Author + information
- Received December 22, 2017
- Revision received February 26, 2018
- Accepted April 3, 2018
- Published online June 18, 2018.
- Philipp Ruile, MDa,∗ (, )
- Jan Minners, MDa,
- Philipp Breitbart, MDa,
- Simon Schoechlin, MDa,
- Michael Gick, MDa,
- Gregor Pache, MDb,
- Franz-Josef Neumann, MDa and
- Manuel Hein, MDa
- aDepartment of Cardiology & Angiology II, University Heart Center Freiburg-Bad Krozingen, Bad Krozingen, Germany
- bDepartment of Radiology, Section of Cardiovascular Radiology, University of Freiburg, Freiburg, Germany
- ↵∗Address for correspondence:
Dr. Philipp Ruile, Clinic for Cardiology II, University Heart Center, Südring 15, 79189 Bad Krozingen, Germany.
Objectives The aim of this study was to investigate medium-term outcomes in patients with leaflet thrombosis (LT).
Background The clinical significance of early LT after transcatheter aortic valve replacement, diagnosed by computed tomography angiography in approximately 10% of patients, is uncertain.
Methods In this observational study, computed tomographic angiography was performed a median of 5 days after transcatheter aortic valve replacement and assessed for evidence of LT. Follow-up consisted of clinical visits, telephone contact, or questionnaire.
Results LT was diagnosed in 120 of 754 patients (15.9%). Patients with LT were less likely male (36.7% vs. 47.0%, p = 0.045), with a lower rate of atrial fibrillation (28.3% vs. 41.5%, p = 0.008). Peri- and post-procedural characteristics were comparable between groups (e.g., valve implantation technique; p = 0.116). During a median follow-up period of 406 days, there were no significant differences in the primary endpoint of all-cause mortality and the secondary combined endpoint of stroke and transient ischemic attack between patients with LT and those without LT (18-month Kaplan-Meier estimate for mortality 86.6% vs. 85.4%, p = 0.912; for stroke- or transient ischemic attack–free survival 98.5% vs. 96.8%, p = 0.331). In univariate and multivariate analyses, LT was not predictive of either endpoint, whereas male sex (p = 0.03), atrial fibrillation (p = 0.002), and more than mild paravalvular leak (p = 0.015) were associated with all-cause mortality.
Conclusions In this prospective observational cohort undergoing post–transcatheter aortic valve replacement computed tomographic angiography, LT was not associated with increased mortality or rates of stroke over a follow-up period of 406 days.
Transcatheter aortic valve replacement (TAVR) is the current standard therapy for severe symptomatic aortic stenosis in patients with increased perioperative risk (1–3). Early leaflet thrombosis (LT) as diagnosed on computed tomographic angiography (CTA) after TAVR was first reported in 2013 (4). Subsequent studies have confirmed these findings and described an incidence of LT ranging from 4% to 40% for various aortic transcatheter heart valves (THVs) (5–10). During short-term follow-up, LT does not seem to influence outcomes, with unchanged rates of stroke or heart failure reported in a number of smaller studies (6,8–10). However, a recent report from Chakravarty et al. (8) suggested an increased rate of transient ischemic attacks (TIAs) but no influence on the rate of stroke or mortality. The purpose of the present study was to investigate medium-term outcomes in patients with LT from a large, single-center cohort.
This observational cohort study was approved by the local Institutional Review Board and complied with the Declaration of Helsinki. All echocardiographic, computed tomographic angiographic, and clinical data were obtained from our institutional database. Routine post-TAVR CTA was performed before discharge in all patients who underwent TAVR between May 2012 and June 2017 at our institution. Reasons for not performing CTA have been described previously and include renal insufficiency, frailty, and others (5). The following commercially available balloon- or self-expandable THVs were implanted: SAPIEN XT, SAPIEN 3 (both Edwards Lifesciences, Irvine, California), CoreValve, Evolut R (both Medtronic, Minneapolis, Minnesota), Lotus (Boston Scientific, Natick, Massachusetts), Portico (St. Jude Medical, St. Paul, Minnesota), Symetis ACURATE (Symetis, Ecublens, Switzerland), and JenaValve (JenaValve Technology, Munich, Germany).
