Single or dual antiplatelet therapy after transcatheter aortic valve implantation. A meta-analysis of randomized controlled trials

INTRODUCTION

Atrial fibrillation (AF) increases with age and raises the risk of ischemic stroke, systemic embolism, and death, with stroke or transient ischemic attack (TIA) often being the initial presentation. 1,2 Oral anticoagulation (OAC) with direct OAC (DOAC) or vitamin K antagonists (VKA) is effective in reducing these risks.1,2 However, recent randomized clinical trials suggest a 1.0%–1.5% annual stroke rate in AF patients on OAC.3-5 Up to one-third of AF patients who develop ischemic stroke are on OAC at stroke onset, with 8.8% up to 20% being on DOAC,6-9 which poses a significant challenge for secondary prevention purpose. Additionally, these patients tend to be older and have more comorbidities, sometimes requiring off-label low-dose DOAC.7 The use of off-label low-dose DOAC, atrial enlargement, and increased AF burden further raises the risk of stroke despite OAC therapy, suggesting a possible association with advanced cardiac disease or inadequate anticoagulation.10,11

After an ischemic stroke—despite OAC—there are no clear guidelines on the management of OAC, such as switching drugs, targeting a higher international normalized ratio (INR) with AVK, or adding antiplatelet agents. Network meta-analyses show no differences in stroke risk across different DOAC. Observational studies also found no benefits in changing OAC, suggesting additional stroke mechanisms.7 Stroke workups should identify alternative embolisms or sources of thrombosis, such as large-vessel atherosclerosis or small-vessel disease, which may need specific antithrombotic strategies.7,9,10,12 Although adding antiplatelets may reduce platelet activation, it is associated with a higher bleeding risk.13 Therefore, the best secondary prevention for AF patients after stroke despite OAC remains unclear, highlighting the need for new treatments, including novel non-pharmacological approaches such as left atrial appendage occlusion (LAAO).

Recent data from the STR-OAC cohort—an international LAAO registry—showed promising findings in AF patients with stroke despite OAC. In this study, the LAAO cohort had a 2.2% per patient-year stroke rate vs the 9.8% reported in the non-LAAO cohort (HR, 0.33; 95%CI, 0.19-0.59).14 Despite this positive outcome, no RCTs have assessed the feasibility, safety and efficacy profile of using OAC plus LAAO as adjuvant therapy in this complex population. Herein, we propose oral anticoagulation alone vs oral anticoagulation plus left atrial appendage occlusion in stroke patients despite ongoing anticoagulation (ADD-LAAO trial). This is a pragmatic randomized controlled trial (RCT) designed to evaluate the superiority of a hybrid strategy combining transcatheter left atrial appendage occlusion (LAAO) and long-term continued oral anticoagulation (OAC)—either direct oral anticoagulants (DOAC) or VKA when clinically appropriate—vs medical management with OAC alone. The trial aims to assess the effectiveness of this approach in reducing recurrent ischemic stroke in patients with atrial fibrillation (AF) who have experienced an AF-related acute ischemic stroke despite being anticoagulated at the time of stroke onset.

METHODS

Study design

We conducted a multicenter randomized controlled trial with patients with a past medical history of ischemic stroke within the past 6 months despite OAC. A total of 6 teaching hospital centers are involved in this study. Eligible patients will be screened and included based on specific criteria (figure 1 illustrates the inclusion and exclusion criteria). In general, patients with stroke despite OAC will include the following data: those with VKA with correct or labile INR and those with DOAC with poor compliance. Poor compliance will be defined as missing a maximum of 1 day dose (1 pill for VKA and 2 for DOAC) over the past week before the index stroke. Poor compliant patients will be included as the risk of non-compliance is an inherent characteristic of OAC. In contrast, patients missing > 1 day dose over the past week before the index procedure will not be included as these patients cannot be considered anticoagulated. A crucial inclusion criterion for randomization purposes will be the absence of an absolute contraindication to OAC, as this population is considered a high-risk cohort for thrombotic events, and post-LAAO OAC discontinuation has been reported to increase thrombotic risk.15 The study flow diagram is shown in figure 2.

Figure 1. Inclusion and exclusion criteria. AF, atrial fibrillation; CT, computed tomography; DOAC, direct oral anticoagulant; Hb, hemoglobin; LAAO, left atrial appendage occlusion; NYHA, New York Heart Association; OAC, oral anticoagulation; PFO, patent foramen ovale; TEE, transesophageal echocardiogram; TIA, transient ischemic attack.
^a Cardioembolic ischemic stroke definition: AF-related stroke after ruling out symptomatic ipsilateral great vessel/intracranial vascular disease, and small vessel disease and active endocarditis or neoplasm.
^b > 1 dose per antagonist of vitamin K and > 2 doses per DOAC.
^c > 50% lumen diameter narrowing on CT, magnetic resonance imaging, or transcranial Doppler with symptoms of ipsilateral transient or visual TIA.
^d If general anesthesia is planned for the study procedure.
^e If the patient needs CCTA and cannot undergo TEE.
^f The active treatment group may confound the results of this trial.

Figure 2. Central illustration. Study protocol. CT, computed tomography; DOAC, direct oral anticoagulant; LAAO, left atrial appendage occlusion; OAC, oral anticoagulation; TEE, transesophageal echocardiogram.

Subject screening, enrollment, and randomization

Patients meeting all inclusion and no exclusion criteria will be approached for the study. Prior to inclusion, transesophageal echocardiography (TEE) or coronary computed computed tomography will rule out the presence intra-cardiac thrombus and assess the LAA anatomy. If suitable for LAAO, the study will be explained, and informed consent from the patients will be obtained. Upon consent, patients will be randomized on a 1:1 ratio to the interventional group—receiving LAAO and long-term OAC—or the control group, receiving the optimal medical therapy as decided by neurologists. The interventional procedure will occur within 2 weeks post-randomization. The optimal medical therapy may involve intensified antithrombotic therapy, switching OAC regimens, reinforcing drug compliance, or a combination of these strategies. Randomization will be managed online, allocating patients in groups of 10 to ensure balanced inclusion across both groups.

OAC strategies

The treating neurologist team will decide on the OAC strategy and dosage for the interventional and control groups. Any DOAC (apixaban, dabigatran, rivaroxaban, or edoxaban) will be accepted. DOAC dosages will be down titrated if the patient’s bleeding risk is high or per product label recommendations. In the interventional group, although DOAC is preferred to minimize bleeding risk,16 VKA will be allowed if necessary, such as for patients with mechanical cardiac valves. The protocol does not restrict other concomitant drugs; each will be assessed for additional hemorrhagic risk by the physician. The OAC duration will be indefinite in the 2 groups unless a new formal contraindication emerges, such as in cases of major bleeding. Furthermore, each situation will be managed individually by investigators at each center.

Interventional group

Patients from the interventional group will undergo LAAO following standard practice. ACO will be discontinued prior to the procedure per product label recommendations for DOAC, and with bridging therapy using low-molecular-weight heparin in patients with an indication for VKA. A prophylactic antibiotic—cephalosporin or vancomycin for beta-lactam allergy—will be administered 2 hour prior. Procedure will be performed under general anesthesia or deep sedation with TEE or intracardiac echocardiography guidance. The femoral vein will be used for vascular acces, followed by a transseptal puncture. A 100 IU/kg bolus of IV heparin will be administered to achieve an activated clotting time ≥ 250 seconds. Fluoroscopy and TEE or intracardiac echocardiography will guide the procedure, allowing for LAA measurement, catheter and device positioning, and early complication detection. Approved LAAO devices include Amulet (Abbott Medical, United States), Watchman FLX (Boston Scientific, United States), and LAmbre (Lifetech Scientific [Shenzhen] Co. Ltd., China). Six to 24 hours after the intervention, transthoracic echocardiography will check for any pericardial effusions or device embolizations that may have occurred. If no complications are found, the patient will be discharged the same or the next day (institution protocol). Anticoagulation therapy (DOAC at the same dose as before LAAO or VKA with the initial dose determined by the thrombosis clinic assessment) will be resumed the day after the procedure. No antiplatelet therapy will be added to anticoagulation therapy in the LAAO group.

