Low voltage is classically reported as an electrocardiographic (ECG) finding in cardiac amyloidosis (CA). We evaluated electrocardiograms to determine the prevalence of low voltage and its association with outcomes. Electrocardiograms in 200 patients with CA were reviewed. The presence of low voltage was assessed by all limb leads ≤0.5 mV, all precordial leads ≤1.0 mV, or Sokolow index ≤1.5 mV, and the association with time to adverse outcomes, including hospitalization, orthotopic heart transplant, and death, was assessed by the Cox proportional hazards model. Low voltage prevalence was 60% when using Sokolow index ≤1.5 mV, 34% by QRS amplitude ≤0.5 mV in each limb lead, and 13% when ≤1.0 mV in each precordial lead with no differences in prevalence noted by the type of amyloid. Apart from atrial fibrillation and second-degree atrioventricular block being more common in wild type transthryretin cardiac amyloid (ATTRwt), the prevalence of ECG findings was similar among the 3 types of amyloid. Sokolow ≤1.5 mV (HR 1.690; 95% CI of 1.069 to 2.672; p = 0.0246) was independently associated with adverse outcomes. In conclusion, among the 3 main types of CA, the prevalence of low voltage is dependent on the method used for defining low voltage. Sokolow index ≤1.5 mV indicated the highest prevalence and was associated with adverse outcomes in CA. Our data suggest that low voltage is a relatively late finding in CA and may not be useful for early identification.
The prevalence of specific electrocardiographic (ECG) findings in patients with cardiac amyloidosis (CA) is important to determine the diagnostic utility of this readily performed test. ECG findings in patients with light chain (AL) have been relatively well reviewed; however, less is known about ECG findings in patients with tranthyretin amyloidosis (ATTR). Classically, low voltage is often cited as a clinical clue for CA, especially in the setting of increased wall thickness on echocardiography; however, the sensitivity of low voltage is still uncertain. There is much variation in the reported prevalence of low voltage in the literature, ranging from 46% to 70%. This discrepancy may in part be because of the variability in the degree of cardiac involvement in patients on presentation; however, the varying definitions of low voltage may also play a significant role. In the clinical setting, the presence of all limb leads ≤0.5 mV or all precordial leads ≤1.0 mV are the most commonly used methods to define low voltage because of the convenience of visually estimating the QRS amplitude. A recent study, however, showed that the prevalence of low voltage in a cohort of patients with AL amyloid is dependent on the method used to define this parameter. Therefore, comparisons of these 2 definitions to other easily calculated measurements of low voltage or even different ECG variables may help further determine the ECG parameters with the best diagnostic value. We sought to determine the prevalence of various ECG findings in a combined cohort of patients with AL, mutant transthyretin amyloid (ATTRmt), and wild type transthryretin cardiac amyloid (ATTRwt) and their prognostic significance. We hypothesized that for low voltage to be a useful early marker of cardiac amyloid, it would be present in most patients with various types of cardiac amyloid and as an early marker of disease would not have prognostic significance.
Methods
We identified 200 consecutive patients (AL, n = 110; ATTRmt, n = 45; ATTRwt, n = 45) with diagnosed CA seen at the Center for Advanced Cardiac Disease. Inclusion criteria for the diagnosis of CA were one of the following: (1) biopsy-proved CA; (2) in the absence of an endomyocardial biopsy, histological documentation of Congo red staining in at least 1 involved organ with echocardiographically defined evidence of amyloid cardiomyopathy (thickness of the left ventricular septum or posterior wall of >12 mm without another cause of LV hypertrophy along with biatrial enlargement, low tissue Doppler velocities, short deceleration time of the E wave and low-velocity A wave, thickened intra-atrial septum, reduced strain, or strain rate); or (3) documented amyloidogenic transthyretin mutation by DNA analysis and echocardiographically defined evidence of amyloid cardiomyopathy without evidence of a plasma cell dyscrasia.
