Authors: Davide Gabbieri, Italo Ghidoni, Greta Mascheroni, Matteo Chiarabelli, Giuseppe D’Anniballe, Paolo Pisi, Marco Meli, Clorinda Labia, Alice Barbieri, Francesca Spina, Federico Giorgi
Categories: Clinical, TAVI, SAVR, Permanent pacemaker implantation, Leaflet, LVOT, DLZ, Calcifications
Source: Heart Rhythm O2
Authors: Davide Gabbieri, Italo Ghidoni, Greta Mascheroni, Matteo Chiarabelli, Giuseppe D’Anniballe, Paolo Pisi, Marco Meli, Clorinda Labia, Alice Barbieri, Francesca Spina, Federico Giorgi
Permanent pacemaker implantation (PPI) after transcatheter aortic valve implantation (TAVI) remains a significant challenge despite new-generation devices.
This study aimed to identify predictors of PPI during preoperative evaluation and assess its mid-term impact in a single-center, real-world setting.
From 2010 to 2020, 1486 patients with aortic stenosis underwent surgical aortic valve replacement or TAVI using balloon-expandable transcatheter heart valves. The PPI rate was estimated using the hospital discharge record for each patient in the Emilia-Romagna region.
The 30-day incidence of PPI in the TAVI group was 3.4%. While 30-day PPI did not affect overall survival (log-rank, P = .494 NS), it predicted rehospitalization due to cardiac causes (hazard ratio 10.28; 95% confidence interval [CI] 95% 3.41–31.00; P <.001). Calcifications in the leaflet (odds ratio [OR] 4.66; 95% CI 1.41–15.47; P = .012), left ventricular outflow tract (OR 4.51; 95% CI 1.48–13.76; P = .008), and device landing zone (OR 2.52; 9% CI 0.86–7.40; P = .093) were associated with a higher risk of PPI.
A low 30-day PPI incidence was observed, primarily because of the exclusive use of balloon-expandable SAPIEN transcatheter heart valves and high implantation techniques. Baseline factors such as leaflet, left ventricular outflow tract, and device landing zone calcifications, as well as right bundle branch block, highlight the need for comprehensive preoperative analysis to reduce PPI incidence and mitigate its associated longer hospital stays and rehospitalizations due to cardiac causes.
Key Findings ▪Balloon-expandable SAPIEN (Edwards Lifesciences, Irvine, CA) transcatheter heart valves have resulted in a low 30-day permanent pacemaker implantation (PPI) rate of 3.4%. In addition, performing a high implantation technique has been confirmed to reduce the risk of PPI.▪PPI has been strongly associated with rehospitalizations due to cardiac causes during follow-up; its influence on overall survival has not been reported; and it might require further investigation.▪Preoperative computed tomography scan factors, such as leaflet, left ventricular outflow tract, and device landing zone calcifications quantified using a quantitative method, have emerged as predictive for PPI.
Over the last decade, transcatheter aortic valve implantation (TAVI) has revolutionized the management of symptomatic severe aortic valve stenosis (AS). Initial TAVI trials demonstrated its effectiveness in inoperable and high-risk patients. Subsequently, excellent results in intermediate- and low-risk patients have broadened TAVI indications. The current European and American guidelines recommend both transfemoral TAVI and surgical aortic valve replacement (SAVR) as class I treatments for most patients with severe symptomatic AS.^1^^,^^2^
However, from the outset, it became apparent that new-onset conduction disturbances were more frequent with TAVI than with surgery, making postprocedural pacemaker (PM) implantation the “Achilles’ heel” of the TAVI procedure.^3^ The development of new-generation devices has significantly reduced this gap, though it still exists.^4^ Available data are insufficient to determine whether PM implantation affects patients’ mid- and long-term outcomes.^5^^,^^6^ As TAVI use expands to younger and low-risk patients with longer life expectancies, understanding the impact of permanent PM implantation (PPI) after TAVI becomes increasingly important. Recently, some authors have argued that early PM implantation after TAVI may be unnecessary, showing a low rate of PM stimulation during follow-up.^7^
In this retrospective study, we analyzed mid- and long-term mortality after SAVR and TAVI with balloon-expandable (BE) technology. In addition, we investigated the anatomical and echocardiographic predictors of 30-day PM implantation after TAVI and the influence of PM implantation on patient outcomes.
In Hesperia Hospital Modena, Italy, 1486 consecutive patients with native AS underwent SAVR with all types of commercially available biological and mechanical prostheses or TAVI with BE SAPIEN (Edwards Lifesciences, Irvine, CA) transcatheter heart valves (THVs) from February 2010 to December 2020. Exclusion criteria were active endocarditis, previous SAVR or TAVI, mild or moderate AS, asymptomatic AS, or isolated aortic valve regurgitation. Since 2010, inoperable or high-risk patients, after Heart Team discussion, were selected for TAVI, similarly to PARTNER 1A and 1B^8^^,^^9^; moderate- or low-risk patients were considered for TAVI in case of multiple comorbidities or severe frailty. Demographic data, comorbidities, preoperative status, procedural and in-hospital outcomes, and postoperative and long-term outcomes were collected in a mean follow-up of 5.7 years. All patients signed written informed consent, and the research reported in this article adhered to Helsinki Declaration.
