Author + information
- Received October 21, 1998
- Revision received May 27, 1999
- Accepted June 30, 1999
- Published online November 1, 1999.
- Philippe Pibarot, DVM, PhD, FACC∗,†,* (, )
- Jean G. Dumesnil, MD, FRCPC, FACC∗,
- Jean Jobin, PhD, FACSM∗,
- Paul Cartier, MD, FRCSC, FACS∗,†,‡,
- George Honos, MD, FRCPC, FACC‡ and
- Louis-Gilles Durand, PhD, Eng†
- ↵*Reprint requests and correspondence: Dr. Pibarot, Quebec Heart Institute/Laval Hospital, 2725 Chemin Ste-Foy, Ste-Foy G1V-4G5, Quebec, Canada
The objective of this study was to compare stentless bioprostheses with stented bioprostheses with regard to their hemodynamic behavior during exercise.
Stentless aortic bioprostheses have better hemodynamic performances at rest than stented bioprostheses, but very few comparisons were performed during exercise.
Thirty-eight patients with normally functioning stentless (n = 19) or stented (n = 19) bioprostheses were submitted to a maximal ramp upright bicycle exercise test. Valve effective orifice area and mean transvalvular pressure gradient at rest and during peak exercise were successfully measured using Doppler echocardiography in 30 of the 38 patients.
At peak exercise, the mean gradient increased significantly less in stentless than in stented bioprostheses (+5 ± 3 vs. +12 ± 8 mm Hg; p = 0.002) despite similar increases in mean flow rates (+137 ± 58 vs. +125 ± 65 ml/s; p = 0.58); valve area also increased but with no significant difference between groups. Despite this hemodynamic difference, exercise capacity was not significantly different, but left ventricular (LV) mass and function were closer to normal in stentless bioprostheses. Overall, there was a strong inverse relation between the mean gradient during peak exercise and the indexed valve area at rest (r = 0.90).
Hemodynamics during exercise are better in stentless than stented bioprostheses due to the larger resting indexed valve area of stentless bioprostheses. This is associated with beneficial effects with regard to LV mass and function. The relation found between the resting indexed valve area and the gradient during exercise can be used to project the hemodynamic behavior of these bioprostheses at the time of operation. It should thus be useful to select the optimal prosthesis given the patient’s body surface area and level of physical activity.
Recent studies (1,2)showed that stentless aortic bioprostheses have better hemodynamic performance at rest than their stented counterparts, and this is believed to be related to the stentless design that provides a larger effective orifice area (EOA) for the same prosthesis outer diameter size. Also, our team recently showed that during maximal exercise in stented bioprostheses, the increase in gradient is mainly determined by the indexed EOA at rest (3). Therefore, we hypothesized that stentless bioprostheses would have less increase in gradient during exercise in comparison with stented bioprostheses due to their larger resting indexed EOAs. We also hypothesized that the stentless bioprostheses might have the added advantage of increasing their EOA during exercise to a relatively greater extent than stented bioprostheses due to a possible expansion of the prosthesis valvular annulus. To our knowledge, there have been few studies of the hemodynamic behavior of stentless valves during exercise (4,5).
The objective of this study was, therefore, to compare the hemodynamic behavior of stentless and stented aortic bioprostheses during maximal exercise in comparable groups of patients. Because the hemodynamic behavior of the prosthesis could also influence the remodeling of the left ventricle (LV) after valve replacement, a secondary objective of the study was to compare the parameters of LV morphology and function in these two groups of patients.
