Leaflet Immobility and Thrombosis in Transcatheter Aortic Valve Replacement

Arnold C. T. Ng; David R. Holmes; Michael J. Mack; Victoria Delgado; Raj Makkar; Philipp Blanke; Jonathon A. Leipsic; Martin B. Leon; Jeroen J. Bax


Eur Heart J. 2020;41(33):3184-3197. 

In This Article


Table 1 and Table 2 summarize all recent publications on THV thrombosis by transthoracic echocardiography and 4D MDCT, respectively. The incidence of THV thrombosis ranged from as low as 0.6% and up to 40% depending on the imaging modality utilized and diagnostic criteria employed.[20–24,26–28,33,37–40,43–48]


By virtue of its ubiquitous availability and ease of use, echocardiography is the first imaging modality of choice. The main advantage is the quantification of valvular dysfunction based on the haemodynamic severity of obstruction. However, previously published mean gradient/peak velocity cut-off values to define bioprosthetic valve dysfunction due to THV thrombosis were highly variable (see Table 1). Despite the variable definitions, THV thrombosis by echocardiography is most frequently defined as a mean gradient of ≥20 mmHg, mean gradient increase by >50% compared with baseline, or effective orifice area <1.2 cm2. Some studies have also incorporated other morphological abnormalities detected on echocardiography or 4D MDCT such as immobile/restricted leaflet motion, thrombotic mass or response to anticoagulation therapy.[33,39,44,47] However, echocardiographic evaluation of increased leaflet thickness, RLM or identification of mobile mass suggestive of thrombus are frequently more challenging due to acoustic shadowing and ring-down artefacts arising from the valve struts. Overall, the published incidence of THV thrombosis detected by echocardiography ranged from 0.6% to 7.6% (Table 1).

Cardiac Multidetector Computed Tomography

Due to its superior spatial resolution, 4D MDCT imaging is often used to detect THV thrombosis. The first case report of THV thrombosis detected by MDCT was published in 2013.[46] Since then, the reported incidence of THV thrombosis by 4D MDCT ranged from 4% to 40% (Table 2).[27,45] The anatomical diagnostic hallmark of THV thrombosis is leaflet thickening that is meniscoid in configuration and extends from the base to the tip of the leaflet, usually recognized as HALT on 4D MDCT (Figure 5). Hypoattenuated leaflet thickening can functionally lead to RLM, usually without severely elevated transvalvular gradients on echocardiography. The assessment of RLM is based on maximal leaflet opening in the systolic phase, thereby further stratifying patients into normal, mild (<50% RLM), moderate (50–70% RLM), severe (>70% RLM), or immobile (100% RLM) leaflets.[12] Finally, HALT with ≥moderate RLM is defined as hypoattenuation affecting motion (HAM).

Figure 5.

Examples of hypoattenuated leaflet thickening in transcatheter aortic valve replacements (top row) and surgical aortic valve replacements (bottom rows). HALT, hypoattenuated leaflet thickening; SAVR, surgical aortic valve replacement; TAVR, transcatheter aortic valve replacement.

One important methodological issue with MDCT is the variability in temporal resolution due to different gantry rotation times of different CT scanners. Theoretically, a lower temporal resolution may reduce the diagnostic accuracy of RLM as maximally reduced leaflet excursion may not be imaged within the acquired dataset. Second, the need for contrast administration may also limit its use. Finally, due to the need for time-resolved imaging to visualize leaflet motion, 4D MDCT scanning protocol usually demands retrospective electrocardiogram gating without pulse modulation and occasionally at higher energy levels (140 kV compared to 'standard' 120 kV).[12] Therefore, radiation exposure should be minimized by limiting scan range to only the aortic valve.

Echocardiography vs. Cardiac Multidetector Computed Tomography

Given the differences in echocardiographic vs. 4D MDCT definitions for THV thrombosis, the timing of imaging post-procedure, and the different patient populations studied, there is a substantial variability in the overall reported incidence of THV thrombosis or valve leaflet dysfunction. Often, the anatomical diagnosis of THV thrombosis by 4D MDCT does not equate to haemodynamic obstruction on echocardiography. For example, in the largest 4D MDCT series to date that included 890 patients from the RESOLVE and SAVORY registries, patients with at least moderate RLM (n = 106) had significantly higher mean aortic gradient (13.8 ± 10.0 vs. 10.4 ± 6.3 mmHg, P = 0.0004).[21] However, only 16% of patients with ≥moderate RLM had moderately elevated mean aortic gradient >20 mmHg and only 4% of patients had severely elevated mean aortic gradient >40 mmHg. This means that 96% of patients with ≥moderate RLM did not have haemodynamically severe valvular obstruction, and 84% had normal transvalvular aortic gradient on echocardiography. This was consistent with several other studies showing that, although patients with THV thrombosis on 4D MDCT had statistically higher mean transvalvular gradients compared with controls, the mean gradients are often still within normal ranges.[21,23,26,38] In the latest PARTNER 3 CT substudy, there was a non-significant trend towards higher mean aortic gradients in all TAVR and SAVR patients with HALT vs. those without HALT (13.2 ± 0.81 vs. 11.7 ± 0.24 mmHg, P = 0.08).[41] Patients with RLM had significantly higher mean aortic gradient than those without RLM (13.3 ± 0.83 vs. 11.4 ± 0.23 mmHg, P = 0.04). However, the degree of increased mean aortic gradient was not associated with clinically significant consequences. Other published MDCT studies showed no statistical differences in transvalvular gradients on echocardiography between the two patient groups.[27,28,40,48]