Clinical Factors Related to Cardiac Conduction Disturbance in Patients With Wild-Type Transthyretin Amyloid Cardiomyopathy Undergoing Tafamidis Treatment

Abstract

Background

Cardiac conduction disturbances are frequently observed in patients with transthyretin amyloid cardiomyopathy (ATTR-CM). Because their relationships with other clinical factors have not been fully investigated, in this study we evaluated changes in atrioventricular conduction disturbances (AVCD) and intraventricular conduction disturbances (IVCD) in patients with ATTR-CM undergoing tafamidis treatment, with the aim of identifying clinical factors, including cardiac magnetic resonance (CMR) parameters, related to AVCD and IVCD.

Methods and Results

We studied 50 patients with wild-type ATTR-CM who underwent baseline CMR and subsequently commenced treatment with tafamidis. Data of cardiac biomarkers and left ventricular (LV) parameters on CMR were acquired at baseline. Changes in the PQ interval (∆PQ) and QRS duration (∆QRS) on ECG were assessed over 1 year. Despite tafamidis treatment, both the PQ interval and QRS duration significantly increased over 1 year. After adjusting for medications that might affect the cardiac conduction system, the native myocardial T1 value (T1_native) was shown to determine the baseline PQ interval (β=0.31, P=0.039), baseline QRS duration (β=0.30, P=0.038), and ∆QRS (β=0.29, P=0.035), whereas Ln (serum high-sensitivity cardiac troponin T [hs-cTnT]) was a determinant of ∆QRS (β=0.32, P=0.036).

Conclusions

In patients with wild-type ATTR-CM undergoing tafamidis treatment, AVCD and IVCD progress over 1 year. Baseline T1_native and hs-cTnT are predictors of progression of IVCD under tafamidis treatment, and T1_native is associated with AVCD and IVCD in disease-modifying therapy-naïve conditions.

Central Figure.

Transthyretin amyloid cardiomyopathy (ATTR-CM) is a cardiac amyloidosis (CA) characterized by the deposition of transthyretin amyloid in the myocardium. Transthyretin has a concentration-dependent neurotoxic effect that may cause disorder of intracellular calcium signaling and induce cellular apoptosis, which is considered to be the mechanism of cardiac conduction abnormalities, including atrioventricular conduction disturbances (AVCD) and intraventricular conduction disturbances (IVCD).^1,2 Atrioventricular block has been reported in 33–49% of patients with ATTR-CM,^3,4 and pacemaker implantation is required more frequently in patients with CA and heart failure (HF) with preserved ejection fraction (EF) than in those with HF with preserved EF without CA.⁵ The progression of IVCD is clinically important because it often correlates with dyssynchronous ventricular contraction, which may cause chronic HF exacerbations and require cardiac resynchronization therapy (CRT). Electrical desynchrony is observed in approximately 20% of patients with CA.⁶ Prognostic classifications for ATTR-CM based on serum high-sensitivity cardiac troponin T (hs-cTnT) concentrations, serum N-terminal pro-B-type natriuretic peptide (NT-proBNP) concentrations, and the estimated glomerular filtration rate (eGFR) are proposed as the Mayo staging system or National Amyloidosis Center staging system.^7,8 However, the association between cardiac conduction disturbances and other clinical factors has not been fully investigated in this disease population.

Editorial p 1540

Tafamidis is the first disease-modifying therapy (DMT) to become available in the clinical setting. Tafamidis reduces all-cause death, the incidence of cardiovascular-related hospitalizations, and the decline in functional capacity and quality of life.⁹ Observational studies from Japan and other countries show that tafamidis is associated with favorable cardiac function, improved survival, reduced HF hospitalizations, stabilization of cardiac structure and biomarkers, and preservation of physical function, especially when therapy is started early.^10,11 Nonetheless, there is a paucity of data regarding alterations in cardiac conduction system in patients with ATTR-CM undergoing tafamidis treatment. Nakamura et al. reported no significant changes in the QRS duration from baseline to 6 months in 6 patients who underwent tafamidis treatment.¹² By contrast, Shahi et al. observed a significant increase in the QRS duration from baseline to 1 year of follow-up for patients with ATTR-CM treated with tafamidis.¹³ Therefore, the effects of tafamidis on cardiac conduction disturbances remain debatable.

Cardiac magnetic resonance (CMR) enables the assessment of cardiac morphology and function through cine imaging and distinguishes various cardiomyopathies using techniques such as late gadolinium enhancement, T1 mapping, and T2 mapping. In particular, the native myocardial T1 value (T1_native) offers an intrinsic signal from the myocardium without the need for contrast agents and serves as a useful surrogate marker for myocardial amyloid deposition, with good diagnostic accuracy.¹⁴ However, there is a lack of studies examining the CMR parameters associated with cardiac conduction disturbances in ATTR-CM.

In this study, we investigated cardiac conduction disturbances in patients with ATTR-CM undergoing tafamidis treatment by evaluating changes in the PQ interval and QRS duration over a 1-year follow-up period. Additionally, we aimed to identify clinical factors, including CMR parameters, related to AVCD or IVCD.

Methods

The study complied with the Declaration of Helsinki. The Ethical Committee for Epidemiology of Hiroshima University approved the study protocol (approval number: E2018-1364). Informed consent was obtained using an opt-out method.

Study Participants

We retrospectively examined 60 patients diagnosed with ATTR-CM who had not previously received DMT and who underwent CMR between April 2019 and January 2025. The diagnosis of ATTR-CM was based on histochemical detection of amyloid and transthyretin deposition in the myocardium or extracardiac tissue with ^{99 m}Tc-labeled pyrophosphate uptake in the myocardium. According to Japanese Circulation Society guidelines for the diagnosis of CA,¹⁵ the confirmation of histologically amyloid deposition in cardiac or extracardiac tissue is necessary to definitively diagnose ATTR-CM. Diagnosis of wild-type ATTR-CM was based on the absence of mutations in the transthyretin gene, which was assessed using genetic testing. Patients were excluded if they lacked ECG data, had received a permanent pacemaker or CRT, or did not survive during the 1-year follow-up. In accordance with the statement of DMT for ATTR-CM in Japan,¹⁶ tafamidis 61 mg or tafamidis meglumine 80 mg was initiated in all patients immediately after baseline evaluation. In a previous clinical study, tafamidis 61 mg was shown to work in the same manner and as effectively as tafamidis meglumine 80 mg.¹⁷ Patients’ characteristics, including ECG and blood test data at baseline, were obtained from medical records. Blood test data included the eGFR and serum concentrations of hs-cTnT and NT-proBNP. All patients were clinically stable without evidence of congestive HF at the time of imaging, ECG, and blood testing (New York Heart Association class I or II).

ECG Measurements

Standard 12-lead ECG was performed at the time of baseline CMR in all patients, and follow-up ECG was performed 1 year later. Changes in the PQ interval (∆PQ) and QRS duration (∆QRS) were calculated by subtracting baseline values from those at the 1-year follow-up. The PQ interval and QRS duration were measured manually using digital calipers by 2 experienced cardiologists. The measurement was verified by consensus. Furthermore, we assessed the prevalence of complete right bundle branch block (CRBBB) or complete left bundle branch block (CLBBB) at baseline and at 1-year follow-up. Definitions of CRBBB and CLBBB were based on previously reported criteria:¹⁸ CRBBB=QRS duration ≥120 ms, plus an rsr′, rsR′, or rSR′ pattern in leads V1 or V2, and a broad terminal S wave in leads I and V6; CLBBB=QRS duration ≥120 ms, plus broad, notched, or slurred R waves in leads I, aVL, V5, and V6, absence of q waves in leads I, V5, and V6, and an R-wave peak time >60 ms in leads V5 and V6. Patients were categorized into 3 groups: (1) those with pre-existing CRBBB or CLBBB at baseline; (2) those with new-onset CRBBB or CLBBB during the 1-year follow-up; and (3) those without CRBBB or CLBBB throughout the 1-year follow-up.

CMR Protocol

The CMR protocol has been described previously.¹⁹ It was performed with the patient in the supine position, using an Ingenia 3T CX scanner (Philips Medical Systems, Best, The Netherlands) equipped with dS Torso coils. The CMR protocol included cine CMR and T1 mapping sequences. Cine images were obtained in the short-axis and 3 long-axis views (2-, 3-, and 4-chamber views) using a cine steady-state free precession sequence to assess left ventricular (LV) morphology, mass, and function. T1 mapping was performed using a 13-heartbeat steady-state free procession, single-breath-hold modified Look-Locker inversion recovery sequence in 5 short-axis slices (base to apex). T1 mapping sequence parameters were as follows: slice thickness, 10 mm; TR/TE, 2.2/1.0 ms; flip angle, 20°; field of view, 300×300 mm; sampled matrix size, 160×155 mm; SENSE factor, 2; and 5 images from 3 inversions (3+3+5) with 1 heartbeat pause before the second and third inversions and an adiabatic prepulse. In the cohort participating in this study, a limited number of patients were able to undergo contrast-enhanced CMR because of the presence of chronic kidney disease or a history of allergy to contrast agents. Therefore, we focused on clinical implications of precontrast CMR parameters in this study.

CMR Image Analysis

CMR images were independently analyzed by experienced cardiologists using commercially available software (ShadeQuest/ViewR and ViewC; Yokogawa Medical Solutions Corp., Tokyo, Japan). LV parameters on cine images and T1 mapping were evaluated as described previously.¹⁹ In the analysis of cine images, papillary muscles were included within the LV cavity volume, and tracings of the endocardial borders were performed at end-diastole and end-systole. LV end-diastolic and end-systolic volumes were quantified with Simpson’s rule, and the LVEF was calculated as follows: (end-diastolic volume−end-systolic volume)/end-diastolic volume. LV end-diastolic volume was normalized to body surface area (mL/m²) to be expressed as the LV end-diastolic volume index (LVEDVI). LV mass was measured at end-diastole from 5 short-axis slices spanning the LV myocardium from the base to the apex, and was normalized to body surface area (g/m²) and expressed as the LV mass index (LVMI). To determine T1_native in each patient, a region of interest (ROI) was drawn to enclose the entire LV myocardium in each short-axis slice. Each ROI was carefully placed to avoid the papillary muscle, LV cavity, and epicardium. Myocardial segments affected by motion artifacts that could not be clearly identified were excluded from the analysis by consensus between 2 readers. T1_native in each patient was defined as an averaged T1 value of 5 short-axis slices of base-to-apex LV myocardium, which represents the global characteristics of LV myocardium. The normal value of T1_native in our institution was previously measured in 10 subjects with no clinical evidence of LV cardiomyopathy and normal CMR findings (1,283±34 ms).¹⁹

Statistical Analysis

The PQ interval, QRS duration, and serum hs-cTnT and NT-proBNP concentrations are expressed as the median with interquartile range (IQR). Other continuous variables are expressed as the mean with standard deviation. Categorial variables are shown as the number with proportion. Interobserver reproducibility of the PQ interval and QRS duration was determined by calculating the 95% confidence interval (CI) for the mean difference and 95% limits of agreement. The PQ interval and QRS duration were compared between baseline and at 1-year follow-up using the Wilcoxon signed-rank test. Correlations of the baseline PQ interval and QRS duration, ∆PQ, and ∆QRS with other clinical factors were assessed using Pearson’s correlation coefficient. Linear regression was used for the univariate and multivariate analyses, and the standardized β value was calculated to evaluate the associations between ECG parameters related to cardiac conduction disturbance and other clinical factors. Serum hs-cTnT and NT-proBNP concentrations were natural logarithmically transformed (Ln) to normalize their distribution (expressed as Ln (hs-cTnT) and Ln (NT-proBNP)). Clinical factors were compared among the 3 groups based on the presence of CRBBB or CLBBB using the Kruskal–Wallis test followed by the Steel–Dwass test for multiple comparisons. A two-sided P value <0.05 was considered statistically significant. All analyses were conducted using JMP Student Edition 18 statistical software (SAS Institute, Cary, NC, USA).

Results

Baseline Characteristics of the Participants

All patients had wild-type ATTR-CM confirmed by genetic testing. Among the 60 enrolled patients, 10 were excluded from the study on the basis of the exclusion criteria, and thus 50 patients were included in the final analysis (Figure 1). No patients died during the 1-year follow-up; 4 patients required permanent pacemaker implantation and 1 of them subsequently received CRT. Among the analyzed patients, 15 had atrial fibrillation or flutter on the baseline ECG, and thus 35 patients were available for evaluation of the PQ interval. At the 1-year follow-up, 2 patients showed atrial fibrillation; therefore, ∆PQ could be evaluated in 33 patients. The patients’ characteristics are shown in Table 1. Regarding the medication profile, which might affect the cardiac conduction system, 40% of patients were on β-blockers (BBs) and 16% were on calcium-channel blockers (CCBs) at baseline.

Figure 1.

Flowchart of the overall study design. ATTR-CM, transthyretin amyloid cardiomyopathy; CMR, cardiac magnetic resonance; TTR, transthyretin.

Table 1.

Baseline Characteristics of the Entire Cohort

Characteristics
Total, n	50
Age (years)	77.4±5.66
Male sex	44 (88.0)
BMI (kg/m2)	22.8±2.91
Hypertension	17 (34.0)
Dyslipidemia	11 (22.0)
Diabetes mellitus	11 (22.0)
Blood tests
Hemoglobin (g/dL)	13.6±1.71
eGFR (mL/min/1.73 m2)	51.5±16.9
hs-cTnT (ng/mL)	0.065 (0.043–0.093)
NT-proBNP (pg/mL)	1,949 (1,247–3,170)
Mayo stage
I	17 (34.0)
II	20 (40.0)
III	13 (26.0)
CMR imaging markers
LVEF (%)	55.4±12.2
LVEDV (mL)	125.2±27.8
LVEDVI (mL/m2)	77.1±18.1
LVM (g)	151±35.8
LVMI (g/m2)	92.6±22.1
T1native (ms)	1,427±53.0
Medications
β-blocker	20 (40.0)
Calcium-channel blocker	8 (16.0)

Serum hs-cTnT and NT-proBNP concentrations are expressed as the median (interquartile range). Other data are expressed as the mean±standard deviation or number (%). BMI, body mass index; CMR, cardiac magnetic resonance; eGFR, estimated glomerular filtration rate; hs-cTnT, high-sensitivity cardiac troponin T; LV, left ventricular; LVEDV, LV end-diastolic volume; LVEDVI, LV end-diastolic volume index; LVEF, LV ejection fraction; LVM, LV mass; LVMI, LVM index; NT-proBNP, N-terminal pro-B-type natriuretic peptide; T1_native, native myocardial T1 value.

PQ Interval and QRS Duration During the 1-Year Follow-up

We found excellent interobserver agreements for the baseline PQ interval (mean difference, −0.61 ms; 95% CI, −1.8 to 3.0 ms; 95% limits of agreement, −12.5 to 13.7 ms), baseline QRS duration (mean difference, −0.15 ms; 95% CI, −1.6 to 1.3 ms; 95% limits of agreement, −9.8 to 9.5 ms), ∆PQ (mean difference, −2.83 ms; 95% CI, −5.6 to −0.1 ms; 95% limits of agreement, −17.5 to 11.8 ms), and ∆QRS (mean difference, −0.11 ms; 95% CI, −1.9 to 1.7 ms; 95% limits of agreement, −11.8 to 11.6 ms). The median PQ interval (209 ms (IQR 168–235 ms) to 218 ms (IQR 188–244 ms)) and QRS duration (106 ms (IQR 92–130 ms) to 118 ms (IQR 97–141 ms)) significantly increased from baseline to 1 year (both P<0.001) (Figure 2). When the patients were divided into 2 groups based on the median T1_native (1,418.9 ms), both the high and low T1_native groups showed significant increases in the PQ interval and QRS duration over 1 year (Supplementary Figure).

Figure 2.

Changes in the PQ interval and QRS duration over the 1-year follow-up period.

Correlations Between the PQ interval/QRS Duration and Clinical Factors

Table 2 shows the correlations of the baseline PQ interval and ∆PQ with clinical factors, including age, sex, eGFR, Ln (hs-cTnT), Ln (NT-proBNP), LVEF, LVEDVI, LVMI, T1_native, and the use of BBs and CCBs. The baseline PQ interval positively correlated with Ln (hs-cTnT) (r=0.46, P=0.0059) and T1_native (r=0.37, P=0.027), but none of the baseline clinical factors was associated with ∆PQ. Table 3 shows the correlations of the baseline QRS duration and ∆QRS with clinical factors. The baseline QRS duration positively correlated with LVEDVI (r=0.28, P=0.047) and T1_native (r=0.29, P=0.038), whereas ∆QRS positively correlated with Ln (hs-cTnT) (r=0.33, P=0.018) and T1_native (r=0.33, P=0.020), and negatively correlated with the LVEF (r=−0.31, P=0.029). Figure 3 is scatter plots illustrating the baseline PQ interval, QRS interval, ∆PQ, and ∆QRS in relation to T1_native and hs-cTnT concentrations.

Table 2.

Correlations of the Baseline PQ Interval and ΔPQ With Clinical Factors

	Baseline PQ interval (n=35)		ΔPQ (n=33)
	Univariate β	P value	Univariate β	P value
Age (years)	0.34	0.042	0.11	0.55
Male sex	−0.096	0.58	−0.30	0.092
eGFR (mL/min/1.73 m2)	−0.31	0.073	0.10	0.57
Ln (hs-cTnT)	0.46	0.0059	0.00	1.00
Ln (NT-proBNP)	0.26	0.13	0.019	0.92
LVEF (%)	0.14	0.42	–0.10	0.57
LVEDVI (mL/m2)	−0.040	0.82	−0.063	0.73
LVMI (g/m2)	0.15	0.39	−0.16	0.38
T1native (ms)	0.37	0.027	0.058	0.75
β-blocker	−0.077	0.66	−0.022	0.90
Calcium-channel blocker	0.40	0.018	0.22	0.21

Abbreviations as in Table 1.

Table 3.

Correlations of the Baseline QRS Duration and ΔQRS With Clinical Factors

	Baseline QRS duration (n=50)		ΔQRS (n=50)
	Univariate β	P value	Univariate β	P value
Age (years)	−0.058	0.69	0.25	0.082
Male sex	−0.28	0.052	0.013	0.93
eGFR (mL/min/1.73 m2)	−0.043	0.77	−0.25	0.081
Ln (hs-cTnT)	0.26	0.073	0.33	0.018
Ln (NT-proBNP)	0.17	0.23	0.26	0.072
LVEF (%)	−0.22	0.13	−0.31	0.029
LVEDVI (mL/m2)	0.28	0.047	0.026	0.86
LVMI (g/m2)	0.20	0.16	0.25	0.079
T1native (ms)	0.29	0.038	0.33	0.020
β-blocker	0.14	0.34	−0.064	0.66
Calcium-channel blocker	−0.11	0.44	0.29	0.043

Abbreviations as in Table 1.

Figure 3.

Scatter plots of baseline PQ interval, QRS interval, ∆PQ, and ∆QRS against T1_native and high-sensitivity cardiac troponin T (hs-cTnT) concentrations.

Determinants of the Baseline PQ Interval, Baseline QRS Duration, and ∆QRS

Two models were used in the multivariate analysis to ascertain the clinical determinants of ECG parameters related to cardiac conduction disturbances. Model 1 was adjusted for age, sex, and the presence or absence of medications that might affect the cardiac conduction system (BBs and CCBs) as independent variables. Model 2 was developed by incorporating the variables that showed significant correlations in the univariate analysis into Model 1 (Ln (hs-cTnT) and T1_native for the baseline PQ interval; LVEDVI and T1_native for the baseline QRS duration; Ln (hs-cTnT), LVEF, and T1_native for ∆QRS). Table 4 shows the results of the multivariate analyses. In Model 1, T1_native was identified as the determinant of the baseline PQ interval (β=0.31, P=0.039), baseline QRS duration (β=0.30, P=0.038), and ∆QRS (β=0.29, P=0.035), whereas Ln (hs-cTnT) was identified as the determinant of ∆QRS (β=0.32, P=0.036). In Model 2, none of the variables were identified as determinants of ECG parameters. Figure 4A presents a representative patient showing an extended PQ interval and wide QRS duration at baseline, as well as progression of the wide QRS. Figure 4B presents a representative patient showing a normal PQ interval and mildly widened QRS duration at baseline, and almost no progression of the wide QRS.

Table 4.

Multivariate Analyses of the Determinants of the Baseline PQ Interval, Baseline QRS Duration, and ΔQRS

(A) For baseline PQ interval, n=35	Multivariate Model 1		Multivariate Model 2
	Standardized β	P value	Standardized β	P value
Ln (hs-cTnT)	0.35	0.056	0.22	0.29
T1native (ms)	0.31	0.039	0.22	0.19
(B) For baseline QRS duration, n=50	Multivariate Model 1		Multivariate Model 2
	Standardized β	P value	Standardized β	P value
LVEDVI (mL/m2)	0.27	0.057	0.19	0.18
T1native (ms)	0.30	0.038	0.23	0.12
(C) For ΔQRS, n=50	Multivariate Model 1		Multivariate Model 2
	Standardized β	P value	Standardized β	P value
Ln (hs-cTnT)	0.32	0.036	0.15	0.40
LVEF (%)	−0.27	0.056	−0.16	0.30
T1native (ms)	0.29	0.035	0.17	0.29

hs-cTnT, high-sensitivity troponin T; LVEDVI, left ventricular end-diastolic volume index; LVEF, left ventricular ejection fraction; T1_native, native myocardial T1 value. Multivariate Model 1 was adjusted for age, sex, use of a β-blocker or calcium-channel blocker (medications). (A) Multivariate Model 2 was adjusted for age, sex, medications, hs-cTnT concentrations, and T1_native. (B) Multivariate Model 2 was adjusted for age, sex, medications, LVEDVI, and T1_native. (C) Multivariate Model 2 was adjusted for age, sex, medications, hs-cTnT, LVEF, and T1_native.

Figure 4.

Representative cases. (A) Patient with elevated T1_native showing an extended PQ interval and wide QRS duration at baseline, as well as a progression of the wide QRS. In this patient, high hs-cTnT concentrations (0.081 ng/mL) were also observed. (B) Patient with modestly elevated T1_native showing normal PQ interval and mildly widened QRS duration at baseline, and almost no progression of the wide QRS. The patient showed a relatively low hs-cTnT concentration (0.032 ng/mL). hs-cTnT, high-sensitivity cardiac troponin T.

Bundle Branch Block and Clinical Factors

At baseline, 12 patients had CRBBB and 6 had CLBBB. At the 1-year follow-up, 5 new cases of CRBBB and 1 new case of CLBBB were observed. In the cohort of patients categorized into 3 groups based on the presence of CRBBB or CLBBB (i.e., pre-existing at baseline (n=18), new-onset during the 1-year follow-up (n=6), and absent throughout the 1-year follow-up (n=26)), the T1_native values were recorded as 1,461 ms (IQR 1,417–1,493 ms), 1,450 ms (IQR 1,380–1,500 ms), and 1,410 ms (IQR 1,380–1,426 ms), respectively. Concurrently, the hs-cTnT concentrations were 0.075 ng/mL (IQR 0.047–0.108 ng/mL), 0.087 ng/mL (IQR 0.048–0.130 ng/mL), and 0.053 ng/mL (IQR 0.040–0.081 ng/mL), respectively. T1_native differed significantly among the 3 groups by Kruskal–Wallis test (P=0.014), and the Steel–Dwass test showed a significant difference between patients with pre-existing CRBBB or CLBBB at baseline and those without CRBBB or CLBBB throughout the 1-year follow-up (P=0.009). No significant difference in hs-cTnT concentrations was observed among the 3 groups (P=0.083).

Discussion

In this study, we focused on the clinical factors related to cardiac conduction disturbance in patients with ATTR-CM undergoing tafamidis treatment. Our findings are as follows.

1. There were significant increases in the PQ interval and QRS duration from baseline to 1 year of follow-up following the initiation of tafamidis treatment.

2. The baseline PQ interval showed a positive correlation with T1_native and Ln (hs-cTnT), and T1_native remained a significant determinant of the baseline PQ interval after adjusting for medications that might affect the cardiac conduction system. ∆PQ did not correlate with the baseline clinical factors under investigation.

3. The baseline QRS duration positively correlated with T1_native and LVEDVI. ∆QRS positively correlated with T1_native and Ln (hs-cTnT), while showing a negative correlation with LVEF. After adjusting for medications that might affect the cardiac conduction system, T1_native remained a significant determinant of the baseline QRS duration, and T1_native and Ln (hs-cTnT) remained significant predictors of IVCD progression.

4. In the models that incorporated variables that showed significant correlations in the univariate analysis, no significant clinical determinants were identified for baseline ECG parameters, and there were no predictors of AVCD or IVCD progression.

To the best of our knowledge, this study is the first to document the baseline serum and imaging markers associated with cardiac conduction disturbances in patients with ATTR-CM undergoing tafamidis treatment. Our study provides novel clinical insights that may contribute to the management of cardiac conduction disturbances associated with ATTR-CM.

We found significant increases in the PQ interval and QRS duration over 1 year following the initiation of tafamidis. Tafamidis is thought to suppress amyloid deposition in the myocardium by inhibiting transthyretin tetramer dissociation and amyloid formation. However, our data suggest that tafamidis treatment does not effectively inhibit the progression of cardiac conduction disturbances. Our findings indicate the importance of ECG monitoring and assessment of symptomatic bradycardia and syncope in clinical practice, even after initiating tafamidis treatment.

In this study, T1_native was a significant determinant of the baseline PQ interval after adjusting for medications that might cause cardiac conduction disturbance. Therefore, AVCD in DMT-naïve patients appears to be associated with myocardial tissue degeneration, as reflected by elevated T1_native. We found no clinical factors associated with the progression of AVCD. Notably, patients with atrial fibrillation or flutter were excluded from the AVCD analyses. If these patients had considerable atrial myocardial damage, their exclusion could have affected the negative findings. Furthermore, the 1-year follow-up period for assessing the changes in the PQ interval may have been insufficient. Therefore, this evaluation should be extended over a longer duration.

In this study, T1_native positively correlated with the baseline QRS duration, and elevated T1_native and hs-cTnT concentrations were associated with progression of IVCD. T1_native is a marker that comprehensively reflects the intracellular and extracellular characteristics of the myocardium in CA. An elevation in the hs-cTnT concentration reflects myocardial degeneration or damage. The patients with pre-existing CRBBB or CLBBB had significantly higher T1_native than those without throughout the 1year followup, suggesting that the presence of CRBBB or CLBBB may reflect more advanced myocardial degeneration. After adjusting for medications that might cause cardiac conduction disturbance, in this study the LV parameters remained significant determinants of the baseline QRS duration and IVCD progression. Our data suggest that elevated T1_native is indicative of the severity of IVCD under DMT-naïve conditions, whereas progression of IVCD after the initiation of tafamidis treatment is subsequently accompanied by elevated T1_native and hs-cTnT concentrations. We reported previously that T1_native correlated with the LVEDVI, LVEF, and hs-cTnT concentrations in patients with ATTR-CM.²⁰ On the basis that LV parameters are mutually associated, significant relationships may have disappeared in the multivariate analyses models that included both T1_native and hs-cTnT.

In CA, an increased QRS duration has been reported as an independent prognostic factor, and duration >120 ms is associated with a higher risk of death.²¹ Kuyama et al. evaluated disease progression at 1 year using a 3-domain score consisting of clinical/functional, biomarker, and imaging and ECG domains, and demonstrated that higher scores were associated with an increased risk of all-cause death. In that scoring system, the imaging and ECG domain included new-onset conduction disturbance on ECG, defined as new-onset CRBBB, new-onset CLBBB, or IVCD with a QRS duration >120 ms.²² In patients who have risk factors associated with an increase in the QRS duration, identified as elevated T1_native and hs-cTnT concentrations in this study, comprehensive follow-up with meticulous attention to potential adverse prognoses is imperative. In addition, the usefulness of CRT in patients with ATTR-CM who have HF with a reduced EF remains controversial, and implantation is generally performed after comprehensively considering the benefits and disadvantages.^23,24 Identifying patients at risk of IVCD progression would be beneficial for appropriate ECG monitoring and for early and prolonged consideration of the indications for CRT.

Previous studies have identified clinical and ECG predictors of pacemaker implantation in ATTR-CM. Kawahara et al. reported that interventricular septal thickness ≥16.9 mm, PR interval ≥220 ms, and bifascicular block were identified as independent predictors of future pacing device implantation in wild-type ATTR-CM.²⁵ In our study, 4 patients required permanent pacemaker implantation during the 1-year follow-up. Our study results demonstrated that elevated T1_native and hs-cTnT were associated with an increase in the QRS duration in ATTR-CM patients undergoing tafamidis treatment. As device implantation was not a prespecified endpoint in our study, we could not evaluate whether these parameters predict the risk of pacemaker or CRT implantation. Further studies with longer follow-up and device implantation as a prespecified endpoint are needed to reveal the clinical factors, including CMR parameters and biomarkers, to predict the need for device therapy.

Study Limitations

This study was conducted at a single center and had a relatively small number of patients. In the multivariable models, the associations with T1_native and Ln (hscTnT), which were significant in the univariable analyses, were no longer statistically significant in Model 2. This outcome may be attributed to the limited size of our dataset, complicating the interpretation of these parameters as definitive, independent determinants. T1_native is dependent on several factors, including field strength, vendors, and sequences, requiring an institution-specific normal value to be determined for evaluation. T1_native differs among institutions, making the design of our study as a multicenter investigation difficult. Furthermore, establishing a control group without tafamidis treatment was ethically impossible, which caused difficulty in distinguishing whether the progression of cardiac conduction disturbances occurred despite tafamidis treatment or resulted from a natural course of the disease. Another significant limitation of our study is that the findings reveal only statistical associations without identifying any thresholds or cutoff values that could be directly applied to clinical decision-making. Moreover, as this study did not assess whether tafamidis responders, as defined by changes in T1_native or hs-cTnT concentrations, showed less worsening in ECG parameters, the association between treatment response and the progression of conduction disturbances remains to be elucidated in future research. The follow-up period was limited to only 1 year, and long-term follow-up of the progression of cardiac conduction disturbances may be warranted. In accordance with our study design, we were required to restrict the analysis to patients in whom quantitative ECG assessment was available at both baseline and 1-year follow-up, and we excluded 4 patients who required implantation of a permanent pacemaker or CRT during the 1-year follow-up period. These patients exhibited relatively elevated baseline T1_native and hs-cTnT concentrations of 1,471 ms (IQR 1,439–1,498 ms) and 0.107 ng/mL (IQR 0.067–0.151 ng/mL), respectively; however, our study did not yield evidence regarding parameters that could predict the risk of pacemaker or CRT implantation. Exclusion of these patients may have biased our cohort toward relatively stable patients, leading to underestimation of the true burden and clinical consequences of conduction system disturbances. Because of the procedural demands of CMR, our study participants were limited to patients in a relatively stable condition (NYHA class I–II), and patients with NYHA class III–IV were excluded. Thus, our findings may not apply to patients with more advanced HF. We did not analyze post-contrast CMR parameters, such as the extracellular volume fraction or late gadolinium enhancement burden. This limitation arose because a substantial number of patients were unable to undergo contrast-enhanced CMR because of chronic kidney disease or a history of allergy to contrast agents. Previous studies have shown a correlation between T1_native and the extracellular volume fraction in patients with ATTR-CM.^20,26 Furthermore, T1_native can be obtained without the administration of gadolinium contrast, which is a considerable advantage in the clinical setting. Regardless of the limitation of not analyzing post-contrast CMR parameters, this study provided promising insights into the clinical implications of non-contrast CMR parameters in the management of ATTR-CM. Finally, recent studies have demonstrated the efficacy of acoramidis, a next-generation oral transthyretin stabilizer, in patients with ATTR-CM.^27,28 Our findings warrant validation in patients undergoing DMT other than with tafamidis.

Conclusions

In patients with wild-type ATTR-CM, we observed considerable progression of AVCD and IVCD during a 1-year follow-up after the initiation of tafamidis treatment. After adjusting for medications that might affect the cardiac conduction system, baseline T1_native and hs-cTnT concentrations emerged as predictors of IVCD progression under tafamidis treatment, and T1_native was associated with current AVCD and IVCD under DMT-naïve conditions. Our study highlighted clinical factors that may aid in identifying patients with ATTR-CM who are at risk of current and future cardiac conduction disturbances.

Funding / Disclosures

None.

Author Contributions

K.H. designed the study and wrote the first draft. K.H. collected the data and performed the statistical analyses. T.K., D.O., Y.S., F.T. and Y.N. gave substantial contributions to improve the first draft. T.K. supervised the whole study. All authors reviewed the manuscript.

IRB Information

The study was approved by the Ethical Committee for Epidemiology of Hiroshima University (approval number: E2018-1364). Informed consent was obtained using an opt-out method.

Supplementary Files

Supplementary File 1

Supplementary Figure.

Data Availability

The datasets analyzed during the current study are available from the corresponding author on reasonable request.