Aortic stenosis, transcatheter aortic valve replacement and transthyretin cardiac amyloidosis: are we progressively unraveling the tangle?

This article refers to ‘Unveiling outcomes in coexisting severe aortic stenosis and transthyretin cardiac amyloidosis’ by H. Rosenblum et al., published in this issue on pages 250–258. The topic of coexisting aortic stenosis (AS) and cardiac amyloidosis (CA) has become a classic of contemporary cardiology due to the spread of transcatheter aortic valve replacement (TAVR) as a treatment for AS and the wide availability of bone scintigraphy as a non-invasive diagnostic standard for transthyretin-related amyloidosis (ATTR). In this issue of the Journal, Rosenblum et al.1 report the results of a prospective study, conducted at two academic US centres, aimed to evaluate the implications of coexisting transthyretin cardiac amyloidosis (ATTR-CA) on outcomes after TAVR, in moderate-risk patients with severe AS. The Authors must be congratulated since they were able to collect 204 patients and to follow them for 2 years (median value) showing that ATTR-CA does not apparently affect mortality but increases the risk for heart failure hospitalization. Although the paper helps to clarify some aspects of the topic, it nonetheless leaves a series of questions open. In the study of Rosenblum et al.1, the prevalence of dual pathology (AS and ATTR-CA) is 13%. It should be noted that the diagnosis was confirmed in all cases by technetium-99m pyrophosphate (99mTc-PYP) scintigraphy (diffuse, with visual score ≥2 and a heart-to-contralateral score ≥1.5) even if the cases studied in one of the two centres from which the study came were partly filtered by an initial magnetic resonance imaging screening which had to be judged suggestive of CA. In early studies on this topic, dual pathology was not systematically searched and only patients with ‘diagnostic red flags’ fostering the suspicion of CA were studied with bone scintigraphy leading to higher prevalence values.2 Considering only the prospective studies with a systematic search for CA associated with AS, the prevalence ranges from 4% to 16% (Table 1).1-6 41 (25–84) 21 (14–34) 28 (20–49) 3702 (1286–5626) 1254 (598–2769) In general, the prevalence of CA among patients with AS is mainly influenced by patients' age, the criteria with which they were initiated in the diagnostic process for the search for CA, and the diagnostic methods used. For example, magnetic resonance imaging has a lower sensitivity than scintigraphy in detecting ATTR [but bone scintigraphy has a limited ability to recognize light-chain (AL) CA]. The lowest prevalence (4%) is reported in cardiac surgery series of patients undergoing aortic valve replacement (AVR), in which the diagnosis was systematically made by biopsy during the intervention.6 In the vast majority of the reported cases, the aetiology of amyloidosis is wild-type transthyretin (wtTTR), with a minimal (<5%) epidemiological contribution from AL-CA. It is also interesting to observe the association of CA with AS from the ‘point of view’ of amyloidosis (not of AS). Among a cohort of 1240 consecutive patients with ATTR-CA from the National Amyloidosis Center (the largest available study on this topic), only 1.8% of patients had severe AS on echocardiography.7 In a smaller cohort (171 patients with CA), Sperry et al.8 described a higher prevalence (15.7%). These patients with dual pathology were more likely to be older and anaemic and had a larger left ventricular end-diastolic diameter. Notably, in this study, mortality was the same irrespective of AVR, suggesting that ATTR-CA may have driven mortality more than the haemodynamic effects of AS. The question of whether amyloidosis is a cause or a consequence of AS or whether they are ‘epidemiologically related’ is still open. The hypotheses are broadly three: (i) the association could be likely attributable to the age-dependent penetrance of both conditions that belong to the same epidemiologic stratum of the population; (ii) amyloidosis could be the common cause of cardiomyopathy and AS; (iii) cardiomyopathy could be the consequence of AS. The first hypothesis is reasonable and partly supported by epidemiological data that record a progressive increase in the prevalence of both AS and ATTR with age, with peaks over 75–80 years. Indeed, severe AS affects >3% of people over 75 years, with similar echocardiographic features to that of CA.9 The prevalence data for ATTR in the general population are less certain. Approximately 25% of people over 85 years of age have ATTR deposition in their heart at autopsy,10 but not necessarily these deposits constitute a true and widespread cardiomyopathy. Some studies assessed the prevalence of unexpected myocardial positivity among subjects undergoing bone scintigraphy for non-cardiac reasons and reported values ranging from 2% to 14% in males ≥85 years11, 12 (estimated prevalence for the European standard population ≥75 years 4.15% in males and 1.03% in females12). These values are lower than the prevalence of 8–16% reported in most of the series with dual pathology. Therefore, the purely epidemiological explanation cannot be entirely convincing. That systemic amyloidosis plays a pathogenic role in aortic valve stenosis currently remains a hypothesis in search of evidence. Kristen et al.13 analysed 100 surgically removed stenotic aortic valve specimens and found amyloid deposits in 74 cases. However, by immunohistochemistry, none of the most common amyloid proteins was identified and amyloid deposition appears secondary to atheroinflammatory conditions and high shear-stress haemodynamics. On the other hand, degenerative AS could plausibly be a contributing cause or acceleration factor of CA. The increased myocardial strain of AS could act as a mechanical trigger on an underlying amyloidogenic substrate related to circulating and tissue wtTTR (mechano-enzymatic hypothesis).14 Given the similarity in echocardiographic features between severe AS and CA, it is challenging to identify those patients with dual pathology. Although in the study of Rosenblum et al.1 the subgroup with CA showed older age, lower modified body mass index, higher troponin I, lower left ventricular ejection fraction, lower stroke volume index, and larger left atrial dimension, no cut-off value or score is