Background: The importance of moderate patient-prosthesis mismatch (PPM) for the prognosis of patients who undergo aortic valve replacement is unclear.
Methods: The presence of PPM was assessed in 361 consecutive patients undergoing valve replacement for isolated severe aortic stenosis and related to perioperative and postoperative mortality. Indexed effective orifice areas (EOAi) were estimated for each type and size of prosthesis.
Results: Using the previously proposed cut-off of EOAi ⩽0.8 cm2/m2, PPM was present in 54% of patients. Patients were followed for 4.1 (2.0) years. Survival tended to be slightly, but insignificantly, worse in the group with PPM (1-year, 3-year and 5-year survival 89%, 86% and 76% vs 92%, 88% and 82%; p = 0.21). However, patients with PPM were also older (p<0.0001), more often female (p<0.0001), more symptomatic (p = 0.001), more often had coronary artery disease (p = 0.04), triple vessel disease (p = 0.03) and hypertension (p = 0.01) and presented with a higher EuroSCORE (p<0.0001). By multivariate analysis only EuroSCORE and diabetes but not PPM were independent predictors of survival.
Conclusions: Moderate PPM is a frequent finding after aortic valve replacement. In our patient population it had no impact on short-term and long-term survival. It may therefore not be justified to recommend complex surgical interventions to avoid moderate PPM in patients undergoing aortic valve replacement for isolated severe aortic stenosis.
Statistics from Altmetric.com
In 1978 Rahimtoola first introduced the term “patient-prosthesis mismatch” (PPM) to describe a condition in which the orifice area of an implanted aortic prosthetic valve is smaller than that of the original human valve.1 2 More recently moderate PPM was defined as an indexed prosthetic effective orifice area (EOAi) smaller or equal to 0.80 cm2/m2 body surface area (BSA).3 4 In most studies on PPM the proportion of patients who are subject to some degree of mismatch after aortic valve replacement (AVR) is substantial, ranging from 20% to 70%.3–14
The importance of moderate and severe PPM for outcome is a controversial issue. Several studies indicated that moderate PPM significantly increases morbidity and mortality3 6 7 and also leads to reduced exercise capacity,15 while others showed no impact on long-term outcome but elevated short-term mortality.5 10 On the basis of these results it has been proposed that PPM should be avoided if possible and that even left ventricular outflow tract enlargement may be justified to prevent moderate PPM.4 5 16 17 On the other hand a considerable number of studies failed to show an effect of PPM on short-term as well as long-term mortality8 9 11–14 18 or quality of life19 after aortic valve replacement.
From currently available data it seems likely that PPM has an impact on outcome after aortic valve replacement in some patients but probably not in all. However, the inconsistent study results are difficult to interpret since PPM was not uniformly defined, patient groups were not clearly specified and in most papers causes of death were not specified.
Considering the large number of patients affected it appears highly important to define very precisely which individuals are at high risk and may benefit from more complex surgical techniques to prevent moderate PPM in the setting of AVR.20 21 In the present prospective study we have investigated the outcome in a strictly defined group of patients, referred for surgery for isolated severe aortic stenosis (AS).
Between January 1998 and January 2005, 361 consecutive patients were seen in our outpatient clinic for valvular heart disease and were referred for subsequent AVR for isolated severe AS. Patients with more than mild concomitant aortic regurgitation were excluded. Clinical, echocardiographic, operative and outcome data were prospectively collected on a computerised database.
At study entry the following clinical data were collected: age, gender, body mass index and body surface area, New York Heart Association (NYHA) and Canadian Cardiovascular Society (CCS) functional class, history of hypercholesterolaemia (cholesterol >200 mg/dl or patient undergoing lipid-lowering therapy at study entry), diabetes mellitus (fasting blood glucose >110 mg/dl or use of antidiabetic medication/insulin), hypertension (blood pressure >140/85 mm Hg on the basis of the average of repeated measurements), chronic renal insufficiency (creatinine ⩾2 mg/dl), presence of chronic obstructive pulmonary disease (long term use of antiobstructive therapy), coronary artery disease (angiographically documented coronary artery stenosis) and an ECG. Furthermore the EuroSCORE was calculated for each patient.22 23 Good systolic left ventricular function was defined by an ejection fraction >50% and/or fractional shortening >28%. The baseline characteristics of all patients are presented in table 1.
All patients underwent a comprehensive echocardiographic examination before operation, 3–9 months postoperatively and at each follow-up appointment including M-mode echocardiography, two-dimensional echocardiography and conventional and colour Doppler ultrasonography conducted by experienced echocardiographers. Echocardiographic data were obtained with commercially available ultrasound machines.