Data were collected from RERIC Hesperia Hospital. The “Regione Emilia Romagna Interventi Cardiochirurgia Registry” is a prospective regional database collecting preoperative, intraoperative, and postoperative data from patients undergoing cardiac surgical procedures in the 6 regional cardiac surgery departments (academic n=2; private n=4). The registry management is centralized; every 3 months, the Cardiac Surgery Department, Hesperia Hospital, is required to dispatch the data to the Regional Health Care Agency for quality/completeness control and to monitor cardiac surgery results in the Emilia-Romagna region. Demographic data, such as sex, age, and residency region at the time of the procedure, and clinical data, such as medical history, including other cardiovascular risk factors (diabetes mellitus, obesity, hypertension, dyslipidemia, and smoking habit) or other extracardiac conditions (chronic obstructive pulmonary disease [COPD], frailty, chronic kidney disease, whether on dialysis or not, cerebrovascular events, and extracardiac arteriopathy), were considered.
Preoperative, procedural, and postoperative variables are listed in Table 1, Table 2, Table 3, Table 1, Table 2, Table 3 and Table 1, Table 2, Table 3, respectively.Table 1Baseline parametersPreoperative variableTAVI(n = 437 [29.4%])SAVR(n = 1049 [70.6%])PEuroSCORE I (logistic)15.9 ± 9.47.0 ± 4.9<.001EuroSCORE II5.2 ± 4.02.0 ± 1.7<.001STS-PROM score4.5 ± 2.82.0 ± 1.2<.001Age (y), mean82.2 ± 6.272.7 ± 9.7<.001Age (y), median83.275.0<.001Female235 (53.8)498 (47.5).027BMI (kg/m^2^)27.3 ± 5.027.5 ± 5.0.4292BSA (m^2^)1.7 ± 0.21.8 ± 0.2<.001Diabetes101 (23.1)194 (18.5).042Hypertension433 (99.1)993 (94.7)<.001Hypercholesterolemia296 (67.7)659 (62.8).072Dialysis6 (1.4)8 (0.8).267Preoperative creatinine (mg/dL)1.5 ± 7.51.1 ± 3.6.145Stroke19 (4.4)21 (2)TIA23 (5.3)32 (3.1).004Smoking habit167 (38.2)544 (51.9)<.001COPD76 (17.4)60 (5.7)<.001Extracardiac arteriopathy118 (27.0)152 (14.5)<.001Bicuspid anatomy of the aortic valve7 (1.6)182 (17.4)<.001Previous cardiac surgery61 (14.0)29 (2.8)<.001Previous CABG49 (11.2)6 (0.6)<.001Previous surgery of cardiac valves14 (3.2)18 (1.7).072Previous PCI216 (49.4)111 (10.6)<.001History of myocardial infarction63 (14.4)37 (3.5)<.001Recent myocardial infarction (EuroSCORE II definition)14 (3.2)18 (1.7).072CAD260 (59.5)150 (14.3)<.001Preoperative AF83 (19.0)100 (9.5)<.001NYHA class III or IV383 (87.6)304 (29.0)<.001EF < 30%113 (25.9)185 (17.6)<.001Aortic regurgitation82 (18.8)119 (11.3)<.001Mitral regurgitation76 (17.4)52 (5.0)<.001Tricuspid regurgitation69 (15.8)27 (2.6)<.001Pulmonary hypertension22 (5.0)11 (1.1)<.001PM dependency30 (6.9)13 (1.2)<.001Liver cirrhosis3 (0.7)5 (0.5).700Active cancer5 (1.1)11 (1.1).871Senile aortic valve degeneration431 (98.6)1009 (96.2).019Degeneration of the tricuspid valve3 (0.7)33 (3.1)<.001Degeneration of the bicuspid valve3 (0.7)7 (0.7)<.001Aortic valve area (cm^2^)0.70.7nsMean aortic transvalvular gradient (mm Hg)4948.5nsValues are presented as mean ± SD, median, or n (%).AF = atrial fibrillation; BMI = body mass index; BSA = body surface area; CABG = coronary artery bypass graft; CAD = coronary artery disease; COPD = chronic obstructive pulmonary disease; EF = ejection fraction; NS = not significant; NYHA = New York Heart Association; PCI = percutaneous coronary intervention; PM = pacemaker; SAVR = surgical aortic valve replacement; TAVI = transcatheter aortic valve implantation; TIA = transient ischemic attack.Table 2Procedural variablesProcedural variableTAVI(n = 437 [29.4%])SAVR(n = 1049 [70.6%])Mechanical valve62 (5.9)Bioprosthetic valve987 (94.1)Cross-clamp time (min)60.3 ± 17.9ECC time (min)83.0 ± 21.8Edwards SAPIEN THV9 (2.1)Edwards SAPIEN XT THV123 (28.1)SAPIEN 3 THV213 (48.7)SAPIEN 3 Ultra THV92 (21.1)Transfemoral approach266 (60.9)Transapical approach105 (24)Transaortic approach60 (13.7)Transaxillary approach6 (1.4)20-mm THV6 (1.4)23-mm THV173 (39.6)26-mm THV187 (42.8)29-mm THV71 (16.2)Values are presented as mean ± SD or n (%).ECC = extracorporeal circulation; SAVR = surgical aortic valve replacement; TAVI = transcatheter aortic valve implantation; THV = transcatheter heart valve.Table 3Postoperative variablesPostoperative variableTAVI(n = 437 [29.4%])SAVR(n = 1049 [70.6%])PBlood transfusion166 (38.0)529 (50.4)<.001Mechanical ventilation time (h)13.1 ± 58.814 ± 62.4<.001ICU length of stay (d)1.7 ± 3.52.2 ± 6.6.129Acute renal failure stage 19 (2.1)25 (2.4).751Acute renal failure stage 22 (0.5)5 (0.5)Acute renal failure stage 32 (0.5)11 (1.1)Minor bleeding4 (0.9)5 (0.5).017Major bleeding17 (3.9)17 (1.6)Minor vascular complications31 (7.1)0 (0)<.001Major vascular complications9 (2.1)0 (0)Percutaneous vascular complications13 (3.0)0 (0)All stroke2 (0.5)8 (0.8).732TIA1 (0.2)4 (0.4).999Postoperative AF12 (2.8)332 (31.7)<.001PM implantation10 (2.3)24 (2.3).999In-hospital length of stay days, mean9.6 ± 12.011.6 ± 11.9.004In-hospital length of stay days, median79<.001Valve embolization1 (0.23)Annular injury2 (0.46)Coronary obstruction3 (0.69)More than moderate perivalvular leak9 (2.1)Values are presented as mean ± SD, median, or n (%).AF = atrial fibrillation; ICU = intensive care unit; PM = pacemaker; SAVR = surgical aortic valve replacement; TAVI = transcatheter aortic valve implantation; TIA = transient ischemic attack.