Study endpoint and outcome definitions

The primary endpoint of the study is the occurrence of a cardioembolic event—ischemic stroke or peripheral arterial embolism— within the first 12 months after inclusion. Secondary endpoints include evaluating the safety and efficacy profile of the strategies and combining cardioembolic events (efficacy) and major bleeding (safety) within the same period. A stroke is defined as the sudden onset of a focal neurologic deficit consistent with a major cerebral artery territory, categorized as ischemic, hemorrhagic, or unspecified, and confirmed through imaging using computed tomography or magnetic resonance imaging. Systemic cardioembolic events are acute vascular occlusions in an extremity or organ, confirmed by imaging, surgery, or autopsy. Major bleeding will follow the Bleeding Academic Research Consortium (BARC) criteria for types 3 and 5.17

Additional endpoints will assess major and minor bleeding (using the BARC classification), all-cause and cardiovascular death, recurrent stroke severity (using the modified Rankin scale), procedural major adverse events in the interventional group, device-related thrombus, additional hospital admissions, and OAC compliance. All clinical events, including primary and secondary endpoints, will be independently allocated.

The success of the intervention is defined as the implantation of the LAAO device without major complications, such as death, stroke, or those requiring surgical or endovascular treatment.18 Procedural safety will be evaluated by including all clinical events within the first 7 after the intervention. Events will be defined following the Valve Academic Research Consortium (VARC) guidelines, including mortality, myocardial infarction, stroke, systemic embolism, major bleeding, and procedural complications.19 Major bleeding will follow the BARC criteria for types 3 and 5.17 A neurologist will independently grade disabling strokes with an m-RS score of ≥ 3.20 Device-related thrombus will be any thrombus > 1 mm on the LAAO device, and peri-device leaks will be classified by TEE jet width (> 3 mm being significant).21 Complete LAAO is defined as the absence of any leaks > 3 mm on the final TEE.22

Clinical and imaging follow-up

Patients will be followed for 12 months. Clinical visits with neurological assessment and modified Rankin scale evaluation will occur on months 3 and 12. A phone follow-up will be conducted on month 6. The interventional group will undergo post-LAAO additional imaging modalities. Two imaging modalities with TEE or coronary computed tomography angiography will assess device-related thrombus and peri-device leaks 3 (2-4 months) and 12 months (10-12 months) after the intervention. Additionally, standard blood tests, including complete blood count and renal function will be performed on the same day as the imaging modalities to detect hidden hemorrhagic events.

Sample size and statistical analysis

Sample size is determined based on observed event rates in major registries, with an estimated 10% rate of recurrent cardioembolic events in the control group and 2% in the interventional group within the first year. To detect an 8% difference between the 2 groups, with a 5% type I error and 90% power, 183 patients per group are needed for a total of 366. Accounting for potential dropouts (~5%), the study will include 380 patients. The primary analysis will be conducted following the intention-to-treat principle, making sure that all randomized patients are analyzed in the group they have been initially allocated to, regardless of protocol deviations, dropouts, or crossover. This approach will provide an unbiased estimation of the treatment effect under real-world conditions and preserve the benefits of randomization. If a patient from the interventional group experiences a primary endpoint prior to the procedure—as they will be on OAC treatment—they will be allocated to the interventional group. An interim analysis will be performed after including 50% of the population (190 patients), and the study will be stopped if significant differences are detected.

Categorical variables will be expressed as frequencies and compared using the chi-square or Fisher’s exact test. Continuous variables will be expressed as mean ± SD or median (IQR), using the Kolmogorov-Smirnov test for normality. Comparisons will use the Student’s t test or the Mann-Whitney U test. Composite endpoints will be assessed as a time-to-first event. Cumulative incidence will be evaluated using the Kaplan-Meier method and compared using the log-rank test, followed by Cox proportional hazards modeling. Treatment effects will be estimated with hazard ratios and 95% confidence intervals, with 2-sided P-values ≤ .05 considered statistically significant. Analyses will be performed using STATA (Version 14.0 (Stata Corp., United States). The trial has been registered on ClinicalTrials.gov, and the registry No. is pending registration approval.

Current study status

Study recruitment is set to commence. The study is expected to be completed in 23 months. The projected study timeline is shown in figure 3.

Figure 3. Projected study timeline.

DISCUSSION AND CLINICAL IMPLICATIONS

The current study is expected to have a significant impact on the scientific community, particularly in the fields of cardiology and neurology. As previously mentioned, the prevalence of AF and the number of patients experiencing ischemic strokes despite being OAC are on the rise,3-5 which presents a significant challenge due to the absence of clear and uniform treatment recommendations for these patients. Intensification of antithrombotic regimens in this population often leads to an elevated risk of major bleeding, especially among elderly and frail individuals.13 Therefore, assessing a novel therapeutic strategy that combines LAAO with DOAC is essential. LAAO targets the left atrial appendage, which is responsible for more than 90% of thrombus formation in non-valvular AF,16 potentially enhancing the efficacy of OAC while maintaining a lower bleeding risk with DOAC. Recent registries evaluating the LAAO + DOAC strategy have reported promising outcomes, demonstrating a significantly reduced rate of recurrent strokes and major bleeding vs optimal medical therapy alone.23-25 Given these encouraging preliminary data, the timing is ideal for a randomized controlled trial to evaluate this combined approach rigorously.

As far as we know, the proposed study is one of the first randomized clinical trials to compare LAAO + DOAC and optimal medical therapy, as determined by a neurologist, in patients who have experienced ischemic strokes despite being on OAC. The ELAPSE trial has also started recruitment with a similar design (NCT05976685). Should trial results confirm the superiority of the LAAO + DOAC strategy over current medical management protocols, our findings will contribute to a paradigm shift in the treatment guidelines for this patient cohort. Specifically, the LAAO + DOAC combo could offer a more effective and safer therapeutic option, addressing the unmet need for reducing stroke recurrence while minimizing bleeding risks, which could impact future clinical practice and guideline recommendations, ultimately improving patient outcomes in those with AF and a history of ischemic stroke. This trial successful completion and positive outcomes can potentially establish a new standard of care, thus significantly impacting clinical practice and patient quality of life alike in this high-risk population.

To achieve these goals, the centers participating in this study have been carefully selected based on their annual volume of candidates for LAAO and their expertise in managing these procedures. These centers are recognized as referral centers for both LAAO and stroke code management. Although randomizing 380 patients across 6 hospitals within a reasonable timeframe may appear challenging, we have implemented measures to streamline the process. Dedicated teams for the early identification of eligible patients and standardized follow-up strategies have been established to facilitate recruitment.

We recognize that this is an ambitious undertaking; however, the study design and the cumulative experience of participant centers provide confidence in achieving the goals set within the anticipated timeline. If successful, this trial has the potential to establish a new standard of care, significantly impacting clinical practice and improving the quality of life of this high-risk population.

FUNDING

This work has been funded by a grant from Fundació La Marató de TV3.

ETHICAL CONSIDERATIONS

The study is being conducted following the recommendations outlined in the Declaration of Helsinki on clinical research, has been approved by Hospital Clinic de Barcelona Research Ethics Committee, and endorsed by the remaining ethics committees of all participant centers. Informed consent acceptance and signature are required prior to performing any elective procedures for the study of the non-culprit lesions. Potential sex and gender biases are considered.

STATEMENT ON THE USE OF ARTIFICIAL INTELLIGENCE

No artificial intelligence was used in the drafting of this manuscript.

AUTHORS’ CONTRIBUTIONS

X. Freixa and E. Flores-Umanzor drafted this document. The remaining signatories reviewed the document and made changes at their discretion. All the authors revised and approved the final version of the manuscript.

