This proof-of-concept study showed that despite a significant decrease in left ventricular mass and hence increase in V/M at 6 months after SAVR or TAVR, FFRCT remained constant.
Because AS has such a profound influence on coronary physiology, the 2 most widely used invasive coronary indices FFR and instantaneous wave-free ratio (iFR) have to be interpreted with caution in this subset of patients. So far, 2 studies have explored the long-term variations of coronary physiology after AVR. Scarsini et al. showed that FFR can decrease after TAVR at a median follow-up of 14 months in lesions with abnormal baseline value, whereas values remained stable when baseline FFR > 0.80. (5) In contrast, Vendrik et al. showed a significant and consistent reduction of FFR values immediately post-TAVR, further lowering after 6 months follow-up; iFR however did not change significantly. (6) Since its value can lower after TAVR, FFR could possibly ‘underestimate’ the severity of a coronary artery stenosis in AS. This is caused by a blunted hyperemic response that is restored after relief of AS through recovery of microcirculatory function and coronary flow reserve resulting in more hyperemic systolic flow. (7) Therefore, a practical solution would be to measure FFR after AVR. The question here remains when to do this: directly after TAVR which is most straightforward since the patient is still on the table (for obvious reasons this is cumbersome after SAVR) or after a longer period of time i.e., after full restoration of coronary and microcirculatory physiology? iFR is less subject to change after AVR since this is a resting index measured in the diastolic wave-free period only. However, the ischemic threshold of iFR seems to be lower probably because of the higher coronary resting flow associated with severe AS altering the pressure-flow relationship during the diastolic wave-free period. (8, 9) Therefore it is important to investigate if and how physiological measurements change after relief of AS. This informs us about the proper applicability of the test in question.
FFRCT uses computational fluid dynamics to calculate 3-vessel virtual FFR based on CCTA, presuming an intact microcirculatory function with a preserved vasodilatory capacity mimicking hemodynamics as they would be after AS treatment. (10) In a recent study it was demonstrated that FFRCT is feasible and safe with a moderate to good diagnostic accuracy compared to invasive FFR in patients with severe AS. However, there was a tendency of higher mean FFR values compared to mean FFRCT values probably explained by the mechanisms described above. (3) Until now, it has not been investigated if FFRCT changes after AVR. Theoretically, this could be because the estimation of coronary blood flow used for FFRCT calculation is based on LVM that often regresses after treatment of AS. (10) Lower LVM causes lower absolute coronary flow (ml/min) with the same perfusion (ml/min/g) thus theoretically leading to a lower pressure loss and hence a higher FFRCT value. (11) This means FFRCT could ‘overestimate’ the severity of a coronary artery stenosis in AS compared to FFR. Regression of LVM after AVR is translated into a higher V/M ratio. V/M has already been studied in other patient populations with somewhat conflicting data concerning its effect on FFRCT values. A low V/M on CCTA in patients with stable coronary artery disease was an independent predictor of invasive FFR < 0.80 irrespective of the presence of obstructive coronary stenosis. (12) V/M was lower in hypertrophic cardiomyopathy compared to controls, especially when there was important septal hypertrophy. This resulted in lower cumulative 3-vessel FFRCT values but somewhat surprising, no difference in FFRCT was seen in the mid portion of the left anterior descending coronary artery. (13) Patients with primary microvascular angina had lower V/M (intrinsically smaller caliber coronary arteries versus impaired vasodilatory response to nitroglycerin) compared to controls. This did not lead to lower FFRCT values however. (14) Recently, it was shown that females compared to males have higher V/M for the same degree of coronary stenosis which was associated with a higher FFRCT value. (15) Concerning our study, V/M could be regarded as a quantitative metric of imbalance between coronary blood supply and myocardial demand in AS that improves to a certain extent after AVR. It could therefore serve as a future marker of ‘maladaptive’ (low V/M) versus ‘adaptive’ (high V/M) left ventricular hypertrophy in AS. A low V/M ratio could also provide one of several explanations why a high LV mass independently conveys a worse prognosis in AS. (16, 17)
Despite a significant increase in V/M at 6 months after AVR, FFRCTAUC remained constant however. FFRCTAUC quantifies total epicardial conductance and pressure loss along the vessel. (18) Previous work has shown an excellent correlation between this virtual FFRCT pullback curve and the invasive FFR pullback curve with nearly identical curves in non-obstructive segments. (19) We used this parameter rather than a single point FFRCT value for comparison because these were all patients without significant focal stenosis. Also, we did not use the distal or nadir FFRCT value (defined as the lowest FFRCT value in a given system) since these are sometimes prone to false positivity. (20) We did see however a gradual decline of FFRCTAUC from the proximal to the distal part of the vessel (without ever reaching absolute FFRCT value < 0.75) reflecting the diffuse coronary disease that was present in our patients which is in concordance with previous data using invasive FFR pullbacks. (21) See Fig. 3.