The length of stay, 30-day mortality, and in-hospital mortality were recorded. Using the ANPR (Anagrafe Nazionale della Popolazione Residente) database, we were able to obtain the date and cause of death. Follow-up of the patients living in Emilia-Romagna was possible because of the affiliation of Hesperia Hospital Modena to the UNIMORE TAVInAVEN study (Sostituzione valvolare aortica per via percutanea TAVI – Area Vasta Emilia Nord), approved by the Institutional Review Board of Area Vasta Emilia Nord on January 4, 2019, with protocol AOU 0000184/19. Using the hospital discharge record (scheda dimissione ospedaliera), the cardiovascular disease diagnosis ICD 9 codes (390–459), and cardiovascular procedure ICD 9 codes (35–39), especially 37.8 “Insertion, replacement, removal, and revision of pacemaker device,” and excluding Emilia-Romagna residents who underwent PPI at the time of the index procedure, we could freedom from PM implantation estimate for each patient living in Emilia-Romagna (Figure 1). The impact of PM implantation on rehospitalizations for stroke, heart failure, and all-cardiac causes (Valve Academy Research Consortium 3 definitions^10^) was evaluated. Angiographic and computed tomography parameters were identified and evaluated for their influence on PM implantation after TAVI.Figure 1Study population and pacemaker (PM) implantation incidence. SAVR = surgical aortic valve replacement; TAVI = transcatheter aortic valve implantation.
Protocols for angiographic analysis and computed tomography analysis are reported in the Online Supplement.
Leaflet (global and by cusp), left ventricular outflow tract (LVOT), and device landing zone (DLZ) calcifications were categorized as significant or not, according to other previous studies.11, 12, 13 Leaflet calcification was considered significant if >235 mm^3^ and if >4528 Agatston unit (AU). A >10-mm^3^ threshold was used for LVOT calcification, while DLZ calcification had a >1000-AU cutoff.
Continuous variables are reported as mean ± SD or as median, while categorical variables are expressed as frequency and percentage. Differences between groups for continuous variables were tested using the Student t test (unpaired 2-tailed t test) or Mann-Whitney test, as appropriate, while comparisons for categorical variables were done using the χ^2^ test or Fisher exact test, as appropriate.
All preoperative and intraoperative variables were first analyzed using univariate analysis (unpaired 2-tailed t test, Mann-Whitney test, χ^2^ test, or Fisher exact test, when appropriate) to determine whether any single factor influenced mortality. Variables that achieved P < .05 in the univariate analysis were examined using multivariable analysis by forward stepwise logistic regression to evaluate independent risk factors for mortality. Survival curves were estimated at 1, 2, 5, and 8 years using the Kaplan-Meier method and compared using the log-rank test. To adjust survival outcomes for possible influencing factors, Cox models were built and hazard ratios (HRs) have been calculated.
Univariate logistic regression analysis was then applied to the TAVI cohort to relate a broad range of preoperative parameters to the study end point (ie, 30-day PM implantation after TAVI). Variables with a P value of <.05 were selected and included in a multivariable regression model using forward stepwise selection to identify the covariates with the highest predictive value. The associations were expressed as odds ratio (OR) with their 95% confidence interval (CI). All tests were 2-sided, and P values < .05 were considered statistically significant.
To evaluate the discriminatory utility of some features, a sensitivity analysis using a receiver operator characteristic curve (ROC curve) for the risk of PPI after TAVI was performed. The area under the curve was used to summarize the performance into a single measure. An area under the curve >0.70 was considered good. Statistical analysis was performed using STATA software (StataSE.19, StataCorp, College Station, TX).
The mean EuroSCORE I, EuroSCORE II, and STS-PROM score in TAVI and SAVR were 15.9% vs 7%, 5.2% vs 2%, and 4.5% vs 2%, respectively. The mean age of TAVI and SAVR populations were 82.2±6.2 vs 72.7±9.7 years, respectively. Pulmonary hypertension (pulmonary artery pressures ≥60 mm Hg), COPD, coronary artery disease, extracardiac arteriopathy, PM implantation, previous cardiac surgery, permanent atrial fibrillation, moderate-severe mitral regurgitation, worse New York Heart Association (NYHA) class (III or IV), and reduced ejection fraction (EF < 30% and/or 30% < EF < 50%) were more frequent in the TAVI cohort. In the SAVR cohort, smoking habit and bicuspid anatomy of the aortic valve were more frequent. Regarding aortic valve area (0.7 cm^2^ vs 0.7 cm^2^; P = .65) and mean aortic transvalvular gradient (48.5 mm Hg vs 49 mm Hg; P = .73), no differences were observed (Table 1).
