Impact of leaflet thrombosis on hemodynamics and clinical outcomes after bioprosthetic aortic valve replacement: A meta-analysis.

Abstract Background Leaflet thrombosis (LT, also called cusp thrombosis) detected by multidetector computed tomography (MDCT) is common in bioprosthetic aortic valve replacement (bAVR). However, it remains contradictory whether MDCT‐defined LT following bAVR is associated with hemodynamic deterioration and stroke. Thus, we performed the first meta‐analysis to assess hemodynamic outcomes and updated the latest researches on the clinical outcomes of MDCT‐defined LT after bAVR. Hypothesis MDCT‐defined LT might be associated with worse hemodynamic and clinical outcomes after bAVR. Method MEDLINE, EMBASE, Cochrane Library, and http://clinicaltrial.gov were searched from inception to 15th April 2019. The fix‐effect model was utilized to calculate odds ratio (OR) and 95% confidence interval (CI). The primary outcomes were hemodynamic stability indexes, including mean pressure gradient (MPG), left ventricular ejection fraction (LVEF), paravalvular leak (PVL), and clinical heart failure. The secondary endpoints were major adverse cardiovascular and cerebrovascular events (MACCEs), which consisted of myocardial infarction, all‐cause death, stroke, and transient ischemic attack (TIA). Results Twelve studies with 4820 patients were included. The total prevalence of MDCT‐defined LT was 9.7%. MDCT‐defined LT was associated with a significantly increased risk of MPG (inverse variance 0.43, 95% CI: [0.30, 0.57]), MACCEs (OR 2.43, 95% CI: [1.45, 4.06]), stroke (OR 1.79, 95% CI: [1.03, 3.11]), and TIA (OR 4.09, 95% CI: [1.59, 10.54]). There were no differences for other outcomes. Conclusions MDCT‐defined LT after bAVR is associated with increased MPG and increased risk of adverse cerebrovascular events, including TIA and stroke. While LVEF, PVL, and clinical heart failure were similar between patient with and without LT.


| INTRODUCTION
Surgical aortic valve replacement (SAVR) or transcatheter aortic valve replacement (TAVR) have been recommended in both American and European guidelines for patients with symptomatic severe aortic stenosis. 1,2 Bioprosthetic aortic valve is an implanted device made of non-synthetic origin to replace severe dysfunctional aortic valve during SAVR or TAVR. Recently, reports have shown that TAVR is suitable not only for surgical-prohibitive or high-risk patients, [3][4][5][6] but also for intermediate and low-risk patients, [7][8][9][10] and therefore, the number of patients undergoing TAVR is expected to be up to 270 000 in Northern-America and Europe annually. 11 Bioprosthetic aortic valve replacement (bAVR) during TAVR is fast becoming a key treatment for relieving symptomatic severe aortic valve stenosis. Leaflet thrombosis (LT, also called cusp thrombosis) is a common complication in bAVR with a prevalence of about 7% to 15% detected on multi-detector computed tomography (MDCT). [12][13][14] MDCT has been proven to be a more sensitive method for detection of (LT), which is characterized by hypoattenuated leaflet thickening (HALT) and a reduction in leaflet motion (RELM). 15 Some studies report contradictory findings about the association of adverse clinical events with MDCT-defined LT in bAVR. [16][17][18][19][20] On the one hand, the impact of MDCT-defined LT on hemodynamic indexes remained unidentified. On the other hand, it remains controversial whether MDCT-defined LT following bAVR is associated with stroke. Two previous meta-analysises 21,22 with 5 and 6 included studies respectively conclude inconsistent conclusions on this topic. As new evidence emerges in recent years, a new metaanalysis is needed to reassess this problem. Therefore, we performed a meta-analysis and systematic review of the incidence, and the hemodynamic and clinical outcomes of MDCT-defined LT following TAVR or SAVR to explore whether MDCT-defined LT induced deteriorating hemodynamics and adverse cerebrovascular events.

