Measure or predict prosthesis–patient mismatch?

This editorial refers to ‘Reclassification of prosthesis–patient mismatch after transcatheter aortic valve replacement using predicted vs. measured indexed effective orifice area’, by J. Ternacle et al., pp. 11–20. No valve prosthesis can fully correspond to native valvular area. Prosthesis–patient mismatch (PPM) occurs immediately after the procedure when the size of the prosthesis orifice is too small for the body size of the patient.1 This results in persistence of increased afterload and hence to high transvalvular pressure gradients in resting conditions that can increase significantly during exercise. The parameter used to define PPM is the effective orifice area (EOA) divided by the patient’s body surface area (BSA), called indexed EOA.2 As the gradients have an exponential increase when the indexed EOA is ≤0.85 cm2/m2, PPM for aortic prostheses occurs when the indexed EOA is lower than this value. PPM is considered to be moderate when indexed EOA is between 0.85 and 0.65 cm2/m2 and severe when indexed EOA is <0.65 cm2/m2.3,4Cut-off values have been adapted for obese patients (body mass index ≥ 30 kg/m2): moderate PPM between 0.7 and 0.55 cm2/m2 and severe when it is ≤0.55 cm2/m2.5 Severe PPM is associated with less regression of the left ventricular (LV) mass,6 persistence of reduced coronary flow reserve, higher incidence of heart failure, and early and late mortality.7 The incidence of PPM was high in patients submitted to surgical aortic valve replacement (SAVR), occurring in 20–70% of patients.3,8 Prevention of PPM before SAVR consists in the selection of the most appropriate prosthesis with at least a minimum EOA, obtained by multiplying BSA by 0.85 cm2/m2,1 using stentless bioprosthesis9,10 or even in the case of small annulus perform aortic root enlargement which can increase the operative risk.11Supra-annular implantation has also be performed to avoid PPM.12 The most frequently used method to assess EOA and thus PPM was based on Doppler echocardiography, using the continuity equation. This method has however several pitfalls. The most important is the underestimation of the diameter of the LV outflow tract (squared in the equation) and thus the stroke volume which can overestimate PPM. Indeed, the LV outflow tract is assumed to be circular, although it is frequently elliptic, the smaller diameter being measured in the LV long-axis view, the view recommended for its measurement. Another limitation of the Doppler method occurs in the patients with low flow and low stroke volume as the consequence of LV dysfunction and hence a limited opening of the leaflets of the prosthesis. The pressure recovery phenomenon can also overestimate PPM. Another method, intuitively more accurate and reproducible is the predicted indexed EOA, using the size and the model of the prosthesis divided by BSA.5 The incidence of PPM is high with SAVR, not transcatheter aortic valve replacement (TAVR): a large series of 59 779 patients who were submitted to SAVR showed moderate PPM in 54% of patients and severe PPM In 11%, associated with 1.08-fold and 1.32-fold increase in mortality.13 TAVR is increasingly used in patients with high, moderate, and even low risk of surgery. The incidence of PPM in patients submitted to TAVR has been found to be lower than those who underwent SAVR in the PARTNER 1A (Placement of AoRTic TraN cathetER Valve 1A).14 Severe PPM was associated with an increased mortality in the SAVR arm but not in the TAVR-arm. In the CoreValve Pivotal trial, severe PPM occurred in 19.2% of patients in the SAVR group and only 7% in the TAVR group, but with increased mortality in the two groups.15 Several meta-analyses have been performed. Takagi and Umemoto16 included 4000 patients undergoing TAVR: they found overall PPM incidence of 35.1% and severe PPM in 8.0%. A more recent and much larger report by the Society of Thoracic Surgeons/American College of Cardiology (STS/ACC) found in 62 125 patients moderate PPM in 25% of patients and severe PPM in 12%. PPM occurred more frequently in patients with large body mass index and those who underwent valve-in-valve procedure. In this registry, severe PPM was associated with higher mortality and re-hospitalization for heart failure at 1 year.17 However, the measurements were not adjucated in a core lab of echocardiography and patients who died early after the procedure were not included, a risk of underestimating the role of PPM. A meta-analysis of 71 106 patients showed an incidence of PPM after TAVR of 36.3%. Severe PPM was associated with a significant increase in 1-year mortality; it was not observed in patients with moderate PPM.18 Another meta-analysis of 21 studies found severe PPM prevalence of 8.0% and a moderate PPM prevalence of 26.7%; the presence of PPM did not impair late survival. These results remain controversial. Indexed EOA was obtained in most studies and centres by the echocardiographic method which could imply an overestimation of the frequency of PPM. Ternacle et al.19 used the database of the 1088 patients with severe aortic stenosis or failed bioprosthesis who underwent TAVR at the Québec-University Laval between May 2007 and December 2018. The group of Pibarot developed the definition, haemodynamic and clinical impact of PPM since more than 30 years. The study of Ternacle et al. was a retrospective study of prospectively included variables. The authors tested the hypothesis that the predicted indexed EOA was a better parameter than the indexed EOA calculated with Doppler echocardiography. All-cause mortality until 8 years and 1-year re-hospitalization for heart failure were available. The main results of this study indicate a much lower incidence of both moderate and severe PPM when identified by the predicted indexed EOA compared with the measured parameter. Nine percent of patients had a high residual gradient (cut-off value ≥ 20 mmHg) which was more strongly associated with the predicted than the measured PPM. Logically, valve-in-valve procedure was a factor associated with PPM. Another factor was the use of balloon-expandable vs. self-expanding device. A large proportion of the included patients (83%) received balloon-expandable valves. The balloon-expandable device requires precise positioning for deployment, with the use of rapid pacing. With the self-expanding prosthesis, minimization of PPM occurs with a narrow window and also needs precise optimization of the final device position.20 In the study of Ternacle et al.,19 although predicted PPM was more associated with the haemodynamic consequences, both predicted and measured PPM were not associated with events: all-cause mortality at 30 days, at 1 year, and on the long term until 8 years, cardiovascular mortality at 1 year, heart failure readmission at 1-year, and the composite outcomes were not different that in the patients without PPM. Of note, a large proportion of the cohort (45%) was in low-flow condition, a factor which could lead to overestimation of measured PPM. In this setting, a dobutamine Doppler echocardiography could be useful to determine if the opening of the leaflets can increase during the test and lead to an increase of the measured EOA and indicate pseudo PPM. Exercise Doppler echocardiography can provide important information on the extent of trans-prosthetic gradients during exercise, thus in the real life. Such studies have mainly been performed after SAVR.21,22 Long-term survival and morbidity are warranted to confirm the absence of guarded prognosis in patients treated by TAVR who had PPM.23 This is particularly important in young age patients. Despite the reassuring results of the study by Ternacle et al.,19 the Heart Team should include in their discussion the predicted indexed EOA, and identify patients at risk of PPM, such as large body size, young age, small aortic annulus for the decision of TAVR vs. SAVR and for selecting the most appropriate device and its size. Although the predicted indexed EOA does not require an echocardiogram, comprehensive Doppler echocardiography is required in patients with severe aortic stenosis who require an intervention. Pre-procedural cardiac computed tomography (CT) is essential to select the size of the prosthesis; it permits a precise measurement of the annulus area and avoids the underestimation of the area obtained by echocardiography. Optimal valve deployment is essential and can be assessed by post-procedural CT. Conflict of interest: none declared. The opinions expressed in this article are not necessarily those of the Editors of EHJCI, the European Heart Rhythm Association or the European Society of Cardiology.