Fifty patients, 11 female and 39 male, with either a stented or a stentless bioprosthetic valve in the aortic position were randomly selected for the study. Of the 27 patients included in the stented valve group, 22 received a Medtronic Intact valve (Medtronic, Minneapolis, Minnesota), 2 received a Medtronic Mosaic valve, 2 received a Carpentier-Edwards pericardial valve (Baxter, Irvine, California) and 1 received a St-Jude X-Cell valve (St-Jude Medical, St. Paul, Minnesota). All the patients (n = 23) in the stentless valve group received a Medtronic Freestyle valve. The valve was implanted using the subcoronary technique in 22 patients and the root replacement technique in one patient. A complete Doppler echocardiographic evaluation was first performed at rest with the patient in the supine position to verify that all patients had a normal LV systolic function (ejection fraction >50%) and no evidence of prosthetic valve dysfunction. The protocol was approved by the authors’ institutional review board, and all subjects gave informed written consent. Some of the data from the patients with a stented valve have been reported in a previous study (3).
Echocardiographic assessment of LV morphology and function at rest
Doppler echocardiographic examinations were performed with a Hewlett Packard Sonos 2000 Ultrasound system (Hewlett Packard, Andover, Massachusetts). From a parasternal LV long-axis view, standard recordings of two-dimensionally directed LV M-mode echocardiograms were obtained at rest with the patient in the left lateral decubitus position. Left ventricular minor axis internal dimension, posterior wall and septal thickness were measured at end-diastole and end-systole. Left ventricular mass was calculated using the corrected formula of the American Society of Echocardiography and was indexed for body surface area (6). The shortening fraction of the internal radius, the mid-wall radius and the length of the LV, as well as the fraction and velocity of the posterior wall thickening, were determined using the model proposed by Dumesnil et al. (7). The ejection fraction was calculated using the mid-wall radius and the longitudinal shortening fractions as previously described (8).
The peak systolic wall stress was estimated by the method of Grossman et al. (9): where LVIDs and LVPWTs are the internal dimension and the posterior wall thickness of the LV at end-systole, respectively, and LVPs is the approximated LV systolic pressure, calculated by adding the systolic blood pressure and the peak transprosthetic gradient obtained from continuous wave Doppler recordings of the aortic jet velocity.
Patients were submitted to a maximal ramp upright bicycle exercise test beginning with an initial workload of 0 W, followed by continuous increments of 10 W/min. A 12-lead electrocardiogram was continuously recorded and blood pressure was measured every 2 min. The patients were encouraged to exercise until exhaustion or the appearance of symptoms. The test was stopped if there was an abnormal rise in blood pressure (diastolic pressure >110 mm Hg), ECG evidence of ischemia (>1 mm horizontal or downsloping S-T depression 0.08 s after the J point compared with the resting ECG) or significant arrhythmia. Respiratory gas analysis was carried out with a Q-Plex 1 system (Quinton Instrument Co., Bothel, Washington) to measure oxygen consumption (VO2) and CO2production. Results were averaged over five breaths. Peak oxygen consumption was defined as the highest value of VO2obtained at the end of the test. The percentage of the predicted peak VO2for a given patient’s age and gender was calculated from the mean values of peak VO2in the normal population (10). The anaerobic threshold was determined by two experienced observers using the V-slope method (11). Patients were considered to have reached their cardiorespiratory maximum if the respiratory exchange ratio was ≥1.10 or if it was ≥1.05 and the VO2had leveled off at the end of the test (12).
Doppler echocardiographic measurements during exercise
Doppler echocardiographic examinations were performed with the patient sitting on the bicycle: 1) at rest, and 2) at peak of maximal exercise just before stopping the test or within 2 min after test termination. Measurements performed at rest in the upright position included the transvalvular flow velocity using continuous wave Doppler, LV outflow tract (LVOT) flow velocity using pulsed-wave Doppler and LVOT diameter, as previously described (1,13,14). The same measurements were performed during exercise except for LVOT diameter which was assumed to have remained constant (15). The velocity recording was first performed in the transvalvular jet and then in the LVOT. From these measurements, we calculated the LV stroke volume from the product of the LVOT velocity-time integral and cross-sectional area, the cardiac output from the product of stroke volume and heart rate, the mean transvalvular flow rate from the quotient of stroke volume and systolic ejection time, the mean transvalvular pressure gradient using the modified Bernoulli equation with inclusion of subvalvular velocity and the EOA using the standard continuity equation. If the difference in heart rate between the time of transvalvular velocity recording and that of LVOT velocity recording was >5%, the data was rejected to avoid potential errors in EOA and mean gradient due to changing hemodynamics in the early postexercise recovery period. The EOA was then indexed for body surface area.
