OUP user menu

Left and right ventricular function in aortic stenosis patients 8 weeks post-transcatheter aortic valve implantation or surgical aortic valve replacement

Lena M. Forsberg, Éva Tamás, Farkas Vánky, Niels Erik Nielsen, Jan Engvall, Eva Nylander
DOI: http://dx.doi.org/10.1093/ejechocard/jer085 603-611 First published online: 24 June 2011

Abstract

Aims Knowledge of longitudinal left and right ventricular (LV and RV) function after transcatheter aortic valve implantation (TAVI) is scarce. We hypothesized that the longitudinal systolic biventricular function in aortic stenosis (AS) patients is affected differently by TAVI and surgical aortic valve replacement (SAVR).

Methods and results Thirty-three AS patients (all-TAVI group, age 81 ± 9 years, 18 female), with EuroSCORE 18 ± 9%, were accepted for TAVI. Seventeen of these patients were matched (by gender, age, and LV function) to 17 patients undergoing SAVR. Conventional echocardiographic parameters, systolic atrioventricular plane displacement (AVPD) at standard sites and peak systolic velocity (PSV) by pulsed tissue Doppler at basal RV free wall, LV lateral wall, and septum were studied before and 8 weeks after the procedure. Procedural success was 100%, and 30-day mortality 9%. In all TAVI patients, AVPDlateral, PSVlateral, AVPDseptal, and PSVseptal increased (P< 0.001, 0.003, 0.006 and 0.002). When studying the matched patients postoperatively, both the SAVR and TAVI patients had increased PSVlateral and AVPDlateral (SAVR: P=0.03 and P=0.04, TAVI: P=0.04 and P=0.01). The PSVRV increased in the all-TAVI group (P=0.007), while the AVPDRV was unchanged. SAVR patients had decreased AVPDRV (P=0.001) and PSVRV (P=0.004), while the matched TAVI patients had unchanged RV function parameters.

Conclusion An improvement in regional longitudinal LV function in the septal and lateral wall could be seen after TAVI. Among the matched patients, both the TAVI and SAVR patients seemed to improve LV function in the lateral wall. RV systolic function increased in TAVI patients, but was impaired in the matched SAVR group at the 8-week follow-up.

  • Transcatheter aortic valve implantation
  • Aortic stenosis
  • Left ventricular function
  • Right ventricular function
  • Echocardiography

Introduction

Surgical aortic valve replacement (SAVR) is the treatment of choice in patients with symptomatic severe aortic stenosis (AS) since, following symptom onset and without SAVR, this patient group has an average survival of only 2 to 3 years.1 Some of these patients have not been offered conventional open heart surgery because of a high mortality risk rate due to age and comorbidities. Transcatheter aortic valve implantation (TAVI) has rapidly developed into a feasible alternative for treating these high-risk patients with severe AS.2

An abnormal longitudinal left ventricular (LV) function has been observed in AS patients before a decrease in LV ejection fraction (LVEF) has appeared.3 Pulsed tissue Doppler imaging (TDI) has proven to be informative in a variety of cardiac disorders, with the potential of identifying a reduced longitudinal LV function before more established LV function parameters, such as EF, exhibit deterioration.4,5

RV function assessed by tricuspid annular motion decreases after open heart surgery and cardiopulmonary bypass. Pericardial disruption and myocardial ischaemia have been hypothesized as explanations of this phenomenon.68

The aim of this study was to determine changes in LV and RV global and longitudinal function in patients undergoing TAVI assessed by echocardiography and pulsed TDI. By matching a group of AS patients undergoing conventional aortic valve surgery, we also aimed to study whether the two procedures affect biventricular function differently.