Definition of PPM
The projected EOA was derived from the published normal in vivo effective EOA values,24–31 which has been shown to be the most reliable method for the determination of PPM.32 The reference values that we used in our study are those which have been proposed by Pibarot and coworkers, because these values have been used in most publications on PPM.
Table 2 shows reference EOA values for each size and model of prosthesis.
To obtain the EOA indexed to the patients’ body surface area (EOAi), the effective EOA was divided by body surface area (BSA), and PPM was defined as EOAi ⩽0.8 cm2/m2. The selection of this value was based on the results of previous studies.3 33 Severe PPM was defined as EOAi ⩽0.60 cm2/m2.34
Patient evaluation and follow-up
Patient assessment included history, physical examination and echocardiography at study entry. All patients had isolated severe AS (peak velocity ⩾4 m/s and mean pressure gradient ⩾50 mm Hg in the presence of normal left ventricular function, calculated valve area <1.0 cm2) and presented with symptoms (exertional dyspnoea ⩾ functional NYHA class II, exertional angina pectoris ⩾ functional CCS class II, syncope) or with reduced left ventricular function (ejection fraction ⩽50%) when asymptomatic. Patients who were primarily referred to surgery for coronary artery bypass grafting (CABG) and additionally received an aortic prosthesis for non-severe AS were not included in this study.
Patients were re-evaluated 3–9 months after AVR, and at intervals between one and three years thereafter, including history, echocardiography, physical examination and evaluation of functional status (NYHA and CCS class) and an ECG.
Short-term mortality was defined as death from any cause within 30 days after operation if the patient was discharged from hospital or within any interval if the patient was not discharged.35
For the assessment of outcome, the primary end point was death. Perioperative as well as late deaths were included in the analysis. Deaths were classified as cardiac or non-cardiac on the basis of review of medical records, including autopsy records and death certificates, which were available in all but one case (table 3). For the assessment of outcome with regard to quality of life, postoperative symptoms and functional status at last follow-up were assessed.
Statistical analysis was performed using SAS 9.1 for Windows (SAS statistical software, SAS Institute, Cary, NC, USA). Kaplan-Meier estimates were used to calculate the 1-year, 3-year and 5-year survival. Differences between survival curves of patients with EOAi <0.8 (PPM) and EOAi ⩾0.8 (no PPM) were analysed using log rank from PROC LIFETEST. Univariate Cox regression was used to identify factors influencing survival and with these factors a multivariate Cox regression was carried out. Additionally, the propensity score method was conducted to reduce bias in the observational study examining the relation between PPM and survival. Differences in these and all other tests were considered significant at p⩽0.05.
Two patients were lost to follow-up, thus follow-up information was complete for 359 patients (99%). Patients were either seen in our outpatient clinic for valvular heart disease (74%) or phoned (8%) or they returned a questionnaire (18%).
Table 1 shows the baseline characteristics of the 361 patients enrolled in the study.
Using the previously proposed cut-off of EOA ⩽0.8 cm2/m2, PPM was present in 54.3% of patients. Groups differed significantly with regard to baseline characteristics.
Patients with PPM were older (p<0.0001), more often female (p<0.0001) and presented with a higher EuroSCORE (p<0.0001). They had a larger body surface area (p = 0.003) and body mass index (p<0.0001). Furthermore, patients with PPM more often had coronary artery disease (p = 0.04), triple vessel disease (p = 0.03) and hypertension (p = 0.01), and were more symptomatic preoperatively compared to patients without PPM (p = 0.001).
Operative data are presented in table 4. Emergency operations and additional CABG were equally distributed between patients with and without PPM. However, patients with PPM significantly more often received a biological valve (p <0.0001) and a prosthesis ⩽21 mm (p <0.0001). The mean prosthesis EOAi in patients with PPM was 0.69 (0.07) cm2/m2 versus 0.94 (0.14) cm2/m2 in patients without PPM (p <0.0001). Only 3% of patients received an aortic prosthesis with an EOAi ⩽0.6 cm2/m2, a condition previously defined as severe mismatch. Aortic root enlargement was only performed in seven patients with PPM (3.6%).
Short-term mortality was 8.0% (predicted operative mortality by additive EuroSCORE 6.3%).