Data and statistical analysis
For each patient, the change in valve EOA and mean gradient per change in transvalvular flow rate during exercise was calculated to assess and compare the flow dependence of these parameters in stentless bioprostheses versus in stented bioprostheses (16–18).
Data were expressed as mean ± SD and compared using a two-way analysis of variance for repeated measures followed by the Tukey test to evaluate the effects of exercise and bioprosthesis type (stentless vs. stented). If the normality test or the equality of variance test failed, the analysis of variance was performed on the logarithmic transform of the data. Statistical analysis of the association of variables was performed with the Pearson correlation coefficient or the determination coefficient adjusted for degrees of freedom when the relation was linear or nonlinear, respectively. Graphs were constructed with the corresponding regression equation using a curve-fitting software (Table Curve, Jandel Scientific, San Rafael, California). Values of p < 0.05 were considered significant.
A forward stepwise regression analysis was performed to identify factors that significantly influenced the change in valve EOA during exercise. The relevant variables tested for this analysis were: patient age and gender, prosthesis size, type of bioprosthesis, time interval since prosthesis implantation, mean flow rate, EOA and indexed EOA at rest and changes in mean flow rate and systolic blood pressure with exercise.
Of the 50 patients recruited for the study, 38 patients achieved maximal exercise. The exercise test was stopped prematurely because of excessive blood pressure rise in seven patients and ST-segment depression in one patient. In addition, one patient stopped because he did not tolerate the mouth piece and three patients stopped because of claudication. Transvalvular flow velocity could be adequately measured in 37 of the 38 patients who completed the maximal exercise test, whereas it was possible to measure LVOT flow velocity (and therefore EOA and mean gradient) in 30 of the 38 patients.
Patients clinical characteristics
Table 1shows the preoperative and operative data for the 38 patients who completed the maximal exercise study. These data show no significant difference between the two groups and as well, the patients operated on for dominant aortic stenosis had no difference in preoperative mean gradient (50 ± 21 vs. 47 ± 18 mm Hg) or indexed EOA (0.36 ± 0.07 vs. 0.37 ± 0.07 cm2/m2), and the patients operated on for dominant aortic regurgitation or mixed dysfunction had no difference in the severity of regurgitation.
Left ventricular morphology and function
Table 2shows that, at the time of the exercise study, several parameters of LV morphology and function were significantly better in the patients with a stentless bioprosthesis as compared with the patients with a stented prosthesis. Moreover, on the basis of the criteria (LV mass index >110 g/m2for women and >134 g/m2for men) proposed by Devereux et al. (19), 1 of 19 (5%) patients with a stentless bioprosthesis had LV hypertrophy compared with 6 of 19 (32%) of patients with a stented bioprosthesis.
Maximal exercise capacity
Maximal exercise capacity as estimated by maximal workload, peak VO2or anaerobic threshold did not significantly differ between patients with a stentless bioprosthesis and those with a stented bioprosthesis (Table 3).
Validation of echocardiographic measurements
There was a strong correlation and agreement between the resting EOA measured in the supine position and that measured in the upright sitting position: EOA upright = 0.10 + (0.95 × EOA supine), r = 0.97, standard error of estimate (SEE) = ±0.14 cm2, p < 0.001). There also was a very good correlation (r = 0.83, SEE = ±1.19 L/min, p < 0.001) between VO2and the cardiac output measured by Doppler echocardiography at peak exercise.
Baseline Doppler echocardiographic data (Table 4)
In the upright rest position, patients with a stentless bioprosthesis had similar cardiac index and mean transvalvular flow rate but higher valve EOA and lower mean gradient when compared with patients with a stented bioprosthesis. Despite a similar indexed prosthesis size in both groups (Table 1), the indexed EOA was significantly higher in stentless bioprostheses, thus reflecting their superior hemodynamic performance at rest.