Method

Patients

TAVI group

Between September 2008 and January 2010, 33 patients with severe symptomatic AS were accepted for TAVI. These patients were assessed by a team of surgeons and cardiologists as not being candidates for conventional surgery, due to a high surgical risk or contraindications to AVR. Exclusion criteria were: an aortic annulus diameter of less than 18 mm, or more than 25 mm, or if life of a reasonable quality or duration of life was unlikely. Preprocedural investigations included transthoracic and transoesophageal echocardiography, coronary angiography, iliofemoral angiography, and CT scan of the aorta with 3D reconstruction. The transfemoral approach was the first choice of treatment and in the case of contraindications to this approach, the transapical approach was chosen.

SAVR patients

Seventeen AS patients, referred to our centre for SAVR, could be matched to 17 TAVI patients by gender, age (±10 years), and LV function.

Echocardiography

All patients were examined by echocardiography (Vivid 7 ultrasound system, GE Vingmed Ultrasound, Horten, Norway) 1 day before their procedure and were re-examined 8 weeks post-procedurally. Echocardiographic images were saved for offline analysis and parameters were measured as recommended in the appropriate guidelines.9,10 LVEF was visually estimated by an experienced investigator according to a four-grade scale where normal systolic function corresponds to EF >50%, slightly reduced EF 40–50%, moderately reduced EF 30–39%, and severely reduced <30%. LV mass was calculated by using the equation: 0.8 × (1.04 × [LVEDD + PWTd + SWTd]3 − LVEDD3)+0.6g and was indexed to body surface area (BSA). Left ventricular outflow tract (LVOT) dimension was measured from the parasternal long-axis view and LVOT area was calculated. Right and left atrial area was calculated by close outlining of the endocardial border in an apical four-chamber view.

Using Doppler echocardiography, peak aortic velocity, peak LVOT velocity, aortic and LVOT velocity time integral (VTIAo and VTILVOT), and mean pressure gradient (▵Pmean) were determined. The effective orifice area (EOA) was calculated from the continuity equation: EOA = (AreaLVOT×VTILVOT)/VTIAo, and was indexed to BSA (iEOA). The presence and degree of aortic regurgitation and mitral regurgitation (MR) were recorded in all patients. Post-procedurally, aortic regurgitation was further evaluated as paravalvular or transvalvular. In the case of tricuspid regurgitation (TR), the peak TR velocity was inserted into the Bernoulli equation and the pressure gradient between the right atrium and right ventricle was estimated.

Peak velocities of early (E) and late (A) diastolic filling, E/A ratio, isovolumic relaxation time (IVRT), deceleration time, and pulmonary venous systolic and diastolic flow velocity were derived from Doppler recordings of the mitral inflow at the mitral leaflet tips and the venous inflow of the right upper pulmonary vein near the orifice. These variables were used for classification of LV diastolic function into four groups: normal, impaired relaxation, pseudonormal, and restrictive.11 Patients with atrial fibrillation formed a fifth group.

Pulsed tissue Doppler imaging (TDI)

To estimate regional myocardial function, peak systolic (PSV), early diastolic (é), and late diastolic (á) myocardial velocity were measured in the LV septal, LV lateral, and RV free wall. The 6 × 6 mm sample volume was placed directly underneath the mitral or tricuspid annulus in the basal myocardium in an apical four-chamber view. The mean value of three beats was calculated.

Systolic atrioventricular plane displacement (AVPD), measured by M-mode echocardiography, was determined at the LV septal, LV lateral, and RV lateral annulus as previously described.12

TAVI procedure

The transfemoral and transapical procedures were performed in a catheterization laboratory under general anaesthesia and with guidance of transoesophageal echocardiography (TEE) and fluoroscopy. Cardiopulmonary bypass was not used. Edwards SAPIEN bioprostheses (Edwards Lifesciences; Inc, Irvine, California), 23 mm or 26 mm were implanted. The bioprostheses were deployed during rapid ventricular pacing.