| METHODS
This systematic review was based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. We The search strategy utilized terms synonymous with the TAVR, SAVR, bioprosthetic aortic valve, and thrombosis, including "transcatheter aortic valve replacement," "transcatheter aortic valve implantation," " bioprosthetic aortic valve," "computed tomography," and "thrombosis." A manual search for all references of included studies was performed simultaneously.
The inclusion criteria were as follows: (a) report of LT in TAVR or SAVR patients, (b) post-TAVR or post-SAVR MDCT or four-dimensional computed tomography (4D CT) imaging performed during follow-up, (c) clinical outcomes reported between patients with and without MDCTdefined LT, and (d) retrospective or prospective cohort studies. Studies were excluded if: (a) there was no control group, (b) diagnosis of LT was based on other imaging modalities, (c) only the abstract was published, and (d) there was less than a 6-month follow-up. We chose the newest papers with the largest population if there were several studies reporting on the same cohort of patients. In this study, LT was specifically defined as evidence of HALT or RELM of more than 50% in at least one leaflet on MDCT.
Two investigators performed the meta-analysis independently, including study selection, risk-of-bias assessment, data extraction, and analysis. If there were any discrepancies, they were resolved by consensus. The risk of bias was assessed according to the Newcastle Ottawa Scale. 23 The primary outcome of this study was hemodynamic stability, as measured by mean pressure gradient (MPG), left ventricular ejection fraction (LVEF), clinical heart failure, and more than moderate using a fixed-effect inverse variance (IV) model to determine standard mean differences with 95% confidence intervals (CI). Statistical heterogeneity was rated as high, moderate, or low based on I 2 values of 75%, 50%, and 25%, respectively. Egger's linear regression tests were performed to assess publication bias. A P-value of <.05 was considered significant.

| RESULTS
The search strategies yielded a total of 25 citations and 12 studies met the criteria (Figure 1). These 12 studies included reported results from the PORTICO IDE trial 23 16 The other 10 studies were single-center observational cohort studies. In total, 4820 patients were included in this analysis, and 4636 patients received TAVR and 184 patients received SAVR.
The demographic details are presented in Table 1. Overall, 452 patients (9.7%) were identified with evidence of LT on MDCT. Table S1 shows the quality assessment characteristics of the included studies.
Eight studies compared MPG in patients with and without LT during follow-up, and the MPG was significantly higher in patients with LT There was no publication bias regarding clinical outcomes (P = .284, Figure S8) or hemodynamic outcomes (P = .618, Figure S9). ous meta-analysis demonstrated that MDCT-defined LT was associated with increased risk of cerebrovascular events and TIA, but not stroke, 21 while another meta-analysis showed that LT detected by echocardiography or MDCT was associated with increased risk of stroke. 22 While we find similar results for TIA, we find that there is an Moreover, the rates of AF vary between included studies and evidences showed that one-third of ischemic strokes attribute to AF 36 .

| DISCUSSION
But it seems influent little on the result because previous studies find an impact of new-onset but not pre-existing AF on the early stroke in TAVR. 37,38 And the incidence of stroke are still higher in MDCT- Given that it is possible that MDCT-defined LT is related to hemodynamic deterioration and MACCEs, this is an important issue for physicians. Several reports have shown that HALT and RELM regress after anticoagulation. 39,40 The initiation of anticoagulation depends on the discretion of physicians. Anticoagulation seems to reduce the occurrence of subclinical LT more effectively than antiplatelets. Despite the RESOLVE and SAVORY registries mentioned in this article, the FRANCE-TAVI registry experiment, which investigated subclinical thrombotic events in 12 804 patients, showed that LT rates were significantly lower in patients who received an anticoagulation regimen. 42 Jimenez et al. 43 observed that lack of oral anticoagulant therapy at discharge was an independent predictor of MDCT-defined LT in patients following TAVR. The mechanism behind the lower occurrence of LT in patients administered anticoagulation agents might be that adenosine diphosphate-induced platelet reactivity was not significantly associated with the occurrence of HALT. 44  In conclusion, MDCT-defined LT following TAVR/SAVR is associated with changed hemodynamic stability and a significantly increased risk of MACCEs, especially adverse cerebrovascular events, including TIA and stroke. Therefore, we recommend that MDCT is performed where possible to detect high-risk patients for LT, such as patients with the morbidity of AF and CHA2DS2-VASc≥2, history of thrombolism, low LVEF, the elderly, etc. Further studies are required to explore whether MDCT-detected LT influences other hemodynamic indexes, even valve durability in the long term, and whether the adverse cerebrovascular events can be adequately prevented with the use of anticoagulants.