The SAPIEN 3 THV was implanted in about half the cases (n=213 [48.7%]), followed by the Edwards SAPIEN XT THV (n=123 [28.2%]), the SAPIEN 3 Ultra THV (n=92 [21.1%]), and the Edwards SAPIEN THV (n=9 [2.1%]). We chose the 26-mm THV in 42.8% of cases (n=187), the 23-mm THV in 39.6% (n=173), the 29-mm THV in 16.2% (n=71), and the 20-mm THV in 1.4% (n=6) (Table 2).
The TAVI cohort was characterized by mean mechanical ventilation time, mean intensive care unit length of stay, mean in-hospital stay, need for blood transfusion, and postoperative atrial fibrillation incidence statistically significantly lower than that in the SAVR cohort. By contrast, minor vascular complications and major bleeding were more common. Regarding the incidence of postoperative kidney failure, transient ischemic attack, stroke, and PM implantation, no differences between the 2 cohorts were observed (Table 3).
In the overall population, a 30-day mortality and in-hospital mortality of 1.2% (n=18) and 1.6% (n=24), respectively, were 0.9% (n=4) and 2.1% (n=9) in the TAVI cohort and 1.3% (n=14) and 1.4% (n=15) in the SAVR cohort, with no statistically significant differences between the 2 procedures. Worse NYHA class, major blood procedural transfusion rate, major or minor bleeding, higher risk scores (EuroSCORE I, EuroSCORE II, and STS-PROM), more than moderate mitral regurgitation, and chronic kidney disease whether on dialysis or not were statistically significantly higher in a bivariate analysis.
The 30-day, 1-year, 2-year, 5-year, and 8-year freedom from all-cause mortality was 98.9%, 94.9%, 78.2%, and 61.3%, respectively (Figure 2, Figure 3 and Figure 2, Figure 3).Figure 2Overall survival after surgical aortic valve replacement (SAVR) and transcatheter aortic valve implantation (TAVI).Figure 3Overall survival after transfemoral (TF) transcatheter aortic valve implantation (TAVI) and non-TF TAVI.
Excluding Emilia-Romagna residents who underwent PPI at the time of the index procedure (n=33 [2.76%] of the study population from Emilia-Romagna), we could estimate freedom from PM implantation for each patient living in Emilia-Romagna (n=1163). Permanent PM implantation rates at 30 days and after 30 days were 3.4% (n=35) and 2.9% (n=34), respectively, with no statistically significant difference (P = .081 NS). In the TAVI cohort (n=293), permanent PM implantation rates at 30 days and after 30 days were 3.4% (n=10) and 4.8% (n=14), respectively. In the SAVR cohort (n=870), permanent PM implantation rates at 30 days and after 30 days were 2.9% (n=25) and 2.3% (n=20), respectively. When comparing the permanent PM implantation rate between the SAVR and TAVI cohorts, no statistically significant difference was found (P = .640 NS). In patients who underwent TAVI (n=293), we evaluated PM implantation for each Edwards BE THV used, founding a rate of 0%, 5.7%, 3.8%, and 0% for SAPIEN (n=8), SAPIEN XT (n=88), SAPIEN 3 (n=133), and SAPIEN 3 Ultra (n=64), respectively, with no statistically significant difference (P = .265 NS). When comparing the first-generation Edwards BE THVs (SAPIEN and SAPIEN XT, n=96) with the new-generation Edwards BE THVs (SAPIEN 3 and SAPIEN 3 Ultra, n=197), the permanent PM implantation rate at 30 days was 5.2% and 2.5%, respectively, with no statistically significant difference (P = .237 NS). Overall survival in patients who underwent or not permanent PM implantation in the first 30 days after TAVI was not statistically significantly different (log-rank, P = .494 NS) (Figure 4). Also, in patients who underwent SAVR, the impact of 30-day PM implantation on overall survival was not found statistically significantly different (log-rank, P = .446 NS) (Figure 5). In the TAVI subset, COPD and preoperative atrial fibrillation were more statistically significantly prevalent in patients not implanted after the index procedure while conduction disorders, such as right bundle branch block (RBBB), were more common in implanted patients, similarly to postoperative longer weaning time (Table 4, Table 5 and Table 4, Table 5). In the same population, 49% of patients underwent nontransfemoral TAVI (29% transapical [n=7] and 20% transaortic [n=5]). The most common THV size in patients requiring PM implantation after TAVI was 26 mm in 45% of cases, followed by 23 mm in 29% and 29 mm in 25%, respectively. The univariate analysis showed preoperative RBBB as an independent predictor of PM implantation (OR 6.74; 95% CI 2.65–17.15; P < .001). Undergoing PM implantation was found to be statistically significantly related to rehospitalizations due to all cardiac causes (HR 10.28; 95% CI 3.41–31.00; P < .001). Calcium analysis showed overall leaflet calcification as an independent predictor