Deposition of amyloid proteins in the extracellular space of the atria determines a progressive stiffening of the myocardial wall, resulting in structural fibrosis, loss of compliance and contractile function [1, 2, 3, 14].

The left atrium (LA) has three functional phases in healthy subjects, called ‘reservoir’, ‘conduit’ and ‘pump’ or ‘contraction’ phase, each determining the 50%, 30% and 20% of ventricular filling, respectively [1, 2, 3, 15].

Specifically, during the ‘reservoir’ phase the LA represents a ‘storage’ unit for energy when the LV is on isovolumetric contraction, ejection, and relaxation phase at mitral valve closed.

Conversely, the second or ‘conduit’ phase starts with the opening of atrioventricular (AV) valves at the beginning of the diastole and represents a specific ‘blood pathway’ from the pulmonary veins to the LV [14, 15].

Finally, the last phase, called ‘contraction’ or ‘pump’, permits, with the systolic ejection of the atrial myocardium, a further filling of the ventricular chamber at the end of the diastole.

All the three functional phases are impaired by amyloid deposits, especially the ‘reservoir’ one, thus resulting in decline of compliance and of systolic performance. It has also been hypothesized that ATTR-CA subjects may be more vulnerable to this process [14].

Furthermore, the atria involvement may represent a direct consequence of the high filling pressures secondary to LV diastolic dysfunction, or may occur simultaneously and separated from this, probably due to the direct deposition of amyloid in the extracellular space of atrial walls; this phenomenon, called ‘atrial myopathy’, leads to important consequences, such as the electrical isolation and the onset and manteinance of arrhythmias and AF, or to arterial thromboembolic events (AEs) such as stroke, TIA or peripheral vascular events [1, 14]. Moreover, some authors highlighted that the loss of atrial contraction can lead to rehospitalization, poor prognosis and higher mortality [14], therefore an early diagnosis by advanced ultrasound techniques of atrial dysfunction during the first phases of CA, in particular before the onset of ventricular infiltration, may be crucial to improve the prognosis of these patients [14, 15].

Atrial involvement by CA represents a vulnerable substrate for the formation and maintenance of arrhythmias; amyloid deposits and fibrosis cause an electrical isolation among the myocytes and the onset of supraventricular arrhythmias (Table 1) [1, 2, 3, 4].

An impaired ventricular diastolic function and the increase of filling pressure lead to atria enlargement with further myocardial damage [11].

An altered diastolic relaxation of the ventricular cavity can be associated with an increase of atrial stretch and with the overproduction of atrial natriuretic peptide (ANP). ANP oligomers are therefore deposited between the myocytes and cause electrical isolation, impulse fragmentation and arrhythmias in a vicious circle [1, 10].

AF is very common in patients with CA: in a retrospective study conducted by Sanchis et al. [20], on 238 subjects [123 (52%) with ATTR-CA, 115 (48%) with AL-CA], 104/238 patients (44%) had history of AF; 42/104 patients (40%) had non permanent AF, and 62/104 patients (60%) had permanent AF. Fourty-eight patients had an episode of AF during the follow-up. Specifically, the most common CA subtype linked to AF was ATTRwt-CA (71%), while only 26% with AL-CA and 19% with ATTRv-CA had an episode of AF [20].

Another retrospective study conducted among 133 patients with CA (53% with AL-CA, 41% with wtATTR-CA, 6% with ATTRh-CA), confirmed that the most common subtype linked to AF was ATTRwt-CA (80%) vs. AL-CA (28%) and ATTRh-CA (13%) with a p 0.001 and a 10-fold higher risk of AF. In opposite to CA patients with SR, those with AF were more frequently older (74 vs. 69 years, p 0.001), male (80%) and had usually a NYHA symptom class $\ge$III (66%, p = 0.02). Moreover, patients with AF had more often comorbidities and severe symptoms. Curiosly, a direct correlation was noted between the prevalence of AF and advanced stages of ATTR-CA (47% vs. 74% vs. 94%, p 0.001, for stages I, II, & III, respectively), while an inverse one with advances stages of AL-CA (0% vs. 40% vs. 31% vs. 18%, p 0.001, for stages I, II, IIIa, & IIIb, respectively) [21].