Statistics

Continuous data are expressed as mean ± SD. All categorical variables are given as frequencies or percentages with the exception of the grade of aortic and MR which are reported as median (25th–75th). Pre- and postimplantation data within the entire TAVI group (all TAVI), and within and between the two matched groups, were analysed by Wilcoxon signed rank test. Pearson correlation coefficient was used to explore possible relations between RV long axis function and RA–RV pressure difference. A probability value of P≤0.05 was considered significant. Data analyses were performed with SPSS 16.0. All patients gave their written informed consent prior to their participation. The study complies with the Declaration of Helsinki and the study was approved by the Regional Ethical Review Board in Linköping.

Results

TAVI patients

Clinical characteristics

Baseline characteristics and co-morbidities of the 33 TAVI patients are presented in Table 1. Because of extensive aortic or iliofemoral vascular disease, transapical TAVI was the chosen procedure in 16 patients. The overall success rate was 100% for both the transapical and transfemoral procedure. The prosthesis size was 23 mm in 16 patients and 26 mm in 17. Early postoperative complications were: bleeding requiring transfusion (>2 units) in 16 patients, atrioventricular block in 2, acute renal failure in 4, pneumonia in 2, pericardial effusion in 3, bradycardia requiring pacemaker implantation in 1, and RV failure in 1. Before the follow-up, four patients died; one of multiorgan failure, one of intramyocardial haematoma, and two of non-cardiac reasons. Four patients were not studied postprocedurally because of administrative error,2 planned non-cardiac surgery,1 and stroke.1 Chronic atrial fibrillation was considered to be the most probable underlying cause of the stroke. In total, 25 patients were included in the follow-up 8 weeks after TAVI.

View this table:
Table 1

Baseline characteristics

Feature (n)All TAVI (n=33)Matched TAVI (n=17)Matched SAVR (n=17)
Age (years)81 ± 976 ± 872 ± 6
Female sex1899
Body mass index (kg/m2)25.7 ± 25.326.5 ± 6.226.3 ± 4.1
Body surface area (m2)1.8 ± 0.21.9 ± 0.21.9 ± 0.2
Logistic EuroSCORE (%)18 ± 917 ± 97 ± 5
STS score (%)19 ± 716 ± 611 ± 7
NYHA pre
 II229
 III28138
 IV320
Heart failure1383
Atrial fibrillation1160
Hypertension251312
Diabetes (insulin)5 (4)3 (2)4 (1)
Coronary artery disease151111
Peripheral vascular disease721
Stroke630
PAP >60 mmHg931
Chronic obstructive pulmonary disease110
Creatinine (µmol/L)119 ± 54108 ± 2993 ± 24
Previous heart surgery12101
Previous PCI6411
Chest radiation310
Porcelain aorta410
  • Data are presented as mean ± SD or frequencies. Syst, systolic; diast, diastolic; NYHA, New York Heart Association, STS Society of Thoracic Surgeons; PAP, pulmonary arterial pressure; PCI, percutaneus coronary intervention.

Aortic valve and prosthesis function

Preprocedurally, the iEOA in the all-TAVI group was 0.3 ± 0.1 cm2/m2. Mean pressure gradient, maximal aortic velocity, and VTIAo decreased significantly after TAVI (all P<0.001, Table 2). Postprocedurally, a paravalvular leakage was observed in 60% (n=15) and a transvalvular leakage in 36% (n=9) of the TAVI patients; however, the highest observed degree of postoperative prosthesis insufficiency was mild.