CONFLICTS OF INTEREST

I. Cruz-Gonzalez and X. Freixa are proctors from Abbott Medical, Boston Scientific and Lifetech. R. Estevez-Loureiro and D. Arzamendi are proctors from Abbott Medical, Boston Scientific. L. Nombela- Franco is a proctor from Abbott Medical. The rest of the authors declared no conflicts of interest whatsoever.

REFERENCES

INTRODUCTION

Coronary artery disease is the leading cause of mortality worldwide.1 Despite the safety involved in deferring invasive treatment in patients with stable coronary artery disease without functionally significant lesions,2 a percentage of patients experience cardiovascular events at the long-term follow-up.3 It has been reported that cardiovascular events not only depend on the degree of coronary obstruction assessed by intracoronary physiology4-5 but also on the global atherosclerotic burden and its vulnerability assessed by intracoronary imaging modalities.6-8

The new era of coronary physiology is based on predicting fractional flow reserve by reconstructing the coronary tree using angiography and computational fluid dynamics.9-10 Estimating quantitative flow ratio (QFR) is the most validated method of the ones currently available.

QFR—which predicts fractional flow reserve^10-11—has proven to be a better tool than angiography alone to guide the need for lesion revascularization12 and shown long-term prognostic value13. Furthermore, it provides quantitative information out of the 3D reconstruction of the coronary tree, including minimum diameter and area, reference diameters, luminal volume, and atherosclerotic plaque volume in the studied vessel. However, the prognostic value of this quantitative analysis has not been sufficiently studied.

The main aim of this study was to determine the prognostic value of global plaque volume (GPV) in patients with stable coronary artery disease without functionally significant lesions at a 5-year follow-up.

METHODS

We conducted a retrospective observational study on a cohort of patients from 6 tertiary referral centers.

Study population

Patients who underwent coronary angiography from January through December 2015 for suspected stable coronary artery disease were included. Each participant center retrospectively enrolled all patients who underwent coronary angiography for suspected stable coronary artery disease and met the inclusion criteria. Patients with chronic total coronary occlusions, prior coronary artery bypass graft surgery, or inadequate angiographic quality for analysis were excluded. Additionally, patients whose angiographic analysis revealed a positive QFR study (< 0.80) in any coronary territory were excluded. The principal investigator conducted a retrospective follow-up at each center within the next 5 years following the index procedure. Baseline and procedural characteristics, and events at the follow-up were collected by local investigators. The study fully complied the good clinical practice principles and regulations set forth in the Declaration of Helsinki for research with human subjects. The study protocol was approved by the ethics committee of the reference hospital (Hospital Clínico Universitario de Valladolid) and the institutional review boards, including informed consent obtained from participants or, alternatively, approval for retrospective data analysis under ethical committee supervision.

Angiographic analysis

A blinded angiographic analysis of diagnostic coronary angiograms was performed by trained analysts at a centralized imaging unit (Icicorelab, Valladolid) using specialized software (QAngio XA 3D QFR, Medis Medical Imaging System, The Netherlands). A 3D reconstruction of the 3 major coronary vessels was performed using 2 different projections with > 25° of separation. For the right and left circumflex coronary arteries, the proximal marker was manually placed at the vessel ostium, while for the left anterior descending coronary artery, it was placed at the left main coronary artery ostium. The distal marker was placed at the end of the coronary artery. Plaque volume was estimated by calculating the difference between the theoretical reference vessel volume in the absence of atherosclerotic disease and the estimated vessel volume in angiography using QFR software via quantitative analysis. Reference diameters, minimum diameter, and minimum area were obtained for each vessel. Considering contrast flow through the coronary tree, QFR was calculated according to FAVOR II standards for the physiological significance of coronary lesions. Patients with functionally significant disease (QFR < 0.80) were excluded.

Statistical analysis

Categorical variables are expressed as totals and percentages, and continuous ones as means and standard deviations. GPV was estimated as the sum of plaque volume across 3 coronary territories.

The primary endpoint—major adverse cardiovascular events (MACE)—was a composite of cardiac death, acute myocardial infarction, or all-cause unplanned hospital admission.

An optimal GPV cutoff as a predictor of MACE was determined using the receiver operating characteristic (ROC) curve as the value with the maximum Youden index. Multivariate logistic regression models were used to calculate the odds ratio and 95% confidence interval as independent predictors for MACE. Variables with P < .20 in the univariate analysis were included in the multivariate model as covariates.

Event-free survival was compared using Kaplan-Meier and Mantel-Haenszel analyses. All probability values were two-tailed, and P < .05 was considered statistically significant. Statistical analysis was performed using Stata (16.1, StataCorp, College Station, United States).

RESULTS

Descriptive population analysis

A total of 803 patients were evaluated for inclusion in the registry, 122 of whom (15.2%) were excluded due to chronic occlusions in ≥ 1 coronary territory, 17 (2.12%) due to previous surgical myocardial revascularization, and 159 (19.2%) due to inadequate angiographic analysis in, at least, 1 coronary territory. Among the remaining patients, 228 (45.1%) had significant coronary artery disease (QFR < 0.80) in, at least, 1 coronary territory, which left a final cohort of 277 patients. Patient flowchart is shown in figure 1.

Figure 1. Flowchart of the patient selection process for inclusion in the study. CABG, coronary artery bypass graft; CTO, chronic total coronary occlusion; QFR, quantitative flow ratio.

The mean age of the population was 65.8 years (most were hypertensive [74.4%] men [66.1%]). Table 1 illustrates the baseline characteristics of the population. The median follow-up was 69 months, during which time 5 patients were lost to follow-up.

Table 1. Baseline characteristics of the included population

Angiographic analysis

Mean plaque volume in the study population was 170.5 mm³ (± 16.5); mean QFR was 0.95. Table 2 illustrates the overall means from the angiographic analysis according to the coronary territory studied. Plaque volume was independently analyzed for each coronary territory and was significantly higher in the right (243 mm³) vs the left anterior descending (161.4 mm³) and left circumflex coronary arteries (172.9 mm³). Data on this analysis by coronary territories are shown in table 1 and figure 1 of the supplementary data.

Table 2. Characteristics of the angiographic analysis performed in the 3 coronary territories using quantitative flow ratio

Prognostic value of global plaque volume

The primary event (MACE) occurred in 116 patients, which amounts to 42.7% of the cohort at the follow-up. Among these patients, 11% died, 2.6% suffered an acute myocardial infarction, and 38.1% required unplanned hospitalization. Patients who developed MACE had a significantly higher GPV (231.6 vs 111.8 mm³; P < .001), as well as those with a higher mortality rate (255.2 mm³ vs 154.3 mm³; P = .04) or unplanned hospitalizations (235.0 mm³ vs 125.4 mm³; P < .001). However, there were no significant differences in patients who experienced acute myocardial infarction (235.1 mm³ vs 169.3 mm³; P = .51).

The optimal GPV cutoff to predict events was set at 44 mm³ based on ROC curve analysis (sensitivity, 64%; specificity, 65.8%; LR⁺, 1.9; LR^–, 0.6).

Table 3 illustrates the study of the main determinants of the primary event. Variables with a significance level of P < .10 were included in the multivariate analysis. In the final model, age and GPV were independent predictors. A GPV > 44 mm³ was associated with a 2.8-fold higher risk of events at the follow-up (figure 2).

Table 3. Uni- and multivariate analysis of determinants of the main event

Figure 2. Kaplan-Meier curve showing the patients’ event-free survival based on their global plaque volume.

DISCUSSION

The main finding of this study is that GPV quantification emerged as an independent prognostic factor in patients without functionally significant coronary artery disease, which demonstrated that those with a higher GPV experienced more events at the follow-up. The optimal GPV cutoff for event prediction was set at 44 mm³. This study emphasizes the importance of anatomically characterizing coronary arteries without significant lesions.