Changes in Doppler echocardiographic data with maximal exercise (all patients)
On average, cardiac index increased by 123 ± 41% (+3.11 ± 0.89 L/min/m2, p < 0.001) and mean flow rate increased by 53 ± 24% (+132 ± 60 ml/s, p < 0.001), whereas EOA increased by 10 ± 13% (+0.21 ± 0.27 cm2, p < 0.001) and mean gradient increased by 96 ± 53% (+8 ± 7 mm Hg, p < 0.001) with maximal exercise.
Comparison of stentless versus stented bioprostheses
Despite similar increase in cardiac index and mean flow rate with exercise, the patients with a stentless bioprosthesis had significantly (p = 0.002) less increase in mean gradient (Table 4). The changes in gradient per change in flow rate were also significantly lower in stentless bioprostheses (+4 ± 4 vs. +13 ± 10 mm Hg/100 mls−1; p = 0.002).
The EOA increased significantly with maximal exercise in both groups and there was no significant difference between the groups for the change in EOA with maximal exercise both in absolute (Table 4)and relative terms (stented: 9 ± 13% versus stentless: 11 ± 14%, p = 0.56). There also was no significant difference between stentless and stented bioprostheses in regard to the change in EOA per change in flow rate (+0.15 ± 0.15 vs. +0.08 ± 0.18 cm2/100 mls−1; p = 0.30). In multivariate analysis, the change in mean flow rate with exercise was the only independent determinant of the change in EOA during exercise.
Dependence of exercise transvalvular gradients on resting indexed EOA
A strong inverse exponential relation was found in all patients between the mean gradient either at rest or during peak exercise and the resting indexed EOA (Fig. 1, Panels A and B). As previously found in our study of patients with a stented valve (3), there also was a strong inverse correlation between the increase in gradient with maximal exercise and the resting indexed EOA (Fig. 2). However, in the previous study the relation was linear whereas in this study, that included a group of patients with a wider range of indexed EOA (0.56–1.75 cm2/m2vs. 0.56–0.98 cm2/m2), this relation was exponential. The analysis of Figures 1 and 2shows that patients with a stented or a stentless bioprosthesis were distributed on different sections of the same exponential curves. Most patients with a stented bioprosthesis had an indexed EOA ≤0.85–0.90 cm2/m2and were therefore on the steep portion of the exponential curves. Consequently, in these patients the gradient was relatively high at rest (Fig. 1, Panel A) and increased markedly with maximal exercise (Fig. 2). In contrast, most patients with a stentless bioprosthesis had a larger indexed EOA at rest and were, therefore, on the flat portion of the curves where the resting gradient was low and little increase in gradient occurred during exercise.
Comparison with previous studies
Several studies in patients with an aortic bioprosthesis have reported a significant increase in mean gradient during exercise, but important variations were observed among patients, depending on valve model and size (4,5,20–23). To our knowledge, this is the first study that compared the hemodynamic performance in patients with stented versus stentless bioprostheses during maximal exercise. Previous studies in similar bioprostheses were performed during submaximal exercise, and, in most cases, the EOA was not measured.
Hemodynamic performance of stentless versus stented bioprostheses
This study definitely shows that stentless bioprostheses have less increase in gradient during maximal exercise than stented bioprostheses, despite similar increases in mean flow rate. In both cases, the EOA also increased during exercise thus minimizing the increase in gradient that could have occurred had the EOA remained constant. In Figure 3, the solid lines are the relations between the indexed EOA and the gradients observed at rest and during exercise and the dashed line represents the gradients that would have been observed during exercise given similar increases in mean flow rates and constant EOAs. The reduction in gradient due to the increase in EOA is in the order of 24%. To explain the apparent discrepancy in the order of magnitude between the increase in EOA (+10 ± 13%) and the reduction in gradient (−24 ± 10%), one must remember that gradient is a squared function of flow and EOA.
