysis

Values are presented as mean and SD or median and IQR (25th and 75th centile), depending on variable distribution. Categorical variables are presented as frequencies and percentages. Comparisons between subgroups were performed by unpaired t test, Mann–Whitney U test or χ2 test as appropriate. For all analyses, a two-tailed p<0.05 was used as the criterion for statistical significance. Inter-rater agreement was assessed by calculating a κ-statistics with corresponding 95% CIs. Statistical analyses were performed using MedCalc V. (MedCalc Software, Mariakerke, Belgium).


Study population

We included 65 patients (age, 82±8 years; 38 male) who had undergone TAVI at our institution. The assessment was performed at a median of 69 days after the TAVI procedure. Of the 65 patients, 6 asked for the early termination of the CMR examination due to discomfort/anxiety and, thus, AR severity could not be quantified on CMR. The remaining 59 patients undergoing both TTE and complete CMR on the same day constituted the study population. Of these, 46 had undergone a transfemoral TAVI (34 with an XT-valve). The valve size was 23 mm in 36 patients (61%) and 26 mm in 23 patients (39%). Further demographic and baseline patient characteristics are presented in table 2.

Table 2

Characteristics of the study population

Frequency and degree of postinterventional AR

Overall, 39 (66%) patients presented with at least mild AR post-TAVI based on CMR assessment. Of these, 16 (27% of the study population) had moderate AR, while no patient was classified to have severe AR on CMR.

The mean RF on CMR was 15.1±9.9% and was not significantly different between patients undergoing XT or non-XT valve implantation (14.6±7.4% vs 15.6±11.4, p=0.70, for non-XT and XT valve prosthesis patients, respectively). In addition, no significant difference was found between valve sizes (RF 15.5±11.2% vs 15.0±9.1%, p=0.90, for 23 and 26 mm prostheses, respectively) and transapical versus transfemoral TAVI approach (RF 12.5±5.7 vs 15.8±10.6%, p=0.30). The severity of AR on CMR was not correlated with patients’ age (r=−0.018; p=0.89) and preprocedural mean pressure gradient on TTE (r=−0.04; p=0.77).

Agreement between CMR and TTE

Overall, the agreement between TTE and CMR regarding the severity of AR was only fair (κ=0.33 (95% CI 0.15 to 0.51)). This is illustrated in table 3 and figure 3.

Table 3

Agreement between TTE and CMR regarding the severity of AR

Figure 3

Dotplot comparing regurgitant fraction on cardiac MRI (CMR) with echocardiographic classification of the degree of aortic regurgitation (AR). The red dots represent patients who were misclassified by transthoracic echocardiography according to the CMR cut-off values for AR employed (dotted horizontal lines). The vertical lines represent means and 95% CIs.

In patients with absent or minimal AR on CMR, TTE overestimated the severity of AR in one-third of patients. All misclassified patients were diagnosed to have mild AR by TTE. More importantly, however, TTE underestimated the degree of AR in those patients presenting with moderate AR on CMR. In fact, out of 16 patients with CMR-diagnosed AR, only 3 were correctly classified by TTE (table 3, figure 3). This resulted in a rate of detection of moderate AR by TTE of only 19%.


To our best knowledge, this is the first study to directly compare AR severity assessment by TTE and CMR in a sizeable group of patients who underwent TAVI. The results demonstrate considerable discrepancies between CMR and TTE in quantifying AR severity in this special setting. It is concerning that TTE underestimated the degree of AR in most patients with moderate AR on CMR and had a sensitivity of detecting more than mild AR of only 19%.

The limited accuracy of TTE in quantifying AR after TAVI is of particular clinical interest since postprocedural AR assessment is currently mostly based on echocardiographic measurements. Although the adverse prognostic impact of moderate or severe postprocedural AR has been described, some studies have also demonstrated that postprocedural AR may have a significant adverse impact on long-term outcome even if it is only mild.35 ,26 ,27 While the definitive reasons for these findings remain insufficiently understood, it is hypothesised that a hypertrophied LV that has been exposed to long-term pressure overload may not be able to adapt quickly to volume overload.4 Previous TTE-based studies have suggested that even mild postinterventional AR may be clinically detrimental,7 a finding that is difficult to explain physiologically. Based on the results of the current study, one may speculate that at least part of these patients with AR classified as mild may in fact have had at least moderate AR, which could help explain their worse clinical outcome.