THV implantation procedure
Eligibility for TAVR, procedural feasibility, preferred access route, and THV type and size were determined in consensus by a multidisciplinary, institutional heart team. Implantation procedures were performed under general anesthesia with fluoroscopic and transesophageal echocardiographic guidance via the transfemoral or the transapical approach (11,12). The majority of valves were deployed without pre-dilatation except for the SAPIEN XT THV. In case of small orifice area or extensive calcifications, pre-dilatation was performed as per operator’s discretion. All balloon-expandable valves were deployed under rapid pacing.
Peri- and post-procedural antithrombotic regime
Peri-interventional heparin 5,000 IU, with adjustment for high or low body weight, was administered in all patients. Before implantation, patients were administered either 400 mg effervescent aspirin (ASA) or a combination of 400 mg effervescent ASA and 600 mg clopidogrel. The peri- and post-interventional antiplatelet medication consisted of ASA (100 mg/day) alone from May 2012 to May 2014 and, since June 2014, dual-antiplatelet therapy with ASA (100 mg/day) plus clopidogrel (75 mg/day) for 6 months followed by lifelong ASA 100 mg/day. Anticoagulation treatment was paused as long as needed to achieve an international normalized ratio <2 in patients on vitamin K antagonist (phenprocoumon) or 24 h before TAVR for novel oral anticoagulant agents. One day after TAVR, anticoagulation was readministered in combination with clopidogrel.
Anticoagulation regimen in patients with LT
From May 2012 until May 2015, all patients with LT were treated with a modified antithrombotic regimen including a combination of phenprocoumon (target international normalized ratio 2 to 3) and clopidogrel 75 mg/day. Follow-up CTA was performed after 3 months, and phenprocoumon was stopped when signs of LT on CTA had resolved, unless anticoagulation was indicated for other reasons. Subsequently, patients were switched to dual-antiplatelet therapy. Repeat follow-up CTA was recommended at least 2 months after discontinuation of phenprocoumon. The results of this short-term follow-up have been published previously (5).
From May 2015 onward, patients with LT received the same antithrombotic regimen as those without LT, and anticoagulation was initiated only when indicated for other reasons (e.g., atrial fibrillation). In case of clinical signs of thrombosis or a relevant increase in mean transvalvular gradient, patients received anticoagulation according to the decision of the treating physicians.
Post-interventional echocardiographic assessment
Echocardiographic examinations were performed before discharge at the time of post-TAVR CTA by experienced cardiologists using an iE33 ultrasound system (Philips, Leiden, the Netherlands). According to the grading recommendations of the Valve Academic Research Consortium, paravalvular leakage was classified as follows: 0 = none, 1 = trace, 2 = mild, 3 = moderate, and 4 = severe. Mean pressure gradient was calculated according to the Bernoulli formula.
CTA was performed using a dual-source computed tomographic scanner (Somatom Definition Flash, Siemens Healthcare, Forchheim, Germany). Aortic root imaging was obtained in a craniocaudal direction using retrospectively electrocardiographically gated contrast-enhanced computed tomography with temporal resolution of 75 ms. Computed tomographic angiographic data were reconstructed in either 5% or 50-ms steps throughout the cardiac cycle with a section thickness of 1 mm and an increment of 0.8 mm using a stent-specific convolution kernel (B46f). All datasets were transferred to a dedicated post-processing workstation (Syngo Multimodality Workplace, Siemens Healthcare). Multiplanar reformations were used for THV evaluation. The detailed protocol has been published previously (5).
Computed tomographic angiographic datasets were assessed by an experienced cardiac radiologist as part of routine post-TAVR clinical follow-up. In the presence of findings suspicious for LT, CTA was reassessed by a second radiologist blinded to the initial findings. Prosthesis leaflets were dynamically assessed on multiplanar reformations for the presence of LT throughout the cardiac cycle. LT was defined as hypoattenuated thickening with or without rigidity of 1 or more leaflet segments in at least 2 different multiplanar reformation projections and 2 different reconstruction time intervals.