However, this study did not confirm our hypothesis stating that stentless bioprostheses might have a greater intrinsic capacity for EOA enlargement due to the flexibility of their annulus. The Freestyle bioprosthesis is surrounded by a polyester covering which might limit its potential to expand with exercise.
Given these observations, the superior hemodynamic performance of the stentless bioprostheses appears mainly to be related to a question of sizing, i.e., to their larger EOA in relation to the patient’s body surface area resulting in a larger EOA and a lower gradient at any given flow level. This conclusion is strongly supported by the relations given in Figures 1 and 2. A larger indexed EOA can be due either to a larger EOA in relation to the prosthesis outer diameter size or to the fact that, for the stentless bioprostheses, a larger prosthesis can be inserted into a smaller patient, as previously reported for the Freestyle and Toronto SPV valves (St. Jude Medical, St. Paul, Minnesota) (1,24,25). In this study, however, the indexed EOAs were much larger in the stentless bioprostheses despite a similar average valve size indexed for body surface area, suggesting that the main factor in these patients was a larger EOA in relation to the prosthesis outer diameter size.
Influence on exercise capacity and LV morphology
Although stentless bioprostheses had better hemodynamics than stented bioprostheses at rest and during exercise, there was no difference in maximal exercise capacity between the two groups. For all practical purposes, even in the patients with the smallest indexed EOAs, the resting and exercise gradients were similar to those observed in mild to moderate aortic stenosis (18)and probably not high enough to affect maximal exercise capacity.
Furthermore, our results suggest that the lesser gradients observed at rest and during exercise in stentless bioprostheses might have a beneficial effect on LV mass and function. Indeed, despite similar baseline characteristics, the patients with a stentless bioprosthesis had a significantly smaller LV mass than the patients with a stented bioprosthesis. This result is consistent with the previous study of Jin et al. (2)showing that stentless bioprostheses are associated with greater decrease in LV wall stress and better regression of LV hypertrophy than stented prostheses. The very significant differences in posterior wall thickening and ventricular longitudinal shortening also suggest a better recuperation of subendocardial myocardial function in the patients with a stentless bioprosthesis (7), which is also consistent with the lower residual systolic wall stress found in these patients. Further longitudinal studies are necessary to confirm these findings as well as to determine their potential impact on long-term morbidity and mortality.
Since Doppler echocardiographic measurements were performed with the patient in the upright position and during exercise, one cannot exclude completely that the true maximal transvalvular velocities were not always recorded. However, the very good correlation (r = 0.90; p < 0.0001) obtained between the gradient during maximal exercise and the indexed EOA at rest tends to confirm a certain coherence in the results. The recordings of the LVOT velocities and the measurement of the LVOT diameter, which determine stroke volume, cardiac output and EOA, can be affected by experimental conditions. However, the strong correlation that we obtained between the cardiac output measured at peak exercise by Doppler echocardiography and peak VO2supports the validity of these measurements. Although Doppler echocardiographic recordings during exercise, particularly those of LVOT velocities, are technically demanding, this study shows that they are feasible and reliable.
Most of the stented bioprostheses were Medtronic Intact valves, whereas all the stentless bioprostheses were Medtronic Freestyle bioprostheses. Other bioprosthetic or mechanical valves may demonstrate different changes in EOA with exercise and therefore different increases in gradient given the same indexed EOA at rest. The power to detect differences in EOA behavior during exercise is also limited by the low number of patients in each group, but the results nonetheless suggest that the change in gradient with exercise is much more related to the resting EOA rather than to its relative change during exercise.
Also, this study was carried out in a relatively homogenous group of patients, i.e., sedentary patients between 55 and 75 years old with a normal LV systolic function. In a more heterogeneous group of patients with regard to patient age, LV function and physical training conditions, the correlation between mean gradient and the resting indexed EOA might have been lower than that observed in this study. For instance, given the same indexed EOA at rest, young athletes might have had a higher gradient than predicted at peak exercise due to a supranormal cardiac output, whereas inversely, older patients with depressed LV function might have had lower gradients.