Despite the absence of significant coronary artery obstructions, some patients still experience events during follow-up.14 In patients with a negative QFR functional study, it has been reported that the 5-year rate of events—cardiac death, target vessel myocardial infarction—is 11.6%,3 similar to our findings, where mortality rate was 11% and acute myocardial infarction occurred in 2.6% of patients. Determining the difference between the actual vessel diameter and the estimated diameter obtained through 3D reconstruction from QFR-based angiography has been used in other studies.15 This estimation—previously derived from coronary computed tomography16-17—has demonstrated the prognostic significance of plaque volume differences between normal and non-obstructive coronary arteries. These differences have also been confirmed using invasive imaging modalities such as intravascular ultrasound.18 Although angiography-derived percent luminal stenosis shows poor concordance with myocardial ischemia,19 a greater degree of coronary stenosis (percent diameter stenosis > 50%) is associated with a higher event rate at the 2-year follow-up in patients without functionally significant coronary lesions.20 The present study takes a step further into the minimally invasive characterization of atherosclerotic burden using easy-to-implement 3D coronary tree reconstruction technology as an independent prognostic factor in patients without functionally significant coronary lesions. In this regard, this study is consistent with recent studies which demonstrated that subclinical atherosclerosis burden—measured by vascular ultrasound for carotid plaque quantification and computed tomography for coronary calcium scoring—in asymptomatic individuals is independently associated with all-cause mortality.21

Based on these findings, GPV measurement enables the identification of patients who, despite having no significant coronary lesions, are at risk of developing events within the next 5 years, allowing for intensified treatment and cardiovascular risk factor control. However, this study has limitations, including its retrospective design for patient inclusion and recruitment, the use of indirect methods—such as QFR—to estimate plaque volume, and the inability of this method to describe plaque characteristics, or potential lipid plaque vulnerability. Of note, the estimated plaque volume in each coronary artery was not specifically correlated with events in that territory but rather with overall adverse cardiovascular events. Therefore, further studies are needed to confirm or refute this hypothesis.

CONCLUSIONS

Plaque volume, calculated by 3D coronary tree reconstruction, is an independent predictor of events in patients with suspected stable ischemic heart disease without significant coronary artery disease. The optimal GPV cutoff for event prediction is 44 mm³.

FUNDING

C. Cortés received funding through the Río Hortega contract CM22/00168 and Miguel Servet CP24/00128 from Instituto de Salud Carlos III (Madrid, Spain).

ETHICAL CONSIDERATIONS

The present study was conducted in full compliance with clinical practice guidelines set forth in the Declaration of Helsinki for clinical research and was approved by the ethics committees of the reference hospital (Hospital Clínico Universitario de Valladolid) and other participant centers. Possible sex- and gender-related biases were also considered.

DECLARATION ON THE USE OF ARTIFICIAL INTELLIGENCE

No artificial intelligence was used in the writing of this text.

AUTHORS’ CONTRIBUTIONS

C. Cortés and J. Ruiz-Ruiz participated in study design, data analysis, manuscript drafting, and critical review. C. Fernández and M. García participated in data collection and result analysis. F. Rivero and R. López-Palop assisted in data collection. S. Blasco and A. Freites contributed to statistical analysis. L. Scorpiglione and M. Rosario Ortas Nadal collaborated in data interpretation. O. Jiménez participated in manuscript preparation and initial review. J.A. San Román Calvar and I.J. Amat-Santos conducted the final review and approved the version for publication.

CONFLICTS OF INTEREST

None declared.

SUPPLEMENTARY DATA

REFERENCES

INTRODUCTION

Transcatheter aortic valve implantation (TAVI) revolutionized the treatment of severe aortic valve stenosis over the past decade.1,2 Furthermore, in recent years, there has been a clear preference for TAVI over surgical techniques regarding the treatment of this valvular heart disease, with an increasing number of patients, including low surgical risk ones.3 Although post-TAVI infective endocarditis (IE) is a rare complication reported in 1% up to 6% of patients, it is often associated with grim clinical outcomes and high mortality rates despite diagnostic1,2 and therapeutic4 advances. Complications are expected to rise exponentially as the number of TAVIs continues to rise at a constant rate in our setting as well.5

Few studies have compared the incidence of IE on TAVI and surgical aortic valve replacement (SAVR). However, randomized clinical trials have shown similar annual incidences of IE after SAVR and TAVI.6,7 Due to the strict patient selection of such studies, the results obtained in each center during the routine clinical practice may vary, and few studies like this using dedicated databases have been conducted. A multicenter study8 revealed the characteristics of IE on TAVI vs SAVR and the outcomes of a large patient cohort, yet it did not analyze the incidence of IE. Assessing the outcomes of this severe complication at each center should be a priority due to its potential impact on decision-making for the heart team.3

The aim of the present study was to analyze both the incidence and mortality rates of IE on TAVI vs SAVR patients from a tertiary teaching hospital in a long series of patients with severe aortic stenosis.

METHODS

Study design and population

We conducted an observational, single-center, retrospective study of a prospective cohort including all cases of IE in symptomatic patients with severe aortic valve stenosis treated with TAVI or SAVR using a biological prosthetic valve, diagnosed, and followed at a tertiary teaching referral hospital by an endocarditis team from January 2008 through December 2022.

Endpoints and definitions

The primary endpoint of the study was to analyze the overall incidence and mortality rates of IE on TAVI at our center. Secondary endpoint was to compare both rates in patients with IE on TAVI and SAVR during the same period. Another secondary endpoint was to study the number of patients who underwent surgery for IE in both groups. IE was diagnosed using the Duke criteria9 or the 2015 European Society of Cardiology modified criteria10 depending on the time of diagnosis in the study period. Cases of IE in TAVI patients during this period were identified, and their characteristics were compared with those of IE in SAVR patients.

Follow-up

Follow-up events were defined based on the criteria established by the Valve Academic Research Consortium-2.11 All complications, need for surgery, and mortality at the follow-up were recorded. IE-related cardiac complications included decompensated heart failure, fistula, prosthetic dehiscence, abscess, and complete atrioventricular block. Systemic complications included acute kidney failure, sepsis, embolism, and disorders of the central nervous system. Early IE was defined as that occurring within the first year after TAVI or surgery based on European recommendations.10 Since 1987, a prospective protocol for inclusion and follow-up of all IE cases has been in place at our center, with a systematic registry including in-person visits, at least, annually for all patients, as well as phone consultations when needed.

Statistical analysis

Qualitative data are expressed as percentages, and continuous data as mean and standard deviation or median [interquartile range], depending on whether they follow a normal distribution. Inter-group comparisons were drawn using the chi-square test or Fisher’s exact test for qualitative variables, and the Student’s t-test or Mann-Whitney U test for continuous variables, as appropriate. Time-to-event analyses for all-cause mortality were conducted using Kaplan-Meier curves. All tests were two-sided, and P values < .05 were considered statistically significant. Statistical analyses were performed with SPSS software (version 24; IBM Corp, Armonk, NY, United States).

RESULTS

Study population and incidence rate

During the study period, a total of 778 patients successfully underwent TAVI, 70% of them with self-expanding valves. After a median follow-up of 49 months (p25-p75, 29–108 months), 10 cases of IE were eventually diagnosed, which amounts to an overall incidence rate of 1.29% (an incidence rate of 0.09/100 patient-years). Twenty-four of the 1457 patients treated with surgical bioprostheses were diagnosed with IE for an overall incidence rate of 1.64% (an incidence rate of 0.12/100 patient-years). The hazard ratio for the IE incidence rate between the 2 groups was 0.75 (95% confidence interval, 0.36–1.57). The incidence rate of IE on TAVI remained stable throughout the study period, as did the incidence of IE on surgical bioprostheses. Four of the 10 IE cases in TAVI occurred between 2008 and 2015—incidence of 1.21%—and 6 between 2016 and 2022 (incidence of 1.33%). The incidence rates of IE on surgical bioprostheses for the same periods were 1.62% and 1.67%, respectively. Clinical characteristics, treatment, and mortality rates were also similar across periods in both IE groups.