provided to identify individual patients with dual pathology. In a previous study,3 an average mitral annular S′ ≤6 cm/s was the best independent echocardiographic predictor of ATTR-CA, being 100% sensitive for ATTR-CA. If we consider the published studies all together (Table 1), a characteristic average profile of patients with CA associated with severe AS becomes evident. This ‘phenotype’ includes: low-flow low-gradient pattern, reduced left ventricular ejection fraction and myocardial contraction fraction, reduced systolic longitudinal function (low S′), and restrictive pattern of transmitral flow. Other non-echocardiographic findings help to delineate this profile, including low QRS voltage/left ventricular mass on the electrocardiogram, and clinical history of bilateral carpal tunnel syndrome.15 Note that this profile does not come from sophisticated methods but from absolutely standard exams, interpreted however with an ‘amyloidotic-oriented mindset’. The appreciation of this phenotype (in whole or in part) can guide the clinician to perform a bone scintigraphy and to search for monoclonal protein in order to reach the definite diagnosis. It has recently been reported that computed tomography scan with contrast can also provide a quantitative estimate of myocardial extracellular volume.16 Since computed tomography is an integral part of the pre-TAVR assessment of the patient, this exam can also provide a red flag and should be considered. Notably, two findings considered frequent and specific in isolated CA are much less frequent and lose specificity when AS coexists: the high prevalence of male gender and the apical sparing of the longitudinal strain. There is no clear explanation for the greater frequency of women in dual pathology compared to ATTR-CA alone. However, this phenomenon was also observed in patients with heart failure with preserved ejection fraction and concomitant CA. As for the lack of apical sparing, apex longitudinal strain is impaired probably due to raised wall stress and increased afterload induced by AS that ‘compensate’ reduced apical deposition of amyloid in comparison with other segments. The doubt that the coexistence of ATTR-CA could negatively influence the results of TAVR or AVR arises from early anecdotal reports and small descriptive studies that cannot represent the ‘evidence’. The correct way to tackle with this problem would be a prospective study enrolling patients with coexisting ATTR-CA and AS randomized to TAVR vs. optimal medical therapy alone, perhaps stratifying patients by CA severity. Such a trial is neither available nor ongoing (nor easily feasible). A lower than this but still reasonable level of evidence arises from non-randomized prospective studies (like the one by Rosenblum et al.1) that compare, within the same centre, the results of TAVR in patients with AS with or without CA (Table 1). In none of the prospective studies with follow-up data did ATTR-CA appear to influence mortality despite the fact that patients' age is more frequently higher, the left ventricular systolic and diastolic function more impaired, the hypertrophy more severe and the low-flow low-gradient paradoxical phenotype more frequent. In each study the confidence interval for the hazard ratio for all-cause mortality is wide, suggesting the possibility of an underpowered study due to the small sample size and low event rates. Their overall analysis, however, supports the hypothesis of an absence of significant differences. Another line of reasoning that leads to the same conclusions is the comparison between the mortality rates in the various studies and that reported in the Placement of Aortic Transcatheter Valve (PARTNER) trials.17-19 As properly underlined by the Authors, the cohort described by Rosenblum et al.1 can be considered to be at ‘moderate risk’ for TAVR, with baseline characteristics and mean Society of Thoracic Surgeons risk scores falling between the PARTNER 1A (high risk) and PARTNER 2A (intermediate risk) cohorts. The PARTNER 1A study reports mortality rates of 24%, 34% and 68% at 1, 2 and 5 years, respectively. The PARTNER 2A study reports an overall mortality rate of 12% and 16% at 1 and 2 years, respectively. Accordingly, overall survival rates in the study for patients both with and without ATTR-CA falls between the PARTNER 1A and PARTNER 2A cohorts. The same observation holds true for the other prospective studies included in Table 1. Unlike what happens for TAVR, patients with concomitant ATTR-CA and severe AS undergoing aortic valve intervention exhibit a higher mortality (Table 1), probably due to cardiopulmonary bypass and open-heart surgery. Apart from mortality, two findings appear clinically important: a higher incidence of heart failure during mid-term follow-up (specifically reported by Rosenblum et al.1) and a possible higher frequency of periprocedural atrio-ventricular block requiring pacemaker implantation.20 Recognition of ATTR before any type of intervention is crucial for appropriate risk stratification and guidance of downstream management. Although the available information is relatively limited (no more than 100 patients studied in non-randomized prospective studies) all the foregoing considerations make it reasonable to state that patients with severe AS and associated ATTR should not be denied TAVR when indicated. All other conditions being equal, the coexistence of CA can guide the choice between AVR and TAVR towards the transcatheter procedure. The survival benefit conferred by TAVR seems to persist even in patients with more advanced restrictive cardiomyopathy. However, these patients should be followed up after TAVR with particular care, given the greater probability of developing heart failure secondary to the persistence of infiltrative restrictive cardiomyopathy. Once TAVR has been performed, these patients could probably benefit from a disease-modifying therapy for amyloidosis even if a note of caution is necessary. In fact, these are generally patients with advanced forms of CA. Furthermore, this type of patient was not represented in the single phase 3 trial so far conducted with disease-modifying drugs (tafamidis) in patients with CA21 nor in other non-randomized studies.22, 23 Consequently, a randomized trial on this topic would be desirable. Conflict of interest: none declared.