Donnellan et al. [22] analyzed retrospectively 382 ATTR-CA patients in the period between 2004 and 2008 and 265 of these had AF, especially in an advanced stage of disease (69%). Elderly subjects, patients with higher stages of ATTR-CA, and higher left atrial volume index were all predisposing factors for AF onset and development. Moreover, they observed that a rhythm control strategy was more effective in the early stages of the disease; pharmacological and electrical cardioversion were used in in 35% and 45% CA patients with AF, respectively, while 5% of these was subjected to ablation. Advanced stages of CA were associated with worse prognosis and increased mortality, while maintenance of SR and tafamidis use were linked to better survival [22].

Given the risk of intracardiac thrombosis even in patients already treated with anticoagulant therapy, some authors suggest to perform in these subjects a transesophageal echocardiography (TEE) before electrical cardioversion [23, 24, 25].

It has been proposed also that drugs having a negative inotropic or chronotropic effect should be avoided, or at least used with caution at the minimum therapeutic dosage as rate or rhythm control; conversely amiodarone remains to be the best drug for rhythm control [10, 26, 27].

IAA is an uncommon condition where amyloidosis affects directly the atria without any sign of ventricle involvement. Several authors suggest that this could be related to an abnormal ANP accumulation especially in the hearts of older patients. In a study conducted by Röcken et al. [28] on 245 subjects undergoing open heart surgery, the study of atrial appendages showed that IAA can predispose to AF through the infiltration of atria and conduction system by amyloid deposits: indeed, after Congo red staining and immunohistochemistry, 40/245 (16.3%) patients had amyloid proteins, which were all immunoreactive for ANP. Persistent AF was found in 38/245 (15.5%) subjects. Furthermore, patients with IAA and AF were at higher risk of prolonged P wave as compared with those with SR [28].

Moreover, Yang et al. [29] suggested that preamyloid oligomers formed by natriuretic peptides may have cytotoxic consequences and a proarrhythmic activity, and all of these effects are more evident in older adults, due to the physiological heart accumulation of natriuretic peptides.

A very common finding is the thickening of the myocardial wall ($>$ or = 12 mm), secondary to a diffuse infiltration of the myocardium by amyloid proteins and fibrotic tissue. Another characteristic that is not sensitive (35%) is the granular pattern of the myocardium (‘speckling or sparkling’). Despite a low diagnostic accuracy, the specificity of this finding is relatively high (80%) [30]. The sparkling aspect is secondary to the increased and heterogenous echogenicity related to the deposited amyloid fibrils [31].

Most common echo features are reported in Fig. 1 and in Tables 2,3.

Fig. 1.

Echocardiographic features of atrial amyloidosis. Parasternal long-axis (A,B) and five chamber apical view (C) of cardiac amyloidosis (CA) and atrial amyloidosis characterized by the presence of concentric right and left ventricular thickness (green arrows in A, B, C), dilated and thickened atria (yellow arrows in A, B, C), by the thickening of the interventricular septum (red arrow in C), by sparkling spots in ventricular, atrial and septal walls (yellow dots in B, C) and by deposits in aortic and mitral valves (red dots in A, B, C).

The interstitial accumulation of fibrotic tissue between the myocytes causes rigidity with reduction of ventricular compliance and various degrees of diastolic dysfunction [32, 33]. These abnormalities occur already in the early stages of the disease and tend to evolve gradually until the onset of a restrictive pattern [34] with an increased early diastolic peak (E) to atrial (A) ventricular filling velocities ratio $>$2 [35]. Only in the advanced stages of disease there can be a real loss of systolic function [36], a negative prognostic factor which is associated with a worse outcome [34].

Atrial dilatation is another unfavorable prognostic factor [37], and results from a reduced ventricular compliance, atrial wall and septum thickening from the myofibril deposits. Approximately, 60% of patients with amyloidosis have an increased atrial septum thickness [38]. A reduced atrial contractility is observed on echocardiography as a reduced or absent A wave [39].

Fibrillar deposition of the atrial walls leads to dilatation and thrombosis due to stasis, also found in patients with SR [40].

Heart valves are frequently thickened by amyloid deposits: 42% of patients with light chain amyloidosis had a thickening of $>$3 mm of the mitral valve at echocardiographic diagnosis. In a study, these patients were often elderly with an advanced NYHA class [37, 38, 39, 40, 41]. These alterations can be found also in the other forms of the disease, such as the transthyretin type [42] and have a negative prognostic significance [40, 43]. As discussed above, echocardiography is very useful to detect intracardiac thrombosis, which is located most often in the LAA [1, 13, 16, 17].