of PM implantation (OR 4.66; 95% CI 1.41–15.47; P = .012), as well as LVOT calcification (OR 4.51; 95% CI 1.48–13.76; P = .008). The Agatston score analysis evidenced DLZ calcifications (OR 2.52; 95% CI 0.86–7.40; P = .093) as a postoperative PM implantation predictor. The multivariable regression model showed high implantation as a protective factor (OR 0.32; 95% CI 0.12–0.90; P = .030), while LVOT (OR 4.05; 95% CI 1.44–11.42; P = .008) and leaflet (OR 4.78; 95% CI 1.71–13.37; P = .003) calcifications were both associated with a higher risk of PM implantation after TAVI (Table 6, Table 7 and Table 6, Table 7).Figure 4Overall survival in patients who underwent or not pacemaker (PM) implantation in the 30 days after transcatheter aortic valve implantation (TAVI). PPI = permanent pacemaker implantation.Figure 5Overall survival in patients who underwent or not pacemaker (PM) implantation in the 30 days after surgical aortic valve replacement (SAVR). PPI = permanent pacemaker implantation.Table 4Baseline TAVI and SAVR parameters for patients who underwent PPIBaseline parameterOverall study population (N = 1163)PPI after SAVR(n = 45 [65.2%])PPI after TAVI(n = 24 [34.8%])PEuroSCORE I (logistic) (%)9.4 ± 7.76.7 ± 0.718.8 ± 2.5<.001EuroSCORE II (%)2.9 ± 2.91.8 ± 0.25.3 ± 0.8<.001STS-PROM score (%)2.6 ± 2.11.8 ± 0.24.3 ± 0.5<.001Age (y)75.0 ± 9.6 (median 77)71.6 ± 1.2 (median 72)83.6 ± 0.8 (median 83.5)<.001Female572 (49.2)19 (42.2)11 (45.8).773Body mass index (kg/m^2^)27.4 ± 5.028.1 ± 0.828.0 ± 0.9.449Surface body area (m^2^)1.8 ± 0.21.9 ± 0.31.8 ± 0.4.015Diabetes232 (19.9)14 (31.1)7 (29.2).867Hypertension1120 (96.3)44 (97.8)24 (100).462Dyslipidemia750 (64.5)31 (68.9)17 (70.8).867Dialysis11 (1.0)1 (2.2)0 (0).462Previous stroke37 (3.2)1 (2.2)2 (8.3).097Previous TIA33 (2.8)1 (2.2)3 (12.5).097.093Smoking habit562 (48.3)30 (66.7)11 (45.8)COPD (EuroSCORE)95 (8.2)2 (4.4)0 (0).295Extracardiac arteriopathy (EuroSCORE)206 (17.7)9 (20.0)6 (25.0).632Bicuspid anatomy144 (12.4)7 (15.6)0 (0).042Previous cardiac surgery67 (5.8)0 (0)3 (12.5).015Previous PCI245 (21.1)4 (8.9)13 (54.2)<.001Previous AMI76 (6.5)4 (8.9)4 (16.7).336Coronary artery disease313 (26.9)8 (17.8)15 (62.5)<.001Preoperative atrial fibrillation135 (11.6)1 (2.2)0 (0).462NYHA class III or IV516 (44.4)12 (26.7)20 (83.3)<.001Reduced EF (30%–50%)221 (19.0)7 (15.6)8 (33.3).077Severely reduced EF (<30%)22 (1.9)0 (0)1 (4.2).077Values are presented as mean ± SD or n (%).AMI = acute myocardial infarction; COPD = chronic obstructive pulmonary disease; EF = ejection fraction; NYHA = New York Heart Association; PCI = percutaneous coronary intervention; PPI = permanent pacemaker implantation; SAVR = surgical aortic valve replacement; TAVI = transcatheter aortic valve implantation; TIA = transient ischemic attack.Table 5Baseline and postoperative parameters in patients who underwent or not PM implantation after TAVIParameterTAVI (n = 293 [25.2%])No PPI after TAVI(n = 269 [91.8%])PPI after TAVI (n = 24 [8.2%])PBaseline parameters EuroSCORE I (logistic) (%)15.9 ± 0.618.8 ± 2.5.926 EuroSCORE II (%)5.3 ± 0.35.3 ± 0.8.482 STS-PROM score (%)4.4 ± 0.24.3 ± 0.5.417 Age (y)81.8 ± 0.483.6 ± 0.8.914 Female146 (54.3)11 (45.8).427 Body mass index (kg/m^2^)27.0 ± 0.328.0 ± 0.9.817 Surface body area (m^2^)1.7 ± 0.11.8 ± 0.1.416 Diabetes59 (21.9)7 (29.2).867 Hypertension268 (99.6)24 (100).765 Dyslipidemia184 (68.4)17 (70.8).806 Dialysis5 (1.9)0 (0).501 Previous stroke14 (5.2)2 (8.3).072 Previous TIA9 (3.4)3 (12.5).072.402 Smoking habit100 (37.2)11 (45.8) COPD (EuroSCORE)44 (16.4)0 (0).032 Extracardiac arteriopathy (EuroSCORE)73 (27.1)6 (25.0).821 Bicuspid anatomy3 (1.1)0 (0).603 Previous cardiac surgery41 (15.2)3 (12.5).719 Previous PCI133 (49.4)13 (54.2).657 Previous AMI40 (14.9)4 (16.7).813 Coronary artery disease163 (60.6)15 (62.5).855 Preoperative atrial fibrillation59 (21.9)0 (0).010 NYHA class III or IV240 (89.2)20 (83.3).382 Reduced EF (30%–50%)63 (23.4)8 (33.3).475 Severely reduced EF (<30%)7 (2.6)1 (4.2) Preoperative first-grade BAV32 (11.9)9 (37.5).001 Preoperative RBBB22 (8.2)10 (37.5)<.001 Preoperative LBBB17 (6.3)6 (25).001 Preoperative LAFB26 (9.7)4 (16.7).278Postoperative parameters Mechanical ventilation hours 1–527 (10.0)3 (12.5).027 Mechanical ventilation hours 6–1241 (15.2)2 (8.3) Mechanical ventilation hours 13–24195 (72.5)18 (75.0) Mechanical ventilation hours 25–481 (0.4)0 (0.0) Mechanical ventilation hours >480 (0.0)1 (4.2) ICU length of stay 1–24 h220 (81.8)20 (83.3).877 ICU length of stay 25–48 h23 (8.6)2 (8.3) ICU length of stay >48 h19 (7.1)2 (8.3) Acute renal failure stage 1261 (97.0)24 (100.0).693 Acute renal failure stage 26 (2.2)0 (0.0) Acute renal failure stage 32 (0.7)0 (0.0) Minor bleeding (VARC 2)2 (0.7)1 (4.2).190 Major bleeding (VARC 2)9 (3.4)0 (0.0) Stroke (VARC 2)0 (0.0)0 (0.0)– TIA (VARC 2)1 (0.4)0 (0.0).765 New-onset atrial fibrillation7 (2.6)0 (0.0).424 In-hospital stay (d), mean7.7 ± 0.4111.9 ± 2.5.996 In-hospital stay (d), median67.541Values are presented as mean ± SD, median, or n (%).AMI = acute myocardial infarction; BAV = atrio-ventricular block; COPD = chronic obstructive pulmonary disease; EF = ejection fraction; ICU = intensive care unit; LAFB = left anterior fascicular block; LBBB = left bundle branch block; NYHA = New York Heart Association; PCI = percutaneous coronary intervention; PPI = permanent pacemaker implantation; RBBB = right bundle branch block; TAVI = transcatheter aortic valve implantation; TIA = transient ischemic attack; VARC 2 = Valve Academy Research Consortium 2.Table 6PPI predictorsPPI predictorOR95% Confidence intervalPTotal leaflet calcification (>235 mm^3^)0.700.22–2.19.535LVOT calcification (>10 mm^3^)4.511.48–13.76.008Total leaflet calcification (>4528 AU)4.661.41–15.47.012DLZ calcification (>1000 AU)2.520.86–7.40.093High TAVI implantation (>80:20)1.130.05–28.02.940High TAVI implantation (>70:30)0.300.02–4.07.367Preoperative RBBB6.742.65–17.15<.001AU = Agatston unit; DLZ = device landing zone; LVOT = left ventricular outflow tract; OR = odds ratio; PPI = permanent pacemaker implantation; RBBB = right bundle branch block; TAVI = transcatheter aortic valve implantation.Table 7PPI predictive risk scorePredictive PPI risk scoreOR95% Confidence intervalPHigh TAVI implant (>70:30)0.320.12–0.90.030LVOT calcification (>10 mm^3^)4.051.44–11.42**.008Total leaflet calcification (>4528 AU)4.781.71–13.37.003**AU = Agatston unit; LVOT = left ventricular outflow tract; OR = odds ratio; PPI = permanent pacemaker implantation; TAVI = transcatheter aortic valve implantation.
ROC curve analysis showed a 77% specificity and 55% sensitivity for total leaflet calcification with a 203-mm^3^ cutoff value. The same sensitivity and specificity emerged considering the Agatston score analysis for the same region, with a 3732-AU cutoff value. Considering the calcium analysis by cusp (analysis of calcium performed considering the single cusps) right coronary cusp calcification reached a 70% specificity with a 80.5-mm^3^ cutoff value. Evaluating the calcification impact on PM implantation using the Agatston score, the sensitivity and specificity for left, right, and noncoronary cusp calcifications were 70% and 60% (cutoff 856.5 AU), 75% and 68% (cutoff 808 AU), and 65% and 77% (cutoff 1620 AU), respectively. DLZ calcification reached an 80% sensitivity with a 64% specificity for PM implantation, with a 20-mm^3^ cutoff value (Figure 6, Figure 7, Figure 8, Figure 9, Figure 10, Figure 11). The membranous septum length showed neither a relevant sensitivity nor specificity for PM implantation.Figure 6Receiver operator characteristic (ROC) curve for total leaflet calcification.Figure 7Receiver operator characteristic (ROC) curve for right coronary cusp calcification.Figure 8Receiver operator characteristic (ROC) curve for left coronary cusp calcification Agatston score.Figure 9Receiver operator characteristic (ROC) curve for right coronary cusp calcification Agatston score.Figure 10Receiver operator characteristic (ROC) curve for noncoronary cusp calcification Agatston score.Figure 11Receiver operator characteristic (ROC) curve for device landing zone calcification.
The American and European guidelines now recommend TAVI as a class I treatment for AS, highlighting its central role.^1^^,^^2^ Initially, TAVI was reserved for inoperable and high-risk patients. However, recent trials have demonstrated its excellent performance in moderate- and low-risk patients,^14^^,^^15^ suggesting a potential expansion of the eligible population for this procedure.
Given that new-onset conduction disturbances are more frequently associated with TAVI than with SAVR,^3^ it is crucial to determine the true incidence of PM implantation with new-generation devices and to assess the impact of permanent PM implantation on patient outcomes.
In the second decade of the 21st century, TAVI was primarily used for inoperable or high-risk patients.^8^^,^^9^ Today, the screening process includes younger and intermediate-risk patients.^14^ Data from Hesperia Hospital Modena confirm this trend, showing that patients who underwent TAVI tend to be older and have a higher burden of comorbidities (eg, extracardiac arteriopathy, COPD, coronary artery disease, worse NYHA class, previous cardiac surgery) than did patients who underwent SAVR. Age-related conditions such as preoperative atrial fibrillation and PM implantation are also more common in the TAVI cohort.