Characteristics of IE in TAVI and SAVR

Approximately half of the cases reported in both groups turned out to be early IE: 5 of the 10 TAVI IE cases and 11 of the 24 cases described on surgical bioprostheses occurred within the first year after implantation (50% TAVI and 46% SAVR; P = .56). The remaining cases were late prosthetic IE (table 1). Among TAVI IE cases, the 5 early ones were diagnosed 2, 4, 6, 8, and 11 months after implantation, while the 5 late ones were diagnosed within years 2 (3 cases) and 3 (2 cases). Among the IE reported on surgical bioprostheses, early cases were diagnosed within the first 2 months (2 cases), between months 3 and 6 (3 cases), and between months 6 and 12 (5 cases) after surgery. Regarding late cases, 5 occurred in year 2, 4 in year 3, and 4 > 3 years after surgery.

Table 1. Baseline characteristics of patients with infectious endocarditis after transcatheter aortic valve implantation or surgical aortic valve replacement

A possible source of infection different from implantation or surgery was identified in 2 of the 10 TAVI IE cases (20%) and in 2 of the 24 surgical bioprostheses. IE cases (8.3%). Only 1 of the early IE cases was associated with a possible infection source: a colonoscopy in a TAVI IE patient due to Enterococcus. None of the SAVR IE cases had an identified source, with implantation or the perioperative period being regarded as the probable sources of infection in the remaining patients. Among late IE cases, 1 TAVI IE case was associated with a dental procedure—despite proper antibiotic prophylaxis—and 2 SAVR IE cases with an upper gastrointestinal endoscopy and a dental visit, respectively. These patients’ clinical characteristics are shown in table 1. TAVI patients were older, with a median [interquartile range] age of 76 years [67–85] vs 63 years [49–77] of those treated with surgical bioprostheses (P < .001) and had higher rates of diabetes (70% vs 29%; P = .034). However, there were no significant differences in sex, other comorbidities, or symptoms. As expected, the TAVI group had a significantly higher EuroSCORE II vs the SAVR group (table 1). The profile of causative pathogens was very similar between the 2 groups (table 2). Enterococci were the pathogens most widely identified, followed by coagulase-negative staphylococci and Staphylococcus aureus, also with no significant differences being reported between the 2 groups. In 17% of SAVR patients and 20% of TAVI patients, the causative agent could not be identified (figure 1). Diagnostic echocardiographic findings of IE were also very similar between the 2 groups: transthoracic echocardiography identified IE only in half of the TAVI cases vs 37% of the cases treated with surgical bioprostheses (table 2). Transesophageal echocardiography diagnosed all IE cases in the TAVI group and slightly fewer in the SAVR group (table 2 and figure 1). In TAVI IE patients, the echocardiographic lesions identified included 9 endocardial vegetations—all affecting the leaflets—with 3 stent expansions (figure 2), 1 periaortic abscess, and 1 right ventricular perforation.

Table 2. Microbiological profile of the most common microorganisms and diagnostic lesions in the echocardiogram of patients with infectious endocarditis after transcatheter aortic valve implantation or surgical aortic valve replacement

Disease progression

The course of the disease was similar in the 2 groups (table 3). Half of the TAVI group experienced cardiac complications, as did one-third of the SAVR group, without any significant differences being reported. Half of the patients had a surgical indication due to IE (40% in the TAVI group vs 50% in the SAVR group; P = .49), but among those, only 20% of the TAVI group and 25% of the SAVR group eventually underwent surgery (P = .93) (table 3). The in- hospital mortality rate was similar (20% in the TAVI group vs 25% in the SAVR group; P = .51), without any significant differences being reported in the 1-year mortality rate, which remained high—at approximately 30%—in both groups (table 3, figure 1, and figure 3).

Table 3. Complications, rate of surgical procedures, and 1-year mortality rate in patients with infectious endocarditis after transcatheter aortic valve implantation or surgical aortic valve replacement

Figure 1. Infective endocarditis (IE) after transcatheter aortic valve implantation in a cohort of 778 patients vs a cohort of patients with surgical bioprostheses-related IE. SAVR, surgical aortic valve replacement; TAVI, transcatheter aortic valve implantation; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography.

Figure 2. Transesophageal echocardiogram showing a large vegetation on the leaflets of a transcatheter aortic valve (arrows).

Figure 3. Kaplan-Meier curves showing the 1-year mortality rate in patients with infective endocarditis after transcatheter aortic valve implantation (TAVI) or surgical aortic valve replacement (SAVR).

DISCUSSION

The main finding of our study was that the incidence rate of IE on TAVI is low in our setting and similar to that of surgical aortic bioprostheses, despite being patients with higher surgical risk, older age, and more comorbidities. These results are similar to those reported in the literature (1% up to 6%);4 however, more recent large TAVI trials suggest lower rates. The PARTNER 3 trial reported annual rates of 0.2%,1 and similarly, the low-risk Evolut study2 found incidence rates of 0.1% and 0.2% at 30 days and 1 year, respectivelu, which is more consistent with our rates. Studies directly comparing the incidence of IE after SAVR and TAVI are scarce, and some have produced contradictory results.7,12 However, most observational studies based on large national databases and randomized clinical trials have found no statistically significant differences on this regard,6,13,14 even though TAVI patients are older and generally have more comorbidities, which is consistent with our results. There were also no differences in our study regarding early IE incidence (50%), which was similar in both groups, although slightly lower for IE on TAVI vs what has been reported by the literature (rates up to 64%.)15 However, a multicenter study conducted in Spain found a higher rate of early IE in TAVI vs SAVR patients (78.1% vs 39.3%; P = .001), which is significantly higher than that described in our series for the TAVI group.16 This could be explained by the different antibiotic prophylaxis regimens and procedures used at each center.

Secondly, another relevant finding from our study is that enterococci were the main cause of IE in both patient groups, which is consistent with the literature on TAVI-related IE17,18 but not on surgical bioprostheses-related IE, of which S. aureus is usually the main microorganism involved. These differences are difficult to explain, as the increase in enterococcal incidence as a causal agent in TAVI is primarily associated with older patient age and transfemoral access but still would not explain why it is also the most common pathogen in SAVR-related IE.4 Compared with a multicenter Spanish series of similar characteristics to ours,16 there are some differences: the most common causative microorganism of IE—in both TAVI and SAVR— was Staphylococcus epidermidis, followed by enterococci— less frequent than in our series—and thirdly, S. aureus. These differences could be explained by varying antibiotic prophylaxis regimens used in different settings, underscoring the critical importance of understanding the most common microorganisms involved in prosthetic IE, in general, to apply the most appropriate and effective prophylaxis regimen.

Another noteworthy aspect of this study is that echocardiographic lesions diagnostic of IE could be identified in 100% of TAVI- related IE cases vs 67% of SAVR-related IE cases in surgical pros­theses. This contrasts sharply with most reports, which indicate that the combined sensitivity rate of transthoracic and transesophageal echocardiography was 67.8% in TAVI patients, 73% in SAVR patients, and nearly 90% in native valves.19 However, a very recent study comparing patients with IE after TAVI or SAVR confirmed that vegetations were identified via echocardiogram in up to 82% of the TAVI group, more in line with our findings and significantly higher than the diagnostic rate of IE reported in surgical prostheses (62.5%; P < .001).8 Our results also differ from those reported in the literature in that most lesions found via echocardiography were vegetations, whereas other authors, such as Salaun et al.,20 reported vegetations in only 5 out of 11 cases diagnosed with TAVI-related IE, with the remaining cases being atypical lesions.

