Amyloidosis is a progressive, systemic disease caused by misfolded proteins that progressively accumulate in the interstitial space as insoluble β-sheet amyloid fibrils.1 Cardiac amyloidosis (CA) is caused by the deposition of fibrils in the myocardium, the conduction system of the heart, and the valvular tissue.2 A significant proportion of patients with aortic stenosis (approximately 16%)— especially those undergoing transcatheter aortic valve implantation (TAVI)—have been found to have transthyretin associated (ATTR) cardiomyopathy3,4 and, with a lower prevalence, light chain CA.5

However, the effect of amyloid fibrils on structural valve deterioration and on the clinical outcomes of patients who undergo TAVI is still largely unclear. The few studies on this topic have yielded conflicting results.6–8 For example, Scully et al.6 reported no differences in the periprocedural complications of TAVI between patients with and without amyloidosis, whereas a recent meta-analysis showed that amyloidosis was associated with an increased risk of all-cause mortality and other common complications, such as stroke, heart failure (HF), and acute kidney injury (AKI).7 Given these discrepancies, further studies are needed to elucidate these questions.

We therefore aimed to investigate the 1-year risk of adverse events following TAVI in patients with amyloidosis using a large global federated health research network.

TriNetX is a global network of health care organizations compliant with the HIPAA (Health Insurance Portability and Accountability Act) and the United States federal law which protects the privacy and security of health care data. All data are deidentified in accordance with the deidentification standards outlined in the HIPAA Privacy Rule. A data-sharing agreement is needed to access the data in the TriNetX network. As a federated research network with non-identifiable patient data, studies conducted using the TriNetX health research network do not require approval from an ethics committee. TriNetX does not modify, impute, or estimate missing clinical values to address gaps in patient records. Additional details regarding data extraction from TriNetX are provided in the supplementary data.

Our study included 589 TAVI patients with amyloidosis (mean age±SD 78.9±8.2 years, 31.9% female) and 52 296 TAVI patients without amyloidosis (mean age±SD 78.1±8.8 years, 40.3% female). Before PSM, patients with amyloidosis were more likely to be male, Black, or African American, or Asian and had a significantly higher prevalence of cardiovascular risk factors, HF, AF, conduction system diseases, previous cerebral infarction, pulmonary embolism, and chronic kidney disease, compared with those without amyloidosis (table 1). These patients had a higher prevalence of bilateral carpal tunnel syndrome (3.7% vs 1%, P<.0001), lumbar spinal stenosis (16.5% vs 7.4%, P<.0001), spontaneous rupture of tendons (1.7% vs 0.3%, P<.0001), orthostatic hypotension (8% vs 3.5%, P<.0001), and macroglossia (1.7% vs 0.01%, P<.0001), which are considered the most common clinical red flags for CA9 (table 2).

After PSM, with 589 patients in each cohort, no significant differences were observed between the 2 groups in terms of demographics, prescribed medications, and diagnoses, except for the prevalence of conduction system diseases, which remained higher in patients with amyloidosis compared with those without (29.9% vs 22.1%, respectively; ASD: 0.179) (table 1). Figure 1 provides a visual assessment of how PSM reduced differences between the 2 groups, ensuring that the effects observed were less likely to be due to confounding factors.

Figure 1.

Love plot of the propensity score distributions before and after matching showing the degree of overlap between the groups. ACE inhibitors, angiotensin-converting enzyme inhibitors; AV, atrioventricular.

(0.21MB).

Table 3 shows the number of events following TAVI in patients with and without amyloidosis and the HRs, with relative 95%CI, for the primary and secondary outcomes at the 1-year follow-up. As shown by the Kaplan-Meier curves, patients diagnosed with amyloidosis had a significantly higher 1-year risk of the composite outcome compared with controls (HR, 1.27; 95%CI, 1.08-1.49) (figure 2). However, when analyzing the proportional hazards assumption, we observed substantial variability in the risk estimates compared with those expected (chi-square 5.75; P prop.=.017). This violation of the proportional hazards assumption suggests that the impact of amyloidosis on outcomes after TAVI may vary over time. When dividing the follow-up into 2 periods, we found that the risk of the composite outcome in patients with amyloidosis was greater during the late phase of follow-up (HR, 1.34; 95%CI, 1.11-1.61) compared with the early phase (HR, 1.17; 95%CI, 0.95-1.43), with no violation of the proportional hazards assumption in either period (table 3).