The Heart Team at Hesperia Hospital Modena, consisting of a heart surgeon, an interventional cardiologist, 2 clinical cardiologists, and an anesthetist, remained consistent throughout the 2010–2020 decade. Patients were typically referred by local cardiologists and screened by the Heart Team. Those deemed inoperable or high risk for surgery were considered for TAVI. All TAVI procedures were performed by the same team, which had extensive experience in transcatheter procedures from the outset, resulting in a favorable learning curve.
In our experience, the overall 30-day mortality rate of 1.2% is lower than that reported in the PARTNER 1A trial (4%), PARTNER 2A trial (3.9%), and the US Nationwide Readmissions Database (3.5%).^9^^,^^14^^,^^16^ However, the 30-day mortality rate for each procedure was similar to that reported in other real-world 0.9% for SAVR and 1.3% for TAVI. Di Eusanio et al, the UK national database, and Takeji et al reported 30-day mortality rates of 2.2%, 1.9%, and 1.3% for SAVR, respectively.17, 18, 19 For TAVI, the UK TAVI Registry and the Nationwide Inpatient Sample Registry reported 30-day mortality rates of 1.8% and 3.17%, respectively.^20^^,^^21^
The overall 5-year survival rate was 78.2% for both SAVR and TAVI. All-cause mortality has already been discussed previously.^22^
New-onset conduction disturbances requiring PM implantation are among the most common complications of TAVI, with an incidence of 9.3%–42% for self-expandable THVs and 2.5%–11.5% for BE THVs.^23^ SAVR can also lead to PM implantation, with rates ranging from 3.2% to 8.5%.24, 25, 26, 27, 28, 29, 30
Bagur et al^25^ reported a permanent PM implantation rate of 7.3% after TAVI with SAPIEN and SAPIEN XT in a population of 411 patients with AS and similar preoperative electrocardiographic characteristics. This rate was higher, though not statistically significant, than the 3.4% rate after SAVR. Conduction disturbances after TAVI and SAVR included complete atrioventricular (AV) block (5.6% vs 2.7%; P = .039) and symptomatic severe bradycardia (1.7% vs 0.7%; P = .097). Preoperative RBBB was an independent predictor of PM implantation after TAVI (OR 8.61; 95% CI 3.14–23.67; P < .0001), but not after SAVR.^25^
The introduction of new-generation Edwards BE devices, SAPIEN 3 and SAPIEN 3 Ultra, was initially not associated with a lower incidence of permanent PM implantation. In the PARTNER 3 trial, PM implantation and left bundle branch block rates were higher in the TAVI group than in the SAVR group, though not statistically significant (6.6% vs 4.1%; HR 1.65; 95% CI 0.92–2.95 and 22.0% vs 8.0%; HR 1.65; 95% CI 2.13–4.72, respectively).^15^
The strong association between preoperative conduction disturbances, especially RBBB, and postprocedural PM implantation has led to the inclusion of RBBB in preoperative risk scores, such as the Emory risk score^31^ and the Midas strategy.^32^ In our study, RBBB was a consistent independent predictor of PM implantation after TAVI (OR 6.74; 95% CI 2.65–17.15; P < .001). This may be due to an increased incidence of bradyarrhythmia and high-grade AV block in cases of left bundle branch injury during valve deployment.^33^
Recent years have seen the adoption of a different implantation strategy for new-generation Edwards BE devices, SAPIEN 3 and SAPIEN 3 Ultra, which involves reducing the implantation depth.34, 35, 36, 37, 38, 39 This approach aims to reduce the THV frame in the LVOT by positioning the balloon marker a few millimeters above the virtual basal ring. Mailey et al^40^ found that a high implantation strategy, using the radio-transparent line on the THV crimped frame as a reference, resulted in a permanent PM implantation rate of 4.4%. However, the effectiveness of this strategy needs further confirmation in clinical practice, as it may affect coronary reaccess and valve-in-valve feasibility.
In our population, the permanent PM implantation rate was 3.4% for TAVI and 2.9% for SAVR, with no statistically significant difference. When comparing the old-generation Edwards BE THVs (SAPIEN and SAPIEN XT, n=96) with the new-generation devices (SAPIEN 3 and SAPIEN 3 Ultra, n=197), the 30-day PM implantation rates were 5.2% and 2.5%, respectively, with no statistically significant difference (P = .237). The introduction of new-generation devices led to a reduction in the 30-day PM implantation rate (2.5%), possibly because of higher THV implantation techniques.
In the last decade, the depth of TAVI implantation has become a crucial topic of discussion. Initially, an implantation depth ratio of <70:30 was recommended by industry standards.^41^ However, recent studies have highlighted the benefits of more aortic implantation.^32^^,^^40^
In our analysis, we categorized THV implantation depth into 2 one with a ratio of >80:20 and another with a ratio of >70:30. The statistical analysis showed no significant difference for the >80:20 class, while the >70:30 ratio was identified as a protective factor, reducing the risk of PM implantation by up to 68% in our predictive risk model. Ramanathan et al^42^ and Jilaihawi et al^32^ reported similar findings, supporting the evidence that higher implantation acts as a protective factor for PPI.
This technique reduces the need for a PM because the valve is deployed high in the aortic annulus, with only a minimal portion of the valve in the LVOT, thereby avoiding the conduction system. The AV node is located in the right atrium, near the septal part of the tricuspid valve and the muscular ventricular septum, which is continuous with the aortic valve cusps. At a lower level, the LVOT is closely related to the left bundle branch.^43^ Direct mechanical trauma or compression of the AV node or the left bundle branch by balloon dilation or prosthesis implantation can cause a high-degree AV block or left bundle branch block during or after TAVI. Preexisting damage to the conduction system, such as RBBB, increases the risk of advanced conduction abnormalities after TAVI, necessitating PPI.