Lastly, this study also highlights the course of the disease, with a 1-year mortality rate of 30%, which is lower than that reported in other published studies (between 33% and 66%).4,12,21-23 The in-hospital mortality rate was also lower than that reported in other studies, such as the international multicenter registry of post-TAVI endocarditis21 (36% vs 20% in our series) and much lower than the Spanish multicenter trial (35%).16 These differences may be due to the wide variability in patient characteristics across studies. Of note, there were no significant differences in mortality when comparing surgical bioprostheses-related IE, unlike other series reporting lower mortality for the latter vs TAVI-related IE.23 However, a study published by Panagides et al.,8 comparing early and 1-year mortality in IE after TAVI and SAVR using propensity score matching found no significant differences between the 2. In our series, only 20% of TAVI patients with a surgical indication underwent surgery, which is consistent with other studies where surgical rates were similar to ours, with optimal medical therapy being the most common strategy.4,12,16,22,23 Some studies found no improvement in prognosis for these patients undergoing surgery, with similar mortality rates associated with optimal medical therapy,24 while a meta-analysis published by Tinica et al.25 showed that the surgical strategy was significantly superior to conservative therapy. In any case, given the lack of strong evidence, the treatment strategy for TAVI-related IE remains unclear, even if the presence of complications warrants surgical indication, and decisions should be based on local expertise.

Study limitations

Our study has the inherent limitations of its observational design, with data collected over many years, during which diagnostic criteria for IE, valve types—towards better technical models—and procedural approaches—towards less invasive methods—have evolved. The small number of IE cases in the 2 groups means results should be interpreted with caution. As this is a crude analysis, the presence of confounding bias cannot be ruled out; nevertheless, data show real-life outcomes. Additionally, patients with mechanical valves were not included, so results cannot be extrapolated to this group.

CONCLUSIONS

In our setting, TAVI-related IE has a low overall incidence rate, with no significant differences vs SAVR-related IE, despite involving older and more comorbid patients. Regarding the causative microorganism of IE—both in surgical and percutaneous surgical bioprostheses—enterococci were the most common pathogen. There were no differences in mortality between the 2 types of aortic valves, and treatment for IE was predominantly conservative.

FUNDING

None declared.

ETHICAL CONSIDERATIONS

The study was conducted in full compliance with the Declaration of Helsinki and approved by the Clinical Research Ethics Committee at the beginning of the registry in 1987. However, for patients with TAVI or SAVR, all participants signed an informed consent form authorizing the collection and analysis of their data for research purposes. Since this was an observational study, patient treatment was not impacted. For this new registry analysis, approval was obtained from our center ethics committee. Data were handled completely anonymously in full compliance with the provisions of Organic Law 3/2018 of the Spanish Data Protection Authority.

STATEMENT ON THE USE OF ARTIFICIAL INTELLIGENCE

No artificial intelligence was used in the development of this work.

AUTHORS’ CONTRIBUTIONS

A. Roldán participated in data collection and analysis. C. Urbano participated in data collection and analysis. N. Aguayo participated in patient identification, data collection, and analysis. M. Crespín, J. López, and J.C. Castillo participated in the conception of the article and its interpretation. R. González contributed to statistical analysis and result interpretation. D. Mesa and M. Ruiz contributed to data processing, analysis and result interpretation, and collaborated in the revision and preparation of the manuscript for publication. J. Perea, I. Gallo, J. Suárez de Lezo, and S. Ojeda participated in the conception of the work, helped gather patient information, and provided guidance on literature review and manuscript drafting. M. Pan and M. Anguita supervised all stages of manuscript drafting, from conception, data collection, and result interpretation, to revision, correction, and preparation of the article for submission.

CONFLICTS OF INTEREST

S. Ojeda is an associate editor of REC: Interventional Cardiology. The journal’s editorial procedure to ensure impartial handling of the manuscript has been followed. The remaining authors declared no conflicts of interest whatsoever.

REFERENCES

INTRODUCTION

Transcatheter aortic valve implantation (TAVI) has become an established option for symptomatic severe aortic stenosis across risk groups.1,2 The implantation technique is moving towards a less invasive procedure with shorter hospitalizations.3,4 Because the length of hospital-stay has been related to major adverse cardiovascular events during follow-up,5 optimizing the procedure remains a concern. Different studies have demonstrated that patients can be safely discharged 72 and even 48 hours after TAVI, and various protocols with a minimalist approach have been described so far, mainly with balloon-expandable valves.6-9 Nevertheless, the potential need for a permanent pacemaker implantation (PPI) frequently prolongs hospital admissions for the purpose of rhythm monitorization after TAVI.10,11 This issue is particularly relevant for self-expanding valves,12-14 since the rate of PPI is highest amongst these models and most early-discharge strategies have been reported for balloon-expandable prostheses.9,15

The rate of PPI can be reduced by guiding TAVI by the cusp-overlap projection (COP) to isolate the non-coronary cusp and better control prothesis implantation depth.16,17 In addition, sequential rapid atrial pacing (RAP) after deployment has been proposed to screen for atrioventricular (AV) conduction disorders.18 However, the effect of combining both preventive and diagnostic techniques on clinical outcomes after self-expanding TAVIs remains unknown. Therefore, the present study was designed to determine whether compared to conventional care, the combined use of COP and RAP can improve the management of conduction disturbances and the early outcomes after self-expandable TAVI procedures.

METHODS

Study population

This is a one-center retrospective analysis of consecutive patients who underwent transfemoral self-expandable TAVI between January 2018 to January 2022. All patients had symptomatic severe aortic stenosis and were deemed eligible for TAVI by a multidisciplinary heart team. We included all transfemoral cases irrespectively of surgical risk. From this cohort, we selected patients in sinus rhythm undergoing a self-expandable prothesis implantation procedure. We excluded patients with permanent atrial fibrillation or permanent pacemakers, and patients who did not undergo follow-up in our center. Procedures performed before 2018 were excluded from our analysis to test our hypothesis on a homogeneously contemporary series.

The institutional Ethics Committee approved the study, and all subjects provided written informed consent.

TAVI procedure

Pre-TAVI workup included baseline electrocardiogram (ECG), transthoracic echocardiography, coronary and peripheral angiography, as well as computed tomography. Implanting technique was standard among operators, following the conventional “minimal approach” previously described3,4,19: all invasive lines were minimized, and procedures were performed under minimal sedation and local anesthesia. Femoral access was obtained using ultrasound guidance and double percutaneous vascular closure (Perclose, Abbott, MN, United States) was placed before non-fractionated heparin administration. During the procedure, all patients received a temporary pacemaker via femoral venous access using a balloon-tipped bipolar catheter (5 French, Arrow, Teleflex, USA) positioned in the right ventricular apex. Pre and post valvular dilation during the intervention were performed according to the operator’s judgement. After implantation, a supravalvular angiography was routinely performed to check for paravalvular leak. At the end of the procedure, an angiography of the main vascular access was performed on all patients to verify arteriotomy closure success as well as ilio-femoral vessel integrity. Until May 2020 (standard-of-care cohort), the temporary pacemaker was kept for 24 hours with monitorization in the cardiac intensive care unit, with later transfer to the cardiology ward. Permanent pacemaker was implanted in the presence of persistent or new-onset complete or high degree AV block.

Since June 2020, the procedure was modified to routinely include: a) COP during valve deployment to assess implantation depth, and b) RAP at the end of the procedure in the absence of complete AV block, at rates of 70 to 120 beats/min (or until pacing-induced Wenckebach heart block was observed) in 10 beats/min increments for a total of 20 beats at each increment.18 In patients with persistent complete AV block, permanent pacemaker was placed in the same day. In the absence of immediate complications, patients were transferred to a standard monitored ward bed minimizing ICCU stay unless clinically needed, and temporary pacemaker surveillance was only maintained for 24 hours in the presence of Wenckebach phenomenon or in patients with AV block until permanent pacemaker was implanted. Final determination of patient discharge was based on patient’s characteristics and clinical status by the referring physician.