All patients with TAVR are monitored using a standardized follow-up protocol, including at 30 days, 1 year, and then yearly after the TAVR procedure, with contacts by questionnaire or telephone calls. For patients reporting events or symptoms suggestive of stroke or TIA, referring cardiologists and/or general practitioners were contacted for further information. The classification of TIA or stroke was based on the diagnosis of the attending neurologist. The attending neurologist was not aware of the diagnosis of LT, and all cerebral events were diagnosed and treated in external hospitals.
Until May 2016, all patients with LT were scheduled for follow-up CTA after 3 months, independently of their treatment regime. These results were published previously (9). Thereafter, no specific technical follow-up was performed, neither with CTA nor with echocardiography. In case of symptoms suggestive of obstructive valve thrombosis, patients were examined by echocardiography either in hospital or by the referring cardiologist. At the discretion of the treating physician, repeat CTA was performed.
Hemodynamic valve deterioration was defined as an increase of at least 10 mm Hg in mean gradient or a mean gradient >20 mm Hg with new onset of dyspnea, aggravation of dyspnea, or new onset of angina pectoris without progress of coronary heart disease (13).
Patients with and without LT after TAVR were compared with respect to the primary endpoint of all-cause mortality and the secondary combined endpoint of stroke and TIA. All statistical analyses were performed using SPSS version 23.0 (SPSS, Chicago, Illinois). Continuous variables are reported as mean ± SD, ordinal variables as median (interquartile range), and categorical data as number (percentage). We tested for a Gaussian distribution using the Kolmogorov-Smirnov test. For comparison between groups, continuous variables were tested using the unpaired Student's t-test in case of a normal distribution and the Mann-Whitney U test if the assumption of a normal distribution was rejected. Differences in categorical variables were tested by using the Pearson chi-square test. Kaplan-Meier curves and the log-rank test were used for visualizing and calculating the cumulative outcomes between both groups. Univariate and multivariate Cox proportional regression models were constructed to generate hazard ratios. We included all variables with p values <0.20 in univariate analysis (and introduced LT as a forced variable) into the multivariate regression.
Study population and study flow
From May 2012 to June 2017, 1,424 patients received THVs at our institution. In 754 of these patients (53%), CTA was performed a median of 5 days (interquartile range: 4 to 6 days) post-procedure: 604 patients (80.1%) received balloon-expandable valves (SAPIEN 3 or SAPIEN XT), 109 (14.5%) received self-expandable valves (CoreValve or Evolut R), and 41 (5.4%) received various other valve types (Lotus, Portico, JenaValve, and Symetis ACURATE). In the remaining 670 patients, CTA could not be performed, because of renal failure or severely impaired renal function, denial of repeat CTA, death before CTA, logistic reasons, or severely reduced general state of health (Figure 1).
Post-TAVR CTA revealed LT in 120 of 754 patients (15.9%) (Figure 2). Table 1 summarizes the baseline and procedural characteristics of patients with and without LT, as does Online Table 1 for patients with LT with and without anticoagulation. Among patients with LT, there were slightly fewer men (36.7% vs. 47.0%; p = 0.045) and a lower rate of atrial fibrillation (28.3% vs. 41.5%; p = 0.008). There were no significant differences in peri- and post-procedural characteristics (e.g., valve implantation technique, prosthesis size) between patients with and those without LT. Mean pressure gradient was similar between groups at the time of CTA (11.3 ± 4.9 mm Hg vs. 12.0 ± 5.1 mm Hg, p = 0.229), and there was no significant difference regarding anticoagulation at discharge (261 [41.2%] vs. 51 [42.5%]; p = 0.840). Between May 2012 and May 2015, 16 of these patients were temporarily treated with anticoagulation therapy for 3 to 6 months because of LT findings on post-interventional routine CTA.
The median follow-up period for the overall cohort was 406 days (interquartile range: 124 to 762 days). The average follow-up rate for the first year in our cohort was 97%. The mortality per year for the entire study cohort was 11.1% (n = 124), 11.1% (n = 15) in patients with LT and 11.2% (n = 109) in those without LT.
During follow-up, 3 patients (2.5%) with LT and 5 (0.8%) without LT developed symptomatic hemodynamic valve deterioration (log-rank p = 0.006). All patients with LT and obstruction were treated with anticoagulation, and in none of these were reintervention necessary. Two patients without LT revealed late obstructive thrombosis and were successfully treated with anticoagulation. In 2 of the other patients, valve-in-valve procedures were performed because of valve degeneration, and 1 patient developed hemodynamic valve deterioration due to endocarditis.
