The impact of patient‐prosthesis mismatch on early and long‐term survival after aortic replacement with the Edwards Perimount valve: A propensity score‐matched analysis

To investigate the impact of severe patient‐prosthesis mismatch (PPM) related to the Edwards Lifesciences Perimount (EP) bioprosthesis in the aortic position on early in‐hospital outcomes and long‐term survival.


| INTRODUCTION
In the last two decades, the use of bioprosthetic aortic valve implantation has increased significantly, far exceeding the use of mechanical prostheses 1,2 The Carpentier-Edwards Perimount (EP) bovine pericardial bioprosthesis (Edwards Lifesciences, Irvine, CA) is widely recognized as one of the most commonly used bioprosthetic valves for many years. This valve design was developed to overcome the difficulties observed with previous pericardial prosthetic models with the aim of optimizing the hemodynamic profile with improved durability. 2,3 Since its introduction, multiple reports have demonstrated excellent clinical outcomes and durability, especially when implanted in the aortic position. [4][5][6][7][8][9] One of the most common issues after valve replacement, especially after implantation of a biological prosthesis, is a valve with a reduced effective orifice area compared to the native valve. This condition is referred to as patient-prosthesis mismatch (PPM) 10 and its main hemodynamic consequence consists of an elevated transvalvular gradient through a normal functioning prosthetic valve. 11 After aortic valve replacement (AVR), PPM can vary in terms of severity according to the effective orifice area (EOA) of the prosthetic valve indexed to the patient's body surface area (EOAi). In the literature PPM is described as mild or absent, having no clinical impact, when EOAi is >0.85 cm 2 /m 2 ; moderate PPM when EOAi is between <0.85 cm 2 /m 2 and >0.65 cm 2 /m 2 and severe when EOAi < 0.65 cm 2 /m 2 . 2,3 The prevalence of PPM ranges between 20% and 70% for moderate and between 2% and 20%, for severe. 12 The impact of PPM on short and long-term mortality is still the subject of debate as many studies have reported very conflicting outcomes. [12][13][14][15][16][17] In this study, we report a single UK center experience with the use of an EP valve implanted in the aortic position. The aim of the study was to investigate the incidence and the clinical impact of severe PPM on early in-hospital outcome and long-term survival in patients undergoing aortic valve replacement with EP valves.

| Patient selection and data collection
We performed a single center retrospective analysis of prospectively collected data of patients undergoing AVR. Data were prospectively collected, validated, and stored by a data management team at our institution, as part of the UK National Institute for Cardiovascular Outcomes Research (NICOR) Registry. The study protocol was in compliance with the local Institutional Clinical Audit Review Board and patient consent was waived.
From January 1998 to December 2014, a total of 5964 consecutive adult patients underwent AVR for Aortic valve disease. Of these, 2667 patients received EP valve for both isolated and combined procedures and they represent the cohort of this study ( Table 1).
The baseline data collected includes clinical characteristics, symptom status, and past medical history. (Table 1).
Indications for surgery were based: on referral for surgery by a cardiologist, baseline clinical details, echocardiographic results, coronary angiography, and when necessary following case discussion by the Heart Team. Late survival data after discharge were obtained from the UK Office of National Statistics.

| Outcomes measurement
All patients underwent transthoracic echocardiography preoperatively; LV function was assessed by the ejection fraction (LVEF) accordingly to the Simpson rule.
Early outcome collected included: in-hospital mortality, cerebrovascular accident (CVA), defined on the basis of a focal or global neurological impairment at a physical examination or at CT scan/ magnetic resonance imaging, postoperative acute kidney injury (AKI) (requiring haemofiltration), low cardiac output (requiring IABP support), postoperative bleeding (requiring surgical re-exploration), deep sternal wound infection (defined as a surgical site-related infection affecting the median sternotomy wound and requiring antibiotics and/or surgical re-exploration) and in-hospital length of stay.

| Propensity matched analysis
After propensity score matching, the overall number of patients was reduced to 320 patients with non-severe PPM (group Am) and 320 patients with severe PPM (group Bm), with similar preoperative characteristics ( Table 1). The early postoperative outcome of the matched analysis is shown in and 57.3% vs. 58%, group Am vs. group Bm p = .9; Figure 2).

| DISCUSSION
During the last two decades, a generational change in aortic valve prosthesis implantation has been observed with biological prostheses implanted progressively more often than mechanical ones. 1 Freedom from anticoagulation, better outcome in redo surgery and, furthermore, the increased performance of the new The effects of PPM on short and long-term outcomes continue to be a matter of debate, a recent study by Blais et al. 13 suggested that one of the strongest independent predictors of mortality after AVR was the presence of postoperative PPM.
Similarly, Pibarot et al. 21 in a subanalysis from the PARTNER trial cohort showed that severe PPM was an independent predictor of 2-year mortality in the surgical AVR cohort. Hong et al. 22 analyzed data from patients who underwent isolated AVR and showed that early survival was not significantly affected by PPM, however, in the group with severe PPM the overall survival as well as the cardiac-related-death-free survival were significantly reduced at 12 years. 20 Interestingly, other studies report no significant differences in outcome ascribed to the presence of postoperative PPM. A recent study by Swinkels et al. 23  As previously demonstrated by some studies these results suggest that within balanced groups the same risk factors had a similar impact on early and long-term mortality independently from the presence of moderate PPM. 20 In that study, the classification of PPM has been extracted from previously published reference values 14 followed by echocardiographic validation.
The question of PPM remains a matter of debate as there are many unanswered questions that need to be prudently addressed before reaching any solid conclusions that impact surgical practices. 26,27 Most of the data available are based on tables provided by manufacturers which often report the best evaluation of the EOA; moreover, data are usually presented only as mean, while standard deviation is known to be a significant parameter to take into account. 28,29 Since the determination of the EOA can substantially differ when investigated in-vivo, the EACTS-STS-AATS Valve Labeling Task Force has been advocating the regulation and standardization of the information provided in the valve prosthesis charts. 30 Furthermore, the specifics of the PPM assessment that should be conducted in vivo need to account for the optimal time frame, the need and the rate of repeated examinations, and the findings required to reach a conclusive result. 27 Lastly, it is still under debate whether BSA is the most accurate parameter to investigate the surface area in the context of PPM assessment, as lean mass has also been suggested as a potentially effective alternative. 27,31 Our study has many limitations and the results should be interpreted carefully as we showed that PPM based on EOAi is not associated with the increased adverse outcome; however, as argued above this may not be the correct method to look at PPM.
The EOAi assessment was extracted following previous publications [14][15][16][17][18] and not based on echocardiography. Nonetheless, even though the majority of studies in the literature have broadly described the evaluation of PPM based on the calculation by echocardiogram and Doppler of the continuity equation across the LVOT, the accuracy of this assessment has been questioned for several reasons, such as anatomical and methodological factors. 32,33 Additionally, this was a retrospective single-center analysis on prospectively collected data in a limited cohort. The allocation of patients to the study group was by surgeon expertise, and this might have led to undetected difference in risk profile between groups. The study included a patient cohort treated over a long-time period, hence with possible confounding factors due to changes in clinical practice over time.
Finally, the evaluation of long-term outcomes after surgery was limited to all-cause mortality as no data were available on structural valve deterioration or the need for re-do surgery.
We conclude that the presence of EOAi < 0.65 cm 2 /m 2 based on published valve areas measurement in patients undergoing AVR using EP valves does not affect early or late outcomes.