Follow-up

After implantation, assessment included ECG, echocardiogram, in-hospital, and 90-day clinical follow-up. All patients’ clinical, procedural, in-hospital information, and follow-up data were collected from the electronic medical records.

Study endpoints

The primary safety and efficacy outcomes were complete AV block after implantation with COP and the diagnostic performance of RAP added to COP in self-expandable valves.

The secondary endpoints included the length of hospitalization following TAVI and the composite major adverse cardiovascular events, defined as cardiovascular mortality, PPI, rehospitalization for heart failure decompensation, stroke, and major bleeding and vascular complications occurring 90 days after hospital discharge. In addition, rehospitalizations for healthcare-associated infections within the same period were analyzed separately.

Statistical analysis

Categorical variables are presented as counts (percentages) and were compared using the chi-square test. Continuous variables were tested for normality of distribution by using the Shapiro-Wilk test. According to their distribution, continuous variables are expressed as mean ± standard deviation or median values with interquartile range [IQR], and were compared using Student t test, ANOVA or Wilcoxon test, as appropriate. For the multivariate analysis, we first selected baseline variables that showed a significant association with the outcome (P < .2) and performed single logistic regression. Variables with a P-value < .05 were entered into a multiple logistic regression analysis. Odds ratios (OR) were calculated using uni and multivariate logistic regression, along with 95% confidence intervals. Statistical analyses were conducted using RStudio 4.1.20 P-values < .05 were considered significant.

RESULTS

Baseline characteristics

A total of 273 patients met the inclusion criteria for the analysis. Among them, 134 underwent standard implantations, and 139 underwent the combined use of COP and RAP (COP+RAP) (figure 1). Mean patient age was 81 ± 7 years old and 51.3% were female, with similar prevalence of cardiovascular risk factors and comorbidities between groups. Patients in the standard-of-care group showed a worse NYHA functional class, and there were no differences in the surgical risk (table 1). Both cohorts presented similar rates of conduction disturbances at baseline. The mean transvalvular pressure gradient by ultrasound was 47 [41-58] mm Hg with no significant differences on baseline findings between groups. The COP+RAP group showed a higher valvular calcium score (2429 [1577-3557] Agatston units) and larger left ventricular outflow tract and annular perimeters (P = .006 and P = .04 respectively) (table 1).

Figure 1. Patient flow chart. AF, atrial fibrillation; COP, cusp-overlap projection; RAP, rapid atrial pacing TAVI, transcatheter aortic valve implantation.

Table 1. Demographics, baseline electrocardiographic and imaging characteristics

Procedure and in-hospital outcomes

Procedural details are listed in table 2. All patients underwent TAVI under conscious anesthesia, with more frequent use of midazolam and fentanyl in the COP+RAP cohort in place of dexmedetomidine (P < .001). Also, in line with current recommendations for self- expandable valves, the rate of predilation was significantly higher in the COP+RAP group (P < .001).

The occurrence of intraprocedural conductance disorders was lower with COP in the COP+RAP group (25.8% vs 14.4%; P = .007; OR, 0.49; 95%CI, 0.27-0.91; P = .026) mainly due to a reduction in temporary complete AV block (table 2), whereas rates of new-onset first degree AV block and atrioventricular block (LBBB) remained similar between groups. Same day permanent pacemaker was implanted in 16 patients of the COP+RAP group with persistent AV block. RAP revealed Wenckebach phenomenon at a median stimulation rate of 115 bpm in 24 patients and allowed to reduce the need for temporary pacemaker surveillance from 91.8 % in the standard-of-care group to 28.1 % of the cases in the COP+RAP group (P < .001). Among the patients in the COP+RAP cohort who did not develop Wenckebach phenomenon (n = 98/122, 80.3%), permanent pacemaker was required in only 3% of patients (OR, 0.16; 95%CI, 0.03-0.76; P = .02; negative predictive value = 97%; 95%CI, 91-99%. Patients with pacing-induced Wenckebach had a higher risk of PPI during admission (OR, 6.33; 95%CI, 1.30-34.29; P = .021; positive predictive value 79% 95%CI, 58-93%).

Table 2. Procedural characteristics, electrocardiographic and imaging outcomes

Incidence of AV block and timing of PPI after TAVI is shown in figure 2. Delayed AV block (developed > 48 hours after TAVI) during admission occurred in 8.2% patients in the standard group and 3.6% of the COP+RAP group (P = .14) mostly in the presence of new-onset LBBB (68% of the cases). Adding both maneuvers did not impact the procedure length or led to adverse events, with a low and similar rate of in-hospital complications among groups (table 3). Additionally, reduction of the implantation depth did not lead to a higher rate of residual significant aortic regurgitation (P = .32).

Figure 2. Immediate complete AVB and on-admission permanent pacemaker implantation trends. Upper: intraprocedure cusp-overlap projection and Wenckebach phenomenon during atrial pacing. Midd: comparison of immediate AVB in standard and protocol groups. Bottom: trend in pacemaker implantation during admission. AVB, atrioventricular block; COP, cusp-overlap projection; RAP, rapid atrial pacing TAVI, transcatheter aortic valve implantation.

Table 3. Immediate, and short-term outcomes

The reduction of stay in the intensive care unit along with the lower rate of 24-hour temporary pacemaker surveillance, significantly decreased the length of hospitalization in the COP+ RAP cohort (−3.0 [−2.5 to −3.5] days, P < .0001), enabling to discharge patients within 48 hours in 56.8% of the cases compared to 7.5% in the standard-of-care group (P < .001).

90-day outcomes

At 90-days, the cumulative rate of PPI in the COP+RAP cohort persisted lower than in the standard-of-care group (29.5% vs 18.7%; P = .049). Of all 65 pacemaker implantations, 58 (89%) occurred during the index admission and 7 (11%) following hospital discharge. Among the 3 patients from the COP+RAP requiring late PPI, 2 developed with new-onset LBBB (QRS width 127 and 136 ms respectively), and both had displayed Wenckebach phenomenon at 120 bpm. They were monitored with temporary pacemaker for 24 hours without events but developed high degree AV block within 48 hours after discharge (day 4 after TAVI). The third patient had new-onset first degree AV block (PR interval 240 ms), Wenckebach phenomenon at 120 bpm and narrow QRS (118 ms) but developed high-degree AV block at day 55. No patients with a negative RAP test required late PPI. The univariate logistic analysis of potential baseline and procedural factors of PPI after TAVI is reported in table 4. Interestingly, differences in sedation, predilation, valve model (P = .98) and bicuspid valves (P = .71) did not influence the risk of PPI after TAVI. In the multivariate analysis, prior 1st degree AV block, baseline right bundle branch block, prosthesis posdilation and new-onset LBBB were associated with PPI. The combination of COP with RAP decreased the need of PPI on a 90-day follow-up (OR, 0.48; 95%CI, 0.24-0.92; P = .031) (figure 3).

Figure 3. Odds plot of multivariate analysis of 90-day permanent pacemaker implantation predictors. 95%CI, 95% confidence interval; AVB, atrioventricular block; COP, cusp-overlap projection; LBBB, left bundle branch block; PPI, permanent pacemaker implantation; RAP, rapid atrial pacing; RBBB, right bundle branch block.

Table 4. Main factors associated with PPI after TAVI: unadjusted logistic regression analysis of PPI predictors at 90 days

The COP+RAP group presented a significantly lower percentage of the secondary composite endpoint than the standard-of-care cohort (20.1% vs 10.1%; P = .031). The rates of HF readmission, and mortality were low and similar across groups during the follow-up period (table 3). None of the 3 deaths during follow-up were due to rhythm disturbances. At 90 days, there was a reduction in the number of readmissions due to admission-related infections in the COP+RAP cohort (OR, 0.35; 95%CI, 0.12-0.89 P = .036) (figure 4), with an 8.6 % reduction in the odds of infection for each day of reduced hospitalization (OR, 0.91; 95%CI, 0.85-0.99; P = .023). The most common were urinary (47.4%) and respiratory (42.1%) infections. None of the patients presented infections related to PPI.