Some authors described alterations of atrial strain in all its functions (reservoir, conduit and pump) in 124 patients with ATTRwt-CA (27 patients), AL-CA (68 patients), ATTRm-CA (n = 29) and SR compared to twenty healty controls: all the functional phases of LA (longitudinal strain, early and late longitudinal strain rate or LSR, peak LSR) and LA active emptying fraction were altered, especially in the ATTRwt-CA subtype (p 0.05). These alterations correlated with left ventricular deformation and strain: peak LA LS and late LSR were related with LV global LS (R = –0.60, p 0.001) and with A wave at LV inflow (R = –0.69, p 0.001), respectively [44].

Evaluation of atrium strain can be useful to differentiate CA from other disease with thickened myocardium from unclear cause, such as hypertensive heart disease. In a study conducted by Brand et al. [45], 54 subjects with thickened septal wall (17.8 $\pm$ 3.5 mm) were included: of these, CA was confirmed through biopsy in 35 patients (20 AL-CA, 8 ATTRm-CA, 6 ATTRwt-CA, 1 AA-CA) while left ventricular hypertrophy (LVH) in the remaining 19 patients. Specifically the reservoir, conduit and contraction strain of LA were reduced in CA group and were more accurate than apical sparing [45, 46].

A recent retrospective observational study performed from January 2019 to December 2022 by Monte et al. [15] highlighted how the function of the LA studied throught STE is significantly altered in CA patients if compared to healthy control group and in those with hypertrophic cardiomiopathy (HCM): they recruited a total of 100 patients (34 HCM, 33 ATTR-CA and 33 controls); the CA subgroup had impaired reservoir, conduit and contraction strain of the LA (median values of –9%, 6.7% and –3%, respectively) if compared to HCM patients and control group. LA volume index, LV mass index, E/e’, LV-global longitudinal strain correlated with the strain of LA and were strictly connected with dyspnea and AF [15].

The recent position statement of the ESC Working Group on Myocardial and Pericardial Diseases confirmed the role of tissue doppler imaging (TDI) as non invasive echocardiographic criteria for diagnosis of CA, in particular the reduced tissue Doppler s’, e’, and a’ waves velocities (5 cm/s) [2]. Koyama et al. [47] recruited 97 subjects with biopsy confirmed AL-CA and they divided them into 3 groups: those without cardiac infiltration (n = 36), those with heart infiltration and congestive heart failure (CHF) (n = 29) and those without CHF (n = 32). Main TDI finding are listed in Tables 2,3.

Tissue velocity, strain and strain rate imaging were measured by TDI: this technique showed significant differences among basal strain in all groups, with an early impairment of contractility before the onset of CHF in patients with cardiac infiltration by amyloid proteins. Specifically, transmitral flow peak a’ velocity, which expresses the end-diastolic ventricular relaxation secondary to atrial contraction, was generally reduced due to loss of atrial elasticity [47, 48].

Similar results were obtained in the study by Palka et al. [49], where 36 patients with biopsy proven CA were divided into two subgroups, those with non restrictive (n = 22) and those with restrictive (n = 14) LV filling pattern, and compared to a control group. All patients were subjected to TDI examinations that showed the reduction of mitral annulus velocities, incluse a’ wave velocity, the reduction of the mean myocardial velocities and of the myocardial velocity gradient, if compared to control group [49].

As with the other cardiac structures, the atria chambers appear thickened, enhanced, and dilated on CMR [50, 51]. Most typical CMR features are reported in Fig. 2 and in Tables 2,3.

Fig. 2.

Cardiac magnetic resonance (CMR) of cardiac amyloidosis (CA) and atrial amyloidosis. The CMR images show a left atrium with thickened walls (yellow arrows in a,b) and a diffuse late gadolinium enhancement (LGE) of LA myocardium (yellow arrows in c,d). LA, left atrium.

In a study performed by Di Bella et al. [52] on 28 patients (53 $\pm$ 12 years) with familial amyloid polyneouropathy (FAP), they observed that approximately half (14 patients) of them had LGE on CMR analysis if compared to 22 healthy control subjects (49 $\pm$ 11 years).

In a similar way, on twenty-two patients with biopsy-proven CA who underwent CMR, 17 of them (78%) showed a diffuse LGE of LA myocardium; the same authors postulate that LA LGE and atrial dysfunction may be correlated with the risk of thrombus formation [53].

Moreover, in another study which recruited 32 patients with ATTR-CA and 15 healthy controls, authors observed that left atrial dimensions were larger in patients with CA, and 9/10 patients with ATTR-CA vs. 0/8 HCM patients had LGE after CMR imaging (p 0.001) [54].

LGE may be useful in the differential diagnosis with other cardiomyopathies, such as non-ischemic dilated cardiomyopathy (NIDC) or systemic hypertension (SH) [53]. Finally, CMR appears to have optimal resolution in detecting intracardiac thrombi [55].

CMR represents a valid and useful technique to evaluate the myocardial deformation; different methods have been developed to this function; one of these was the use of magnetic tags applied to the myocardium: tracking tag during cardiac cycle and visualization during CMR scans can provide useful information about the strain of the cardiac muscle, the strain rate and velocity of myocytes.

The same method can be used for atrial myocardium, diagnosing CA and monitoring treatment response.

Another technique is calculating the LA emptying functions, which is derived from the % reduction in LA volumes obtained by the validated biplane area-length method at the three diastolic phases (end-ventricular systole, pre-atrial contraction, and post-atrial contraction). These phases were characterized by aortic valve closure, second diastolic opening of mitral valve, and mitral valve closure [51, 52, 53].

CA patients recruited in the study by Kwong RY et al. [53] had a reduction of total LA emptying function (19 $\pm$ 14) if compared to SH and NIDC, respectively (40 $\pm$ 14 in SH subjects and 33 $\pm$ 20% in NIDC subjects, p = 0.0006) with a global reduction of active and total atrial emptying (r = –0.69, p = 0.001; r = –0.67, p = 0.01, respectively). These values were inversely proportional to LGE [53].

A study conducted on 44 patients with biopsy-proven CA, 19 with HCM and 24 healthy control subjects showed a reduction of reservoir left atrial strain (LAS), conduit LAS and booster LAS in CA and HCM subjects as compared to healthy control group (p 0.001). Specifically, reservoir LAS and booster LAS were lower in CA patients in opposite to HCM group (p 0.001) [56]. Similar results were obtained by Zhang et al. [57]: they included a total of 25 CA patients, 30 sex and age-matched hypertensive patients, 20 sane subjects, and studied LAS and LVS. Booster LAS, reservoir LAS, left atrial strain rate (LASR) were all impaired in CA if compared to healthy group (p 0.001) [57].

Another retrospective study by Palmer et al. [58] on 54 patients (mean age 67 $\pm$ 11 years, 68.5% male) with CA (30 with AL-CA and 24 with ATTR-CA) compared to 15 age-marched control confirmed a marked reduction of the strain values of LA at CMR analysis vs. control group: in particular it was lower in ATTR-CA, with left atrial reservoir value of 7.4 (6.3–12.8) in ATTR-CA group if compared to 13.8 (6.90–24.8) in AL-CA group with p = 0.017; moreover booster strain value was 3.6 (2.6–5.5) in ATTR-CA patients and 5.2 (3.6–12.1) in AL-CA patients, with p = 0.039.

Moreover, LAS and left atrial emptying fraction (LAEF) were strickly correlated with amyloid burden in 43 AL-CA patients: in opposite to control group, AL-CA subjects had impaired LAEF and LAS, and larger LA volumes; this was more evident in patients at higher risk of disease (n = 27, high levels of troponine I and N-terminal pro-B-type natriuretic peptide (NT-proBNP)) [59].

In addition, some authors highlighted an impairment of both left and right atrial reservoir strain of CA vs. HCM subjects and control group [for RA; HCM group: 33.5 $\pm$ 16.3% vs. CA group: 10.6% (5.6; 19.9)], p 0.001; [for LA; HCM group: 14.7 $\pm$ 7.1% vs. CA group: 7.0% (4.5; 11.1)], p 0.001 [60].

Finally, in a study performed on 51 patients with proven CA, Benjamin et al. [61] also showed, in a median follow up of 4.9 months, a lower LA strain and higher LA volumes at CMR scans in opposite to 51 age-, gender-, and race-matched SR subjects and without cardiovascular disease (CVD).