View this table:
Table 2

Echocardiographic variables

VariableAll TAVIP-valueaMatched TAVIP-valueaMatched SAVRP-valuea
Pre (n=33)Post (n=25)Pre (n=17)Post (n=17)Pre (n=17)Post (n=17)
Heart rate (bpm)66 ± 1169 ± 90.10168 ± 970 ± 100.40863 ± 669 ± 110.017
Systolic blood pressure (mmHg)146 ± 27150 ± 240.351135 ± 31140 ± 240.624145 ± 19138 ± 160.284
Diastolic blood pressure (mmHg)78 ± 1169 ± 130.03371 ± 1167 ± 120.10981 ± 1178 ± 90.306
Aortic valve
 EOA (cm2)0.5 ± 0.21.4 ± 0.3<0.0010.6 ± 0.11.5 ± 0.30.0010.7 ± 0.21.2 ± 0.20.007
 iEOA (cm2/m2)0.3 ± 0.10.8 ± 0.2<0.0010.3 ± 0.10.8 ± 0.20.0010.4 ± 0.10.6 ± 0.10.007
 ΔP mean (mmHg)60 ± 2111 ± 5<0.00153 ± 1511 ± 60.00147 ± 1912 ± 30.001
 Aortic velocitymax (m/s)4.7 ± 0.92.2 ± 0.6<0.0014.4 ± 0.72.0 ± 0.6<0.0014.5 ± 0.92.4 ± 0.3<0.001
 Aortic VTI (cm)124 ± 3047 ± 11<0.001113 ± 2345 ± 12<0.001111 ± 2948 ± 9<0.001
 LVOT VTI (cm)20 ± 522 ± 70.15319 ± 420 ± 60.69021 ± 520 ± 40.279
Left atrial area, s (cm2)28 ± 726 ± 70.04728 ± 827 ± 70.16821 ± 323 ± 50.138
Left ventricle
 Septal wall thickness, d (mm)13 ± 312 ± 30.04413 ± 312 ± 30.19213 ± 212 ± 30.100
 Posterior wall thickness, d (mm)13 ± 312 ± 20.58412 ± 312 ± 20.69212 ± 211 ± 20.106
 LV end-diastolic dimension (mm)51 ± 649 ± 70.16452 ± 650 ± 60.14049 ± 645 ± 50.115
 LV end-systolic dimension (mm)34 ± 732 ± 80.34735 ± 733 ± 80.10032 ± 833 ± 50.059
 Fractional shortening (%)33 ± 834 ± 80.73334 ± 835 ± 90.68335 ± 926 ± 60.009
 Indexed LV mass (g/m2)157 ± 55134 ± 380.036147 ± 54129 ± 320.551126 ± 3098 ± 240.041
Ejection fraction (n)
 >50%25210.08314141.00015130.564
 50–40%421102
 40–30%211112
 <30%101010
Aortic regurgitation (0–3)1 (0.5–1)0.5 (0.5–1)0.1711 (0.5–1)0.5 (0–1)0.2501 (0–1)0.5 (0–0.5)0.040
Mitral regurgitation (0–3)1 (0.5–1.5)1 (0.5–1)0.0241 (0.5–2)1 (0.5–1.0)0.1020.5 (0.5–1)0.5 (0.5–0.5)0.660
  • Data are expressed as mean ± SD, median (25th–75th percentile) or frequencies. D, diastolic; iEOA, indexed effective orifice area; LV, left ventricle; LVOT, left ventricular outflow tract; n, number; ▵P, pressure gradient; s, systolic; VTI, velocity time integral. Remaining abbreviations as in Table 1. Regurgitation scale; 0, none; 0.5, trivial; 1, mild; 2, moderate; 3, severe.

  • aPreoperatively vs. postoperatively.

Global LV function

Among the 25 TAVI patients who were included in the follow up, 19 (76 %) had an EF of 50 % or more before TAVI. At follow up, EF was unchanged but septal wall thickness and indexed LV mass had decreased significantly (Table 2).

Seven TAVI patients had a moderate MR preprocedurally. At follow up, there was a significant reduction in MR (P=0.024, Table 2).

Longitudinal LV function

When considering all TAVI patients, AVPDLateral (9 ± 3 vs. 11 ± 3 mm, P< 0.001) and AVPDSeptal had increased (6 ± 2 vs. 7 ± 3 mm, P=0.006) at follow up (Figure 1). In addition, the PSV had increased significantly in the lateral and the septal walls; 5 ± 2 vs. 6 ± 3 cm/s, P=0.003 and 4 ± 1 vs. 5 ± 2 cm/s, P=0.002, respectively.