The reasons for the limited accuracy of TTE in quantifying AR after TAVI could be related to poor acoustic windows with reduced image quality in some of the elderly patients, the eccentricity of paravalvular AR jets and the difficulty to deal with multiple paravalvular leaks diverging into different directions into the LV outflow tract. In this setting, CMR appears to be a very attractive alternative. Quantification of forward and backward flow in the ascending aorta with calculation of the RF is not affected by the above-described specific phenomena occurring at the valve level and in the LV. It may therefore be considered the gold standard and should be considered for prospective studies in the field of TAVI. Unfortunately, some difficulties to integrate CMR in large multicentre trials or registries are obvious. CMR is still not ubiquitously available and is resource intensive. In addition, we demonstrated that in this particular group of elderly patients a considerable proportion may not tolerate CMR well. This is due to general physical discomfort in the scanner, anxiety and the relatively long duration of the examination. In addition, previous research has indicated that severity of AR may be dynamic,7 thus requiring repeated follow-up investigations, which are logistically challenging using CMR for the reasons discussed above.

Comparing the spectrum of AR severity on TTE in our study with previous publications, we find a similar distribution of AR severity suggesting that the cohort investigated here is likely to be representative of the general TAVI population.25 ,28 ,29 We could not identify patient or prosthesis-related risk factors for more than mild AR in the current study. Finding such predictors would be, however, clinically appealing as it would potentially allow selecting patients who should undergo CMR in addition to routine TTE.

Although the current study highlights the limitations of echocardiographic quantification of AR post-TAVI, it should be emphasised that the studies associating AR with worse prognosis postintervention have relied on echocardiographic assessment.19 Therefore, further studies using both modalities with a longer duration of follow-up and morbidity or mortality endpoints are required to assess whether CMR-based AR assessment carries additional prognostic information in this patient population.

Study limitations

We opted to include patients of all ages and functional status but cannot fully exclude the possibility of selection bias as only patients presenting to our follow-up clinic and agreeing to participate could be enrolled. The accuracy of CMR flow measurement with the Q-flow method is well validated in experimental models and clinical studies.23 However, if the imaging slices are not truly perpendicular to flow, inaccurate velocity data may be obtained. We presented no data of AR assessment during the implantation procedure like TEE or angiography because our study was mainly designed to compare AR grading between TTE and CMR during routine follow-up (in this study, a median 69 days after TAVI). In addition, it is unclear how AR severity and other haemodynamic parameters change between implantation and CMR/TTE and this may confound any associations between these parameters measured at different time points.

We also have not quantified the degree of aortic valve calcification pre-TAVI, which could represent an important determinant of postinterventional AR. Echocardiograms and CMR studies were performed on the same day but at different times of the day, and we cannot completely exclude changes in physiological variables between the tests. However, this reflects normal clinical practice and we do not expect this to have a major impact on the results.

Due to the lack of quantitative TTE parameters of AR such as TTE RF, we cannot exclude the possibility that the observed differences in AR severity, as estimated by both techniques, is due to a lack of correlation between haemodynamic measures or is—at least in part—influenced by the system of categorisation of AR used.


AR is the most frequent complication after TAVI and has been shown to be an independent predictor of morbidity and mortality. Its negative impact increases with the extent of regurgitation based on previous echocardiographic studies. Thus, accurate assessment of postprocedural AR appears crucial. Echocardiography that is generally used for this purpose, however, significantly underestimates the severity of AR compared with CMR flow measurements. CMR should therefore be considered in future studies that evaluate this interventional technique.

Key messages

  • What is already known on this subject?

  • Paravalvular aortic regurgitation (AR) after transcatheter aortic valve implantation (TAVI) has been associated with increased mortality, but quantification of postprocedural AR with transthoracic echocardiography (TTE) remains challenging.

  • What might this study add?

  • The postinterventional degree of AR was assessed by cardiac MRI (CMR) and TTE. CMR allows for accurate assessment of the severity of AR by calculation of regurgitant fraction.

  • How might this impact on clinical practice?

  • TTE is generally used for the assessment of postprocedural AR but significantly underestimates the severity of AR compared with CMR flow measurements. CMR should therefore be considered in future studies that evaluate this interventional technique.


View Abstract


  • Contributors SO, G-PD and HB designed the study. SO, G-PD, AK and RMR analysed the data. SO, G-PD and HB prepared the first draft of the manuscript All authors have revised the manuscript critically for important intellectual content and have provided final approval of the manuscript.

  • Competing interests HB acts as a consultant and proctor for ‘Edwards Lifesciences’; all the other authors have no disclosures.

  • Patient consent Obtained.

  • Ethics approval Institutional Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles