Validation of red flags in cardiac amyloidosis across diagnostic modalities

Abstract Background Cardiac amyloidosis (CA) is a relevant and treatable aetiology of heart failure. Despite substantial developments in the therapeutic options for CA, its time-sensitive identification often remains challenging. Red flags (RF) have been proposed to facilitate early detection of CA; however, large-scale studies validating these RF are lacking. Our study aims to validate a comprehensive group of clinically relevant RF in combination. Methods We conducted a retrospective analysis of all patients from our amyloidosis centre with known TTR gene status and a definite diagnosis or exclusion of CA in accordance with ESC guidelines. A holistic analysis of the available diagnostic data at the time of initial presentation was performed and included multidisciplinary clinical evaluation, ECG, laboratory tests, transthoracic echocardiography (TTE) and cardiac magnetic resonance imaging (CMR). Comparative analysis included CA and a control group (CG), comprising patients with CA exclusion exhibiting left ventricular hypertrophy (LVH). For each modality, sensitivities (SE), specificities (SP) and RF with the highest diagnostic effectiveness, defined as Youden’s index (YI), were determined. Results Of the 618 available patients, 554 (90%) were included. 347 (63%) patients had CA, with transthyretin CA being the most common form (94.8%). In 207 (37%) cases CA was excluded. Of these, 104 (50%) had LVH and comprised the CG. Patients in the CA group were 10 years older on average (79 vs. 69) and showed a higher proportion of men (88.2% vs 62.5%). Heart failure and conduction disorders were the most effective clinical RF (YI 52.6% and 43.7%, respectively), followed by polyneuropathy (YI 31.1%) and bilateral carpal tunnel syndrome (CTS, YI 22.2%). QT interval prolongation and atrial fibrillation achieved the highest YI among ECG RF (40.7% and 40.2%, respectively). Peripheral low voltage, although highly specific (94.6%), was found in only about one third (32%) of CA patients. NT-proBNP 1000ng/L (YI 58.8%) and high sensitivity Troponin T (hsTnT) levels ≥ 2 times upper limit of normal (ULN, YI 58.2%) were the most effective laboratory RF. The combination of hsTnT 14ng/L and NT-proBNP 180ng/L ruled out CA in 99.7% of cases. As a marker of the "apical sparing", Apical-to-Basal strain ratio ≥ 2.0 was the most effective echocardiographic parameter (YI 61.3%), followed by biatrial dilatation (YI 53.3%). Further effective TTE-RF included diastolic dysfunction II-III°, LV-strain reduction, and relative wall thickness (RWT) 0.6 at YI of 47-48%. Typical LGE pattern on CMR and ECV ≥ 40% achieved the best YIs among all evaluated RF (YI 66.9% and 63.1%, respectively). Two clinical scenarios identified CA with approximately 80% sensitivity and specificity (Fig. 1). Conclusion CA is frequently associated with distinctive cardiac and extracardiac features, which allow for a robust suspicion in routine clinical practice (Fig. 2), facilitating early diagnosis and treatment.Two clinical scenarios suspicious for CA Diagnostic characteristics of CA-RF