Figure 1

Longitudinal left ventricular function evaluated as atrioventricular plane displacement (AVPD) and peak systolic velocity (PSV) in all transcatheter aortic valve implantation (TAVI, dark blue), matched TAVI (light blue), and SAVR (green) patients pre- and postoperatively. The box and whiskers represent values within the 25th–75th percentile and 1.5 of the interquartile range, respectively. Significant differences preoperatively vs. postoperatively within each group are presented as actual P-values. Matched SAVR group vs. matched TAVI group; P<0.05, ††P<0.01, †††P<0.001.

RV function

Postprocedurally, RV systolic and diastolic function measured by AVPD, é, and á were unchanged in the TAVI group (Table 3 and Figure 2), while PSVRV had increased (8 ± 2 vs. 10 ± 3 cm/s, P=0.007). The change in RV long-axis function was independent of the decrease in RA–RV pressure difference.

View this table:
Table 3

Right ventricular systolic and diastolic function

VariableAll TAVIMatched TAVIMatched SAVR
PrePostP-valueaPrePostP-valueaPrePostP-valuea
RV free wall
 AVPD (mm)15 ± 515 ± 40.61315 ± 614 ± 40.47620 ± 3*14 ± 30.001
 PSV RV (cm/s)8 ± 210 ± 30.0068 ± 29 ± 30.12011 ± 3**8 ± 20.004
 é RV (cm/s)10 ± 49 ± 40.9369 ± 49 ± 40.50810 ± 26 ± 20.005
 á RV (cm/s)13 ± 512 ± 60.45911 ± 610 ± 60.52617 ± 49 ± 40.001
 é/á0.8 ± 0.60.9 ± 0.80.9431.0 ± 0.81.1 ± 1.00.7210.6 ± 0.20.8 ± 0.40.133
RA area (cm2)20 ± 619 ± 50.07621 ± 520 ± 60.54317 ± 318 ± 30.124
RA–RV pressure difference (mmHg)44 ± 1333 ± 160.03041 ± 1232 ± 140.04033 ± 1028 ± 70.044
  • RV, right ventricle; AVPD, atrioventricular plane displacement; PSV, peak systolic velocity; é, early peak diastolic velocity; á, late peak diastolic velocity; RA, right atrial.

  • aPreoperatively vs. postoperatively,

  • *matched SAVR vs. matched TAVI group P<0.05.

  • **P<0.001.

Figure 2

Longitudinal right ventricular (RV) function evaluated as atrioventricular plane displacement (AVPD) and peak systolic velocity (PSV). Abbreviations and symbols as in Figure 1.

Diastolic LV function

The classification of LV diastolic function, showed in Figure 3, was unchanged postoperatively. By TDI parameters, higher é-septal, á-septal, and á-lateral velocities were observed postprocedurally (all P<0.03, Table 4). Moreover, there was a significant reduction in the average E/é ratio (27 ± 15 to 20 ± 9, P=0.001, Table 4). Also the pressure difference between the right atrium and ventricle, reflecting pulmonary artery pressure, had decreased at the follow up (P=0.03, Table 3).

View this table:
Table 4

Left ventricular diastolic variables by tissue Doppler imaging before and 8 weeks after TAVI and SAVR