Figure 4. Central illustration. Main immediate and short-term findings of combining cusp-overlap projection and rapid atrial pacing when implanting self- expandable valves. OR, odds ratio; TAVI, transcatheter aortic valve implantation.

DISCUSSION

In the present study we demonstrate that combining the COP with RAP, COP reduces the risk of complete AV block after self- expanding TAVI and negative RAP after implantation adequately dismiss underlying AV conduction disturbances requiring PPI.

This strategy reduces the need of temporary PPI and the appearance of late onset AV block, and safely allows an early discharge in sinus rhythm patients receiving a self-expandable valve. Reducing hospital stays lead to a significant reduction in admission-related infections at short-term follow-up (figure 4).

A minimalistic approach in TAVI has become progressively relevant as a way of simplifying the intervention and promoting early recovery.4,6-8 Multiple studies have shown that early discharge after TAVI is possible, preserving the high standards of effectiveness, patient safety and outcomes.3,9,21 However, most of these studies included highly selected patients, with a predominance of balloon-expandable valves, and might not be representative in the current TAVI scenario. Evidence at the beginning of minimalistic TAVI inception found self-expanding prothesis to be a predictor for rhythm disturbances and delayed discharge.5,6 Since their risk of AV conduction disturbances remain in the range of 5% to 30% even with contemporary transcatheter heart valves10,12-14 and there is no appropriate way to assess the risk of AV block development, self-expandable valves have been underrepresented in early discharge protocols so far.3,7,22,23

Postimplantation conduction disturbances remains one of the current weaknesses in TAVI, and great heterogeneity of recommendations persist regarding their management.24,25 In this regard, the aim of our study was to minimize the appearance and improve the assessment of conduction disturbances without increasing invasiveness in patients receiving self-expandable valves, by combining 2 well-described maneuvers.

Guiding the procedure using the COP has already been established as an effective method to guide implantation depth.16,17 Isolating the non-coronary cusp allows a better identification of the most inferior point of the aortic annulus, leading to a higher (more aortic) valve positioning, which in turn reduces the risk of post-TAVI conduction disturbances. Various observational studies have described the COP technique in distinct types of self-expandable valves,15-17,26 and highlighted its potential benefits in decreasing the PPI rate in the Evolut (Medtronic, United States) and the Portico (Abbott Structural Heart, United States) valves.27 In line with these studies, in our cohort, COP significantly decreased the emergence of intra-procedural transient and late onset ( > 48 hours) high-grade AV block irrespectively of the self-expandable valve type, with no noticeable impact on the rate of new-onset LBBB, that remained around 29% in both cohorts.

After implantation, current consensus documents recommend temporary pacemaker surveillance for at least 24 hours. Further management depends on baseline intra-procedure, and post-procedure conduction disturbances,25,28 that can result in longer duration of temporary pacemaker placement or prolonged inpatient rhythm monitoring. In this context RAP, emerges as an easy and interesting strategy to discriminate which patients may require monitoring. Since its publication in 2020,18 different works have recognized the underuse of this technique, although is an important component of routine electrophysiological studies to “stress” the AV conduction.10 It has the limitation of assessing specifically at the level of the atrio-His interval but seems to be a good indicator of patients with underlying conduction disorders with a described strong negative predictive value for PPI in the absence of Wenckebach phenomenon. In our study, the rate of Wenckebach phenomenon during dynamic atrial stimulation was 19.8% and showed a moderate positive predictive value. However, most importantly, the absence of RAP-induced AV block displayed a very high negative predictive value of PPI during follow-up. These results are in agreement with the previous report of RAP on balloon-expandable valves,18 showing adequate reliability of AV Wenckebach testing over delayed conduction disturbances immediately at the end of the procedure in self-expandable valves. The improvement in valve deployment and risk stratification of AV conduction disturbances allowed us to reduce temporary pacemaker surveillance in 63.1% of cases, with less late-onset high-degree AV block and no impact in procedure length or complications, leading to early recovery and minimizing functional decline. Moreover, although early discharge protocols described so far have focused on demonstrating safety, it should be noted that in our simplified protocol cohort there was a significant reduction in the healthcare-related infection readmission rate at 90 days. The reduction in the incidence of admission-related respiratory and urinary infections was likely attributable to the shortened length of hospitalization, that has been associated with adverse events.5 Observed differences in major adverse cardiovascular events, although significant, should be interpreted cautiously, as part of the findings might be a result of technological advances and cumulative expertise.

The adequacy of these combined maneuvers was patent in the multivariate analysis and in the maintained differences in PPI rate at a 90-day follow-up. Nevertheless, special caution should be recommended in patients developing persistent LBBB after TAVI, as it was a predictor of delayed AV block and PPI. These patients would probably benefit from longer monitoring, and in the presence of new-onset LBBB or Wenckebach-induced phenomenon, a formal electrophysiological study or continuous ECG monitoring might have the potential to risk-stratify patients with unclear indications of PPI following TAVI.29-31

Limitations

This is a single-center study with a modest sample size and an observational design; therefore, there is an inherent potential bias due to patient selection. First, both cohorts are not contemporary, which may introduce bias related to technological advancements or increased operator experience over time. Second, RAP is restricted to patients in sinus rhythm, and the cutoff of 120 beats/min for maximum RAP rate was selected according to previous literature but may be insufficient as low risk patients tend to be young and might request testing at higher heart rates. Also, RAP stresses the AV conduction at the atrio-His level, and therefore our results might not properly assess patients with broadened QRS, as high-grade conduction block occurs at the His-Ventricular segment.18,30,31 Third, our protocol has the inconvenience of requiring temporary pacemaker so it can be moved to the atrium at the end of the procedure, which is not feasible in procedures performing pacing through the left ventricular wire. However, we believe it might be a safe and efficient way to risk-stratify patients immediately after the procedure. Finally, our study was focused on patients in sinus rhythm and self-expanding prothesis and therefore the results and conclusions should be interpreted in this context.

CONCLUSIONS

Combining COP during deployment with RAP immediately after TAVI diminishes the incidence of postprocedural AV block and improves the screening of underlying atrio-His conduction disturbances in patients receiving a self-expandable valve. In turn, this allows to reduce the length of temporary pacemaker surveillance and admission days, with differences in PPI rate maintained at a 90-day follow-up. Optimizing strategies of care after TAVI has a significant impact at short-term, driven by a significant reduction in readmissions due to infections.

FUNDING

M. Tamargo was partially supported by grants from the Fundación para la Investigación Biomédica Gregorio Marañón, Spain, and Rio Hortega CM20/00054 from the Instituto de Salud Carlos III, Spain. J. Bermejo was partially supported by INT22/00025 from the Instituto de Salud Carlos III, Spain.

ETHICAL CONSIDERATIONS

The institutional Ethics Committee approved the study, and all subjects provided written informed consent. This manuscript complies with the guidelines of the SAGER guidelines, and possible gender and/or age differences have been taken into account, without significant differences being observed.

STATEMENT ON THE USE OF ARTIFICIAL INTELLIGENCE

No artificial intelligence has been used in the development of this paper.

AUTHORS’ CONTRIBUTIONS

All authors comply with the international characteristics of authorship of scientific articles. Conception and design: M. Tamargo, J. García Carreño, E. Gutiérrez, J. Bermejo, F. Fernández-Avilés. Data collection: M. Tamargo, M. Huanca, E. Ludeña, J. García Carreño. Analysis: M. Tamargo E. Gutiérrez, J. Bermejo. The manuscript was reviewed and approved by all authors.

CONFLICTS OF INTEREST

Nothing to disclose.

REFERENCES