Figure 2.

Kaplan-Meier curves for the risk of the composite outcome following TAVI in patients with (blue line) and without (dark red line) amyloidosis. Dark lines, survival probability; faded borders, 95% confidence interval; TAVI, transcatheter aortic valve implantation.

(0.14MB).

When considering the secondary outcomes, patients with amyloidosis were more likely to experience acute HF episodes (HR, 1.37; 95%CI, 1.10-1.70, chi-square 0.124 P prop.=.724) throughout the follow-up period, with no violation of the proportional hazards assumption.

The risk of PM implantation showed an increasing trend during long-term follow-up. Specifically, patients with amyloidosis had a higher overall risk of PM implantation (HR, 1.41; 95%CI, 1.02-1.95), but with a proportional hazards assumption at the threshold of significance (chi-square=2.92, P prop.=.087). When considering only the late follow-up period (31 days to 1 year), the risk of PM implantation tended to rise, with no violation of proportionality (HR, 2.25; 95%CI, 1.15-4.41, chi-square=0.175, P prop.=.676).

Patients with amyloidosis also demonstrated an increased risk of stroke following TAVI. Specifically, the risk was higher for the entire observation period (HR, 1.47; 95%CI, 1.07-2.03, chi-square=4.624 P prop.=.032) with a slight violation of proportionality. However, when considering the long-term follow-up period, the risk of stroke remained consistently higher with proportionality respected (HR, 1.67; 95%CI, 1.16–2.40, chi-square=1.447 P prop.=.229).

No statistically significant differences were observed between the 2 groups regarding the risks of AKI (HR, 1.07; 95%CI, 0.83–1.38), with no violation of proportionality (chi square=0.008; P prop=0929). Similarly, no significant difference was found for the risk of all-cause mortality (HR, 0.77; 95%CI, 0.54–1.10), although this outcome showed high variability over the observation period (chi-square=3.21; P prop.=.073) (figure 3).

Figure 3.

Kaplan-Meier curves for the risk of all-cause death following TAVI in patients with (blue line) and without (dark red line) amyloidosis. Dark lines, survival probability; faded borders, 95% confidence interval; TAVI, transcatheter aortic valve implantation.

(0.11MB).

As shown by the Aalen-Johansen curves, the cumulative incidence at the end of follow-up for the different outcomes in patients with amyloidosis was 3.4% for all-cause death, 25.7% for HF, 10.6% for PM implantation, 8.6% for AKI, and 11.1% for ischemic stroke. In comparison, the cumulative incidence for patients without amyloidosis was 3.8% for all-cause death, 16.5% for HF, 8.3% for PM implantation, 6.7% for AKI, and 6.3% for ischemic stroke (figure 4).

Figure 4.

Aalen-Johansen estimator curves for events in patients with amyloidosis (A) and patients without amyloidosis (B) after TAVI. PM, pacemaker; TAVI, transcatheter aortic valve implantation.

(0.58MB).

In the present observational study, we provide clinical insights into a patient population undergoing TAVI with a diagnosis of amyloidosis. Our main findings are as follows: a) amyloidosis was associated with an increased risk of adverse events within the first year post-TAVI; b) the risk was not constant over time and appeared to increase progressively during the later phase of the follow-up period; c) the risk of HF remained steady throughout the follow-up period; d) over the long-term, amyloidosis was associated with a higher risk of PM implantation and stroke (figure 5).

Figure 5.

Central illustration. Amyloid deposition in the heart might increase the risk of adverse events in the first year following TAVI. Patients with amyloidosis are at higher risk of heart failure, stroke, and pacemaker implantation. 95%CI, 95% confidence interval; HR, hazard ratio; TAVI, transcatheter aortic valve implantation.

(0.44MB).

In our population, the prevalence of amyloidosis in patients undergoing TAVI was <1%. This value is significantly lower than those reported by other studies, where the prevalence of CA in TAVI candidates was approximately 8% to 16%.3,10 This discrepancy could be explained by various factors. First, screening methods for amyloidosis detection are not uniformly applied, not only among different health care organizations, but also over time. Indeed, in the past, there was a lack of awareness that a large proportion of patients undergoing TAVI have concomitant amyloidosis; echocardiographic red flags for amyloidosis have only recently been extensively recommended11 and specific ICD-10-CM codes for ATTR-CA and AL-CA have been introduced only since 2018, leading to underreporting of the disease. Second, imaging techniques applied for screening have different sensitivities and specificities. For example, it has been reported that cardiac magnetic resonance has lower sensitivity than bone scintigraphy.12 When patients are systematically screened for ATTR-CA using bone scintigraphy before the procedure, the prevalence of CA drastically increases.6,8,13

Despite the low prevalence of amyloidosis in the analysis, our results are in line with those of previous studies supporting the hypothesis that, although TAVI improves outcomes independently of the presence of amyloidosis,14 CA in TAVI candidates is associated with a higher rate of adverse events after valve replacement.

In our study, patients with amyloidosis were more susceptible to worsening HF, experiencing a higher rate of acute episodes during both short- and long-term follow-up after the intervention. Although the lack of information about left and right ventricular ejection fraction at baseline and during follow-up may limit the generalizability of our results, this finding appears to be supported by previous reports. In a prospective multicenter observational study of 27 patients with ATTR-CA undergoing TAVI, Rosenblum et al.8 reported that, although mortality was similar to that in controls, patients with amyloidosis had an increased risk of hospitalization for HF 1 year and 3 years after TAVI. These findings may be explained by amyloid deposits in the myocardial extracellular space, which progressively stiffen the ventricles, leading to both diastolic and systolic dysfunction. This impairs the ability of the heart to reverse remodeling after valve replacement and increases the risk of postoperative HF.15

In this context, as reported by histopathological studies,16 it should be considered that amyloid fibrils frequently deposit in the valvular tissue, causing a proinflammatory state that induces the differentiation of valvular interstitial cells into myofibroblasts and osteoclast-like cells, which perpetuate fibrosis and calcification of the extracellular matrix.17 The same pathological mechanisms could be associated with the degeneration of biological valve prostheses, but these hypotheses require further investigation.18

The AMYLOCARTESIAN trial (Prevalence and Post-surgical Outcomes of Cardiac Wild-type Transthyretin Amyloidosis in Elderly Patients With Aortic Stenosis Referred for Valvular Replacement; NCT02260466) is an ongoing multicenter prospective study that might help clarify the precise effect of amyloid fibrils on the myocardium and prosthetic valves.

We found that following TAVI, patients with amyloidosis have an increased risk of stroke—not periprocedural, but in the long-term follow-up. In a retrospective observational study on 273 TAVI patients with concomitant amyloidosis, Elzeneini et al.,19 found that CA was associated with a 3-fold increased risk of acute ischemic stroke during index hospitalization (odds ratio [OR], 3.08; 95%CI, 1.41-6.71). Conversely, Ismayl et al.20 did not find a significant association between amyloidosis and stroke following TAVI. A meta-analysis that included these 2 studies21 concluded that there was a significant increase in stroke risk in patients with a diagnosis of amyloidosis undergoing TAVI compared with those with lone aortic stenosis (OR, 1.97; 95%CI, 1.05-3.69).

Several factors might explain this increased risk. First, amyloid deposition in the valvular tissue, together with valve calcifications, could increase the amount of debris released from the native valve during prosthesis implantation.17 Second, due to amyloid deposition in the atrial myocardium and subsequent atrial cardiomyopathy, CA itself carries per se an increased risk of thrombus formation in the cardiac chambers.22,23 The use of cerebral embolic protection during the TAVI procedure is still a matter of debate regarding which patient subgroups might benefit from this strategy.24 Moreover, antithrombotic treatment following TAVI could be tailored to patients’ specific comorbidities, including amyloidosis.25 To date, the antithrombotic strategy used after TAVI in patients who do not have an indication for oral anticoagulant therapy involves single antiplatelet therapy with aspirin.26 However, it is still debated whether more effective antithrombotic therapy is necessary in specific categories of patients at greater ischemic risk. Although data from randomized trials support anticoagulant therapy only in patients diagnosed with AF,27–29 early diagnosis of CA associated with severe cardiomyopathy during the pre-TAVI screening could influence antithrombotic therapy decisions to reduce the incidence of thromboembolic events after the intervention.

Another relevant finding in our study was that amyloidosis increased the risk of permanent PM implantation mainly within 31 days to 1 year after TAVI. This could mainly be due to the higher prevalence of conduction system disorders at baseline, which we observed in patients with amyloidosis compared with those without, even after PSM. However, it should be noted that complete atrioventricular blocks and new-onset left or right bundle branch block are often the result of the valve replacement itself,30 potentially further exacerbating the need for PM implantation in a significant proportion of patients with amyloidosis. Our results appear to support this hypothesis, suggesting that amyloidosis may increase the long-term risk of conduction system diseases due to amyloid fibril deposition in the conduction system, as recently demonstrated by a systematic review and meta-analysis.31 The diagnosis of amyloidosis, often associated with aortic stenosis prior to intervention, could affect valve choice. Knowing that a patient has amyloidosis might lead to the choice of balloon-expandable valves in this category of patients, as some studies suggested that self-expandable valves could be associated with a higher risk of conduction disorders.32–34 Moreover, patients at high risk of atrioventricular block, such as those with CA, might require more careful and prolonged electrocardiogram monitoring following the procedure.

The risk of AKI, a significant complication of TAVI, was not significantly higher in patients with amyloidosis in our cohort during the first year after valve replacement compared with those without. These findings differ from earlier studies that reported a higher AKI rate in this population. Further research is needed to determine whether patients with amyloidosis are more susceptible to contrast-induced nephropathy.

In addition, mortality was not significantly affected by the presence of amyloidosis, as reported in a previous observational study.6 However, this could be related to the small number of patients considered and the relatively short follow-up period after the index event, which may have limited the statistical power of our study.

Lastly, in the sensitivity analysis, which evaluated the cohorts in 2 different periods (2005-2019 and 2020-2023), we demonstrated that the results from the main analysis were consistent over time and that, while TAVI techniques have undoubtedly improved, the underlying association between amyloidosis and the outcome of interest remained stable. The slightly lower HR in the 2020 to 2023 period may reflect better procedural outcomes due to technological advances, but it does not significantly alter the overall conclusion.

Patients with CA have a significantly higher risk of adverse events in the first year following TAVI, particularly HF, as well as an increased risk of PM implantation and stroke in the long-term. Further studies are needed to fully understand how amyloidosis affects TAVI outcomes and to investigate the specific role of amyloid fibrils in the deterioration of valve structure.

No funding was received for the preparation of this article.

This retrospective study is exempt from informed consent. The data reviewed are a secondary analysis of existing data, do not involve intervention or interaction with human participants, and are deidentified as per the deidentification standard defined in Section §164.514(a) of the HIPAA Privacy Rule. The process by which the data were deidentified is attested to through a formal determination by a qualified expert as defined in Section §164.514(b)(1) of the HIPAA Privacy Rule. This formal determination by a qualified expert was refreshed on December 2020. In the present study, we considered only the sex variable and, unfortunately, not the gender variable, as recommended by the SAGER guidelines.

Artificial intelligence has not been used in the preparation of this work.

L. Gerra, T. Bucci: concept, design, manuscript drafting, literature search, statistical analysis and interpretation. All authors: critical revision and final approval of the manuscript. L. Gerra and T. Bucci share first authorship on this manuscript. L. Gerra, T. Bucci, and G.Y.H. Lip are responsible for the data published in this study.

The authors have no conflicts of interest.