Analyzing overall survival with and without 30-day PM implantation after SAVR and/or TAVI showed no statistically significant differences. The literature reports conflicting results on this topic. Some studies associate 30-day PM implantation with worse overall survival and more frequent rehospitalizations due to heart failure,^44^^,^^45^ while others do not find a negative impact on patient prognosis.^46^ In our study, PM implantation did not significantly affect overall survival but was associated with a higher risk of rehospitalizations due to cardiac causes (HR 8.65; 95% CI 4.87–15.36; P < .001). This trend, observed by Sá et al,^44^ was confirmed in our population. The higher rehospitalization rate may be related to PM implantation, as right ventricular pacing can lead to dyssynchronous systolic left ventricular function, higher degrees of heart failure, and lower left ventricular EF.^47^ The absence of mechanical synchrony after afterload relief (induced by preprocedural left ventricular hypertrophy) combined with ventricular dyssynchrony (electrically induced by right ventricular pacing) may not only compromise the postprocedural normalization of cardiac function but also cause further decline in function.^44^
The impact of aortic valve calcification on procedural success and outcomes is a well-discussed topic in the literature. Its association with annular rupture^48^ and paravalvular aortic regurgitation^12^^,^^49^ has been established. However, the role of aortic valve calcification in PPI remains unclear.
Nai Fovino et al^50^ found that patients who received a PM had a higher grade of LVOT calcifications (36% vs 32%; P = .045) using a qualitative calcification grade. Latsios et al^51^ proposed DLZ calcification, using a semiquantitative method, as a predictor of PPI (OR 1.06; 95% CI 1.02–1.11; P = .004). Fujita et al,^52^ using a quantitative method, found an association between left coronary cusp calcification (>209 mm³) and a higher rate of PPI after TAVI. However, Gama et al,^53^ in a quantitative calcium analysis, failed to prove any correlation between aortic root calcification and PPI.
In our experience, using a quantitative calcium grading system, overall leaflet calcification > 4528 AU, LVOT calcification > 10 mm³, and DLZ calcification >1000 AU (even if with P = .093) were identified as independent predictors of PM implantation. However, we did not find a correlation between regional calcium location and 30-day PM implantation.
We also failed to prove any correlation between membranous septum length and 30-day PM implantation after TAVI, unlike other authors. However, we did find a correlation between a high implantation technique and a lower PM implantation rate. Chen et al^54^ analyzed the preprocedural assessment of membranous septum length in coronal view compared with infra-annular membranous septum length from the stretched vessel image, highlighting the need to determine the optimal implantation depth to mitigate conduction disturbances after TAVI. Jung et al,^55^ in a large cohort of consecutive patients, confirmed the significantly independent predictive value of membranous septum length for the occurrence of postprocedural AV block with the need for PPI.
Using ROC curve analysis, we were unable to identify any variable with relevant sensitivity or specificity. For isolated sensitivity, left coronary cusp (70%), right coronary cusp (75%), and DLZ (80%) calcifications were significant, while for specificity, total leaflet (77%), right coronary cusp (70%), and noncoronary cusp (77%) calcifications were adequate.
Our results support the need for further investigation into the impact of aortic valve calcification on PM implantation rates. If confirmed, this should become part of routine preprocedural evaluation.
The study presents all the intrinsic limitations of observational retrospective studies, such as discrepancy in data collection, changes in definitions of comorbidities, and loss of patients during follow-up. However, between 2010 and 2020, all data were collected by the same physician, comorbidity definitions did not change, and patients were registered in the same database (RERIC). Moreover, adopting the Italian ANPR database and the hospital discharge record (scheda dimissione ospedaliera) system for Emilia-Romagna residents, no patients were found to be lost at follow-up, regarding mortality and rehospitalizations during follow-up. During the study period, the choice of TAVI or SAVR, made by the local heart team, was influenced by outcome data from PARTNER trials 1A and 1B.^1^^,^^2^ Inoperable and high-risk patients were selected for TAVI, while SAVR was performed in all others, introducing selection bias, which can be observed considering the remarkable heterogeneity of our 2 populations. All patients who underwent TAVI were treated with BE THVs, allowing the comparison with procedural outcomes from PARTNER trials, but preventing from translating our outcomes with self-expandable THVs.
During the first 2 years of the study, electrocardiogram-gated computed tomography scans were not available, introducing a possible bias in imaging analysis.
Mechanical trauma during TAVI can injure the conduction system, leading to permanent PM implantation. Anatomical features, particularly the variation in the length and location of the penetrating segment of the bundle of His and the depth of the proximal portion of the left bundle, affect susceptibility to injury.
In our population, a 3.4% permanent PM implantation rate was reported. This outcome is likely due to the exclusive use of Edwards BE devices and thorough preoperative evaluation. However, given the negative impact of PPI on patient prognosis, it is essential to explore new tools in preoperative screening and operative techniques, such as high implantation, to further reduce this event.
While the impact of PPI on mortality remains debated, it is still associated with the increased length of hospital stay and rehospitalizations. A comprehensive preoperative evaluation, considering factors such as the amount of calcium in the LVOT (LVOT calcification >10 mm^3^) and aortic root (overall leaflet calcification > 4528 AU), the length of the membranous septum, the implantation depth in the LVOT, and a precise computed tomography–scan simulation before the procedure, could be crucial in lowering the PPI incidence and preserving the benefits of TAVI.
Dr Gabbieri declares his role of consultant for Edwards Lifesciences. Mr Chiarabelli is an Edwards Lifesciences employee. The other authors declare no conflict of interest.