An older study showed a decrease in coronary size after AVR with an even more important regression of LVM resulting in normalization of V/M compared to controls without AS. We did not see any change in coronary size after AVR, keeping in mind that we used total coronary volume instead of proximal coronary artery diameter. (22) Furthermore, our patients did not have strictly normal coronary arteries but diffuse disease with often high calcium burden and extensive atherosclerosis possibly hampering coronary vasoreactivity.
The main reason we excluded patients with angiographically moderate or severe stenosis is the duration of follow-up during which revascularization would not have been possible due to the concept and the protocol of the study. After percutaneous coronary intervention or coronary artery bypass grafting the repeat CCTA would not be interpretable. Nevertheless, because of the propensity for myocardial ischemia in severe AS, it is worthwhile and interesting to examine coronary physiology even in the absence of angiographically significant coronary artery disease. Impairment of coronary flow reserve with microvascular dysfunction, low aortic valve area, high rate pressure product, low coronary perfusion pressure and short diastolic perfusion time are all hemodynamic factors causing myocardial ischemia even without coronary artery stenosis. Many of these are at least partially reversible after AS relief. (1, 23) See table 2.
A possible explanation for the stable FFRCT values is that our patient population did not have the severe left ventricular hypertrophy sometimes seen in severe AS since mean LVM at baseline was 151.7 ± 40.7 g. (24) This could mean that the absolute change in mass and thus V/M was not large enough (although a very substantial mass reduction of 16% was seen) to induce a significant change in FFRCTAUC. Also, since LVM regression after AVR is an ongoing process that can continue for years, we do not know what the evolution of coronary physiology will be after a longer period of time. (24, 25) Another possibility is the absence of focal stenosis with all baseline FFRCT values > 0.75 making our findings concordant with those of Scarsini et al. mentioned before. (5)
Importantly, FFRCT showed low time-to-time biological variability and proved to be a robust and reproducible test in the same patient on 2 totally different occasions with a 6 months’ time interval. Despite massive changes in the heart first induced by the presence of severe AS and afterwards by AVR, FFRCT was not vulnerable to these confounding factors. The most important confounders in comparing AS and post-AVR patients are most likely CCTA image quality and change in LVM. None of them caused distortion in the performance of FFRCT. This could be inherent to the technique which uses fixed inlet and outlet boundary conditions making FFRCT blinded to the presence of AS or an aortic bioprosthesis, with coronary anatomy being the main determinant of FFRCT. (26, 27)
Further research is needed in patients with AS and concomitant focal coronary artery stenosis, as well as validation of FFRCT against the other physiological indices (invasive versus non-invasive and hyperemic versus resting). A particularly appealing question is whether FFRCT before AVR could predict invasive FFR after intervention i.e., after full recovery of microcirculatory function and coronary flow reserve. As such, CCTA could become a one-stop test in patients with AS to evaluate severity of valve disease (aortic valve calcium score), severity of coronary artery disease (Agatston score, percentage stenosis, FFRCT), feasibility of TAVR (sizing, vascular access), left ventricular mass and V/M ratio. When FFRCT is clearly negative or positive far away from its ischemic cut-off point of 0.80, one could safely defer from revascularization or proceed to revascularization respectively without further testing. In case of borderline values, invasive physiological testing would be indicated by measuring iFR or FFR after AVR but further studies with clinical endpoints are needed to validate this hypothesis.