Standard 12-lead electrocardiograms closest to the time of presentation were reviewed in addition to medical histories, baseline symptoms, and echocardiographic findings on presentation. Low voltage was defined as a QRS voltage amplitude ≤0.5 mV in all limb leads, ≤1 mV in all precordial leads, or the criteria defined by Carroll et al : the sum of the S wave in V 1 and R wave in V 5 or V 6 <1.5 mV. Left ventricular hypertrophy was defined by the Sokolow-Lyon criteria : S wave in V 1 and R wave in V 5 or V 6 ≥3.5 mV, and the Cornell criteria : R wave in lead aVL and S wave in V 3 >2.8 mV in men or >2.0 mV in women or R wave in aVL >1.2 mV. Patients with left bundle branch block were excluded from the analysis of left ventricular hypertrophy and pseudoinfarct. Total QRS was calculated using the method described by Roberts et al. Patients with paced rhythms were excluded from the analysis of total QRS voltage. A pseudoinfarct pattern was defined as pathologic Q waves or QS waves in any 2 consecutive leads in the absence of previous myocardial infarction and left bundle branch block. Poor R-wave progression was defined as R wave ≤3 mm by V 3. Prolonged QTc was considered to be >440 ms in men and >460 ms in women. JTc interval was calculated if QRS duration was >120 ms. LV mass was calculated using the method of Devereux et al and classified as increased when >110 g/m 2 in women and >130 g/m 2 in men as previously reported. To compare voltage to mass ratio, we calculated the sum of R wave in V 1 and the largest of S wave in V 5 or V 6 (Sokolow index) divided by the cross-sectional area of the LV wall as performed by Carroll et al. Normal values were calculated as 1.5 to 5.8 based on the normal parameters reported.
All analyses were performed using the SAS 9.2 (Cary, North Carolina). Categorical data are presented as frequencies and percentages. Comparison of categorical data was performed using the chi-square and Fisher’s exact tests as appropriate. We used the 1-way analysis of variance test to compare the continuous variables among AL, ATTRmt, and ATTRwt groups with a Bonferroni post hoc correction to control for multiple comparisons. Univariate analysis of outcomes was performed using the Cox proportional hazard model with a time for first event (death, hospitalization, or orthotopic heart transplant). Multivariate analysis was performed in a forward stepwise method using a model with age, gender, and type of amyloid as the baseline model. In each step, all potential variables were added individually, and the variable with the smallest p value was retained; this process was repeated until no additional variable had a p value <0.05.
Results
The baseline characteristics of all the patients are summarized in Table 1 . The 3 groups differed in age, gender, and ethnicity consistent with known demographics of the populations. Expectedly, the most common transthyretin mutation was Val122Ile (28 of 45, 62%), followed by Thr60Ala (7 of 45, 16%). The presence of low voltage based on limb leads ≤0.5 mV, precordial leads ≤1.0 mV, and pseudoinfarct pattern were infrequent in this cohort of cardiac amyloid and did not differ by amyloid type. Low voltage based on Sokolow criteria ≤1.5 mV (60%), and poor R-wave progression (64%), was more common and was similar among the 3 groups (see Table 2 , Figure 1 ).
| Parameter | AL | ATTRmt | ATTRwt |
|---|---|---|---|
| Age (years) | 60 ± 11 | 66 ± 11 | 78 ± 6 |
| Male | 74 (67%) | 35 (78%) | 41 (91%) |
| Female | 36 (33%) | 10 (22%) | 4 (9%) |
| White | 58 (53%) | 14 (31%) | 36 (80%) |
| Black | 10 (9%) | 19 (42%) | 1 (2%) |
| Other | 42 (38%) | 12 (27%) | 8 (18%) |
| Body mass index (kg/m 2 ) | 27.2 ± 5 | 26.3 ± 4 | 26.3 ± 6 |
| New York Heart Association Class III or IV on presentation | 67 (61%) | 22 (49%) | 16 (36%) |
| Systolic blood pressure (mm Hg) | 106 ± 17 | 111 ± 15 | 114 ± 15 |
| Diastolic blood pressure (mm Hg) | 67 ± 11 | 70 ± 9 | 68 ± 10 |
| Creatinine (mg/dL) | 1.8 ± 1.7 | 1.4 ± 0.5 | 1.5 ± 0.5 |
| Echocardiographic parameters | |||
| Pericardial effusion ∗ | 4% | 0% | 2% |
| Left ventricular end diastolic diameter (cm) | 4.2 ± 0.6 | 4.4 ± 0.6 | 4.4 ± 0.7 |
| Posterior wall thickness (cm) | 1.5 ± 0.3 | 1.7 ± 0.4 | 1.7 ± 0.4 |
| Ventricular septal thickness (cm) | 1.6 ± 0.4 | 1.7 ± 0.4 | 1.9 ± 0.4 |
| Left ventricular mass (g) | 266 ± 89 | 333 ± 127 | 355 ± 119 |
| Left ventricular mass index (g/m2) | 140 ± 48 | 172 ± 68 | 186 ± 59 |
| All Patients | AL | ATTRmt | ATTRwt | p Values | |
|---|---|---|---|---|---|
| Rhythm | |||||
| Sinus | 75% | 86% | 76% | 48% | <0.001 |
| Atrial fibrillation | 15% | 6% | 17% | 38% | <0.001 |
| Paced | 8% | 4% | 5% | 23% | 0.001 |
| Atrioventricular block | |||||
| PR interval (ms) | 186 ± 42 | 186 ± 45 | 219 ± 67 ∗ | 0.027 | |
| Primary | 30% | 26% | 33% | 45% | 0.220 |
| Secondary | 1% | 0% | 0% | 10% | 0.004 |
| Pseudoinfarct pattern | |||||
| Any | 22% | 25% | 18% | 18% | 0.507 |
| Anterior | 13% | 16% | 10% | 9% | 0.454 |
| Inferior | 9% | 7% | 13% | 12% | 0.638 |
| Lateral | 2% | 3% | 0% | 3% | 0.538 |
| Poor R wave progression | 64% | 62% | 60% | 71% | 0.610 |
| Low voltage | |||||
| Limb | 34% | 37% | 38% | 18% | 0.116 |
| Precordial | 13% | 16% | 10% | 6% | 0.279 |
| Sokolow <1.5 mV | 60% | 64% | 53% | 58% | 0.557 |
| Sokolow (mm) | 14 ± 9 | 15 ± 9 | 14 ± 8 | 0.760 | |
| Total QRS (mm) | 101 ± 43 | 112 ± 44 | 118 ± 39 | 0.808 | |
| Left ventricular hypertrophy | 9% | 8% | 8% | 13% | 0.715 |
| QTc/JTc (ms) | 432 ± 56 | 445 ± 58 | 431 ± 55 | 0.696 | |
| Prolonged QTc | 59% | 55% | 63% | 66% | 0.553 |
| Voltage to mass ratio | 1.0 ± 0.7 | 1.1 ± 0.9 | 0.9 ± 0.5 | 0.746 | |
| Abnormal voltage: mass | 77% | 78% | 73% | 80% | 0.771 |
Over a mean follow-up period of 616 days (range 6 to 4,660 days), 72 patients required hospitalization, 78 died, and 25 underwent orthotopic heart transplant. Overall, 125 (63%) had at least one of these events. On univariate analysis, predictors of adverse outcomes included AL amyloid, low voltage in the limb leads, Sokolow ≤1.5 mV, and abnormal voltage to mass ratio. Only Sokolow ≤1.5 mV was independently associated with outcomes on multivariate analysis in a model including age, gender, type of amyloid, and low limb lead voltage ( Table 3 ).