The Kaplan-Meier 18-month estimate for survival showed similar survival rates (86.4% in the LT group vs. 85.4% in the group without LT; p = 0.912) (Figure 3).
Stroke was reported by 13 patients (2.1%) without LT and by 1 patient (0.8%) with LT and TIA by 3 patients (0.5%) without LT and no patient with LT (p = 0.320 for stroke and p = 0.594 for TIA). The Kaplan-Meier estimates for 18-month stroke- and TIA-free survival were 98.5% for patients with LT and 96.8% for patients without LT (p = 0.331) (Figure 3). After the exclusion of 16 patients with temporary anticoagulation (change of therapeutic regime until May 2015), we did not observe any differences for primary and secondary endpoint.
Predictors of mortality and stroke or TIA
In the univariate Cox regression analysis, predictors of all-cause mortality were atrial fibrillation (hazard ratio [HR]: 1.84; 95% confidence interval [CI]: 1.29 to 2.62; p = 0.001), more than mild paravalvular leakage (HR: 3.71; 95% CI: 1.35 to 10.16; p = 0.011), and male sex (HR: 1.50; 95% CI: 1.05 to 2.13; p = 0.026) (Table 2). In contrast, LT was not significantly associated with an increased risk for all-cause mortality (HR: 0.88; 95% CI: 0.97 to 1.67; p = 0.913) (Table 2). In multivariate analysis atrial fibrillation, male sex, and more than mild paravalvular leakage remained predictors of all-cause mortality (Table 2).
In univariate analysis neither LT (HR: 0.38; 95% CI: 0.05 to 2.88; p = 0.350) nor the other tested variables (age, male sex, atrial fibrillation) were predictive of stroke or TIA. The incidence of stroke or TIA in our cohort was too low to perform a meaningful multivariate analysis.
To the best of our knowledge, the present study is the largest to date to systematically investigate medium-term outcomes in patients with LT after TAVR. The main finding of our study is that LT does not appear to be associated with increased mortality or risk for stroke or TIA over a median follow-up period of 406 days (Figure 3).
The incidence of LT has been reported at approximately 4% to 40%, depending on valve type and timing of computed tomography, with little variation among studies (6,9,10,14).
However, the clinical relevance of LT post-TAVR remains unclear. In line with a number of smaller studies, the data presented here add to the growing evidence indicating that LT has no impact on short- and medium-term mortality (6,8–10). The fact that LT represents a relatively benign condition is further supported by comparable rates of stroke in patients with and without LT, in line with previous studies, including a recent report by Chakravarty et al. (8,10,15). In the latter study, however, higher rates of TIA were reported in patients with LT (8). This discrepancy may be explained by high variability in the incidence of TIA depending on whether symptoms were self-reported or the result of a neurological evaluation of a whole cohort (16,17). A study with systematic analysis of patients with LT regarding neurological symptoms, including magnetic resonance imaging and complete neurological assessment, is currently missing but urgently needed considering the competing risks for stroke in a population with a high prevalence (approximately 40%) of atrial fibrillation.
Predictors of all-cause mortality in our study were male sex, atrial fibrillation, and more than mild paravalvular leakage. This is in line with prior studies (18–20).
In the present study, computed tomography was performed a median of 5 days post-TAVR, with an incidence of LT (16%) similar to previously published studies in which CTA was performed between 1 and 3 months after the intervention (6,8,10). Besides different time points, differences in computed tomographic angiographic techniques (e.g., prospective vs. retrospective electrocardiographic gating) might explain large differences of prevalence among published studies.
As demonstrated in prior studies, echocardiography is incapable of detecting LT in most cases, and CTA is the suitable diagnostic tool (4,21). Because of the assumed minor clinical relevance of LT, a general recommendation for routine post-TAVR CTA seems inappropriate. In this cohort of primarily elderly patients, additional contrast media exposure should be avoided, whereas the potential harm of additional radiation exposure seems negligible. In contrast, in patients with increased pressure gradient during routine echocardiographic follow-up or reporting symptoms suggestive of valve thrombosis, CTA is advisable.