VariableAll TAVIMatched TAVIMatched SAVR
PrePostP-valueaPrePostP-valueaPrePostP-valuea
Doppler echocardiography
 E (m/s)1.2 ± 0.41.1 ± 0.30.4501.1 ± 0.41.1 ± 0.30.6920.8 ± 0.2*0.8 ± 0.30.077
 A (m/s)0.8 ± 0.40.9 ± 0.30.1310.8 ± 0.40.9 ± 0.40.3800.8 ± 0.30.9 ± 0.30.504
 E/A (m/s)1.5 ± 0.61.4 ± 0.90.6171.3 ± 0.51.6 ± 1.00.5751.2 ± 0.81.0 ± 0.50.670
 Deceleration time (ms)194 ± 87211 ± 860.322192 ± 102206 ± 990.326284 ± 99*232 ± 890.055
 IVRT (ms)54 ± 2273 ± 350.01457 ± 2567 ± 370.12178 ± 25*77 ± 210.959
 PVs/PVd (m/s)1.0 ± 0.81.0 ± 0.90.9110.9 ± 0.70.9 ± 1.00.1911.3 ± 0.51.2 ± 0.50.142
Tissue Doppler imaging
 Septal
  é septal (cm/s)4.0 ± 2.05.2 ± 2.60.0284.9 ± 2.45.8 ± 2.60.1224.8 ± 1.65.6 ± 2.00.167
  á septal (cm/s)5.0 ± 2.05.9 ± 2.30.0156.0 ± 1.96.0 ± 2.60.3668.6 ± 2.2*6.9 ± 2.10.033
  é/á septal0.9 ± 0.50.8 ± 0.40.5320.8 ± 0.50.9 ± 0.50.1390.6 ± 0.20.9 ± 0.50.013
 Lateral
  é lateral (cm/s)6.6 ± 3.17.2 ± 2.60.1187.3 ± 3.27.6 ± 2.80.4185.9 ± 1.98.9 ± 3.10.005
  á lateral (cm/s)4.8 ± 2.56.6 ± 3.50.0065.1 ± 2.76.5 ± 3.40.0587.0 ± 3.28.0 ± 3.70.166
  é/á lateral1.4 ± 0.81.4 ± 1.00.5381.5 ± 0.81.5 ± 1.00.6101.3 ± 1.6*1.6 ± 1.50.116
  E/é average27 ± 1520 ± 90.00123 ± 1320 ± 90.03116 ± 513 ± 6*0.093
  E/é septal34 ± 1826 ± 150.04531 ± 1926 ± 240.23418 ± 618 ± 100.868
  E/é lateral24 ± 1518 ± 90.01320 ± 1218 ± 100.17715 ± 811 ± 4*0.016
  • E, early diastolic filling; A, late diastolic filling; IVRT, isovolumic relaxation time; PVs, systolic pulmonary venous flow velocity; PVd, diastolic pulmonary venous flow velocity; é, early diastolic myocardial velocity; á, late diastolic velocity. Remaining abbreviations as in Table 1.

  • aPreoperatively vs. postoperatively.

  • *Matched SAVR vs. matched TAVI group P<0.05.

  • **P<0.01.

Figure 3

Classification of LV diastolic function pre- and postprocedurally in all TAVI patients, matched TAVI patients, and matched SAVR patients. Neither TAVI- nor SAVR patients significantly changed classification postprocedurally. TAVI patients had a more severely disturbed diastolic function, both pre- and postprocedurally, than the SAVR group.

Left atrial area had decreased from 28 ± 7 to 26 ± 7 cm2, P=0.047 (Table 2). When analysing the change in left atrial area in patients with atrial fibrillation and sinus rhythm separately, a significant decline was only possible to detect in the atrial fibrillation group (P=0.047). However, in this group, a majority of the patients had a moderate mitral insufficiency.

The matched SAVR patients

Baseline characteristics are presented in Table 1. Ten patients had undergone combined SAVR and coronary artery bypass surgery and seven patients isolated AVR. There were perioperative complications in three patients: two needing a pacemaker implant due to a third-degree AV block and one need blood transfusion due to bleeding. The preoperative iEOA was 0.4 ± 0.1 cm2 and the postoperative haemodynamic function of the prosthetic aortic valves, evaluated according to guidelines, was as expected for the prostheses used (Table 2).