According to the differences in baseline characteristics and predicted operative risk mortality also differed between groups: 20 patients (10.2%) in the group with PPM and nine patients (5.5%) in the group without PPM died. However, this difference did not reach statistical significance (p = 0.14). Among these early deaths, nine (4.6%) patients in the PPM group were defined as cardiac deaths, while this was the case in five (3.0%) patients in the non-PPM group (p = 0.60). Causes of death are shown in table 3.
Patients were followed for 4.2 (2.0) years (up to 8.5 years). Mean follow-up time was similar in both groups (4.1 (2.1) years in the PPM group versus 4.3 (2.0) years in patients without PPM, p = 0.34). Causes of late deaths are listed in table 3; 11.5% of patients without PPM and 9.7% of patients with PPM died during follow up (p = 0.52). Symptoms and functional status at last follow-up remained slightly worse in patients with PPM (NYHA 1.67 (0.7) in PPM patients versus 1.52 (0.6) in non-PPM patients, p = 0.021; CCS 1.12 (0.4) versus 1.04 (0.19), p = 0.032, respectively). Mean transvalvular gradients at last follow up were significantly higher in the PPM group (20.7 (8.7) mm Hg in PPM patients versus 15.9 (6.5) mm Hg in non-PPM patients, p <0.0001).
Overall survival as estimated by Kaplan-Meier analysis tended to be slightly worse in the group with PPM. 1-year, 3-year and 5-year survival rates were 89%, 86% and 81% in PPM patients versus 93%, 89% and 86% in patients without PPM. This difference did not reach statistical significance (p = 0.21) (fig 1).
A univariate Cox regression analysis was performed with the following variables: PPM, EuroSCORE , age, sex, body surface area, hypertension, diabetes, reduced left ventricular function, chronic obstructive pulmonary disease, renal failure, coronary artery disease (CAD), additional coronary arterial bypass grafting (CABG), additional CABG ⩾two grafts, and previous cardiac surgery.
All univariate significant variables (EuroSCORE, age, diabetes, reduced left ventricular function, CAD, additional CABG, additional CABG ⩾two grafts, previous cardiac surgery) were used to perform a multivariate stepwise Cox regression. By multivariate analysis EuroSCORE and diabetes were the only independent predictors of long-term survival (table 5).
For calculating a propensity score, we performed a stepwise multivariate logistic regression analysis. The significant variables from the univariate Cox regression were taken as independent variables (EuroSCORE, CABG, age, previous cardiac surgery, diabetes, coronary artery disease, CABG ⩾two grafts). The rescaled R2 was 0.18 and age was the only significant parameter (p value <0.0001). We divided the sample into age quintiles and conducted a Cox regression for each quintile. In the quintiles, PPM showed no significant effect (p values 0.96, 0.22, 0.34, 0.53, 0.48).
The impact of moderate and even severe patient-prosthesis mismatch (PPM) on outcome after aortic valve replacement (AVR) remains unclear. A considerable number of studies have yielded conflicting results.3–15 17 18 However, these studies are difficult to interpret since PPM was not uniformly defined, patient groups were not clearly specified and in most papers causes of death were not discussed. Since a substantial number of patients who undergo AVR are affected by PPM and since the use of more complex surgical techniques to avoid PPM may put patients at an additional operative risk,21 36 it is important to define patients groups whose outcome is worsened by PPM.
The present study evaluated the impact of moderate PPM on survival after AVR in stringently defined patients who had been referred for surgery for isolated severe aortic stenosis (AS).The results show that PPM is strongly associated with unfavourable preoperative conditions compared with patients who finally receive an “adequately sized” prosthesis (EOAi >0.80 cm2/m2). A significantly higher preoperative EuroSCORE in PPM patients already indicates this fact (table 1). Patients with PPM were older, had a higher incidence of CAD and triple vessel disease, more often had hypertension and were more symptomatic. This apparent over-representation of high-risk patients in the PPM group could be accidental but may also be caused by the fact that facing a very sick patient may prompt a surgeon to a fast “get in and out” strategy,8 37 38 while paying less attention to sizing of the valve, particularly because aortic root enlargement increases aortic clamp and cardiopulmonary bypass time by approximately 20 minutes.20 The unfavourable baseline characteristics of PPM patients were, surprisingly, not associated with a worse outcome. Moderate PPM by itself, however, did not appear to have a significant negative impact in various analyses.
In a recently published study3 PPM was defined as a strong and independent risk factor for cardiac events and midterm mortality after AVR for AS. The difference between these results and our data is not easy to explain. However, we followed a larger number of patients (361 versus 315) and our mean follow-up time was longer (4.2 years versus 3.7 years, respectively). In addition, our patients presented with significantly more advanced aortic valve disease as indicated by higher initial mean aortic valve gradients (67 mm Hg versus 50 mm Hg) and smaller EOAs (0.61 cm2 versus 0.84 cm2). Differences in baseline characteristics may explain the discrepancy of results.