These findings further confirm the importance of valve sizing with regard to the hemodynamic behavior of a prosthesis. Indeed, our results show that, regardless of the type of bioprosthesis, the resting indexed EOA is the best predictor of postoperative gradients whether at rest or during exercise. Furthermore, the relations found in this study can be used to project postoperative gradients at the time of operation and should thus be useful to select the optimal prosthesis given a patient’s body surface area and level of physical activity. Figure 4illustrates the relations between the resting and exercise gradients observed in this study and the indexed EOAs calculated from the patient’s body surface area and the published normal values for the prosthesis being implanted (1), a calculation that can be made in the operating room. The correlations are slightly less significant than those shown in Figure 1, probably due to differences in prosthesis manufacturing and other unknown factors. Nonetheless, they show that these postoperative gradients could have been predicted from information available at the time of operation. They also further confirm that the projected indexed EOA should ideally be no less than 0.90 cm2/m2in order to minimize postoperative gradients. It should be emphasized that the normal values for the EOA of stented bioprostheses were taken from in vitro data whereas for the stentless bioprostheses, these values were taken from in vivo data because it has been shown that, contrary to stented bioprostheses, in vitro data for stentless bioprostheses grossly overestimate in vivo EOAs (1).
Stentless bioprostheses have a better hemodynamic performance than stented bioprostheses during exercise mostly because of their larger resting EOA rather than a greater intrinsic capacity for EOA enlargement. This is associated with beneficial effects with regard to LV mass and function. The strong exponential relations found between the resting indexed valve EOA and the gradients at rest and during exercise further confirm that the postoperative hemodynamic behavior of a bioprosthesis can largely be predicted from this parameter at the time of operation. These relations should therefore be useful to select the optimal valve given a patient’s body surface area and level of physical activity.
We thank Louise Côté, RN, Jacinthe Aubé, RN and Desneiges Beauvais, RN, for the recruitment of the patients, and we thank Jocelyn Beauchemin, RT, Madeleine Cloutier, RT, France Gilbert, RT, Louise Marois, RT, Marie Josée Blais, RT, Carole Daoust, RT, Denis Boucher, PhD, Joakim Allaire, MS, Lise-Hélène Caron, MS and Isabelle Laforest, MS, for their technical assistance.
☆ This study was supported by grants from the Medical Research Council of Canada, Ottawa (MT-10929), the Heart and Stroke Foundation of Canada, Ottawa, the Fonds de recherche en santé du Québec, Montreal, Canada and Medtronics Company. Dr. Pibarot is the recipient of the Senior Personal Award from the Heart and Stroke Foundation of Canada.
- effective orifice area
- left ventricular or ventricle
- left ventricular outflow tract
- oxygen consumption
- Received October 21, 1998.
- Revision received May 27, 1999.
- Accepted June 30, 1999.
- American College of Cardiology
- Dumesnil J.G.,
- LeBlanc M.H.,
- Cartier P.,
- et al.
- ↵Pibarot P, Dumesnil JG, Jobin J, et al. Usefulness of the indexed effective orifice area at rest in predicting an increase in gradient during maximal exercise in patients with a bioprosthesis in the aortic valve position [In press]. Am J Cardiol 1999;83.
- Dumesnil J.G.,
- Shoucri R.M.,
- Laurenceau J.L.,
- Turcot J.
- Dumesnil J.G.,
- Shoucri R.M.
- Mahler D.A.
- Dumesnil J.G.,
- Honos G.N.,
- Lemieux M.,
- Beauchemin J.
- Rassi A.,
- Crawford M.H.,
- Richards K.L.,
- Miller J.F.
- Shively B.K.,
- Charlton G.A.,
- Crawford M.H.,
- Chaney R.K.
- Takeda S.,
- Rimington H.M.,
- Chambers J.B.
- Burwash I.G.,
- Pearlman A.S.,
- Kraft C.D.,
- et al.
- Devereux R.B.