TAVI patients vs. SAVR patients: LV and RV function

Among the matched TAVI patients, there was a significantly increase in both AVPD and PSV in the LV lateral wall postprocedurally (9 ± 3 vs. 11 ± 3 mm, P=0.01 and 5 ± 2 vs. 6 ± 3 cm/s, P=0.04, Figure 1). Also the SAVR patients increased their LV lateral wall function estimated by AVPD and PSV (11 ± 2 vs. 13 ± 3 mm, P=0.04 and 6 ± 2 vs. 8 ± 3 cm/s, P=0.03). In both groups, the AVPDSeptal and PSVSeptal were unchanged. The septal long-axis function was significantly lower in the TAVI patients, when compared with the AVR patients both pre- and postprocedurally (Figure 1).

Postoperatively, the RV function was unchanged among the matched TAVI patients. The SAVR patients, however, decreased both in AVPD and PSV. Although the TAVI patients had lower AVPDRV and PSVRV than the SAVR patients preprocedurally, there was no difference between the two groups after the procedure (Figure 2).

With regard to the LV diastolic function classified after filling pattern, the TAVI patients had signs of more severely disturbed diastolic function, i.e. increased filling pressures, than SAVR patients, both pre- and postprocedurally (Figure 3). Also the average E/é was higher among the TAVI patients than the SAVR patients postprocedurally (20 ± 9 vs. 13 ± 6, P=0.041). E/é was unchanged in the SAVR group after surgery (P=0.093), while the TAVI patients decreased significantly (P=0.031).

Discussion

This prospective study reports a detailed evaluation of LV and RV function by means of echocardiography and TDI in patients who have undergone TAVI. The study also includes a group of SAVR patients matched to TAVI patients by gender, age, and LV function. Although the included number of matched SAVR patients is limited, these data allow a unique possibility to compare the two treatments regarding ventricular function.

Our main findings were those of increased longitudinal systolic LV and RV function postprocedurally among the TAVI patients. Both the matched SAVR and TAVI patient groups increased their LV lateral longitudinal function parameters, while the SAVR patients showed a decreased RV function postoperatively.

Concerning prosthesis function, the mean gradients were consistent with the findings from previous studies of both TAVI and SAVR.13,14

Global LV function

Following SAVR and TAVI, improvement in LV function, estimated by EF, has predominantly been seen in patient groups with preoperatively reduced or severely reduced EF.13,15,16 Ewe et al.16 showed an immediate increase in EF in TAVI patients with preoperative reduced EF, while TAVI patients with normal LV function were unchanged. At follow-up, patients with preoperative normal EF continued to be unchanged while there was a further improvement of EF in the reduced EF patient group. Because the majority of our patients had normal EF preoperatively, unchanged EF was expected in both patient groups.

Longitudinal LV function

An impairment in long-axis function can be seen early in the natural history of AS patients with normal EF.3 The long-axis excursion is due mainly to subendocardial myocytes.17 These cells are considered more vulnerable to ischaemia and stress than cells of the intermediate layers of the myocardium.3,18 After SAVR, an improved longitudinal function in the lateral wall has been reported.19 In this study, we observed an improvement of LV function shown by increased AVPD and PSV in the lateral and septal wall 8 weeks after TAVI. These findings are supported by earlier studies which have shown increased PSV 24h and increased AVPD 6 weeks after TAVI.20,21 The most probable explanation of the absence of an increase in AVPDSeptal and PSVSeptal in the SAVR patients would be paradoxical septal motion (PSM), which is often seen after open heart surgery and has been proposed as being more frequent after valve surgery.22 Although the aetiology of PSM is debated, the most accepted explanation is that PSM is an artefact caused by postoperative sternal-cardiac adhesions.23