A second study, which also showed an increased long-term mortality in PPM patients after AVR, did not specify preoperative patient characteristics or causes of death.7 These data are therefore difficult to compare with our results.
Another recent study33 reported a negative impact of PPM on long-term outcome in a subgroup of patients with impaired left ventricular function undergoing AVR. In the present series, however, only 11 PPM patients (5.6%) presented with left ventricular dysfunction, therefore this subgroup appeared too small for separate analysis.
A recent paper, which evaluated the outcome in patients with small St Jude mechanical prostheses, reported a worse long-term outcome in PPM patients.6 However, it may be an important limitation of the whole concept to rely fully on Doppler estimates of EOA in bileaflet mechanical prostheses. It has been argued that reference values for EOAs should ideally be derived from in vivo measurements.16 32 Thus, the EOA reference values for the definition of PPM have been derived exclusively by echocardiography.17 This approach is acceptable in bioprostheses and in tilting disc mechanical valves as in these valves Doppler EOAs quite accurately reflect effective valve areas.39 40 In bileaflet mechanical valves, however, EOAs are significantly underestimated by Doppler echocardiography. This is due to localised high velocities between the leaflets, which do not reflect the mean velocity distribution across the orifice. We have previously demonstrated that the Doppler continuity equation underestimates orifice areas by approximately 50% in bileaflet valves.39 When Doppler data are used to define PPM in these prostheses significant overestimation of the presence of PPM must be expected. The percentage of mechanical valves in previous studies varies considerably from 14% to 54%.3 5 7 8 11 18 33 41 This may also explain discrepant results. In the present study the percentage of bileaflet prostheses was 24.8%. When we analysed our data after excluding these patients, there was still no difference in outcome between PPM and non-PPM patients.
Apart from the impact on long-term results, the influence of PPM on perioperative mortality is a matter of concern. Looking at the uncorrected data in the present study, operative mortality was markedly higher in the PPM group, although this difference did not reach statistical significance. However, this group had significantly less favourable baseline characteristics with significantly older age and more comorbidity. This is also reflected by a significantly higher EuroSCORE predicting a higher mortality. Although the difference in mortality between patient groups with and without PPM observed in the present study appears to be primarily due to differences in patient characteristics resulting in significantly higher operative risk, a minor impact of PPM on outcome cannot entirely be excluded. Previous investigators have indeed reported such results.5 10 41 Blais et al5 recently reported significantly elevated surgical mortality in 1260 PPM patients after AVR. Rao and coworkers10 showed a significant difference in perioperative survival rates between patients with and without prostheses ⩽21 mm. However, small prosthetic size was also not an independent predictor of outcome by multivariate analysis. In a retrospective analysis, Adams and coworkers41 showed an increased operative mortality in elderly men after implantation of 19-mm aortic valves.
The present study is dedicated to the subgroup of patients undergoing aortic valve replacement—namely, patients referred to surgery for isolated severe aortic stenosis. We are therefore not able to draw a general conclusion about the impact of moderate PPM in patients undergoing AVR for other reasons.
Every aortic prosthesis should be chosen with great care to guarantee adequate effective orifice area and haemodynamic performance. However, more complex surgical procedures to avoid PPM such as aortic root enlargement, considerably prolong aortic clamp time and cardiopulmonary bypass time,20 21 putting the patient at additional risk. It has been shown that aortic annulus enlargement increases operative mortality,21 and that a longer cross-clamp time in the setting of AVR is a risk factor for postoperative heart failure.36 Although it has been reported that aortic root enlargement can be carried out with acceptable results,20 this approach currently is not generally accepted in all patients at risk for moderate PPM.
The present data indicate that a more risky operative strategy may not be justified in patients undergoing AVR for isolated severe AS to prevent moderate PPM.
Moderate PPM as currently defined is a frequent finding in patients with aortic valve prostheses. From the presented data, it has no major impact on perioperative and long-term survival after valve replacement for isolated severe aortic stenosis. A recommendation of more complex surgical interventions such as aortic root enlargement to avoid moderate PPM may therefore not be justified. Furthermore, it may be a major limitation of this concept to rely fully on Doppler estimates of EOA, which have been shown to markedly underestimate EOA in bileaflet prostheses.
Competing interests: None.
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.