The incidence of symptomatic obstructive thrombosis, presumably resulting from LT, varies between 0% and 18% (6,9,10,14). With early LT, the risk for developing hemodynamic valve deterioration in our study was significantly higher than without early LT. This may suggest that early LT heralds an increased risk for subsequent hemodynamic valve deterioration, thus challenging the view that early LT is entirely benign. Nevertheless, as numbers are still low, this finding must be interpreted cautiously, and further studies are clearly needed.
The role of oral anticoagulation in this setting needs further evaluation because on one hand, it has been demonstrated that anticoagulation resolves LT with potential relapses after discontinuation (6,9,10). On the other hand, approximately 30% of untreated LTs are no longer detectable on follow-up CTA, and anticoagulation is associated with a substantial bleeding risk in this population of mostly elderly patients. Thus, although anticoagulation prevents LT and resolves LT in most cases, a general recommendation seems inappropriate. Here further studies investigating the optimal treatment strategy in case of LT are necessary. At least 2 ongoing trials (ATLANTIS and POPular-TAVR) are targeting the issue of anticoagulation after TAVR and may provide further evidence concerning patients with LT.
Because of its observational character, our study has some limitations. First, selection bias cannot be excluded, as only 53% of TAVR patients underwent post-TAVR CTA. Second, the follow-up data were acquired mainly by written or telephone interviews, which may be insufficient in this cohort of elderly patients. Third, the diagnosis of the secondary endpoint of TIA or stroke was made by the treating neurologist based at the referring hospital, without independent adjudication. Fourth, the treatment regime for patients with LT changed during the course of the study. Until May 2015, all patients were treated with anticoagulation for at least 3 months. Subsequently, anticoagulation for LT was at the discretion of the treating cardiologist, resulting in inconsistent antithrombotic treatment in the LT group.
Because patients with LT do not seem to be at increased risk for stroke or TIA and death, and findings on follow-up CTA are highly variable with a significant proportion of spontaneous resolutions, it appears reasonable to withhold anticoagulation in patients with typical findings on post-TAVR CTA. We recommend follow-up CTA after 3 months and oral anticoagulation in case of a relevant increase in mean gradient and/or clinical signs of valve obstruction.
In our prospective, observational cohort study systematically investigating medium-term outcomes in patients with LT after TAVR, LT was not associated with increased mortality or rates of stroke or TIA over a follow-up of 406 days.
WHAT IS KNOWN? LT is present in up to 40% of patients undergoing TAVR. Currently there are limited follow-up data available regarding the clinical relevance of these findings.
WHAT IS NEW? This is the largest study to date to systematically investigate medium-term outcomes in patients with LT after TAVR. LT does not appear to be associated with increased mortality or risk for stroke or TIA over a median follow-up period of 406 days.
WHAT IS NEXT? Large randomized controlled trials are needed to confirm these findings.
Dr. Pache is a consultant for Edwards Lifesciences. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- computed tomographic angiography
- hazard ratio
- leaflet thrombosis
- transcatheter aortic valve replacement
- transcatheter heart valve
- transient ischemic attack
- Received December 22, 2017.
- Revision received February 26, 2018.
- Accepted April 3, 2018.
- 2018 American College of Cardiology Foundation
- Leetmaa T.,
- Hansson N.C.,
- Leipsic J.,
- et al.
- Chakravarty T.,
- Søndergaard L.,
- Friedman J.,
- et al.
- Ruile P.,
- Jander N.,
- Blanke P.,
- et al.
- Hansson N.C.,
- Grove E.L.,
- Andersen H.R.,
- et al.
- Webb J.G.,
- Chandavimol M.
- Capodanno D.,
- Petronio A.S.,
- Prendergast B.,
- et al.
- Yanagisawa R.,
- Hayashida K.,
- Yamada Y.,
- et al.
- Vollema E.M.,
- Kong W.K.,
- Katsanos S.,
- et al.
- Howard V.J.,
- McClure L.A.,
- Meschia J.F.,
- Pulley L.,
- Orr S.C.,
- Friday G.H.
- Biviano A.B.,
- Nazif T.,
- Dizon J.,
- et al.