Right ventricular function

Longitudinal RV function, measured by TDI and M-mode, has shown a correlation with “gold standard” measurements of RVEF.24 The SAVR patients included in this study had decreased AVPD, PSV, é, and á in the free wall of the RV at follow-up, implying RV dysfunction, while the TAVI patients showed an improved or unaltered longitudinal RV function by the different parameters used. These findings are supported by the results from Zhao et al.,21 which showed a decreased AVPD in the RV free wall 1 week after SAVR but unchanged RV AVPD after TAVI. Furthermore, in our study, with a matched SAVR group, we could show that despite a superior RV function before surgery in the SAVR group, the RV function parameters did not differ between the groups at follow-up. The aetiology of RV dysfunction after open heart surgery is unknown but intraoperative ischaemia, myocardial damage, and pericardial disruption have been suggested as underlying mechanisms.68 However, an altered pattern of RV contraction and, consequently, a mechanical explanation of the decreased longitudinal RV function postoperatively have also been proposed.7,25 In line with this, TAVI would, due to its less invasive nature, probably cause both less intraoperative trauma to the RV and less postoperative pericardial adhesions and influence on cardiac motion. However, the exact mechanism to preservation of RV function after TAVI but not SAVR is unknown. These results have their clinical relevance by expanding our understanding of RV function after aortic valve intervention.

LV diastolic function

Short-term follow-up of diastolic function after TAVI has not been extensively studied earlier.

Although the LV diastolic classification based on filling patterns remained unchanged in the TAVI patients postprocedurally, the decrease in the left atrial area and E/é could be interpreted as a reduction in LV filling pressure; however, the latter continued to be elevated postprocedurally.

Among patients undergoing SAVR, LV diastolic function estimated by TDI variables have demonstrated prognostic value.26 The TAVI group had increased é-velocity of the septal LV wall postoperatively and this pattern has been shown earlier after AVR by Giorgi et al.19 The significant increase in LV é- and á-velocity and decreased E/é among the TAVI patients in our study may imply an early recovery of LV diastolic function. Bauer et al.20 found an increase in é and á measured by pulsed TDI in the LV septal wall but no difference in LV diastolic function measured by blood flow Doppler echocardiography when they examined TAVI patients 24 h postprocedurally. Gotzmann et al.,27 who studied the CoreValve prosthesis, were only able to establish a significant change in é-velocity and decrease in E/é at the 6-month follow-up.

Our data suggest a more pronounced elevation of postoperative LV filling pressures in TAVI when compared with SAVR patients, despite a comparable EF. One reason could be that the TAVI patients were, at first, not accepted for SAVR due to co-morbidities and therefore had a delay in receiving aortic valve intervention. Normalization of diastolic dysfunction after correction of AS takes years, and further follow-up of data on LV diastolic function and filling pressures after TAVI is therefore of interest.14

Limitations

A majority of the SAVR patients had concomitant coronary bypass surgery and long-axis function could have been affected by their coronary disease. On the other hand, an equal number of TAVI patients also had coronary disease and previous heart surgery in these patients may also have influenced their longitudinal function.

Conclusions

An improvement in LV systolic and diastolic function assessed by echocardiography and TDI could be seen already 8 weeks after TAVI. Although the matched TAVI patients had significantly lower PSVSeptal and PSVLateral than did the SAVR patients, a similar reaction with increased LV longitudinal lateral function could be observed in both groups postprocedurally. Postoperatively, RV function assessed by tricuspid annular motion and PSV was unaffected or improved in the TAVI patients: a more favourable alteration of RV function than the SAVR patients had. We conclude that patients with severe AS and a high surgical risk profile have a favourable change in longitudinal LV and RV function 8 weeks after TAVI.

Funding

This work was supported by the Swedish Heart Lung Foundation, Medical Research Council of Southeast Sweden and ALF Grants from the County Council of Östergötland, Sweden.

Acknowledgements

The authors gratefully acknowledge Inger Huljebrant, research nurse, and Mirjam Johansson, medical technician, for their valuable assistance in patient inclusion and performing the echocardiographic studies.

Conflict of interest: none declared.

References

View Abstract